ORNL-2106.txt

 

RIETTA ENERGY SYSTEMS LIBRARIES

IR

 

 

 

yvou wish someone else to s ihis docment,
i werie with document and the horary il

Viaiigs o sumii,

 

 
 

 

This report was prepcred as an occount of Government sponsored work, Neither the United ,S}utes_,

nor the Commission, ner a‘ny’pe‘rscm acting on behalf of the Commission: L -

A. Makes any warranty or represerntoflon, express or implied, with respect ‘to-the q{:;fu_rocy,
completeness, or usefulness of the informotion contained in this report, or ,thdi:}_h_e use of

any -~ information, apparatus, method, or proces_s' disclosed in this repott may not infringe
privately owned rights; or _ o
As_sum.es any liabilities with respect to the use of, or for damages. resulting from the use of =
any informcflon, apparatus, method, or process disclesed in this report.

As used in the above, '‘person acting on behalf of the _Co,rprpis_s'ion_'_'Vrin'clip'dé;f. any employee of

contractor of the Commission to the extent that such employee or contractor prepares, handles

or distributes, or provides access to, any information pursuant to his employment or confract

with the Commission,

 

 

 

 

 

 

 

 

 

 

 
 

 

Part 1

AIRCRAFT REACTOR ENGINEERING

S. J. Cromer

 
 

 
 
 
 

  

 

"

 
 

—— e b

 

 

+¥

1.1. AIRCRAFT REACTOR TEST DESIGN
A. P. Fraas

STATUS OF ART DESIGN

Design work on the Aircraft Reactor Test (ART)
reactor, heat-exchanger, pump, and pressure-shell
assembly is nearing completion. Layouts on all
the major subassemblies have been completed,
along with the major portion of the drawings of the
detailed parts. Drawings for the remaining parts
should be completed during the coming quarter.
The applied mechanics and stress analysis work
is accompanying the design, with rough first ap-
proximations being completed, usually, shortly
after completion of the layouts and with better,
second approximations following closely, in most
instances, upon completion of the detail drawings.
In cases in which component tests have been
deemed essential, the results are being anclyzed
and modifications in details made when essential.
Of course the analyses have not been completed
for many very complex situations, and many key
component tests have not yet been run. It is
believed that modifications that will be required
as the results of this work become available will
probably involve only relatively minor reworking
of partially fabricated parts. Such a calculated
risk- is necessary and inherent in design work in-
volving such exceptional extrapolahons of avail-
able technology.

The preliminary ‘layouts for the shield have been
modified to include provision for a substantial
amount of instrumentation and special equipment.
A one-half-scale model of the top portion of the
reactor {(commonly referred to as the “*north head''),
including the NaK manifolding, has been completed,
and models of the instrumentation components,

the lead shielding, and associated parts are well.

under way. Such models are used to investigate
assembly ‘and interference problems. - Preliminary
layouts have also been prepared for the arrange-
ment of the lube-oil, hydraulic-fluid, water, gas,

~ electrical, and instrumentation lines in the reactor

cell. A one-sixth-scale model of the entire ART
assembly is being kept closely abreast of this
work to ensure accessibility, freedom from inter-
ferences, etc. The fuel fill-and-drain tank design
has been completed, preliminary layouts for the
associated supports, shielding, plumbing and in-
strumentation have been prepared, and the con-
sequent assembly, accessibility, etc. problems

cre being studied in the one-sixth-scale model.
The detail design of the plumbing and equipment
installation outside the cell is well along and
should be largely completed during the coming
quarter;

APPLIED MECHANICS AND STRESS ANALYSIS
R. V. Meghreblian

North-Head Pressure Stresses

The stress analysis of the composite double-
deck structure of the north head, mentioned in the
previous report,! was completed. The analysis
was based on the pressure loads to which the
structure will be subjected during full-power oper-
ation. Since the actual design consists of two

- circular flat-plate decks joined by a complex
pattern of vertical baffles and walls arranged both

radially and circumferentially (Fig. 1.1.1), it was
not possible to carry out an analysis of this com-
posite structure which would yield an exact dis-
tribution of the elastic stresses. Moreover, this
structure is to be exposed to various operating
conditions at temperatures of 1200°F and above

_for about 1000 hr, and it is expected that thermal

distortions and creep will cause redistributions
of stresses which will differ markedly from any
predicted elastic stresses. It is not entirely
meaningful therefore to think in terms of an exact
stress distribution, and, for this reason, precise
analyses of this structure were not attempted.
So long as the proposed design is capable of
supporting the operating loads at relatively low
stress levels, the details of the exact distribution
are not important. From the viewpoint of creep
limitations, it would suffice to know the general
focation and magnitude of the highest stresses
in the system,

This information has beéen obtamed from a series
of calculations based on simplified geometric con-
figurations of the north-head structure, and these
results will eventually be checked by an experi-
mental stress analysis of a full-size aluminum
model, The calculations consisted of three parts:
a very elementary analysis in which the various

 

IR. V. Meghreblian, ANP Quar. Prog. Rep. March 10,
1956, ORNL-2061, p 22.

19

 
 

 

 

 

ANP PROJECT PROGRESS REPORT

 

of shells and plates subjected to a complex pattern

20

SRe—
jl PHOTOD 26471 .

 

 

 

 

 

 

 

 

 

 

A
X
-
Fig. 1.1.1. Plastic Model of ART North-Head Structure. _ :
deck and shell areas were treated as individual ,
plate segments with assumed edge conditions, : ORNL-LR- WG 14944
an analysis of two circular flat plates of annular _l_ - PRESSURE kL -
“shape joined by a continuous circumferential baffle UPPER DECK - N CONTOUR
(Fig. 1.1.2), and an analysis of a composite system / ! E S ™~
of plates and shells to represent the primary struc- . -
) . . - q N Y
tural members, which was based on the require- / N j H H ‘\\ \\
ment that the deformations of adjoining members / 4 N ' A D) N\
be matched along the various junction lines (welds). _ \
The first two analyses served to give a very crude CIRCULAR PLATES LOWER DECK
estimate of the stress levels involved. The pur- BAFFLE (CYLINDRICAL SHELL)
pose of the third analysis was to determine the *
stresses produced in the north head duve to com- ¢
patibility requirements of the various segments .

Fig. 1.1.2. Circular Plates with Baffle.
 

 

I
LT

of pressure loads. The idealized model used in
this calculation is shown in Fig. 1.1.3, and the
net pressure loads at various points are indicated,
This model represents approximately the cross
section indicated in Fig. 1.1.4., It includes some
‘of the longest spans which appear in the design,
and the results obtained from this model are there-
fore conservative,

The configuration of Fig. 1. l .3 was analyzed
by writing the deflection equation for each member
in terms of its load and edge conditions and solving
the resulting system of eleven simultaneous equa-
tions on the Oracle.
in terms of the moments and reactions at the gclnts'
and edges of the members, The stresses due to
these loads were then computed. The largest
stress, as indicated on Fig. 1.1.3, was found to
be 2100 psi. Since the highest temperature which
will occur in the north-head structure during full-
power operation will be approximately 1300°F,
this 2100-psi stress value is to be compared to

FUEL EXPANSION TANK WALL -

VERTICAL AXIS OF REACTOR

Sy

"'/4 in.

I | %
W////////////

50 psi (PRESSURE)

+— 8in.R

UPPER -DEC_K_ '

_‘_\fi%

wm

[

\w&\\\\\&m

SRR zOps. (PRESSURE)
FUEL SWIRL CHAMBER

 

—+— 9in. R

o uowsn DECK ' -

b _ 50ps.(pnessuas.) 1225ps: :
o~ R

 

 

890 psi

Fig. 1.1.3.

The results were obtained

7

PERIOD ENDING JUNE 10, 1956

the creep properties of Inconel in the fuel mixture
at about 1300°F, Creep tests have indicated that
‘the tensile stress required to produce rupture in
1000 hr at 1300°F is about 10,000 psi. The design
criterion for creep which has been selected for
the ART requires that the total deformation in
any member not exceed 0.2% strain in 1000 hr,
At 1300°F in the fuel mixture this corresponds to
a tensile stress of about 2000 psi.

The experimental program, designed as a check
on the calculations for the northehead structure,
is under way at the University of Tennessee. It
is believed that the combined results of the ana-
tytical studies and the model tests will reveal
any defects in the propesed design.

North-Head Thermal Stress

With the completion of the analysis for the
mechanical stresses in the north-head structure,
attention is now being directed to the determinae-
tion of the thermal stress disfl:ibutions. For this

-ECRER
ORNL-LR=DWG 14948

1500 p;\\\\ CONTOUR OF PRESSURE
| < \SH\ELL AND LINER
30psi (PRESSURE) N
_ ~ |
20 psi { PRESSURE}\ :
~ 2100 psi .~ 1600 psi \ ';} |
%// %’%//// /// ////////////////// -
o™ PN

—— e 4Tin R
‘/z in -

aaoo psi : 1o psi (PRESSURE) ',,

////%

3
Va in. 1270 psi—

Ideclized Configuration of North-Head Composite Structure (Sect%on A-A of Fig. 1.1.4).

21

 
 

 

 

 

ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
ORNL-LR-DWG 14946

 

Fig. 1.1.4. Cross Section Used in Composite-
Deck Analysis.

purpose, calculations have been undertaken to
obtain preliminary estimates of the energy-
deposition rates throughout this region (see Chap.
1.2, **Art Physics’). The locations in the north
head at which these rates might lead to relatively
large temperature rises are indicated in Fig. 1.1.5.
The calculations of the associated temperature
distributions and the thermal stresses are presently
under way, |f these calculations yield excessive
temperature gradients, it is planned to review the
initial energy-deposition estimates and, if neces-
sary, perform more precise analyses.

Sodium Expansion Tank Design

The design analysis of the sodium expansion
tank was completed, and the configuration which
was found to be acceptable from both the pressure
and thermal stress viewpoints is shown in Fig.
1.1.6. The proposed design consists of a vertical

wall of more or less elliptical shape joined to
an end cap of slightly cylindrical curvature. This

cap (or roof) is welded to the control-rod thimble,
“which passes through the center of the ellipse.
The stress analysis of this design was based on
an idealized model consisting of a short, elliptical
cylinder with a flat-plate cap subjected .to an

22

internal pressure of 30 psi. The stresses in the
cap were computed from the relations for ‘an
elliptical plate with various edge conditions. The
stresses at the joint and in the vertical walls
were computed from the relations for an equivalent
circular cylinder with a flat head. These calcu-
tations indicate thet the moximum stress (1000
psi) is due to the bending moment at the cylmder-
cap junction. :

During full-power - operation the tank wfll be
partially filled with sodium at 1270°F to an

assumed depth of 2 in., and the upper portion of

the side walls and the roof of the tank will be ex-
posed to direct gamma radiation from the sodium
(Fig. 1.1.7). The outer surfaces of the tank are
to be surrounded by insulating material, and, if
no cooling is provided for these surfaces, the
metal temperature in the roof will rise to 1420°F.
Since the side walls are to be welded to the
pressure shell, which will be at 1250°F, there
will be differential thermal growth between the
roof and the shell, which will give rise to a'thermal
stress of 80,000 psi (based on an elastic analysis)
at the roof-wall intersection. This stress is con-
sidered to be excessive; and therefore cooling is
to be provided for the roof. Sodium at a tempera-
ture of 1250°F will be taken from the pressure-
shell cooling circvit and fed into a system of
tubes welded to the bottom surface of the roof.
A total flow of 3 gpm will pass through this circuit
at a pressure drop of 15 psi. The coolant sodium
will leave the roof circuit at 1270°F and spill
into the expansion tank volume, With this supply
of coolant, the average roof temperature will be

reduced to 1280°F.. The temperature profiles in

the roof in the vicinity of a’cooling tube are shown
in Fig. 1.1,8. The maximum thermal stress due to
this temperature structure is 10,000 psi.

ALUMINUM NORTH-HEAD MOCKUP FLOW
STUDIES

E. R. Dytko?

R. E. MocPherson
D. Ward

A fullesize aluminum mockup of the fuél"syr;tem

R. Curry?

components in the north head of the ART has been

set up with external piping to complete the

 

20n assignment from Pratt & Whitney -A_ircr-clft.

M
 

PERIOD ENDING JUNE 10, 1956

scomet
ORNL—LR-DWG 14947

 

 

 

 

"

Fig.r 1.1.5. Locations in ‘N-orth Hécd at Which Enargy bepdsition Might Cause Large Tempe

Increases.

 

 

 

 

     

 

 

 

 

 

 

  

 

  

 

 

 

PRESSURE SHELL/ TANK
"‘/
LINER — ] :
@___ FUEL
EXPANSION !
TANK
UPPER DECK :
—\ 5:) -3 P |
T
!
> _
'II/II/I//II//IZ c M
1
¢

   
 

 

 

 

 

 

 

%
.

== Na EXPANSION

 

rature

- 23

 
 

 

|
|

 

ANP PROJECT PROGRESS REPORT

——— i —— —
———
-

   

-
— -
——
-——

   
     

- -

by

——
-

j

N\
TTTTOAN

1
A
\
\
~
——— e e =

 
     

Iy
—

/
/
/

#
/ K

i

!
/d.____.__._.(

COOLING TUBES

  

e —_———

—_—-
[ iy N
\ N
N N N
\ON NSNS
\\ \ “ q\
N Y

L
k
\
\

UNCLASSIFIED
ORNL-LR-DWG 14948

—— -
-
~

    

  
  
       
 

.(
S

p——————

-
—_—————

-
\
\
\c
1

I
\
\

1
\

\

\

—

o ——

 

RELAXATION SECTION

/ (SEE FIG. t.1.8)

T“T

LEVEL OF SODIUM IN TANK |

\1270°F Na
SIDE WALLS — ——

 

 

 

LINER
PRESSURE SHELL

EXPANSION
- LOOP TANK ROOF

COCLING TUBES

1270°F Na

 

 

 

 

\CONTROL ROD HOUSING 1250°F Na

 

SECTION A-A

Fig. 1.1.6. Sodium Expansion Tank Design.

circuits,® ond flow tests with water being pumped

by the two fuel pumps under simulated reactor flow

conditions are under woy.

Twin fuel pump operation has been demonsitrated
up to 3000 rpm, although 2400 rpm produced approxi-
mately rated conditions of head and flow, De-
gassing of the system with pump speeds of 2400
rpm required less than 30 sec with the water level
1 in. or more above the bottom of the surge chamber.
When the speeds of both pumps were lowered in-
unison from 2400 rpm, the first signs of ingassing

 

3p. R. Ward, ANP Quar, Prog. Rep. Dec. 10, 1955,
ORNL-2012, p 65, Fig. 2.29.

24

occurred at 1200 rpm. The amount of gas entrained
during operation with the pump speeds matched
was believed to have reached a maximum, not
exceeding 0.1 vol %, at 500 rpm, At speed levels
above 1200 rpm, ingassing could not be detected

with mismatching of up to 5%. Mismatching the

speeds by 20% produced ingassing estimated to
be 0.5 vol %. When the power to one of the hy-
draulically driven fuel pumps was cut off while
both pumps were running uniformly at 2400 rpm,
about 13 sec was required for the one pump to
stop. This one-pump-stopped condition created
ingassing of roughly 2 vol %. - "

 
 

 

 

 

 

tw

UNCLASSIFIED
ORNL~LR—-DWG 14949

pt———— INCONEL TANK RQOF ———=

SODIUM
LEVEL

  
   

 

 

 

 

 

 

 

 

06 N

 

 

ENERGY DEPOSITION (w/¢m3)

 

 

 

 

 

 

 

 

02 X
0 0.5 1.0 1.5 20
DISTANCE FROM BOTTOM SURFACE OF ROOF {cm)

 

Fig. 1.1.7. Energy fieaosition in .Sadi.am Ex;
pansion Tank Roof Due to Gamma Rediation.

Surging and insfabillity' were noted in pump

speeds, flow rates, pumpmg power, and mlef and
discharge pressures. It is believed that both the .
centrifuge hardware and the ‘unfavorable pump ‘inlet
pipe conflgurahons ‘were - conmbutmg to the ‘ob-"
served mstabihty. The ‘addition of stralghtemng,
vanes to’ the pump inlet. reglons rnaternally reduced
the surging. - ~This, ‘coupled with some modlflca-':.L_i;:‘surface immeduately below the guide vanes and

tions_to the xenon-removal system," _appears to again on the shell surface immediately below

____,_:"‘j__have reduced the* pressure fluctuaflons ‘to an-ace . the equafor, ‘however, the iimprovement was only
jceptable level fhat |s, about 0 5 psl m fhe full-_j_.-

S *’scale reactor; ” :
e Water bypassecl mta fhe surge chamber for sumu-,_- :
o lated Xenon removal was very turbuient, and splash-

wetting~ of “all- _surfaces ‘and entrainment of fine

“bubbles ‘resulted.” Methods ‘being . cansndered for__,, S

reducing the extreme turbulence in this region
include the addition of baffles, o reduction in

PERIOD ENDING JUNE 10, 1956

bypass flow rates, and o redesign of the passages
from the pumps to the surge chamber.

CORE FLOW STUDIES

W. T. Furgerson W. J. Stelzman
D. B. Trauger

Changes in design of both the center volute of
the axial-flow type of header and the island expan-
sion bellows located within the header of the
proposed ART core resuited in an unsatisfactory
core flow pattern being generated by the previously
satisfactory inlet guide vane, designated GS-2,

-and the conical baffle plate, designated GS-2-

P3 (ref. 4). Under the revised design, this par-
ticular guide vane and baffle plate combination
generated flow reversal ot the island surface in
the region of the equator. Systematic relocation
of the conical baffle plate only ond analysis of
the resulting core flow pattern yielded baffle plate
GS-2.P10, which, in combination with guide vone
GS-2, again generated a flow pattern containing
no flow reversal along either the outer core shell
or the islond surfaces. Brief periods of miner
flow reversals did occur at the equator; however,

‘these occurred in midstream ond seemed to be

caused by the turbulent condition of the fluid mass
in this region. In general, the flow generated in
the upper half of the core by this combination was
extremely unstable, but it exhibited excellent sur-

- face scrubbing, with very good transfer of fluid

from the walls, and excellent fluid niixing. Below

- the ‘equator, some improvements were noted in
~ the fluid flow properties; however, the streamlines
"again tended to hug the inner and outer surfaces
" as they approached the core outlet, - Attempts had
- previously been made with the criginal axial-flow
- “header - to improve the flow in the lower half of

the core by means of turbulators on the island

- minor, : This latest design is being evaluated in
~_the half-scale ART volume-heat-source apparatus
_{see Chap. 4. 'I,'v,"Heat Transfer and Phys:cal
- Propemes") - ' S o |

 

. ‘GUb-'Whumaa,‘ w. J. Ste'liz‘l_'han,"an‘iiiw.'T.. Fdrgerson,
AI;P Quar, Prog. Rep. March 10, 1955, 0RNL-206'|,
p 24.

25

 
 

 

 

ANP PROJECT PROGRESS REPORT

DEVELOPED COOLING TUBE
1250°F

UNCLASSIFIED
ORNL—LR—-DWG {4950

UNDERSURFACE OF ROOF

_~ZERQ HEAT TRANSFER SURFACE

 

1250°F
1254
1260
1265
1270
1275
SYMMETRY SURFACE —3| -
1280

1283

1285

 

4287

 

- SYMMETRY SURFACE

1299°F
1298
1297
1296
1295
1293

 

 

UPPER SURFACE

/ 1288 1290 1292

INSULATED SURFACE

Fig. 1.1.8. Temperature Profile in Sodium Expansien Tank Roof.

PERFORMANCE REQUIREMENTS FOR PUMPS
IN THE NaK SYSTEMS

M. M. Yarosh

A study was completed to determine the speed
and power range requirements for the main and
auxiliary NaK pump motors. Because of unequal
system lengths, the pressure drops in each of the
four main NaK systems for an equal NaK flow are
different, and thus o different pump speed is re-

 quired for each system. In addition to the vari-

ations between systems, the individual system
resistances will vary as a function of operating
time and operating conditions. These variations
will be attributable to the. mass-transfer buildup
anticipated - in the colder sections of the NaK
systems, principally in the radiator tubes.

Pump speed requirements were established from
the pump performance curves., The NaK system
curves (head vs flow) showing expected ranges
of system resistance as a function of flow were
plotted on the pump performance curves, and the
operating speeds were established at design flow
to be those speeds falling between the minimum

26

and maximum system resistance curves. Thus,
a range of operating speeds was established for
each individual NaK system. Power requirements
were then computed for the corresponding speed,
head, and flow requirements.

In order to reduce warmup time for the NaK
systems, it will be desirable to operate the NaK
pumps ot, or near, full speed during the warmup
period. Stress considerations, however, will pro-
hibit operation at full NaK pump speed for extended
periods of time when there is no fuel in the reactor.
Therefore an intermediate pump speed was estab-
lished at which reduced warfnup periods can be
attained ot reduced stress. The operating speeds
established and corresponding power requirements
for the main and auxiliary NaK pumps are given
in Table 1.1.1.

CONTROL~-ROD COOLING SYSTEM
J. Foster

The ART control rod is designed to move verti-
cally in a well containing static sodium, this
sodium being deep enough to cover the control

W

 
 

e g ——_

 

 

17

TABLE 1.1.1. NaK PUMP SPEEDS AND
HORSEPOWER REQUIREMENTS

 

 

Main System Auxiliary System
Pump Power Pump Power
Speed Required Speed Required
{(rpm) (hp) (rpm) (hp)
1900 27 1900 7
2300 . 42 2300 14
2650 61 2800 26
2800 72 2950 N
3000 87 3100 37
3200 102 3250 41
3400 118 3400 46
3500 127 3550 55

 

rod in its fully withdrawn position. This places
the level of the sodium-free surface in the well
just a few inches above the top of the reactor
north head. The well extends up about 5 ft above
this fevel so as to place the control-rod drive
mechanism outside the reactor shield. The sodium
at and near the free surface must be cooled to
below 500°F to minimize the vapor pressure and
hence the diffusion and deposition of sedium vapor
on the components of the control-rod drive mecha-
nism, where such deposits might create operating
difficulties such as shorting of electrical circvits.
Tests have shown that sodium vapor evolution and
deposition are negligible at 500°F.

The lower portions of the sodium in the well

will be exposed to temperatures of about 1200°F,
and therefore the cooling system includes con-

vection baffles to still the upper few inches of

the sodium and @ water jacket around this baffled

sodium zone. As a precaution to ensure against
any possibility of water entering the sodium

- chamber, the jacket will be a completely water- -
R 'hght assembly, -An Inconel sleeve will be shrunk
- over the outside of the controlerod well pipe to

form a double wall. Water will be circulated. at

220 to 240°F (sodium melts at 208°F) through the

jacket and will serve to remove heat or supply

PERIOD ENDING JUNE 10, 1956

heat as required by the condition of the reactor
system. No flow or pressure controls will be
provided other than an orifice in the water line,
designed to give a flow of 1 gpm,

The water-jacketed and baffled sodium zone will
be separated from the hot sodium in the lower well
by a solid Inconel plug inserted in the sodium as
a heat dam to keep the thermal gradient along the
Inconel well to .a reasonable value from the thermal -
stress standpoint. This Inconel plug is a 4-in.-
high cylinder with a central hole drilled along
the cylindrical axis, through which the 9-|n.-d|a
control-rod drive is free to move and posmon the
rod as required.

The effluent hot water from the jacket will pass
through an economizer, in which it will heat the
entering water stream. This will reduce the water
heating load and cool the effluent stream to prevent
flashing in the drain.

SODIUM SYSTEM STUDIES
R. |, Gray

Recent tests showed that the flow resistance
in the annuli around the core in which sodium will
be circulated will be somewhat smaller than ex-
pected with the spacers in place. This will effect
a slightly lower over-all pressure drop and more
nearly balanced flow between the cooling holes
and the core annuli. Pressure drop calculations
indicated the need for increasing the thickness
of the control-rod cooling annulus, in which sodium
will circulate, from 0,080 in. to about 0.125 in.

Stress calculations indicate the need for cooling
the top of the sodium expansion tank and for the

~ addition of a flexible bellows to the island sodium
- inlet line (see previous section of this chdpter 'on
 "“Applied Mechanics and Stress Analysis’’).

auxiliary sedium expansion tank of upprommately
0.6 ft has been added to the system so that in
the event of a major reactor shutdown sodium can
be added as the sodium temperature is lowered

from 1200°F to 300°F to avoid a loss of prime
in the sodium pumps (which would otherwuse occur

at about 800°F).

27

 
 

 

ANP PROJECT PROGRESS REPORT

1.2, ART PHYSICS
A. M. Perry

RADIATlo_fl’ HEATING ON THE ART
'EQUATORIAL PLANE IN THE VICINITY
OF THE FUEL-TO-NaK HEAT EXCHANGER

H. W. Bertini

The results of calculations - of the radiation
heating on the ART equatorial plane in the outer
3 cm of the beryllium reflector and in the Inconel
and the boron-containing shells on both sides of
the fuel-to-NaK heat exchanger are presented in
Figs. 1.2.1 end 1.2.2. The total gamma-ray
heating in- each region is given in Fig. 1.2.1,
as well as the heating from the sources which
are the main contributors to the total in each

shell. The encircled numbers on Fig. 1.2.1 refer
to the sources described in Table 1.2.1.

The data on heating in the copper-boron layer
by alpha particles from the B19(n,a)li? reaction
are plotted in Fig. 1.2.2. The heating goes to
infinity at the face of the layer closest to the
core because the heating at various points is
governed by an E function, |

o daA
Eyw) = [ e —,
1

where A is the mean free path, The integral under

the curve will be finite.

TABLE 1.2.1. SOURCES OF RADIATION HEATING CONSIDERED IN CALCULATING
THE RESULTS PRESENTED IN FIG. 1.2.1

 

Source

 

 

No. Source Source Strength
1* Prompt gamma rays in the fuel region of the core of the 28.3 W/C,"‘a
reactor _
2 Decay gamma rays in the fuel region of the core of the 6.84 w/cm?
reactor
3 Gamma rays from inelastic scattering of neutrons in the 10.1 w/cm®
fuel region of the core
4 Capture gamma rays in the outer core shell 41.4 w/cm?
Capture gamma rays in the reflector (average) "~ 0.5 w/emS
Capture gamme rays in first Incone! shell outside of 22.5 w/cm>
beryllium reflector
7 _ Beron capture gamma rays in copper-boron layer 1.8 w/em?
8 Alpha particles from the Bw(n,a)Li7 reaction in the 42 V_l_/cm3
copper-boron layer (average) 7
’ 9 Decay gamma radiation from the fuel in the heat exchanger 2.3 w/cm®
10 Gaemma rays from inelastic scattering of neutrons in first 0.7 w/cm3
9 cm of reflector (average) '
1N Capture gamma rays from delayed neutrons in the heat ex- <0.1 w/cm3_
changer and Inconel shells (including the pressure shell)
12 Capture gamma rays in the copper of the copper-boron layer 0.5 w/cm?
13 Gamma rays from inelastic scattering in both core shells ~4 w/cm?
14 Capture gamma rays in the island core shell 41.4 w/cm?

 

*In Fig. 1.2.1 the data for heating from sources 1, 2, 3 are combined and labeled a.

28
T4

CALCULATED HEATING (w/cm3)

 

w

’ ‘ (

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SRR
ORNL—-LR-DWG 14916
|~ STAINLESS STEEL
\ TOTAL HEATING.,
D= HEATING FROM SOURCE x. (SEE
FUEL 6AP —— . TABLE 1.2.4).
g , =—{(@)—— HEATING FROM SOURCES 1, 2, AND
S = o 2 3 (SEE TABLE 1.2.1).
= - = 2 W
5 S 2 W ALL DIMENSIONS IN cm.
w - L w Q I @
B - . 2 = 51| @ x
x | w 3 & a1 15 S
.5 o o z :
3. 7 E : 5 2 5 le— HEL UM
& | & S @ onp | l—s0DIUM
8 o Y @« |
L@ e S |a Ll 0157 I~ |
S = 2 N
‘ n L 0.457 0.318
g lwlal /1 I ™[N o b
o S | o w o o
=2
@ : 0 x &
‘_ 2
|
=
Wl
T —+
\ % @
ul I
% w
@ ¥
\, W o
N a Q
o a
. B =
STAINLESS STEEL Ysl \ g
=
& e,
rS o
o e ey
. ' . S © G \’9“ \
' | & & roFo— ~o——|
#\ ——— ‘_@_

Fi‘g.‘l 1.2.1. Gamma-Ray Heating in the Vicinity of the Fuel-to-NaK Heat Exchanger on the Equatorial Plane of the ART.

9§61 ‘01 INNF ONIGNI aol¥3d

 

 
 

 

 

ANP PROJECT PROGRESS REPORT

SEORETY
ORNL=LR~-DWG 14917

103

3
"N

o

CALCULATED HEATING {w/cm)
o

 

0 0.04 0.08 a2 016 0.20 0.24
THICKNESS OF LAYER {cm}

Fig. 1.2.2. Heating in Copper-Boron Layer By
Alpha Particles from the B'%(n,a)Li7 Reaction.

The heating from sources 10 to 14 was neg-
lected.  Their combined contributions to the
heating in the region being considered was
estimated to be about 5% of the total heating.

The fuel region of the core of the reactor was
assumed to be o spherical shell 5.125 in. thick
with an outside radius of 10.5 in.! This region
was divided into 13 spherical shells of thicknesses

varying from 0.27 cm to 2 em, The source strength -

(in watts/cm3) of the prompt gamma rays in each
shell was assumed to be proportional to the
average fission power in each shell, which was
calculated (from ref. 2) at the equatorial plane of

 

w. L. Scott, Jr., Dimensional Data for ART, ORNL
CF-56°'I =186 {March 13, 1956).

25, M. Perry, Fission Power Distribution in the ART,
ORNL CF+56+1-172 {Jan. 25, 1956).

30

the reactor. The source strength of the decay
gamma rays was assumed to be the same for each
fuel shell,?

The average source strength of gamma rays
resulting from inelastic neutron scattering in the
fuel was calculated4 by using the output of
multigroup calculations® performed by the Curtiss-
Wright Corp. on ART-type reactors with spherical
symmetry. The inelastic cross sections used
for the fuel were those reported in ref, 3. This
calculation had been performed before all the
data in the latter reference had been accumulated,
so it was assumed that, for each inelastic
collision, one-half the average neutron energy in
each energy group was given off as 1-Mev gamma
radiation, Caleulations made by using the more

recent data indicate that the source strength

used was too high by about 50%. The total heating
values given in Fig. 1.2.1 may therefore be about
5% too high. This error is partially compensated
for, however, by the neglect of sources 10 through
14.

The prompt and decay gamma-ray spectra’
were divided into four energy groups with average
energies for each group of 0.5, 1, 2, and 4 Mev,
The last four groups listed in ref, 3 were combined
into one group with an average energy of 4 Mev.
The heating at the vorious places described in
Fig. 1.2.1 was calculated by summing the contri-
butions from each energy group from every fuel
shell.

It was assumed for the calculations that each
fuel shell was replaced by an infinitely thin
spherical-shell source embedded in an infinite
homogeneous medium so that the standard transfor-
mation from a spherical-shell source to two
infinite-plane sources would apply, that is, so
that the heating at R, b(R), would be given by

r
(1) A(R) =-E- [H(R ~ 1) - HR + 1) ] ’

 

3H. W. Bertini et al.,, Basic Gamma-Ray Data for
ART Heat Deposition Calculations, ORNL-2113
{in press). ‘

4Calculations performed by R. B. Stevenson, Pratt &
Whitney Aircraft, private communication to H. W.
Bertml.

5s, Strauch, Curtiss-Wright Corp., prlvate communi-
cation to H. W. Bertini.

H. Reese, Jr.,, S. Strauch, and J. Michalcozo,

Geometry Study for an ANP Circulating Fuel Reactor.
WAD-1901 (Sept. 1, 1954).

TH, w. Bertini, C. M. Copenhaver, and R, B,
Stevenson, ANP Quar, Prog, Rep. March 10, 1956.
ORNL-2061, p 35.

9
 

 

 

[

where _

r = radius of source (taken as the average

radius of each fuel shell),

R = distance from center of fuel shell to
field point,
heating at a field point due to an infinite
plane source of monocenergetic gamma
rays a distance x away from the field
point,
The second ferm of Eq. 1 was dropped for these
calculations because of the large radii involved.

The assumptions made were certainly not
consistent with the geometry, inasmuch as the
region between the source shells and the field
point ‘is not everywhere homogeneous. The justi-
fication for this approach was that it appeared to
be as good as could be done without going to a
much more detailed numerical integration aver all
source points for each energy group to calculate
the heating at one field point. ,

In deriving H, it was assumed that the attenu-
ating medium between the plane source and field
point consisted of infinite slabs of materials
whose thicknesses were determined by their
thicknesses at the equatorial plane of the ART.!

The buildup factor used was of the form

H(x)

' | al pit, _
(2) Aje ? -1) +1,
where ,
A,a = parameters of the equation,
p; = linear total gammo-ray absorption cross

section {cm=1) for material i,

t, = thickness of material { (cm)

1

Wl

Under fhese condlhons,
s

| - +(]—A)E(Ep.itt)}

wheré

oy
' m : m

“source strength (w/cmz), T

. t . “ - !
o
"

section (cm=1) for the field pomt
fhlckness of the ith siab (cm)

LY
m 

The s term was determmed for each fuel sheII -

and each energy group by multiplying the source
strength for each group (in w/cm3) by the thickness

heuf;ng (w/cm3) ot the fleld po:nt '.

PERIOD ENDING JUNE 10, 1956

of the shell. The buildup factor parameters, A and
a, were taken to be those for beryllium. The
i's, A's, and o's were evaluated at each energy
group. :

The spectrum of prompt gamma rays. reported
in ref. 7, 8.8 e=10VE photons/Mev-fission, is
different from that reported in ref, 3, 9.61 e~1-01E

‘photons /Mev+fission, The former value, which

neglects the variation of (¢_/0 ) with energy,
was used in these. calculchons be?ore the cor-
rection was pointed out. Use of the latter spectrum
would change the results repor_ted here by less
than 2%. .

The heahng by the capture gamma rays in the
outer core shell was calculated with the use of
the same assumptions as those used for the calcu-

lations of the heating by the fuel-region sources,

that is, a sphere-to-plane transformation was
made and slab geometry was used for the inter-
vening mediums betweeh the plane source and
field point. The absorption rate in the outer core
shell was calculated from the output date of
multigroup calculations.’ :

The spectrum of capture gamma rays in Inconel
was divided into seven energy groups in ref. 3,
but, for this calculation, the first three groups
were combined into one group, and the average
energy of this combined group was taken to be
2 Mev. The fourth ond fifth groups in ref. 3 were
taken as the second and third energy groups for
this calculation, and the average energies were
tcken to be 4 and 6 Mev, respectively. The sixth
and seventh energy groups in ref, 3 were combined
into one group with an average energy of 8 Mev.
Thus  a total of four energy groups was used in
this calculation. The buildup factor used was that

for beryllium at ecch energy group. The outer
- core shell has an msnde radius of 26 ¢m and ¢
' thlckness of 0.381 cm. B

~The capture gamma rays in the island core shell

- were neglected because of the shielding properties
. of the fuel. The heating by the _capture gamma
“rays in the beryll:um reflector was calculated by
- using the sphere-to-plane tranisformation and the

“7 " other assumptlons given above. ‘The reflector_
linear gomma-ray energy absorphon cross

region was divided into five spherlcal shells, that

~ is, the same shells as those used by the Curtiss-
 Wright Corp._m their mulhgroup culculanons of

reactor No. 675 6 The capture gamma-ray source

 

8Calcu|cflons performed by C. M, Copenhaver,
ORNL, private communication to H. Bertini.

31

 
 

 

"ANP PROJECT PROGRESS REPORT

strengths in each region were calculated® from the
output - of the multigroup calculations, The
spectrum of the capture gamma rays in beryllium
is divided into two energy groups in ref, 3, and
these groups were combined into one group with

‘an average energy of ‘6 Mev, for this calculation,

The bunldup factor for - berylllum at 6 Mev was
used,

In calculahng the gamma-ray sources in the
first Inconel shell around the beryllium reflector,

- it- was assumed that 32% of all neutrons born in

the core ‘escape from the reflector as thermal
neutrons.® It was further assumed that these

‘neutrons esc¢ape with an isotropic angulor distri-

bution from the surface of the reflector. A source
strength, S, in thermal neutrons/cm2.sec escaping
from the surface of the reflector, was calculated
by using a reflector radius of 55.04 cm (ref. 1).
Because of the large radii of curvature in this

 

 

 

(6)h-% ]f{ [1-@2! ]

2”1 a =2

neutron absorption cross sectlon is 0. 18 cm=1,
Then, for the Inconel,

(5) absorption rate in Inconel

S[1 -~ P_(Na)1P, (lnconel)

The capture gamma-ray source strength per unit
volume was then calculated by using the di-
mensions given in ref. 1, the energy per capture
given in ref. 3, and the assumptions given above,
The macroscopic neutron absorption cross sections
at average velocity at 700°C were cclculafed from

values given in BNL-325.°

It was assumed that the gamma-ray source was
constant in the Inconel. Also, because of the
large radii involved, slab geometry was assumed
for the source and for the mediums between the
source and field points. o

For @ slab source and the buildup factor inén
in Eq. 2, the heatmg for monoenergetlc gamma
rays is given by

- Ey [““a) L ‘i#i}}
i=1

+(|_A)[E2<§2 t,-#,-) ~ E, (E} t#,)} '

 

 

region of the reactor, the neutron absorption rates
were calculated on the basis of slab geometry.

Between the Inconel shell and the reflector
there is a %-in. layer of sodium coolant.! If it is
assumed that a neutron leaving the surface source
will be absorbed only on a first-flight absorption
collision, the probability of absorption in the
s_odium,_P'a(Na), is derived to be

(49 PNa) =1 - E,[Z (Na):t],

where

oy =" e
| Ll A2
t = thickness of sodium layer (in cm),
2 J(Na) = macroscopic  thermal-neutron  ab-
sorption cross section for sodium
= 5.6 x 10~3cm=1,

It was assumed that the angular distribution of
the neutrons reaching the Inconel was still
isotropic, and a similar expression was obtained
for the Inconel, for which the macroscopic thermal-

32

-

where
¢ = source strength (w/cm3),
p; = linear total gamma-ray cross section
(cm=1) for the ith slab,
t, = thickness of ith slab (cm); i = 1 designates

the source region.
The spectrum of capture gamma rays from Inconel
was divided into seven energy groups?® with
average energies for each group of 0.5, 1, 2, 4,

6, 8, and 10 Mev. In calculating the heating in

the shells odjocent to the Inconel source, the
buildup factor for inconel was uvsed. Test calcu-

~ lations have shown 10 that the heating is relatively

insensitive to the type of buildup factor used for
materials which have nearly the same equivalent
Z, The heating in the beryllium reflector was
calculated by using the buildup factor for be-
ryllium. For the shells on the pressure shell side
of the heat exchanger, the bui!dup fccfor for th‘e

 

. J. Hughes and J. A, Harvey, Neutron Cross
Sect:ons, BNL-325 (July 1, 1955).
H. W. Bertini, C. M. Copenhaver, and R, B.
Stevenson, ANP Quar. Prog. Rep. March 10, 1956,
ORNL-2061, p 36.

)

o

©
 

heat exchanger was used., The heating at all
field points was found by summing over the
contributions of every energy group.

The self-heating of this Inconel shell for each
energy group was calculated by using the ex-
pression

(7) H ="€'%i {2 - (Ez(l-’t) + E2 [ l’-(to - t)])} ’

where

H = heating at ¢ (w/cm3),

¢ = source strength of the shell (w/cm3),

B, = linear energy absorption cross section

~ (ecm=1) of the Inconel shell,
p = linear total cross section (cm"‘) of the
inconel shell,
t, = thickness of the Inconel shell (ecm),
t = distance from the surface of the Inconel

shell (cm).
No buildup factor was used for this self-heating
calculation. _ S

Sample calculations were made in order to
compare the hedting from this shell when it was
assumed to be a plane source with the heating
when the shell was assumed to be a slab source;
the results indicated a difference of 30% between
the two values of the heating at o distance of
2.5 mean free paths from the source.1® Thus the
approximation of a plane source was not used in
this case.

The source strength of the gamma rays resulflng
from neutron captures in the boron of the copper-
boron layer was calculated by assuming that oll
the neutrons. leaving the reflector and escaplng
from the  sodium and
above are obsorbed at the fronf surface of \‘he
copper-boron layer, -
that only 90% of the ‘neutrons smkmg the" layer
are cbsorbed there, and they also indicate. ‘that
the distribution of. 1he sources  of gamma rays
would ‘be the ‘same as: the dlstrlbu’non of . hecmng

by alpha pcrhcles, as tllusfrated in- Flg. 1.2.2. -

- The source strength of the gamma rays-used for -

lthls caiculatlon -was ‘therefore ™ about ]0% ‘too

o rh_lgrh It was ‘assumed that there was ‘a0, 48-Mev

‘gamma
sorphons.‘? e

 

 

 

Vg, Ajzenberg and T. Lauritsen, Rev. Mod. Phys.
24, 351(1952).

“Actual calcuhnons :ndlccte-,

ray. ossocmted wnh 93% of the ab--.

PERIOD ENDING JUNE 10, 1956

The heating in the shells for this source was
calculated by using Eq. 3. The buildup-factor
parameters and the gammaeray cross sections
were evaluated .at 0.5 Mev.3 The buildup factors
used were the same as those used in the previous
calculation of the sources in the Inconel shell
outside the reflector.

The hecting in the copper-boron layer by alpha
particles from the B10(n,a)Li? reaction was
calculated by using techniques similar to those
used in estimating the neutron captures in Inconel.
It was assumed that all neutrons escaping from
the reflector, sodium, and Inconel were incident

~on the surface of the copper-boron layer with an

isotropic angular distribution. By assuming slab
geometry and by assuming that the neutrons make
only first-flight absorption collisions, the proba-
bility of absorption at ¢ per unit thickness in the
ith constituent of the layer is given by the ex-
pression ‘

i T
(8) P,=320 E, (X, 1),
where
Ei‘ = macroscopic thermal-neutron absorption

cross section of the ith constituent,

2: = total macroscopic thermal-neutron ab-
sorption cross section of the copper-boron
layer,

t = distance from the front face of the slab
(cm). omt ;

In this case the copper-boron layer was assumed
to consist only of copper and B19, and the atomic
density of each was calculated for a matrix con-

_;s:shng of 16 vol % natural B 4C ond 84 vol %

ccappt‘:r..],2 The mlcroscoplc cross sechons used
- were. those for the. average velocaty at o temper-

;_oture of 700°C ‘and they were calcuiated from the
room-temperafure cross sections given in BNL-325.?

~The neutron. absorphon rate per cublc cenhmeter

in B‘o czf tis glven by

(‘9_)5:““.7"__*' 23 g (th) ,

'_where S is glven in neufrons/cm sec tmpmgmg on
 the front face of the matrix, The heating resulting:
- from- neutron cupfures in B10 was ‘calculated by
‘- i'-assumlng thot in 93% of the absorphons in- Bw-,

 

12A, M. Perry, ORNL,
H. W. Bertini.

private communication to.

33

 
 

 

 

 

ANP PROJECT PROGRESS REPORT

@ 2.32-Mev alpha particle is given off and that in
the remaining 7% a 2.8-Mev alpha particle is given
off.1! The heating from beta decay resulting from
neutron captures in copper was calculated, and it
was found to be negligible compared with the
alpha-particle heating.

The gamma-ray heating resulting from the
emission of decay gamma radiation in the heat
exchanger was calculated on the basis of slab
geometry by using Eq. 6 for each energy group.
The spectrum of gamma rays was divided into the
seven energy groups given in ref. 3, with average
energies of 0.5, 1, 2, 4, 6, 8, and 10 Mev. The
buildup factor for the heat exchanger was used
for each energy group. The total heating at
various points was calculated by summing over
the contributions of each energy group.

The source strength of the gamma rays resulting

from inelastic neutron scattering in the beryllium

reflector was also calculated. These sources are
appreciable only in the first 9 cm of beryllium,
However, their contributions to the heating in the
regions around the heat exchanger are small
compared with those of the beryllium capture
gamma-ray sources, The reason for this is that
the gamma rays from inelastic scattering were
taken to be 1-Mev gamma rays, which are at-
tenvated much more strongly by the beryllium than
are the 6-Mev capture gamma rays. {n addition
the total source strength of the gamma rays from
inelastic neutron scattering is much smaller than
that of the capture gamma rays because of the
volumes of beryllium involved in each,

The source strength of the capture gamma rays
resulting from delayed neutron captures was
calculated by assuming that 50% of the delayed
neutrons were given off in the heat exchanger and
that ‘all these neutrons were captured in the
neighborhood of the heat exchanger. By assuming
that an 8Mev gamma ray was given off for each
capture, the average source strength per unit
volume was found to be small as compared with
the other sources in this region.

The capture gamma-ray sources in the copper
were calculated in the same general way that the
alpha-particle heating in the copper was calcu-
lated, with the source converted to a surface
source of gamma rays. In itself this is not an
insignificant source, but the heating resulting
from this source would be only cbout 2% of the
total from all the sources considered here,

34

The gamma.rays resulting from inelastic neutron
scattering in the core shells were calculated by
using the Curtiss-Wright multigroup fluxes> and
an inelastic microscopic cross section of 1.5
barns (ref. 13) for the constituents of Inconel.
It was assumed thaf a 2-Mev gamma ray was given

off for each inelastic collision. The results .

gave a source strength of about 10% of that
resulting from the capture gamma rays in these
shells,

As has oalready been remarked, the capture

gamma rays in the island core shell were neglected -

because of the shielding properties of the fuel.
Where it has not been stated otherwise, the
dimensions used in these calculations were those
from ref. 1 for the equatorial plane of the reactor.

RADIATION HEATING IN VARIOUS REGIONS
OF THE NORTH HEAD -

H. W. Bertini D. L. Platus 14

Calculationis of the radiation heating to be .

expected in various regions in the north head of
the |ART were undertaken in order to supply
numbers from which thermal-stress calculations
could be made. Because of the complexity and
the time that would be involved in calculating
accurately the heating in all the regions of the
north head, it was decided to make preliminary
estimates of the deposition rates, More accurate
values calculated for other regions of the reactor
were used as guides. In all cases the tendency
was to overestimate the heating. These estimates
can be used to identify the thermal-stress
problems, and where the calculated thermal
stresses are marginal, the heating will be recalcu—
lated. '

Calculations were mcde of the heuf-deposifion
rate in a slab of Inconel bounded on one side by

on infinite fuel region containing the sources of

radiation. This heat-generation rate was used in
all regions in the north head which are bounded by
finite fuel volumes,

The heat-deposition rates in a slab of Inconel
bounded on one side by slabs of sodium of various
thicknesses were calculated, and the resuits were
extrapolated and interpolated to obtain the heat-
g’enerm‘ion rates in the Inconel regions of the n‘orth

 

By, L. Taylor, O. Lonsjo, and T. W, Bonner, Phys.
Rev. 100, 174 (1955).

140n assignment from USAF,

+)

v
 

 

 

ek

head which are bounded by various thicknesses of
sodium,

Fairly accurate calculahons have been made of
the heat-deposition rates in the -Inconel filler
plates below the island and in the vicinity of the
fuel-to-NaK heat exchanger on the equatorial
plane of the reactor. These results were used
as a guide in estimating the heating in some
north-head regions, and new values were obtained
by compensating (by simple exponential attenu-
ation) for decreased beryllium thicknesses,
penetrations through additional fuel layers,
increased thermal-neutron leakage currents into
the north head, etc.

A neutron current of 7 x 1013 neutrons/cm2-sec
was assumed to be escaping uniformly from the
upper portion of the core,'? and it was assumed
that 1 Mw of fission power was being generated in
the fuel regions of the north head by neutrons
escaping into this region. The latter increased
the gamma rays in the fuel by about 30%.

The sources of gamma radiation considered
were those from the sodium and fuel in the north

head, the heat exchangers, boron, the fuel in the

core, core shells, beryllium, and Inconel shell
capture gammas. The sources of beta particles
considered were those from the gases in the
fuel-expansion tank, and the sources of alpha

particles were taken fo be those from boron

captures. The average values of heat generation
obtained in these calculations are presented in
Table 1.2.2. The configuration of the north head
is shown in Fig. 1.1.5 of Chap. 1.1 of this report,

BETA- AND GAMMA-RAY ACTIVITY IN THE.
FUEL EXPANSION CHAMBER AND THE
OFF-GAS SYSTEM

R. B. .'5tev¢;enson“5 |

The power-source dzstrlbuhon of the ochwty of

the gases in the space above the fuel in the fuel
expansion chomber and in the off-gas - line has

 

CA5A, oM Perry, ORNL, prlvafe communicahon to
: _H- W- Berflrflo T

6On ass:gnmenf from Prafl & Whltney Anrcroft. :

, 75,30 Newgard, Fission Product Activity and Deéay |
, Heat Distribution in the Circulating Fuel Reactor with
_F:sszon Gas Strippmg, TIM-205 (Sept. 28, 1955)

. {approximately)

 

PERIOD ENDING JUNE 10, 1956

been determined. The results obtained are to be
used . in the calculation of the radiation heating
and the thermal stresses in this region of the
reactor, _
The radicactive constituents of the gas in this
space are the gaseous fission products, xenon
and krypton, and their daughter products. There
is also a possibility that some volatile fission-
product fluorides will be formed in the fuel and
will escape into this area. However, it has been
shown17 that if all the fission-product fluorides
entered this space, they would add very little
activity to that already caused by the gaseous
fission products and their daughters, Thus, their
effect has been neglected here. Also, there is
some question as to whether the daughter products
of the fission gases will actually be carried
downstream by the off-gas system or whether they
will be deposited on the enclosing walls as they
are formed. In order to get a conservative estimate
of the power-source distribution, it was decided
to treat the daughter products of xenon and krypton
as gases (except insofar as their purging from the
fuel into the fuel expansion chamber is concerned).
The equilibrium amount of the gaseous fission
products in the fuel expansion chamber is given by

‘yl‘Ap
N, = ,
oA 4 Ay (A +A)

where N is the total number of atoms of type 7 in
the gas volume per fission per second, y, is the
saturation fission yield of the ith nuclide, A_ is
the ‘‘purging’’ constant, A_ is the ‘'sweeping”
constant, and A. is the decay constant of the
ith nuclide. The **purging’’ constant is determined
by the volume flow rate of the fuel through the

 

- purging pumps and the total volume of fuel. This
constant determines the amount of the gaseous

fission products which are purged from the fuel
into the expansion chamber. " The "sweepmg

constant is determined by the flow rate of helium
through the expansion chamber and the volume of
the gas space.  This constant determines the
dwell time of the radicactive nuclides in the gas

"space and thus the number of disintegrations they

suffer there. The *‘purging’’ and . ‘'sweeping”’

.. ".constants are given by

volume flow rate of fue| through purging pumps

 

=

p 7 total volume of fuel

volume flow rate of helium through expansion chamber

 

s

volume of gas space

35

 
 

 

 

ANP PROJECT PROGRESS REPORT

TABLE 1.2.2. AVERAGE HEAT-GENERATION RATES IN MEMBERS OF ART NORTH HEAD

 

 

Member Deserioti 'Heat Generation
No.* escription (w/cma)
1 Pressure shell 4
(below sodium expansion tank)

2 Liner 6 w/cm® + 16 w/cm? on ex-
pansion-tank surface due to
beta rays

3 Fuel-expansionstank baffle 3

4 Fuel-expansion-tank wall 6

Upper deck
(regions with sodium on both sides)
é Upper deck 15
(regions with fuel on both sides)
7 Swirl=chamber baffle 3
Swirl-chamber wall 8
Lower deck 8
{regions with fuel below and sodium above)
10 Lower deck
(regions with fuel on both sides) 12
1" Copper-boron tiles 25 w/cm2s + 6 w/cms, where
t = thickness of tiles (cm)
12 Filler block 3
13 Beryllium support struts 10
14 Filler block 1
15 Copper-boron tile 30
16 Flat section of lower support ring 15
17 Strut part of lower support ring 3
18 Lower support ring 1.5

 

*See Fig. 1.1.5, Chap. 1.1, this report, for location of member.

The power-source distribution of the radiocactive
nuclides is found by multiplying their equilibrium
concentration by their decay constant A, and their
average energy per disintegration.

The total power and the power density in the
gas space of the fuel expansion tank as a function
of the volume of the gas and the helium flow rate
are given in Fig. 1.2.3. For these calculations,
A, was taken to be 5.82 x 10=3 sec, which
corresponds to a fuel flow rate of 22 gpm through
the purging pumps. If the purging device is

36

assumed to be 100% efficient, this means that the
reactivity effect of the xenon is reduced to about
0.1% at equilibrium.'® The sweeping constant,
A, is dependent on the helium flow rate and the
gas volume, and thus it is different for each point

on the curves. In converting the STP values
of the helium flow rate, the temperature of the

gas was assumed to be 1200°F, and the pressure
in the expansion chamber and off-gas line was

 

18, . Meem, The Xenon Problem in tbe ART,
ORNL CF+54-5-1 (May 3, 1954).

‘.

1)
 

 

 

 

 

SaeReT
ORNL—-LR—-DWG 14918

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

35 — - — 140
POWER DENSITY
\ o - TOTAL POWER
30 [ [ 120
HELIUM FLOW RATE, 1000 liters/day(STP) L
- //
o 25 . — = 100
5 - z
£ -
3 _~T1000 liters/day (STP) >~
= % N . 80 &
= L
5 7 g
tzu -t Q.
a5 z 3000 Ju==" 60
° A - 5
§ \ 74/ <] _>_Zooo - F
-
© §< S T e
|
A [ ]
4o 5000 T T———1
5 , — - 20
o
//
o 0

 

 

 

0 100 200 300 400 500 600 TOO 8OO
VOLUME OF GAS SPACE (ind)

Fig. 1.2.3. Total Power and Power Density in
the Gas Space of the ART as a Function of the
Gas Volume and the Helivm Flow Rate for a Fuel
Flow Rate of 22 gpm.

taken to be 1.3 atm. The power of the reactor was
assumed to be 60 Mw.

In the calculation of the curves, the very short-
and very Iong-llved nuclides of xenon and krypton
(along with their decay products) were neglected,.

Since the fuel circulation time in the ART is

less than 3 sec, nuclides with half lives less
than this value will decay mosfly in the fuel
before it reaches the purging pumps. Thus very
few atoms with half lives less than about 3 sec
would get into the gas space. Also, for nuclides
with long half lives (greater than about 100 ht),

the number of disintegrations tokmg ploce in the
fuel expansion chamber .and off-gas"line WOuld be
small, since the dwell time at the assumed helium -
flow rates is very short, Thus, these nuchdes' -

may be neglected. -

“In this' study, 32 nuclides were cons;dered N
16 of these being tSotopes of xenon and krypton
“and 16 bemg their daughter products. The main
“contributors  fo the power " distribution: are the_k,_' :
daughter products and not ‘the nuchdes of xenon -
and krypton themselves, -In all cases, the daughter
products contribute about 50 to 60% of the total
 power dlstrlbutlon. Of the total power, about 90%

is due to the beta-ray decays, with only 10% being
due to gamma-ray decays. Thus, in determining
the heating caused by these gases, it can be

PERIOD ENDING JUNE ‘10, 1956

assumed that the heat deposition will occur
mainly in a small surface layer of the materials
surrounding the gases in the fuel expanswn
chamber and the off-gas line.

The power density in the off-gas line as a
function of time and gas veolume for helium flow
rates of 1000 and 3000 liters/day (STP) is given
in Fig. 1.2.4. The time axis can be converted
into lengths along the off-gas line by dividing the
volume flow rate of the helium gas by the cross-
sectional area of the off-gas pipe. Thus, Fig. 1.2.4
gives the power-source density of a cubic centi-
meter of the gas at any position in the off-gas line.

These plots were made by using the well-known
equations of the decay of parent products and the
buildup of their daughters as a function of time.
The initial conditions at the beginning of the
off-gas line were taken as the equilibrium con-
ditions prevailing in the fuel expansion tank.

ACTIVITIES OF NIOBIUM, MOLYBDENUM,
RUTHENIUM, AND THEIR DAUGHTER
PRODUCTS AFTER SHUTDOWN

R. B. Stevenson

The activities of the materials which will plate-
out on the walls surrounding the fuel channel in
the ART during reactor operation will, along with
other factors, determine how long @ period must
elapse before reactor disassembly can proceed.
The activities of the various radicactive nuclides
of niobium, molybdenum, and ruthenium, and their
daughter products have been determined for 100
and 300 days after shutdown and for reactor oper-

_ation periods of 500 and 1000 hr. These three
fission products are expected to plate-out in large
‘quantities, and therefore it has been assumed in
~this . study. that all the atoms of these elements
'rkcreated as fission products are plated-—out There
is’ evidence that other fission products may plate-
 out; however, it is felt that the three taken into
; ‘_.uccount here are the primary ones. -

The only isotopes of these elements that need

" to be considered at times greater than 100 days
‘after shutdown ‘are Nb?3, Ru193, and Rul06,

All the other isotopes have  sufficiently short"

~ half lives that they will decay appreciably in this
“time, and thus they may be neglected. The only
-daughter products that will have large activities
after shutdown for more than 100 days will be

Rh103m gnd Rh196 (daughters of Ru193 and Ru 106,
respectively).

37

 
 

 

ANP PROJECT PROGRESS REPORT

SEORET
ORNL—~LR—DWG 14219

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

[ TTTITT T 1 P
\\ —— —  HELIUM FLOW = 3000 liters/day (STP)
\\\ HELIUM FLOW = 1000 liters/day (STP)
30 N L
GAS VOLUME =20in3
. - N
_ ‘——__'-.____.- 90 ina \
0 St
£ - B \
2 \
=~ 20 \
3>
b N\
2 NN
a \ N ,
x N
wl N
2 T , \ ' \\ '
e L \ \
Ny T L 20003 \\ \
[T S~ N
300in3 T TN b %
s ol \\ \ \
-—-_-___-.. l\.\~~ \
=~ e \\N\ - §\
U NN
) "'--§-':\ \\-.\..::\
S
0 - -
1 2 5 10 2 5 102 2 -5 1003 - 2 5 108
TIME (sec) - '

Fig. 1.2.4. Power Density in the Off-Gas Line as @ Function of Time and Gas Volume in the Ex-

pansion Tank for ¢ Fuel Flow Rate of 22 gpm.

The activities to be expected at shutdown and
100 and 300 days aofter shutdown for reactor oper-
ating periods of 500 and 1000 hr at 60 Mw are
presented in Table 1.2.3, which also gives the
average beta-ray energies, the gamma-ray energies,
and the gammc-ray yields in photons per 100
disintegrations. It has been assumed that, at
shutdown, the fuel is dumped instantaneously so
that no more plating takes place. Also, the
activities of Rh1937 gnd Rh196 gre zero at shut-

38

down, since these nuclides are carried away with
the fuel and have a chance to build up only when
their precursors decay.

The gamma activity at 100 days after shufdown
for 500 hr of operation at 60 Mw is approximately
2 x 10° curies, and at 300 days after shutdown,
for the same .operating conditions, it is about
8 x 103 curies. These high activities at times
fong after shutdown must be contended with in
any disassembly procedure suggested for the ART.

7]
 

 

 

 

 

 

 

PERIOD ENDING JUNE 10, 1956

Q TABLE 1.2.3. ACTIVITY OF PLATED-OUT MATERIALS IN THE ART AFTER 500 AND 1000 hr
OF OPERATION AT 60 Mw

 

 

 

 

. . Activity 100 Days Activity 300 Days
Average - Average Ga Ray Yield Activity at_ Shutdown After Shutdown After Shutdown
. : Beta-Ray Gamma-Ray mma-Ray Yie ______fl’is)_____ {curies) (curies}
Nuclide Energy ‘Energy {photons per 100 After After ‘
(Mev) (Mev) disintegrations) 500 hr of 1000 hr of After After After After
Operation Overation 500 hr of 1000 hr of 500 hr of 1000 hr of
' . pe! Operation Operation Operation Operation
Nb%5 0.053 0.745 100 129105 4.08x105 1.82x10%  5.65x104 3.36x102 - 1.06x10%
Ry103 0.074 0.498 %9 5.65 x 10° 9.57 x105  1.00x10%  1.68x105  3.06x103  5.16 x 103
Ru106 0.0131 0 0 7.84x10%  1.53x10*  6.49%10%  126x104  4.44x103  8.65x103
Rh103m 0 0,04 100 0 0 9.20x104  1.60x105  292x103  4.92x10%
. Rh106 1.05 2.4) 025 0 0 6.49x10°  126x104  4.44x103  865x10°
1.55 0.5 '
1.045 2
0.87 : 1
z - 0.624 12
‘ - 0.513 25
-
&

39

 
 

 

 

 

ANP PROJECT PROGRESS REPORT

1.3. ART INSTRUMENTS AND CONTROLS

E. R. Mann

LOUVER CONTROL SIMULATION

J. M. Eastman! F. P. Green
E. R. Mann :

The ART simulator was used for further study
of the problem involved in closing the main heat-
dump louvers to prevent the fuel from freezing as
the result of a fast insertion of the control rod, Im-
proved transport-lag simulators were incorporated in
the system. The control technique to be used will
be that of closing the louvers to {imit the minimum
NoK temperature.

Two louver-closing speeds have been found to
be necessary to keep the temperature of the NaK
returning from the radiator from dropping below the
fuel freezing temperature. One speed will auto-
matically close the louvers from the open position
to the 10% open position in 9 sec after a fast
controlerod insertion. With the louvers at the 10%
open position and with the four main blowers at
design point speeds, 15 Mw of power will be
removed.

The slow-speed louver actuators will operate on
o temperature signal from the radiator outlets,
When this temperature drops below 1070°F, the
slow actuator will start to close the louvers, The
slow actuator will be capable of closing the 10%
open louvers in 3 sec.

For the simulation study it was assumed that all
four blowers remained in operation, two being
supplied from the TVA circuit and two from the
diesel circuit. With the four blowers operating, a
fast control-rod insertion will automatically shut
off the power to one of the blowers on the TVA
circuit and one on the diesel circuit. A power
failure of either the TVA circuit or the diesel
circuit would then leave only one blower in opera-
tion. The NaK-temperature undershoots would then
be somewhat less, the overshoots somewhat more,
and the ultimate cooling rate lower. Although this
NaK-temperature limit system seems to be func-
tionally acceptable, the system will be further ex-
plored on the simulator. The fail-safe character-
istics are being studied, and modifications may be
made which will require further simulator work for
evaluation, |

 

10n assignment from Bendix Products Division.

40

C. S. Walker

LOUVER ACTUATOR
J. M. Eastman

Specifications for the louver actuators were
established that are based on the following safety
considerations: first, if the actuators should fail,
the louvers must lock into the position they are in

at the time of failure; second, the components

which are not considered to be of first-order re-
liability must be located in an area accessible for
repair without the NaK systems having to be cooled
or drained; and, third, the over-all system relia-

bility and dependability must be high. Since the

actuators and any locally mounted associated equip-
ment will not be accessible during the test, these
components must be of first-order reliability,
Hydraulic-actuator cylinders were selected. They
will be operated in conjunction with spring-loaded
clamps arranged to grip the actuator-output shafts,
For steady-state louver conditions, both ends of
the actuator cylinder will be vented to the hydraulic
return, or drain, The louvers will be moved by
valving hydraulic pressure to either end of the
hydraulic cylinder. This pressure will be simul-
taneously vented to a piston which will release the
spring-loaded clamp. The rate of louver movement
will then be controlled by an crifice in the line
through which return hydraulic fluid will flow from
the cylinder to the drain.

The fast louver-closing rate will be obtained by
opening an orifice in paralle! with the one used for
slow closing. The hydraulic pressure supply unit

is to be located in an area accessible for repair -

work. Two pumps will be used that will operate
in parallel and be driven from different electric
power sources. They will be arranged so that one
can be repaired or replaced without interrupting
the fluid-pressure supply. Solenoid valves will be
used for control of the hydraulic fluid. When not
energized, these valves will hold the louvers fixed,
and they will be energized to move the louvers,
Low-viscosity hydraulic fluid and thin-plate
orifices will be used so that the flow rates and the
corresponding louver-actuation rates will be rela-

tively insensitive to fluid temperature., Valves

ond orifices will be located near the hydraulic-
power-supply unit in an area accessible for repair
work,

 
 

 

't

ENRICHER ACTUATOR
J. M. Eastman-

Specifications were also established for the
actuator for the fuel enricher. The actuator is to
consist of a gear-motor drive equipped to provide a
multiple-synchro indication of the position of the
enricher piston, Enrichment is to be at the rate of
1 Ib of U235 jn approximately 22.5 sec. Three
synchros will be used to indicate 1, 10, and 150 Ib
of U235 per revolution. Thus the synchro dials
will be read in gas-metér fashion to note the pounds
of U235 displaced (equivalent volumetric measure-
ment of enriched fuel mixture, Na UF6) in terms of
piston position. _ _

Data obtained from calibration tests of the
enricher during operation of the high-temperature
critical assembly have been adjusted for changes
in the geometry of the ART, The data indicate
that the enriching uncertainty which will result
from a fuel-nmieniscus effect at the spillover weir
will be equivalent to approximately 4% of 1 Ib of
U235,  The enricher will be able to deliver about
130 Ib. -

FLUX SERVO SIMULATION

F. P. Green E. R. Mann
‘ C. S. Walker

The ARE micromicroammeter and servo amplifier
were used in the preliminary simulation of the ART
flux servo, but the system is unstable because of
the faster reactor response that results from the
lower delayed-neutron contribution. The latter effect
is caused by the differences between the ART and
the ARE in core-residence time of the fuel and in
ratio of core-réesidence time to loop-flow time, The
system is being studied to ascertain methods for
correcting the mstabllcty

CONTROL ROD AND DRIVE MECHANISM
'S, C. Shuford?

The design of the control-rod drive mechanism
‘was completed, and 95% of the parts to be supplied

by vendors have been obtained. -A mechanical

- shakedown test of the mechanism at operating tem-
“perature is plonned Ruggedness, reliability, and
~ ease of servicing were the prime objectives in the

design of the mechanism. A single rod that will

operate - in a thimble passage through the north
head and down the center line of the island to

 

20n assignment from Pratt & Whitney Alrcraft,

PERIOD ENDING JUNE 10, 1956

10 in. below the mid-plane is to be employed to
furnish normal shim control for changing the fuel
mean temperature and to provide for emergency
reactor shutdown, , _

- The neutron-absorbing section of the rod will
consist of twenty-three l-in.,<long, 1.275-in.-0D,
0.775-in.-ID, rare-earth oxide (Lindsay Code 920)
hollow cylinders. (The composition of these rare-
earth oxide compacts and the fabrication methods
used in producing them are presented in Chap. 3.6,
*‘Ceramic Research,'’)

The cylinders are to be retained by Inconel inner
and outer tubes with holes at the top and bottom
to permit sodium permeation and communication
with the stagnant sodium in the thimble. Heat
generated in the rod will thereby be conducted to
the cooled thimble wall.

The drive mechanism is to be powered by two,
Diehl, 6-pole, 200-w, 2-phase, low-inertia servo-
motors. One motor will be bidirectional and is to
be used for servo and manual control. It will give
a rod speed of 0.01% (Ak/k)/sec. The other motor
will be unidirectional and is to be used for fast
insertion. It will give a rod speed of 0.125%
(Ak/k)/sec.  The fast-insertion motor will be
energized through the safety circuit by the follow-
ing emergency-condition signals:

1. any power outage,

2. fuel- or sodium-pump stoppage,
3. flux level above normal,

4, drop in fuel or sodium level,
5. a 3-sec reactor period.

The two motors will be connected through non-
reversing worm gears on concentric shafts that
will, in turn, supply power separately to the sun
gear and to the arm of a differential gear box. This

- arrangement precludes the necessity for clutches

or gear shifts, Both motors will be continuously

‘capable of inserting the rod when either circuit or
both circuits are energized.

‘A high-resolution potenhometer (0 1% |lnear|ty)
will feed a strip-chart recorder to make a permanent
od-posmon record. A spare pofenhometer and
means for external switchover will be provided.

It will be possible to read the rod position to
0,10 in. and to estimate n to 0.05 in. on the re-

corder chart,
Two mdependently driven synchro transmltters

‘will be used for position indication. One will

rotate 300 deg for full-rod stroke and the other will
rotate one revolution per inch of stroke. Resolution
from the two-synchro system will be approximately

41

 
 

 

i
i
|
|

 

 

ANP PROJECT PROGRESS REPORT

0.020 in. To minimize backlash and to prevent

“instrument pinion run-off (with attendant loss of

synchronization), both the synchros and the poten-
tiometers will be driven by a fine-pitch rack 1 in.
longer than the power-tfransmitting rack.

Two plunger-sealed limit switches, which will
operate through the motor control circuits, will be
used to prevent rod overtravel at each end of the
stroke, and one limit switch will be used for giving
a 1% Ak/k reserve indication for filling operations.

The drive mechanism, which will be canned, will
be operated under a 15-psi helium pressure., The
helium consumption will be limited to that lost in
periodic venting of the container.

INSTRUMENT DEVELOPMENT
R, G, Affel
~ Thermocouple Data Reduction
J. T. DelLorenzo W. R. Miller

Equipment suitable for transmitting 100 thermo-
couple signals on two conductors has been as-
sembled and tested. As the simplified block
diagram shows (Fig. 1.3.1), the input signals are

{800-rpm MOTOR

 
  

THERMOCOUPLES
AT REACTOR

scanned by a high-speed mercury-jet switch. This
switch is actually a centrifugal pump, driven by a
motor, which directs a fine stream of mercury at
adjacent pins to complete a circuit. Synchronism
between the two switches is assured by driving
them with synchronous motors that operate from
the same power source. Such a SyStem‘offers
several advantages. For example, o large number
of temperatures may be scanned, visually, on one
large, 17-in.-screen oscilloscope; alarm or moni-

toring systems that scan several hundred readings

a second are possible (available systems are
limited to 1 point/sec or slower); temperature
gradients or profiles on equipment may be directly
displayed; essentially all time lag in the system
lies in the sensor, and therefore transient studies
are limited only by the sensors; commercial null-
balance potentiometric recorders can be used with
such a system, '

The equipment being tested scans data at the
rate of 2500 points/sec. Conventional, Brown,
multipoint, strip-chart recorders have been modified
to operate on the switch output, which is 30
pulses/sec, with each pulse 200 usec in duration,

UNCLASSIFIED
ORNL~LR-DWG 1495}

1800-rpm MOTOR

RECORDERS IN
CONTROL ROOM

 

 

 

17-in.-SCREEN
OSCILLOSCOPE

 

L

 

 

 

 

o D
12~ POINT

 

 

 

 

 

BROWN RECORDER

 

 

S

 

 

 

 

 

0
°00000

 

 

 

 

 

 

 

 

MERCURY JET
SWITCH NO.1§

 

 

 

 

 

 

 

 

 

 

 

 

 

 

12-POINT
BROWN RECORDER
MERCURY JET L
SWITCH NO.2

Fig. 1.3.1. Thermocouple Data Transmitting and Recording Apparatus.

42

hy
 

 

 

 

. f[

Thus, even with an input signal only 0,006% of
the time, the available, conventional equipment
may be used in an integrated alarm or monitering
system,

Cell Bulkhead Penetrations
J. T. Delorenzo

Fittings suitable for reactor cell penetrations for

heater wiring, control wiring, chamber leads, and
thermocouples have been examined. Preliminary
specifications and drawings have been prepared
for review, and a contract is to be negotiated with
¢ vendor,

Fue'aExpansion-Tunk Level Indicator
R. F. Hyland

Tests are under way to determine the practica-
bility of using a helium bubblér to obtain a con-
tinuous fuel-level indication in the ART fuel-
expansion tank in the temperatures range 1100 to
1500°F with o varying pressure cbove the fuel.
The tests were designed so that information could
be obtained on whether the bubbler tubes would
become clogged with fuel or ZrF ; vapor deposits,
to determine the minimum hel:um flow that would
provide an accurate level measurement, to study
the effect of changes in bubbling rate on the

accuracy of measurement, to learn the response of.
the system to pressure surges, and to ascertam“{%

the accuracy of measurement.

The test apparatus consists of a convenhonal'
purge system that utilizes a constant-differential

relay and purgemeter to obtain constant flow and a
pneumatic differential-pressure transmitter. = One

side of the pressure transmitter is hfiped'bff the
purge line, and the other side is connected to a .
gas-pressure tap ubove the fuel.” The outpu! of
the transmitter is fed into a pneumuhc tecorder,

the apparatus, was 2% in the level range being
measured. However, on fuel runs, errors of as

PERIOD ENDING JUNE 10, 1956

large as 0.43 in. of fuel in 5 in, (8.6%) and as
small as 0.12 in. in 5§ in. (2.4%) were noted. - The
average error appears to be of the order of +5%.
This error can probably be accounted for by the
fuel-density change with temperature, which is
known to an accuracy of only t5%, difficulties in
the precise placement of the bubbler tube and the

‘spark plug, and the slight inaccuracy resulting
from the use of a spark-plug probe to indicate the

fuel level in the vessel,
It was found that a purge flow rate as small as
0.2 scth of helium would yield consistent flow

‘measurements, but the system response to pressure

surges at this purge rate was sluggish. Purge
rates of the order of 1.5 scth (~1000 liters/day) of
helium gave both consistent level measurements
and fairly rapid system response to pressure
surges. There is no observable change in level
indication for any purge flow rate between 0.2 and
1.5 scth.

On the basis of 1850 hr of cperation of the appa-
tatus in one run, it appears to be practical to
measure fuel level at a fuel temperature of 1150°F
by the helium-bubble method on a static system,
There has not yet been a sufficient number of
hours of a 1500°F test to evaluate the results of
level indication at this higher teniperature.

. High-Temperature Pressure Transmitters

W, R, Miller

- Pressure  transmitters manufactured by four
- different vendors have been evaluated in tests in
_the temperature range 1000 to 1400°F. Three of

the four units tested operated similarly, They were
of the force-balance type that requires an external
gas supply to maintain a zero pressure differential
- aeross a . bellows or ‘diaphragm. The fourth unit
“utilizes o~ diaphragm-isolated, NaK-filled tube
The bottom of the bubbler ‘tube is located exacflyf'*- system in which-the NaK hydrostatically transmits

5 in. from the surface of the fuel, and a spark plug‘f‘j’ the applied pressure to an external, low-temperature, .

s msfalled to provzde a level reference when the - dlsplacement-type tronsducer. The tube-system

rig is bemg filled,” A reasonably accurate level . units were found to have 0.6% or better, average

indication “is" cssured dunng fillmg by the spark-"‘flrri accuracy. Accordmgly, orders _have been placed

" plug probe “giving off a_flickering. light when fhe__j-,-:i;’flfor 81 of these units. " Development work is con-

" helivm bubbles npple the surface of the fuel. The R t"“"“"9 on other types of pressure tr ansmmers,

‘vessel is maintained at o static pressure of 5 psug 1'7f‘-""_‘;‘_‘"CIUd'"9 e'“"'ca”y operuted devices,
by ventlng off excess hehum. C
"~ The “accuracy “and reproducuhulsty of the Ievel T

V’mcllcahons, os obtained by water calibration of =

ngh-Temperuture Turbme-Type F Iowmeier
L ~G. H. Burger

A high-temperature turbine-type flowmeter has
been developed for measuring flow in fused-salt

43

 
 

 

ANP PROJECT PROGRESS REPORT

and in NaK systems at high temperatures. This
type of flowmeter does not require penetrations to
the process fluid, and it gives very accurate flow-
rate data.

The flow-sensing unit, shown in Fig. 1.3.2,
consists of an Inconel turbine, or rotor, supported
by titanium carbide bearings, and flow-straightening
vanes of Inconel, which, in turn, support the turbine
bearings. The flvid enters the venes and then
impinges on the turbine blades and causes the
turbine to rotate. The turbine contains, within its
body, a small cqbalt vane, which has the unique
property of retaining its magnetic characteristics at
temperatures up to 2100°F, A permanent magnet
assembly is used outside the piping to establish a
uniform magnetic field perpendicular to the axis
of rotation of the turbine and the cobalt vane con-
tained therein, A pickup coil, with a large number
of turns, is mounted so that it intercepts the
magnetic lines of flux of the magnet. When the

turbine is rotated by the fluid, the cobalt vane
rotates and cuts the lines of flux of the magnet;
this changing flux generates a voltage in the pickup
coil. Thus, the number of pulses per unit time is
o direct measure of the turbine speed, which, in
turn, is a linear measure of the flow rate. The
speed of the turbine (in counts, or pulses, per
second) can be measured quite readily by using
appropriate standard electronic equipment, such as
a linear rate meter or scaler. Since the turbine
speed is a linear function of the flow, the flow rate
(in gpm) can be plotted against counts (pulses) per
second, This device is not dependent upon pres-
sure, temperature, or specific gravity but is de-
pendent upon viscosity. The viscosity effect,
which causes turbine ‘‘drag,’’ is not appreciable,
however, at viscosities below 5 cp. A water-
calibration curve of turbine speed vs flow rate is
presented in Fig. 1.3.3. Calibration of this type
of flowmeter with water is valid for the fused-salt
fuel mixtures. ‘

UNCLASSIFIED
PHOTO 26144

~w——— FLOW DIRECTION

 

 

TURBINE

RETAINER RING

  

 

INLET SIDE GUIDE AND BEARING ASSEMBLY

 

  

o2 3

 

INCHES -

  

Fig. 1.3.2. ORNL-Designed Turbine-Type Flowmeter for Flow Measurements in the Range 2.5 to

50 gpm at Temperatures Up to 1600°F.

44
 

 

 

it e oAl 5 i ] b

UNCLASSIFIED
ORNL-LR-DWG 14952

 

t40

/|

 

120

/

/

 

100

o
o

®

/

/

 

 

TURBINE SPEED {counts /sec)
o .
Q

b
o

 

A

 

 

 

 

 

 

 

 

 

 

0 2 4 6 8 10

Fig. 1.3‘.3. Weter-Calibration Curve for a Turbine-

Type Flowmeter.

FLOW RATE {gpm)

12

14

PERIOD ENDING JUNE 10, 1956

Experimental models of the turbine-type fiow-
meter are being tested in gas-fired forced-circulation
loops. The flowmeter is welded into the pump-
discharge section of the loop, and the magnet
assembly is cooled with plant air to keep the pickup
coil at a low temperature. The experimental results
obtained thus far indicate that the principle of
operation of the turbine-type flowmeter is sound
and valid for high-temperature operation, The
cobalt vane has been satisfactory at temperatures
up to 1400°F and should be satisfactory at 1600°F.
Information is to be obtained in further tests on
bearing corrosion and self-welding and on bearing
tolerances at various temperatures. It is felt that
after sufficient operating data are obtained, a
satisfactory design can.be worked out for scaling
the flowmeter to any desired size. It is further
believed that this method of flow measurement can
be made accurate to within 1%, which is much
better accuracy than is attainable with the con-
ventional magnetic flowmeters used on liquid-
metals systems. The magnetic flowmeters speci-
fied for use on the 3.5-in., sched-40 pipe in the
ART NoK system must be calibrated individually
to achieve 4% accuracy at a flow rate of 1200 gpm.
In the larger, 10-in. pipe, the mognetic flowmeters
can give accuracies of only about 12% at 8000 gpm.

45

 
 

 

 

 

 

 

ANP PROJECT PROGRESS REPORT

1.4. COMPONENT DEVELOPMENT AND TESTING
H. W. Savage

| PUMP DEVELOPMENT
E. R. Dytko! A. G. Grindell

- 'VBe'qrin‘g-and'-Sedl Development
W, L. Snapp! V. K. Stair2

- The maximum operating temperature expected in
the lubricating and cooling fluid used in the ART
reactor pumps is approximately 220°F. For an
operating period of 500 hr, with an oil dilution
factor of approximately 0.03, the integrated radia-
tion dose absorbed by this pump lubricant is cal-
culated to be 3 x 108 rep.: Therefore it will be
possible to use a lubricant with o mineral-oil base
for the ART pumps. '
~ Since a petroleum product could not be used if
the operating conditions given above became more
severe, tests were continued3 in an effort to find
a liquid more suvitable for higher performance.
Important factors affecting the choice of liquid are
-.compatibility with the materials of construction
(that is, elastomers) and process fluids, radiation
‘stability, thermal stability, heat removal capacity,
and lubricating properties. The loads that will be
imposed on the pump bearings will be light, and
therefore the lubricating properties are of only
secondary importance, The other considerations
listed are all of primary interest ond are closely
interrelated.

After the failure of the UCON fluids to perform
satisfactorily,® the next synthetic liquid investi-
gated was an organic-phosphate-based material
manufactured under the trade nome ‘‘Cellulube’’ by
the Celanese Corp. of America. Cellulube 150
was the particular grade selected for evaluation
testing, since it most nearly approximated the
operating properties of the liquids previously
tested.

The initial test of Cellulube was made at a tem-
perature of 210 to 220°F in a fuel pump (model MF)
rotary element in a mechanical shakedown test
stand, Over the entire operating period of 636 hr,
the leakage of the lower seal was too small to be
measured. The total leakage of the upper seal was

 

t0n assignment from Pratt & Whitney Aircraft.
2Consultunt, affiliated with the University of Tennessee.

3W, L. Snapp ond W. K. Stair, ANP Quar. Prog. Rep.
March 10, 1956. ORNL-2061, p 43.

46

430 cm3. The unit was disassembled after this
636-hr period of satisfactory operation, and inspec-

tion revealed fully developed uniform wear patterns

on both seals.. However, the wear path on the
upper seal rotor exhibited & ring, which, upon
examination, proved to be a copper deposit. Al-
though it was known that Cellulube would attack
copper at high temperatures, ‘attack . was not ex-
pected at the operating temperature of this test.

With the exception of the upper seal, which was
replaced entirely, the unit was reassembled, in-
stalled in the same stand, and operated under the
same conditions for an additional 480 hr. During
this period the operation was, again, completely
satisfactory, with the total leckage of the upper
seal being 70 cm3, Since all the leakage occurred
during the first 24 hr of operation, it may be con-
sidered as ‘‘run-in’’ leakage. The. lower seal
leakage totaled 40 cm3 for the 480-hr period. This
low leakage rate is particularly significant because,
when seals tested with other fluids were dis-
assembled, inspected, reassembled, and tested
again, with the same fluid, the retest leakage rates
were always higher than those measured initially,
inspection of both seals after the test revealed a
very light reflective film of copper on all surfaces
exposed to the Cellulube,

A second test of Cellulube, again at a tempera-
ture of 210 to 220°F, was conducted on a model MF
fue! pump rotary element mounted in a bearing-and-
seal test stand which permitted the application of
transverse shaft loads, The total operating time
was 595 hr, with an equivalent journal bearing
load of 150 Ib being applied for the last 504 hr,

The journal bearing, which was designed for a

load of only 70 b, suffered no measurable wear,
and seal performance throughout  the test was -

satisfactory. Both the upper and the lower seals
had well-developed wear patterns, with a very

slight trace of copper plating visible on all sealing .

surfaces. It should be noted that all the copper
plating experienced during the testing of Cellulube
was far less than that resulting from operation with
the UCON fluids,® and it did not appear to have @
marked effect on seal performance.

All the elastomer O-ring seals used in the
Cellulube tests had a Buna-N base. Such. com-
pounds have a tendency to swell, become tacky,

rj

o
 

 

W

&

and lose their hardness in the presence of Cellu-
lube. These particular problems could be solved
by using Oe-rings mode from butyl rubber, but,
unfortunately, butyl rubber suffers severe radia-
tion damage. Tests fo determine the radiation
stability of Cellu!ube are planned for the near
future.

Dowtherm A whrch is a good heot tronsfer.

medium but which has poor lubricating properties,
was also selected for study because it is rela-
tively stable upon exposure to radiation. Dowtherm
A is a eutectic mixture containing 26.5% diphenyl
and 73.5% diphenyloxide that is manufactured by

‘the Dow Chemical Co. of Midland, Michigan, The

only test run to date was terminated after 575 hr
due fo a motor failure. There was no transverse
load applied during the test. The. Dowtherm A

operating temperature was 210 to 220°F. Both

seals had well-developed wear patterns, The upper
seal showed no leakage ofter the first 4 hr, The
lower seul had no leakage for the first 150 hr, at
which time it began to leak at g neul_'Iy uniform
rate of 20 cm3/day. Upon examination, the lower
seal rotor was found to have a light copper deposit.

This is contrary to the information provided by Dow. |

The system used for this test had previously been
operating with Cellulube, und, although it was

carefully and ‘thoroughly cleaned, it is possible-
that some Cellulube may have remained and caused
the plating. This problem will be investigated in

future tests. )
The process side of the seal nose had a black,

‘gummy deposit,. which proved to be virtually all

carbon. Since there was no evidence of physical

damage to the Grophrtor seal nose, it is probable -
that this deposit was the residue from- Dowtherm A
“which thermally decomposed in Ieokmg across the

seal interface. Again, all the elastomer O-ring

lubricants.™
results shou]d be forthcoming soon, -

As ongmoliy des:gned the reactor pumps utilized,
as the upper bearing, a spherical roller type of

to the lubricating oil.

PERIOD ENDING JUNE 10, 1956

antifriction bearing. In addition to providing ample
thrust and radial load-carrying capacities, this
type of bearing permitted shaft alignment with
respect to the journal bearing without attendant
shaft deflections. This design oppeored to be ade-
quate unti! tests of long duration on pump rotary
elements revealed that the initial bearing end play
of 0.0045 in, (maximum) had increased by a factor
of 2 to 3. Such excessive end play is not con-
ducive to good dynamic seal performance and it
also interferes with the close axial 1oleronces
required in these pumps. The easiest solution to

~ this problem, with respect to minimizing pump

redesign, was to change to o double-row angular-
contact bearing having convergent angles of con-
tact. This has been done and, to date, all per-

formonce has been satisfactory. However, more

operating time will be necessary to prove the
reliability of the redesigned bearing arrangement,
A temporary impasse has been reached in the
development of seals for the NaK (primary and
ouxlllary) pumps. All the seals purchased for
comparative testing have failed in operation, The
seal obtained from the Sealol Corporation failed
because the unhardened stoinless steel wear ring
rotor and the mating carbon ring wore out in less

~ than 20 hr. The ceramic rotor of the seal obtained
- from- the Durametallic Corporation -disintegrated

upon application of o 70°F temperature differential
An attempt to correct this
condition will be made by vutilizing a thin ceramic
facing on a steel-bodied rotor. The Byron Jackson
Co, seals never developed full wear patterns, even
after runs in excess of 300 hr. This condition

“apparently resulted from distortion of the relatively
- thin rotor, - At no time during test operation with

any of these secls was @ satisfactory. leakage rate

“established. Therefore efforts are being made to
seals used had a Buna-N base,. and they were not ..
compatible with Dowtherm A, ~This difficulty can
be overcome by using natural rubber, which should :
be odequate from a rodlonon stcmdpolnt. The only
~difficulty with the natural rubber may be its thermal
} ,;stablllty, which wu!l be determmed in future o

- testlng. N :

U A test stond is now’ ovclloblc for. testmg the'
S ;_elasfomerlc compounds recommended for O-rings in
' a thermally hot environment in contact with various
“These tests are now under ‘way ond B

adapt @ seal configuration. Slmlldl‘ to that being

used on the fuel and sodium pumps. " In addition,
- other promising types of commercmlly ovolloble
,seals will be investi goted - :

Pump Lube-OnI System Gos-AHenuchon Tests
~S. M. DeComp ‘

The leokoge tests of the seals on - the model

 MF-2 fuel pump, started previously,4 were con-
tinved with Gulfspin 60 oil being used as the
lubricant, * For these tests the temperature of the

 

45, M. DeCamp, ANP Quar. Prog. Rep, March 10, 1956
ORNL-2061, p 48.

47

 
 

 

 

ANP PROJECT PROGRESS REPORT

system was 1000°F, the temperature of the lubri-
cant was 160°F {max), the pump speed was 2400
rpm, and the gas pressure in the system was 5 psig.
There was no liquid in the main pump loop. The
seal leakage data were obtained by bleeding 550
liters of helium through the pump into the toop
system each day and at the same time bleeding
500 liters of argon into the loop system directly.
Fifty liters of gas was bled from the pump catch-
basin drain each day and 1000 liters of gas was
bled from the pump loop system, Gas samples
were taken by evacuating a glass container and
then filling it through a line connected to a sample
joint. The sample joints were at the catch-basin
drain, the off-gas line from the loop system, and
the gas space over the lube-cil reservoir. The
samples were analyzed ~spectrographically for
argon, helium, oxygen, and nitrogen. The appear-
ance of argon in a sample taken at the catch-basin
drain was an indication that the argon had back-
diffused up the annulus between the pump shaft
and shield-barrier plug. The appearance of argon
in the sample taken at the lube-oil reservoir gas
space was an indication that the argon was leaking
across the lower pump seal. The data obtained
from this test are presented in Table 1.4.1.

During the tests it became apparent that the
control of the small gas flow rates ond low pres-
sures was more difficult that had been anticipated
and that the adequacy of the method of gas sampling
was questionable, Therefore the data presented
in Table 1.4.1 give only qualitative indications of
the leakage. Future seal leakage tests will be
made by bleeding argon down the pump shaft and
using helium as the trace gas. This should improve
the accuracy of the data obtained spectrographically.
The masking effect of air will also be less. The
sampling technique will be improved by continuously
bleeding gas through the sample bottles to elimi-
nate the possibility of drawing gas from regions
other than the sampling points.

Fuel Pump Endurance Tests

S. M, DeCamp

An endurance test of an ART fuel pump (model
MF) was started on April 10, 1956 with the fuel
mixture (No. 30) NaF-ZrF -UF, (50-46-4 mole %)
as the pump fluid. The first 1488 hr of operation
of the system is described here, The operating
conditions are indicated on Fig. 1.4.1, which gives
the system temperatures, oil flow rates, and seal

TABLE 1.4.1. FUEL PUMP (MODEL MF-2) SEAL LEAKAGE AS MEASURED BY THE
AMOUNT OF ARGON DIFFUSION FROM THE PUMP LOOP INTO GAS
SAMPLES TAKEN AT POINTS EXTERIOR TO THE PUMP SEALS

 

Argon Found (ppm)

 

 

Date of Sample ‘
Sampling No. At Catch- At Off-Gas At Lube-Oil
, Basin Drain Vent Reservoir
3-24-56 1 3,200 260,000 280
3-26-56 2 16 218,000 344
3-27-56 3 15,000 230,000 400
3-27-56 4
3-28-56 5 Samples taken but lost through mishandling
3-29-56 6 |
4-3-56 7 124 290,000 1830
4-4-56 8 7 247,600 137
4-5-56 9 15 174,600 | 48
4-8-56 10 0 193,600 86
4-9-56 1 738 152,000 ' 84
4-9-56 12 9 106,400 201

 

48
 

BONFBEMNTFHL
ORNL-LR-DWG 14953

OPERATING TIME (hr)

 

o 96 192 288 384 . 480 576 672 . 768 864 960 1056 152 1248 1344
400 e e 1
RUN 8-SEAL LEAKAGE TEST, NO FUEL, HIGM TEMPERATURE, ~ 1000°F RUN 9 - FUEL BEING PUMPED
360 |
320
. 280 ‘
s FUEL TEMPERATURE
-
E
2 240
w
< N
~
<
Wl .
25— - 200
v
o
g
o LUBE OIL FLOW RATE
~20F ¢ t60
E
Q.
= -
w
e s 120
3 o
3 _
v _ BOTTOM SEAL LEAKAGE
5 1.0 — 80
w
.
2 — ‘
_ INLET OIL TEMPERATURE
0.5 |— 40
. TOP SEAL LEAKAGE
0l o 7 ‘ .
0 4 8 12 16 20 24 28 32 36 40 44 a8 52 56

OPERATING TIME (days)

Fi'g.'.'l.4.l- ART Fuel Pump (Model MF) Endurance Test Performance.

6y

1440

 

60

1400

1200

1000

800

600

400

200

TEMPERATURE {°F}

9561 ‘01 INNF ONIANI 40Ol¥3d

 

 
 

 

 

ANP PROJECT PROGRESS REPORT

leakage as a function of operating time. The speed
of the pump was maintained at 2400 rpm duwring the
period described by Fig. 1.4 1. In addition to the
temperature fluctuations shown on the graph, the
temperature of the system was cycled 17 times

" between 1300 and 1100°F. It takes approximately

”/2 hr to lower the temperature to 1100°F and
2Y% to 3 hr to increase it to 1300°F. The lubri-

cating oil being used in the system is Gulfspin 60.

Helium is being bled down the pump barrel at a
rate of 550 liters/day, and 50 liters/day is bled
off through the catch-basin drain. Present indi-
cations are that low leakage rates and satisfactory
removal of seal oil leakage from the catch basin
can be obtained..

~ Thus far very little trouble has been experienced
in the operation of this pump. - The hydraulic-drive
equipment has functioned satisfactorily, except for
some trouble with the hydraulic pressure indicator,

Fuel -Pu.mp Development Tests
J. J. W, Simon®

The two loops designed for high-temperature
performance testing of the ART fuel pump and
centrifuge assembly were completed and are being
used for development tests. Water performance
data were taken on these loops to determine whether
the addition of the xenon-removal system would
have an effect. on pump performance,

The head and flow curves obtained for these
loops substantiated the original water-test data.?
During the tests it became apparent, however, that
the system pressures were not stable. The fluc-

tuations in system pressure appeared to be «

function of the water level in the fuel expansion
tank and the loop resistance. In some instances
these fluctuations appeared to be large enough to
be detrimental to reactor operation,

Two types of modification will be made to the
system in an attempt to locate and remove these
fluctuations. . First, the pump suction inlet con-
ditions will be improved, and, second, the xenon-
removal system will be altered to study the effects
of each of the various flow passages.

~ Sedium Pump Endurance Tests
S. M. DeCamp

Preliminary water tests of a sodium pump that

had been instailed in a test stand for endurance

50

testing indicated that leakage, or bypass flow,
around the impeller was causing ingassing of the
system. The pump and the loop were therefore
modified, as described in the following section,
by providing a pressure-breakdown labyrinth along
the shaft and by removing the bypass vanes from
the back face of the main impeller. The system is

‘now being prepared for high-temperature operation,

As in the fuel pump endurance fests, described
above, seal leakage tests will be run before the
system is filled with the test fluid.

Sodium Pump Development Tests
S. M. DeCamp

The original design of the ART sodium pump
(model MN) called for the bypass flow around the
main impeller to be from the expansion tank to the
pump discharge. This was to be accomplished
with the aid of vanes on the back side of the main
impeller, During the first series of tests of this
pump, however, it became evident that the system
was ingassing. In an effort to correct this situation
the bypass flow was reduced by reducing the vane
diameter on the back of the impeller, but ingassing
still occurred. It was decided then to reverse the
direction of bypass flow, that is, to allow the flow
to be from the pump discharge to the expansion
tank. This flow condition allows the gas bled
down the pump shaft to flow to the expansion tank
in the same direction and through the same opening
as that through which the bypass liquid flows.
This reversal of flow was accomplished by com-
pletely removing the bypass vanes from the back
of the sodium impeller, The amount of reversed
bypass flow was restricted by adding a pressure-
breakdown labyrinth around the shaft. Tests of the
modified pump - indicate satisfactory operation,
Tests are now under way to determine whether the
expansion tank and the bypass flow lines can
handle reactor bypass flow requirements without

ingassing of the system fluid. The best sodium

pump performance data obtained to date are pre-
sented in Fig. 1.4.2, '

 

30n assignment from Pratt & Whitney 4Ai|_'cr'uft. »
6R. Curry and H. Young, ANP Quar. Prog. Rep. Sept,
10, 1955, ORNL-1947, p 44. -

7R. L. Brewster et al,, ANP Quar. Prog. Rep, Sept. 10,
1955. ORNL-1947, p 29. ‘
 

 

 

 

 

PERIOD ENDING JUNE 10, 1956

BEHPIDUNTIRE ™
ORNL-LR-OWG 44854

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

140
120 335070,
e
100 -~ <
0-0%9& ZoEsion poiNT "k.\"
e 2%00rpm “\‘
g 2790em —.
- ' -._."-'- -
W 60 2400 rpm
—
\%
2000 rpm
40 g
1600 rpm .
20 1200 rpm
Sty ve s
800rpm .
0
50 {00 150 200 250 300 350 400 450 500 550 600 650
FLOW (gpm)

Fig. 1.4.2. ART Sodium Pump (Model MN) Performance Curves.

Primary NaK Pump Development Tests

H. C. Young® J. G. Teague®
M. E. Lackey

Water tests on the primary NaK pump (model
PK-P) volute and impeller, initiated previously,?
were completed. The final performance curves
which show head vs flow, efficiency, and the volute

pressure balance line at several constant speeds
from 1225 to 3550 rpm are presented in Fig. 1.4.3,
and o plot of pump shaft horsepower vs flow for -
the same head and flow condmons is. shown in

Fig. -1.4.4, The original test design point of

1220-gpm flow and a 340-ft head has been changed -

to 1220-gpm flow over the head range indicated by

~the “vertical line on. Fig. 1.4.3. The volute’ hy-
“draulic balance “line crosses this head range at -
(cpproxnmotely the m:d-pomt, und thus the normal

‘operating conditions appear to be in the runge of

.'occeptable hydruuhc balance. - -

- Tests were run at speeds above 2900 rpm by,
. using a ‘200-hp ‘wound-rotor motor, ‘which was ‘re-
- ploced with. '@ 60+hp dec ‘motor in. order to obtam'

 

80n assignment from Pratt & Whliney Aircraft.

H. C. Young and M. E. Lackey, ANP Quar, Prog.
Rep, March 10, 1956, ORNL-2061, p 48.

dataat the lower speeds required for reactor startup
operation.

In order to determine the correct volute-tongue
angle, a directional probe was installed in the
volute at the periphery of the impeller to measure
the angle at which fluid left the impeller. The
data obtained indicated that the original tongue
angle was correct.

- A weir was installed in the pump pot to measure
‘leakage flow through the labyrinth seal around the
impeller hub. Some leakage flow is needed to
remove ‘gas from the system fluid; however, ex-
cessive leakage, parhculcrly at low liquid levels,
--would cause agitation of the liquid level in the

pump pot and possibly result in ingassing through
" the pump suction. = A series of three tests was
 -conducted in which the radial clearance between
- the impeller hub and the top labyrinth seal and the
axial  clearance befween the dynamic seal vanes
- on the back of the impeller hub and the stationary
element were varied. Bypass flow was measured
“~-during these tests, and the pumped fluid was
observed through Plexiglas ports in the loop.
There was no visible indication of ingassing. The
tests showed the optimum axial clearance to be

51

 
 

 

 

ANP PROJECT PROGRESS REPORT

SOTPIT,
ORNL—LR—DWG 14955

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

450
‘ 070 0.7% o076
400 2?5,._ e e } 077
060 _~-T1 0 —p— y 0.775
. / ”-_- ‘ L
055 LA~ _~1 \ 3550 rpm
05;3_ /,J'/ __;______. —t _ 3500 rpm
’ Iy
/ "
1 1
350 /.? 'y ‘/1/ i_- ] 7 3400 rpm
4t i — ' 1 r
27T ! ____...._.:.____ * 3300 rpm
o ‘fl ' L~ 1 1
300 7 e i 3200 rpm
LT AT i /
// oo ! ,' ,' /’_——DESIGN RANGE
~ |t 3 ! g :
~ 1o ! ,——-',""'" ,' / / ———{3000 rpm
250 | e e
— . am— oy
= T 4—"',”,1— — =7 e || 2900rem
e LT LA / 27330
w # -—r""_ - 7
I L~ o _E— / fi\-;
L= a7 / / / .
- ! '/I_-—-——'l 7 .
~ 1 _4 ’l ” / 2400 rpm
. 1 ] / J /"‘-—VOLUTE PRESSURE BALANCE LINE
L~ i 1 / Y /
i1 / .
T / 7
00 T 1 ) , A e 2000 rpm
¥ ¥ ¥ 7
by i \ z / 7 . ;
_L‘ v\ M < — — — — PUMP EFFICIENCY
N e LT
— Vi NS P <OI540 rpm
>0 NS~
— e
- B _"'";-v_/;—" 1225 rpm
/
. '//’
0 .
o 200 400 600 800 1000 1200 1400 1600 1800 - - 2000

FLOW (gpm)

Fig. 1.4.3. 'ARTv Primary NaK Pump (Model PK-P) Performance Curves of Head vs Flow Showing

Efficiency and Volute Pressure Balance Lines.

0.030 in. and the optimum radial clearance to be seal radial clearance of 0,015 in. should be suffi-
0.015 in. These clearances will result in a bypass cient to prevent rubbing of the impeller hub against
flow rate of 3 to 6 gpm at 1220-gpm pump-discharge  the top labyrinth seal. '

flow. - ‘ ' ' These water tests of the primary NaK pump indi-
, R cate that the pump will meet the head and flow
‘Static  deflection tests were made on the pump  conditions required for the ART. A pump tank gas
- shaft used in the water test rig and on the Inconel  pressure of 10 psigshould be sufficient fo suppress
pump shaft to be used in the final pumps. The  cavitation at the design flow rate and up to a NaK
.information obtained, along with the calculated  temperature of approximately. 1380°F. The design
deflection of the Inconel pump shaft during high- conditions are in the range of minimum volute
temperature operation, indicated that a top labyrinth  hydraulic unbalance. '

52

o
 

£S

HEAD {ft)

 

T

 

SOMNADENTHY, -
ORNL—LR—DWG 14956
200

 

500,

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

      
 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. ".)0 (9‘“
5%
& _
450 P — 180
: 95?’ h/ o
o _a00 v
. _ _|HEAD AT 3550 rpm | ] 2 o
| ——l T | 7
HORSEPOWER / T
. == == = HEAD ' ' / 340
250 = e 00 TOML 140
}_300 rpm
. 3200 rpm
300 - 120
250 fo 10
2610 rpm / -
) —-——-—-———n——_—._d___-
200 — 80
2400 rpm —_— ‘
— S e — ——'_—_--“—_?-——7—&_
150 2610 ("] 60
"
____.—-———7(._._.__‘_2_0_0.9@)”1 '___
| ) | / --h)):"
_ 0o | -~
100 -+ 240 — 0
: m
ot '2.000 P /
1540 rpm |
. _—-—b__--——_ -—-——uq-—_--—_-.--
50 1228 rpr ' L 20
s 1225 rpm
0 - - 0
200 300 400 500 600 700 800 900 1000 1100 1200 130C 1400 1500 1600 1700
' o FLOW {gpm)

Fig. 1.4.4. ART Pi‘iinory NaK Pump (Model PK-P) Performance Curves Showing Shaft Horsepower vs Flow.

SHAFT HORSEPOWER (BASED ON WATER FLOW}

956L 01 INNF ONIGNI QOId3d

 

 
 

 

ANP PROJECT PROGRESS REPORT

Auxiliary NaK Pump Development Tests

H. C. Young J. G, Teague
M. E. Lackey

 Water tests were conducted on the auxiliary NaK
pump (model PK-A), The volute used for these
tests consisted of two brass volute halves bolted
together. The volute flow passage was accurately
milled in each volute half, The test impeller was
fabricated of brass, with vanes silver-soldered to
the hub and soft-soldered to the inlet shroud.

The pump assembly was tested in the test loop
used for the water tests of the primary NaK pump.?
A 60-hp d-c drive motor was used. No changes
were necessary in the instrumentation, except
that a new orifice was installed to measure the
lower flow rates,

Performance. and cavitation data were obtained
for as wide a flow range as possible within the

practical limits of volute hydraulic unbalance, and

the final head vs flow curves are presented in
Fig. 1.4.5. A plot of pump shaft horsepower versus
flow for some head and flow conditions is shown in
Fig..1.4:6. The volute hydraulic unbalance over the
design range was found to be quite satisfactory,

SR
. ORNL-LR-0WG 14957
450

J PUMP
EFFICIENCY RANGE
! VOLUTE PRESSURE
400 _ BALANCE LINE
SPECIAL rpm
DESIGN POINT S :
350 3450rpm
3350rpm
300
3250 rpm
3MS0rpm
= 250
3 _ 3050rpm
T 200 2950rpm
2850rpm
150 7 om
2250rpm
100 -
: /
1900 pm
7
50

000 rpm

 

0 100 200 300 490 500 600 700 800
FLOW (gpm})

Fig. 1.4.5. ART Auxiliary NoK Pump (Model
PK-A) Performance Curves of Head vs Flow Show-
ing Efficiency and Volute Pressure Balance Lines.

54

and therefore this pump will be satisfactory for

use as the special NaK pump for cooling the ART
fuel fill-and-drain tank. This pump requires 300-gpm

flow at a 275-ft head. The impeller diameter, the

dynamic seal vanes, and the top labyrinth seal are
the same as those for the primary NaK pump, and
the speed ranges of both pumps are the same.
Therefore the radial and axial clearances determined
by the primary pump tests were used for the auxiliary
pump. The pump showed satisfactory degassmg
characteristics with these clearances.

After the initial performance tests were com-
pleted, a 0.023-in. spacer was installed between
the volute halves, and tests were conducted to
determine whether the addition of the spacer would
change the volute hydraulic unbalence. The pur-
pose of these tests was to ascertain the allowable
tolerances for use in welding the lnconel volute
halves together,

The volute hydraulic unbalance was measured
during these tests by eight static pressure taps
equally spaced in the volute near the periphery of
the impeller. The static pressure readings at any
of the eight taps, for a given discharge head and a
set flow rate, did not change over 1 psi from the
measurements taken without the spacer,

In general, the water tests indicate that the pump
will meet the head and flow conditions required for
use in the ART, both for the auxiliary and the
special NaK pumps., The design conditions are in
the range of minimum volute hydraulic unbalance.
Cavitation characteristics of the pump, bosed on
extrapolations from water to NaK, indicate that a
pump tank gas pressure of 14.4 psia, or ~0.3 psig,
should be sufficient to suppress cavitation at
design flow for a NaK temperature of 1282°F, as
compared with a required pressure of 6 psig for the
primary NaK pump at the same NaK temperature.

Primary and Auxiliary NaK Pump Test Stands

H. C. Young J. G. Teague

Fabrication of all parts for two primary and two
auxiliary NaK pump high-temperature test loops is
under way. At present, the pump volutes are the
most critical delivery item. A method of profile-
machining the volute halves from a pattern is being
used. Several primary pump volute halves have
been received, and the vendor is now setting up for
profile-machining of the first auxiliary pump volute
halves.

A series of welding tests were conducted to

determine the joint preparation and the welding

1]

Cy

&

 
qS

HEAD (ft)

450 (——

400

a50

200 -

100

50

ERNEIBEAT b
ORNL—~LR—DWG 14958

 

 

HEAD A
— o 21 3550 rpm

 

.| ~———HORSEPOWER|
.| === == HEAD

3450rpm

o — ——v——

AUXILIARY PUMP ]
n | DESIGN RANGE—

s,
l—-.__-

P S

 

   
 

80

 

i

—~~.

  
 

70

 

300 fi

anE-

; S

 

250" f

L s e s

_‘
“. oo

~3050 rpm
o sl o

 

60

 

2850 rem 1T

I 850 M

.. SPECIAL PUM
.\~ DESIGN RANG

A

——— o o— —

50

 

2250 ?fim

gl Uy s oS

40

 

150"

    

rpm

A

 

30

 

]
_—

\
J
/

1900 rpm

\

20

 

e -

1000 rpm |

-—."‘—--

 

 

 

 

 

 

 

1000 er

 

 

 

 

 

 

 

 

 

. 200

250

300

350

FLOW (gpm)

400

500

550

" Fig. 1.46. ART Auxiliory NaK Pump (Model PK-A) Performance Curves Showing Shaft Horsepower vs Flow.

SHAFT HORSEPOWER (BASED ON WATER FLOW)

9561 ‘0L INNr ONIGNT QOl1¥3d

 

 
 

 

 

ANP PROJECT PROGRESS REPORT

techniques required to control weld shrinkage so
that final volute dimensions would be within ac-
ceptable tolerances (see Chap. 3.4, *‘Welding and
Brazing Investigations’’). The first set of primary
pump volute halves has been welded. '

In the first tests in the new loops the pumps will
be operated with water in order to check the pump
assembly, test-stand vibration, loop resistance,
and operation of the system throttle valve, and,
of primary importance, to correlate performance
and cavitation data with the data previously ob-
tained on the water-test rig.

Some engineering changes are being made in order
to lengthen the piping on the primary NaK pump
endurance-testing loop and also on the guxiliary
NaK pump endurance-testing loop and to relocate
these loops for use in calibrating electromagnetic
flowmeters. -

HEAT EXCHANGER DEVELOPMENT
E. R, Dytko

R. E. MacPherson J. C. Amos

Intermediate Heat Exchanger Tests
J. W. Cooke H. C. Hopkins'

The information obtained from intermediate heat
exchanger (IHE) test stand operation during the
quarter is summarized in Table 1,4,2, York radiator
No. 9, which is being tested in intermediate heat
exchanger test stand A, is currently undergoing a
thermal-cycling program consisting of 16 hr of
power operation followed by 8 hr of isothermal
operation. During the power phase of the cycle,
the radiator NaK inlet and cutlet temperatures are
1500 and 1100°F, respectively. Isothermal opera-
tion is at 1500°F. Transition from isothermal to
power operation is accomplished by dropping the
NaK outlet temperature at a rate of 20°F/sec for

TABLE 1.4,2, SUMMARY OF INTERMEDIATE HEAT EXCHANGER TEST STAND OPERATION

 

 

Hours of b
f R
Test Unit? Nonisothermal Total Hours Number o eason f'or
. of Operation Thermal Cycles Termination
Operation
Test Stand A
.York radiator No. 9 471 973 20 Test continuing
(revised design)
Circulating cold trap No. 2 1218 Test continuing
(4 in. in digmeter®)
NaK screen filter 264 Test continuing
Test Stand B
Black, Sivalls and Bryson heat 362 403 31/2 Test continuing
- exchangers Nos. 1 and 2 '
" (type IHE-3)
Cambridge radiators Nos. 1 ‘ 820 1345 6'/2 Test continuing
and 2 {modification 3)
' Circulating cold trap No. 3 630 " Removed for exami-
(4 in. in diameter) nation after heat
exchanger failure
Circulating cold trap No. 5 715 Test continuing

 

(4 in. in diameter)

%ncludes only units tested during this report period.

bFor tests in progress the total operating time is shown as of May 15, 1956.

©This type of cold trap was praviously referred to as 80-gal system.

56

)

4/

"
 

 

 

 

 

@k e o rmcknin 1

re

3 .

$

“-r..

&

the first 200°F and 8°F/sec for the next 100°F,
The final 100°F drop is at a slower rate in order
to allow the outlet tempersture to level out
properly. The transition from power to isothermal
operation is accomplished at a rate of 10°F/sec

for the first 100°F, 5°F/sec for the final 100°F,

The radiator has been cycled 20 times, and the,

cycling program is continuing.
The ait pressure drop datal? obtained for York

radiator No. 9 substantially agree with the data

for York rodiators Nos. 1 and 2, and the heat
transfer data substantially agree with the data!®
for Cambridge radiators Nos. 1 and 2. Thus far
the NaK pressure drop has increased 119% above
the initial level (219% of the initial value). This
radiator is the first 500-kw radiator of the revised
design!! to be tested, and it is identical to York
radiator No, 7 illustrated in Fig. 1.4.8 of the sub-
sequent section, ‘‘Small Heat Exchanger Tests,"’
of this chapter, _ ' .

ORNL  heat exchangers Nos. 1 and 2, type
IHE-3 (ref. 12), were removed from intermediate
heat exchanger test stand B and are undergoing
metallurgical inspection. Preliminary results of
this inspection reveal that heat exchanger No. 1
(NaK-to-fuel heat exchanger) failed in the hot end

~ between the header weld and tube bends (see

Chap. 3.4, *“Welding and Brazing Investigations®').
Severe corrosion was evident on the fuel side of the
tubes, Five tubes had obvious cracks in the
tension. side. The frequency and severity of the
cracks were most pronounced in the end row of
tubes where the distance between the tube bends

and header was the shortest. A maximum of 15 mils

of mass-transferred deposit was measured on the
NaK side of this heat exchanger {(Fig. 1.4.7). No

evidence of mass transfer was detected in heat ex-
- changer No. 2 {fuelto-NaK heat exchanger), Mass
~ transfer buildup would account for the NaK pres-

sure drop increase experienced during nonisothermal

operation. Pressure drop was not measured in the

individual heat exchangers. However, if it is
assumed, on the basis of the mass-transfer evi-

dence, that all the pressure-drop increase occurred
~in heat exchanger No. 1, the total pressure drop

 

10y, €. Amos, ANP Quar, Prog. Rep. March 10, 1956,
ORNL-2061, p 54. B .

- VIE, R. Dytko et al,, ANP Quar. Prog. Rep, March 10,
1956, ORNL-2061, p 52.

12, D. Pedk et al., ANP Quar, Prog. Rep., Dec. 10,

PERIOD ENDING JUNE 10, 1956

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Fig. 1.4.7. Maximum Mass Transfer Found in
NaK Circuit of NaK-to-Fuel Heat Exchanger (ORNL
No. 1), Type IHE-3, Which Operated- 1825 hr in
Intermediate Heat Exchanger Test Stand B, (Seeret

with-eaption)

increase for this heat exchanger was approximately

180%.

Black, Sivalls and Bryson Nos. 1 and 2 heat ex-
chdngers, type IHE-3, were installed in test stand

B, and test operotions were resumed. The stand is

currently operating on a constant-power endurance
run to provide corrosion information for comparison
with the data for ORNL heat exchangers Nos. 1 and

2. During shutdown of the test stand, instrumenta-

tion was added to allow separate measurements of
heat exchanger pressure drop. After 362 hr of
power operation, the NaK pressure drop has in-
creased 140% in heat exchanger No. 1 (NaK-to-fuel).
No increase has occurred in heat exchanger No, 2,
Cambridge radiators Nos. 1 and 2, which have

~ operated for 820 hr under nonisothermal conditions, -

have experienced a NaK pressure drop increase
of 84%. A tcbulation of the NaK pressure drop-
variations that have occurred in the various heat
exchanger tests is presented in Table 1.4,3.

Construction work on stand C, to be used as an
ART prototype radiator test stand, is continuing.

57

 
 

 

ANP PROJECT PROGRESS REPORT
Small Heut Exchanger Tests P stand B was shut down when the ORNL heat ex- u

L. H. Devl ind : J. G, Turner!3 changer No. 1, type SHE-2, had operated success-

fully £ total of 2071 hr, and th hat hanger
A summary of -small heat exchanger (SHE) test vily for adotal o f, and The heal exchang

stand operanon is presented in, Tabie 1. 4 4, Test '30n assignment from Pratt & Whltney Aircraft. ..

 

TABLE 1.4.3. SUMMARY OF NaK PRESSURE DROP VARIATIONS THAT HAVE OCCURRED
~IN HEAT 'EXCHANGER TESTS

 

Maximur_n NaK  Minimum NaK  Pressure Drhop ~

 

T;st'Unif ' : . Test PTe':"d | Operufing‘ Temperature Temperature : . Change _
S . Stand - —::; Condition _ in System in Test Unit (%, based on
' , r A - (°F) (°F) j,niiia! level) . .
ORNL radiator No. 3~ SHE-B 200  Nonisothermal 1500 1270 '
York radiator No. 4~ SHE-B 140  Nonisothermal 1270 ms 4 |
B | - R 580 Thermal cycling 1270 1005 : .‘8._5, - -
90 Nonrisothermcl 1255 1020 ‘ 8.2
50  Nonisothermal 1500 1300 EREN:
20 isothermal - 1270 1270 : 3.2
- York radiator No, 9 IHE-A 174 Nonisothermal 1500 1100 - 107
o * 72 isothermal 1200 1200 -0 o
160 Isothermal 1400 ' 1400 -16
25 Isothermal “1500 1500 , 0 -
30 Isothermal 1400 1400 0
512 Thermal cycling 1500 1100 28
Cambridge radiator IHE-B 552 Nonisothermal 1600 A 1100 56
Nos. 1and 2 274  Nonisothermal - 1600 1100 28
ORNL heat exchanger IHE-B 360  Nonisothermal 1600 1275 st '
No. ']o type |HE"3 . NaK dumped and leop cooled 62
to room temperature _ o
448  Nonisothermal 1600 12752 -
Black, Sivalls . IHE-B 274  Nonisothermal 1600 1275 140
and Bryson heat '
~ exchanger No, 1
Black, Sivalls IHEsB 274  Nonisothermal 1600 1100 0

and Bryson heat
exchanger No. 2

 

“Percentage shown is within experimental error.

bDurmg this run the cold trap and the plugging indicator were inoperative. Oxide contamination level of the NaK
was below 150 ppm in all other cases, as determined by a plugging indicator. '

€This culculnted value was based on the assumption that all the heat exchanger pressure drop chonge occurred in ‘ e

heat exchanger No. 1. , . U

58
 

ey A g 5

 

PERIOD ENDING JUNE 10, 1956

TABLE 1,4,4, SUMMARY OF SMALL HEAT EXCHANGER TEST STAND OPERATION

 

 

Hours of Total Number of
Test Unit® Nonisothermal Hours of Thermal Reason for Termination
Operation Operafionb Cycles
" Test Stand B -
ORNL Heut exchanger No. 1 . 1041 2071 36 Test completed
{type SHE-2) '
Process Englneering heat exchanger : 0 120 o Test continuing
No. 1 (type SHE-2)
York radiator No. 4 748 1356 31 Test completed
(medificction 2) v
Yorl.('.irodiotor No. 7 (re_vised design) - 0 - '120 Test continuing
. Circulating cold trap No. 1 995 - Test completed
(4 in. in diameter)®
Circulating cold trep.“No. é 216 Test continuing
{4 in. in diameter, modification 1)
Test Stand C
ORNL heat exchanger No. 2 102 634 15 Removed when restriction
(type SHE-2) to fuel flow developed
Struthers-Wells heat exchanger No, 1 4 280 2 Test terminated because
{type SHE-2) of restriction to fuel flow
York radiator No. 5 106 914 17 Test continving
(modification 2)
Circulating cold trap No, 4 914 Replaced by new cold trap

(4 in. in diameter)

 

“lncludes only units tested durmg fl'us report permd. )

- bFer fesfs in progress ‘lhe total operating fime ls shown as of May 15, 1956. |

:and York radlafor No. 4 were removed from the

stand for metallurgucoi exammohon.

~Heat hcnsfer data, fuel pressure dl'op data, and

' irNaK pressure drop data for the heat exchanger

-~ _.were in substantial agreement with data previously

. reported'? and did not change throughout the test,

. The radlotor heat transfer data’ substantially agreed

o mth the datc for ORNL radiater No. 3, and the air.

.pressure ‘drop data: subsfanhally agreed with the

" data' for Cambridge radiators Nos, 1-and 2, afl .
previously reported.19 The radiator NaK pressure’

drop increased approximately 30% during the test
operation.  Much of the test program on these

N c"l'hu; type of cold frap previously referred to es 80-ga| system.

units consisted of thermal cycling operahoris. A

- complete cycle consisted of 16 hr of power opera-

tion, with maximum and minimum NaK_temperatures

of 1275 and 1005°F, -respectively, and 8 hr of

-|sothermal operation at 1285°F, The rate of NaK
‘temperature change durmg the transition from one
.condition to the other was approximately 7°F/sec._

The heat exchanger log-mean temperature difference

changed from 0°F during isothermal operuhon to

74°F during power operation,

York radiator No. 7 (Fig. 1 A, 8) und Processl
Engineering Co. heat exchanger No, 1, type SHE-2,
were installed in stand B, and test operations were

59

 
 

 

 

 

ANP PROJECT PROGRESS REPORT

. UNCLASSIFIED
- PHOTO 23875

 

Fig. 1.4.8. York Radiator No. 7 (Revised Design).

resumed. A new, 4-in.-dia cold trap was installed
in this loop. This cold trap is identical to the
previously used cold traps except that the copper
cooling coif has been replaced by a stainless steel
coil and additional thermocouple wells have been
provided to determine the temperature profile inthe
cold trap. Operation thus far has been devoted to
cold-trap evaluation under simulated ART operating
conditions, ‘ ' _
Operation of small heat exchanger test stand C
was terminated cfter 634 hr when ORNL heat ex-
changer No. 2, type SHE-2, developed a high re-
sistance to fuel flow during power operation. This
heat exchanger was replaced by Struthers-Wells
Corp. heat exchanger No. 1, type SHE-2, and test
operations were resumed. When heat transfer con-
ditions were initially established, this heat ex-
changer also began to develop a high resistance
to fuel flow, similar to that experienced by ORNL
“heat exchanger No. 2. At this time, @ leak de-
veloped in the resistance heater and the stand was

60

shut down. The resistance heated section was
removed for metallurgical examination,
Examination of ORNL heat exchanger No. 2
disclosed extensive buildup of metal particles
throughout the fuel side of the heat exchanger that
would account for the marked increase in resistance
to fuel flow. It is felt thot examination of
Struthers-Wells heat exchanger No. 1 will reveal
the same condition but to a lesser degree, ‘Evidence
of mass transfer and self-welding in certain arecs

of the resistance heater indicated that extreme hot

spots had occurred in the heater, This would

account for the buildup of metal particles in the

heat exchanger. A section of the resistance heater

at which severe overheating occurred is shown in

Fig. 1.4.9. The metal particles which built up on
one of the heat exchanger tubes are shown ‘in
Fig. 1.4.10. : L

A new heater design has been completed, and o
heater is being fabricated. The test stand is
currently being operated as a NaK loop to obtain

@

of
 

 

 

 

- PERIOD ENDING JUNE 10, 1954

UNCLASSIFIED §
“PHOTO 26114 8

Fig. 1.4.9. Section of Resistoance Heater Removed from Small Heat Exchanger Test Stand C Showing

Effects of Overheating.

cold-trap-evaluation data. Upon receipt of a new

‘25-tube’ small heat ‘exchanger (type SHE-?),“ the |
Struthers-Wells heater exchanger No. 1 will be-

replaced, ‘and- the test stcmd wufl be put back into
full operanon.

Cold Trap Evuluaflon in Hect Exchanger
Test Stands

J. C. Amos

| Six sumllar 4-m.-dm cwculaflng cold trups huve"'
been operated in the various heat exchanger test .
stands. These cold. traps wete previously referred
to as 80-ga|-sysfem curculuhng cold traps.'5 - l_t

 

_ ,MJ. C.','.-A'mbs, .’L._H. ,'De\i.iin,“ rund J; G.. Turnér,'_ANP
Quar, Prog. Rep, Dec, 10, 1955, ORNL-2012, p 49.

15k, A, Anderson and J. J. Milich, ANP Quar., Prog.
Rep: Sept, 10, 1955, ORNL-1947, p 54.

has been found that care must be exercised duwring
initial cold-trap operation to prevent oxide plugging
during main system heatup.  The original design
specified copper cooling coils which necessitated

-maintaining the cold-trap temperatures below 700°F
ot all hmes to prevent excessive oxidation of the
. -copper. This meant that at system temperatures’
above  800°F, with high contaminant saturation

levels (800 to 1100°F), ‘it was possible to pre-
cipitate material in the economizer which tended
to settle out and plug the cold-trap inlet line,

One cold trap.with a stainless steel cooling coil
has been tested. Thls trap was brought up to
1200°F with the main system: The temperature of

the trap was then gradually reduced by supplying

maximum - air cooling and reducing the cold-frap
NeK flow. By using this method of startup, no
indication of serious plugging was observed.” No

61

 
 

 

 

 

ANP PROJECT PROGRESS REPORT

 

 

 

Fig. 1.4.10. Fuel-Side Wall of Tube Removed
from ORNL Fuel-to-NaK Heat Exchanger ORNL
No. 2, Type SHE-3, Showing Mass-Transferred
Metal Particles. (Seeret~withcaption)

difficulty with cold-trap circvit plugging has been
encountered with any of these cold traps during
routine loop operation after initial heatup by the
procedure described here.

In all cases, the cold traps have been capable
of reducing the NaK contomination level of a
77-gal NaK system to less than 100 ppm and, in
several instances, to below 50 ppm, as determined
by a plugging indicator. All cold traps tested on
18-gal NaK systems have been capable of reducing
the contamination to below 50 ppm and, in most
cases, to below the sensitivity of the plugging
indicator (<30 ppm).

A longer period of time is required to clean up a
77-gal system than an 18-gal system. The cleanup
time appears to be related to the surface area in
the system, and it is appreciably reduced when a
large percentage of the system surface area has
previously been cleaned at maximum system tem-
perature, The rate of cleanup appears to increase
with an increase of flow through the cold trap. The
available cold-trap cooling limits the NaK flow
through the cold traps to approximately 1.5 gpm
while maintaining a 300°F cold-trap outlet tempera-
ture with a system temperature of 1500°F, There-
fore, the effects of cold-trap flow rates of above 1.5
gpm have not been studied. The minimum attainable
contamination level has been found to be dependent
on the minimum attginable cold-trap temperature.

62

However, these two variables are not necessarily
closely related. In some instances the equilibrium
plugging temperature of the plugging indicator may
be as much as 400°F above the minimum cold-trap
temperature, while, in other instances, " ihls dlf-
ference may be less than 100°F,

At operating temperatures above 1350°F the con-

~ tamination level has been found to increase rapidly

when the system temperature is increased to beyond

‘the temperature at which the system was previously

cleaned. Sudden variations in system operating

" conditions, such as flow stoppage or q'brupf'flpw

changes, also appear to increase the contamination
level. After a system has been cleaned at the
maximum system temperature, there is slow con-

~ tamination buildup if the cold trap is turned off.

However, this increase appears to become slower
with system age.

Cold-trap evaluation tests are contmmng on the
heat exchanger stands in an attempt to obtain
quantitive data regarding the effects of cold-trap
NoK flow, cold-trap temperature, rate of cooling,
etc, Various methods of cold-trap operations are
being investigated under simulated ART operating
conditions to establish an operating procedure for

the ETU and ART cold traps.

AUXILIARY COMPONENT DEVELOPMENT
D. B. Trauger J. J. Keyes

Dump Yalve
L. P. Carpenter

Prototype dump valve No. 2 was tested to de-
termine its adequacy for use as the dump valve
between the reactor and the fuel fill-and-drain tank
in the ART, The variables investigated were seat
leakage, self-welding of seat materials, and galling
of the stem on the stem guides, Seat leakage
requirements are less than 2 cm3/hr with a 90-psig
differential pressure across the valve. The valve
temperature durmg power operation is expected to
be 1250°F.

The seat matericls were the some as those for
prototype valvelé No. 1, that is, Kennametal 152B
(64% TiC~30% Ni~6% NbTaTiC, for the plug and
Kennametal 162B (64% TIC—257 Ni-5% Mo—6%
NbTaTiC,) for the seat ring. Prototype valve
No. 2 differed from valve No. 1 in that the stem-to-
stem-guide clearance was increased from 0.001 to

 

16} . P. Carpenter, J. W. Kingsley, and J. J. Milich,
ANP Quar. Prog. Rep. March 10, 1956, ORNL-2061, p 60-
 

 

 

 

it

0.005 in. on the diameter and @ 30-deg conical
bevel was ground on the seat ring,

The valve was tested for seat leakage at 1200°F
in the fuel mixture (No, 30) NaF-ZrF +UF  (50-46-4
mole %). The minimum leakage rote obtamed was
2.13 cm3/hr, with a 5-psig ‘differential pressure
across the seat and a valve stem thrust of 1200-1b
force, which gave an estimated seating pressure
of 12,000 psi. - Failure of the seat ring braze
occurred when the valve was opened after having
been closed for 72 hr with the above mentioned
force applied to the Stem.__ Examination of the seat
ring and plug revealed that self-welding had
occurred, and the seat ring had pulled from the

valve at the brazed joint.  The brazing technique

has been revised for future valve fabrication.
There was no evidence of galling of the stem
guides, such as that. which occurred in the tests
of prototype valve No. 1,. |

Prototype valve No. 3 has been recewed from the
manufacturer for testing. The flame plating on the
stem cracked during a stress-relief operation, and
the leakage past the seat, as measured with
helium, increased. The valve will be accepted,
however, for seat material testing.

Cold Trap and Plugging ln:hcator
R. D. Peak!”

The cold-trap evaluation test stand, prewously
described, 18 has been operated only intermittently
because of continual equipment difficulties. Two
different types of plugging indicators were tested,
and the results correlated poorly with the data
obtained with the Argonne sampler, described

previously, 19:. Two methods have ‘been’ fested for
‘coolmg the ‘cold. trap - ‘with ‘air, but it is ‘opparent
that air: coohng clone is msuff;clent. A ‘combina- -
- tion of ‘air: with_water |n|ect|on und finally, full -
-+ water flow should achleve the - Iow-temperuture;;'

:..';;.ZrF . The helium ‘and- ZrF, vapor then passed

R '-cold-trap operaflon des;red for the ART. systems-}l_”f-jthrough a section “of. y-m.-dm tubing extending

S Other mefhods of °°°l""9 are belng mveshgcted through ‘@ hole in- fhe center of a stamless steel

The second cold-trap evaiuuf!on ‘fest stand. W"s::flf-fblock whnch was 6 in. in.diometer and 10 in,. Iong.

completed and will ‘be placed in operation s°°“°"i“'i-i'i_The _entrance face of the stainless jocket was
~ heated by a Colrod heater and the rest of the trap,
't'___',except ‘the exit face, .was msulofed o prowde o
“linear - thermal gradlent. - The ‘temperature of -the

““tube_at the point where the ZrF, first premputated _

-~ This_test. srond is -much . more -versatile than the -
. first stand, in that 'several means have been pro--
o ._,_.vnded for sampling and for. recontommanng the NaK' =
o -wnh the various |mpurttles of mterest. j-jf

 

nOn nssignment from Protf & Whltney A;rcraft. S ;-‘."_:.7
184,73, Milich and R. D, Peak,-ANP Quar, Prog, Rep,

Marcb 10, 1956, ORNL-2061, p 62.

%A, s, Meyer, W. J. Ross, and G. Goldber ANP

Quar. Prog. Rep. Sept. 10, 1955, ORNL-1947, p 1 2

PERIOD ENDING JUNE 10, 1956

Zirconium Fluoride Yapor Trap
M. H. Cooper20

- The program for the development of a trap for
zirconium fluoride vapor has included an investi-
gation of high-temperature adsorbents for ZtF , a
determination of the plugging temperature of ZrF

in helium in equilibrium with the fuel as a function
of the fuel temperature, and the testing of prototype
traps. High-temperature adsorbents were tested by

~passiig helium partially saturated with ZrF,
- through a l-in.~dia pipe packed for 24 in. with the

test material. The temperature of the test section

was controlled by clamshell heaters. From the

trap, the helium flowed first through water bottles
to ‘collect any residual ZrF 4+ through wet-test gas
meters, and then to the atmosphere. The results of
these tests are summarized in Table 1.4.5. Alumina
is the most promising adsorbent tested, while UF,,

which has not yet been tested, may also prove
satisfoctory, since ZrF, and UF, form a solid
solution with a high melting point. Charcoal, when

~used in a trap, enhanced the formation of ZrF

nuclei, which grew into large individual crystals,
Charcoal thus may be a svitable packing for a
thermal trap, since the formation of individual
crystals does not-plug the trap. Suitable operating

“conditions have not yet been established for char-

coal traps,
- The saturation temperature of ZrF, in helium

in equilibrium must be known in order to design

the fuel off-gas system of the ART for tempera-
tures which will prevent the precipitation of
ZrF A schematic drawing of apparatus for de-

3 fermmmg the plugging temperature of ZrF, in
~ helium- pcssmg ‘through - '/-m. tubing is shown in
~“Fig. ‘1.4 In each run, helium was bubbled
~through - “the - ‘molten fuel mixture at ¢ measured
~'temperature, and. the hellum became saturated with

was assumed to be the p!uggmg temperature of
ZrF, d correspondlng to the fuel femperature.- The

 

200, assignment from Pratt & Whitney Aircraft. - -

63

 
 

 

 

ANP PROJECT PROGRESS REPORT

TABLE 1.4.5. SUMMARY OF TESTS OF ZrF‘ ADSORBENTS

 

 

Ad Fuel Trap Helium - Weight of Zr
MS:TT Temperature  Temperature  Flow Rate o r:r: 4 Not Trapped Comments
ateria : erate :
©Rn R hos) " (mg)
NeF 1300 1300 0.0573 260 5 NaF pellets melted in front part
' ' 0.105 119 5 of trap because of formation of
0.214 98 67.1 ZrF 4-NaF eutectic
1320 1400 1o 8007 0.0520 96 18.4 NaF peliets melted in inlet -
1200 to 800% 0.0910 112 section; contents of water trap
of second run lost because of
leok in water trap '
ALO, 1320 1400 0.0568 96 28.4  Plug in off-gas line believed to
‘ 0.0930 112 46.0 be Al F3 from HF reaction with
A|203: water found in helium
, supply during this run
1400 1400 10 800° 0,098 480” SHll in operation; no sign of
plugging '
CaCl, 1315 1285 to 13355  0.0978 40 CaCl, melted and hence is un-
satisfactory
Charcoal 1315 1385 to 1415° 0.0978 40 62 Inspection of trap showed that
1308 to 1321° 0.0978 66 703 ZrF4 had precipitated in large
1110 to 1160° 0.0978 96.4 1496 individual crystals; plugs

formed in offegas line from trap
in all tests

 

2Temperature gradient afong the trap.
bHours operated as of May 15, 1956.
€ Tomperature varlations during test.

experimental plugging temperatures obtained as a
function of fuel temperature, as well as the satu-
ration temperatures predicted from both ORNL and
Battelle Memorial Institute vapor pressure data, are
given in Fig. 1.4.12. - The experimental plugging

.temperatures agree to within 5% with the tempera~

tures predicted from the vapor-pressure data. The
experimental plugging temperatures and the pre-

dicted temperatures differ because the helium was

not actually saturated with ZrF , in the experiments,

Two prototype vapor traps, of the type described
previously,2! have been tested. Prototype trap
No. 2, which was connected to the fuel pump of
intermediate heat exchanger test stand B and was

cooled by natural convection and radiation, plugged

 

21, 3, Milich and J. W. Kingsley, ANP Quar. Prog.
Repo DeCQ 10. 1955; ORNL'20]2, P 60-

64

with entrained fuel after 12 hr of operation. The
pump off-gas line had been installed to closely
approximate one designed for the ART. The fuel
entrainment indicated the need for redesign of the
off-gas line. Prototype trap No. 3, which is water
cooled and mounted on a fuel sump through which
helium is bubbled, was placed in operation on
April 27, The pressure required to maintain the
helium flow at 3.0 liters/min increased from 3.0 psi
to 5.0 psi in 50 hr, There was no further rise in
the pressure required to maintain the flow until
the test was terminated after 420 hr, Two layers
of ZrF, were found in the trap entrance just up-
stream of the tube header. The first layer was
rather dense and amorphous, and the second layer
was more porous and crystalline, A total of 984 g
of ZrF, was found in the layers, which totaled
 

 

 

o

at

PERIOD ENDING JUNE 10, 1956

UNCLASSIFIED
"ORNL~LR—-DWG 4959

 

f—— INSULATION

 

CALROD HEATER—

CALROD HEATER

 

 

CLAMSHELL
HEATERS

 

 

 

THERMOCOUPLE
WELL —]

 

La——STAINLESS STEEL JACKET

—=—TO WET-TEST GAS METER

Yo-in. DIA TUBING

 

 

 

 

 

 

HELIUM-\

 

 

 

 

ZrF,-CONTAINING MIXTURE

 

\\\\\\

R T R T A T T =SS

USRS S R R

\\\\\\\\\\‘\\ R AN RN N AN AR N

R Y, AN \\\\\\\\\\\\\_——*FIRE BRICK
> .

xxxxx

 

Fig. 1.4.11. Apparatus for ZsrF, Tube Plugging Test.

- SEORETe
ORNL—LR~DWG 14960

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1600
& 1500
2 ® RESULTS OF TUBE-PLUGGING TESTS
Q :
.z 1400
uy
S
N
& 1300 i L2
W BMI VAPOR PRESSURE DATA .~
g ' L2
g L
@ // [ .
é o | ## ORNL VAPOR PRESSURE DATA ~ | .~ -
g 0 A | |
g 1 -
. 9 1100 .
1000 s e . —
o 100 | 1200 _ 1300 - 1400 1500 1600

- FUEL TEMPEhATURE F)

Flg. '|4 12., Plugging Temperoture of Zl'F4 in
Hehum as a Funchon of Fuel errure Temperature.

23’ in. in thrckness. The werghf of ZrF, was only
30% of the amount that would have vaporized if
the helium had been saturated with ZrF,. The
dense deposn found in this trap near the inlet
renders the design unsuitable for the ART,

OUTER CORE SHELL THERMAL
STABlLlTY TEST
R. Curry A M. Sm:th

The first test of ihe one-quarter-scale model ofk

- the ‘lower half of the outer core shell23 was termi- -
_nated ‘ofter - the completion of 57 thermal cycles,‘-
- because a sodium leak developed in the Ioop piping
. near the-test piece. "It wds decided fo partially -
B 'examme the shell before continuing the test. The
. center island was cut loose and lifted out of the
‘holsing | assembly to expose the inner surface of

the ';h_ell f_qr,i_nSpectio'n; “A,rdye' ;beck ,bf_ 1he inrrer

 

220:1 nssignment from Prcm & Whifney Aircraft. o
23p, W, Bell, ANP Quar. Prog. Rep. Sepz. 10, 1955,

-’.‘,r:ORNL-1947. p

65

 
 

 

 

 

ANP PROJECT PROGRESS REPORT

sutface revealed no cracks or other flaws. Xeray
photographs taken through the shell wall and the
thick outer housing wall likewise did not indicate
cracks or flaws, but it is judged that this inspec-
tion technique was not sufficiently sensitive to

‘reveal hairline cracks, if such existed, on the

exterior surface of the core shell.

The internal contour of the shell was mapped by
centering the shell on a lathe and measuring 264
radii. For these measurements, 15 axial stations
]/2 in, apart were designated, The radius was then
measured for each 30 deg of rotation from an
assigned zero position at the eight stations nearest
the smail-diameter end. At the seven stations
nearest the large-diameter end, the radius was
measured for each 15 deg of rotation from the zero
position. The measurements obtained were com-
pared with measurements taken after the initial
machining of the model. It was found that the
average radius at each station in no case differed
from the original radius by more than 0.012 in.
At no station did the spread from minimum to
maximum radius exceed 0,021 in., and the average
spread for the 15 stations was 0,014 in. The
““crests’’ and ‘‘valleys’’ in a plot of radius vs
rotation indicated that the basic deformation at
each axial station was from circular to elliptical
or pear shaped. This comparison between final and
initial shape is invalid to some degree, because
the initial measurements were taken before the
shell was welded at euch end and installed into
the outer housing.

The shell is now back in the test rig and it is
to be cycled a total of 300 cycles or to failure,
Parts are being fabricated for a second 300-cycle
test on another core-shell model,

FORCED-CIRCULATION CORROSION AND
MASS-TRANSFER TESTS
W. B. McDonald -

Fused Salts in Inconel and Hastelloy B

J. W. Kingsley P. G, Smith
: A. G, Smith24

Nine forced-circulation loops, seven fabricated
of Inconel and two of Hastelloy B, were operated
with fused salts as the circulated fluids during

 

240, assignment from Pratt & Whitney Aircraft.

66

this quarter, The operational data for these loops
are summarized in Table 1.4,6. Eight of these
loops were electrically heated and one (loop 4935-6)
was heated with a natural-gas furnace. Two of the

Inconel loops circuloted a new fuel mixture (No. 70)

NaF-ZrF ,.UF, (56-39-5 mole %), and the Hastelloy
B Ioops cnrculated the fuel mixture (No. 107)
NaF-KF-LiF-UF, (11,2-41-45.3-2.5 mole %). = All
other tests were made with the fuel mixture (No. 30) -
NaF-ZrF ,-UF , (50-46-4 mole %).- Other parameters
mvestlgoted were operating temperature of the
fluid, surface-area-to-volume ratio, temperature
differential, and Reynolds number, When com-
pleted, the results of metallurgical examinafions of
the loops are reporfed in Chap.-3. Dynomlc
Corrosion Studies.”’

The first Hastelloy B loop failed duri_ng attempts
to put it into operation. Therefore no corrosion
data were obtained, but valuable mformahon per-
taining to the engineering, fabrication, and opera-
tion of Hastelloy B loops was obtained. As o
consequence, the second Hastelloy B loop operated
satisfactorily for the scheduled 1000 hr,

The one gas-fired loop was terminated at 8319 hr
as a result of a leak in the cooled section, This
leak occurred while the cooled section was being

 thawed out aofter a freezeup which occurred when

the pump-drive unit failed.

Liguid Metals in Inconel and Stainless Steel

J. W, Kingsley P. G. Smith
A. Ge Smith

Twelve forced-circulation loops were operated
with sodium or NaK as the circulated fluid, and one
was operated with water, Ten of these loops were
electrically heated and three were heated by
natural-gas furaces. The operational data for .
these loops are summarized in Tab!e 1.4.7. The
three loops that are heated by gas furnaces are -
being operated as life tests.” Small centrlfugal ,
pumps (model LFB) are used on the gas-heated

loops; the other loops that circulate liquid metals -

have electromagnetic pumps. The model LFB
centrifugal pump was used on the loop that circu-
lated water. :

When completed, the results of metallurglcal
examinations of the loops are re_ported in Chup. 3.1,
**Dynamic Corrosion Studies,”’

 
 

L9

 

 

A o 4 T CM o T .M . R

’fiéi;s 1.4.6. SUMMARY OF OPERATING CONDITIONS OF INCONEL AND HASTELLOY B LOOPS IN WHICH
U | FUEL MIXTURES WERE CIRCULATED

 

*Time with '.fii"np.e_rafuro dif#ér_enfi&l ‘elé't‘nlyjli’s‘héd'.

 

Co A : T Maximum Maximum Op & '..

L ; Gl o pproximate emperature Fluid . Tube Wall erating . :
Loop ; oop R Circ'inldtéd'.Fluld.f "~ Reynolds Differential o ahe T4 . Time* Comments
‘ No.’«‘w' Mahna! b bt T _ o Temperature  Temperature ‘

74259 inconel NoF-ZrF‘-UF4 - 5750 1300 1600 1700 3000 Life test; terminated
LT (50_46,4 mole % i ' at 3000 hr to make

Na,,’,‘o . ‘ T roomforuddifuonoi .. :

L S o ‘ tests )
"':"’C “v‘ : ' .‘ s \ J ' ‘ . ‘ ‘ ! . " ' . o !

10,000 200 ']500‘\:.‘,‘.7‘ . 1580 o , 786 . Loop‘vlnl_.urrlef._fém."’_"ti\;'nas:

T e T that of Sid'ndord'l'pdp'

  

  

~ Nd F'K F-LI F-UF - 10,000 200 1500 ’ Three vnsuccessful
ey (1].2*4‘-4‘5-3-2¢5"1‘*7; - . ' aftempts were made to
',:<.’  L mole 9’), No. 107 _ - | ‘ operate this loop

12 Hastelloy B. -

13 Hastelloy B ‘Same;grs;ui:‘:_t_:v‘e_‘ 710,000 200 1500 1555 1000 Trouble-fras operation
T e ' | terminated on schedule
-14  inconel - NaF-ZrF <UF, " 10,000 200 1500 1600 1022  Terminated on schedule
S (seeases mols %),.-‘: - - ‘ SRR

No. 70 L :
-15: ! Inconel &:m. as obov'a © 7,000 200 1640 1700 In test
-16 laconel NaF-ZeF, SUF, 10,000 145 1540 1595  Intest
S (50-46-4n1o|a %),

.'\ _No. 30 ‘ o

-17  Inconel . Same.asabove . 5000 150 1500 1555 In test

49356 Inconel ' Same os above 8,000t 200 1500 1575 8319 Life test terminated by
Lo T 10,000. pump-drive failure

9561 0L INNC SNIGNT Aol 3d

 

 
 

89

TABLE 1.4.7, SUMMARY OF OPERATING CONDITIONS OF INCONEL AND STAINLESS STEEL

LOOPS THAT CIRCULATED LIQUID METALS AND WATER

 

FORCED.CIRCULATION

 

 

i : Moximfim
Loop Cold Loop Reynolds Temperature Fluid Operating
No. Trap Material - Number Differential Temperature Fluid - Comments
. (°F) o
(°F)
7426-6A Yes Inconel Variable Variable Sodium Plugging-indicator tests
6B Yes Inconel Variable Variable Sodium Plugging-indicator tests
-7 No Inconel >15,000 300 1250 Sodium Box of beryllium plates inserted in
hot leg; operated 1000 hr
-8 Yes Inconel >15,000 300 1250 Sodium Box of beryllium plates inserted in
hot leg; coldstrap temperature was
300°F; pump failed after 900 hr of
operation
<9 Yes Inconel >15,000 300 1300 Sodium Beryllium insert in hot leg; operated
1000 hr
-10 Yes lncone! >15,000 300 1250 Sodium Beryllium insert in_hoi leg; cold«trap
temperature was 280°F; operated
1000 hr
11 Yes Inconel >15,000 300 1350 Sodium Cold-trap temperature, 280°F; in test
12 Yes {ncone! >15,000 400 1500 Sodium Cold«trap temperature, 300'0 F; in test
=14 Yes Type 316 >15,000 300 1650 Sodium Coldetrap temperature, 260°F; in test
stainless . :
steel
=51 Yes Inconel >15,000 750 1500 Sedium Life test; 1824 hr accumulated
‘ 7432-'IA Yeos lnconel Variable Water  Operated to obtain date on flow vs
' : pump speed; surface=to-volume ratio
. of main NaK circuit of ART simulated
- 7;\4‘39‘-51  VYes ~|ncone_'.‘l ' >15,000 665 1500 - _ Noneutectic NaK | Llfo_-tes.t; 2192 he accumulated
<52 Yes  Inconel ©  >15,000 875

1600 ' Noneutectic NaK '

* Life test; 1968 hr accumulated

 

C

. .C

LY0d3IY SSTHI0Vd LID3F0¥d dNY

 

 
 

 

 

't

‘\

5!

PERIOD ENDING JUNE 10, 1956

1.5. PROCUREMENT AND CONSTRUCTION
W. F. Boudreau

"ART FACILITY
F. R. McQuilkin

Construction work is nearing completion on the
contract portion of the Aircraft Reactor Test {ART)
facility in Building 7503. Package 1 work on the
building additions, building alterations, and cell
installation was approximately 7% behind schedule
at the 92,5% completion point on June 1, 1956.
About one-half the deficiency may be accounted
for by lack of breckers and reactors for the elec-
trical switchgear. During the quarter the principal
work accomplished included final erection of the
cell tanks; erection of the stack; placement of
concrete for the main air duct, the penthouse, the
radiator pit, the blower house floor and equipment
bases, the cell encasement, and the main building
floor at an elevation of 852 ft; painting; grading;
fencing; and general job and site cleanup.

Package ‘2 work on the installation of diesel
generators and facility, electrical control centers,
and Spectrometer-room electrical and air-condition-
ing equipment progressed to the 79% completion
point during the quarter. The installation of the
electrical control centers and the diesel generators

will complete this work, The estimated date for

shipment of the diesel generators is June 25, 1956.
The contract for Package A work on the installa-
tion of auxiliary piping was negotiated with the

Y. L. Nicholson Company at a contract price of -
-$50,351,62, with completion scheduled for June 15, .
1956, unless material delivery schedules interfere,

Some of the Packdge ‘1 and Package 2 work may -
be seen in Fig. 1.5.1, which is a view from a -
location southwest of the bu:ldmg. The dark-sided
portion of the main bualdmg is the addition which
will house the reactor cell, the heat dumps, and -
, the spectrometer tunnel; The appurtenances to

= the main: bmldlng ShOW" in"this view ore, ‘from . for supporting these vnits and the radrators, and
left to right, the generator house, blower house,
~vent house, ond stack. ‘In front of the vent house
" can be seen the top portlon of the 10-ft-dia-and -

- 23-ft-deep concrete. tank that wnll Contaln the
water-submerged off-gcs piping. - o -

" The west-and north faces of the main- bmldmg'l '
_and the generator house may be seen in Fig. 1.5.2.
* On the left may be seen the newly installed 32-ft-

wide door that is required in order that the reactor
and the tops of the cell tanks may pass into the
building. The cylinder bank in the left foreground,
which served as helium storage for the Aircraft
Reactor Experiment (ARE) will be used for storage -
of nitrogen for the ART. ' In the foreground, in
front of the truck, may be seen the 1500-kva trans-
former and substation that will serve the purchased
power (TVA) portion of the power service to the
facility. The openings in the generator house,
on the right, will receive the radiators for the
five diesel generators that will have 300-kw con-
tinuous ratings and will supply the locally gener-
ated power service to the facility. In Fig. 1,53
may be seen the pressure vessel and water tank
that make up the containment cell. The 24-ft-dia,
36-ft-high pressure vessel is within the 30-ft-dia
water tank. The top portion of the water tank had
been removed and was sitting on the main floor
when the photograph was taken. Although the
pressure vessel manhole is 5 ft in diameter and
will pass most of the items to be installed within
the vessel, the contractor will remove the top head
to provide the required access for the reactor.
The pressure vessel was field stress-relieved and
hydrostatically tested at 300 psi during the quarter.

Inspection of the welded joints between the 24.in.-

dia junction panel sleeves and the 4-in.-thick shell
walls revealed o failure of a joint in the heat-

“affected zone of the sleeve base metal. Therefore
-all eight of the one-piece sleeves were replaced
- with two-piece units and a revised joint design

and welding procedure was used. The field stress-
relief procedure and the hydrostatic pressure tests

- w:il be repeated

“The coricrete penthouse ad|ocent to the cell is.

‘shown in Fig. 1.5.4. This structure will house.

the NaK- pumps, pump motors, and the mechanisms

~will _also house the NaK piping to  the main air
“duct and to-the radiator pit below, The special
;‘equnpment room -is dlrectly ‘below and covered by
fhe roof plugs shown in the open doorway.

“Design ‘work continued on Package 3, which

- ‘l,comprsses ‘the process equipment, process piping,

etc. to be installed by ORNL, The design will
probably be completed in July 1956.

69

 
0L

R R

UNCLASSIFIED
PHOTO 17604

 

W
3
X
]

s

 

 

L3043y SSIII03d 133rodd dNV

 
 

 

 

 

 

 

 

 

 

 

"

”

a

-

PERIOD ENDING JUNE 10, 1956

UNCLASSIFIED
PHOTO 17548

 

 

 

 

 

 

 

 

 

 

 

 

Fig. 1.5.2, View of West and Norfh Faces of Buiiding 7503 and fhe Genemtor House. Photograph

taken May 4, 1956,

ETU FACILITY ;-

M. Bender W, R Osborn
G. D. Whnmon A

- Design work on the facuhty in Bunldmg 920]-3‘4'_
for the. Engineering Test Unit {(ETU) is well under

way, and some preliminary ‘drawings of service

piping have been: completed, Line and equipment. |
heating requirements have been established, and
‘the design of the NaK: piping and structural sup-:

ports has been started. Preliminary layouts have

been completed for the heat-dump system, includ.’ -
“ing ducting, louvers, barrier doors, and the blower,

The building structurcxl ‘modifications were com-

‘pleted, but additional design information has shown
the need for further support for the floor structure.

The lube-oil pumps, the NaK tanks, the ETU

- fuel dump ‘fdnk,'-an‘d"miseelluneoows,?Ificonel ma-
terials have been ordered. A decision to expedite

assembly of the ART with respect to ETU com-

~pletion -has resulted in procurement of uddmonal

items of equipment so that ART construction can

- proceed concurrently wnh ETU operatlon.

ART-ETU REACTOR PROCUREMENT AND
ASSEMBLY '

_ W R, Osborn
G.D. Whlfman )

Detmled study is under way on - methods for
assembling the reactor. A preliminary procedure
for assembling the reflector-moderator has been
completed, and work is continuing on procedures

M. Bender

71

 
 

ANP PROJECT PROGRESS REPORT -

UNCLASSIFIED.§. - . - e
PHOTO 17605 8 :

 

 

~ Fig. 1.5.3." Pressure Yessel and Water Tank That Make Up the Reactor Cell. Photograph taken
May 11, 1956. L - SR

 

72

 
 

 

 

 

PERIOD ENDING JUNE 10, 1956

UNCLASSIFIED
PHOTO 17547

- Fig. 1.5.4. Concrete Penthouse Adjacent to Reactor Cell. Photograph taken Mdy 4, 1?56.

for assembling other parts of the reactor. Of par-
ticular concern in these studies are the problems
caused by weld shrinkage during assembly, Recent
data obtained from test weldments (see Chap. 3.4,
**Welding and Brazing Investigations’') have shown
the need to modify some assembly techniques
originally considered for the reactor in order to
control the dimensions of the final assembly,
Stress problems which will occur in handling com-

"ponents durmg assembly have also imposed re-
- strictions on assembly methods and are now being
‘evaluated with respect to the|r effect on the tech-

niques.’
Virtually all of the long-delwery materials for

the - reactor . ‘assemblies have been ordered, but

actual. fabncatlon of most items has been delayed
by lack of firm drawings.

,A review of the design indicated the need for
flow tests on the reflector-moderator and island

assemblies. While the extent of this test program
has not been cléarly defined, it is expected to
cause a significant. delay. in the c0mp|etlon of
the reactor assembly. |

As stated above, construchon of the ART i

 being accelerated with respect to construction of

the ETU. A co!culated risk is being ‘taken that
the ETU will not show a need for correction of
the ART reactor, and thus-the ART reactor will
be assembled as soon as possible after the ETU
assembly is completed " Parts for a third reactor

“are being procured as insurance qgamst excessive

time loss if the risk fails. .

Fabrication of the north head of 1he ETU has
been started. Some delay has occurred because of
redesign, and welding progress appears to be

" slower than anticipated. Delays in the receipt

of the sodium-to-NaK heat exchanger and some
forged parts may hold up work on this item. All

73

 
 

 

 

 

ANP PROJECT PROGRESS REPORT

the necessary information for two of the six Inconel
shells is available, and Lycoming is expediting
work on these two shells. Additional design in-
formation is still needed for the remaining shelis,
but mandrels are being rough machined for pro-
ducing them.

The beryliium reflector for the ETU is being
machined by The Brush Beryllium Co., but some
information on this item is still lacking because
of design modifications. The islond holes are
now being drilled at ORNL and the island will
be sent to The Brush Beryllium Co. for contouring
when this operation is completed.

The heat exchanger suppliers are still studying

assembly methods and techniques for fabricating
the channel sections of the fuel-to-NaK heat ex-
changers. Potentially, this equipment still appears
to be one of the controlling factors in the assembly
schedule. ‘

Orders for the boron carbide tiles and boron-
copper cermets have been placed with the Norton
Company aond the Allegheny Ludlum Steel Corp.,
respectively. A fabricator has not yet been se-
lected to form the boron-copper cermets into the
desired shapes,

Fabrication of the heovy l«in.sthick pressure
shells, which must be specially formed, wili re-

74

quire equipment that is not widely available.
Because of the special dies which will be re-
quired, the delivery of these heavy shells may
also affect the reactor completion date.

Knepp Mills, Inc., has been sele;ted as the
supplier of the lead shielding, which will be pre-
formed and bolted to the reactor. It is proposed
that only the side pieces be mounted on the ETU
because of space and time considerations. No
design drawings are available to describe the
water shield, but it is antncapated that this, too,.

will be prefabricated.

A number of jigs and fixtures for use in assembly
have been designed and are being procured, These
include fixtures for assembling the reflector-
moderator, assembly and checking fixtures for
the main heat exchangers, templates and a gage
for checking the shells, and several handlmg
devices.

An area in Building 9201-3 is now being modi-
fied for use as a reactor assembly area. The
area will be enclosed and air-conditioned, and
it will be serviced by an overhead crane. It is
anticipated that the area will be ready for use
when the first reactor parts arrive in August,

 
 

 

¥

 

PERIOD ENDING JUNE 10, 1956

1.6. ART, ETU, AND IN-PILE LOOP OPERATION

ART OPERATING MANUAL
W. B, Cottrell

An Operations Committee, consisting of repre-
sentatives of the physics, design, control, engi-
neering, construction, and operahng groups, wdas
formed for the purpose of preparing an Operating
Manual for both the ART and the ETU, Thus far,
the efforts of the committee have been directed
toward the preparation of the ART operating pro-

‘cedures, which, in first rough-draft form, is about

20% complete. Since the ETU is to serve as ¢
proving ground for the ART, the ETU operating
procedures will subsequently be adapted from the
ART operating procedures and will include other
tests which may be desired.

The bases for the operating procedures are the
experimental objectives of the ART, the reactor
design, the reactor characteristics, as determined
from simulator data, and the flow diagrams and
instrumentation lists for the various systems. The
information being used in the formulation of the
procedures is presently available in the form of
design memorandums and data sheets, reports
prepared by stoff members, and mmutes of desrgn
and Operations Commmee meetings.

The monual is to ‘provide detailed procedures,
including check lists, for all operations beginning
after the installotion of the ART equipment and
terminating with the orderly shutdown of the re-
actor. As presently outlined the manual will in-
clude a check list of equipment and detailed

'procedures for cleaning the ART, for loading the
‘NaK- systems, for heating the reactor, for loading

the sodium and fuel systems, for shakedown check-

~ing the fluid systems, for enriching the fuel , for
~ operating ‘at low power (10 to 100 w), for operating
at mtermedlate power (2 to-20 Mw), for operatmg :
-at high power, and for termingating the test,’ S
By detailed consideration of procedures,: the
_._‘Commnttee has - found sncompahbllmes in the

- performance of “some system - ‘components - and
- - limitations -in ‘the design of others which would
‘impair ‘operation during unscheduled *shutdowns.

These. d:fflcuhles fall ‘into two categories: those

'V'wh:ch will require. design changes to permit opera- - -
‘tion and those which will not require changes but

which will restrict the method of operation. Pro-

posed design modifications for resolving these
difficulties are currently being evaluated.

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

Operation of Loop No, 4

C. C. Bolta? R. A. Dreisbach’
J. A, Conlin W. T. Furgerson
C. W. Cunningham D. M. Haines!

W. L. Scott

In-pile loop No. 4, which was described in the
previous report,2 was cut from the shielding plug,
and the nose end is being sectioned for metallo-
graphic examination. This loop was operated for
a total of 501 hr, including 316 hr at the maximum
design temperature differential of 200°F and 80 hr
at lower temperature differences. The high-
temperature point in. the loop was maintained at
1500°F throughout the periods of operation with
a temperature differenticl imposed on the system.
The loop was operated under isothermal conditions
at lower temperatures during the remaining 105 hr,
while the reactor was shut down for refueling and
for other maintenance. The power density in the
loop was 778 w/cm3 during operation with the
maximum temperature differentiol, as determined
from the heat removal through the heat exchanger
and the total volume of fuel in the nose from the
heat exchanger outlet to inlet.

 Functionally, the loop operation was good. The
heater circuits, which had better insulation than
~-the insulation used for the heater circuits of loops
‘tested previously, were free of failure. The pump
_operated smoothly “throughout- the test, and con-

trols for the heat exchanger and other components'

_were satisfactory. However, 7 of the 14 nose
‘thermocouples failed in operahon, _e:ther by the
“junction coming off the pipe or by breakage of one

of the wires (usually the Chromel wire). These
failures are now believed to have been caused by

 poor mstallahon ,techmques.' Slx of the seven

 

1on osmgnment frorn Profl & Whhney Alrcraft,

2C C. Bolta et al,, ANP Quar. Prog Rep. March 10,
1956, ORNL.-2061, p 41.

75

 
 

 

 

 

thermocouple lead wires which failed had, in-
advertently, been clamped rigidly to the end of the
outer wall of the double-walled heat exchanger.
This could have resulted in strains on the thermo-
couple wire from the thermal expansion of the
pipes. The loop design has been modified to
correct this condition.

As with loop No. 3, both the bearing-housing and
the pump-sump purge outlets plugged. The bearing
purge outlet plugged five days after startup, ond
the sump purge outlet plugged four days later.
The cause of this plugging is still being investi-
gated, The fission-gas absorption traps from loop
No. 3 have been sectioned. An extensive black
deposit was found in the inlet of one and a brownish
film was found in the other. This black deposit,
believed to have come from the pump lubricating
oil in the bearing housing, is to be given radio-
assay and spectrographic analysis to determine
its composition (see Chap. 4.2, ‘‘Radiation
Damage'’). It is planned to operate future loops
with little or no purging of the bearing housing
and reduced purging of the pump sump to minimize
the probability of plugging. Previously, the higher
purging rates were considered necessary to mini-
mize hydraulic-motor-0il contamination, Experience
indicates that, although some contamination results
from operating with the purges plugged, such con-
tamination is not a serious problem during dis-
assembly.

Other major difficulties encountered were a leak
in the pump bulkhead, probably through a glass
heater or thermocouple-wire seal, and an excessive
radiation level in the cubicle after shutdown. A
new type of electrical seal for the pump bulkhead
is being investigated. The excessive activity of
the cubicle after shutdown was caused by the

deposition of material on the purge-outlet-tube .

walls during operation. Radiation levels as high
as 50 r/hr were measured on the l/‘-in. pump-purge-
outlet tube. This caused considerable difficulty
in loop removal because it necessitated a relatively
large amount of preparatory work. The design has

76

of 200°F.

now been modified so that all tubes may be pinched
leak-tight, cut, and removed remotely to make

possible the removal of the activity prior to the

entry of personnel into the cubicle. The only work
which will require entry to the cubicle will be the
removal of the air and water lines.

Loop No. 5
D. M. Haines

In-ptle loop No. 5 was completed- and mserted
in the MTR, but could not be filled. This loop
was to have operated for the duration of two
MTR operating cycles with a maximum fuel tem-
perature of 1600°F and a temperature differential
‘The improved thermocouple installa-
tion mentioned above was used in the fabrication
of this loop, and the rear section was modified to
simplify removal and to provide a second hermetic
seal to back up the seal at the intermediate bulk-

head to prevent fission-gas leakage. The loop is

being returned to ORNL for salvage.

Horizontal-Shaft Sump Pump for In-Pile Loops
W. So Kcrns

An improved prototype {Mark ll) of the horizbntal- '

shaft sump pump designed for in-pile {oop operation
completed 1000 hr of a 2000-hr endurance test at
4500 rpm and 1500°F, The test was terminated by
shaft seizure, and disassembly showed that the
seizure was caused by the buildup of zirconium
fluoride on the shaft slinger. This was the first
pump operated at 1500°F, the previous pumps.being
operated at 1400°F. The vapor pressure of zir-
conium fluoride in the fuel mixture (No. 44) NaF-
ZrF -UF ;- (53.5-40-6.5 mole %) used in this test
is 2 ‘mm Hg at 1400°F and 4.5 mm Hg at 1500°F,
The increase in vapor pressure is considered to
have been the cause of the .buildup of zirconium
fluoride on the slinger. A new pump is to be buiit

with increased clearances at the slinger.

 

30n assignment from Pratt & Whitney lAirért:l-ft. )

 
 

 

 

Part 2

- CHEMISTRY

W. R Grimes

 
 

 
 

 

 

 

 

2.1. PHASE EQUILIBRIUM STUDIES!

C. J. Barton R. E. Moore

R, E. Thoma

H. Insley, Consultant

The several methods described in previous
reports of this series were used for further phase
equilibrium studies of a variety of binary, ternary,

and quaternary systems, Although the phase

diagrams are not considered to be final in every
respect, a compilation of the diagrams of binary
systems consisting of UF, or ZrF, with each of
the alkali fluorides is presenteg here. The
striking differences observed in the diagrams for
these systems indicate a need for detailed study
of some of the complex crystal structures that
characterize certain of these materials.

The RbF-ZrF, and RbF-UF, systems now
appear to be moderately well described., A thermal
effect in the RbF-ZrF, system at 375°C that was
previously reported to be a eufectic temperature
has been shown to be a lowered inversion temper-
oture of the 2RbF.ZrF, compound; the eutectic
contains 42 mole % ZrF , and melts ot 410°C.

Solvents with compositions near 35 mole % NaF,

25 mole % RbF, and 40 mole % ZtF , apparently |

dissolve up to 4 mole % UF, at Ilquldus temper-
atures below 480°C and up to 7 mole % UF,
below 510°C. Such mixtures are of interest os
reactor fuels, _

Detailed examinations of some ternary systems
containing BeF, are being made. While extremely
low melting pomts can be obtained (the lowest
eutectic observed to date in the NaF LiF-BeF,
system melts at 318°C), it is not yet npparent
that these materials offer fuel mixtures that are

substanhully better -than those avallable in the,
ZrF (~containing systems. |

Sfudles of the phase. behavior of CeF in bmary

systems with alkali flucrides have been conflnued .

These systems are of interest because of concern

- over -the behavior of flssmn-product fluondes in _.'
high-power long-duranon reactor OPe"G*W“- -

GENERAL COMPARISONS OF THE BINARY
SYSTEMS MF-ZrF "AND MF-UF4 i

R E Thoma L .

The two porullel famllles consnstmg of bmory
systems of UF, and of ZrF, with each of the
alkali metal fluorides have been under investi.

gation here for several years, and the character-
istics of these systems are relatively well known,
In the course of this research mony people have
contributed to an understanding of these phase
relationships, and the information on phase
behavior is scattered throughout a large number of
reports in this series. Accordingly it has seemed
worth while to prepare a brief summary of the
differences and similarities in these several
phase systems, No attempt has been made in this
concise compilation to give credit to those who

did the work,

Both these families of binary systems are much
more complex then a cursory examination would
indicate.2 It is obvious that fundamental studies
of the structures of the complex compounds
involved would be of general value. The large
number and wide variety of complex compounds
observed in these systems are indicated in
Table 2.1.1. (The ratios shown are the ratios
of the alkali fluoride component to the ZrF, or
the UF, component.,) Comparisons of the phase
dmgrams presented in Figs. 2.1.1, 2.1.2, ond
2.1.3 display several of the striking character-
istics of these materials.

A stable, congruent, high-melting-point compound
(Fig.2.1.1) with the formula 3MF-ZrF, or 3MF UF,
characterizes all the diagrams, except those for
LiF-UF , and NaF-UF,. For the LiF-UF system,
which represents the lowest ratio of radlus of
M* ion to radius of M4t ion, the 3:1 compound
does not exist; for this system, and this system

alone, an incongruent compound 4LiF-UF, is
~observed, In the NaF-UF, system the compound
3NaF.UF, is congruent but relatively low melting,

and it |s unstable at temperatures below about

- 530°C.

 

MThe- petrographic _examinations reported here were
performed by G. D. White, Metallurgy Division, and

-~ Te. Ne McVay and H. Insley, consultants, The x-ray

examinations were performed by R. E. Thome and
B. A. Soderberg, Materials Chemlstry Division.

2The case of the related system KF«ThFy is similar.
The early concept of the system reported by E. P.
Dergunov, Doklady Akad. Nauk S.5.5.R. 60, 1185—1188
(1948), was @ simple one, but a more recent report by
w. J. ASkGf, E. R, Segnlt, and A. W, Wylse, ] Chem,
Soc. 1952, 4470, shows the system to be relatively
complex.

79

 
 

 

 

ANP PROJECT PROGRESS REPORT

TABLE 2.1.1. COMPOUNDS* or'z:g‘on UF,

WITH ALKALI FLUORIDES

 

3:1 2:1

3:2

5:3 5:4

 

3LiF°ZrF4 {c)
3N¢:F-ZrF‘ (<)
SKI"'-ZrF4 (<)

3RbF-ZrF4 (<)

2LIFZrF ()

2KF-ZeF , (1)
2RbF-ZrF, (1)
ALiF-UF o
3NaF-UF‘ {c)
3KF4UF4 ()
3RbF-UF, (c)

2NaF-UF, (1)
2KF-UF, (1)
2RbF-UF, (1)

2NaF-ZeF, (1) 3NaF-2ZrF, (S)
3KF-2ZeF (1)

TNaF+6ZrF , (c)

.'SRbF-AZrF4 {c)
7LiF-6UF, (1)
7NaF+6UF , (c)
7|(F'6UF4 {c)
7RbF-6UF , (f)

S5NaF-3U F4 {1

 

 

1 3:4 2:3 132 1:3 14 1:6
3LIF-4ZeF , (1)
NaF-ZrF, (M) 3NaF-4ZeF, (1)
KF-ZrF, (c)
RbF-ZeF, (c) RbF-22rF, (1) |
LiF-4UF, (1)

NaF+2UF , (S)
KF-2UF, (1)

RbF-UF, (c) 2RbF+3UF (1)

RbF+3UF , (1) RbF+6UF , (1)

 

*(c) ~ Congruent compound,
() Incongruent compound.
(M) Metastable compound.
(S) Subsolidus compound.

A series of incongruent compounds, 2MF.ZrF
and 2MF-UF4,' are observed in all the diagrams,
except those containing LiF.  The compound
2LiFUF, does not exist, and 2LiF.ZrF, has a
relatwely low melting point but is congruent. All
the incongruent 2MF.ZrF, and 2MF UF, com-
pounds, except the 2NaF.UF, material, melt to
form the compounds 3MF ZrF or 3MF-UF, and
liquid. :

The |ow-me|fmg-pomt central portion of the
binary mixtures, that is, the mixtures with 35 to
55 mole % ZrF , or UF4, as shown in Fig, 2.1.2,
exhibits many vorlctlons in compound types. The
occurrence of the unusual compound 7MF-6ZrF,

“or 7MF-6UF , is remarkable, The compound hos'

80

the same crystal system wherever it oceurs, and it
displays no polymorphism. One member, the
compound 7NaF 6ZrF forms an extensive
interstitial solid solutlon series with the -next
lower ZrF 4 compound.

Compounds formed from mixtures with high
concentrations of ZrF, or UF, are invariably
incongruent, but they display wide variations in
degree of incongruency, These compounds, as
shown in Fig. 2.1.3, occur at widely varying
compositions in the range 55 to 95 mole % ZrF,
or UF,. Several of the diagrams in this series
are considered to be preliminary, and the systems
those diagrams depict are still being studied.

 
 

 

 

 

 

Al

o

PERIOD ENDING JUNE 10, 1956

SEORET"
ORNL-LR-DWG 14626

 

1000

CsF-ZrF, RbF—ZrFy KF~Zrf, NaF-Zrf, LiF-ZrF4

P a ¥

 

900

AMINZN NN

 

800

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

? AN
o \
lfil:j 700 / ‘ ‘
2
g \ \
o ——
W
: \
Z 600 _\/
’...
N
N
u
500 j 3
400 ] ol St — &1 S
N N 5| & N N
Gl A ol @ 5l % “
Qf o | x x| ¥ Z} = 3
M} N " o MmN mf N o
300 | i [ I I
O 10 20 30 0 0 20 30 O {0 20 30 O 10 20 30 O 10 20 30
ZrF, (mole %)
1000
CsF-UF, /\ RbF —UF, /\ KF -UF, NaF -UF, LiF - UF,
900 f 1\ / \\ /
/ L | \
& 700 - -
e NI | \
<z
5 |/
S .
& 600 5 i A
-~ T o TNt .
M m
00, & TRy
2
1 0
- . C. < Q: o g . N ) - -« .
400 55T S5 o 51 i
ul w wlwl 2t 2 .. N
v w aF . o) - wi W o
ol o e x| ¥ = .
S ol | R E ol o Y |
300 1 | P p o g E i
.0 10°20°30 O © 20 30 . 0 10°20 30 O 40 20 30 0 10 20 30
. ST © T UR, (mole %) o '
Fig. 2.1.1. The Systems MF ZrF cmd MF UF‘, Vhere MF Is an Alkali Fluoride, in the Composmon

Range 0 to 40

mole % ZrF4 or UF4.

81

 
 

 

 

ANP PROJECT PROGRESS REPORT

- oconey-
ORNL-LR-OWG 14627

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1000
CsF-ZrF, RbF - ZrF, KF-ZrF, NaF-2rFf, - |uiF-2eR,
900
800
o
<
w700
>
.—
«
&
a
- ! \\
'1"\1/ \/
500 \ N \ \/‘.
\/' [
so0 [ ot | s
N ol & & of o
b w w ™ 'i'i ué
300 o ml 5
35 40 45 50 35 40 45 50 35 40 45 50 35 40 45 50 35 40 45 50
ZrF4(mole %)
1000
CsF-UF, RbF -UF, KF-UF, NoF-UF, LiF-UF,
a a a a 4
900
800 \\
—\
~ \// \
e [\ s ~ i
w 700 N .
e = V
=2 - .
k2
o
&
600
B S 7
M
Ww
o
=
500 .0
- -
L < w |.I?
400 - o 1= 5 2 5
o wl> o b w
U gl% £ z 5
o ~|lx ~ ~ ~
300 ] | | |
45 50 35 40 45 50

 

 

 

 

 

50 35 40 45 50 35 40
Uf, {mole %)}

H
o,

35 40 45 S0 35 40

Fig. 2.1.2. The Systems MF-ZiF, and MF-UF,, Where MF Is an Alkali Fluoride, in the Composmon
Range 35 to 55 mole % Z«F , or UF4.

82

 
 

 

 

 

 

PERIOD ENDING JUNE 10, 1956

PEGRES-
ORNL~LR-0OWG 14628.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

CsF-2rfy RbF-ZrFy KF-ZrFy T NaF—ZrfFy LiF-2ZrFq
900 - ' : ‘
800 // . // /
o
&
5 / /
<
o
& /
= 600
< ¥ /
500 / : : o =
- N
. " T
400 " H s %'t
. : 3
o -
300 : & ' n m
55 65 75 85 55 65 75 85 85 65 75 85 55 65 75 85 55 65 75 B85
2rf, {mole %)
1000 _ '
CsF-UF, / RbF-UF, / KF-UF, / / NaF~UF, LiF-UF, /
900 / 7/ / /
/ e / /
800 7 - / /
/| / /
- & - /[
1&' 700 ) — . ] J
S
«
Lt
£ )
. = 600 |
-
500 s —
a0 &
" 400: —— et
' 5 5 5 u¥ g ‘g
.l mi© 2 - o <
wif wl o o u o
300 ® :

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

"55 '65-75.8 55 65 75 85 .55 65 75 B85 55 65 75 8 55 65 75 85
: , - UF, (mole %) o

Fig. 2.1 3 The Systems MF- ZrF und MF. UF4, Where MF Is an Alkuh Fluorlde, in the Composmon
Range 55 to 95 mole % ZrF, or UF‘. :

o

 
 

 

 

"ANP PROJECT PROGRESS REPORT

THE SYSTEM RbF-UF,

H. A, Friedman R. E. Moore

The system RbF-UF, has been investigated
recently because of its importance in the qua-
ternary fuel system NaF-RbF-ZtF .UF,. The
earlier work of Blakely et al.3 has been largely
confirmed by thermal analyses of slowly cooled
melts, In addition to the compounds originally
reported, however, there is an incongruent com-
pound 7RbF-6UF . :

Quenching expenments on the system are now
under way, but the work is not yet far enough
advanced to warrant presenting a phase diagram of
the system, Petrographic examinations of a
series of quenched samples containing 33.3 mole
% UF, indicated that a rapid inversion of the
compound 2RbF.UF, occurs at some temperature
below 572°C, The congruent melting point of
RbF.UF, (727°C) was confirmed by petrographic
examination of a series of gradient-quenched
samples containing 50 mole % UF,. The compo-
sition of an incongruent compound previously
thought to be RbF-4UF has now been established
os RbF:3UF, by the observahon of single-phase
material in u series of equilibrated and quenched
samples containing 75 mole % UF,. This phase
is not stable above 714°C. In mixtures con-
taining between 50 ond 100 mole % UF,, two
phases, in addition to RbF.3UF,, have been
observed in both slowly cooled and quenched
samples. One phase contains more than 80 mole
% UF ,, while the other phase contains between
50 ond‘66 7 mole % UF .

THE SYSTEM RI:F--ZIF4
R. E. Moore

Several revisions have been made in the tenta-
tive diagram presented previously for the RbF-
ZrF, system,® The revised diagram is given in
Fig. 2.1.4.

Visual observation experiments have shown that
the eutectic between 2RbF.ZrF 4 and 5RbF 4ZrF
contains about 42 mole % ZrF and melts at about
410°C. Results of quenchlng experiments had
indicated previously that the eutectic contained

 

3. P, Blakely et ale, ANP Quar, Prog. Rep. June 10,
1951' ANP.65' P 87' Flg. 4.'.

4H. A. Friedman and R. J. Sheil, ANP Quar. Prog.
Rep. March 10, 1956, ORNL-2061, p 72, Fig. 4e1s

84

38 mole % ZrF, and melted at 375°C. It is now
apparent that 375°C is the lowered rapid inversion
temperature of 2RbF.ZrF , rather than the eutectic
temperature, In the region between 37 and 44.5
mole % ZrF,, 2RbF.ZrF, appears optically to be
like quench growth at all temperatures above the
lowered inversion temperature. This accounts for
the misinterpretation of the quenching data.

. Petrographic examination of a series of samples
containing 48.3 mole % ZrF, established that
5RbF4ZrF , is the primary phuse, the liquidus
temperoture is about 396°C, and the solidus
temperature is 390°C, Thus, the eutectic between
5RbF-4ZrF, and RbF.ZrF, contains about 48.5
mole % ZrF and melts at 390°C

Exammchons of slowly cooled preparations have
shown the existence of a compound containing
more than 50 mole % ZrF . This compound has
been tentatively |dent|hed as RbF-2ZrF , by the
observation of nearly single-phase matenal in a
series of equilibrated and quenched samples
containing 66.7 mole % ZrF,. This compound
melts incongruently to ZrF , and liquid ot 447°C,

THE SYSTEM Nt‘IF'--RhF-ZI'F'4
R. E. Cleary? H. A, Friedman

A phase diagram of the fuel solvent system
NaF-RbF-ZrF,, based on thermal analysis of
slowly cooled melts, differential thermal analysis,
and quenching experiments, is shown in Fig. 2.1.5.
Some of the compatibility triangles and quasi-
binary mixtures were reported previously.® The
present status of the information on the com-
patibility triangles, eutectics, peritectics, and
boundary curves is shown in Fig. 2.1.5. The
locations of the eutectics and the exact paths of
the boundary curves are not considered to be
final in this diagram; they are contingent on
confirmation by quenching experiments.

The phase relationships of the recently dis-
covered compound RbF2ZrF, in the termary
system have not yet been determmed The ter-
nary compounds of the solvent system are:
NaF -RbF.ZrF ,, which melts congruently at 615°C;
3NaF 3RbF+4ZrF ,, which melts incongruently at
450°C; and NaF:RbF.2ZrF, which melts congru-
ently at 460°C. A fourth ternary compound with

 

50n assignment from Pratt & Whitney Aircraft.

6H. A. Friedman, ANP Quar. Prog: Rep. March 10.
1956, ORNL-2061, p 72.

 
 

 

 

PERIOD ENDING JUNE 10, 1956

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 COMPEENTITI
ORNL~LR-DWG t4629
1000
900 Pt
\ /
\
\ _
1
800 - 1
| ’
S I
[
& /
5 700 /
<
o
w
a
= 600 !
) .
\
! \
|
! | /
I { \ [
< ! ’\l
4 SNIVINAS/
N ~NE
400 o H— i 5 v/ L
u 4
(I A ;
"M ' ~ | "_.'5 'T'\ 0
,l P 3 W =
300 i &
RbF 10 20 30 40 50 60 70 80 90 Zrf,
ZrFy (mole%)

Fig. 2.1.4. The System RbF-Z(F,.

the approximate composition 3NaF 3RbF 24eF
may exist at high temperatures. Quenching sfudnes

have not yet established temperature -limits on its
Preliminary quenching studies ‘show
compound -~ 3NaF. 3RbF-42rF - melts
mcongruently to: NoF-RbF-ZrF ~and hqund The -
optical ‘and x-ray data for. the compound are
presented in the followmg section. B -

stability.
that - the.

OPTICAL PROPERTIES AND X-RAY
PATTERNS FOR 3NuF-3RbF-4ZrF

> R.E. Thoma H.lnsley

| “Listed below are the identifying characteristics
of 3NaF 3RbF 4ZrF o O new compound encounfered

“in the phose studies described above, The symbol, _

| d(»&) means the distance between reflecting planes

measured in angstroms. The term I/1] refers to
the relative intensity of the lines as compared
with an arbitrary value of 100 for the strongest

line. Under optical properties, N, and N refer
to the ordinary and extraordmary lndrces of

“refraction of uniaxial crystals, The compound s
~ believed to form an mcomplete Solld solution with .
~ the compound NqF-RbF-?Zr_F_ ¢ - -

Optical datu '

Unioxial negatlve ;':'

= 10436
Ng = 1f430<.':'_,

Xeray data:
o dA) S,
677 s
611 15
5.72 15

85

 
 

 

 

ANP PROJECT PROGRESS REPORT

o
d(A)

5.37
4.67
4.21
4.11
3.98
3.82
3.72
3.56
3.52
3.44
3.31
3.22
3.12
3.11
3.07
3.01
2.95
2.571

2.482

2.356
2.243
2.164
2.111
2.088
2.060
2,025
2.004
1.976

1.947

1.909
1.861

1.812

1.727
1.638
1.591

/1

30

9
37
10
15
22
30

5
11

100

30
63
100

20

10

10
15
15
33
30

28
18
85

10
10
10

THE SYSTEM NcF-RbF-ZrF‘-U Fa

H. A. Friedman H.- Davis?

Two areas of the quaternary system NaF-RbF-
ZiF ;~UF, offer mixtures which may be svitable
fuels for circulating-fuel reactors, Solvents with
approximately 10 mole % NaF, 52 mole % RbF,
and 38 mole % ZrF, with the addition of 4 mole %
UF, give a fuel with a liquidus temperature of
approximately 500°C, With the addition of 7 mole
% UF , a fuel with a liquidus temperature of ap-
proximately 545°C is obtained. These fuyel
mixtures have RbF.UF, or RbF.3UF, as the
primary phase. No serious segregation of the
uranium phase has been evident in the experi-
mental work, These mixtures are not optimal
choices for the fuel mixture because the uranium
phase contains a high concentration of'UF4' and
there is o large difference between the temper-
atures at which the uranium phase and the primary
phase of the solvent separate,

Solvents with approximately 35 mole % NaF,
25 mole % RbF, and 40 mole % ZrF4- with the
addition of 4 mole % UF4 give a fuel with o
liquidus temperature of approximately 480°C.
With the addition of 7 mole % UF , a fuel with a
liquidus temperature of approximately 510°C is
obtained, At both UF, levels the uranium is con-
tained entirely in the solid solution 7NaF -6Zr (U)F ,.
Mixtures of this general type should have physical
properties and corrosion characteristics thot
would make them of definite interest as fuels,

THE SYSTEM NaF-L!F-Ber
R, E, Meadows

Quenching studies of the system NaF-LiF-BeF,
were continued, and it is now possible to present
a diagram (Fig. 2.1.6) of the triangle having LiF,
NaF, and NaF:BeF , at the apexes. Work on the
other portion of the system is under way, Petro-
graphic identification of phases in this system is
very difficult, because several of the compounds
have nearly the same indices of refraction and are
isotropic, or nearly so. As a result reliance was
placed on identification by means of x-ray dif-
fraction patterns; even this identification is

 

70n assignment from Pratt & Whitney‘Airc'i-i_;fh‘

 
 

 

 

 

- 2NaF- BeF .

 

 

  

 

. PERIOD ENDING JUNE 10, 1956

ORNL-LR-DWG 14630

 

 

 

 

ZrFy
912
TEMPERATURES ARE IN °C
RbF- 2 ZrF,
/s
?
" 3NoF-4ZrF, /7 403
537 Z
RDF-NaF-2ZrF,
510 RbF - ZrF,
7NaF - 6ZrF, 460 385
5RbF- 4ZIF,
505 - 395
438 -\ 400
3NaF . 3RbF - 4ZrFy 475 _ E
630 . =\ 65 620 .
2NoF - ZrF, 2RbF-2rF,

 

  
 

 

- 3NaF-ZrF,

747

NaF

 
  

3RbF-2ZrF,

  

RbF

 

298

- er3 - 790

Fig 2 I 5 The Sysfem NuFaRbF-ZrF

- dsfflcult in some cuses, since the s!andord paflernsj

i were obtained from sumples whlch are not smgle- ‘
S phose materials, . ;_
- The diagram ‘shows the followmg companblhty'
© friangles: " LlF-ZNGF-LiF-25eF2-2NaF-BeF LiF. - ©
"~ 2NaF. Ber-NaF ‘and 2NaF.LiF+2BeF . -NaF-BeF -
‘In-addifion to the congruently melnng
~ .compounds which- form - the apexes  of these
. .?compatlblhfy frnangles, the data show a subsolidus

compound, NaF.LiF.BeF,, which decomposes to
LiF and 2NaoF.LiF 2Bei= above about 240°C
(ref. 8), and another subsohfius compound, probably

5NaF L|F~3BeF2, wh:ch decomposes ubove ubout '
- 320°C to 2NaF . LIF 2BeF 2NaF:BeF ., and LiF.
. Below - -320°C,
g '5NaF-L|F 3BeF ., are the apexes ofa compahb:hty o
tnangle. ‘No otfner subsolidus relatlonshlps have
- been  definitely established,
" (Figs 2.1.6) the blnary and quasa-blnary eutectics -

LiF, . 2NaFL:F2§ o and

‘In-the diagram

. are indicated, and the dotted lines are the approxi- -

- mate pnmclry phose boundoraes which meet at the

 

8w. Jahn, . anorg, u. allgem., Chem. 276, 113—127
(1954).

87

 
 

 

 

ANP PROJECT PROGRESS REPORT

TP TDEN i
ORNL-LR-DWG 14631

  
 

 

NGF' Ber
EUTECTIC 380
343
EUTECTIC . '
332 EUTECTIC
2NaF - LiF -28eF, -\ 345
355 - - TERNARY EUTECTIC
EUTECTIC 348
340 ~ 2NoF- BeF,
NaF -LiF -BeFp ~ 595 -
~a EUTEGTIC
TERNARY EUTECTIC — 570
328 :
485 _ ”
7 TERNARY EUTECTIC
7 480 .
/
/
LiF ~~ NoF
845 EUTECTIC 995

649

Fig. 2.1.6. The System NaF-LiF-NaF.BeF,.

ternary evutectics shown in each of the three
compatibility triangles. There is a limited solid
solution region in the ternary system for a high-
temperature modification of 2NaF:BeF,, which is
not indicated on the diagram. It is not likely that
the limits of this region can be accurately de-
termined with the experimental methods now being
used.

The phase which was previously? believed to
be a ternary compound and which was tentatively
assigned the formula LiF-7NaF-.4BeF, was
identified by G, D, White through the use of high-
temperature x-ray diffraction as a polymorph of
2NoF-BeF,. This phase, which is stable above
320°C, had not been reported in previous investi-
gations of the NaF-BeF_ system!¢:11 and had not
been. found in quenched samples of NaF-BeF
mixtures examined in this laboratory because |ts
inversion is too rapid for it to be quenched in the

 

 

e M. Bratcher, R. E. Meadows, and R. J. Sfieil,
ANP Quar. Prog, Rep. March 10, 1956, ORNL.-2061, p 75.

106, Thilo and F. Liebau, Z. physik. Chem. 199,
_ 125..141 (1952).

p. M. Roy, R. Roy, and E. F. Osborn, J. Am.
Ceram. Soc. 36(6) 185 (1953).

88 -

binary system. It is, however, stabilized by the
addition of LiF, with which it undoubtedly forms
a limited solid solution.

Quenching experiments with compositions on the
LiF-2NaF.LiF-2BeF, join on both sides of the
mixture NoF-LlF-Ber (37-26-37 mole %) show
this mixture to have approximately the compo-
sition of the eutectic which melts at about 340°C.
Petrographic and x-ray diffraction examinations
of quenched samples and slowly cooled prepa-
rations show that the 2NaF -LiF.2BeF -2NaF.BeF,
join is a quasisbinary system wnh a eutectlc
that contains 45 mole % NaF, 16.5 mole % LiF,
and 38.5 mole % BeF, and melts at about 343°C..

A phase that appecred at a temperoiure below.
about 320°C in quenched samples was assigned
the formula 5NaF-LiF-3BeF, because the x-ray
diffraction pattern of a quenched sample of this
composition was not found to contain lines that -
could be assigned to any other known phase .in the
system.  The conclusion that this compound
exists only below the solidus temperature is
based on the following evidence. First, slowly
cooled preparations within the 2NaF.LiF:2BeF,-

 
 

 

NaF-BeF ,-2NaF.BeF, triangle contained only
the three phases at the apexes, since the reaction
rate in the solid state is too slow to permit
equilibrium to be obtained. ‘Second, examination
of quenched samples in the LiF-2NaF.LiF 2BeF -
2NaF BeF,. triangle showed that the ternary
eutectic (mp, 328°C) is above the upper limit of
stability of the compound and that the phases just
below the solidus temperature are those at the
apexes of the triangle,

The x-ray diffraction pattern of 5NaF.LiF -3BeF,
is related to that given by Jahn12 for low-femper-
ature 3NaF-.LiF- 2BeF2 Quenched compositions
corresponding  to 3NaF.LiF-2BeF, contained
2NaF.LiF-2BeF,, 2NaF-BeF,, and LiF just
below the solldus temperature and 5NaF.LiF -3BeF,,
2NaF.LiF. *2BeF,, and LiF below 320°C. No
evidence has been found in this laboratory to
suggest that 3NaF. LiF-2BeF , exists.

The locations of the phase boundaries and
ternary eutectics were deduced from the primary
and secondary phases observed in quenched
samples of the compositions within the compati-
bility triangles and the temperatures at which the
phases appeared, as well as from previously
obtained thermal analysis data,

THE SYSTEM NnF-RbF-BeF2
L. M. Bratcher

Thermal analysis data were obtained for o
number of compositions in the NoF-RbF-BeF
system containing 50 mole % BeF,, and some of
the slowly cooled melts were examined petro-

graphically and by x-ray diffraction. The avmlable .
data - show that the 'NaF-3RbF. BeF, join is a
quasi-binary system which has a eutectic with -

NaF-RbF- BeF

the . approxlmate composmon

join NaF-2RbF - BeF2
a quosu-bmary system which has a eutectic . with
-~ the composmon NaF RbF- BeF2 (43-38-]9 mole %)
that melts ot 655 + 5°C.

melts mcongruenfly. ,
mdncate that  the - add;tlon ‘of NaF lowers the
incongruent . melhng point of RbF-BeF

 

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

The NaF- RbF. BeF -
~join is’ compllcated because one_or. more. 1ernary '
: compounds exist on or near the join and RbF. Berr

“Thermal “anclysis data:

None ofr,’
the slowly cooled melrs was completely homo-_ -

PERIOD ENDING JUNE 10, 1956

geneous, but the mixtures containing 45 and 50
mole % NoF were predominantly one phase, and
therefore there may be a compound with the compo-

sition 2NaF-Rb_F-BeF2.

THE SYSTEM NGF-KF-L!F-UF4
Ro Ja Sheil
'Data reported previously on the liquidus temper-

atures of mixtures formed by adding UF, to NaF-
KF-LiF (11.5-42.0-46.5 mole %) indicated that the

liquidus temperature varied quite rapidly with

changing UFA concentration, at least with mixtures
containing approximately 4 mole % UF, (ref. 13).
Visuval observations of liquidus temperatures in
this system during this quarter gave values of
550, 570, and 595°C for mixtures containing 5.0,
7.5, ond 10 mole % UF respectively. The
values obtained by this method for the mixture
containing 10 mole % UF, agreed quite well with
earlier thermal analysis data. |t now appears that
the liquidus temperatures of mixtures in this
system containing 0 to 10 mole % UF, are not so
strongly dependent upon UF, concentrations as
the earlier datal3 had indicated.

ALKAL1 FLUORIDE-CeF; SYSTEMS
L. M. Bratcher
Preliminary data on the NaF-CeF ; and RbF- CeF3

systems were given in the prevrous report, 14
Study of the LiF-CeF, system was initiated
recently, and thermal analysis data obtained with
five compositions in this system showed that
there is a eutectic confaining approximately

19 mole % CeF ‘that melts at 755 £ 5°C. For the

NoF-Ce F

system, extrapolahon ‘of thermal data

obtained - wn‘h mixtures containing 10, 15, and

(45-41-14 mole %) that mehs at 640 + soc The-:' 20 mo!e % CeF3

is likewise believed to be

indicates that the eutectic which
melts at 725 % 10°C contains - ‘approximately 28
mole % CeF3. A compound reported to be present

Clin this ‘system has not yet been |denhf|ed The
- RbF-CeF, system has been studied more than the

oother two systems mentioned, but thermal analysis
~data and studies of slowly cooled melts have so.
~ far failed to give a clear picture of phase relations.

|t appears that at least two compounds are formed
' thot p055|bly have the- formulas 3RbF.CeF, and

 

]3R0 Jo Sheil, ANP Quar. Prog. Rep. Dec. 10, 1954,

ORNL-1816, p 59.

14 . M. Bratcher, ANP Quar. Prog. Rep. March 10,

89

 
 

 

 

ANP PROJECT PROGRESS REPORT

RbF.CeF,. In addition to liquidus and solidus
thermal e&ects on cooling curves, solid transitions
were noted at about 425 to 450°C with mixtures
in the 10 to 35 mole % CeF, range and at temper-
atures ronging from about 500 to 580°C with
mixtures containing 33 to 60 mole % CeF,. These
transitions probably account for the fact that
particle sizes of crystalline phases are so small
that microscopic identification is difficult and
some phases even give . poor x-ray diffraction
‘patterns, The minimum liquidus temperature of
approximately 615°C apparently occurs between
42.5 and 50 mole % CeF,. Study of all three

systems mentioned is continuing.

THE SYSTEM LIF-CsF
L. M. Bratcher

The recent availability of pure CsF prompted a
re-examination of the LiF-CsF system which is
of interest because a compound is formed such as
that formed in the LiF-RbF system.1% The results
of the earlier studies were not published because
of the scatter in the thermal analysis data. While

the existence of other: ulkali-hdlide binary
compounds is well established, no mention of

. alkali-fluoride binary compounds has been found

in the literature, The freezing point of the CsF
used in the recent investigation was found to be
700 * 5°C, which agrees with recently published
values of 705°C (ref. 16) and 703°C (ref, 17).
The thermal analysis data shown in Fig. 2.1.7
include some data from the earlier studies. The
binary compound, believed to have the compo-
sition LiF.CsF, was easily recognized under the
microscope because it is birefringent, whereas:
both LiF and CsF are isotropic. However, for
reasons that are not well understood at present,
the mixtures examined to date gave rather poor
x-ray diffraction patterns for the compound. The
large number of lines observed suggests that the
compound is either monoclinic or rhombohedral.

 

15 . M. Bratcher et als, ANP Quar. Prog. Rep. June
10, 1954, ORNL-1729, p 44.

160, Schmitz-Dumont and E, Schmitz, Z. anorg.
Chem. 252, 329 (1944).

174, A, Bredig, He R. Bronstein, and Wme T. Smith,
Jre, J. Am. Chem. Soc. 77, 1454 (1955).

SOMNPOENTA
ORNL-LR-DWG 14632

 

900

T

 

800

~

 

700

.

 

600

TEMPERATURE (°C)

 

500

 

 

400

 

 

 

 

 

300

LiF-CsF

 

 

 

 

 

 

 

 

 

LiF fo 20 30 40

50 60 70 80 90 - CsF

CsF (mole %)

Fig. 2.1.7. The System LiF-CsF.

90
 

 

THE.SYSTEM,Mng-Can
L.. M. Bratcher

In the course of an investigation of the system
NaF-MgF -Can, thermal data were obtained with
two MgF ,-CaF, mixtures which indicated that the
literature value for the melting point of the eutectic
mixture in this system mlght be too low.?8 There-
fore six additional mixtiores were prepared which
covered the composition range 37.5 to 50 mole %
CaF,.  Cooling curves obtained with these
mixtures showed that the eutectic mixture con-
tained approximately 48.5 mole % CaF . and melted
at 985 + 5°C, as compared with the literature value
of 940°C.

THE SYSTEM RbF-Cqu
L. M. Bratcher

Preliminary thermal analysis data indicate that
there is a eutectic between RbF and a binary
compound (probably RbF-CaF ) that contains less
than 10 mole % CaF, and has a melting point of
765°C. Cooling curves with mixtures containing
10 to 45 mole % CaF, did not show reproducible
liquidus thermal effects, but the thermal effect at
765°C became less marked with increasing C0F2
concentration in this range. Petrographic exami-
nation of samples containing 10 to 20% CaF,
showed increasing amounts of a cubic compound
having a refractive index of approximately 1.410.
A 1:1 compound {molar ratio) has been reported!?
to be present in the similar system KF-CaF,,.

THE SYSTEM LlF-NIF2
L. M. Bratcher

Data obtained with three mixtures in the LiF-
NiF, system covering the composition range 10 to

30 mole % NiF, were given in the previous

report.29 The compound found in this system was
tentatively assigned the formula 3LiF. NxF

‘mainly on the basis of petrographic and x-ray
diffraction studies of slowly cooled melts. The
thermal ~analysis data obtained to. date are
dlfhcuh to interpret, but it appears llkely that
the compound melts ‘incongruently to NiF., and
liquid ot o temperature -only slightly aboye.the

 

18€, Beck, Metallurgie 5, 504 (1 908).

19p, Silber and ‘M. Ishaque, Compt. rend. 232, 1485

(1951).

20, M. Bratcher, ANP Quar. Prog. Rep., March 10,
1956, ORNL.-2061, p 78.

alkali fluoride-ZnF,

PERIOD ENDING JUNE 10, 1956

minimum liquidus  temperature. Well-crystallized
NiF, was found in slowly cooled mixtures con-
taining 30 mole % NiF, or more. The first attempt
to determine the melting point of NiF2 was
unsuccessful, but the melting point was evidently
below 1200°C, the maximum temperature to which
the material was heated.

'THE SYSTEM UF,-U0,
R. J. Sheil

Preliminary thermal analysis data and the
results of studies of a few slowly cooled mixtures
in the UF ,-UO, system were reported previously,?1
Thermal analysis data obtained during this quarter
were not very satisfactory because of the previ-
ously mentioned undercooling difficulty and a
container problem. The use of graphite con-
tainers for the UF ,-UQ, mixtures was abandoned
when it was dlscoverej thdt mixtures containing
4 wt % UO, or more penetrated both ordinary
(C-18) and high-density graphite when heated to
a maximum temperature of 1100°C.  Mixtures
containing 2 wt % UO2 or less did not penetrate
ordinary graphite to a noticeable extent when
heated to the same temperature. Surface tension
measurements of UF ,-UO, mixtures will be made
in an effort to account for this interesting
phenomenon. The recent thermal analysis studies
were conducted in sealed nickel capsules to
minimize changes in oxide content of the mixtures
while the thermal analysis data were being
obtained. This closed apparatus does not permit
the use of seeding, which apparently will be
required in order to obtain reliable liquidus values
in this system, Heating curves and some cooling
curves obtained with mixtures containing 4, 6,
8, and 10 wt % UO show evidence of a eutectic
which melts at approximafely 910°C. Extrapolation

of available liquidus temperature data for these
mixtures involves considerable uncertainty, but
it appears that the eutectic probably contains

9t0 11 % UO (10 3to 12. 6 mole % U02)

THE SYSTEM ZnF,-ZnO
L. M. Bratcher H. A. Friedman

During the course of earlier investigations of
systems,?2 a variation in

 

21R, ), Sheil, ANP Quar. Prog. Rep. March 10, 1956,
ORNL-2061, p 71.

22) , M. Bratcher, R. E. Traber, Jr., and C. J. Barton,
ANP Quar, Prog. Rep. june 10, 1952, ORNL-1294, p 94.

91

 
 

 

 

 

ANP PROJECT PROGRESS REPORT

liquidus temperatures was noted in attempts to

determine the melting point of pure ZnF,. Cooling
curves showed liquidus effects ranging from
885 to 940°C and a second break at or near 845°C.
Since ZnO was found in all the somples that
. were examined subsequent to the melting-point
determinations, it seemed probable that the
liquidus temperature was affected by the amount
of ZnO present, Confirmation of this postulate
was obtained when a cooling curve obtained with
oxide-free ZnF,, prepared by NHF.HF treatment
of ZnF, previously dried in an HF atmosphere,
showed a thermal effect only at 940 + 5°C, Be-
cause of the large discrepancies between this
value for the melting point of ZnF, and those
reported in the literature (872°C by Puschin and
Baskow, 23 and 875 + 3°C by Haendler, Patterson,

 

23N, Puschin and A. Baskow, Z. anorg. Chem. 81,
359 (1913).

and Bernard?4), the slowly cooled ZnF, was
analyzed chemically and spectrographically. 1t

~ was also carefully examined petrographically and

by x-ray diffraction, along with a highly precise
determination of cell parameters. = Since no
evidence was found of the presence of impurities
in more than trace amounts, the melting points for
ZnF, reported in the literature are believed to be

~erroneous, possibly because of oxide contami-

nation,  Thermal analysis data obtained with
mixtures formed by adding 5, 10, 15, and 20 mole %
ZnO to purified ZnF, show that the eutectic
which melts at 850 t 52°C contains approximately
21.5 mole % Zn0O. Only the pure components were
found in the slowly cooled mixtures, and thus it
appears that solid solutions do not occur, at least
at room temperature,

 

244, M, Haendler, W. L. Patterscn, Jr., and W. J,
Bernard, J. Am. Chem. Soc. 74, 3167 {1952).
 

 

 

PERIOD ENDING JUNE 10, 1956

2.2, CHEMICAL REACTIONS IN MOLTEN SALTS

F. F. Blankenship
R. F. Newton

ACTIVITY OF CHROMIUM IN ,CHROMIUM;NICKEL
ALLOYS

M. B. Panish

Further measurements were made of the electro-
motive  forces of electrode concentration cells in
in order to.determine the activity of chromium-nickel
alloys. The cells used were the same as those
described previously,! and they contained as an

‘electrolyte a eutectic mixture of sodium chloride

and rubidium chioride with about 0.2 to 0.5 wt %
chromous chloride added. '

Activity determinations were made for alloys
containing from 11.2 to 53.0 mole % chromium, It
was found that there was a marked tendency for
the electromotive forces of the cells to drift down-
ward because of the reaction

Ni + GCl,=NiCl, + C

This effect was reduced markedly by packing the
lower end of the cell with crushed quartz in order
to prevent the transfer of nickel by convection and
diffusion. For several cells in which the quartz
packing was not used, the equilibrium electromotive
force was approximated by extrapolating the steadily
drifting electromotive force to zero time.

With cells in which the alloy electrode contained
over 35 mole % chromium, erratic results were ob-
tained after raising and lowering the cell tempera-
ture, whereas the electromotive forces obtained
should be reproducible. The reasons for this be-
havior have not yet been ascertgined, but it is
highly probable that the diffusion rate in these elec-
trodes is very low and that surface effects play a

very important role. It should also be noted that ac-
tivity values for the high-activity region will be
- approximations because of the low electromotive

force produced by cells containing these electrodes
vs a pure chromium electrode. If the activity of the

chromium in the alloy electrode of such a cell is
0.90, then the elécfrom_otive' force will be about -
~ 0.004.: -The nonreproducibility of the cells in this
~ region is of the same order of magnitude as ihe

electromohve forces measured

 

M. B. Panish, ANP Quar. Prog. Rep. March 10, 1956,
ORNL-2061, p 92,

L. G, Overholser
G. M, Watson

The activities obtained for various chromium-
nickel alloys at 750°C are shown in Fig. 221
along with the curve obtained by Grube and Flad?
at 1100°C. The two curves are not inconsistent, in
that it is quite possible that the differences are
due entirely to the difference in the temperatures at
which the measurements were made,

UNCLASSIFIED
ORNL—LR—-DWG 14633

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

100
|
—_ l |
i
[ |
TWO PHASE REGION |
0.75 of AT 750°C A __[—IF
-
, I’
v\/
> / \QQ’ /
e .
>
Z 0.50 _ .
< /" TWO PHASE REGION
AT 1100°C
1o
0.25 £ THIS WORK AT 750°C
(ELECTROCHEMICAL)
o DATA OF GRUBE AND FLAD AT
' 1100°C (2Cr,05+3H, == 3H,0+2Cr)
0
Ni 25 50 75 cr

CHROMIUM (at. %)

Fige 2.2.1. Activity of Chromium in Nickel-

- Chromium Alloys at 750 and 1100°C.

It should be noted that the chromium activity in
the region near 15% chromium is slightly below the
ideal activity which would have been predicted for
a  perfect solid solution, This is considerably
lower than the activity which might have bteen ex-

pected from an inspection of the chromium-nickel
phase diagram, and it justifies the assumption of

near- “‘ideal’ activity for chromium in Inconel.

- This assumption is usually made in discussions

regarding the chemical equilibria involved in the
corrosion of Inconel by molten salts,

 

26 Geube and M. Flad, Z. Elektrochem. 48, 377 (1942).

93

 
 

 

ANP PROJECT PROGRESS REPORT

_REDUCTION OF UF, BY STRUCTURAL METALS

Jo D. Redman

Further studies were made, by use of the filtration
method, of the reduction of UF, by chromium and
by iron in reaction mediums that dlffered from those
used previously. In the earlier studies, the re-
action mediums used were NaF-ZrF , (50-50 mole %e
53-47 mole %4 59-41 mole %3), NaF-LiF-ZrF,
(22-55-23 mole %),é and NaF-LiF-KF (11.5-46.5-42
mole %).7 Other alkali fluoride mixtures containing
ZrF, have been employed as reaction mediums in
the recent studies in order to determine the effects
of various alkali fluorides on the interaction of
UF, with chromium or iron, The results obtained
by using KF-ZrF or LiF-ZrF , (both 52-48 mole %)
as reaction medlums are presented in Tables 2.2.1

and 2.2.2.

 

3J. D. Redman and C. F, Weaver, ANP Quar. Prog,
Rep. June 10, 1954, ORNL-1729, p 50; ANP Quar.
Prog. Rep. Sept. 10, 1954, ORNL-1771, p 60.

4), D. Redman and C, F. Weaver, ANP Quar. Prog.
Repo Ime 10. 195" ORNL‘1896’ P 60-

5). D. Redman, ANP Quar. Prog. Rep, March 10, 1956,
ORNL-2061, p 93.

6), D. Redman ond C. F. Weaver, ANP Quar. Prog
Rep, Sept. 10, 1955, ORNL-1947, p 74.

75, D. Redman ond C. F, Weaver, ANP Quar, Prog.
Rep, March 10, 1955, ORNL-1864, p 56.

TABLE 2.2.1., DATA FOR THE REACTION OF UF
WITH CHROMIUM IN MOLTEN KF.ZrF (52-48 MOLE %)
AT 600 AND 800°C

Data for the reaction of UF  with chromium at
600 and 800°C in the reaction medium KF- ZrF
(52-48 mole %) are given in Table 2.2,1. In these
runs approximately 2 g of chromium was reacted
with UF, (10.6 wt %, 4.0 mole %) dissolved in
approx:mafely 40 g of the KF-ZrF mixture cone-
tained in nickel. Similar studies were made with
LiF-ZrF , (52-48 mole %) as the reaction medium at
a UF, concentrahon of 12.0 wt % (3.9 mole %).
these studies chromium from two different sources
was employed; the results are given in Table 2.2,2.
- As seen from the data in Table 2,2.2 the equi-
librium chromium concentrations cre the -same at
both temperatures for the two different chromium
metals used — hydrogen-fired electrolytic chromium

TABLE 2.2.2. DATA FOR THE REACTION OF UF <
WITH CHROMIUM IN MOLTEN LIF-ZrF, (5248 MOLE %)
AT 600 AND 800°C

 

 

Conditions of
Present in Filtrate

 

 

 

Equilibration
Temperature Time Total U Total Cr*  Total Ni
(°C) () (wt%)  (ppm) {ppm)
600 3 8.0 1030 35
3 9.4 1060 55
5 9.7 1180 55
5 8.5 1060 235
800 3 8.2 1250 35
3 8.4 1190 25
5 8.1 1150 295
5 8.1 1150 285
16 8.2 1060 30
16 8.2 1140 5

 

*Blank of 290 ppm of Cr at 800°C.

94

Conditions of Present in Filfrote:

 

 

Equilibration

Temperature Time TotalU Total Gr*  Total Ni
(°C) (hr)  (wt %) (ppm)  (ppm)

600 3 9.9 2550 30

3 9.8 3000 75

S 10,0 2980 70

5 9.4 3060 45

5 9.5 2800** 50

5 9.5 2790** 50

12 9.6 2910 30

12 9.6 3060 35

800 3 10.1 3820 55

3 9.2 3830 30

5 B9 4070 35

5 8.6 4040 4;

5 9.2 3810** - 40

5 9.3 3780** 60
12 9.5 -~ 3780 60 |

12 9.5 3830 50

 

*Blank of 250 ppm of Cr at 800°C. Electfolyfl'c-_chro-
mium hydrogen-fired at 1200°C was used in all runs ex-
cept those noted.

**Blank of 190 ppm of Cr at 800°C. Very pure iodide

chromium, not hydrogen-fired, was used in these runs.

 
 

 

 

 

and . iodide chromium. The iodide chromium was
obtdined from Battelle Memorial Institute and con-
tained 10 ppm or less of oxygen, This pure chro-
mium was not hydrogen-fired, and therefore a com-
parison of the blanks obtained for this metal with
those found for the electrolytic chromium (used in
all - previous studies), which was hydrogen-fired
under the usual conditions, should demonstrate the
effectiveness - of the hydrogen freatment, The
chromium values of 190 and 250 ppm obtained for
the unfired iodide chromium and the hydrogen-fired
electrolytic chromium, respectively, suggest that

the hydrogen-firing is successful (unfired electro-

lytic chromium gave a blank of 900 ppm). Evidently

the major portion of the blank arises from oxidizing

materials (H,0 and HF) present in the LIF-ZI’F4
mixture,

The equtllbrlum chromium concentrations and the
equilibrium constants calculated from mole frac-
tions for the reaction of UF , with chromium in the

various solvents are presented in Table 2,2,3 for

comparison, The effect of varying the NaF-to-ZrF,
ratio in the various NaF-ZrF, mixtures on the
chromium concentration has been shown and dis-
cussed previously. The values are included in
Table 2.2.3 for comparison with the KF-ZrF, and
LiF-ZrF, data. It is evident from the volues

PERIOD ENDING JUNE 10, 1956

given .in Table 2.2.3 that the particular alkali
fluoride used in combination with ZrF , influences
the reaction markedly, The different alkali fluorides
affect the activity of the UF ; to varying degrees.
The activity of the CrF, is also influenced through
complexing of the CrF, by both the alkali flucride
and the ZrF . Studles will be made shortly with
an I‘?I:\F—ZrF4 mixture as reaction medium to com-
plete the alkali fluoride series,

Studies of the reduction of UF, by iron at 600
and 800°C with LIF-ZrF and KF-ZrF (both 52-48
mole %) as reaction medlums were made at UF
concentrations of 12,0 wt % (3.9 mole %) for the

Li F-Zrl':i mixture and 10,6 wt % (4.0 mole %) for

the KF-ZrF, mixture, The data are presented in
Table 2, 2.4 As may be seen from the data the
equilibrium iron concentrations are not significantly
changed by replacing LiF with KF., |t may be
noted that somewhat smaller values result at 800°C
than at 600°C for both solvents. These values
also are in good agreement with those found when
NaF-ZrF, (50-50 mole %), NcF-ZrF (53-47 mole
%), and NaF-ZrF (59-41 mole %) were used as
the reaction medlums. The iron values obtained
by using NaF-LiF-ZrF, (22.55-23 mole %) as the

solvent fell in the same range, but in this case the

values were slightly larger at 800°C than at 600°C,

TABLE 2.2,3, EQUILIBRIUM CONCENTRATIONS AND CONSTANTS FOR THE REACTION
| Cr° + 2UF ;=2 2UF + CrF; IN VARIOUS SOLVENTS

 

 

. _ Temperature - UF, Cr
Solvent - {°0) (mole %) ~ {ppm) K.*
LiF-ZrF Lo 00 40 2900 7 x 10~4
(52-48 mole %) - 800 40 3900 S 7x 1073
NaF-ZrF, 600 4.1 2250 4x 1074
(50-50 mole o 800 41 2550 5 x 10~4
“NaF-ZrF, 00 4.0 Cowoe o ax et
(53-47 mole %) B 800 4.0 2100 - 3x 10=4
© NeF-ZeF, 600 a7 975 lLax10-S
 (59-41 mole %) - 800 37 10507 T 16 x 1073
KF-ZeF, - 600 3.9 080 0 2.4x 1075
) (52-43 mole %) 800 3.9 o M0 - 32x 1073
NaF-LiF-ZeF, | 600 25 s 1x 108
(225523 mole %) 800 25 75 4x 10~5
" NaF-LiF-KF _ 600 2.5 1100
-(Il.5-46.5-42 mo|e % 800 2.5 2700

 

*Kx = XEIF3 XCer/XCr XfiF“, where X is concentration in mole fractions.

95

 
 

ANP PROJECT PROGRESS REPORT

TABLE 2.2,4. DATA FOR THE REACTION OF UF,
WITH IRON IN LIF-ZsF (52-48 MOLE %) AND IN
KF-ZrF , (52-48 MOLE %) AT 600 AND 800°C

 

Conditions of

. . P . .
 Equilibration resent in Filtrate

 

Temperature Time Total U Total Fe* Total Ni
CO - () wt%)  (pem)  (ppm)

 

 

 

Solvent: LiF-ZrF4 {52-48 mole %)

_50(_)',_ .

3 9.1 450 )
3100 550 35
5 97 460 o
5 9.9 520 65
800 3 9.8 360 15
3 9.6 9 2
5 9.6 400 35
5 9.5 350 35

Solvent: KF-ZrF4 (52-48 mole %)

600 3 7.2 520 1
3 8.0 630 10

5 7.7 490 5

5 7.9 510 10

800 3 8.2 330 65
3 8.1 280 30

5 8.1 310 70

5 8.1 330 35

 

*Blanks of 100 and 40 ppm of Fe at 800°C for LiF-Zr F4
and KF-ZrF ., respectively,

The indifference exhibited by the UF ,-Fe® reaction
to changes in the reaction medium is in marked
contrast to the behavior noted for the UF ;-Cr re-
action. No satisfactory explanation can be offered
at this time for these differences.

SOME OBSERVATIONS ON MASS TRANSFER
OF CHROMIUM BY MOLTEN SALTS

-W. R, Grimes

When the corrosion of Inconel by NaF-ZrF -UF,
mixtures and by NaF-KF-LiF-UF, mixtures is
compared, some striking dlfferences are observed.
The corrosion, as measured by depth of void forma-

96

tion in the hot region of a system with a temperature
gradient, is much worse when the circulated fluid
is the alkali fluoride mixture than when it is the
ZrF ~bearing fuel. In addition, discrete crystals

of nearly pure chromium are found in the cold’

zones of Inconel loops in which ‘the NaF-KF-LiF-
UF , mixture has been circulated. Metallic deposits
are not usually observed in loops which have cir-
culated the NaF-ZrF -UF4 preparations, Depth of

void formation does increase with time, however,

for both classes of systems, even though the con-

centration of chromium compounds in solution -

reaches an equilibrium concentration early in the
corrosion test and remains essentially constant
thereafter. The amount of chromium removed from
the loop walls, as estimated from the volume of
voids in the hot zone, is considerably larger than
can be explained by the amount of chromium com-
pound in solution in the melt ofter the test. More-
over, the corrosion seems to be nearly independent
of the ratio of the.surface area of the loop to the
volume of the melt and not strongly dependent on
the concentration of uranium in the melt.

The processes that occur are undoubtedly very
complex, and it is not possible fo explain in
quantitative fashion all the effects observed,
However, by the use of equilibrium data obtained
for the chemical reactions and by making some
reasonable assumptions it is possible to rationalize
much of the available data and to indicate a pos-
sible explanation for the observed similarities and
differences in these two fuel classes. For ZrF -
bearing mixtures, of which the mixture with 50 mole
% NaF, 46 mole % ZrF,, ond 4 mole % UF is
typical, the important corrosion reaction is known
to be

NuF-ZfF4 S ) :
2UF + Cr fi CI‘F2 + 2UF3

The equilibrium constant for this reaction is given

by

(CrF,) (UF5)?
K(a) =

r

 

(UF ,)2(Cr)

where the quantities in parentheses represent the
activities of the materials involved. The equi-
librium constant, K('a), cannot be determined by

experiment, but the equilibrivm concentration,

 
 

 

 

 

 

 

 

K('c), which is'defined by the equation

2
o Crerrp Clury)
Kiey = '

2 .
“lur o Cien

‘where C U )Eepresents the ethllbnum concentration

of the designated component in mole fraction, can
be evaluated for pure chromium and has been shown
to be constant at a given temperature,

For cases in which the CrF, and UF; present
are formed only by the reaction described above,

C = 2C

(UF,) (CeFy)

and
3
4C (Cer)

Kiey = —0———— -
' C(UF )C(Cr)

‘lfrthe amount of UF, added inifiully is dlways the

same and if the fractlon of this UF4 reduced to
UI"'3 is small, then

Cluey =47
L
4C(CtF2)'

K/ & ——
(c) ’ ’
€ A C(c”

, , o\ l_/.'.’; i, o
c Kier Cien] 7~
(C'Fz) =- S Co- . i

and

) Accordmgly, the CrF2 concenh'ahon ct equuhbrlum,_
if all other factors are constant, depends on the
~.cube “root of the mole fraction. of chromlum in the
~ alloy. Therefore it should be possiblie to calculate,f_j i
~ from experlmental data for pure chromium, the equi~
hbnum concentrations to be expected- from Inconel, -
~ The corrosion reaction for the NaF-KF-LiF-UF
;mlxtures is considerably - more complex. In \‘hls‘ ;
- case two consecutlve reactzons uppurently proceed

NcFoK F-LIF
S ,

2UF + Cr CrF + 2UF

’ "3crl='2*'-——--—a' NP KL 2o, +Cr

As a result of these reactions abouf 80% of fhe

PERIOD ENDING JUNE 10, 1956

chromium compound in solution is trivalent, . The
net reaction is {approximately)

 

 

14UF, + 5 KL 4oE 4 GF, + 4UF,
and , - |
o (chz)(chs)HUFz)_H-
K(a) =
(UF §14(Cr)5
Cieer ) C?cn-'_s) C(IJF:,)
K(c) = — .

14 5
S Clen

If CrF,, CrF,, and UF, all arise solely from this
reaction and if C(UF4) is constant, then

C(Cst) = 4C(CtF2) '

Cc

' 4ch 14
CieeF ) 4C1cer ) MC(Cer )

 

K, =
(c} ’
+ ~5
A' C(cr)
»r ,L 1/19
c (C)A C(Cr)
C F = LAt ettt e,
(CrF2) T 4w 144

Accordingly, with all other variables held constant,

~~the concentration of CrF, + CrF, is a function of
-~ the 5/19 power of the mole fraction of chromium in
the alloy, Therefore, from equnllbrwm dota for

pure chromlum, correspondmg concentrcmons ccm'
be culculoted for lnconel & - ' '

 

81y should be emphasized that accuracy of the come -

lpllcated equation -

s not an essential port of the crgumenf. “The same
qualitative conclusions are uppurenf Whether ‘one. uses'
~as the corrosion reaction . oY

. er

S 2UF +Cr—'"2UF3+CrF2
- T 2 UF, + Cremt BUF, + cu=3 .

In feacflon 1, as shown for the ZrF -bearmg ffiel the i

Csz concentration varies as (Cr)]/3, and in reaction 2, -
the CrF concentration varies as (Cr)l 4,

97

 
 

 

 

 

 

ANP PROJE_CT PROGRESS REPORT

Inconel contains 15 wt % Cr or a chromium mole
fraction of about 0.16. The assumption that the
activity of chromium in Inconel is roughly equal to
its mole fraction appears to be rather good. The
compounds in the melts, which were treated with
pure chromium at 600 and 800°C, are shown in
Table 2.2.5, along with corresponding valuves cal-
culated for equilibration with Inconel. Experimental
values obtained from corrosion experiments with
Inconel and these melts are usually slightly lower
than these calculated values, presumably because
in the experiments equilibrium is dttained with an
Inconel surface layer significantly depleted in
chromium. - When Inconel powder is used in experi-
mental equilibration apparatus, side reactions that
involve oxides of the metals complicate the situa-
tion; equilibrium concentrations of chromivm come
pounds higher than those calculated are often

- observed. In general, it appears that the calcu-

lations are good only for the idealized case de-
scribed.

From the data in Table 2.2.5 it is obvious that
Inconel exposed to the NaF-KF-LiF-UF , melt will
support o higher concentration of CrF 5-CrFgy
equilibrium ot 800°C than pure Cr° IS oble fo
support at 600°C, Accodingly, chromium is re-
moved from lnconel in the hot zone of o loop and
deposited as essentially pure chromium in the cold
zone. Since no diffusion process is necessary in
the cold zone (the chromium can deposit at the sur-
face of the Cr° crystais), the rate of attack is con-
trolled simply by the rate of diffusion of chromium
to the metalesalt interface in the hot zone.

The data in Table 2.2,5 also reveal, however,
that Inconel exposed to NaF-ZrF ‘-UF « mixtures is
in equilibrium with much lower CrF_ concentrations
than pure Cr° is in equuhbrwmwfl% when exposed
to the fluoride mixture under the same conditions.
Accordingly, it is not possible for chromium to
dissclve from 800°C Inconel ond to deposit at
600°C as Cr° when NaF-ZrF +UF , mixtures of
this general composition are c:rculated Loops
of pure Cr® would, of course, mass-trunsfer in
this medium; moreover, mass transfer can occur
if a sufficiently dilute alloy of chromium can be
formed in the cold zone. This suggests that the
mass-transfer process takes place in the followmg
general way.

The molten salt in the hot zone reaches equi-
librium with the 800°C Inconel and passes with
the dissolved CrF, to the cold zone. In the cold
zone, equilibrium is established by deposition of
a small amount of the chromium (by reversal of the
reaction) in the surface layer of the metal to form
an alloy containing slightly more chromium than is
normally present in Inconel. If no diffusion of
chromium were possible, c true equilibrium would
be reached shortly, with the hot surfaces slightly
depleted and the colder surfaces slightly en-
riched in chromium, The process would then
stop (except for the exchange process which has
no net effect), with negligible corrosion of the
metal. Since diffusion does toke place, however,
the process continues. In the hot zcne the con-
centration gradient causes a flow of chromium to

TABLE 2.2.5. EQUILIBRIUM CONCENTRATIONS OF CHROMIUM FLUORIDES WITH
ALKAL! FLUORIDE AND ZrF -BEARING FUEL MIXTURES

 

Chromium Concentration {ppm)

 

In NaF-KF-LiF-UF, In NaF-ZrF -UF ,

 

Experimental resuits for melt treated with
pute chromjum, (Cr) = 1,0*

At 600°C
At 800°C
Resulits calculated for equilibration of melt
with Inconel, (Cr) = 0.16*
At 600°C
At 800°C

1100 2400

2600 2550
710 1320
1660 | 1400

 

tConcentration of chromium in mole fraction.

98

 
 

 

the salt-metal interface, and the octivity of chro-
mium ot the surfoce is maintained at some level
appreciably less than 0.16. In the colder zones
the diffusion gradient is away from the salt-metal
interface, and the chromium activity is maintained
at some activity slightly higher than 0,16, Since
diffusion at the lower temperature is slower thon
that at the higher temperature, the rate-controlling
step is presumably the low-temperature diffusion
process. It may be that the intermediatestemperature
regions, in which the chemical driving force is less

but the diffusion rate is faster, are the most im-

portant regions; there is no a priori way of telling.
If the dynamicecorrosion and mass-transfer phe-
nomena cre examined in this light, it is obvious why
neither flow rate nor uranium concenfration of the
fuel mixture is an important factor affecting corro-
sion attack and mass trensfer in Inconel systems.,
If it is assumed that most of the corrosion observed
invelves mass fransfer to a dilute alloy, the lack of
an effect of the surface-to-volume ratio becomes
comprehensible. If the relative effect of diffusion
rate vs driving force is noted, the reason for the
poor correlation of corrosion with temperature drop
can be understood; with a ZrF ,-bearing fuel, corro-
sion may be worse with a top temperature of 1500°F
and a 200°F drop than with a top temperature of
1500°F and o 400°F drop.
- The argument points up the fact that, if the low
temperature were made sufficiently low, deposition

of Cr® might be possible in ZrF (bearing systems;

the temperature coefficient of the reaction is not
known below 600°C. It is possible that the light
frosting of chromium often observed in the cold

_dead-leg sump of oId-ster thermul-convechon loops e

may have been due to such-an effect, It would also -

- be ‘possible to -have alloys’ suffaclently high in
o t_'chromwm concentrohon (ond octivny) to cause
j?,chromwm deposmon to occur, These considera~
- -tions’ point to. the conscderoblo supericrity of alloys

OVer pre mefaIs fO'I' conIdIrlment Of moIIen saIfs.r':!-_f- fa"-.ed by f.ltrohon mefl-.ods. The change |n solu.

" bility at 600°C with total quantity of NiF, added
~ may be ascribed to the saturating phase being a
- complex compound rather than Nan. . The complex

EQUILIBRIUM REDUCTION OF NIF BY H2
IN NcF-ZrF

: C M. BIood |
Inveshgohon of fhe equshbnum S

NIF @ + H2 (g)‘-"-——‘Nl (s) + 2HF (g)

'when Ihe mixture NoF-ZrF (53-47 mole %) is used- -'

as the solvent was mmoted. The equipment and
experimental fechnlques for this study were very
similar to those described previously for the study

mixture NaF- ZrF (53-47 mole %)

PERIOD ENDING JUNE 10, 1956

of the reduction of FeF, by H, in NaF-ZrF . 911

Equipment changes mcfiuded the subsmunon of
nickel as the container material in place of the
mild steel used in the previous study. Operational
changes consisted in the use of hydregen-helium
mixtures of known composition instead of pure
hydrogen in the preparation of the equilibrating
gaseous streams. This was found to be necessary
after the results of preliminary experiments indi-
cated the necessity for using small partial pressures
of hydrogen and high partial pressures of HF in
order to maintain measurable quantities of NiF, in
solution.

The initial experiments were made in order to
ascertain the solubility of NiF, in the solvent
The values
obtained by fnltrohon of the saturated solution and
chemical analysis of the filtrate when o total of
0.17 wt % Ni was added as NIF are summonzed
in the following tabulation:

Temperature Solubllity of NII-'2
(°C) (ppm)
550 - 400
575 600
-~ 600 980
625 _ 1450

| Solubility values were also obtained by filtration

by Redman (see following section on *‘Sclubility

_ and Stability of Structural Metal Fluorides in Molten
NaF-ZrF "), who showed that the solubility is a
function of the quantity of NiF, added. Topol 12
‘also obtained solubility vulues by measuring the
“electromotive forces . of electrolytic cells. The
“values for the squblhty at 600°C that were ob-
_tained in the different investigations are shown in
- Fig. 2.2.2, There appears to be satisfactory corre-
" lation of these solubility values with those ob-

 

- 9C, M. Blood and G. M. Watson, ANP Quar. Prog. Rep.

Sept. 10, 1954, ORNL-1771, p 66.

. ‘°C. M. B!ood. ANP Qudf. Progo Repo Dec. 10' 19”0
ORNL-2012, p 85.

Ye, m. Blood and G. M. Watson, ANP Quar. Prog.

12 E, Topol ANP Quar. Prog. Rep. March 10, 1956.
ORNL-2061, p 89.

99

 
 

 

 

 

ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
ORNL-LR-DWG-44634

 

4000

 

 

/
3000 7

 

 

 

 

 

 

 

 

CONCENTRATION OF NiF, IN SOLUTION (ppm Ni)

 

2000
e REDMAN
= TOPOL
A BLOOD
1000
0
0 2 4 6 8 10

TOTAL NiF, ADDED {wt % Ni)

- Fig. 2.2.2. Comparison of Values of Seolubility
of NiF, in NoF-ZrF, (53-47 mole %) at 600°C
Obtained By Three Investigators.

compound ‘is probably, though not necessarily,
idenfical with the substance NiF .ZrF, previously
synthesized by Sturm. 13

Thermodynamic calculations performed prior to
the experimental work indicated that if NiF, ex-
hibited activity coefficients similar in magnitude
to those previously determined for FeF,, equili-
bration at 600°C with H,-HF mixtures containing
about 3% HF would allow sufficient NiF, to remain
in solution for satisfactory analysis. ly traces
of NiF, (<10 ppm) were found after repeated and
prolonged equilibration under these conditions and
even after the reactor temperature had been re-
duced to 550°C and the H,-HF equilibrating mix-
ture had been changed to contain 30% HF. Attempts
to increase the NiF, concentration in solution by
simply raising the HzF concentration of the equili-
brating gas were unsuccessful beyond NiF, con-
centrations of about 20 ppm Ni.

Nickel fluoride concentrations of sufficient mag-
nitude for an acceptable degree of precision in the
analytical determinations were finally obtained by
lowering the partial pressure of hydrogen, in.addi-
tion to maintaining a high HF concentration in the
equilibrating gas mixtures. This was accomplished

by the use of hydrogen-helium mixtures rather than

pure hydrogen. The mixtures were prepared by
injecting helium gas into a partially empty hydrogen

 

13g, J. Sturm, ANP Quar, Prog. Rep. Dec. 10, 1955,
ORNL-2012, p 91.

100

cylinder. The compositions of the mixtures thus
prepared were determined by the proper adaptation
of the hydrogen combustion analysis technique to
the existing experimental assembly., The experi-
mental results are summerized in Table 2,2.6.

As is customary, equilibrium was approached
from both directions in obtaining the series of
measurements described. It is of interest to note
that the calculated values of the equilibrium con-
stants for the reaction range from 11 to 45 atm at
625°C when the solid and the supercooled liquid,
respectively, are used as standard states. The
experimental values denote activity coefficients
of NiF, a thousand times greater than those ob-
tained for FeF,. Detailed numerical calculations

TABLE 2,2,6. EQUILIBRIUM CONSTANTS FOR THE
REACTION NiF,(d) + H,(g)==Ni(s) + 2HF(g)

 

 

AT 625°C
Nickel Content Hydrogen HF -
of Melt Pressure Presswe Kx*x 10-4
(ppm) (atm) (atm)
77 0.0628 0.502  3.06
75 0.0661 0.475  2.68
75 0.0640 0.492 = 296
75 0.0595 0.528 3.67
75 0.0617 0.510 3,30
85 0.0628 0.501 2.76
75 0.0640 0.493 2.97
75 0.0643 0.490 2.92
185 0.0251 0.456 2.74
165 0.0253  0.460  2.98
160 0.0255 0.457 3.00
150 0.0259 0.448 3.03
140 0.0250 0.447 3.22
120 0.0270 0.425 . 3.28
140 . 0.,0270 0.425 2.81
125 0.0269 0.427 3.19
135 0.0274 0.416 2.74
125 0.0276 0.412 2.89

Av 3.01 10.18

 

*K, = PElF/xNIszfiz‘ where X is mole fractior? and

P is pressure in atmospheres.
 

 

 

of this reaction will be made ot the conclusion of
the experimental work.

Experimental work is now in progress at 575°C,
Efforts will be made to extend the NiF, concen-
tration ranges and to obtain the temperature de-
pendence of the equilibrium. Particular attention
will be given to the determination of the activity
of NiF, in the saturating phase.

SOLUBILITY AND STABILITY OF STRUCTURAL

METAL FLUORIDES IN MOL TEN NuF-ZrF4
| J. D. Redman

The results of studies of the stability and solu-
bility of CrF, in NaF-ZrF, (53-47 mole %) at

600°C were presented previously.'®# From these

experiments it was concluded that the solubility
of CrF, increased as the zirconium-to-chromium
ratio decreased, and, accordingly, the solubility

of CrF, in this solvent was a function of the

amount of CrF, added, It was postulated that the
change in so|u2|:ili1y was due to the separation of
CrF,.ZeF, as the solid phase, with a resultant
increase in the NaF concentration in the melt,
Some of these experiments have been rerun in order
to obtain more precise values at the larger zirconium-
to-chromium ratios, and the values thus obtained,
along with some of the earlier ones, are given in

Table 2.2.7. .

 

145, D. Redman, ANP Quar. Prog. Rep. Dec. 10, 1955,

ORNL-2012, p 88.

"PERIOD ENDING JUNE 10, 1956

Similar studies have been carried out on the
solubility of FeF, and NiF, in this solvent, and
the data are presented in Tables 2,2,8 and 2.2.9,
respectively, The data for these three structural
metal fluorides also are plotted on Fig, 2,2.3.

An examination of the data presented for 600°C
shows that the solubility of CrF, increases quite
rapidly as the zirconiumeto-chromium ratio de-
creases over the entire range studied. A less
pronounced effect is noted for FeF,, and in the
case of NiF, @ slight increase is observed for
the higher zirconium-to-nickel ratios. At the lower
zirconium-to-nickel ratios the solubility is con-
stant, These differences in behavior are shown
quite clearly in Fig. 2,2.3. It appears that the
solid phases present for the GF, and FeF, sys-
tems contain ZrF ., probably as MF,.ZrF ,, whereas
the solid phase in the NiF, system appears to tie
up only very small amounts of ZtF,, It may be
noted that at 800°C the solubility of NiF, is inde-
pendent of the zirconium-to-nickel ratio, which
suggests that the solid phase present is NiF,,

There is some question regarding the reliability
of the data given in the zirconiumsto-sodium ratio
columns in the tables, but for Cer and for FeF,
at 600°C it is believed that the decrease in the
ratio as the excess of metal fluoride added was

_increased is real. The ratio of zirconium to sodium

is 0.88 in the starting material, and the consider-
ably lower values found for the large additions of

TABLE 2.2.7.. SOLUBILITY AND STABILITY OF CeF, IN MOLTEN NaF-ZcF , (53-47 MOLE %) AT 600°C

 

Conditions of .

‘Found in Filtrate

 

 

- Equilibration _

Ce*t ZrtoGr It ‘Na Zr-to-Na F ot Total Cr
(wt %) Rctfip* (wt %) (wt %) Ratio* (wt %) {wt %) Awt %)
1.4 w7 - 415 13,3 0.79 - 45.3 0.66 0.73
1.4 7 2.3 3.4 0.78 44.8 0.65 0.83
40 58 8 127 0.79 45,1 2.1 e
4.0 58 9.5 129 0.78 452 . 20 1.9
64 34 38.4 6 07 8 3.5 3.5
64 34 389 141 e 443 34 36
9.6 21 7.9 . 13.8 069 . 440 .50 52
9.6 2.1 36.3 13.4 0.68 44.2 5.7 5.9

 

*Ratio calculated from mole fractions.

101

 
 

 

 

ANP PROJECT PROGRESS REPORT

 

 

 

 

_TABLE 2.2.8. SOLUBILITY AND STABILITY OF FeF, IN MOLTEN NaF-Z¢F , (5347 MOLE %)
AT 600 AND 800°C :
- Conditions of Equilibration Found in Filtrate
Temperature Fe*t  Zrto-Fe  Zr Na. . Zreto-Na F  Fe*t  Total Fe

{°0. (wt %)  Ratio* (wt %) (wt %) Ratio* (wt %) (wt %)  (wt %)
600° 0.5 50 415 13.7 0.78 45.7 0.26 0.26
| 0.5 0 419 12.9 0.82 45.3 0.25 0,26

1.0 25 41.6 13.8 0.78 452 . 0.33 . 0.32

1.0 25 4.6 13.6 0.78 45.0 0.35 0.35

5.0 4.8 40.3 1.5 0.76 44.3 1.3 1.4

5.0 4.8 39.8 13.8 0.73 44.4 1.4 1.5

10,0 2.1 37.9 14.0 0.69 45.0 2.9 30
100 2.1 37.5 3.7 0.70 446 3.5 3.5

800 - 6.0 3.9 38.9 10.8 0.91 45.2 6.1 6.0

| 60 3.9 38.5 10.6 0.91 441 4.8 5.8

120 17 34.7 0.0 - 0.8 87 07 1.7

12.0 17 34.5 9.4 0.93 47 N3 1.9

18.0 1.0 32.2 8.8 0.93 41 4 14.2

18.0 1.0 32.1 9.2 0.88 440 128 14.0

 

*Ratio calculated from mole fractions.

TABLE 2.2,9. SOLUBILITY OF NiF, IN MOLTEN NaF-ZrF , (53-47 MOLE %) AT 600 AND 800°C

 

Conditions of Equilibrdfion

Found in Filtrate

 

 

Temperature Ni Zr-to-Ni Zr No Zr-to-Na F Ni
(°C) (wt %) Ratio* (wt %) (wt %) Ratio* (wt %) (wt B
600 0.5 43 4.2 13.0 0.80 45.2 0.14

| 0.5 8 418 13.6 0.78 451 0.12
1.5 18 41.4 12.8 0.82 45.5 0.17
1.5 18 41.4 12.5 0.84 45.0 0.19

5.0 5.2 40,4 14.1 0.73 45.3 0.32

5.0 5.2 39.8 14.0 0.72 44,7 0.31
10 2.4 40,0 14.0 0.73 451 0.32

10 2.4 40.8 12.8 0.81 45.3 0.31

800 1.5 18 41.6 12.0 0.88 45.2 1.0
1.5 18 41,5 12.1 0.87 45.3 - 1.1

5.0 5.2 41,6 12.3 0.87 45.6 1.2

5.0 5.2 41,8 11.8 0.89 45.4 1.3

10 2.4 417 12.6 0.84 44,6 1.0

10 2.4 0.3 12.2 0.87 447 13

 

*Ratio calculated from mole fractions.

102

C
 

 

 

PERIOD ENDING JUNE 10, 1956

UNCLASSIFIED
ORNL-—-LR—-DWG 14635

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6
5 pd
Oé, /
o
Y
i 4 7
: .
s /
e .
. 4
> / "
5 . 0090/
3 / gefz)
Q 2 ]
m / ~ /
A/
_ NiF,; BOO°C
OF‘// i
A | NiFy; 600°C
o et ] 1
5 6 7 8 9 10

0 { 2 3 : 4

METAL FLUORIDE ADDED (wt % M)

Fig. 2.2.3. Eflect of Amount of Metal Flucride Added on Solubility of NiF,, FeF,, and Cer

NeF-ZrF  (53-47 mole %) at 600°C.

metal fluoride are odditional evidence for the
belief that ZrF is combined with the metal fluo-
ride in the solrd phase,

 

mole %) were centinued through the use of the con-
cenfration cells discussed previously,15 Cells of
the type

MIMF 2% ZrF (cy), N_aF-ZrF4| NaF-ZF,, MF xZrF (c )M

 

A comparlson of va Jlues given for Fe""" and total
iron shows that Fett
tions, The slight devnahons noted are attrrbutoble
to expenmentcl errors. The ' ogreement between

it is evident that more than 90% of the chromium s
present as’ Cr'“' Since the starting materaal does
not contain over 95% of the chromium as Crtt, it -
- s evrdent that no opprecrable dlsproportlonotron of
".Cer occurs under the condltlons employed

SOLUBILITY DETERMNAT'ONS BY MEASURE"’ to ‘measure the temperature of either of the half

cells - durectly. T ‘ :
. thermocouple placed in g well that was permanently

o MENT or= ELECTROMOTIVE FORCES
OF CONCENTRATION CELLS '

L. E. Topol

Solubility studies of structural metal fluorides
and their complexes in molten NaF-ZrF, (53-47

. celis. .
as a functlon of temperature, the saturation points

* where M was Fe; Nl, o‘r Cr,and x = 0 or 1, were
.is stable under these condi=

measured at 525 to 750°C in a helium atmosphere,

~ The "containers ‘were alumma, platinum, or the
- metal M when M was iron and-nickel; a Zr0, bridge
- Cr** ond total chromtum is not so- satrsfactory, but

lmpregnoted with NaF-ZrF (53-47 mole %) served
as the: electrlcal contact between the two half
‘When the eiectromotwe force was measured

were apparent from the points of dlscontmuny in

: | the P|°f of yoltage vs temperature. -

ln the cell ussembly ernployed it was not feaslble

The temperature recorded by a

 

151, E. Topol, ANP Quar. Prog. Rep, March 10, 1956,
ORNL-2061, p 89.

103

 
 

 

 

ANP PROJECT PROGRESS REPORT

fixed midway between the two half cells was
- formerly used as the cell temperature. In order to

refine the temperature measurements, a calibrated
thermocouple, enclosed in a nickel tube, was sub-
stituted for an electrode in a dummy cell. In the
temperature range of interest (550 to 750°C) the
melt temperatures found were 10 + 2°C higher than
the melt temperatures at the center between the two
crucibles, Thus all the temperatures reported in
connection with previous potential measurements

-should be revised upward by 10°C, Only the solu-

bility data are sufficiently temperature sensitive
to be appreciably affected.

Application of the temperature correction to the
solubility data previously reported!5 gave the
following new expressions for the solubility of

MF 5. Z¢F , in NaF-ZrF , (53-47 mole %):

n3

log N = = 00X 197 (67, for FeF o2eF,,
40.1 x 10°

|og N = —m—* + 8.36 ’ fOl‘ Cer'ZI'F4,
28.8 x 103

log N = 'Ts;is_r + 472, for NiF 2 ZcF 16

where N is the mole fruchon of MF -ZrF and T is
temperafure in °K. The corrected hects of solution
and *‘ideal” melting points are now 36.0, 40.1, and
28.8kcal/mole MF ,«ZrF, and 875, 775, ond 1060°C
for the Fe, Cr, ond Ni compounds, respectively.
Solubility data obtained by analyses of filtrates
from saturated solutions, as described in the pre-
ceding sections of this chapter (‘‘Equilibrium
Reduction of NiF, by H, in NaF-ZrF '’ and *‘Solu-
bility and Stublhty of Structural Metol Fluorides in
Molten NaF-ZrF '), did not agree with the elec-
trometrically determined solubilities, particularly
in the case of NiF,, Since the filtration measure-
ments were carried out by adding NiF, rather than
NiF,ZrF, as solute (and similarly with FeF, and
CrF 2)s the electrometric determinations were re-
peated with Nth as solute so that the data could

 

' léFréh this fnvestigation and xeray and petrographic
examinations of quenches of NaF-ZrF, melts, NiF, is
now believgd to exist in a ternary compound, rather than
as NiFyeZrF,, in saturated solutions. The equation
for the solubility was revised to include additional

meus'uremeriis, 'qs well as the temperature correction.

104

be compared. X-ray and petrographic examinations
of the solidified melt from the half cells should
have sufficed to identify either or both of the joins
as quasi-binary; in practice, however, both
NiF.‘,-ZrF4 and a new phase were found in the
slowly cooled half cells. The new phase is thought
to be a ternary compound, since it has never been
found in the binary systems.

Gradient quenches (510 to 540°C) along the two
joins revealed only the new phase, with no
NiF,ZrF,.  Such ambiguous results were dis-
appointing, since on a quasi-binary join the solu-
bility does not vary with the amount of solute
added, as it does along a random join.

Cells containing MF, in NaoF-ZrF, (53-47 mole
%), with no addition of ZrF,, gave the solubilities
of the structural metal solts shown in Table 2,2,10
and Fig. 2.2.4 for NiF, and Table 2.2.11 for FeF,,.
The solubilities of the nickel salts along the joins

NiF - ZrF -7NaF+6ZrF, and NiF,-7NaF+6ZrF,, as

we!l as fhe values for NiF, obtomed by filtration

 methods, are plotted in Fig. 2 2.4. The discrepancy

between the results of the various experiments
has not been explained. The electrometrically
determined solubilities appear to be virtually the
same whether the solute is added as NiF2 or as
NiF,.ZrF,, and the same seems to be true for

TABLE 2,2,10. SOLUBILITY OF NiF, IN
NaF-ZrF , (53-47 MOLE %)

AH’;M = 28.8 kcal

Ideal melting point = 1060°C

 

 

 

Temperature | NiF, Solubility
(°C) wt % mole % .
553 - 0.12 0.12
564 0.14 0.15
574 " 0.9 0.20
576 0.20 o 0.21
650 0.79 | 0.83
650 0.93 0.97
655 0.88 0.92
669 0.98 1.0
706 1.97 2,05
725 21 2.29

 
 

 

- SOLUBIITY (mole fraction}

%

FeF,. The FeF., saturation points were not so
2 2

well defined or reproducible, however, as those for

FeF,.ZrF,, especially ot temperatures above

650°C. The potentials of many of the cells flux-

tuated greatly and were much lower than the values

predicted from the Nernst equation (where complete

UNCLASSIFIED
0.0 ORNL-LR-DWG 14636

EMF DATA:
. NF
0.05 : © NiFg-2rf4
FILTRATION DATA FOR NiF2:
A VARYING TEMPERATURE (BLOOD)
A° VARYING AMOUNT OF SOLUTE (REDMAN)
0.02
FROM EMF DATA

Q.04

0.005

FROM FiLTRATION DATA ™

0.004

 

0.0005

800 750 700 650 600 550 . 500
TEMPERATURE (*C)

Fig. .2.2.4. = Solubility. of NiF, and NiF,.ZrF
in NaF-ZrF, (53-47 mole %).

TABLE 2.2,11. TENTATIVE VALUES FOR THE
SOLUBILITY OF Fer IN NaFZrF , (53-47 MOLE %)

 

 

 

Temperature FeFy Solubllity :
°c oWt % " mole % '
53 013 o4
561 025 027

%5 . 031 034
8 . 072 . o018
e . 18 . 116
60 w0 s

66 244 263 .
665 . o 323 34y
70 122 775
742 9.9 10.59

 

PERIOD ENDING JUNE 10, 1956

solution in both half cells was expected). X-ray
and petrographic examinations revealed a better-
defined temary complex in low-temperature (535°C)
quenches of FeF, melts than in high-temperature
(672°C) quenches; the FeF, ternary complex gives
the same x-ray pattern as that given by the NIF
ternary complex,

FREE ENERGIES OF FORMATION OF COMPLEX
"METAL FLUORIDES MF ,+Z+F , '

L. E. Topol

For a saturated solution of MFZ.ZrF4 in a molten
electrolyte, the following equilibrium holds:

MFZ.ZrF4(crys!)=MF2(so|n)
+ ZrF y(gotny + AF = 0.

In NaF-ZrF, (53-47 mole %) as the solvent, the

activities of MFz("ln) and ZrF“uln) are known,

- at least approximately, in a few cases, and, since

the activity of solid MF,.ZrF, is unity, the free
energy of formation of M‘F2-ZrF4 is the only un-
known in the following relations as applied to the
saturation equilibrium:

aManZrF4
= 0 = AFO + RT |n

 

a
MFZ'ZUF4
= AF%_ 4 AFS - AF§
MF2 ZrF4 MFz'ZrF,4

-where N is mole fraction and a and y are, respec-
- tively, the activity ond the activity coefficient
" based on the pure solid as the standard state. The
" free ‘energy of complexing of MF -ZrF from solid
o 'MF and solid Zrl":4 is o

omp

_ . : A; o ° .
'.AF AFMFQ-ZrF" - AFfiE’? -AF,Z'F‘

- 2.‘3 RT |09 yME NMF#“Z:F ‘

o Hence a knowledge of lhe acflvn‘y of the dissolved
- MF, ‘and ZrF , suffices to yield the free energy of |

‘complexing, AF this quantity has been cal-

comp'

" culated for the proposed family of complexes

GF, ZrF FeF,ZrF,, and NiF,-ZrF,, as shown
in Toble 2 2 12. Smce for NIF ZrF . AF is

comp

positive, there is additional evudgnce that this

105

 
 

 

 

 

ANP PROJECT PROGRESS REPORT

TABLE 2,2.12. FREE ENERGY OF FORMATION OF MF »ZrF

 

Free Energy of

 

 

, S T Saturation A o Free Energy of
o, lemperature ofs ctivity _ . o
MPrZFe g Solobilits N e tonty  CoPlexing AF o, Formation, AF
) - : {mole fraction) (kcal/mole) (kcal/mole)
700 0.0550 2.20 -8.3 - 520 _
NiF . ZrF, 600 - 0.00331 1900 +0.9 ~510
700 0.0180 1100 +1.6 ~500
CFpZeF, 600 0.0210 115 -8.8 ~545
700 - 0.183 0.87 7.8 -535
solid complex is not stable with respect to solid 4 P

The activities of MF, in the saturated solution
were obfained by combining solubility and activity
coefficient measurements. Henry’s law is expected
to hold quite well for dilute solutions in molten
electrolytes because the environment of a solute
ion is not significantly influenced by the presence
of other solute ions unless appreciable concentra-
tions are present. This has been demonstrated for
FeF, in NaF-ZrF, (53-47 mole %) over o range of
concentrations from 0.03 to 0.06 mole % by meas-
urements of the activity coefficient based on equi-
librium studies at 600 to 800°C.!7 The values of
the activity coefficients for FeF, (y at 600°C is
3.28; y at 700°C is 2.20) were combined with
electrometrically determined activity coefficient
ratios 18 (from bimetallic couples) to give the
activity coefficients of CrF, and NiF, shown in
Table 2,2,12. Previously published measurements15
showing a twofold change in yg, g, with concen-
tration in the range’ 0.06 to 1.5 mole % FeF, are
tentatively regarded as in need of reinterpretation.

The saturation concentrations listed in Table
2.2.12 were obtained from the equations presented
in the preceding section on ““Solubility Determina-
tions by Measurement of Electromotive Forces of
Concentration Cells."

The activities of ZrF , were computed from vapor
pressures by using the expression

 

17¢, M. Blood and G. M. Watson, ANP Quar. Prog.
Rep, March 10, 1956, ORNL-2061, p 84, -

- 18, E, Topol, ANP Quar. Prog. Rep, Dec. 10, 1955,
ORNL-2012, p 97. ‘ 3

106

where p is the vapor pressure of ZrF in NaF-ZrF,
(53-47 mole %) and p, is the vapor pressure of
solid ZrF . Since

3
logp - 9_243 - M
and
12.376 x 103
log p = 13.400 ~ ——r— .
3
Iog_p_= 415 3.124 x 10 o
" b T

where T is temperature in °K, which gives

log £— = —0.578 at 600°C ,
o

The equation for p, was obtained from the work of
Sense et al.'? The activity of ZrF , in the satu-
rated solution was assumed to be the same as the
activity in the pure solvent NaF-ZrF 4 (53-47
mole %). This assumption is probably valid in all
cases except that of CrF, ot 700°C; here the solu-
tion is too concentrated for much reliance to be
placed on the calculation. If the saturating phase
were simple NiF, in the NiF, experiments, then

 

19k, A. Sense et al, Vapor Pressures of the Sodium
Fluoride—~Zirconium Fluoride System and Derived In-
formation, BMI-1064 (Jan. 9, 1956).
 

aNle' = N should be unity (a pure solid has unit
activity). The apparent activity is 9 at 600°C and
20 at 700°C, however, which indicates that the
saturating phase is something less soluble than
NiF,. The discrepancy amounts to a factor of
10 m the activity coefficient or the solubility and

to about 5 kcal/mole in the value of AFNI‘Fz; In

computing the AF® for the formation of the complex
compound from the elements, the free energy esti-
mates given in Table 2,2,13 were employed. 20

 

201, Brewer et al, p 104=115 in The Chemistry and
Metallur of Miscellaneous Materials, Thermod ymamics,
NNES 1| 'I 'B, ede by L. L. Quill, McGraw-Hill, New
York, 1950. ,

TABLE 2.2,13. SELECTED VALUES OF
FREE ENERGY OF FORMATION

 

 

 

AF® {kcal/mole)
At 600°C - A+700°C
FeF,  -137.5  -1339
CeFy ~152.5 -148.9 -
NiF, 1279 1248
ZrF, -383.4 ~376.8

 

-PERIOD ENDING JUNE 10, 1956

An independent check on the differences in the
standard free energies of formation of the MF . ZrF
complex from the elements is afforded by previously
published data obtained for bimetallic couples in
saturated half t:ells.z'I To a first approximation,
the cell reachon was ’ ’

M+MF‘.ZrF = MFZrF, + M" .

'Thlsreactlon neglects transport and junction poten-

tials, which are small if the saturated solutions

are sufficiently dilute. From the relation

End = -AF

for the cell reaction, where E is the electromohve

force, n is the number of equivalents transferred,
ond J is the number of Faradays passed, an approx-
mate value of the difference in the free energy of
formation from the elements is found, The com-
parison is shown in Table 2.2.14,

The values of AF, , in Table 2.2.14 were ob-
tained by subtrachng the values of AF® given for
these complexes in the last column of Table 2.2.12,
White the agreement is good, the effect of the high

-solubility of CrF, at 700°C is probably responsible

for the larger of the discrepancies.

 

211, E, Topol, ANP Quar. Prog. Rep. Sept. 10, 1955,
ORNL-I947, P 85.

TABLE 2,2.14. COMPARISON OF FREE ENERGIES OF FORMATION OF MF,+ZsF,
OBTAINED BY TWO METHODS '

 

~ Difference Betwee_n AF

and AF,, » '
(keal/mole)

 

 

 

At 600°C : ~ Ar700°C’
~End 155 | 6.8
AFcalc ' 15 o ' _ 'IS
. =End -19.5 2200
cale -20 , :‘ - =20
. -End 35,1 . =363
AFeulc =35 =35

 

107

 
 

 

 

ANP PROJECT PROGRESS REPORT

PREPARATION OF SOLUTIONS OF LaFg,
BaF,, AND RbF IN NaF.Z¢F (-UF

F. L. Ddley : - W, T. Ward

‘A special preparatlon containing LaF, -BaF ,°RbF
(2,8-1.0-1.1 mole %) dissolved in a sfandard NaF-
ZrF ,-UF 4 (50-46-4 mole %) mixture was requested
for use in determining whether the introduction of
the simulated fission products would produce sig-
nificant viscosity changes, The usual methods of
purification and filtration were used, and the equip-
ment was modified so that small filtered samples
could be dispensed directly into the tubes used for
the viscosity measurements, The resulting preparc-
tion was studied to obtain some indication of the
solubilities of the fission products at 800°C in
this solvent. It is believed that the solubility of

LaF, may be typical of the solubilities of the rare-

earth fluorides, since LaF, presumably forms solid
solutions with other rare-earth fluorides.
Two “filtered samples of the NaF-ZrF -UF , mix-

ture without the additive were obtamed ond two

mote samples were obtained cfter the fission.

product mixture was added. A small portion of
solution "which remained unfiltered in the reactor
after withdrawal of all the samples was also re-
covered for analysis. All filirations were performed

at approximately 800 + 10°C, During the course of
the ‘viscosity measurements, the presence of a
finely divided solid material was detected in the
molten mixture, and a small sample of the solid
was obtained by decantation. The solid was
tentatively jdentified by x-ray diffraction analysis
as free LaF,, with occluded solvent. The resulits
of chemical analyses of all the samples taken
are presented in Table 2,2,15.

The results indicate no greater concentration of
odditives in the unfiltered portion (reactor heel)
than in the filtered samples, and consequently it
appears that the additives were completely dis-
solved at 800°C. From visual observations, how-
ever, the solution seemed to be essentially satu-
rated at this temperature, The saturating phase
appeared to be free LaF,, and the probable solu-
bility value for this compound at 800°C seems to
be about 2,5 wt %. During the preparation of this
material no particular effort was made to control
the filtration temperature at exactly 800°C, and
visual observations could not be made conveniently
enough to fix a liquidus temperature; however, the
order of magnitude of the solubility of LaF, in
NaF-ZrF -UF, (50-46-4 mole %) appears to be
about 2 mole % at 800°C.

TABLE 2.2,15. CHEMICAL COMPOSITION OF NaF-ZrF ,.UF, SAMPLES BEFORE AND AFTER
INTRODUCTION OF LaF4-BaF,-RbF TO SIMULATE FISSION PRODUCTS

Filtration temperature, 800°C

 

 

 

Major Constituents (wt %) Impurities (ppm)
Na Zr U Ba La Rb F Ni Cr Fe
NGF-ZI'F4-UF4 -
Sample 1 . 10.7 38.8 8.8 0 0 0 41,9 9  50- 65
Sample 2 10.3 38.8 8.7 0 0 42.5 35 40 45
Theoretical 10,41 37.98 8.2 O 0 0 42.99
Na F‘Zde-UF“ plus
Lqu-Bon-Rb.F
Sample 3 9.2 35.7 8.0 .23 2.62 0.86  41.7 80 50 90
Sample 4 9.5 34.6 7.7 L16 2.5 .00 41,8 100 65 160
Reacfor hee' 8.9 37.8 8.2 lc '8 2.3‘ 0087 4‘-2 25 ) 65 150
Theoretical 9.64 35.17 7.98 1.25 3,40 0.82 41.74 '
Solid detected during 18.1 0.62 2.6

viscosity measurements

 

108
 

 

 

-

Although it appears that the solubilities of the
rare earth fluorides in the NaF-ZrF .UF, mixture
are sufficiently high to prevent prec:pltutlon upon
their formation as :fission products during ART
operation, some studies of solubility of individual
rare earths and rare-earth mixtures are currently
under way. One possible. development from such
a study might be the demonstration of a simple
means for removal of the rare-earth fluorides from
the fuel mixture if the magnitudes and temperature
dependences of the solubilities are adequate.

PRELIMINARY SOLUBILITY VALUES OF CeF,
IN NGF'Z’F4-UF4

C. M. Blood

As a start of a systematic determination’ of solu-
bilities of the rare-earth fluorides in ART-type

fuels, such as NaF-ZrF +UF, (55-40,-57mo!e %),

one experiment was performed to determine the
approximate solubility of CeF, in a typical solvent,

A previously purified mixture of the NaF-ZrFA-UF

solvent and approximately 2,5 mole % CeF, were
enclosed in a nickel container and heated to 850°C.,
The mixture was continuously stirred by bubbling
hydrogen and helium, alternately, through o dip leg.
Two hours prior to filtration, the temperature of the

solution was lowered to 800°C.  Samples of liquid

were then withdrawn through a nicke!l filter tube at
800, 700, and 600°C. A second set of filtrates was
obtained in the same manner at 700 and at 600°C
after the mixture had been given an odditional

“hydrofluorination and hydrogenation trecitrhérih The =
. are used and the mixture of salts contains m moles
of. AgCl per mole of NoC! N faradays of electricity -
are pussed ond the *‘idealized” anode compart-

results are summarized in Table 2.2, 16

Smce the CeFa' used in these experlmenfs was -
) nof es.pecmlly pure ond certam dxffu:ulties werei_";;”'r;; ,
L e o Rl RISt o f_r_‘ment .gains N moles of Ag* by electrode reaction
TABLE 22,16, l*APéizdxmrrE sbLU'Bl’Lr‘rv.op"f"" *' '."“1 N‘c; moles of CI= by transfer and loses Nt}

L . , : A e gy 0. moles ©
_~“This may be summcnzed as a gain of N moles of

 

GF N NaF P (U (S5ABSHOLE )

 

 

Temperuture

 

 

to W'M; T MelenCs
o ey 0k
e0 05 05
0.59

 

" onode ‘compartment has lost N’

- PERIOD ENDING JUNE 10, 1956

apparent in analysis of the filtrates obtained, these
date should be considered as preliminary. |t is
anticipated  that more relioble values will be

. available within the next quarter,

CONICEN"I'.RATION CELLS AND TRANSFERENCE
NUMBERS IN FUSED SALTS

R. F. Newton

It is well known that transference numbers are
needed for correlating the electromotive forces of
concentration cells with activities or other thermo-
dynamic properties of the solutions involved, While
measurements have been and are being made of the

- transfer of ions of fused salts across porous dia-

phragms, some -investigators doubt that these data

"can be properly interpreted as the ‘'true’’ trans-
ference numbers of the ions, since the diaphragm
_itselfmay cause a flow of salt across the boundary,
1t is of interest to note, however, that relative

transference numbers, which are the values of
thermodynamic interest, can be measured in an
ordinary Hittorf experiment,

At is well known that the transport of water as
part of a hydrated ion in aqueous systems cannot

be  determined without some rather questionable

assumptions, but standard transference numbers of

aqueous ions are given relative to the water. For

example, in a mixture of two salts, such as AgCl
and NaCl, the transference number of Ag* may be
defined relative to the NaCl. The fictions of the
ideal transference numbers, ¢ , 7 , and #g, are
useful in this analysis. In an ordinary Hittorf
transference experiment in which silver electrodes

Na*.and NtA ‘moles of Ag by transfer.

‘ 7 CeF, Solubility St T AgCl- by electrode - reaction and a loss of NtA
. moles  of- AgCI and of Nt’

moles of NaCl by |

frcmsfer. The best iha'r can: Ee deone exper:mentally,
S ":however, is to compore ‘the AgCl in a given quantity
-of NaCl ot the end with that in the same amount of
" NaCl at the start,’ 1, for example, the anode portion
i - taken for onalysw at the end of ‘the experiment
. ‘contains one mole of NaCl, and the *‘idealized””

Na moles of NaCl,

this represents 1 + Nt;, moles of NaCl at the

_ start. The AgC! at the start was - m + mNt(, ¢ and

109

 
 

 

 

ANP PROJECT PROGRESS REPORT

the AgCl at the end was
(n N + m + mNtj,_ - Neg o,

where m is the AgCl in one mole at the start and
N is the AgCl from the electrode reaction. If m and
N are subtracted from expression 1 and the re-
mainder is divided by N, the relative transference

of Ag¥ is seen to be

(2 fag = thg = Mg o
By a similar argument it may be seen that

(3) . tE"—']—tAgcr

These reluhve transference numbers are all that
can be obtained unambiguously from the Hittorf
experiment, in which N faradays of electricity are
passed, the electrode compartments are analyzed,
and the AgCl content of the electrode compartment
is compared with that initially present in the same
amount of NaCl. They are also the only numbers
needed for thermodynamic use.

Useful correlations of the relative transference
numbers with electric conductance may aisc be
made, For these carelations, A (the ionic con-
ductance) is defined so that, for pure NaCl,
Aya + Mgy = k/C, where k is the specific con-
ductance and C is the concentration in equivalents
per cubic centimeter, For the NaCl-AgCl mixture,

K = Cualna + Cagrag + Coirey
CNuxNu
K

CA gA'Ag

~
>
o
H
-

K

Cerrel

tl, =
cl -

and, when these values are substituted in Eq. 2,
it may be seen that

CAg
C

 

r - 37 ’
4 the = Thg = tq
Na

110

c AAg = A'Nu
= Ag -
I\Ag = I\Nd
= 7R —,
K/CNu '

From Eq. 4 it is evident that the relative fransfer
of the silver ion is propertional to the concentra-
tion of the silver ion and to the difference in the
mobilities of Ag* and of Na*, and hence the relo-
tive transfer may be either positive or negative. If
it is negative, t,, will be greater than unity.

For apphcahon to a concentration cell with
transference, consider the cell

Ag, AgCl in . . AgCl in NaCl,
NeCl; m® mole | . Ag; mB mole
AgCl/mole NoCl . - Ag/mole NaCl

tT T+ dt”

At the anode, for each faraday pass.ed,
Ag » Agt(m?) + e,

A ,
A o mole Ag*(m4) is lost by transfer, and

A
(l - t'Ag) mole C| = is gained,

At the cathode the opposite processes occur, with
each A replaced by B. In the section with a con-
centration gradient between A and B, ‘in each
element, such as between the dotted lines, the
gain of AgCl is 7, and the loss is £} + th '
which gives a net oss of dt' mole AgCl in the
element, and

AF::(] —!rAAg)FA- (l—tr )

= FA _ B B
=FA-FB+ fBy dF .

fBth'

If ¢’ is constant this simplifies to
TA EB) .
AF = (1 -1, ) (FA - FB);

however, as shown in Eq. 4, " is not expected to
be constant,
 

 

 

 

it

PERIOD ENDING JUNE 10, 1956

2.3. PHYSICAL PROPERTIES OF MOLTEN MATERIALS

F. F. Blankenship

G. M, Watson

E. R. Van Artsdalen’

PRESSURE-COMPOSITION-TEMPERATURE
RELATIONS FOR THE SYSTEMS
KF-ZrF, AND RbF-ZrF,

S. Cantor

Among the important guiding principles in the
search for improved or modified fuel mixtures are
predictions regarding the changes in properties
that will occur as a result of the substitution of
one species of ion for another in an otherwise
similar melt. To a very considerable extent these
predictions can be based on a determination of
how the activity coefficients of the constituents
change with composition. One of the easiest
activities to measure, and a fairly important one
to know, is the activity of ZrF in mixtures with
alkali fluorides at composmons such that the
vapor is essentrally pure ZeF . Mlxtures con-
taining 45 mole % or more ZrF4 produce a vapor
which, for practical purposes, is pure ZrF,, and
the activity of ZrF, is readily obtained from the
total vapor pressure. Accordingly, measurements
are being made of the pemnent vapor pressures
in alkali fluoride~ZrF, systems. It has been
found that the volafll:ty of the ZtF “varies in-
versely with the size of the alkali cahon. This
is a consequence of the more pronounced com-
plexing of ZrF, in the presence of the alkali
cations, which have less attraction for fluoride
ions. Lithium fluoride represents the case of an

,outsfandmgly strong attraction of ‘a cation for -
fluoride ions because of the large charge-fo-radws -
ratio “of ihe llfluum ion, Correspondmg1y, the =
vapor - pressures from 'the LIF-ZrF4 muxturearef
_the _highest; “this 'system is curreritly belng,'_
: ‘measured at: the Battelle Memorlai Tnstitute. - "The -
- NaF-ZrFy - mixtures - have ‘been - measured at
~ORNL1=3 qgnd at. Batteile.‘ Vapor pressures for
- ’KF-ZrF and RbF-ZrF‘ are currenfly bemg de-

 

‘R. E. Tfaber, Jr., R, E.. Moore, ond C. J Barfon, _
‘ 'Z,ANP Quar. Prog. Rep. Dec. 10, 1953, 0RNL-]649, P99

o 2RE, Moore and C."J, Barton, ANP Quar. Prog. Rep -
.'_}une 10, ‘1954, ORNL-1729, p 101. - e '

B . E. Moore, ANP Quar. Prog._Rep. Sept 10 1954, S
0RNL"|77]| p129. -

K. A. Sense ef al., Vapor Pressures of tbe Sodium

Fluoride~Zirconium Fluoride System and Derived Infor-
mation, BMI1-1064 (Jan. 9, 1956). :

termined by the Rodebush-Dixon method that has
been used for other ZrF, mixtures at ORNL.
This method depends on t11e *‘valve'’ action of
the salt vapor above a liquid held at constant
temperature; the valve action is manifested by a
differential manometer which registers the re-
sistance to flow of inert gas through a region

occupied by the refluxing salt vapor. The

‘*valve' becomes suddenly much more definite

in effect when the inert gas pressure is lowered

to a value corresponding to the salt vapor pressure,
An absolute manometer is used to measure the
pressure at which this occurs,

The System KF-ZrF

The vapor pressure experiments initiated pre-
viously on the system KF-ZrF, were continved. 5
The Rodebush-Dixon method,® prevnously de-
scribed,®*7 was employed.

In the present study the temperature determi-
nations were carried out potentiometrically by
using calibrated platinum—platinum-thodium ther-
mocouples. Pure ZrF, was prepared by subliming
hafnium-free ZrF, at 720°C under a vacuum,
Transiucent crystals were then hand-picked from
the sublimate. Spectrochemical analyses for ten
possible metallic contaminants showed, in every
case, less of the impurity than of the available
standard, Potassium fluoride was purified by
heating the reagenf-grade product to 50°C above

~the melting point, cooling it slowly, and selechng
: clear fragmems of \‘he fused salt,

“In the KF-ZrF composmons studied, the as-

' ?sumphon that the ZrF, -is responsible for the
‘Yapor pressure was }usflfled by chemical ‘analyses

and x-ray “and petrographlc studies of -sublimed

'.;matena? obtalned from the apparatus. after a run

was completed The vapor pressures . obtamed for

~ the various mixtures are. summarized in. Table )
' "2 3 1 The constants A and B are: those expressed '

 

Ss.- Cantor, ANP. Quar. Prog. Rep. Marcb 10 1 956,

L ..tORNL-zom. p 103

T bw. H. Rodenbush and A. L. Dixon, Pbys. Rev. 26,

851 (1925).

R E. Moore and C. J. Barton, ANP Quar. Prog. Rep.
Sept. 10, 1951, ORNL-1154, p 136.

111

 
 

 

ANP PROJECT PROGRESS REPORT

TABLE 2,3,1, SUMMARY OF VAPOR PRESSURES FOR THE SYSTEM KF-ZiF

 

Composition

Heat of Yaporization

 

Experimental T:mpemture (mole %) Constants of Zer, _AHU
Range (°C) ;:—'7 A B (kcal/mole)
750-900 659 34.1 9.754 8,813 | 40.4
750-900 59.6 40.4 9.142 8,321 38.1
775-900 55,0 45.0 8.403 7,778 35.6
825-950 49.9 50,1 7.722 7,363 33.7
875-950 47.3 52.4 7.679 7,467 34,2
9501000 45.3 54.7 . 7.649 7,721 35.3
975-1100 ' 40.4 59.6 7.612 8,465 38.8
1125-1300 34.9 - 65.1 7.387 9,020 41.3

 

 

in the relation
B
logP(mmHg) A -7

where T is in °K, These constants are obtained
by treating the data by the method of least
squares. The heat of vaporization, AHv, of ZrF,
is obtained by multiplying the constant B by
2.303 (°R) or 4,576. The values listed in Table
2.3.1 are considered to be more reliable than
those previously published.® Apparently there
was some free Z¢F  present in some of the earlier
experiments, and for the compositions containing
65.9 to 49,9 mole % ZrF4 a reanalysis by least
squares gave the new constants. A plot of the
data from which the equation for each mixture was
obtained is shown in Fig. 2.3.1.

A very odd phenomenon is observed when the
apparent heat of vaporization, AH_, is plotted
against composition (Fig. 2.3.2). A sharp minimum
occurs at 50 mole % ZrF,, and the heat of vapori-
zation increases linearly with composition in both
directions from the minimum. The change below
50 mole % ZsF, is in the expected direction, since
ZrF, becomes more tightly complexed by fluoride
ions. The change above 50 mole % was not
expected; qualitatively it can be explained by
assuming that ZrF, exhibits positive deviations
from Raoult’s |uw when dissolved in KZrF_.
(This version of Raoult’s law calls for a ‘‘two-
particle” depression at the Z¢F, end of the
composition range and for zero pressure at the
KZrF; end.) In practice, the vapor pressure of
supercooled liquid ZrF, is not sufficiently well

112

known for a numerically reliable Raoult law to
be established for ZrF, systems at 600 to 800°C.

The System RbF-ZF,

The vapor pressures of various mixtures in the
system RbF-ZrF, were determined in the same
manner as in the KF-ZrF, system. Reasonably
pure RbF was obtained by selecting clear crystals
from a slowly cooled melt. Analyses showed that
the other alkali metal ions were present in the
following weight percentages: Li, 0.023; Na,
0.005; K, 0.16; Cs, 0.24.

The dutu for various mixtures are complled in
Table 2.3.2 and shown graphically in Fig. 2.3.3.
The constants A and B are from the relation

B
log P (mmHg) = A -

where T is in °K,

As in the case of the KF-ZrF, system, the heat
of vaporization goes ihrough ct minimum at 50
mole % ZrF (Fig. 2.3.4). The linearity with mole
fraction is also similar to that of the. KF-ZrF,
system, Thus the effect seems to be a real one
in spite of its unnatural appearance.

A comparison of the effects of the various alkah
fluorides is given in Fig, 2.3. 5, in which pressure
is plotted as a function of composition at 912°C,
the melting point of pure ZrF,. Moore® has
studied two compositions for the system LiF.ZrF ,
and Sense et al.* of Battelle Memorial lnstltute
and Moorel =3 have studled the NoF-ZrF system.

 

-8R, E. Moore, ANP Quar. Prog. Rep. ]une 10, 1955,
ORNL-I 896, p 81.
 

 

wl

 

"PERIOD ENDING JUNE 10, 1956

ORNL~LR~DWG 14637

TEMPERATURE (°C)

 

 

 

1200 1100 | 1000 900 800
200
T T T Y,
| %
\-a% v.o
\\Sc’) 07:.“
»
100 \ .\"»*% X

 

 

 

 

 

 

 

 

 

 

 

PRESSURE (mm Hg)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

10 el —
065 = 070 075 . 080

q/rt"m

0.85  o0s0 095 . 1.00
x 103 ' " '

F:g. 2 3. 1 Vupor Pressures of KF-ZrF Mixtures.

| "The value of 905 mm Hg for ZrF at the melflng
gpomt was obtamed from the su1:
- pressure equahon '

: on,3604
iOgP (mm Hg) = ]2 54]7 - -—-—.F-—-—-- .

_ Thrs equuhon represents a least-squares treatmenf ;
- systems was measured by the fiodenbush dixon

of 33 experimental points obtained with the present
apparatus, The results conclusively show that
in solutions of ZrF, and alkali fluorides the vapor

hmahon voper: i - _ _
VAPOR PRESSURES AND PHASE EQU]L'BRIUM :

pressure of ZrF decreases as the ulkoh ion slze
_'mcreases. L e

R DATA |N Tl‘lE FeClzoKC| SYSTEM
| ' C.C. BeuSman '

The vupor—pressure of FeCl in FeCI -KCI.

 

%0n assignment from the Oak Ridge lnsfltute of

Nuclear Studies.

113

 
 

 

 

 

 

ANP PROJECT PROGRESS REPORT

GO
ORNL—LR-—-DWG 14638

 

a6

 

 

a4

 

 

42

s O =1

 

 

40

 

 

38

‘ \\"--.

 

 

36

HEAT OF VAPORIZATION, AH,, CF Zrf, (kcal/mole) .

e
. -_’_/

 

34

 

i

 

 

 

 

 

 

 

 

 

 

 

 

 

32 L

KF

Fig- 203-2. _

20

Vaporization of Z‘H"4

40 .60 80

ZtF, {mole ?'o-)

in KI"’-ZrF4 Mixtures.

ZIrfy

Effect of Composition on Heat of

- method in céncentrcltions from 100 to 45 mole %

FeCl, in KCI. Pressure vs composition curves

" obta med at 850 and 900°C are shown in Fig. 2.3.6;

these are typical of the curves obtained at temper-

‘atures from 700 to 1050°C. The strong devigtion
~ from ideality suggests that the FeCl, is complexed

in the liquid at these temperatures. Since the

experimental evidence was obtained in phase equi-

librium studies of the compound K, FeCl, in the
solid state, it is interesting that the same stoi-
chiometry may apply to the complex in ‘the liquid.
The strong depression of the vapor pressure below
45 mole % precludes further experimental efforts
with the' Rodenbush-Dixon technique. The fran-
spiration method will be used to complete the
experimental curve on the KCl-rich side of the

~ graph.,

The phase diagram for the FeCI -KCl system,
shown in Fig. 2.3.7, was deiermmed by two ex-
perimenta} techniques. First, both convenhonol '
cooling and differential cooling curves were ob-
tained on small melts, and, second, quenched.
samples were prepared by the grodient-quenching
techniques used in this laboratory. - Petrographic
examination of the quenched and slowly cooled
samples revealed two well- defined compounds in
the solid state, K:!FeCI4 and KFeCl Both
compounds appear to have solid-state transfor-
mations, and K,FeCl, melts incongruently.

TABLE 2.3.2, SUMMARY OF VAPOR PRESSURES FOR THE SYSTEM RbF--ZrI‘-’4

 

 

Experimental Temperature ‘ c‘;mp:s“wn Constants Heat of Vuporizufion
| | mole %) of ZrF,, AH
Range (°C) A B 4 v
ZrF, RbF ‘ r(kca 1/mole)
800-850 81.6 18.4 10.778 9,506 435
725-825 74.6 - 25.4 10,359 9,189 42.1
750-825 70.0 30.0 9.808 8,790 403
. 750-875 . 65.1 34.9 9.143 ° - 8,358. 38.3 .
775-875 . 60.0 40.0 8.789 8,078 . 37.0
 775-950 . 548 45.2 8.307 7,902 362 ¢
850~975 '~ 50.4 49.6 7.555 7,310 33.5
900-1050 45.2 54.8 7.512 7,794 35.7
1000-~1200 40.0 60.0 7.290 8,146 - 373
1125-1275 - 35.0 65.0 7.296 8,790 - .- 40.3.

 

114

 
 

 

PERIOD ENDING JUNE 10, 1956

ORNL-LR—DWG 13012A

TEMPERATURE (°C)

750

850

950

1150 1050

1250

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

500

 

o
n

200
100

(bH ww) 3YNSSIYH4

 

20

10

o8 oo

0.60

Yrek) x 10°

Fig. 2.3.3. Yapor Pressures of RbF-ZrF , Mixtures.

115

 

 
 

 

 

ANP PROJECT PROGRESS REPORT

SOMNSEENTTIL

ORNL-LR-DWG {13011A
46

 

 

H
H

 

b
/
/
J
A
/
/
/
\

9

 

H
~n

 

 

 

H
o
/lD/

 

 

2]
m

 

 

 

Od
N
.—4/
@

HEAT OF VAPORIZATION, A H,, OF ZrF, (kcal/mote)

 

W
H

 

 

 

 

 

 

 

 

 

 

 

 

 

 

32
RbF 20 40 60 80 ZrF,

Zrf, {mote %)

Fig. 2.3.4.
Yaporization of ZrF, in RbF-ZeF ,

SURFACE TENSION AND DENSITY
OF NOF-ZI'F4

F. W. Miles

Preliminary values of the density and surface
tension of NaF-ZrF, (53-47 mole %) were obtained
by the maximum-bubtle-pressure method previously
described.'?  The results are summarized in
Table 2.3.3. Difficuities were encountered as a
result of plugging of the nickel tube above the
capillary tip. In every case the concentration of
oxides and oxyfluorides in the plug material
appeared to be quite high, and thus it is thought
that there was air contamination of the helium
stream. The assembly is being examined in an
attempt to locate and eliminate the source of

5

wF. W. Miles, ANP Quar. Prog. Rep. March 10, 1956,

 

116

Effect of Composition on Heat of

BN
ORNL~LR-DWG 14639

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

900 o LiF-2rF, {R.E.MOORE) .
4 NoF—2ZrF, (R.E.MOORE) ‘
800 o NaF~ZrF, (K.A.SENSE, BMI) ‘
4 KF~ZrF, (S.CANTOR) . // 4
700 * RbF—ZrF, (S.CANTOR) 1
E A M
£ 600 -
- 7
& /
2 . 'Vg/h 1%/ '
n 500 \ g 87 )
o &7 / |
& &, J/{/RbF=2rF,
: > /)
" 300 S /1
QSQ/" NaF —ZrF,1 1/ ,
200 - § /
7 LiF~ZsF, /// #KF"Z'&
100 / / / .
7 %4
0 -—w -
MF 10 20 30 40 50 .60 TO B0 90 ZIrfy

ZrFy (mole %)

Fig.2.3.5. Vapor Pressures of MF-ZI'F4 Mixtures
at 912°C. :

 

 

 

 

UNCLASSIFIED
ORNL-LR-DWG 14640
240
4
220 /
/o
200 :
180 J
160 /

 

 

 

120

140 | /. ]
/

 

60 . .’/ /
I

TOTAL PRESSURE (mm Hg)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

40
7
//Qescrc
20 L -
/
o _
K 0 20 30 40 50 60 70 80 80 FeCl,

FeCly (mole% )

Fig. 2.3.6. Vapor ‘Pressures of FeClz-KCl
Mixtures at 850 and 900°C.
 

 

o

TEMPERATURE {°C)

 

PERIOD ENDING JUNE 10, -195‘6

UNCLASSIFIED
ORNL-LR~DWG $46414

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

800
"N —
500 \ /
asec 71
- 7 o
400 385°C ~ I i
#| 358°C
QRO
300 298%C
255°C
200 |
K,FeCl, KFeCl,
400
KCt 10 20 30 40 50 60 70 80 90 _ FeCI2

FeCl, (mole %)

Fig. 2.3.7. Phase Diagram of the FeCl_-KCl System,

TABLE 2.3.3. PRELIMINARY VALUES OF DENSITY AND SURFACE TENSION
OF NaF.ZtF, (53-47 MOLE %) OBTAINED BY MAXIMUM-BUBBLE-PRESSURE METHOD

 

Wall Thickness

Inside-RadiUsf .
. Density

 

. ams o S Temperature Surface Tension
f;f Capul!ary(. _ of Cd.pl lor? ey (a/em3) (dynes/cm)
o lem) (mils) S : o o
0758 4 o s0 3.049 131
10,0513 9 600 3.014 e
00513 9 | 00 - 3,051 . . 125
9

- 0.0513 700 2916 o120

 

117

 
 

 

 

 

ANP PROJECT PROGRESS REPORT

contamination, and changes have been made in
the design of the capillary tips.

The four measurements listed in Table 2.3.3
were obtained on the same batch of material.
Although the melt was repurified, in place, after
each determination, the formation of plugs in-
dicated contamination during measurement. Further-
more, the batch was kept at high temperatures for
several weeks, When the reactor was opened, it
was observed that some ZrF, had sublimed out
of the solution, These factors indicate that the
values given in Table 2.3.3 must be considered
to be only of a very preliminary nature. |t is
encouraging, however, to note that values of the
same order of magnitude were obtained with
capillary tubes of different wall thicknesses,
This behavior seems to indicate contact angles
of less than 90 deg. Consequently, the inside
radius of the capillary controls the bubble size,
and no particular uncertainty exists as to whether
the value of the inside or of the outside radius
should be used in the calculations.

GROUND-GLASS LIGHT DIFFUSER
PYREX TUBE

  
 
 
 
  
  
  
 
  
  
  
 
  
 

z
‘21

PROFILE OF A SESSILE DROP

\

Additional measurements will be made on a
fresh batch of similar composition. When more
reliable values are obtained, the measurements
will be extended to other mixtures of interest.

SURFACE TENSIONS OF MOLTEN SALTS
S. Langer

The sessile-drop technique for measuring surface
tension, briefly discussed in the previous report,!!
was applied to a number of samples of an NaF-ZrF
(53-47 mole %) mixture to establish the importance
of a number of experimental variables. Since it
has been demonstrated that NaF.ZrF, mixtures
of this general composition do not wet C-18
graphite, all the experiments described here were
conducted with this material as the supporting
plaque. '

A schematic drawing of the sessile-drop appa-
ratus is shown in Fig. 2.3.8. The sample on its

 

Vg, Langer, ANP Quar. Prog. Rep. March 10, 1956,
ORNL-2061, p 105.

. UNCLASSIFIED
ORNL-LR-DWG 14642

FURNACE TUBE SUPPORT
SOLENOID-OPERATED SAMPLE-MANIPULATING RODS
NICKEL TUBE

COOLING-WATER JACKET

THERMOCOUPLE WELL

REFERENCE SCALE
FURNACE
SESSILE DROP
SAMPLE HOLDER
CONSTANT-TEMPERATURE 8LOCK

COOLING-WATER
JACKET

- QUARTZ WINDOW

SOLENOQID-QPERATED
WINDOW SHIELD

Fig. 2.3.8, Diagram of Sessile-Drop Apparatus for Measuring Surface Tensions of Molten Salis,

118
j',-_f";surrounds the drop._,

 

 

supporting plaque is loaded into the tube through

the sample-loading flange. - The central portion

of the tube is then baked out under high vacuum

to remove volatile materials which might con-
taminate the surface of the sample.  After the
fumace is adjusted to the operating temperature,
‘the sample is pushed to the center of the constant-
temperature block * with the solencid-operated
sample-manipulating rods. Photographs are taken
of the sample as a function of time to ensure the
establishment of an equilibrium droplet. A photo-
graph of the sessile drop of NaF at 1025°C, taken
during calibration work, is shown in Fig. 2,3.9.

UNCLASSIFIED] -
PHOTO 1822

 

Fig. 03-9o

Sessule Drop of NeF at 1025°C on Gruphlfe P"’q“e':-,ff-.af,-teporfed previously.

H_"i'_j_so’nsfactory theoretical interpretations of the re..

The dlmensmns

‘'made under vacuum,

of fhe sessnle drop (fl‘le;_r:j‘__- sults .

Cor-

PERIOD ENDING JUNE 10, 1956

culated by using the tables of Bashforth and
Adams!2 and the densities obtained by Miles.!3

The initial ‘experiments indicated no change in
surfoce tension when an atmosphere of helium was
admitted to the apparatus in which tests had been
When tests were conducted
in vacuum, however, volatilization of ZrF, caused
large changes in composition of the drop; helium
atmospheres have been adopted as standard for
future tests,

Data . typical of those obtained to date for
NaF-Z¢F, melts whose initial composition was

- 53 mole % NaF and 47 mole % ZrF, are shown in
‘Table 2.3.4. While the data of |11e earlier ex-

periments are viewed with some uncertainty, the

relative agreement of the surface tension values
' obtained from droplets of widely varying sizes
i‘,_,_"--(as evidenced by the spread of sample weights
. ondx/z rot.tos), is gratifying. '

"SPECIFIC CONDUCTANCE AND DENSITY
'OF FUSED LITHIUM, POTASSIUM,
AND cssnum FLUORIDES“

E. R. Van Artsdalen

“Electrical _cond_udance and density have been

" measured as o function of temperature for pure,
~fused lithium, potassium, and cesium fluorides, A
- ‘newly designed bridge, in conjunction with a
S'pei:iei cell
eiloy, was used in the measurements. The results

fabricated from platinum.rhodium

of - thé measurements are expressed by the

- f‘”-e'quafions listed in Tables 2.3.5 and 2.3.6. Similar
data were also. obtained for several rare-earth

=  halides.
Phofogreph of Reference Scale nnd}.,;_,‘r

“These- dofa hove been correlated ‘with those

 

 

12F Bashforth and J. C. Adcms, An Attempt to Test
the Theories of Capdlary Act:on,,Cambridge Univ. -
Press, London, 1883, ; ,

13¢, v, Miles, prwate commumcahon, to S. Longer.

”Deto:ls ‘of this work will be published in separofe: ‘
reports and arhcles fro_rrp the Ch__emlsfl'y Dtv_lsion. T

119

It ‘is now possible to. give

‘in _terms of the structure and 1ransport L

maximum “diameter, 2x, and " the d;stance frem';j__':"‘%?P’°P°m°5 of the hqu:d 5"“5° :

the. line of the maximum dmmefer to the apex, z) "
o .-—_ore measured dlrecfly from the photographtc nega-"_-;_f'_'.
- tive by using o toolmc:ker s microscope.’
: .]_fi?rechons for amsotroplc shrmkage of the fllm are
~obfaired : by .using the reference - scele “which -

' The surface tensuon is cul-'*'_

 
 

 

 

 

ANP PROJECT PROGRESS REPORT

TABLE 2,3.4, SURFACE TENSION MEASUREMENTS ON NaF-ZrF, (53-47 MOLE %)

 

| Tempércture Sample Weight Weight Loss Final Composition* Density  Surface Tension

x/z

 

(°C) (g) (%) {mole % ZeF ) (g/cm’) (dynes/cm)
601 ** 0.870 28.2 36.6 1.71706 3,043 14
603** 0.870 28.2 36.6 1.66592  3.040 141
611 ~ 0.4502 3.1 43.5 1.53637  3.030 141
610 0.4502 3.1 43.5 1.56345  3.031 134
601 0.4502 30 43.5 1.60542  3.043 122
601 0.4502 3.1 43.5 1.57069  3.043 132
601 . 0.4502 3.1 43.5 1.55482  3.043 136
- 601 : . 0.4502 3.1 43.5 1.56994  3.043 131
610 0.5097 3.7 41.4 1.68795  3.031 . - 17
610 0.5097 3.7 41.4 1.65982  3.031 - - 24
602 70,5097 3.7 41.4 1.64464  3.043 o128

 

 

" *Coleulated from analysis of material after test.
**Tested in vacuum; all other samples tested under helium atmosphere.

TABLE 2.3.5, CONSTANTS OF EQUATION FOR SPECIFIC CONDUCTANCE-
K, OF FUSED LITHIUM, POTASSIUM, AND CESIUM FLUORIDES

K=a+(bx 10~ + (c x 10-5)¢2, ohm™licm=1, where ¢ is in °C

 

 

Sals b Standard Deviation Applicable Temperature
a , @ c (ohm=T.em=1) Range (°C)

_ LiF +3.805 +1.004 -3.516 0.008 8471027
KF —3.493 +1.480 —6.608 | 0.009 . 869-1040
CsF - —4,511 +1.642 ~7.632 . 0.009 725921

 

TABLE 2,3.6, CONSTANTS OF EQUATION FOR DENSITY, p, OF FUSED
' LITHIUM, POTASSIUM, AND CESIUM FLUORIDES

p=a—(bx ]0'3)t, g/cma, where t is in °C

 

 

Slt b Standard Deviation, Experimental Temperature
a 7 o a g (g/cms) Range (°C)

LiF 2.2243  0.4902 0.0003 | 876-1047

KF 2.4685 0.6515 0.0003 ' 8811037

CsF . 4.5489 1.2806 0.0004 712-912

 

120
 

 

 

[

 

 

. PERIOD ENDING JUNE 10, 1956

2.4. PRODUCTION OF FUELS

G. J. Nessle

G. M. Watson

L. G. Overholser

EXPERIMENTAL PREPARATION OF
VARIOUS FLUORIDES

B. J. Sturm

Continued use of structural metal fluorides for
research necessitated the preparation of additional
quantities of these materials by the several meth-
ods developed previously. As in the past, use was
made of chemical, x-ray, and petrographic exami-
nations to establish the identity ond purn‘y of the
materials.

Additional CrF; was prepared by hydrofluorina-
tion of anhydrous CrC|3, as well as by the thermal
decomposition of (NH,), CrF ‘The latter is pre-
pared by the mferacnon of hydrated CrF with
NH,F-HF, and it may also be reduced by hydrogen
to yleld CrF,. Several batches of NiF, were pre-
pared by the- hydrofluorination of either hydrated
NiF, or hydrated NiCl,. One batch of AgF was
prepared by hydrofluormahon of Ag,CO,. Methods
described prevmtJSIy were used for the prepara-
tion of additional quantities of CeF, and of LaF,.

An attempt to remove oxide from a KF-ZrF 4 mnx-
ture by hydrofluorination was unsuccessful. Runs
at 600°C failed to remove the oxide completely
and, while this could be accomplished at 850°C,
the loss of ZrF, was sufficient to alter the compo-
sition to the point where the material was unsatis-
factory for its intended purpose. :

A batch of CuF, was prepared for use as a car-

rier in certain spectrographlc boron analyses. Ma-
terial low in boron and silicon was requested for -
this purpose. Anhydrous CuF, was. hydrofluor-
inated for 4 hr at 400°C, but this short period of -
treatment failed to reduce the boron content to a.
sufficiently low level. However, extending the
period of hydrofluorlnaflon to 20 hr yleided o ma-
_terlal thaf was acceptable. '
A fuel mixture containing Li7 and enrached ura-_f .
nium was prepared. for radiation dumage stud|es.‘lgt
Since Li7 was available only as the carbonate, 0

prelimmary run was made with normal Li,CO.,

UF,, KF, NaF; and excess NH,F.HF to esfabhsh‘_”" '

ture containing Li’7 and U233,

oxide free, but a sulfur content of 170 ppm was
reported, most of which came from the Li7C03.
The fuel was subjected to 5 cycles, each of which
entailed a 10- to 15-hr hydrogen treatment and 1 hr
of hydrofiuorination at 800°C. The product then
contained 45 ppm of sulfur and a trace of oxide.
Hydrofluorination at 700°C for 6 hr then produced
an oxide-free product.

LABORATORY-SCALE PURIFICATION
OPERATIONS
F. L. Daley W. T. Ward

With appropriate modzflcahons, the standard hy-

_drofluorination-hydrogenation process was used

to prepare a number of especially pure materials
requested for various purposes. Of these only an
NaF-ZrF,-UF, mixture containing additives of
Ban, LaF3, and RbF had not previously been pre-
pared by the standard procedure. Operational facil-
ities to transfer and dispense small samples (200
to 500 g), especially into apparatus for accurate
evaluation of physical properties of materials, were
constructed and used successfully during this quar-
ter. There is evidence that direct transfer of the
purified melt into the apporatus has significantly
improved the purity of the test material and, ac-
cordingly, increased confidence in the test results.
Copper-lined stainless steel reactors were used
successfully in the purification operations.

© PILOT- SCALE PURIFICATION OPERATIONS

J P. Blakely C. R. Croft
‘ J. Truitt

The pilot-scale purification facility processed 68

'-"-fbatches totaling approximately 1300 Ib of various
- fluoride compositions for use in small-scale corro-
" sion’ testing, phase equilibrium’ studies, and phys-
.,.’,:cal ‘property studies. " The bulk of the material

was produced during February when the facility

“ was operating on a three-shift, 24-hr-day basis. As
a result, the backlog reported last quarter was

,Z-r'.ehmmated and normal operahons were resumed.
whether fuswn of this” mtxture would yield “an . : : :

oxide-free product After the preliminary run proved -
successful the procedure was applied to the mix-
The material was

 

18, J. Sturm, ANP Quar. Prog. Rep. March 10, 1956,
ORNL-2061, p 100.

121

 
 

 

 

 

ANP PROJECT PROGRESS REPORT

The use of copper-lined stainless steel reactors
in the 5-1b units has proved successful. Therefore,
when the nickel reactors now on hand wear out and
are discarded, they will be replaced with copper-

lined stainless steel vessels. The new copper-

lined vessels, in addition to showing promise of
having longer lives, ore more versatile than the
nickel reactors in that they can be used in the
preparation of BeF,-containing mixtures.

Although the supply of 50-Ib nickel reactor ves-
sels is ot present sufficient, these will also be re-
placed with copper-lined stainless steel vessels
as they are discarded.

PRODUCTION-SCALE OPERATIONS
J. P. Blakely

Seventeen batches totaling approximately 4230 Ib
of fluoride compositions were processed by the
production-scale facility. Two new copper-lined
stainless steel reactors were received about the
first of March and were put into production opera-
tion immediately. To date, nine batches have been
processed in one reactor and eight in the other with
no apparent difficulties. Recent external examina-
tion of these reactors show no deterioration or
faults, and it may be that the life service of these
reactors will be much better than expected.

In spite of the temporary shutdown of this facil-
ity, the processed fluorides inventory has remained
high. Operations were slowed down during the last
three weeks of this quarter because of a shortage
of storage containers, A survey is now being made
to determine the anticipated requirements for proc-
essed fluorides during the next six months. A
recent communication from Pratt & Whitney Aircraft

J. E. Eorgan

gave fairly firm commitments for approximately 1700
Ib of processed fluorides per month for the remain-
der of the calendar year. The ORNL Aircraft Re-
actor Engineering Division has been using an av-
erage of about 1000 |b of fluorides per month.
Although the new survey is not complete, indica-
tions are that Aircraft Reactor Engineering Divi-
sion usage will range around 1000 to 1500 Ib per
month for the remainder of 1956. The demands will
not reach the production capacity of the facility,
but ot feast they will make it economically fea-
sible to maintain continuous operation,

The remaining 7500 Ib of hafnium-bearing zirco-

nium fluoride ordered from an outside vendor was

received and all the material was accepted as
satisfactory.

BATCHING AND DISPENS_|NG OPERATIONS
J. P. Blckely F. A. Doss

The batching ond dispensing facility dispensed
148 batches totaling approximately 4900 ib of proc-
essed fluorides in batch sizes ranging from 1 to
250 Ib. The total amount dispensed was about
1000 b less than was dispensed the previous quar-
ter. Therefore there was another gain in stock
inventory in spite of decreased production. A ma-
terial balance for the quarter is given in Table
24.1,

The production and use of special compositions
is decreasing steadily, but periodic increases can
be expected as new materials are developed for
testing and investigations. The use of NaF-ZrF,
(50-50 mole %) and NaF-ZrF ,-UF , (50-46-4 mole %)
or closely related compositions is expected to in-
crease to about 3000 ib per month during the first

TABLE 2.4.1. MATERIAL BALANCE

 

 

 

Material (1b)
NOF-ZI'F[UF4 NaF-ZrF4
(50-46-4 mole %) (50-50 mole %) Special Total
On hand at beginning of quarter 6,235 1,875 591 8,701
Produced during quarter 3,543 198 1,789 5,530
Total 9,778 2,073 2,380 14,231
Dispensed during quarter 3,828 0 1,078 4,906
On hand at end of quarter 5,950 2,073 1,302 9,325

 

122

 
 

 

 

 

half of fiscal year 1957. This estimate includes
Pratt & Whitney requirements. At this rate, full
production -should still be able to maintain a high
stock inventory and to allow for any necessary
shutdown of the production-scale facility. The ef.
ficienf handling, maintenance, and dispensing of
some 200 storage cans has made it possible to fill
all sizes of orders with a minimum of delay to
operating sections of the ANP program, -

PREPARATION OF ZrF‘ FROM Z.Cl
J. E. Eorgan J. P. Blakely

The facility for the conversion of hafnium-free
Z:Cl, to Z¢F, was operated three times during this
quarter. The mechanical operation of the unit is
now satisfactery. Handling techniques need to be
improved to cut down on raw-material and finished-
product losses. Also, the finely divided nature of
the product ZrF has presented a dust-loss problem
of considerable magnitude. Steps are being taken
to eliminate or greatly improve this situation. Large
quantities of ZrF, dust were carried into the proc-
ess gas exit lines and coused major line-plugging
difficulties, A change in piping arrangement im-
proved this situation somewhat in the last run.

Rough material balances of the three operations
indicated a total loss of about 22% of product in
the first two runs and 11% in the last run. Approx-
imately 75% of this loss was accounted for in ma-
terial removed from the exit gas lines. A new dust
filter has been ordered for installation on the unit
which should cut down the dust loss considerably.
Operation of this facility will be postponed until
this new filter arrives or there is an urgent. need
for more hafnium-free ZrF,. .

Analytically, the preduct material was as good or -
'befler than mqterlol previously or presently being
used. ' Four samples were taken from each batch as
it was being discharged from the unit, each sample -

. representing roughly one-fourth of each batch so -

 that uniformity could be tested. - The analytical -
"~ results are given in Table 2.4.2, Although the first"

" batch was not so good as the last two, either in
- ‘_,umformlfy or quallty, |t was consudered to be us-
:able. ‘ ' : o

SPECIAL ssnv:css o
J P Blckely F. A. Doss

Requests for service in filling, 'draining,. and
sompling operations were at a normal ond steady
level. The greatly accelerated pace set in the

PERIOD ENDING JUNE 10, 1956

TABLE 2.4.2. PURITY OF ZrF,
PRODUCED FROM Z(Cl,

 

 

 

Major | .
. - mpuriti
Run Somple Conshtuents 'z m;s
Zr F Cl Ni Cr Fe
1 1 53,6 450 0.29 90 55 715
2 53.7 44,2 0.40 80 55 595
3 54,3 44,3 0.50 55 45 595
4 542 44.6 0.63 55 45 620
2 1 54,5 45.2 029 65 40 375
2 55.1 44,8 0,33 50 40 400
3 54,6 44,2 034 65 35 375
4 54,6 45.2 0,36 S50 35 375
3 1 54.4 45.4 039 85 50 420
2 54,3 44,3 0.45 70 25 420
3 54,5 44,9 0.43 75 30 410

4 54,8 44.6 039 60 30 400
Theoretical 54,5 45.5

 

*The hafnium content of the material was less than 100
ppm, the boron content was less than 1 ppm, the sulfur
content was about 30 ppm, and the carbon content was
300 ppm.

previous quarter was not maintained. Past expe-

‘rience indicates that filling and draining opera-

tions will be requested in cycles, depending upon
the rate at which tests are started or terminated.
Present indications are that another accelerating
cycle will occur in the next quarter.

Approximately 3000 Ib of processed fluorides and

1500 ib of liquid metals were used to chorge test
o equupmenf durmg the quarter in charge sizes rang-
" ing from 50 500 Ib. -Since the liquid metal (NaK)
“inventory had run below 1500 Ib, another 5000 Ib-
“of noneutectic NaK (56% Na-44% K) was ordered
‘and received. It is estimated that this supply
- should last six to eight months unless consumption
~is sharply increased. - ' -

~A serious comphcohon arose in’ the d|5posul of
NaK during the cold months of this winter. This

‘material, which is normally molten at room temper-

atures, could not be disposed of by underwater
~ “injection because it froze in the transfer lines. It -
-~ was fortunate that sufficient storage cans were:

available to contain the NaK drained from equip-
ment until warmer weather arrived and that a com-
paratively small amount of NoK was drained from

123

 
 

 

 

ANP PROJECT PROGRESS REPORT

equipment during this quarter. It has been re-

quested that electric power be supplied at the
disposal quarry so that disposal of both sodium
and NaK can proceed regardless of atmospheric
temperature. Present rough estimates of the cost
of such an installation range from $250 to $1000.
Firmer figures are fo be obtained, and, if the range
above is correct, adequate power is to be instalied
for liquid metal disposal,

In-pile loop No. 5 was filled with enriched NaF-

124

ZrF -UF 4(53.5-40-'6.5 mole %). Two new batches
for in-pile loops were processed. Dates for filling
loops with these batches have not yet been set.

Two other small batches of special enriched mix-
tures were also processed this quarter. One batch,
NaF-ZrF ,~UF, (62.5-12,5-25.0 mole %), was to be
used at ORNL for special radiation tests. The
other batch, NaF-UF, (66.7-33.3 mole %), was or-
dered by Battelle Memorial Institute. '

 
 

 

PERIOD ENDING JUNE 10, 1956

2.5. COMPATIBILITY OF MATERIALS AT HIGH TEMPERATURES

F. Kertesz

PENETRATION OF GRAPHITE
BY MOLTEN FLUORIDES

H. J. Buttram G. F. Schenck’

It appears very likely that graphite is the only
reasonably effective reactor moderator material that
will remain chemically stable upon direct exposure
to molten fluoride fuel mixtures. The use of graph-
ite to contain such fuel mixtures has been known
for a long time to be practical for simple laboratory-
scale experimental equipment. The graphite most
commonly used for this purpose has the commercial
designation C-18, However, little information is
available concerning the rate of penetration of
reactor-grade graphite by molten fluoride fuel mix-

" tures, and accordingly an investigation of this phe-

nomenon has been initiated. ,
Specimens of APC2 graphite 1/ X l/ X !’2 in. were

socked at 600°C for various perlods in sealed cap-

sules of Inconel containing molten NaF-ZrF or

NaF-ZrF4-UF4 mixtures under helium afmospheres. |

The specimens were then sechoned and examined
under the petrographic microscope. Preliminary
observations indicate that the APC graphite was
completely penetrated by NaF-ZrF (53-47 mole %)
in 1 hr at 600°C but that NaF-ZrF -UF (53.5-40.0-
6.5 mole %) had not detectably penetrated the
graphite in 10 hr at 600°C. The rapid penetration

of the graphite by the NaF-ZrF, melt is conSIdered.

to be surprising, smce such penetraflon is in-con-
trast to the stability of C-18 graphite in this and

similar fluoride mixtures. Other graphites of higher .

density will be examined in these and other fluoride

‘mixtures. If necessary, attempts will be made to -
/ prevent penetration of the graphite by prior lmpreg-
nation with hlgh-melhng-pomf mafenals such as

“NaF or CaF

EFFECT or ATMOSPHERE ON ANOD!C -
' mssowflon OF NICKEL IN
MOLTEN NaOH

F A Knox

An uf’rempt has been made to study the efflc;ency "

of -anodic. dissolution: of nickel in NaOH under

- various. atmOSpheres and various condmons of cur-

 

lOrs assignment from Pratt & Whitney Aircraft,
Nanona? Carbon Co. code designation,

rent density and temperature. The apparatus con-
sisted of ‘a crucible of Morganite (recrystallized
alumina) containing molten sodium hydroxide into
which a nickel anode and a cathode, as well as o
nickel thermocouple well, were inserted. This
electrolytic cell and its furnace were placed in a
large chamber which could be evacuated or filled
with the desired atmosphere. A lead storage bat-
tery served as the power suppiy, the circuit also
contained a voltmeter, ammeter, and ampere-hour
meter, as well as rheostats to control the current
density applied.

- When the assembled electrolytic cell was main-
tained at the desired temperature without an applied
voltage, attack on the electrodes was negligible.
Passage of current resulted in erosion of the anode
and subsequent deposition of crystalline nickel on
the cathode. The weight gain of the cathode did
not correspond precisely to the weight loss of the
anode, since the crystalline deposit was not per-
fectly adherent; in addition some sodium alumi-
nate, formed by the reaction of NaOH with the

Al 0, of the crucible, apparently contaminated the

cathode deposit. Loss in weight of the anode was
accordingly used to evaluate the efficiency of the
cell, ,

The data shown in Table 2.5.1 were obtained at
a current density of 2 amp/cm?. At this current

density the cell. effmency is very poor under all

conditions. The dissolution efficiencyis decreased
slightly by substituting hydrogen for helium as the

- static cover gas and is reduced considerably by

bubblmg hydrogen i'hrough the ‘molten electrolyfe. :

~ TABLE 2.5.1.- EFFICIENCY OF ANODIC
- 'SOLUTIONS OF NI_CI_(EL ll_fil NaOH

Current denséi.ty: 2 _,a_nfipr/cr'nz' o -

 

" Cell Efficiency (%)

 

. Téir;perutdre : - Under . -With Hydrogen . =

ey . U"d" .- Static . - Bubble&rT_hrdfigh o

He_llyupj 7_.Hydr'ogen_ - the NaOH

 

350 0,01 0,006
600 0.5  0.’42-'

800 2,77 2,12 0,66

 

125

 
 

 

 

 

ANP PROJECT PROGRESS REPORT

2.6. ANALYTICAL CHEMISTRY
J. C. White

DETERMINATION OF BARIUM, LANTHANUM,
AND RUBIDIUM IN FLUORIDE FUELS

A. S. Meyer, Jr. B. L. McDowell

Methods were developed for the determination of
barium, lanthanum, and rubidium in NaF-ZrF -UF
as part of a program for determining the effect of
typical fission products on the physical properties
of molten mixtures of fluoride salts. For the barium
determination the fluoride salt mixture is dissolved
in fuming sulfuric acid, and the barium remains
in the insoluble residue, as the sulfate. After
dilution of the sulfuric acid solution with water,
the barium sulfate, which is contaminated with
uranium and zirconium, is separated by filtration
and purified by dissolving it in a hot ammoniacal

‘solution of ethylenediaminetetraacetic acid. It is

then reprecipitated by acidification of the solution
with hydrochloric acid. The reprecipitated barium
sulfate is filtered off. The filter paper is charred,
and the barium sulfate is then ignited to constant
weight at 600°C.

Lanthanum is separated from the filtrate from
the barium determination by precipitation, as the
oxalate, from a neutral solution of ammonium oxa-
late, Coprecipitated zirconium and vranium are
separated from the original oxalate precipitate by
dissolving the precipitate in concentrated nitric
acid and carrying out a second oxalate precipita-
tion from a neutral solution of ammonium oxalate.
The purified oxalate precipitate is converted to
the nitrate by digestion with concentrated nitric
acid and evaporation to dryness. The residue is
taken up in a small amount of water or very dilute
nitric acid. Lanthanum is finally precipitated by
the addition of solid oxalic acid to the hot, slightly
acid solution, The lanthanum oxalate is ignited
at 900°C and weighed as La,0,.

Ammonium salts in the filtrate from the lanthanum
determination are destroyed by digesting the solu-
tion with aqua regia. The atkali metals are sepa-
rated and converted to the hydroxides by passing
the solution through an anion exchange column
in the hydroxide form. After all traces of ammonia
are removed by evaporating the alkaline solution
to dryness, the rubidium is precipitated and weighed
as the sparingly soluble tetraphenylboron salt.
Potassium, which is present as a trace contaminant

126

in the base fuel, is also precipitated as the tetra-
phenylboron salt., The potassium in the tetra-
phenylboron precipitate is determined by flame
photometry, and the weight of the precipitate is
corrected accordingly. ‘

Good precision has been obtained in analyses
of replicate samples of fluoride mixtures ‘which
contained 1 to 3% of each of the above-mentioned
elements. These methods are also being tested
for the determination of cerium and the proximate
determination of the cerium group of the rare-earth
elements in fluoride fuels,

DETERMINATION OF NIOBIUM IN FUSED
MIXTURES OF FLUORIDE SALTS BY
THE THIOCYANATE METHOD

A. S. Meyer, Jr. B. L. McDowell
R. F. Apple

A method for the spectrophotometric determina-
tion of niobium as the thiocyanate complex, re-
ported by Ward and Marranzino,! was applied to
the determination of niobium in NaF-ZrF,-UF .

The method is based upon the reaction of niobium(V)

with thiocyanate in a mixed solution of 4 M hydro-
chloric acid and 0.5 M tartaric acid to produce a
complex which exhibits an absorption maximum
at 387 mp in an ethyl ether—acetone medium. The
addition of acetone to the extract of ethyl ether
containing the complex inhibits the polymerization
of the thiocyanate ion and stabilizes the soluhon
for at least 20 hr.

Tartaric acid eliminates the interference of ura-
nium, which also forms a complex with thiocyanate.
The interference of the red color of the iron(lli)-
thiocyanate complex is eliminated by shaking the

ether extract with a solution of stannous chloride

for 30 sec to reduce the iron(i!l) to iron(ll).

The concentration of the reagents in the aqueous
phase must be held within rigid limits if repro-
ducible absorbance values are to be obtained.
In order to maintain fixed concentrations, the

fluoride sample is dissolved by fusing it with -

potassium pyrosulfate and dissolving the melt in
1 M tartaric acid. A 1- to 10-ml aliquot of the
resulting solution is combined with a sufficient

 

'E. N. Ward and A. P. Murrunzmo, Anal Chem. 27,
1325 (1955).
 

o

 

volume of 1 M tartaric acid solution to yield a
total volume of 10 ml.. Then 15 ml-of the hydro-
chloric acid—tartaric acid reagent {9 M HCl and
0.5 M tartaric acid) and 15 ml of 20% {w/v) am-
monium thiocyanate solution is added, following
which the aqueous phase is extracted wnh 5 mi

of ethy! ether.. , |
Tests of this method have revea!ed thct a Imearr

relationship exists between the abscrbance and
the concentration of the niobium in the range 0.2
to 2.0 pg/ml. The coefficient of vanohon for this
me’rhod bused on standards, is 2%. o '

DETERMINATION OF TRACES OF COPPER
IN FLUORIDE FUEL MIXTURES

A, S. Meyer, Jro B. L. McDowe”

' Since copper is cons:dered to be compatlble wnth
molten fluoride salts containing sulfur’ impurities,
the nickel reactors used for the production of
fluoride fuel mixtures have been:equppped,wuh

‘copper liners to prolong their useful lifetimes.

This modification in production technique neces-
sitated the development of a sensitive method for

the determination of copper in mixturefi'of fluoride

sclts. The spectrophotometric” method .in which

““cuproine’’ (2,2 *.biquinoline) is used as_the chro-
mogenic reagent, reported by Gue512 to be Specrflc--

for copper, was adapted for this determination, - -

Copper is determined as the copper(l)-cupronne"-

corln[:Iex fo;mtehd bl{::u;:b:utnltng an uqueous:(:)ui‘fvatt; - ‘can be separated quanhtahvely from uranium and
soiution ot The ple containing coppe! ‘"M zirconium by precipitation with tannin in a slightly

a solution of 002% (w/v) cuproine in n-omyl

alcohol at pH 6. The copper complex is extracted
in the orgdnic phase.’ The complex exhibits its
~ maximum - absorbance at 550 m.® Hydroxylammef.i_j*
_' fhydrOCh'or'de was, used 1o :educe the copper. to-';—f'{ffi-??by Volatlhzanon. ‘This was accompllshed by o '/
. the monovalent state, and ‘l'crtarlc ucud was used o
- '_':to complex iron cmd uranium, o L
. The extraction procedure recommended by Guestz,] .
, jf—_i_-was mod:fled for. upphcoflon to. the - defermmahorx'
" of traces of - ‘copper in solutaons of hugh ionic.
"'fir'.qistrength “The volume of the aqueous phase was-
o Vmcrecsed from 30 to 100 ml*in order to ensure,::i.-f Clplfdfed wath tannm ct a pH of Sina hor solutlon.
~.an-.optimum: umount ‘of copper in_the uhquot‘: The "«WTh cioitat fr ite 1 ! and f +
- .. extraction’ penod Was increased from 1. 1030 min, e pre 'p' ° e, after | s/ emotfc op igni 4on 7 °
-~ "and- 10 ml of‘an"alcoholic: solut:on of: the reagent;ff'-;.__‘r"3'" e L T
3 J. C. White, Determmauon oj Small Amounts of Tan-

- 'was’ used. -Since fhe uqueous-orgamc ratio WAS e S P PR and in NaF-LzF-KF-UF4, ORNL.

so iarge, 10 g “of ‘ammonium ‘sulfate was added'_'sz- {"CF -56-1-49 {Jan, 10, 1956). -

; "‘1o reduce the solublllty ‘of fhe olcoho| in fl-:e o

 

2R, ). Guest, Anal. Chem. 25, 1484 (1953).

 

 
 

PERIOD ENDING JUNE 10, 1956

aqueous- volume. The ammonium sulfate served

to mask the mass-action effect of ionic constituents
in the sample; however, it reduced the effective
pH range for quantitative extroction from a range

of 4,5 to 7.5 to a range of 6.0 to 6.5.

"~ “The abscrbance of the organic extract is a linear

function of the concentration of copper up to 7.5
pg/ml, The coefficient of variation of the methed,

~as derived from stondard samples, is 2%. Samples

of NoF-ZrF -UF‘ prepared in the copper-lined
nickel reactors have been analyzed for copper by

‘this method. The range of the copper concentration
‘was | to 25 ppm,

DETERMINATION OF TANTALUM
IN NaF.ZrF -UF

J P. Young J. R. French

The determination® of trace amounts of tantalum

in'NaF-LiF-KF has been extended to include the

_ determination of tantalum in NaF-ZrF -UF Since

z:rcomum dnd uranium interfere w;fh the determ:-

. nation of tantalum by the pyrogallol method, 4 it
was necessary o develop some means of removing

these interferences. The separation of tantalum

,_from uranium in’ NcF LiF-KF-UF,, which was
described prev:ously, was accomplished by «
;'fprec:pltahon of the tantalum with cupferron. Zir-
~conium, however, is also precipitated w:fh cup-
,ferron. According to Hillebrand et al.,® tantalum

acidic soluhon of oxalate which is half-saturated

‘with ammonium chioride, Since the samples con-

tained’ alkall fluorldes, it-was necessary to dis-

“solve them ccrefully to prevent ‘the loss of Tqu

_:_‘5_‘4'_-cureful dlgeshon of ‘the fluoride sumple in dilute
7 "sulfuric acid.in order to hydroiyze TaFg to Tc:zo_.-,.w
) 5‘The soluhon was then evoporc?ed to dryness, and

the residue ‘from this’ ‘evaporation was fused: with

f_"i_potcsstum pyrosulfcte. ~The product - of the fusuoni
- was ‘dissolved in ammonium-oxalate and then pre-

 

45 |. Dinnin, Anal. Chem. 25, 1803 (1953).

SW. F. Hillebrand et al., El:ed Inorganic Analyszs. :
p 598, 2d ed., Wiley, New York, 1953, ]

127

 
 

 

 

 

 

ANP PROJECT PROGRESS REPORT

Ta,04, was fused with potassium pyrosulfate, and
the tantalum-pyrogallol color was developed, as
described in the earlier report.3

The coefficient of variation was 5% for the de-
termination of tantalum in NaF-ZrF ,-UF,. In the
separation of tantalum from vranium and zirconium,
a sample which contained at least 0,5 mg of tanta-
lum was taken.

DETECTION OF TRACES OF NaK IN AIR
A. S. Meyer, Jr, J. P. Young

The design of an instrument for the photometric

detection of NaK in air was completed. A func-
tional description and the operational requn‘ements
of the apparatus were given previously.® The
electronic components of the instrument have been
assembled, and the optical system is now being
fabricated.
_ Preliminary engineering drawings of the instru-
ment for the detection of submicrogram quantities
of NaK in air by observation of the sodium reso-
nance radiation have been prepared. Experiments
are being carried out to devise a method for the
introduction of reproducible concentrations of alkali
metal oxides into air streams in order to provide
synthetic samples for testing these instruments,
The alkali metal is injected into the air stream
in a jet of helium which is saturated with alkali
metal by contact with the molten alkali metal.
The temperature dependence of the vapor pressure
of the molten metals provides the method for the
control of the concentration of the metal in the
helium. An apparatus has been assembled to test
the stability of sodium aerosols that have been
prepared at various mixing temperatures and various
flow ratios of helium to air and to determine a
quantitative method for transferring the samples
of air to the detection instruments.

VDET_ERMINATVION OF OXYGEN IN NaK
A. S. Meyer, Jr. - G. Goldberg

Since it has been established that reasonable
precision can be obtained in making the determi-
nation of oxygen in alkali metals with the modified
Argonne distillation sampler, the sampler was
connected to a loop in which NaK was being cir-
culated. The loop operates at a temperature of

 

6a. s. Meyer et al.,, ANP Quar, Prog. Rep. March 10,
1956, ORNL-2061, p 207.

128

1200°F, with a cold trap differential of cbout
600°F. Since no inherent difficulties were en-
countered in the operation of the sampler in the
loop, a detailed, step-wise procedure’ was written
which covers the use of the sampler with both
sodium and NaK loops. The analyses made over
a five-day period while the cold trap was in oper-
ation gave the following concentrations of oxygen,

in ppm: 1330, 1265, 575, 310, 265.

Since the results obtained with the sampler
appear to be satisfactory, additional tests will
be conducted in which the results from a plugging
indicator will be compared with the results from
the sampler. Plans are also being formulated to
add a known amount of oxygen to the NaK to make
further comparisons of the sampler and the plugging
indicator. {n the sampler now being constructed a
sight tube is being placed in the distillation cham-
ber so that the operations of sampling and distilla-
tion may be observed., :

EXAMINATION OF COLD TRAPS FROM
ALKALI-METAL SYSTEMS

A. S. Meyer, Jr. G. Goldberg

A program was initiated to carry out the chemical
examination of cold traps removed from systems
circulating alkali metals, such as heat-exchanger
test stands and corrosion-testing loops containing
sodium or NaK. A typical request for analysis
calls for the determination of oxygen, iron, chro-
mium, nickel, and, in the case of NaK, a sodium-
potassium ratio. Other constituents which are
sometimes requested are uranium, zirconium, and
beryllium. The cold traps range from 3 to 5 in,
in diameter and from 2 to 6 ft in length.

In order to carry out the determination of oxygen
successfully, the trap is drained as completely
as possible. The residual sodium, or NaK; on the
Demister packing within the trap is then reacted
with butyl bromide to form the neutral bromide
salts of sodium ond potassium; oxides -of these
elements do not react with butyl bromide. When
the reaction is completed, the organic material
is drained, and the trap is cut into two parts,
Until the trap is cut open, o blanket of helivm is
maintained in it. '

 

7.}. C. White, Procedure for the Determination of Oxy-

gen in Sodium and NaK by tbe Dzst:l[at:on Metbod'

ORNL CF-56-4-31 {April 5, 1956).
 

 

 

After the trap has been opened, both sections
are washed carefully to dissolve the bromides and
oxides, and the washings are combined and filtered.
A portion of the filtrate is titrated with hydro-
chloric acid, and the concentration of oxygen is
calculated from this titration. Both the filtrate
and the residue are also analyzed for the corrosion
products and any other constituents that were
called for in the request for the analysis. The

results of a typical analysis are presented in
Table 2.6.1.

TABLE 2.6.1. ANALYSIS OF NaK DRAINED FROM
THE CIRCULATING COLD TRAP OPERATED IN
INTERMEDIATE HEAT-EXCHANGER TEST STAND B

 

 

Determination Water-Insoluble | Total
Requested Residue (g} (9)
Na 564
K 200
Cr 0.6 0.8
Ni 0.9 1.2
Fe 0.1 0.1
U Not detected
Zr Not detected
0, 120

 

DETERMINATION OF WATER IN HELIUM
A. S. Meyer, Jr. G. Goldberg -

Traces of moisture in the helium used as a
blanket gas in engineering studies can be con-"
veniently determined by measuring the dew point -
of the gas. However, h|gh and ‘erratic dew-pomt :
temperatures ore occasionally measured. < It was =~
" observed. that when -copper tubmg was used for the .
~ connections from the helium to the dew-point meter,

dew-pomt temperafures below -145°F correspond-
-ing to about 1 ppm-H O ‘were obtumed almost.
~immediately, - Conversely, periods of flushing of -
as long as .15 min at a rate of about 10 f3/hr
were - required - before comporuble readlngs were -
obtained ‘when the gases were passed through 8
2.t length of Tygon tubing. It is therefore recom-r_r o
mended  that, when possible, connections to the -

PERIOD ENDING JUNE 10, 1956

meter be made with metal rather than plastic tubing
when dew points below -40°F are to be measured.
If plastic tubing is used, the tubing should be
adequately purged with test gas before measure-
ments are taken,

COMPATIBILITY OF BERYLLIUM WITH SOME
TYPICAL ORGANIC DEGREASING AGENTS

A. S. Meyer, Jr. W. J. Ross

Tests have been carried out to determine the
compatibility of beryllium with the organic solvents
being considered as possible agents fordegreasing
the machined reactor components. Suggested de-
greasing procedures include 30-min flushing periods
with acetone and ethanol at room temperature and
trichloroethylene and perchloroethylene at their
boiling points. No detectable attack on beryllium
was observed even after a 24-hr period of dynamic
contact with these reagents.

ANP SERVICE LABORATORY
W. F. Vaughan

Analyses of ten samples were performed for the
Wright Air Develooment Center (WADC). The de-
terminations made on the WADC fused-fluoride-salt
samples included total uranium, trivalent uranium,
iron, nickel, and chromium. The hydrogen content
of zirconium hydride liners, which had been en-
cased in Inconel tubing, was also determined.

The butk of the work of the service laboratory
was the analysis of fused-fluoride-salt mixtures
and alkali metals for the Reactor Chemistry ond
Experimental Engineering Groups. A total of 1164

samples was analyzed, which involved 4168 re-
~ported results, an average of 3.6 per sample. The
- backleg consists of 48 samples. ‘A breakdown of
,; ;_fhe work follows:

Number of Number of
- . Samples  Reported Results

Reacfor Chemisfry, 734 2601

‘Expenmenfcl L 337 o 1375
VEr_lgmeerlng_' _ S
WADC e 50
Miscellaneous 83 142
Tetal 1164 4168

129

 
 

 

 

 

 
 

 

 

Part 3

METALLURGY

W. D, Manly

 
 

 

 

 
 

 

 

 

 

3.1. DYNAMIC CORROSION STUDIES
J. H. DeVan

FLUORIDE FUEL MIXTURES IN INCONEL
FORCED-CIRCULATION LOOPS

J. H. DeVun - R, S Crouse

" An lnconel forced-curculahon Ioop, 7425-8, in
which the cold-leg surface area was: !arger than
that in ‘the standard loop, was operated to study
the effect of the cold-zone area on mass transfer
in the fuel mixture {No. 30) NaF-ZrF -UF (50-46-4
mole %). A tube bundle with a fop and a bottom
header and five connecting co;ls of V-in.-dia
tubing was used in place of a- standard-s;zed
cooling coil " of 4-rn.-d|a tubing.,! ~ With this
arrangement the cold-zene area was twice that of
o standard loop. The volume of the cold zone
was held as nearly as possible the same as that
for a standard loop so that the total loop volume
would be unchanged. The standord loop (7425-10)
operated concurrently with this loop was designed

 

'G, M. Adamson and Re S¢ Crouse, ANP Q.uar. Prog.
Rep. June 10, 1955, ORNL-1896, p 86, Fig-'5o2-

as a control loop for this experiment and for

'several previously run experiments in which

variables such as wall temperature, temperature
gradient, and surface-area-to-volume ratio were
studied. Both the loops, 7425-8 and 7425-10,
were to operate at a maximum wall temperature
of 1600°F and with the other conditions given in
Table 3.1.1. However, it was found after operation
that both loops had identical discrepancies in the
recorders used to measure maximum wall temper-
ature, ond thus they actually operated at a maxi-
mum  wall temperature of 1580°F. While good
comparisons can be made between these two
tests, the lower wall temperature’ somewhat

invalidates the use of loop 7425-10 as a stundord

loop for comparison with other test loops.

The maximum attack in the lfoop with the larger
cold-zone area, 7425-8, was to a depth of 4 mils,
which compares quite closely with the attack
to a depth of 4.5mils in the standard loop, 7425-10.
There were no cold-leg deposits in either loop,
and fuel analyses made after the tests showed

 TABLE 3.1.1. OPERATING CONDITIONS FOR FOUR 'INC_ONEL FORCED-CIRCULATION LOOPS THAT
CIRCULATED THE FUEL MIXTURE (No. 30) NuF-ZrF‘-UF‘ (50-46-4 mole %)

 

Loop Number

 

: Operuti_n'g' Cofid itions - '
' Vo : 7425-8

7425-10 7425-41 7425-43

 

Operating time, hr 1000 1000 1000 1000
Maximum fluoride mixture =~ 1500 1500 _ 1650 1500
tempera'ture,', °F' o - e R
Flmd temperature drop, oF LT 200 200 : | " 200 - 2‘00- -
| "‘,Max-mum tube wamempemm, Lm0 .?isso -'f; S 1700
| 'l":?iflevm'ds number L ilo.ooo'f - w,ooo 2950 6,500
L Fludvelecty,fos- o es . es 208 4a1
g Heared surfuce area, m.2 e "_,"262-:‘- o 1‘262'7 i 262 . 262
~ Cooled surface area, im"’ o sa2 241 a1 U7
Raho of cooled surfoce areo to S o 4a2 71: Sl 2.0 C _A ) ) _: 20 -, s 2_0 " ’ )
L total. |oop volume o RS BRI L e S LT
-"/"__\'_V"Raho of heoted surfo::e oreq tol o ].94 24 : 21 _ 2.] o
. tofol loop volume L STy RTENEI S , L
Maximum depth of aftuck mrls . 4 | 4.5 10 9

 

133

 
 

 

 

 

 

   

ANP PROJECT PROGRESS REPORT

equivalent chromium contents, approximately
400 ppm. Thus the increase in the cold-zone
surface crea produced po apparent effect on
corrosion under the conditions of these tests.
Some increase in the emount of mass transfer,
as evidenced by increased hot-leg attack, was

‘expected with increased cold-leg surface area on -

the basis of thermodynamic studies of the reactions
of fluoride.salts with pure chromium and with the
chromium' in Inconel (see Chap. 2.2, **Chemical
Reactions in Molten Salts’'). These studies
predict that mass transfer of Inconel in ZrF .

‘bearing fluoride mixtures should be rate-controlled,

in part, by the amount of chromium which can
diffuse from the circulated fluid into the cold leg.

Such a process would be related directly to the

cold-zone surface area, |

Examination was completed of loop 7425-43,
which was operated as a part of a series of tests
to study the effect of the bulk fluoride temperature
on the corrosion of Inconel. The loop circulated
the fuel mixture (No. 30) NaF-ZrF 4 UF, (50-46-4
mole %), and it operated with a mammum wall
temperature of 1700°F, a maximum fluid temperature
of 1500°F, and a fluid temperature drop of 200°F.
The results of operation of this loop are to be
compared with those for loop 7425-41, which, as
reported previously,? was operated with a similar
maximum wall temperature and temperature drop,
but with a maximum fluid temperature of 1650°F.
Other conditions of operation for these two loops
are compared in Table 3.1.1.

The maximum attack was found in the region of
maximum wall temperature in both loops. As
shown in Figs. 3.1.1 and 3.1.2 the types of attack
were quite similar, as were the depths of attack,
being 9 mils in loop 7425-43 in which the maximum
fluid temperature was 1500°F and 10 mils in
loop 7425-41 in which the maximum fluid temper-
ature was 1650°F. The amounts of chromium in
solution in the fluoride mixtures in both loops,

‘as shown in Table 3.1.2, were also comparable,

with the amount in the fluid circulated at 1500°F
actually being greater. If variations in Reynolds
number are neglected, as argued previously,?
the insignificance of the effect of the differences
in bulk fluoride mixture temperature in these two
tests is apparently the result of the similarity in
maximum wall temperatures {(1700°F in both loops).

 

2y, H. DeVan, ANP Quar. Prog. Rep. Dec. 10, 1955,
ORNL-2012, p 105.

The importance of the wall temperature to.cor-.
rosion in Inconel-fluoride fuel systems was
discussed previously.?

 

 

 

rFig. 3.1.1. Region of Maximum Attack in Inconel
Forced-Circulation Loop 7425-43 Which Circulated
the Fuel Mixture (No. 30) NaF-ZrF ,-UF, (50-46-4

mole %) for 1000 hr with a Maximum Wall Tempera-

ture of 1700°F and a Maximum Fluid Temperature
of 1500°F. Etched with modified aqua regia.
250X, Reduced 32.5%. {Secret—with—caption)

 

Fig.-3.'|.2. Region of Maximum Attack in Inconel

Forced-Circulation Loop 7425-41 Which Circulated.
the Fuel Mixture (No. 30) NaF-Z¢F 4-UI"' 4 (50-46-4

- mole %) for 1000 hr with @ Maximum Wall Tempera-
" ture of 1700°F and a Maximum Fluid Temperature

of 1650°F. Etched with modified aqua regm.
250X Reduced 32.5%. (Seeretwittrroption)
 

 

 

 

PERIOD ENDING JUNE 10, 1956

TABLE 3.1.2. URANIUM AND IMPURITY ANALYSES OF FUEL MIXTURE (No. 30) NaF-ZrF -UF,
(50-46-4 mole %) AFTER CIRCULATION IN INCONEL LOOPS 7425-41 AND 7425-43

 

Loop No. = -

Uranium Content

Impurities Found (ppm)

 

 

Sample-Token o
: (wt %) ‘Ni Cr : . Fe
| 742541 Before filling 8.61 0 105 45
- After draining 10.3* 40 440 70
7425-43 Before filling 9,20 50 55 70
After draining 9.37 100 . 650 80

‘*This value is in doubt,

~ SODIUM AND NaK IN INCONEL
FORCED-CIRCULATION LOOPS

J. H. DeVan R. S, Crouse

‘An Inconel forced-circulation loop completed
1000 hr of operation with sodium which had been
specially treated to removed oxide contamination
prior to testing. During this test the maximum

fluid temperature was 1500°F, and the temper-
ature drop was 300°F. Prior to the actual test,
‘sodium was circulated through the loop at’ 1500°F

under isothermal conditions to remove oxide films
from both the hot and cold legs of the loop. This
sodium was then dumped at 1500°F and replaced

with a charge that had been cold trapped -and -

filtered in a service loop operated at 300°F. A
circulatory cold trap maintained at 300°F was also
used during the test,

trap temperature,

An evaluation of mass.
transfer in this loop showed that the weight of -
deposit was equivalent to that found in loops
- operated with normal sodium either wflh or wufhoutf":
tnrculafory cold traps. -
~-An experiment .to evaluate 1he effect of a lower'
' cold trap . operating temperature, in -which NaK. -

" rather “than sodium- was used, likewise failed
_ to show a reduction in mass transfer, Loop 74393,
- also constructed ‘of Inconel, was operuted 1000 hr

= with the cold trap temperafure held to the minimum
" for which Flow through. the trap could be maintained, -
o npproxmately 100°F. ~The operahng “conditions,

- except for.cold trup temperofure, were identical

. to those for the ubove—rnenhoned loops. A visual

* * . examination of the loop failed to show. a reduction
i deposifs in_comparison ‘with the deposnfs found
~in loop 7439-1, described previously,3 which also
circulated NaK and operated with a 300°F cold-

SODIUM-BERYLLIUM-INCONEL
COMPATIBILITY IN DYNAMIC SYSTEMS

J. .H. DeVan R. S, Crouse

Previous compatibility studies of Inconel-
beryllium-sodium systems, conducted by inserting
small beryllium tubes in the hot legs of toroid,
thermal-convection, or forced-circulation loops,
showed little effect on temperature-gradient mass
transfer at temperatures up to 1300°F that could
be attributed to the presence of the beryllium. To
determine whether increasing the beryllium surface
area relative to the surface area of the Inconel
would affect the compatibility, two forced-circu-

~lation loops, in which equivalent surface areas of

beryllium and Inconel were in contact with the
flowing sodium, were operated at 1250°F. Only
_one of the loops had an oxide cold trap.

The beryllium, in the form of rectangular blocks
with - drilled - holes, was canned in Inconel and

- inserted,_inrthe hottest section of the loop; the
_remaining sections of the loop were constructed of

Inconel, ~ A temperature drop of 300°F was

maintained between the hot. ‘and cold fluid temper- '
atures in each. test. . After 1000 hr of operation,

‘no increase in the amount of mass transfer was

seen in either of these loops as compared with the
‘mass transfer found in similar loops with smaller

beryllium inserts, ~ The mass-transferred deposits

found were not sufficient in quantity to permit

“chemical analysus. They were similar in amount

and appearance to deposits seen in Incone! loops

‘without beryllium inserts.  The addition of the

cold trap had little effect on the test results,

3J0 He DeVan, E. A. Kovacevich, and R« S. Crouse,

ANP_ Quar. Prog. Rep, March 10, 1956, ORNL-ZOG'I, '
p 117,

 

135

 
 

 

 

 

ANP PROJECT PROGRESS REPORT

THERMAL-CONVECTION LOOP TESTS OF
INCONEL CASTINGS

- 'J.V H.VDeVon , E. A. Kovacevich

In order to evaluate the use of Inconel castings

for large, intricate sections required in the ART,

three standard Inconel thermal-convection loops
with cast Inconel inserts in the hot legs were
operated for 500 hr. The castings, whose compo-
sitions are given in Table 3.1.3, contained approxi-
mately 1.2% manganese, 2% niobium, and 1 to 2%
silicon, Inserts 321 and 322, which had the
lowest and the highest silicon contents, re-
spectively, were tested in loops which circulated
the fuel mixture {No. 30) NaF-ZrF 4UF, (50-46-4
mole %) at 1500°F, while the remammg insert,

- 323, with an intermediate silicon content, was

_tested in a loop which circulated sodium at

1500°F.

As shown in Table 3,1.4, very severe corrosion:

of the cast specimens was found in the loops
which circulated the fuel mixture, The attack
occurred not only in the form of the subsurface
voids that are typical of the attack of wrought
Inconel, but, in addition, very deep intergranular
penetrations appeared, which reached in the worst
case to a depth of 70 mils, as shown in Fig. 3.1.3,

These intergranular penetrations, although a
result of rapid attack of grain-boundary con-
stituents, were aided considerably by the porosity
and the shrinkage cracks present in all the castings
in the as-received condition, as shown in Fig. 3.1.4.
The casting containing 1% silicon was found to be

TABLE 3.1.3. CHEMICAL ANALYSES OF CAST INCONEL INSERTS TESTED IN
STANDARD INCONEL THERMAL-CONVECTION LOOPS

 

Cast inconel Insert

Chemical Analyses (wt %)

 

 

No. Ni* Cu Fe Si Mn Cc Cr Nb -8

3N 70.49 0.01 8.20 1.04 1.22 0.23 16.67 2.08 0.005
322 69.79 0.03 8.20 1.93 1.17 0.22 16.51 2.09  0.006
323 70.44 0.02 8.10 1.34 1.6 0.22 16.67 2,00  0.004

 

*Obtained by difference analysis.

TABLE 3.1.4. RESULTS OF METALLOGRAPHIC EXAMINATION OF INCONEL THERMAL-CONVECTION .
LOOPS OPERATED WITH CAST INCONEL INSERTS IN THE HOT LEGS
Operating time: 500 hr
Maximum fluid temperature: 1500°F -

 

Loop ' . Insert

" Metallographic Results

 

 

Circulated Fluid
No, ~ Ne. Hot-Leg Attack Cold-Leg Attack
876 o 321 NoF-ZrF4* Cast section, 25 mils Light surface roughening v\lrifh,
7 (50-46-4 mole %). Weld interface, 70 mils a metal deposit present
877 322 NaF-ZrF4-UF4* Cast section, 23 mils Light surface roughening with -
(50-46-4 mole %) Weld interface, 25 mils evidence of metal cry;fol; |
878 Control NaF-ZrF -UF ,* 10 mils | " No attack
' (50-46~4 mole %) '
879 323 Sodium No attack No attack; no deposifs

 

*Fyel mixture No. 30.

136
 

 

o/

 

 

 

Fig. 3.1.3. Region of Maximum Attack Near
Casting-Weld Interface in-Inconel Casting No. 321
Exposed for 500 hr to the Fuel Mixture (No. 30)
NaF.-ZrF -UF4 (50-46-4 mole %) at 1500°F as an
Insert in the Hot Leg of a Wrought Inconel Thermal-
Convection Loop. Etched with modified aqua

regia. 100X. Reduced 32.5%. {Seeret-witireoptien)

 

Fig. 3."-l.'4._7,_:'Pofes'ity in As Recewed |ncene|:-

Casfing ,N-°' .321. IOOX » Reduced 32.5%_. :

no. befler from the corrosion stondpomt than the.i'
' ;,one contammg 2% suhcon. o S e
" Metal depos:ts were “seen metallographlcally o

und wsually ‘in the hot “legs of  both “loops.

'Spectrogrcphlcally, the . deposits were found to be
:predommantly nickel, with aluminum, ‘chromium,
“and “iron reported as .major elements.  The silicon
and mangonese contents™ in’ these deposns were

low,
Loop 879, which contalned insert 323 and circu-
lated sodium, revealed very little corrosion

PERIOD ENDING JUNE 10, 1956

(<1 mil) in the insert section. Metallic deposits,
which were found by analysis to be predominantly
nickel, were visible in the trap area. The amount
of deposited material was approximately the same

-as that normally found in wrought Inconel loops.

It is apparent that Inconel castings of the compo-
sitions tested are not svitable for use in contact
with the fuel mixture (No. 30) NaF-ZrF -UF,
(50-46-4 mole %). While the castings showed
resistance to sodium, further tests of longer
duration will be required to completely evaluate
mass-transfer effects.

FLUORIDE FUEL MIXTURES IN HASTELLOY
THERMAL-CONVYECTION LOOPS

J. H. DeVan E. A. Kovacevich

Three Hastelloy X (ref. 4) thermal-convection
loops, which were constructed from. ¥.in.~dia
tubing, were operated 1000 hr, two with sodium
at 1500°F and one with the fuel mixture (No. 30)
NaF-ZrF -UF, (50-46-4 mole %) at 1500°F.
Loops 855 and 856, which operated with sodium,
had no hot-leg attack and no evidences of metallic
deposits in the cold leg. Loop 857, however,

which circulated the fuel mixture, showed con-

siderable hot-leg ottack to a depth of 27 mils.
The cold leg of this loop was attacked to a depth
of 1 mil, and there were evidences of metallic
crystals and a metallic layer, as shown in
Fig. 3.1.5. The metallic crystals in the trap area

were analyzed and found to be predomlncntly

chromium,
Two loops, 872 and 873, constructed of 3/-|n.-d|c

~ Hastelloy W (ref. 5) tubing were also operofed for
1000 hr with sodium and with the fuel mixture
(No. 30) NqF-ZrF‘_-UF‘ (50-46-4 mole %), re-
- spectively, . at 1500°F,
- hot-leg attack; however, loop 872, which operated

Neither loop showed

with sodium, had scattered metallic crystals in

_the cold leg. The hot-leg samples of both loops
- were similar metallographically to the as-received

samples of each loop.
The significant difference in attack on the two
alloys, Hastelloys X ‘and W, by the fuel mixture

can undoubtedly be explained by the difference
~in the chromium content of the two alloys, 22

 

4The nominal . composition .of Hastelloy X is 22%
Cr-23% Fe~9% Mo-1.5% Co~balance Ni.

5The nominal composition of Hastelloy W is 5%
Cr—5% Fe-24% Mo=1% Co-balance Ni.

137

 
 

 

 

 

ANP PROJECT PROGRESS REPORT

and 5%, respectively, As may be seen in Table
3.1.5, the chromium content of the fuel mixture
circulated in the Hastelloy X loop was much higher
than that of the fuel mixture circulated in the
Hastelloy W loop., This reflects the greater
tendency toward chromium = removal -and the
resultant void formation in alloys with high
chromium contents, such as Hastelloy X.

FLUORIDE FUEL MIXTURES IN MONEL
" THERMAL-CONVECTION LOOPS

J. H. DeVan

Three monel thermal-convection loops, 806,
808, and 809, completed 500, 545, and 1027 hr of

operation, respectively, with the fuel mixture

B URCLASSIFIED
B y.95a7

BRI

mc{m

: B

3L

T
L

 

 

  

Fig. 3.1.5. Metallic Crystals ond Loyer De-
posited in Cold Leg of Hastelloy X Thermal-
Convection Loop Which Circulcted the Fuel
Mixture (No. 30) NaF-ZrF -UF, (50-46-4 mole %)
for 1000 hr at o Hot-Leg Tempercture of 1500°F.
Etched with HCI-H,CrO,. 250X, Reduced 32.5%.

s ” or)

{(No. 107) NaF-LiF-KF-UF, (11.2-41-45.3-2.5
mole %) circulating at a hot-leg temperature of
1500°F. Loops 808 and 809, which were scheduled
to operate 1000 and 1500 hr, respectively, were
terminated before completion of the scheduled

operating period because of excessive oxidation

of the outside surface of the monel tubing. The
oxidation did not completely penetrate the tube
wall in either loop. Metallographic examination
revealed hot-leg attack to a depth of 1 mil, as
shown in Fig. 3.1.6, in all these loops. Three
other monel loops, 880, 881, and 882, that had
been chromium-plated to provide oxidation pro-
tection were operated under the same conditions,
but, apparently, the platings were imperfect,
and two of these three loops, 881 and 882, were

 

Fig. 3.1.6. Region of Maximum Attack of Hot
Leg of Monel Thermal-Convection Loop 809 Which
Circulated the Fuel Mixture (No. 107) NaF-LiF.
KF-UF, (11 .2-41-45.3-2.5 mole %) for 1027 hr at
a Hot-Leg Temperature of 1500°F. Etched with

HNO,-acetic acid. 250X. Reduced 32.5%. Cocre-

ir e eertion)

TABLE 3.1.5.- CHEMICAL ANALYSES OF THE FUEL MIXTURE (No. 30) NaF-ZrF -UF4 (50-46-4 mole %)
BEFORE AND AFTER CIRCULATION FOR 1000 hr AT 1500°F IN HASTELLOY X AND W
THERMAL-CONVECTION LOOPS

 

Loop . Loop

Uranium Centent

Impurities Found (ppm)

 

 

Sample Taken -
No, Material (wt %) Ni - Cr Fe
857 "~ Hastelloy X Before filling 8.39 260 135 115
After draining 8.68 40 1060 125
873 - Hastelloy W Before filling 8.78 80 60 75
| After draining 8.82 - 85 275 110

 

138
 

 

%)

terminated after 1217 and 1339 hr of the 1500 hr
scheduled because ' of excessive oxidation and
resultont leaks in the heated areas . of the loops,
The ! remcmmg Ioop, 880, completec} its scheduled
1000 hr. of operaticn, - Metallographic - unalyses
of these Ioops hcve not been completed

SPECIAL FLUORIDE FUEL MIXTURES IN
INCONEL THERMAL-CONVECTION LOOPS

J. H. DeVan

- Additional Inconel thermal-convection loops
have been operated to evaluate corrosion properties

PERIOD ENDING JUNE 10, 1956

of. several special fluoride fuel mixtures, . Tests
of 500 hr ‘duration were completed for mixtures. in
the MF-ZrF -UF , system, where MF is KF, LiF,

NaF, or RbF M:xtures with each of these fluorides
and 40 or 46 mole % ZrF , have been tested; each
mixture contained 4 mole % UF,. The results
obtained for the fuels containing 46 mole % ZrF

were reported previously,3 but they are repeated ln
Table 3.1.6 for comparison, Lowering the ZrF4
content, as shown in Toble 3.1.6, had no effect on
attack by the NaF-containing mixture, but the
attack by the KF- and RbF-containing mixtures
increased from 1 to 2 mils, The LiF-containing

TABLE 3. 6o RESULTS OF METALLOGRAPHIC EXAMINATIONS OF INCONEL THERMAL-CONVECTION
LOOPS OPERATED WITH SPECIAL FLUORIDE FUEL MIXTURES
AT A HOT-LEG TEMPERATURE OF 1500°F

 

Metallographic Results

 

 

Loop Fue- 'Mtx‘tur'e“-. Fuel Mixture Comfiosifion OP_::::ng Maximum Hot-Leg .
No. Ceode Designation (mole %) (hr) Attack | Cold-Leg Deposits
(mils)

845 94 KF-ZrF -UF ;; 50-46-4 500 65 Metallic layer

846 94 KF-ZeF ~UF ,; 50-46-4 500 8 Metallic layer

883 94 KF-ZrF ~UF ; 50-46-4 1500 n Metallic layer

884 94 KF-ZrF -UF ,; 50-46-4 1500 9 Metallic layer

930 WR4 KF-ZrF ~UF ; 56-40-4 500 7 Metallic layer

931 WR4 KF-ZrF 4" UF i 56-40-4 500 , 9 Metallic layer

847 93 _ LiF-ZrF -UF‘; 50-46-4 500 17.5 Possible crystals

848 93 . LIF-ZF, UFG 50—46-4’ 500 19 ‘Possible crystals
s WR3  LiF-ZrF CYFG S6-40-4 500 10 Metallic layer and crystals

7 929 WRS  LIF-ZiF-UF i 56.40.475 00 10 Metallic layer and crystals
ey 95 inF-ZrF‘-Ua-. S0464 . S0 9 Metaltic layer
S 840 95 _'fi_RbF-ZrF -UF4;-50-45-4— 8500 - . 9 Metallic lgy"ejl"’ - .
932 . WRS | RI:F-ZrF -ur=4. 56+40-4 < 500 - 10 Metallic layer and crystals
933 WRS RbF-IeF (UFi56-404 500 - 1 Metallic layer and crystals
030 NeRZFaUR s0464 500 105 - Neme

8130 NeF-ZrF CUFiS0-d64- SO0 9 Nome o

926 - WRZ NgE-ZrF_ ~UF i . '40-'4._'*7?".7 500 10 'Metallic layer ond crystuls C

927 - WR2 - NaF-ZrF -UF ; 56-40-4 .. 500 7 Metallic layer and crystals

 

139

 
 

 

 

ANP PROJECT PROGRESS REPORT

mixture showed quite anomalous behavior in that
the attack decreased markedly with the decrease
in ZrF ;. These results are being rechecked.

"The lbobs' containing r_n'ixtures with 56 mole %
of the alkali-metal fluoride, with the exception of
the loop operated with the KF-containing mixture,

‘showed traces of metallic crystals in the cold

legs. These crystals were most pronounced in the
RbF-containing mixturés.  Specimens of the
crystals found in loop 933, which circulated the
fuel mixture RbF-ZrF ,-UF , (56-40-4 mole %), are
shown in Fig. 3.1.7.

" Two additional Inconel loops, 883 and 884,
were operdied under similar conditions to obtain
further corrosion data for the KF-ZrF .UF,
(50-46-4 mole %) system. An increase in the
operating time from 500 to 1500 hr caused an
increase in maximum attack of 2 mils.  Thin
‘metallic layers were noted in the cold legs after
both operating periods. As yet these layers are
unidentified,

140

 

 

Fig. 3.1.7. Metallic Crystals and Layer De-
posited in Cold Leg of Inconel Thermal-Convection
Leop 933 Which Circulated the Fuel Mixture
(No. WR5) RbF-ZeF, -UF, (56-40-4 mole %) for
500 hr ot a Hot-Leg Temperature of 1500°F,
Etched with modified aqua regia. 250X. Reduced
32.5%. (@Eecret-with-capiien;
 

 

PERIOD ENDING JUNE 10, 1956

3.2. GENERAL CORROSION STUDI ES .
E. E. Hoffman

BRAZING ALLOYS IN L1QUID METALS AND
IN FLUORIDE FUEL MIXTURES
D H Jonsen

Two Inconel thermal-convection loops with in-
serts brazed with 70% Ni-13% Ge-11% Cr~6% Si
brazing alloy in the hot-leg section were operated
for 500 hr.
circulated, and in the other the fuel mixture (No.
44) NoF-ZrF -UF, (53.5-40-6.5 mole %) was cire
culated., The hot-leg sections of these loops
consisted of seven Inconel segments brazed with
the alloy being tested, This type of thermal-
convection joop test assembly was illustrated
previously.! The hot legs of both loops were

 

!D. H. Jansen, ANP Quar. Prog. Rep. Dec. 10, 1955,
ORNL-2012, p 113, Fig. 5.9.

In one loop NaK .(56-44 wt %) was

maintained at a temperature of 1500°F for the
duration of the tests. The cold-leg temperature
of the loop that circulated NaK was maintained
at 1165°F, and the cold-leg temperature of the
loop that circulated the fuel mixture was held
at 1150°F.

The inner walls of the Inconel segments and two
samples from each brazed joint were examined
after the tests for evidence of attack. The results
of the examinations are summarized in Table 3.2.1.

The attack of the fuel mixture on the brazed
joints averaged about 5 mils. Typical samples
are shown in Fig. 3.2,1. The ottack of the fuel
mixture on the Inconel tubing adjacent to the
joint ranged from 4.5 mils in the coolest (1450°F)
portion of the hot leg to 9 mils in the hottest
(1500°F) portion of the hot leg, as shown in
Fig. 3.2.1. A slight trace of metallic crystals

TABLE 3.2.1. RESULTS OF THERMAL-CONVECTION LOOP TESTS OF INCONEL SEGMENTS BRAZED
WITH 70% Ni-13% Ge~11% Cr=6% St BRAZING ALLOY

Operating time:

500 hr

 

Brazed Joint

Metallographic Notes

 

No. ' Effect of Fuel Mixture

Effect of NaK

 

1* Brazing alloy attacked to o depth of 4.5 mils;
adjacent Inconel tube attacked to a maxi-

. mum depth of 9 mils

2 Brazing alloy attacked to a depth of 4.5 mils;

No attack apparent; large crack through brazing

alloy to center of tube wall

Attack to o deprth of 0.5 mil

spofly, nonunlforrn attack on tube wall to a

T maximum depth of 7 mils

3 - _’ Brczlng alloy aflacked to a de'fifh of 6 mils;

No attack; large crack, as in joint 1 above
- spotfy oflack on tube wu" fo a depth of
4 ' i-:-_:i'--Brazing al loy uttacked to a depfh of 3 mils; Attaék to a depth of 0.5 mil; large crack in

- the wall uflacked to a depth of 4 5 mils

5 T Brozing alloy afi'acked to @ depfh of 5 mlls,
- - tube woll attacked ‘Io a depth of 4. 5 mils .

- 6** Brazing ulloy aflocked to a depth of 5 mils;
' o tube wull oflacked toa depth of 4.5 mils

brazing alloy

~ Large crack in brazing allay

Attack to'ordepfh'of 1.5 mils; some porosity in
- brazing alloy; no large cracks

 

*Brazed joint No. 1 was located at the hottest (1500°F) portion of the hot leg.
** Joint No. 6 was located at the coolest (1450°F) portion of the hot leg.

141

 
 

 

 

ANP PROJECT PROGRESS REPORT

 

 

Fig. 3.2.1. Specimens of Inconel Joints Brazed with the 70% Ni~13% Ge-11% Cr—6% Si Brazing Alloy
and Exposed for 500 hr to the Fuel Mixture (No. 44) NaF-ZrF ,-UF, (53.5-40-6.5 mole %) in the Hot Leg

of an Inconel Thermal-Convection Loop.

(@) Joint located in coolest (1450°F) portion of hot leg. (b)

Joint located in hottest (1500°F) portion of hot leg. Etched with 10% oxalic acid. 150X. Reduced 19%.

{Seerotwith caption)

was found in the cold leg of the loop that circu-
lated the fuel mixture. Spectrographic analyses
of these crystals showed strong lines for chromium
and nickel and weak lines for iron. There was
no evidence of mass transfer in the loop that
circulated NaK.

The brazing alloy showed good corrosion resist-
ance to NaK, as shown in Fig. 3.2.2, but large
cracks which extended to a depth of one-half the
tube wall thickness were found in some of the
brazed joints, Fig. 3.2,3. It is not known whether
these cracks were caused by thermal stresses or
shrinkage. Anclysis of the NaK after the test
showed 1160 ppm of oxygen. This is admittedly
high, but no efforts were made to purify the NaK
which was received in container lots.

A third Inconel thermal-convection loop was
operated for 500 hr in which the hot leg contained
Inconel inserts brazed with the 82% Au-18% Ni
brazing alloy. The fluid circulated was the fuel
mixture (No. 30} NaF-ZrF -UF, (50-46-4 mole %).
The hot and cold legs were maintained at tempera-
tures of 1500 and 1175°F, respectively. Micro-

142

scopic examination of samples from each brazed
joint showed the attack on the alloy to average
about 8 mils, with a maximum depth of 10 mils.

Tests were also conducted on a series of buttons
of Coast Metals brazing alloy No. 52 (89% Ni-
5% Si—-4% B-2% Fe), which were exposed in see-
saw apparatus to NaK (56-44 wt %) and to the
fuel mixture (No. 44) NaF-ZrF-UF, (53.5-40-6.5
mole %). Since previous tests of this alloy showed
depletion of the second phase at the exposed
edge,?2 these tests were conducted to determine
whether the high-cross-section boron component
was removed and, if so, whether the removal was
time dependent. The buttons were contained in
nickel tubes with hot-zone temperatures of 1500°F
for both tests and with cold-zone temperatures of
1100°F in the tests with NaK and 1200°F in the
tests with the fuel mixture. In all the tests, the
specimens were retained in the hot zone of the
test capsule. The duration of the exposures to

 

2p. H. Jansen, ANP Quar, Prog. Rep. Dec. 10, 1953,
ORNL-2012, p 119, Fig. 5.17. :
 

 

 

 

NaK were 100 and 350 hr, and the exposures to
the fuel mixture were of 100 and 500 hr duration.

The results of the tests are presénted in Table
3.2.2. Boron was found only in very small quan-
tities .in the areas depleted of the second phase
by exposure to NaK and to the fuel mixture. A

microspark traverse on the sample exposed to the.

PERIOD ENDING JUNE 10, 1956

fuel mixture for 500 hr showed that the concen-
tration of boron was less than 1% from the surface
to a depth’of 8 mils. The boron content then rose
sharply to its normal value (4%) and stayed there

for the remainder of the traverse. Samples of the

depleted area, obtained by microdrilling, were

analyzed and were found to contain 0.6% boron.

 

Fig. 3.2.2. Specimens of Inconel Joints Brazed with the 70% Ni~13% Ge~11% Cr-6% Si Brazing Alloy
and Exposed for 500 hr to NaK at 1500°F in the Hot Leg of an Inconel Thermal-Convection Loop. These
two specimens, No. 2 and No. 6, illustrations (z) and (), respectively, were the only samples examined
that d:d not have iarge cracks in the brazmg alloy. Etched wuth 10% oxul:c acid. 150X. Reduced 8%.

TABLE 3.2.2. RESULTS OF SEESAW CORROS!ON TESTS OF COAST METALS
BRAZING ALLOY No. 52 (89% NI-S% Si-4% B-2% Fe)

 

L 7 Durarticzfi"cf_”
Bath . Test

Weight Loss of Depth of Edge Depleted

 

_ Specimen of Second Phase
Ahr) (- {mils)
CONaK (S6-44wt®) 100 0.07 X
‘ B - 350 0.23 4
NuF-ZrF -UF o 1000 0.06 - 3
- (53. 5-40-6 5 mole %) ' ' U
500 0.34 | 6

 

143

 
 

 

 

 

ANP PROJECT PROGRESS REPORT

It was also found that the silicon content had
dropped to about one-third its normal value (5%)
in the depleted area. Evidence of the time de-
pendence of the amount of depletion is presented
in Fig. 3.2.4. None of the buttons tested showed
attack to a depth of more than 0.5 mil.

" Hardness measurements on the interior of the

 

 

 

Fig. 3.2.3. Typical Crack Found in an Inconel
Joint Brazed with the 70% Ni=13% Ge=11% Cr~£%
Si Brazing Alloy After Exposure for 500 hr to NaK
at 1500°F in an Inconel Thermal-Convection Loop.
Etched with 10% oxalic acid. 100X. Reduced 29%.

 

button exposed to NaK for 350 hr gave a value of
716 DPH, while measurements at the edge gave
a value of 145 DPH, as shown in Fig. 3.2.5. The

buttons tested in the fuel mixture gave similar .

hardness traverse results. Chemical analysis of
the fuel mixture from the 500-hr test and of the

NaK used in the 100-hr test showed significant

concentrations of boron.

NIOBIUM IN STATIC SODIU_M
D. H. Jansen

Specimens of niobium were tested in . static
sodium at 1500°F for a period of 1000 hr to de-
termine the suitability of liquid sodium as a pro-
tective environment for niobium during high-
temperature mechanical property tests. The fests
were designed to show whether the niobium would
be corroded by the sodium and whether the hard-
ness of the niobium would be appreciably altered
through the pickup of oxygen from the sodium.

The specimens were contained in type 304 stain-
less steel and Inconel capsules, and the varicbles
such as the volume of the bath, the container size,
and the area of the specimen were adjusted to
obtain o surface-area-to-volume ratio that would
be as close as possible to that found in creep-test
equipment. Cold traps were utilized on the bottom
of each capsule to reduce the amount of oxygen
in the sodium bath. The niobium specimen tested
in the type 304 stainless steel capsule showed
more evidence of surface roughening than did the
specimen tested in the Inconel capsule, Fig. 3.2.6,
and the thickness loss was also more than that
of the specimen tested in the Inconel capsule,
Table 3,2.3. A thin, brittle layer, approximately

TABLE 3.2,3. RESULTS OF TESTS OF NIOBIUM IN STATIC SODIUM

Exposure time: 1000 hr
Temperature of sodium: 1500°F

 

 

 

Hardness Thickness Loss Impurity Analysis of Sample (ppm)
Sample (VHN) (mils) | "
mils H, o, N,y Cc
As-received material 130.5 2.4 40 125 300
Specimen tested In type 304 135.0 3.0 3.8 84 250 370
stainless steel tube
Specimen tested in Incone! 116.4 1.8 3.4 90 92 1180

tube

 

144
 

 

 

 

PERIOD ENDING JUNE 10, 1956

 

Fig. 3.2.4. Specimens of Coast Metals‘ Brazing Alloy No. 52 (89% Ni-5% Si—4% B=2% Fe) After Ex-
posure in Seesaw Apparatus to the Fuel Mixture (No. 44) NaF-ZrF -UF, (53 5.-40-6.5 mole %) for (a) 100
hr and (b) 500 hr. 150)( Reduced 12. 5%._ (See-pat—m-t-h—euphun-}- -

 

 

Fig. 3.2.5. Specimen of Coast Metals Brazing Alloy No: 52 (89% Ni—5% S$i-4% B—2% Fe) After Ex-

posure in Seesaw Apparatus to NaK for 350 he.

acid.

150X.

Note differences in hardness. Etched with 10% oxalic

145

 
 

 

 

ANP PROJECT PROGRESS REPORT

l UNCLASSIFLED |
S Y 18196

 

 

 

 

 

 

 

 

Fig. 3.2.6. Specimens of Nicbium After Exposure to Static Sodium at 1500°F in (2) an Inconel Capsule
and (b) a Type 304 Stainless Steel Capsule. Etched with 25% HF-25% H,SO,-50% H,0. 250X. Re-

duced 17%.

0.3 mil thick, formed on the surface of both speci-
mens during the test, This layer, which is very
hard (~1180 VHN), is probably a niobium-nickel
alloy resulting from dissimilar-metal transfer, but
it ‘has not yet been positively identified.

Vickers hardness traverses were made on the
as-received and as-tested specimens, The hard-
ness of the specimen tested in the type 304 stain-
less steel capsule increased (130 to 135 VHN)
during the test, whereas the hardness of the speci-
men tested in the Inconel capsule decreased (130
to 116 VHN). A correlation of the hardness values
with the impurity analyses indicates that nitrogen
is the major hardening agent, A niobium specimen
previously tested in argon at 1500°F for 2000 hr
showed a hardness increase from 127 to 168 YHN.,
The impurities (H,0, Hy, N 50 and 02) of the argon
used in the test totaled 130 ppm.

THERMENOL IN STATIC SODIUM
D, H. Jansen
Samples of Thermenol (82% Fe-15% Al-3% Mo)
cut from a piece of hot-rolled strip were corrosion
tested in static sodium for 100 hr at 1500°F. The
specimens were contained in AIS|I 1035 steel and

type 430 stainless steel capsules. The specimen
tested in the AlSI 1035 steel capsule showed a

146

weight loss of 0.22%, and the specimen tested
in the stainless steel capsule showed a weight
loss of 0.07%. A slight roughening of the surface
occurred on both samples, Fig. 3.2.7. Neither the
type 430 stainless steel nor the-AlS| 1035 steel

capsule was attacked during the tests.
STATIC TESTS OF ALFENOL

E. E. Hoffman
Specnmens of Alfenol (84% Fe-16% AI), sub-

mitted by The Glenn L. Martin Co., were tested

under static conditions at 1500°F in the fuel
mixture (No. 44) NaF-ZrF «UF, (53, 5-40-6.5
mole %), lead, lithium, and sodlum for 100 hr.
Since no Alfenol container tubes were available,
Inconel wos used as the container material for
all the tests, The Alfenol specimens vused in
these tests were 4 -in. cubes, The specimens
were placed in l/-m.-OD '0.035-in.-wall Inconel
tubes, together wnth sufficient test medium to
give 3 in, of liquid bath at the test temperature.

The specimen tested in the fuel mixture was
covered with black crystals and showed o 30%
weight increase. The crystals were analyzed and
found to be UF,. The specimen was attacked
throughout its thickness along the grain boundaries,

Fig. 3.2.8.
 

A PERIOD ENDING JUNE 10, 1956

 

 

 

Flg. 3. 2 7. Specsmens of Thermenol (82% Fe-15% Al-3% Mo) After Exposure to Static Sodium at
: 1500°F in (a) a Type 430 Stainless Steel Capsule and (5) an AISI 1035 Steel Capsule. Etched with aqua
regia. 250X. Reduced 16%.

 

 

| Fig. 3 278 Spreclmens of Alfenol (84% Fe-]6% Al) ln" (a) As-Recewed Condmon and (b) After Ex-
posure for 100 hr to the Static Fuel Mixture (No. 44) NuF-ZrF UF (53.5-40.6.5 mole %) ot 1500°F. The
- grain-boundary attack shown in (b} is typical of the attack found throughout the specimen. (a) Etched

L/ with aqua regia. (b} Unetched. 250X. Reduced 13.5%. dSecrétwith caption)m

 

147

 
 

 

 

 

ANP PROJECT PROGRESS REFPORT

No weight change data were taken on the speci-
men tested in lead becouse porticles of lead ad-
hered to the specimen. Metallographic examination
revealed attack only in a few scattered areas to
a depth of 0.5 mil.

The specimen tested in lithium was attacked
throughout its thickness along the grain boundaries,
which- were apparently aluminum-rich regions.
Gentle tapping caused the specimen to break up
-inte individual grains, Fig. 3.2.9. The lithium
bath was found to contain 0.2 wt % aluminum after
the test,

The specimen tested in sodium lost only 16 mg
(0.009 wt %) during the test and metallographic
examination revealed no attack. The sodium bath
was found to contain 0.001 wt % aluminum after
the test.

SODIUM-BERYLLIUM-INCONEL COMPATIBILITY
' IN STATIC SYSTEMS
E. E. Hoffman

A test was conducted in order to determine the
thickness of Inconel that would be consumed by

an alloying reaction between Inconel and beryllium
in direct contact under: pressure while immersed
in sodium at 1300°F. The apparatus used for
the tests is shown in Figs., 3.2,10 and 3.2.11.
The surfaces of two of the Inconel specimens used
in this test, as shown in Fig. 3.2.12, were chro-
mium plated to study the effect of chromium in
reducing the extent of the alloy formation between
nickel (from the Inconel) and beryllium. The test
specimens were ‘/4 X yz x 1 in. with the Pz % 1-in.
surfaces in contact. Sufficient load was applied
through a compression rod and bellows to yield
e 500-psi. stress on the specimens, The test
assembly was loaded with sodium and held at
1300°F for 1000 hr. The load was applied to the
specimens when the test temperature was reached.

It was found after the test that the chromiume-
plated Inconel specimens could be separated from
the adjacent beryllium specimens; however, it was
impossible to separate the unplated Inconel speci.
mens from the beryllium specimens. The results
of metallographic examination of the four specimen
interfaces, presented in Table 3.2.4, indicate that

 

Fig. 3.2.9.. Alfenol (84% Fe-|6% Al) Cube That Dlsmtegrafed After 100 hr of Exposure to Static

Lithium ot 1500°F. 12X

148

 
 

 

 

 

N

~ of BeNi adjacent to the Inconel.

UNCLASSIFIED
Y2952

LOAD

 

Fig. 3.2. 10., Apparutus for Studymg the Extent

of A[ioynng Between Beryllium and Various Meiqls -

Under Sh'ess Wl'ule lmmersed in Molten Sodium.

a thin’ chrommm p!ute on. lnconel ‘does “not ‘elimis"

nate alloying - ‘reactions ‘with beryllium when the

two “materials are in contact while lmmersed in

hsgh-femperature sodium, but the plate does sub-

stantially reduce the extent of alloy formation.
Approxlmately 4 mils of Inconel ‘was consumed

in the formuhon of the 24.mil nickel-beryllium

'clloy layer which was found where the Inconel

and - the beryillum were  in direct contact (Fig.
3.2.132). The major porhon of reaction layer
was found to be Be,,Nig, with a small percentage
The reduction
in alloy formation brought about by the 2-mil chro-
mium plate is illustrated in Fig. 3.2.13b, A second
test is now under way in which the effectiveness
of 4- and 6-mil chromium platings will be evaluated.

 

PERIOD ENDING JUNE 10, 1956

UNCLASSIFIED
Y1

«17961

 

Fig. 3.2.11.

Enlarged View of Test Specimens

' Shown in Figs. 3.2.10 and 3.2.12.

UNCLASSIFIED
Y.13508

INCONEL

bqRécr-bomAcT
BERYLLIUM
DIRECT CONTACT
INCONEL

A-mil CHROMIUM PLATE
eERVLLIM
5 nil CHROMIUM PLATE

JINCONEL

Sodium-Beryllium-Inconel Com-

3.2.12.
patibility Test Specimens After Exposure fto
Sodium at 1300°F for 1000 h:.

Fig.

149

 
 

 

 

ANP PROJECT PROGRESS REPORT

TABLE 3.2.4, RESULTS OF METALLOGRAPHIC EXAMINATION OF THE INTERFACES OF THE
SPECIMENS SHOWN IN FIG, 3,213

~ Test environment: sodium
Test duration: 1000 hr
Test temperature: 1300°F
Stress on specimens: 500 psi

 

 

Interfccé o ' | Metallographic Notes -
A-B:* Inconel vs beryllium - dlrect 10 to 24 mils of alloy formation (Be,Nij plus BeNi) along
‘contact . interface; 4 mils of Inconel consumed in the production of

reaction layer

B-C: Beryllium vs Inconel = direct 22 to 24 mils of uniform alloy formation along interface
contact |

C-D: Incone!l with 1-mil chromium 3.5 to 7 mils of alloy formation; 7-mil layer found where
plate vs beryllivm plating was thinnest =~

D-E: Berylllum_ vs 2-ml| chromium 2 mils of interaction between specimens along 90% of inter-
plate on Inconel ' _ face; 6-mil layer detected at one area where plating

appeared to have been defective

 

 

*See Fig. 3.2.12 for location of interface.

 

 

FEERE

¥

mcsae:a;

BITTTELTTTEL

v
1

 

 

 

 

fi_i; ETT l L

T
50

T
1
4

 

 

 

 

 

 

 

 

Fig. 3.2.13. Comparison of Results of Direct Contact of Beryllium and Inconel () with Results of
Contact Between Beryllium and 2-mil Chromium Plate on Inconel (b). ‘Areas designated A, B, and E may
be located on Fig. 3.2.12. (@) As polished. 150X. Reduced 15%. (&) Etched with oqua regia. 1000X.
Reduced 15%.

150

 
 

 

 

- STATIC TESTS_.OF_INCQNEL‘ C_AS'flNGS
" R. Carlander '

In thermal-convection loop tests ‘of Inconel cast-

ings in the fuel mixture (No. 30) NaF- ZrF -UF,
(50-46-4 mole %), much heavier attack of the cast
matericl occurred than is normally observed for
wrought Inconel (see Chap, 3.1, **Dynamic Cor-
rosion Studies’’). Metallographic examination of
the as-received material showed the presence of
cracks, stringer porosity, and large grains, with
the grain boundaries perpendicular to the surface.

In order to obtam further information on the effect

of the composition of the cast material on its
corrosion resistance, static tests of three Inconel

castings that differed primarily in silicon content

were performed in the fuel mixture (No. 30) NaF-
ZrF -UF, (50-46-4 mole %)} in wrought Inconel
capsules for 1000 hr at 1500°F. Examination of
the tested specimens revealed that the casting
with the highest silicon content (1 .93%) was the
least attacked, while the casting with the lowest
silicon'contenf was the most severely attacked,
In all cases, the attack was a combination of
subsurface voids and intergranular penetration.
The depth of subsurface voids on the Inconel
castings was 3 to 4 mils, as compared with 2 to
4 mils on the wrought Inconel capsules, The
intergranular - ‘penetration, however, varied from
3 mils on the high-silicon-content casting to 12
mils on the lowssilicon-content casting. This

-surface of the casting.
lographic examinations of the three cast specimens
~are presented in Table 3.2,5, and the wrought

of the lithium on corrosion resistance,

PERIOD ENDING JUNE 10, 1956

_difference is attributed to the porosity and to the

grain boundaries running perpendicular 1o the
The results of metal-

Inconel capsule and cast Inconel specimen No. 321
are shown in Fig. 3.2.14. In every case the
wrought Inconel container was more corrosion

resistant than was the cast Inconel specimen.

INCONEL AND STAINLESS STEEL IN NaK
CONTAINING LITHIUM

R. Carlander

Type 316 stainless steel and Inconel were ex-
posed to NaK containing lithium in static and in
dynamic systems in order to determine the effect
In the
static tests the specimens were exposed to NaK

| (5644 wt %) with lithium additions of 1, 5, 10,

20, and 30 wt % for 100 hr ot 1500°F. No attack
occurred in any of the tests,
For the dynamic tests, 12-in.-long capsules

B were filled to 40% of their volume with NaK plus

lithium (5 or 10 wt %) and placed in a tilting-
type furnace (1 cpm) at @ hot-zone temperature
of 1500°F (cold zone, 1100°F) for 100 hr. No
mass transfer occurred in ony of these systems.
The ‘Inconel exposed to NaK with 10 wt % lithium

_added was attacked to a depth of 1 mil in the hot

zone, while the other capsules were unotfocked.

TABLE 3.2,5. RESULTS OF METALLOGRAPHIC EXAMINATION OF CAST INCONEL SPECIMENS AFTER
EXPOSURE TO STATIC NeF-ZrF‘-UF‘ (50-46-4 mole %) FOR 1000 hr AT 1500°

 

“silicon

Depfh of Aflack (mlls)

 

 

,-Ccst Inconél _ Mt ’ -hi N f‘.:’ : Cost
Specimen Content Wrought lnconel " Cast’ Inconel e :::::elcs' :c:':e:"/ o8
~ Neo. o Awt %) Capsule Specirnen o Ppecimel
321 o o 1.04 n ‘ . 4 o - 12 ' Subsurface voids 'trp_'ge'pfl'\ of 4 mlls, R
L - o - ' o o intergranul'ar'pe'n-e'tr:dtioh' to "dépth' '
_ 7 ’ '_ of 12 rnlls, no crucks cpparent o
| 322 o 1.93 B 2 » 3 B ',Subsurfcce volds und Intergrunular -
RN o . ‘ penetration to depth of 3 mlls, no .
o crucks cpparenf ' : '
- 323 -_: o ,1.734 B : 2 - 8 . 7___-;.Subsurface voids to depth of 3 mi!s, :

lnfargrcnulot penetrutlon to depth :
of 8 mils; no cracks apparent

 

151

 
 

 

 

 

ANP PROJECT PROGRESS REPORT

8 UNCLASSIFIED §
8883

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Fig. 3.2.14. Speéimens of Wrought Inconel Capsule (a) and Cast Inconel Specimen No. 321 (b) After
Exposure to the Fuel Mixture (No. 30) NaF-Z¢F ,-UF, (50-46-4 mole %) for 1000 hr at 1500°F Etched'
with aqua regia. 250)( Reduced 3%. (€eeret-wfih-up-hvrr)

Thermal-convection loop tests were also con-
ducted. In these tests the hot-leg temperature
was 1500°F (cold leg, ~1250°F), and NaK with
5 wt % lithium added was circulated for 1000 hr.
A slight amount of mass transfer, in the form of
small adherent crystals containing 71.5% Ni, 4.6%
Cr, ond 0.7% Fe, occutred in the Inconel loop,
while no mass transfer was found in the stainless
steel loop. Portions of the hot and the cold legs
of the Inconel and the stainless steel loops are
shown in Figs. 3.2,15 and 3.2.16, respectively.
The Inconel was attacked to o depth of 1.5 mils
in the hot and in the cold legs. The stainless
steel was unattacked in the hot leg but was
attacked to a depth of 6 mils in the cold leg.
Photomicrographs of the cold legs of the Inconel
and of the stainless steel loops are shown in
Fig. 3.2.17. The addition of 5 wt % lithium to
the NaK did not decrease the normal corrosion
resistance of Inconel to NaK, but it did decrease

152

that of type 316 stainless steel, The difference
in attack between the hot and cold legs of the
stainless steel loop may be attributed to the larger
amount of carbides present as a network in the
cold leg than in the hot leg, where more carbides

are in solution because of |ncreased solublhty '
at the higher temperature. '

TRANSFER OF CARBON BETWEEN DISSIMILAR
METALS IN CONTACT WITH MOLTEN SODIUM

R. Carlander

The decarburization of AIS| 1043 sfeel by molten'
sodium in Armco iron and type 304 ELC stainless
steel containers was demonstrated 'in tests at
1830°F of 100 hr duration, as described previously.3
An additional test has been performed to determine

 

3E. E. Hoffman, ANP Quar. Prog. Rep Dec. 10, 1955,

-~
 

 

[

 

UNCLASSIFIED
Y. 18385

 

HOT ZONE —816°C (1500°F)

     

COLD ZONE-—660°C(1220°F)

Fig. 3.2.15. Portions of Hot and Cold Legs of
an Inconel Thermal-Convection Loop in Which
NaK with 5 wt % Lithium Added Was Circulated
for 1000 hr, (Secret with caption) |

it - PERIOD ENDING JUNE 10, 1956

UNCLASSIFIED -
Y. 18384

ot

 

COLD ZONE —665°C(1230°F)

Fig. 3.2.16. Portions of Hot and Cold Legs of a
Type 316 Stainless Steel Thermal-Convection
Loop in Which NaK with 5 wt % Lithium Added

Was Circulated for 1000 hr. (Secret with caption)

 

 

  

 

   

Fig. 3.2.17.  Cold Legs of (a) an Inconel Thermal-Convection Loop and (5) a Type 316 Stainless Steel

Thermal-Convection Loop in Which NaK with 5 wt % Lithium Added Was Circulated for 1000 hr ot ¢ Hot--
Leg Temperature of 1500°F and a Cold-leg Temperature of About 1225°F. Etched with aqua regia. -

250X. Reduced 4%. (Secretwithcuprion)

153

 
 

 

 

 

 

ANP PROJECT PROGRESS REPORT

the effect of time on the extent of decarburization
of the steel specimens.

Two AISI 1043 steel specimens and sodium were

loaded into evacuated Armco iron and type 304
ELC stainless steel containers and tested at
1830°F for 400 hr. The conditions and results
of the tests are given in Table 3.2.6. As expected
- on the basis of the previous investigation, exten-
sive decarburization of the steel specimens oc-

curred in both the iron and the stainless steel
containers. The extent of decarburization of the
AISl 1043 steel specimens was greater in these
400-hr tests than in the previous 100-hr test,
and, in addition, the vapor zone of the stainless
steel container was carburized in the 400-hr test
and not in the 100-hr test. This discrepancy may
be due to the vapor zone sample having been taken
closer to the bath zone after the 400-hr test than

TABLE 3.2,6. CARBON AND NICKEL ANALYSES OF THE YARIOUS COMPONENTS OF SYSTEMS DESIGNED
FOR STUDYING THE DECARBURIZATION OF MILD STEEL BY SODIUM

 

Material Analyzed

Welghf Loss in

Carbon Content Nicke! Content

 

Test System 400-hr Test
o {wt %) (wt %) (g/‘n.2)
AISI 1043 steel 3pécimen in Steel specimen
Armeco iron contalner As received 0.433 0.008
After 100-hr test 0.121 0.0001
After 400-hr test 0.054 10.0408
iron container
As received 0.018
Vapor zone
After 100-hr test 0.019
After 400+hr test 0.018
Bath zone
After 100-hr test 0.035
After 400-hr test 0.02¢4
AlS| 1043 steel specimen in Steel specimen
type 304 ELC stainless As recelved 0.433 0.008
steel container ) S
After 100<hr test 0.100 . _0.9002
After 400-hr tost 0.074 0.090 0.0303
Steel container
As received 0.037 11.12
Vapor zone
After 100-hr test 0.022*
After 400-hr test 0.162** 10.32
Bath zone
After 100-hr test 0.1_28
After 400-hr test 0.200 , 9.98

 

*Sar;\ple taken 4 in. above bath level.
**Sample taken 1 in. above bath level.

154

 
 

 

o

after the 100-hr test. The extent of the decarbu-
rization of the AIS| 1043 steel was greater in the
Armco iron container than in the stainless steel
container. Furthermore, sufficient nickel mass-
transferred to the surface of the steel specimen
in the stainless steel container to cause a phase
transformation. to & depth of 2 to 4 mils, but the
amount of nickel that mass-transferred did not seem
to increase appreciably in comparison with the
amount transferred during the 100-hr test. Photo-
micrographs of the as.-received and as-tested AlSI

1043 steel specimens are shown in Flgs. 32 18
and 3,2.19.

RARE-EARTH OXIDES EXPOSED TO STATIC
SODIUM IN INCONEL CONTAINERS

- W. H. Cook

In considering rare-earth oxides as possible
control rod materials for the ART, sodium was
suggested as a heat transfer medium for cooling
the control rod material, and it was proposed that
Inconel be used as the container material for the
sodium and the rare-earth oxides. Tests were
therefore conducted to investigate the corrosion

 

PERIOD ENDING JUNE 10, 1956

resistance of the rare-earth oxides to sodium ond

the effects, if any, on the Inconel containers.
One specimen of Sm,0, (density, 5.88 g/cm?;
apparent porosity, 25.4%) was tested for 1000 hr

in" static sodium ot 1500°F in an Inconel capsule.

Two specimens from a body of a mixture of rare-
earth oxides (density, 6.58 g/cm3; apparent po-
rosity, not determined) were tested similarly, one
for 500 hr and the other for 1000 hr. The body,
fabricated from a commercially available powder
known as Lindsay Mix, obtained from the Lindsay

Chemical Co., had the composition 63.8 wt %

Sm,04~26.3 wt % Gdzos-baiance primarily other

' rure-earth oxides,

The results of the -tests indicated negligible
corrosion attack by the sodium on the three speci-
mens and their Inconel capsules in 1000 hr, but
there may have been some reduction in their
strengths. The weight and dimensional changes

‘of the three specimens were all positive, as would

be expected with porous materials, and were less
than 0.5%. The original buff colors of the speci-
mens changed to gray black, but this was the
only macroscopically visible change. It was found

 

Fig. 3.2.18. Specimens of AISI 1043 Steel (a) in the As-Received Condition and (5) After Exposure to
Sodium in an Armco lron Container for 400 hr ot 1830°F. Etched with 4% picral. 200X. Reduced 10.5%.

155

 
 

 

 

ANP PROJECT PROGRESS REPORT

that the gray-black color extended throughout the
specimens.
completely penetrated the specimens, probobly
along the pore spaces..

Powder x-ray diffraction patterns of the untested
ond tested pieces of Sm,0, and the two Lindsay

‘Mix specimens did not reveal any reaction products.

Chemical analyses of the sodium baths indicated

This is evidence. that the sodium-

negligible quantities of rare earths, chromium, u
iron, and nickel, which also supports the con-
clusion that there was negligible attack on either
the rare-earth oxides or the Inconel capsules.
The major losses of rare-earth oxides were cal- .
culated on the basis of the chemical analyses,
and the results are summarized in Table 3.2.7.
In comparing the results from the 500- and 1000-hr

 

 

 

 

 

+

Fig. 3.2.19.  Specimens of AISI 1043 Steel After Exposure to Sedium in c.Ty;.:e 304 ELC Stainless Steel

Container at 1830°F for 400 hr. (a) 200X. Reduced 13%. (b) 500X. Reduced 13%. Etched with 2% nital.

TABLE 3.2,7. RESULTS OF TESTS OF VARIOUS RARE-EARTH OXIDES EXPOSED TO STATIC
SODIUM IN INCONEL CONTAlNERS '

 

 

 

 

e " Caleculated Oxide Loss of Specimen* :
Specimen Density Apparent Test Test | (x 10"37) '
, 3 Porosity Temperature Duration °
Material {g/cm®) A , _ )
Sm,04 5.88 23.2 1500 ,10004_ ‘ 2.9 2.6 0.18 043 5.‘8_
Lindsay Mix**  6.58 1500 500 29 108 . 076 ' 0.41 . 5.
6.58 1500 1000 48 ° .08  0.82 067 7.4 .
3.53 53.5 1300 100 132 590 216 17.4 387
*These values were calculafed on the basis of the rare earths found by chermcol analyses in the sodium from each -
test. .
**Composition: 63.8 wt % Sm,04, 26.3 wt % Gd203, balance primarily other rare-earth oxides. u

156
 

 

 

 

i

tests of the Lindsay Mix specimens, no significant
differences were found.

Metallographic polishing and examination of the
rare-earth oxide specimens before and after ex-
posure to the sodium indicated that the specimens
may have been slightly weakened by the tests,
but there was no microscopically visible corrosion,
as shown in Fig. 3.2.20. Metallographic examina-
tion of the Inconel capsules revealed slight sur.
face roughening to a depth of less than 0.5 mil

in spots on the inner surfaces that contacted the |

liquid sodium and the Lindsay Mix specimens.

There was no surface roughening and no attack.

of the inconel capsule that contained the Sm,0,
specimen and sodium. Spectrographic examina-
tions of the inner surfaces of the three capsules
did not revecl any rare earths.

The corrosion resistance to molten sodium shown
by the Lindsay Mix with a density of 6.58 g/cm®
prompted further tests that were designed to ap-
proximate more closely the expected operating
conditions and for which a Lindsay Mix body with
the proposed density and configuration was used.

 

PERIOD ENDING JUNE 10, 1956

The purposes of these tests were to determine
the relationship of the apparent porosity of the
Lindsay Mix body for water to that for moiten

~ sodium and to study the corrosion resistance of

the body and Inconel when separated by only a
small thickness (approximately 0.05 in.} of sodium.

A hollow cylinder of porous Lindsay Mix body
was used that was nominally 0.9 in. OD, 0.5 in,
ID, and 1 in. long, with a density of 3,53 g/cm?®
and an apparent porosity to water of 53,5%. This
body had an apparent porosity of 52% to sodium
after exposure to static sodium for 100 hr at
1300°F in an evacuated Inconel capsule. To
determine the apparent porosity of the body to
sodium, the sodium was allowed to cool and
solidify around the body at the end of the 100-hr
test period. The test capsule was then opened
in a helium atmosphere, the excess sodium was
removed from the surfaces of the body with plastic
scrapers, and the body and its absorbed sodium

were weighed, The absorbed sodium was then

removed by vacuum distillation, and the body was
again weighed. The apparent porosity of the body

‘Fig. 3.2.20. Rare-Earth Oxide Specimens (63.8 wt % Sm,0,-26.3 wt % Gd,0,~Balance Primarily
Other Rare-Earth Oxides) (4) in As-received Conditien and (b) After Exposure to Static Sodium in an
Inconel Container for 500 hr at 1500°F. Unetched. 500X. Reduced 5.5%.

157

 
 

 

 

 

ANP PROJECT PROGRESS REPORT

to sodium was calculated by using the latter weight,
the density of 3.53 g/em3, the weight of the
absorbed sodium, and a sodium density at 97.8°C
(the melting point of sodium) of 0.951 g/em®.

As shown in Table 3.2,7, the losses of rare-
earth oxides by this low density (3.53 g/emd)
body were greater than the losses by the higher
density (6.58 g/cm3®) Lindsay Mix body, which is
in keeping with the general trend for the lower
density ceramics to have lower corrosion resist-
ance. There is no explanation for the relatively
large loss of Y,0,. The exposure to sodium also
slightly weakened the body; its hardness, on the
Mchs' scale, changed from 3 to 2 during the test.
Long-term corrosion tests of finished ART control
rod Lindsay Mix shapes and sodium in Inconel
containers have been initiated to determine the
extent of the corrosion of these materials under
‘the most adverse operating conditions expected
and for periods equal to and greater than the pro-
posed operating time.

SOLID-PHASE-BONDING SCREENING TESTS
W. H. Cook

Molybdenum has been tested for 100-hr periods
for sohd-phase bonding with itself, tungsten,

158

K150A (80% TiC-10% NbTaTiC4~10%  Ni), and
K152B (64% TiC-6% NbTaTiC;~30% Ni) with
calculated contact pressures of 20,000 psi in the
fuel mixture (No. 30) NaF-ZrF -UF, (50-46-4
mole %) at 1500°F. Bonding was absent only
for molybdenum vs tungsten and molybdenum vs
K150A. The surface roughnesses of the contacting
surfaces of all specimens were 1.5 to 2 p:m rms,
as determined with a profilometer.

The contact 'pressure was redueea from the
standard 50,000 psi to 20,000 psi in an effort to
get below the yield strength of molybdenum at
1500°F. The 20,000-psi loading was not low
enough, inasmuch as the molybdenum deformed
slightly in all the tests. All examinations were
made with a low-power microscope. |

In the test of molybdenum vs molybdenum, the
solid-phase bonding was accompanied by some
deformation and considerable upsetting.- In the
bonding between molybdenum and K152B, a thin
film appeared to have been transferred from the
molybdenum surface to the K152B surface in the
contact area. This was probably a nickel-
molybdenum diffusion layer, but it has not yet
been identified. -

 
 

 

 

o .":_extrusaon. =

 

 

PERIOD ENDING JUNE 10, 1956

3.3. FABRICATION RESEARCH

J. H. Coobs

DEVELOPMENT OF NICKEL-MOLYBDENUM
BASE ALLOYS

T. K. Roche

The available supply of Hastelloy B and W
tubing was prepared by redrawing tubing formed
from welded strip, and it is not of high quality.
Since these alloys are characterized by high-
temperature strength and resistance to fused-salt
corrosion, a source of seamless tubing is desired,
and therefore considerable effort has been devoted
to the study of extrusion techniques for the pro-
duction of seamless tubing of these Hastelloys
‘and other related alloys.

H. Inouye

Extrusion of Hastelloy W

Early attempts to extrude tube blanks of
Hastelloy W failed. In these experiments forged
billets canned in Inconel were extruded at 2050
and 2100°F at an extrusion ratio.of 7:1.
fracturing of the tubing occurred during the ex-
trusion process as a result of melting. An
excessive temperature rise as g result of the rapid
deformation of the material is believed to have
caused the difficulty. Recently this alloy was

successfully extruded under conditions identical -

to those used earliéf, as listed below, except for
the extrusion temperature and rate:

Forged Hastelloy W canned in Inconel

Billet

Billet size 3 in, OD, 13’ in, ID 3 ln. long

Die size 1/ In., 30-deg cofe Lo
Mcndre_"-"sii_zr'_e‘ ) in,

_ .Exfrusién.'f;:tio"~. 7-1» :

AT
4

Cperiments; ~
The extrusion temperature of 2200°F used for
the successful extrusion was an mcrease over

Severe

the 2050 and 2100°F used for the unsuccessful
extrusions, It was concluded from this work that
Hastelloy W could be extruded on a laboratory
scale by raising the extrusion temperature and
lowering the extrusion rate to allow time for the
heat generated in the work to dissipate to the
surroundings. The successful and unsuccessful
tube-blank extrusions are illustrated in Fig. 3.3.1.

Exirusi_dn of Hastelloys B, W, and X
on ¢ Commercial Scale

An experiment was performed at the Huntington
Works of the International Nickel Company on
March 27, 1954, to determine the feasibility of
producing tube blank extrusions of Hastelloys B,
W, and X on a commercial scale. The experiment
was carried out as a cooperative effort between

the Oak Ridge National Laboratory, the Inter-

national Nicke! Company, and the Haynes Stellite
Company.

Forged and machined blllets of the Hastelloys
were. prepared by the Haynes Stellite Company.

Based upon extrusion experience on a laboratory

scale at ORNL, as described above, a schedule

-~ was prepared for the extrusion of the Hastelloy B

billets. Since small tube blanks had been prepared
successfully by using a slow extrusion rate at

2200°F, it was proposed that these larger billets

be extruded at as slow a rate as practical. The
extrusion temperature could be lower, of course,

. because of the larger size of the billet. .The
~-results of the Hastelloy B extrusions, which are
" “summarized in Table 3.3.1, were reported to ORNL
“ by the Intemnational Nickel Company upon com-
‘ ; _ - . - pletion of the experiment.
e The rate of exirus:on was conh'olled by regulahngy‘*‘_f. tube blanks of Hastelloy B are shown 1 Flgs.
" the flow ‘of high-pressure water to'.the ram of }3 3.2 and 3.3.3,
" the ‘extrusion press by means of a throttle valve *
- which opened completely in four tums. The rate -
7. seftings, .in terms of: the number of turns that .
. the wvalve ‘was Opened were 3% and 3% for the
,'unsuccessful exfrustons “and ‘2'4 for the successful

o Actucl ‘extrusion rates, in feet per
- minute, - were. not defermlned durlng these ex-
' ‘ . in.slong billets of Hastelloys B, W, and X:

The actual ‘extruded

- Although the Hastelloy-w and X blllets were not

~ extruded under the direction of ORNL, the fabri-
cation of these alloys is of interest to the over-all
“alloy. development program. The avallable results
_ of these extrusiens are presented in Table 3.3.2.

- The followmg general conclusions have been
drown regarding ‘the extrusion of ‘6, 9-m.-0D 10-

1. Forged billets of Hastelloys B, W, and X can

‘be successfully extruded to tube blanks.

159

 
 

091

 

 

TABLE 3.3.1. EXTRUSION OF HASTELLOY B BILLETS BY INTERNATIONAL NICKEL COMPANY FOR OAK RIDGE NATIONAL LABORATORY

Billet dimensions: 6,900 in. OD, 2,500 in, ID, 10 in, long o :
Machined size of billats canned on OD only: 6,650 in. OD, 2,50 in iD 'IO in, Iong
Machined size of billets canned on OD and ID: 6.500 in, OD, 2,75 in. ID, 10 in. long
Machined size of uncanned billets: 6.900 in. OD, 2.50 in. ID, 10 in. long -
Product dimensions: Outsida diameter, as indicated below

Inside diameter, 2,250 in,

 

Extrusion Pressures

 

 

Billet Billet Pfeparofion Nose Extrusion Total Scoking Extrusion Product (tsi) Extrusion Rote Usable Length
No, m Rc:dius Temp:rofure Time (min) Ratio oD (in.) - . of Billet Length  of P.roduct
(in.) (°F) Peak Minimum Running (ips) (in,)
B-1 Uncaonned Uncanned 1 2050 220 5.5:1 3.685 77.1 771 77.1 .2 41
B-2 Canned Uncahned 0.5 2050 201 5.5:1 3.685 7.3  66.8 68,3 1.2 51
B Conned  Canned | 0.5 2050 187 5.5:1 3,685 683 6.0 62,3 1.2 - 47
B-7 Uncanned Uncanned 1 2150°¢ 100 5.5:1 3.685 74.3 623 68.3 1.3 47
8-89 Uncanned Uncanned 1 2150¢ 110 7:1 3.462 771 77 77.1 0 0
B-10 Uncanned Uncanned 1 2150 65 7:1 3.462 758 713 68,3 1.4 - 58
BB? Uncanned Uncanned 1 2150 75 71 3462 77 77 77.1 0 0
BOB Canned Uncanned 0.5 2150 65 7:1  3.462 74,3  68.3 69.7 1.2 62
B-3 Canned Uncanned 0.5 2150 70 7:1 3.462 72,8 68,3 69.7 1.4 70
B-5 Canned Uncanned 0.5 2200 100 10:1 3,135 77,1 743 771 O.S 81
B-4 Canned Conned 0.5 2200 105 12.25:1 3.000 77.1 72,8 1.5 .99

7.3

 

4| isted in extrusion order.

bSoaked in gas furnace, except as indicated,
©Soaked in salt bath after prehecmng to 700 F.

dBillet stalled.

s
=

LAOdIY SSTAI0A¥d L23r0dd dNV

 
 

 

PERIOD ENDING JUNE 10, 1956

 

| l"'-ig.'3.3.'l'.‘ Seamless Tube Blanks Extruded from Forged Hastelloy W Billets Canned in Inconel. Tube
1 was extruded at a fast rate at 2100°F. Tube 2 was extruded at a slow rate at 2100°F.

e

TABLE 3.3.2. EXTRUSION OF UNCANNED HASTELLOY W AND X BILLETS
BY INTERNATIONAL NICKEL COMPANY

" Billet dimensions: 6,900 in. OD, 2.50 in. ID, 10 in. long; Nose radius, 1 in,
Product dimensions: Outside diameter, as indicated; Inside diameter, 2,250 lrlf

 

 

- , A Extrusion - Outside Diameter ~ Usable Length
Bitlet ~  "Extrusion ‘Temperature* Extrusion of Product of Product
Material =~ - No. - (°F) Ratio (i) . {inJ)
Hastelloy W 12¢¢ 2150 12500 3,00 0
13 2150 7.9:1 3.28 67
Hastelloy X u o 2150 7 3.28 &7
s 2150 7.9:1 3.28 67
' 2150 0 3135 85
17%* . 2150 12,251 .00 0
. - 18 2200 12.25:1 3.00 104
19 - 2200 12.25:1 300 104

 

*Soaked in gas furnace.
**Billet stalled.

¥

o/

161

 
 

 

ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
Ya18908

 

 

 

0 S 10 15 20
HASTELLOY 8 TUBE BLANKS EXTRUDED FOR ORNL BY INCO

BILLET EXTRUSION TEMPERATURE . EXTRUSION RATIO
BBB | 2150°F T
B-2 2050°F 5Ya:1
B-7 2150°F 8o
B - 2050°F sl 1
B-1 2050°F ’ 5l 1

Fig. 3.3.2. Hastelloy B Tube Blank Extrusions Fabricated from Commercial-Size Billets by the Inter
national Nickel Company.

UNCLASSIFIED
Y-18907

 

 

o 5 10 5 20 ‘
HASTELLOY B TUBE BLANKS EXTRUDED FOR ORNL BY INCO

BILLET EXTRUSION TEMPERATURE EXTRUSION RATIO
B-4 2200°F 12V - 1
B-5 2200°F 10 :1
B-3 2150°F 7:
B-10 2150°F 7:1

Fig. 3.3.3. Hastelloy B Tube Blank Extrusions Fabricated from Commercial-Size Billets by the Inter-
national Nickel Company.

162
 

 

-

"0

 

2. Canning of the billets with ‘/-m <thick type
316. stainless steel appears to be advantageous
from the standpoint of reducing the extrusion

pressure, Because of the number of variables

involved, however, this pomt needs to ‘be con-
firmed.

3. The larger mass of the commercmlly produced
billets made possible the use of lower soaking
temperatures at the smaller extrusion ratios (i.e.,
5.5:1) than were found to be optimum for the
extrusion of small laboratory billets.

4. For successful extrusion of these alloys it
is necessary that the billets be upset with suf-
ficient pressure fo start the material through the
die and then extruded at a relatively . slow rate to
prevent hot-short cracking. :

These were the first successful attempts to
extrude these materials on a large scale, and it
now appears to be feasible to produce seamless
tubing of these alloys. Three Hastelloy B tube
blanks and one Hastelloy W tube blank are
scheduled to be reduced to 1-in. pipe and small-
dmmeter tubing for use at ORNL.

Extrusion of Special A"oys

The study of smgle-phase alloys in the nickel-
molybdenum alloy system that contain 15 to 20%
molybdenum was continved. It is hoped in this
study to find a solution to the problem of em-
brittlement of the Hastelloy-type alloys as a result
of aging. At present the only readlly apparent
solution is to change the composmon of the
alloys. ’

It was reported previously | thcn‘ the corrosion

‘resistance to fused .sdlts of the bmury alloys of
nickel and molybdenum which contain mone than =

15% molybdenum was exceptionally good but ‘that

their high-temperature strength was not adequate. o
Thus in order to obtain alloys with greater sirength
the special -alloys described . below “are bemgf‘ g
studied, In addition to the binary. ulloys, ternary"'?'
alloys are . being mode ‘that will be-tested to
determine the effect of a thlrd element on corrosion .
- '_-1resnstonce of these alloys in. fused salts, ¢

International - Nickel Company Composztzdns.

o V—V:As reported prevtously, -40-1b -heats of - hve

in Table 3.3.3; -

 

1T, K. Roche and H. Inouye, ANP Quar. Pfog. Rep.
March 10, 1956, ORNL-2061, p 155.

T.23011 2060 -
: 723012 275

‘special afloys were recewed from - INCO- for COre '~ .G
- rosion testing - and - strength evaluahon. The_‘,:  T-23013 2100 .

“nominal - composmons of these a”oys ure glven iT.~230'14"' . 2010

- PERIOD ENDING JUNE 10, 1956

TABLE 3.3.3. COMPOSITIONS OF SPECIAL ALLOYS
PREPARED BY INTERNATIONAL NICKEL COMPANY

 

Composition (wf %)

 

 

Alloy No.
Mo Cr W Nb Al Ti C N
T-2011 15. 5 3 -3 0.5 73.5
T-23012 17 0.5 82,5
T23013 15 3 3 0.5 785
T-23014 15 , 1 15 82,5
0,25 78.25

T23015 15 3 3 0.5

 

“Four extrusion billets, .3 in. in diameter and
3 in. long, were machined from each ingot, three
for tube blanks and one for rod fabrication. The
rod extrusions were for rolling to 0,065-in. strip

for the preparation of test specimens for strength

evaluation. The initial tube blank extrusion ex-
periments involving one billet of each composition
met with little success. When a fast extrusion
rate was used, the alloys demonstrated the hot-
short tendencies of Hastelloy B, with the ex-
ception of alloy T-23013, 15% Mo-3% Nb-3%
W-0.5% Al-78.5% Ni. The results of this initial

~experiment are presented in Table 3,3.4.

 Before the remaining billets of these alloys were
extruded, the results obtained in the fabrication
of Hastelloy W at a slow extrusion rate were

‘sufficiently favorable to prompt an alteration in

TABLE 3.3.4, RESULTS OF PRELIMINARY
ATTEMPTS TO EXTRUDE THE SPECIAL
ALLOYS PREPARED BY THE INTER-
.. NATIONAL NICKEL COMPANY

Extrusion ratio: 5.4:1
- Mandrel size: 32 in, dia

 

.. _Extrusion ‘
A’ l?? Temperufure . Results
T NO. N i (OF) E .

 

 Extruded tube blank shattered

. Exfruded tube blank shofleredr |
‘-Good tube blan obfcuned

¢ 7;- Press stafled

Tube blank cracked on the

inside

T.22015 2080

 

163

 
 

 

 

 

 

 

ANP PROJECT PROGRESS REPORT

the extrusion conditions for the special compo-

sitions. Two billets, T-23012 and T-23014, were

extruded slowly at 2200°F with a ¥.in.-dia
- mandrel at an extrusion ratio of 5.4:1, Bol11 alloys
responded satisfactorily, but the mandrel broke
within the tube blank in each case. With a de-
creased rate of extrusion the billet remained in
contact with the mandrel for an excessive length
of time, and the resultant overheating of the tool
caused a subsequent ‘decrease in its ultimate
strength. The utilization of a 1-in.-dia mandrel
alleviated this problem. The final results of slow
extrusion of these alloys at 2200°F are presented
in. Table 3.3.5. Extrusions 1 through 5, shown
-in Fig. 3.3.4, are representative of the good tube
blanks obtained from these special compositions.
The tube blanks have been sent to the Superior
Tube Company for redrawing to 0,500-in.-OD,
0.035-in.-wall tubing.

"TABLE 3.3.5. RESULTS OF THE FINAL EXTRUSIONS
' OF SPECIAL ALLOYS PREPARED BY THE
| INTERNATIONAL NICKEL COMPANY

Mandrel size: ]-in. dia

 

 

| Alloy Exfru'srion ~ Extruded :u::::rs;:‘:
No. Ratio Shape Extrusions
T-23011  7:1 Tube blank 2
T-23012 7:1 Tube blank 1
T-23013 7:1 Tube blank 2
T-23014 . 7:1 Tube blank 1
T-23015 - 72 Tube blank 2
T-23011 6,251 Rod 1
T.23012 6,251 - Rod 1
T-23013  6.25:1 Red 1
T-23014  6.25:1  Rod 1
T-23015  6.25:1 Rod 1

 

Hot rolling of the extruded rods was only moder-
ately successful. The rolling temperature was
varied between 1925 and 2100°F, and the
scheduled reduction was approximately 10% per
pass. In general, the lower temperature proved
to be more satisfactory for minimizing edge
cracking. Each alloy was hot rolled to 0.250-in.
strip, pickled, annealed at 2050°F for 1 hr, and

164

cold finished to 0.065-in. strip with an inter-
mediate anneal at 2050°F for % hr. Final an-
nealing of each alloy was scheduled to yiel_d a
grain size of ASTM 6-7. Alloy T-23015 edge-
cracked severely during hot rolling. The other
alloys for which edge cracking did not interfere
with the machining of sound test specimens will
be evaluated for strength.

Battelle Memorial Institute Compositions. — In
cooperation with Battelle Memorial Institute in the
development of nickel-molybdenum-base alloys for
high-temperature use, extrusion experiments were
conducted on three promising compositions with
the intention of producing seamless tubing for
corrosion testing. A total of 10 forged billets
representing portions of 220-lb air-meited heats
was fabricated. The results of the experiments
are presented in Table 3.3.6.

Although fast extrusion at a ratio of 5.4:1 re-
sulted in cracking on the inside, all these tube
blanks are being salvaged at the Superior Tube
Company by drilling the as-extruded tube blank
and thus increasing the inside diameter from ¥ to
1 in. before tubing reducing. Tube No. 6, shown
in Fig. 3.3.4, is representative of a good extrusion
obtained from these alloys. As is typical of most
of these high-strength nickel-molybdenum-base
alloy extrusions, a certain amount of surface
roughening was found on the inside, but it was
not so severe as it appears to be in Fig. 3.3.4.
Slight conditioning of the blanks before redrawing
will eliminate these defects. The cause of this
roughening is not completely understood at this
time, but it may be related to the surface condition
of the biliet before extrusion or to lubrication
practice during extrusion. ~All the blanks are
being fabricated to 0.500-in.-0OD, 0.035-in.-wall
tubing. o '

ORNL Compositions. — The basic nickel-
molybdenum alloy containing 15 to 20% mo-
lybdenum possesses corrosion resistance on a par
with that of Hastelloy B, better ductility and
fabricabifity, but poorer strength properties. Con-
sequently, alloy additions to the basic composition
are required for improving the strength properties.
The elements being investigated as strengtheners
are Ti, Al, W, Nb, Cr, Fe, V, and C. In order to
determine the effect of these elements individually
on corrosion resistance and to establish the
maximum quantity of each element that can be
 

Gl

 

N

 

*}

 

UNCLASSIFIED
Y:18794

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Mo | Cr |[Co | W a|Ti | M| ¢ | N Molcr | co{ wla [ 1 M| ¢} ni
1" [ 151 5 3 1]05] — — — | BAL 6 20 | - 1 - - 2 |o80 012 | BAL
2 17| — | =] ~jos5| -] -1 - |ea 7 117 3| -1 - | -1 =1 - loos]|saL

-3 115l - | 3] 3]os| -] -1 ~ |sa 8 | 17| 5| — | = | -] = | - loos|BAL
4t 15 — [ — | ~-11 15| -1 — |saL 9 HASTELLOY B

s || —13f3los| -] — |o2sleAL

 

 

 

 

 

 

 

 

 

 

 

 

Fig. 3.3.4. Extruded Tube Blanks from Special Nickel-Molybd;num Alloys Prepared by the International Nickel Company.

9561 ‘0l INNT ONIANT QOI¥3d

 

 
 

- ANP PROJECT PROGRESS REPORT

TABLE 3.3.6. RESULTS OF EXTRUSION EXPERIMENTS ON SPECIAL ALLOYS PREPARED
BY BATTELLE MEMORIAL INSTITUTE

 

Extrusion Conditions

 

 

 

Alloy Nominal Compositions (wt %)
Ne.  Ni Mo Ti € Mn Tm(?,:; " Ratic  Rate* Results

B-2897 77 20 1 0,12 0.80 2060 5.4:1 3 Back of tube cracked on inside
2100 5.4:1 3 Back of tube cracked on inside

B-2898 76 20 2 012 080 2060 5.4:1 3 Back of tube cracked on inside
2100 5.4:1 3 Back of tube cracked on inside
2150 71 2% Good tube blank obtained
2150 7.1 2% Good tube blank obtained

B-2899 78 20 0.20 0.80 2060 5.4:1 3 Back of tube cracked on inside
2150 541 3 Back of tube cracked on inside
2125 5.4:1 3 Back of tube cracked on inside
2150 721 2% Good tube blank cbtained

 

 

 

 

*Number of turns that the valve on the high-pressure water to the extrusion ram was opened; see discussion above.

present without adverse effect on corrosion re-
sistance, ternary alloys are being prepared that
will contain 17 wt % molybdenum, the elements
listed below in the amounts shown, and the
balance nickel,

Quant ity to

Element to Be Added
Be Added (wt %)

Cr 3,5,7,10

W 2,4

Ti 2,4

Nb 24

Al 2,4

Fe 7, 20

C 0.1. 0.25' 0'50

Vacuum-induction heats of each composition are
being prepared in 36-lb billets. A small amount
of carbon is added to each charge to bring the
resultant carbon level to 0.06%. It is possible
to machine three tube-blank extrusion billets from
each ingot for the fabrication of sufficient tubing
to make three standard thermal-convection loops
for corrosion testing. Thus far the chromium-
bearing alloys have been prepared, as well as the
alloys with 2 wt % tungsten, 2 wt % titanium, and
2 wt % aluminum, Good tube blanks were obtained

166

by extrusion of the chromium-bearing alloys ot
2100 and 2150°F. In all cases, a slow extrusion
rate of approximately 1 in. of billet length per
second at o ratio of 7:1 was used, Tubes 7 and 8
of Fig. 3.3.4 are representative product samples.
It is known that chromium additions above @
certain minimum amount are detrimental from the
standpoint of corrosion by fluoride fuels, but
chromium is a desirable addition for imparting
oxidation resistance to the alloy. Results of
previous tests of chromium-bearing nickel-mo-
lybdenum alloys in the fuel mixture (No. 30)
Nc::F-ZrF“-UF4 (50-46-4 mole %) indicated that
chromium additions in excess of 5% caused de-
creased corrosion resistance; at least 7% chromium
is required, however, to make the alloy resistant
from an oxidation standpoint, The chromium level
that can be tolerated when the alloy is in contact
with the fuel mixture (No. 107) NaF-KF-LiF.UF,
(11.2-41-45.3-2.5 mole %) is to be determined.

In all extrusion experiments carried out this
quarter, difficulty was encountered occasionally
with billets failing to extrude to completion. This
trouble was atiributed to excessive chilling of
the billiets by the cold ram at the slow extrusion
rates being used. Therefore a new billet was
designed that has mild-steel nose and tail plates
tack welded to it. The plates were expected to
 

 

o

Ay

allow the billet to start through the die more
easily, as well as to transfer the chilling action
by the ram to a more easily extruded material.
The mild steel is cropped from the tube blank
after extrusion. Steel washed back on the outside
surface of the tube is machined off before the tube
blank is reduced to tubing.

Improved lubrication of the billet was . obtained

by coating the container of the press, as well as

the mandrel, with Necrolene grease prior to each
extrusion, Low-melting-point Fiberglas mats are
also placed around the hot billets before they are
introduced into the extrusion press. This lubri-
cation practice has resulted in lowering the
pressure required for extrusion of these materials.

Consumable-Electrode Experiments

As indicated previously,I arrangements were

‘made with Battelle Memorial Institute for the prepa-

ration of arc-melted ingots of nickel, Hastelloy B,

Hastelloy W, o 76% Ni~17% Mo-7% Cr alloy, and

an 83% Ni=17% Mo alloy. The melts were to be

made by the consumable-electrode process to take

advantage of the high arc temperatures for vapor-

izing *‘“tramp"’ elements in an effort to improve
the strength and fabricability of these alloys.
Electrodes of the first three alloys were to be
supplied in the form of rolled rods, The special

-mckel-molybdenum alloys were to be prepared

by vacuum melting, and the electrodes were to be
fabricated by threading together extruded rods of
the material. All the electrodes are now ready
for shipment. :

 

PERIOD ENDING JUNE 10, 1956

OXIDATION OF HASTELLOY B
H. Inouye J. E. Spruiell?

It was reported previously® that the oxidation
rate of Hastelloy B in air at 1500°F was not
excessive, When the alloy was thermally cycled
from 1500°F to below about 660°F, however, the
rate was increased by an order of magnitude as
a result of spalling of the protective NiMoO, scale
from the metal surface. This spalling is caused
by a phase transformation in the NiMoO, layer at .
about 660°F.

The oxidation of Hastelloy B in static air has

now been investigated at 1200, 1400, 1600, and

1800°F.  Data were obtained for mechanically
polished specimens exposed to air at the various
temperatures for periods of 168 hr. The increases
in the specimen weights were determined at
frequent intervals to obtain oxidation-rate curves.
The total weight increases of the specimens
during the tests at the various temperatures are
shown in Table 3,3,7, The curves of the weight
gain vs time were parabolic at all test temper-
atures, and thus the oxide scale appears to be
protective,

The oxidation rate of the alloy at 1200°F is
very low, and the superficial scale which forms
does not spall upon cooling. Preliminary x-ray
data show that this scale is principally NiO.,

Thus for in these studies of the oxidation
characteristics of nickel-molybdenum alloys, evi-

. dence has been found that the formation of NlMoo

2Cooperahve-program studenf on asslgnmenf frorn fhe 7

Universlty of Tennessee. - : -

3H. lnouye und Jo Ho Coobs, ANP Quar. Prog. Rep. ,

-depends upon the molybdenum content. However,

its formation may also be a funcflon of temper-

L Vature, as mdl cated in the tests. descrlbed above. _

TABLE 3.3 7. RESULTS OF OXIDATION TESTS OF HASTELLOY B EXPOSED TO STATIC AIR FOR 168 hr

 

 

 

 

L T Tast - “Averuge'Totcl WGilh' o
s Numb" :,Temperuture "o 7 Gain of Specimen in - Remorks
5 ,Of Tfsts GRS 168 hr* (9/cm2) o
4 .‘ 1200 . 0.00Q31-'* : Oxide did not spall on coolmg '
s o' 0,00095 Oxide spalled on cooling |
gt e T ooms Oxde spslled on coslng
S L1800 00,0095 Oxide spalled on cooling

 

*Weight gain of Inconel in 168 hr at 1500°F = 0.00004 g/cm?,

167

 
 

 

 

ANP PROJECT PROGRESS REPORT

COMPOSITE TUBING
M. R. D'Amore?

The fabrication of composite tubing by coex-
trusion was described previously.5¢¢ [Initially
these studies were made in order to develop

H. Inouye

. billets which, when extruded, would result in
-composite tubing with predetermined thicknesses

of the various metal layers. To date, billet
configurations for the fabrication of two- and
three-ply tube blanks have been developed.

An extrusion billet design used for the fabri-
cation of two-ply tube blanks of equal layer
thicknesses is shown in Fig. 3.3.5. The effect
of various alloy combinations on the layer thick-
nesses after extrusion was of particular interest,
since it is known that the thickness of metals with
higher resistance to plastic deformation will be
greater after hot rolling than was calculated.
Billets of the configuration shown in Flg. 3.3.5
were extruded at 2100°F to 1Y -in.-OD, /-ln.-wall
tube blanks. Tube blanks of several combmatlons
of alloys were sectioned longitudinally and the
layer thicknesses determined. The results, as
presented in Table 3,3.8, show that variations in
the layer thicknesses are independent of the

 

40n assignment from Pratt & Whitney Aircraft.

5J. H. Coobs et al., ANP Quar. Prog. Rep. Sept. 10,
1955, ORNL-1947, p 145.

6T. K. Roche and H. Inouye, ANP Quar. Prog. Rep.
Dec. 10, 1955, ORNL-2012, p 155.

UNCLASSIFIED
ORNL-LR-DWG 12903

/—DRILL Ya-in. HOLE Ya n. DEEP

—F

 

 

 

 

-n

-
\

g
2 &

 

3-in.DIA
1

——l 0875
DIA
IR 100

 

 

Lo
1t 23,

 

 

NOTE . ALL DIMENSIONS
ARE IN INCHES

 

 

 

NOSE AND INSIDE CYLINDER OF TYPE 36
STAINLESS STEEL

Fig. 3.3.5. Extrusion Billet Design for the Pro-
duction of Tube Blanks with Equal Layer Thick-

168

TABLE 3.3.8. LAYER THICKNESSES OF
COMPOSITE TUBE BLANKS

 

Thickness (.in.) '
Alloy ' :

Combination Outer Inner

Layer  Layer*

 

Incbnel-type‘3'l6 stainless steel _ ‘
Front 0,122 0,130

Middle - 0,130 0,122

End | 0,106 0.122

Average . 0,119 0,125
Hastelloy Batype 316 stainless steel :

Front . 0.110 0,142

Middle | _ 0,114 0,134

End - : 0,110 0,122

Average 0.111 0,133
Monel=type 316 stainless steel

Front | 0,126  0.130

Middle 0.118  0.134

End 0,138 0.134

Average 0.127 - 0,132
Nickel=type 316 stainless steel

Front e :

Middle 0,138  0.098

End S 0.122 0,126

Average 0.130 0.112

 

*Inner layer was type 316 stainless steel in all the
blanks,

combination of alloys, although the resistances
of the alloys to plastic deformation may vary
greatly. Accurate thickness measurements were
difficult to obtain because of the roughness of the
extruded blank.

Tube blanks of the combinations listed were re-
drawn to 0.187-in.-OD, 0.020-in.-wall tubing with
fair success. As a result of inexperience in proc-
essing duplex tubing, however, a few longitudinal
cracks were present in the Inconel layer of the
Inconel-type 316 stainless steel combination. Al-
so, the Hastelloy B-type 316 stainless steel com-.
bination was rejected during processing because
of longitudinal splitting. Preliminary examinations
of the small-diameter tubing obtained showed that
thermal bonds were maintained at the metal inter-
faces throughout the processing and that the layer
thicknesses were about equal,
 

A

-

NEUTRON SHIELD MATERIAL FOR
HIGH-TEMPERATURE USE

M. R. D’ Amore J. H. Coobs

The neutron shield designed for the ART, as
discussed previously,” incorporates a double layer
of boron-containing materials which have a totdl
boron density of 1.2 g/cm?, A stainless-steel-clad
copper-B ,C cermet layer 0.100 in. thick is fo be
used nearest to the neutron source to absorb radi-
ation damage, The second layer consists of B,C
ceramic tiles 0,265 in, thick contained in stainless
steel cans.

Ceramic B, C Tiles

The evaluation of the B,C ceramic tiles sub-
mitted by the Norton Company and by the Carborun-
dum Company was completed. Two sets of four
sample tiles each were received from the Carborun-
dum Company during the quarter., The tiles were
bonded with silicon and fabricated by a casting
ond sintering process. Radiographs of the first set
of four tiles revedled large voids, The macro-
porosity was eliminated in the second set of tiles
by a revision in the manufacturing process. The
boron densities of the eight tiles, as determined by
chemical andlysis, varied between 0,889 and 1.15
g/em3. The minimum boron density of 1,26 g/em?
specified for the tiles oppears to be difficult to
achieve in a cast and sintered B,C-SiC tile.

The Norton Company submitted tiles of both
technical-grade and high-purity B ,C that had been
hot pressed to densities of 1.9 to 2.34 g/em3. A
high-purity B,C tile with o density of 2.0 g/em’d
was found by andlysis to contain 1.48 g ot boron

per cubic cenhmeter, ‘rather than ‘the minimum of -
1.53 g/cm3 guaranteed by the Norton Company.
The Norton Company has guaranteed to produce
finished tiles of hlgh purity wnh a mmimum boron _
‘content of 1.7 g/em3. :

'Representative specnmens cut from sumple flles '
,'submlfled by each company were irradiated ‘in the
LITR for a six-week period. A cursory examination
of the irradioted specimens indicated that the
ability to withstand radiation damage is equal for
both - moterlals at less thon 3% burnup of the. B0
afoms. No cracking of the specimens or gas evolu- N
. tion was observed after irradiation. - The ~hot-

pressed B C hles are preferable for ART cpplxco-"‘

 

"H. inouye, ANP Quar Prog. Rep March 10, 1956
ORNL-2061, p 151.

PERIOD ENDING JUNE 10, 1956

tion, since the higher boron concentrations attain-
able by this fabrication method will result in a

‘more effective reduction in neutron activation of

the NaK heat exchanger circuit.

Copper-B 4C Cermets

The shield plates made from o stainless-steel-
clad copper-B,C cermet are to be 0.100 in. thick
and to have the following cross-sectional configu-
ration:

8 1o 10 mils of type 430 stainless steel
cladding

1 to 3 mils of copper diffusion barrier

80 mils of 16 vol % B,C particles dispersed
in copper

1 to 3 mils of copper diffusion barrier

8 to 10 mils of type 430 stainless steel
cladding :

The maximum plate sizeis to be about 8 in. square.
Bonding of the components by hot-roll cladding in
an evacuated picture frame has been moderately
successful. Bonding of the copper barrier material
to the type 430 stainless steel cladding is difficult
to achieve at moderate intermediate reductions.
Reductions of 20% per pass cause a bulging of the
B C-copper core at each end which ruptures the
cloddmg. A more promising method of manufac-
turing the clad cermet plates consists in bonding
the components by hot pressing at temperatures be-
tween 1800 and 1900°F and then cold rolling the
plates to the final thickness. The plate uniformity
can be controlled to close tolerances by this
method, and good bonding of the components is
achieved.
~'Small plates have been successfully fabricated,

: _"c‘:n,d'thre_ process is being scaled up for the manu-
~ facture of larger plates, Roé‘mflon-domage speci-
- mens of the clad cermet have been prepared and

Wl" be lrroduated in the LlTR and in fhe MTR.

Boride Partacle Dlspersions ina Meta“ic Matrix

Dlspers:ons of BN and CaB, ‘particles in iron

~and nickel were investigated os possnble substi-
tutes for copper-B4C cermets. Compacts containing
21 vol % CaB, dispersed in iron and 30 vol % BN
dispersed in mckel were successfully fabricated

by cold pressing, sintering, and | coining. the powder

_ mixtures, The coined compacts were encapsulated

in o type 304 stainless steel picture frame and
rolled at 2000°F to a total thickness reduction of -
85%. After hot rolling, the CaB,-Fe and BN-Ni

169

 
 

 

 

ANP PROJECT PROGRESS REPOR1

cores were examined by x-ray diffraction tech-
niques for evidence of reaction. A reaction had
occurred in the CaB-Fe material during the fabri-
cation process that resulted in the formation of
Fe,B. No reaction was detected between BN and
nickel, in confirmation of the compatibility test
results for BN and Inconel reported previously.®
Clad specimens of both BN-nickel and BN-Inconel
have been prepared for |rrad|cmon testing in the
LITR. _

The combincfions,CaBé-Ni ond BN-Fe were also
investigated. The CaB-Ni compacts formed a
liquid phase at 1800°F during sintering. No re-
action was observed between BN and iron. - Diffi-
culties in fobrlcathn were experienced, however,
and work .on the ‘BN-Fe combination was discon-
tinved in fovor of the BN-Ni composition. The

BN-Ni cermet cppears to be attractive as a high-

temperature shield material, since it is easily fab-
ricated and the components are compatible up to a

| _ temperature of 2000°F.

~ Boron Steels

The study of boron steels for use in a compres-
sion ring between the beryllium reflector and the
support strut ring in the ART was continued. The
tensile strength of a cast 0.75% B!%-Fe alloy ot
1300°F was 14,775 psi, with no elongation. Billets
of the dlloy were extruded into rod at 1900°F with
no difficulty by using an extrusion ratio of 6.5:1.
Tensile specimens will be machined from the
wrought material to determine the effect of hot
working on the mechanical properties of the dlloy.

The compatibility of boron-iron dlloys containing
1 and 3% boron and Inconel was investigated in
500-hr tests ot 1300°F. Diffusion couples were
prepared for the tests by hot rolling the composite
material at 1900°F. Metdllographic observation of
the Inconel and boron-iron dlloy interface revealed
that no diffusion of boron into the Inconel had
ocourred during the hot-rolling operation. Photo-
micrographs of the diffusion couple interface after
500 hr at 1300°F are shown in Fig. 3.3.6. The
layers 1 to 1.5 mils wide in the 1% B dlloy and 0.5
mil wide in the 3% B dlloy that were depleted in
boron after 500 hr ot 1300°F indicated that diffu-
sion of boron into the Inconel was not serious at
this temperature.

 

8. K. Roche and H. Inouye, ANP Quar. Prog. Rep.

o March 10, 1956, ORNL-2061, p 157, Table 6.7

170

Radiation-damage specimens were fabricated from
a boron-stainless steel alloy for testing in the
LITR and in the MTR., The composition of the
alloy is basically type 304 stainless steel with the
nickel content increased to about 14% and a boron
addition of 1,07%. The boron addition to the melt
contained 84.6% B10 and 15.4 B, The specimens
will be irradiated unstressed and stressed at 500
psi at temperatures from 1300 to 1700°F in the
LITR ond at 1600°F in the MTR. '

Recent work by other mveshgm‘ors9 has revea!ed
low ductility in boron—stainless steels after irra-
diation, The feasibility of using a duplex ring

- configuration is therefore being investigated. The

primary ring would be fabricated from Inconel ond
the secondary ring from clad copper-B,C,

Other Boron-Containing Materials-

The electrophorehc deposmon of unlform copper

~ coatings on ceramic B,C tiles and the bondmg ofa

copper-B ,C layer to type 430 stainless ‘steel is
being mveshgated by the Vitro Laboratories. Good
bonding of a deposited copper coahng has been
achieved on cast and sintered B ,C ceramic tiles.
The bonding of a copper coctmg to hot-pressed
B,C tiles has been moderately successful, Elec-
trophoretically deposited coatings containing up to
40 vol % of B,C particles dispersed in copper
shrank excesswely during the sintering treatment

at 1000°C.

"~ SOLID FUEL ELEMENTS:

M. R. D'Amore " J. H. Coobs
V. M. Kolba 10

The investigations of simulated seamless tubular
fuel element extrusions, reported previously,!!
were continued. The recovery of material with
vniformly thick layers varied from about 55% in a
tube extruded at a 5:1 ratio to a maximum of about
76% in tubes extruded at a 21:1 ratio. Sections 18

/in. in length from two tubes extruded ot a 21:1 rafio

were redrawn to 0.187-in.-0OD, 0.015-in.-wall tubing.
Preliminary metallographic examination of the fin-
ished tubing revealed o large number of tensile
failures in the 30 vol % Al O,—type 302B stainless

 

79J J. Lombardo, Tensile and Impact Test Results
on Irradiated Boron-Stainless Steel, WAPD—SFR-FE-192
(June 28, 1955).

mOn assignment from Glenn L. Martin Co.

114, H. Coobs and M. R. D’ Amore, ANP Quar. Prog.
Rep. March 10, 1956, ORNL-2061, p 161.
 

 

-

14

 

PERIOD ENDING JUNE 10, 1956

 

Fig. 3.3.6. Diffusion Couples of Boron-lton Alloy and Inconel after 500 hr ot 1300°F. (=) The B-Fe
alloy containing ]% B. (b) The B-Fe alloy containing 3% B. 500X. Reduced 7.5%. '

steel cores, probably because of excesswely high
reductions during redrawing.

Additional three-ply billets are being prepared to
investigate the extrusion of hot-pressed cores pre-

pared in sections rather than as a single cylinder, -
The billet designs have been. rewsed in an aitempt o
fo improve material recovery, = :
The Allegheny Ludium Steel Corporcmon hus ex-
truded two 5% -m.-dm, 12-in.-long, three-ply billets,
- The cermet cores were fractured severely in the'
7 extruded fubes. B :
“A “review of radmhon-domage studles on: solld[
-fuel elements is nearly complete,’ and a summary °
is, bemg prepared that is based on metallographic
,exammaflons -and hardness and - bend test: results.‘_ )
" reported by the Solid- Stcte DIVISIOI'I over the past.
"several years, The data were obtained from minia--

ture fuel elements and subassemblies submitted for
examination by Pratt & Whitney Aircraft, GE-ANPD,
KAPL, and the APPR group at ORNL, as well as
from full-sized MTR and LITR fuel elements,

TUBULAR CONTROL RODS

M. R. D’Amore 3. H. Coobs
R. E. McDonald '2

" The feuslblllfy study on the extrus:on of tubulor
control - rods, described prevnously, . was .con-
tinved. ‘An Inconel billet can wos fabricated, end
the co re was prepared by tamping a foose powder

" mixture of 369 vol % Lindsay Mix rare-earth-oxide

particles dispersed in nickel into the billet can,
Fracturing of the outer layer of this three-ply billet
during extrusion at 2100°F resulted in a nonuniform
core. The core was quite dense cfter extrusion and

~ appeared to be bonded to the inner und outer In-
‘ onel cladding.

 

200 assignment from Pratf & Whitney Aircfoh.

13), H. Coobs, R. E .McDonald, and M. R. D'Amore;
ANP3 Quar, Prog. Rep. March 10, 1956, ORNL-2061,
p 163.

171

 
 

 

 

ANP PROJECT PROGRESS REPORT

The as-received Lindsay Mix is composed of very
fine particles about lu in size. Dry mixing of the

Lindsay Mix and carbonyl nickel powders in an .
oblique blender resulted in segregation and lumping
" of the rare-earth oxide particles. A satisfactory .
dispersion of the oxide and carbonyl nickel pow-

ders could be obtained onfy by wet milling.
One section of a control rod core has been fabri-

cated by hot-pressing a cylinder of 36.9 vol %
Lindsay Mix=63.1 vol % carbonyl nickel to a den-

sity of 88.7% of theoretical. Additional extrusions
of Lindsay Mix~nickel cores clad with either

Hastelloy X or lnconel are plunned

Nb-UO FU EL EL EMENTS

CV. M. Kolbu H. Inouye
' J P. Page

Commercna!ly uvcllable niobium powder is of poor

- qudlity, and therefore attempts were made to pro-

duce high-purity powder from wrought niobium sheet

 

 

for use in fabrication studies on Nb-_UOé fue|'c0rh-;
ponents. The niobium sheet was hydrided, crushed,
ond ball milled to 100-mesh powder. This powder

was then leached to remove iron contamination

from the ball milling and then vacuum annealed at
800°C to remove hydrogen. A Bergsman microhard-

ness determination on this powder gave a result of
~317 VHN. Further vacuum ‘annealing at HOO"C n
lowered the hardness to ~290 VHN, which ‘was
still quite high and indicative of oxygen and nitro-

gen contamination,

A second attempt was made to produce powder by

. using cold-roll ed niobium plate and repeating the |
~hydriding, - crushing, and vacuum-annealing  treat-
‘ments. A polished section of particles of this pow-

der is shown in Fig. 3.3.7, on which the hardness

was found to be ~193 VHN.' Small inclusions.may

be seen in the matrix, which md:cafe some degree

of contamination, Samples of both barches of pow-
der are being analyzed for comparison _wuth the

 

 

 

 

Fig. 3. 3.7. Microstructure ond Hardness of Niobium Powder Produced from Wrought Plcte. 500)( Re-

duced 3.5%.

172

 
 

 

.)

L)

wrought plate used as base material. Further at-
tempts are also being made to secure high-purity
niobium powder from commercial sources.

Two companies were contacted for information on
high-vacuum furnaces capable of temperatures of
2100°C for sintering niobium and other materials.
Literature and quotations have been received and
are being evaluated. ~

Three specimens of Inconel-clad niobium have
been machined to slightly larger than creep speci-
men dimensions and are to be undercut and edge-
protected as described previously.'4 They will be
creep tested in an inert atmosphere. Both highly
purified argon and liquid sodium show promise as

creep-test environments for unclad niobium. Iden-

tical specimens will be run in each environment ot
1500°F, and the results will be compared. Data on
weight and hardness change will also be corre-
lated.

LOW-CONDUCTIVITY GAMMA-RAY
SHIELDING MATERIAL

JO Ho COObS Jo P.__Pdge

Conductivity data received from a commercial
supplier of cemented carbide bodies, correlated
with data obtained at ORNL, indicate the thermal
conductivity of tungsten carbide to be approxi-
mately 0,125 col/sec:°C.cm. This value, coupled
with the *‘inertness’’ on hot pressing, makes tung-
sten corbide the most promising base material -for
a low-conductivity gamma-ray shield,

Constantan was chosen as a binder for the tung-
sten carbide because of its low thermal conduc-

 

4, p. Page, H. Inouye, and V. Kolba, ANP Quar.
Prog. Rep. March 10, 0RNL-2061 p 161.

PERIOD ENDING JUNE 10, 1956

tivity (0.06 cal/sec.°C.cm), its low liquidus tem-
perature (1220°C), which minimizes hot-pressing
difficulties, its low cost, and its brazeability.
Further, the elemental copper and nickel powders
from which it is prepared are readily available.

Six tungsten carbide—constantan thermal-conduc-
tivity specimens, as described in Table 3.3.9, have
been hot pressed and finished to size. The vario-
tion of thermal conductivity with composition
should become evident upon evaluation of these
specimens, The composition showing the lowest
conductivity will be chosen for use as the gamma
shield around the ART pump shafts,

Three cold-pressed and sintered ZrO, thermal-
conductivity specimens, with densities of 3.09,
3.52, and 4.41 g/cm, respectively, were machined
to Battelle Memorial Institute specifications for
thermadl-conductivity determination at that installe-
tion, These pieces will be tested in a helium
atmosphere to simulate service conditions.

LITHIUMc.MAGNESIUM ALLOYS
R. E. McDonald C. F. Leitten, Jr. .
Work continued on the 20% Li—80% Mg alloy for

use as shielding material. Corrosion of the alloy
in air and in water was found to be severe. Freshly

“cleaned specimens gained 0.088 mg/cm,z‘during 6

hr of exposure in qir. Another specimen tested for
various times in boiling water was found to have

lost 0,292 mg/em? in the first 2 min of exposure,

which amounted to 90% of the weight lost. during
the entire 15-min test. The decrease in rate of
weight loss may be attributed to the depletion of
lithium ot the surface, since no protective film
could be detected. Pure magnesium tested in
boiling water for 15 min showed a weight loss of

TABLE 3.3.9. DA‘I‘A'ON'TUNGSTEN CARBIDE-CONSTANTAN THERMAL,—CO.NDUC:TWI"'I'Y,S_PECIMENS |

 

Composition (wt %)

| Composition (Vbl',%)_ - Porosnty -

 

 

 

: . (9/cm®) - WC “Constantan wC Constantan " * (%}
120 L83 57 43 42 . 6
127 e 7 29 53 ® -
24 Ne0 8 s 64 R 7
12-19 NS5 . 94 6 68 R 25
13-5 12.96 79 21 5 30 5
13.15 12.67 93 7 75 10 | 15

 

173

 
 

 

 

 

ANP PROJECT PROGRESS REPORT

0.09 mg/cm2. The weight losses of several freshly

cleaned specimens tested in water at various tem-
peratures for 15 min are presented below:

Temperature of Weight Loss
Water (°C) (mg/cmz)
30 0.243
50 0.271
80 0.309
100 0.326

Work was continued on the development of pro-
tective coatings and on cladding of the material, A
technique was developed for producing a bright
surface, which involves a two-step treatment with
concentrated phosphoric acid and acetone. The
bright surfoce tarnished slightly when exposed to
air, but it was stable at 300°C and for long periods
at lower temperatures. Specimens with surfaces
prepared by this method were roll clad with 2$ alu-
minum and AZ31 magnesium dlloy. The 25 alumi-
num-clad material produced a brittle bond, while

the AZ31 dlloy failed to bond even to itself.

“In order to further evaluate the mechanical prop-
erties of the 20% Li—80% Mg alloy, more sheet
specimens were needed. A 0.250-in, sheet was
mechanically cleaned of mill scale and corrosion
products, given the phosphoric acid-acetone treat-
ment, and hot rolled at 250°C to a thickness of
0.070 in. A 30-min sock at 300°C straightened the
sheet, New specimens were machined from this
sheet and kept under minerdl oil until tested. The
tensile data obtained for the new sheets are com-
pared with the data for the old sheets in Table
3.3.10 and the rupture data are compared in Table
3.3.11. Fixtures are being designed to test this
alloy af constant temperature, in a protective
medium of mineral oil, without the film produced by
phosphoric acid—acetone treatment.

Powder metallurgy was investigated as a dif-

. ferent approach to fabrication of the shield ma-

terial. Lithium oxide was chosen as the lithium
source, being the most stable of the lithium
compounds and possessing the highest melting

TABLE 3.3.10. TENSILE DATA FOR 20% Li~-80% Mg ALLOY SHEET

Density: 1.44 g/c:m3

 

 

) Average Average
Sheet Temperature Elongation Tensile Strength Yield Strength
Specimen (°F) (%) (psi) (psi)
Old Room 35 12,500 . 11,230
New Room 48 12,775 : 10,650
Old 200 45 3,000 2,970
New 200 94 3,200 2,940

 

TABLE 3.3.11. RUPTURE DATA FOR 20% Li-80% Mg ALLOY SHEET

All tests at room temperature

Stress: 3000 psi

 

 

Sheet ) Elongation Rupture Life
Specimen Coating (%) (hr)
Old None 48 125
Grease 90 315
HNO, dip and grease 100 220
New Phosphoric acid, acetone, and grease 32 1050
Grease 35 566

 

174
 

 

 

 

A

-

")

 

point. A mixture of 50-50 vol % Li,O-Al powder
was cold pressed and hot rolled in ¢ standard-size
2S-aluminum picture frame, Composite sheet pro-

duced by this method shows a continuous aluminum

matrix, even distribution of the lithium oxide, and
good bonding. The 50-50 vol % mixture gives 20
wt % lithium in the core ond a lithium density of
0.452 g/cm3, as compared with a lithium density of
0.288 g/cm® in the 20% Li-80% Mg alloy. "It has

been possible to produce composite sheets with

thickness ratios of clad to core of 1:2, 1:4, 1:6,

and 1:8.

FPERIOD ENDING JUNE 10, 1956

On the basis of the successful production of the
Li,0-Al composite sheet, Li,O-Mg was investi-
gated. This combination would produce a lighter
sheet and a higher weight percentage of lithium
for a given volume percentage. The picture frame
was made from AZ31 magnesium because M-1
magnesium alloy was not available. No solid
phase bonding with AZ31 could be obtained. The
Li,0-Mg compact and AZ31 interface was brittle,
and the finished sheet was cracked throughout.

175

 
 

 

B s e o e 48 R0

 

 

ANP PROJECT PROGRESS REPORT

3.4. WELDING AND BRAZING INVESTIGATIONS

P. Patriarca

PUMP FABRICATION EXPERIMENTS
P. Patriarca G. M. Slaughter
Welding Studies

The dimensional changes associated with the
welding of Inconel pump volutes representative of
those designed for the main NaK pumps (PK-2)
of the ART were determined in a preliminary
investigation.! These changes were considered
to be in excess of those cllowable. Accordingly
the welding procedure was modified and the test
was repeated in the interest of improving the
ability to more nearly conform to acceptable
tolerances.

Two test pieces essentially similar to those
used previously were machined from 2-in. Inconel
plate. A shrinkage of 0.100 in. was provided for
in the fabrication of these pieces. Four spacers
were also machined from Inconel plate to act as
rigid supports during welding of the volutes after

 

p, Patriarca, ANP Quar. Prog. Rep. March 10, 1956,
ORNL-2061, p 142.

the initial shrinkage occurred. These are shown'

in Fig. 3.4.], along with the component parts of
the test, *~ The spacers were subjected to an
alumlnlzlng ‘treatment prior to assembly of the

components in order to prevent self-welding during

the subsequent operations. The two volutes were
welded in accordance with the welding procedure

‘described in Fig. 3.4.2.

Micrometer measurements were made at four
radial sections, as described in Fig. 3.4.3, prior
to and after each operation. The results of these
micrometer measurements are summarized in Table
3.4.1. It may be noted that the shrinkage de-
creased with increased dlstonce from the weld,
an effect noted previously.! It appears that an
improvement of the shrinkage prediction at position
3 by reducing the shrinkage allowance can be
accomplished only at a sacrifice in the absolute
shrinkage at positions 1 and 2, the over-all error
being of the order of 0.012 in.

A compromise would be to provide for a 0,090-in.
shrinkage allowance and thereby reduce the ab-
solute error to 0,006-in. The ability to reproduce

UNCLASSIFIED
Y-17332

SPACER

 

BOTTOM

TOP

PK-2 WELD TEST

Fig. 3.4.1. Component Parts for NaK (PK-2) Pump Velute Weld Shrinkagé Test.

176

 
 

 

 

w

-

-

¥

-/

PASS NUMBER PROCESS
' INERT ARC
2-3 METAL ARC
4-8 METAL ARC

* ROOT PASS:

ELECTRODE SIZE (in.)

6 TACKS EACH APPROXIMATELY 2in. INLENGTH ON APPROXIMATELY 6in. CENTERS

60°

PERIOD ENDING JUNE 10, 1956

ot — '/iG in.

e L

 

WELDING PROCEDURE; VERTICAL FIXED, 2G

332
32
Sty

ELECTRODE MATERIAL

INCO 62
INCO t32
INCO 432

FOLLOWED BY TIE -INS BETWEEN TACKS.

CURRENT {amp}
s
60
Ho

UNCLASSIFIED
ORNL-LR-DWG {4749

APPROXIMATE

WELDING SPEED (in./min}

*
2
4

| Fig. 3.4.2. Welding Procedure and Joint Design fof Weld Shrinkage Tesi No. 2 en NaK (PK:2) Pump

 

 

 

Volute.
TABLE 3.4.1. RESULTS OF MICROMETER MEASUREMENTS BEFORE AND AFTER WELDING
OF NuoK (PK-2) PUMP YOLUTE
Dir.nensions' (in.)
Position Before After Chan Deviation from Deviation from . " After A}:Iditional.
Welding - Welding ange Desired Change* Revised Change** = Annealing Churig_e ‘
Al 3916 3.820 0096 - —0.004 +0006 3.818 ~0.002
Bl  3.920 3.824  0.096 - —0.004 ~+0,006 ©3.820 =0.004
c1. . 3920 3.826 . 0.094 - —-0.006 +0.004 0 3.823 -0.003
D1 3916 . 3.822 - 0.094 . = -0.006 - +0.004 '3.818  © —0.004
A2 3914 3.826  0.088 o =-0.012 =0.002 3.822  —0.004
B2 3919 3.830  ° 0.089 - =0.011 —0.001 . 3.826. =0.004
€2 3919 3831 0.088 - =0.012 ~ =0.002 3.828 . - .- —0.003
D2 . ..3.916.  3.829  0.087 . —0.013 - =0.003 3,825 —0.004
A3 0670 0585 0,085 - =0.015 ~ =0.005 - 0581 - —0.004
B3 - 0672 0588 - 0.084 ~0.016 ~0.006 0.585 ~-0.003
€3 . 0.675 058  0.086 —0.014 ~0.004 0.586 -0.003
- D3 0.670 0586 - =0.016 ~0.006 .0,583 -=0,003

0.084

 

®*Shrinkage allowance: 0,100-in, -

**Revised shrinka_ge allowance: 0.090-in,

177

 
 

 

 

ANP PROJECT PROGRESS REPORT

the shrinkage data described herein should be
determined by conducting another test prior to
attempting fabrication of a pump with a complex
volute,

Since the shrinkage at position 3 was less than
expected, it was not possible to use the spacers,
as machined, to prevent further shrinkage during
annealing. The pump casing was annealed at
1850°F for '4 hr in order to determine the necessity
for this operation. The additional shrinkage of
0.002 to'0.004 in. observed indicates that the role
of residual stresses is significant, Properly
machined spacers should be used to prevent further
shrinkage during annealing, as was accomplished
in the previous investigation.!

Brazing Studies

Tests have also been performed to study the
feasibility of a brazed NaK (PK-2) pump volute.
The component parts for the test are shown in
Fig. 3.4.4. The bottom volute is shown with eight
0.187-in.-dia, l-in.-long Inconel shear pins in-
serted into 0.197-in.-dia holes drilled to a depth
of '/2 in, These pins were intended to provide
approximately 5 in.2 of shear area to supplement
the 18 in.2 of area that constitutes the faying
surfaces of the upper and lower volutes in the
brazed assembly, The top volute is shown with
0.200-in.-ID tubes welded over the shear pin holes
to act as reservoirs for the Coast Metals No. 52
brazing alloy. The alloy was applied in the form
of 3.16-in.~dia slugs, which ranged from 3.8 to
'.5 in. in length and were fabricated by casting
in graphite molds. Several of these slugs are
shown in the center of Fig. 3.4.4. A total of
30.8 g of Coast Metals No. 52 alloy was used,
This amount was calculated to supply approxi-
mately 0.19 in of alloy, of which 0.048 in.?
would fill the space between the shear pins and
the shear pin holes and approximately 0.054 in.3
would comprise a 0.0003-in.-thick brazing alloy
layer between the faying surfaces of the volutes.
It was expected that a surplus of brazing alloy
would be desirable to ensure complete brazing.
The excess alloy was to remain in the reservoirs.
A 0.003-in. separation between the parts of the
volutes was achieved by using 0.003-in.-dia nickel
wire as a spacer,

Micrometer measurements were made before and
after brazing at the positions described in Fig.
3.4.5. The assembly was brazed at 1050°C in
dry hydrogen., The heating and cooling rate of
the brazing -cycle was approximately 350°C/hr,

178

and a hold time of !5 hr at 1050°C was used. A
photograph of the completed assembly is shown
in Fig. 3.4.6. As may be seen, excess brazing
alloy flowed out of the capillary and down the
side of the casing. A similor condition was
evident on the .inside of the volute; there was a
fillet of brazing alloy thinly distributed around
the bottom volute,

The failure of the reservoirs to retain the excess
brazing alloy is atiributed to the relatively large
clearance between the Inconel shear pins and the
shear pin holes. It is expected that a clearance
of 0.001 to 0.002 in., coupled with a supplementary
capillary on the periphery of the casing, which
could be removed later, would remedy this dif-
ficulty.

The braze joint was subjected to dye-penetrant
and radiographic inspections and was considered
to be sound. The degree of correlation between
the radiographic interpretation and the joint sound-
ness should be determined by metallographic
examination, ’ |

The results of the micrometer measurements are
summarized in Table 3.4.2. As may be noted,
most of the dimensional changes found are within
the limit of error of the measurements. As was
expected, the data indicate that the brazed con-
struction affords a high degree of dimensional
control and should be given consideration as g
fabrication method. '

TABLE 3.4.2. RESULTS OF MICROMETER
MEASUREMENTS BEFORE AND AFTER BRAZING
OF NaK (PK-2) PUMF YOLUTE

 

Dimensions* (in.)

 

 

Position

As Assembled  After Brazing Change
A1 3.815 3.813 ~0.002
2 3.818 3.817 -=0.001
3 0.629 0.628 -0.001

2 3.816 3.816 0
3 0.629 0.627 ~0.002
C1 3.819 3.820 +0.001

2 3.818 3.818 0
3 0.629 0.628 - =0.001
2 3.816 3.815 -0.001
3 0.629 0.628 ~0.001

 

*Estimated accuracy of measurement £0,001 in.

Y,

 
 

 

PERIOD ENDING JUNE 10, 1956

u UNCLASSIFIED

L ORNL-LR-DWG 14720 =  UNCLASSIFIED
) ¥.17519

 

 

 

 

 

 

 

 

 

 

 

 

 
 
 
 
 
  

 

 

 

f_i ""—sla‘fl.
--—'/all'l.
* 1
. 1
62 T SPACER 0.100-in.
\/ SHRINKAGE
] ALLLOWANCE
3
1 Fig. 3.4.4, Component Parts for Brazing Test on
5 1 S - NaK (PK-2) Pump VYolute, Note the Inconel sheor
) pins on the bottom test piece, the Coast Metals
Fig. 3.4.3. Details of Micrometer Measure- No. 52 alloy slugs in the center, and the Inconel
ments on Welded NaK (PK-2) Pump Yolute, tubing reservoirs on the top test piece.
Measurements made at four radial sections
(A through D) at 90-deg intervals at positions
1, 2, and 3 (see Table 3.4.1).
UNCLASSIFIED
_ ORNL—LR—DWG 14721 UNCLASSIFIED
s 1 2 e
INCONEL TUBING
RESERVOIR
COAST METALS
NO. 52
. BRAZING
ALLOY SLUGS
b—1, § .
> INCONEL. PIN- %t
-

 

 

 

I — . - PK-2 PUMP"BRAZE TEST

: Fig- 3.4.5.- Detmls of Mlcrometer Mecsurements . B F'g' 3 4_'6' _N"K (PK-Z) P‘"“P Volufe Afle'
~ on Brazed NaK (FK-Z) Pump Yolute, Measurements' RS ruzmg. ‘
‘made “at four ‘radial “sections (A throud1 D)ot - = ‘

- 90-deg - mtervals at poslhons 1, 2, and 3 (see
: V'Tnble 3. 4.2) '

179

 
081

 

 

TABLE 3.4,3. RESULTS OF MICROMETER MEASUREMENTS ON DIAMETER OF PUMP BARREL

 

Dimensions (in.)

 

Before

 Additional

_Change Due to

 

Position ofo After Welding hange After Rem?vol Additional :Affelz_l B Retiof of " Total
Welding Shell Flange of Restraints Changes Anfteqlmg Changes - .Residua. Stresses Chonges
A 6.378 6.383 +0.005 6,387 +0.004 6.384 - —0.003 +0.001 +0.006 -
2 6.390 6.391 +0.001 6.392 +0,001 6.393 +0.001 +0.002 +0.003
3 6.378 6.375 ~0.003 6.376 +0.001 6.377 +0.001 +0.002 ~0.001
4 6.378 6.380 +0.002 6.382 +0,002 6.382 0 +0.002 +0,004
B 1 6.384 6.389 +0.005 6.391 +0,002 6.388 ~0.003 ~0.001 +0,004
2 6.390 6.390 0 6.392 +0.002 6.391 ~0,001 +0.,001 © +0,001

3 6.381 6.376 ~0.005 6.380 +0.004 6.381 +0.001 +0,005 0
4 6.382 6.383 +0,001 6,387 +0,004 6.386 ~0.,001 +0,003 +0.004
c1 6.385 6.391 +0.,006 6.392 +0.001 6.391 —0.001 0 +0.006

2 6.386 6.386 0 6.387 +0.001 6.386 ~0.001 0 0
3 6.380 6.375 ~0.005 6.379 +0,004 6.379 0 +0.004 ~0.001
4 6,382 6.385 +0.003 6.385 0 6,385 0 0 +0.003
D1 6.396 6.393 -0.003 6,401 +0.008 6.401 0 +0.008 +0.005
2 6.383 6.382 -0.001 6.384 +0.002 6.385 +0.001 +0.003 +0.002

3 6.380 6.377 -0.001 6.376 ~0.001 6.378 +0.002 +0.001 0
4 6,382 6.381 +0.001 6.383 +0.002 6.385 +0.,002 +0,004 +0.005
E1 6388 6.392 +0,004 6.392 0 6,391 ~0,001 ~0.001 +0.003
2 6387 6.380 ~0.007 6.383 +0.003 6.383 0 +0.003 ~0.004
3 6.380 6.373 ~0.007 6.378 +0.005 6.377 —0.001 +0.004 © ~0.003
4. 6381 6.382 +0.001 6.382 0 o 0 ~+0.001

6.382

 

LI0dIN SSTJ903d LIDIT0¥d ANV

 
 

 

o3

4

‘,1'

FABRICATION OF JOINTS BETWEEN PUMP
BARRELS AND THE PRESSURE SHELL
OF THE ART

P. Patriarca

The ability to weld an accurately machined and
properly stress-relieved pump barrel to ‘o thick-
walled pressure shell without the need for finish
machining or subsequent stress relieving would
simplify considerably the assembly of the *‘north
head” of the ART. A design was suggested to

- permit such a procedure, and a test was conducted

to evaluate the feasibility of the desngn. :

The component ports of the joint used for the
test are shown in Fig. 3.4.7. The,’lmer flange

'is shown attached to the pump barrel sleeve,
which in turn was welded to the pump barrel. The

finer flange is also shown attached to a carbon-
steel pipe, which was in tumn welded to_cZ‘é-in.—
thick carbon-steel base plate.. This auxiliary

weldment was intended to provide a degree of -

restraint comparable to that to be expected from
the north-head expansion tank. '

It may be noted that the shell flange is quite :

large, the intent being to minimize distortion of
the flange and thereby require a realistic pro-
portion of the distortion during welding to occur
in the barrel sleeve and hence in the barrel itself.
The restraint provided by the auxiliary weldment
shown in Fig. 3.4.8 was intended to supplement
the afore-mentioned ,condmen and, hence, to more
nedrly simulate the north-head pressure shelf.

The Shell'flange was welded into the pump-barrel . ..~ -
sleeve by using the weldmg procedure described -~ -
in Fig. 3.4.9. " The completed test weldment is
~shown in Flg. 3.4. 10. - Diametrical changes in the

© pump’ barrel ‘were determined by micrometer meass .

_ urements at the posmons described in Fig.34.11. -
.= The restraints were then ‘removed by cutting the
carbon _steel 'with -an " oxyacetylene “torch, and . -
~~the barrel cssembly .was subjected to a 6-hr soakf,_

“-at 1500°F prior to further micrometer measurements
“ on the barrel diameter, : The results of these . =~ -
_,7;,'1‘";d|ometr|ca! measurements are surnmarlzed in Tubleii
" °3.4.3. It may be noted that the maximum chonges;"—
. are -remarkably’ small considerlng the exient ‘of

weldlng involved.

PERIOD ENDING JUNE 10, 1956

It is interesting to note that at least partial
relief of residual stresses was accomplished by

__cutting the barrel assembly from the restraints
and soaking at 1500°F. This procedure resulted

in additional diametrical changes, which indicate

that some distortion will occur in the north head

in service that may be unacceptable,
~ Axial changes were also determined by mi-

| crometer measurements at the positions described

in Fig. 3.4.12. The results of the measurements

‘are summarized in Table 3.4.4. It may be noted

that an oxial shift occurred that was significant
in magnitude but remarkably small for the amount
of welding involved. These results indicate that

~ some distortion is inevitable and must be either
- accepted or removed by subsequent machining and
- that a stress relief anneal will be necessary to

remove effects of residual stresses during oper-
ation, |

" TABLE 3.4.4., RESULTS OF MICROMETER
'MEASUREMENTS ON AXIS OF PUMP BARREL

 

Dimensions (in.)

 

 

Position Before After
Change
Welding Welding
Al 2.459 2.447 -0.012
2 2.459 2.468 +0.009
3 2,444 2.468 +0.024
4 2.437 2.431 —0.006
B 1 2.492 2.480 -0.012
2 2.466 2.470 +0.004
3 2416 2.433 +0.017
4 2437 . 2429 - -0.008
.c1 2.524  -2.516 =0.008
""""" o2 2.473 . 2.474 - +0.00
-1 2,380 2394 +0.014
L4 2437 2431 - -0.006
D1 2563 2,561 —=0.002
2 2.475 2573 -0.002
'3 2,351 . 2361 . +0.010
4 2440 . 2,437 -0.003
E1 2594 2594 . 0 _
2 . 2484 2,478 -0.006 - -
3 2318 - 2,321 +0.003
4 2.442 2438 -0.004

 

181

 
 

 

81

o UNCLASSIFTED
Y.17244

UNCLASSIFIED
PHOTO 17242

 

PUMP BARREL’

AT
£!

cpesme

Fig. 3.4.8. Test Setup Showing Auxiliory Weldment,

 

 

 

Fig. .3.4..7.&' Component. Parts for Test Fabrication of a-
Joint Between a Pump Barrel and the Pressure Shell of the ART,

LI0dIY 5S3JI0Ud LD3F0Ud NV

 
 

 

 

‘PERIOD ENDING JUNE 10, 1956

UNCLASSIFIED
ORNL~LR-DWG 14722

 

 

 

 

 

 

ELECTRODE MATERIAL

" INCO 62
INCO 62
© INCO 132
INCO 132

=

CURRENT (amp)

130
180
120
{20

‘.§
o
-
¥
WELDING PROCEDURE
PASS NUMBER - PROCESS - ELECTRODE SIZE {in.)
o " INERT ARC - Y,
‘ .2 " INERT ARC ¥,
< © 3= . METAL ARC (NOTE 2) - 42
' . f1—24 ' METAL ARC (NOTE 3} T
.NOTES : , o _ ‘
1, VARIED FROM 45 TO 55 deg. DEPENDING ON POSITION AROUND PERIPHERY, BECAUSE A
* STRAIGHT BEVEL INCLINED 10 deg TO THE HORIZONTAL WAS MACHINED ON THE FLANGE.
2. BUILDUP PASSES INTENDED TO SIMULATE J BEVEL AND MINIMIZE SHRINKAGE.
. ~ - 3. FILLER PASSES IN BUILT UP J BEVEL. | |

| Fig. 3.'4".97.: “Welding 'Prééeduré and Joint Design for Wélding fh-e'l.'-‘ressure_ Shell"Flcngé.'im'ofhe Pump-

o

‘Barrel Sleeve.

183

 
 

 

 

ANP PROJECT PROGRESS REPORT

 

 

UNCLASSIFIED
PHOTO 17243

Fig. 3.4.10. Completed Test Weldment for Joining Pressure Shell Flange to Pump Barrel Sleeve.

INVESTIGATION OF SHRINKAGE OF INCONEL
CORE SHELL WELDS

A. E. Goldman

A series of tests are being carried out to de-
termine the weld shrinkage to be expected during
fabrication of the Inconel core shell welds.
Comparisons of the weld shrinkage formulas given
in the literature with preliminary results obtained
at ORNL revealed wide discrepancies. It was
decided that actual experiments that would yield
empirical data would be required. A program of
welding Inconel plates under controlled conditions
and close observation was therefore carried out.
Based upon the results of these tests, the welding
of large Inconel hoops was begun, These tests
were performed in a manner that as nearly as

P. Patriarca

184

possible duplicated the fabrication problems and
restrictions of the actual core shells.

Each test of the initial progrum consisted i
the inert«arc welding of two /-m. Inconel plates,
each 6 x 20 in., in accordance wuth the established
procedure specifications (PS-1). A 50-deg beve!
with a Y% _-in. land was machined on one long edge
of each pfafe. A total of nine tests was performed.

Each pair of plates was assembled as shown
in Fig. 3.4.13. The root gap was fixed at 1/ in,
by using four l/s-m tool-steel spacers. The plctes
were held ogainst a flat horizontal plate by means
of C-clomps. Two large clamps were used to draw

the plates tightly against the tool-steel spacers. |

The edges of the plates were securely taped to
prevent air leakage into the gap, since only the
 

 

 

A

L

"

PERIOD ENDING JUNE 10, 1956

UNCLASSIFIED
ORNL-LR—-DWG t4723

 

  

SHELL FLANGE

 

SLEEVE

PUMP BARREL

 

 

 

 

 

 

PUMP BARREL -

OO

SHELL FLANGE

 

 

 

SLEEVE ~—

 

 

T 7T

 

   
 

 

 

 

 

AL AR ANNN

 

 

e |
ysilfl-J | B

SIS

 

 

 

- N

Fig. 3.4.".' rbetails of Micrometer Meusuremefits on .Diulfnefer of Pump Burrel (s?e Table 3.4.3).

185

 
 

 

 

ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
ORNL-—LR~-DWG 14724

 

—

PUMP BARREL

—

SHELL FLANGE
i

 

 

 

 

—

SECTION AA
. . i

 

Yy in.-}

 

13’3 in.

 

SHELL FLANGE

 

 

 

SNUONMOUONNNNNNNONNNNANN

 

 

 

 

 

i

 

 

 

  
  

 

 

=
l r'—POST §
B g""——PUMP BARREL

N
o
. \
\E
iR

N ,

\

N

%

N

\

N

AN

 

 

 

 

 

 

 

 

 

186

 

- Fig'; 3.4.12. Details of Micrometer Measurements on Axis of Pump Barrel (see Table 3.4.4).

 

 

7

O

 
 

 

 

UNCLASSIFIED
ORNL-LR-DWG 1472%

/4 x 6 x 20-in. INCONEL PLATES

2 Y/g-in. TOOL STEEL SPAGERS

(4 USED)
100° INCLUDED ANGLE

   

 

2
-
' Fig. 3.4.13. Joint Design for Shrinkage Test of
Inconel Plate Welds.
UNCLASSIFIED
. ORNL-LR-DWG 14726
® BASE PLATE
%‘% ROLLER
\ ¢ HOLD-DOWN PLATE
/]
® — Y~in. TEST PLATES .
Ny . PASS SEQUENCE‘V

74— HOLD ~DOWN PLATE .

| roLier

3

———TACK WELD -

 

 

 

0

 

 

- SIDE VIEW OF VERTICAL WELDING JIG -

.

into Vertical Welding Jig.

-

‘were again fastened to the flat base,

-F‘ig- 3.4.14, '-.As'seinbly.of Shrinkage Test Plates

PERIOD ENDING JUNE 10, 1956

torch gas was used to supply backup gas and weld
coverage.

A 14-in.-|ong tack weld was placed at each end

~of the root gap, and two more lé,--in.-long tack

welds were equally placed alorig the root gap.
The clamps were removed after the tack welds
were made, and the spacers were driven out,
Shrinkage measurements were taken, and the plates
The root
pass was applied affer the tack welds had been
wire brushed and the edges had been feathered.
Measurements of the root pass shrinkage were
then taken.

The plates were then assembled in the vertical
welding jig as shown in Fig. 3.4.14. The re-
maining five weld passes per plate were deposited
in accordance with the sequence shown. The
completed weldment is shown in the vertical jig
in Fig. 3.4.15. Dial-gage readings were taken
ot l-min- intervals during the deposition of the

UNCLASSIFIED
PHOTO 17245

 

Fig. 3.4.15. Completed Weld in Vertical Welding
Jigo )

187

 
 

 

 

 

AN’P' PROJECT PROGRESS REPORT

final passes, and micrometer measurements were
taken after the final poss had been made. The
shrinkage measurements and welding data are
presented in Table 3.4.5; the results of tests 106,
107, 108, and 109 were in close agreement,
aithough two different welding operators made the
welds. : :

The welding procedure used for these tests of
plates was then used for test welds of hoops.
For these tests ¥ -in. Inconel plates, 6 x 138 in.,
were bent into Lbops approximately 44 in, in
diometer. One edge of each hoop had a 50-deg
bevel and a ¥ .-in. land. The hoops were placed
on the weld-positioner bed in a horizontal plane
-in the manner shown in Fig. 3.4.16. Tool-steel
spacers, 4 in. long and ¥ in. thick,- shown in
Fig. 3.4.17, were placed between the beveled
edges at 6-in. intervals to maintain the root gap.
Large C-clamps were used to draw the two hoops
tightly against the spacers, Smaller C-clamps
were halved and tacked to the bottom hoops to
aid in the alignment of the upper hoop. Asbestos
string was used to seal the gaps between the

~ spacers prior to tacking. The area behind the
root gap was sealed with a cover formed from

0.010-in. annealed brass sheet and masking tape,
and the enclosed space was purged for 30 min
prior to tacking, Alignment was checked con-
stantly as the tack welds were placed between
the spacers. After removal of the tool-steel
spacers, the root pass was deposited. No dressing
of the land or feathering of the tack welds was
permitted prior to the root-pass deposition.

After completion of the root pass, the weld was
wire brushed and the five final weld passes were
deposited by using the sequence described in
Fig. 3.4.14. For these welds, two welders worked
180-deg apart around the hoop while the positioner
was slowly rotated. As shown in Fig. 3.4.18,
four dial gages were used to record the shrinkages
of the final passes. The welders worked on a
15-min-work, 5-min-rest cycle. A typical plot of
dial-gage readings is shown in Fig. 3.4,19. The
micrometer and dial-gage measurements obtained
are summarized in Table 3.4.6. '

TABLE 3.4.5; WELDING CONDITIONS AND RESULTS OF SHRINKAGE MEASUREMENTS
ON WELDS OF INCONEL PLATE

 

Total Shrinkage (in.) Dial-Gage Shrinkage* (in.)

 

Maximum Minimum Average Maximum Minimum Average

 

' Time Welding Current
T Wel
est Welder o equired Rod Used  Used
No. No.
{min) - (in.) (amp)
100 1 72 - 2135 70 (root)
-' o 105-110
101 1 66 23 70 (root)
- ' 105-110
102%* 1. 74 180 80 (root)
105-110
106 1 61 198 80 (root)
105-110
107 - 1 60 226 75-80 (root)
) | 105-110
108 2 79 231 80 (root)
L 105-110
109 2 64 236 80 (root)
105-110

0.197  0.123 0,126  0.058  0.048  0.053
0.167  0.130  0.146  0.070  0.050  0.060
0.142 0114 0130  0.081  0.079  0.080
0.157 0121 0141 0.092 0,082  0.087
0.148  0.132 0139  0.104 0.096 0.100
0157  0.106  0.138 0103 0.9  0.098

0.154  0.104  0.136  0.102  0.092  0.097

 

*Measurements made on last five of the six passes;dial-gage readlngs on tests 100 and '|01 are erroneous because

the dial-gage actuating arm slipped durmg the test,

**Test 102 plates were machined with an 80-deg included angle rather than a 100-deg included ungle.

188
 

 

-

»

*y

UNCLASSIFIED
PHOTO 17248

 

Fig. 3.4.16. Inconel Hoops on Weld Positioner
Bed.

Within the limits of this mveshgohon, the
following conclusions can be drawn. -
1. The results of the last four tests on plates

indicated that, under controlled conditions, the
effect of the welding variables could be minimized .
so that a welding. operator could, essentially,
' jduphccte his performance from test to test when
welding manually. Also, -for. a given set of con-
~ ditions, o welder could nearly. duphcate another
- '~:.welder s performance. ‘
12, The use -of various lengths of tool-steel .
. spacers m ‘rhe fcck weldmg of the plutes caused
“‘wide . variations - the resulting - tack-weld
~~shrinkage. -
" was found to be mversely propomonol to the '
~ length | of the spacers.. B
73, The results of the tests on the hoops '“‘_,-'
dicated that the inevitable variations in conditions

“The. qmount of tack-weld shrinkage

and techniques while the operators were pro-
gressing around the circumference caused greater

PERIOD ENDING JUNE 10, 1956

UHCLASSIFIED
¥-18648

 

Fig. 3.4.17. Tool-Steel Spacers Used to Main-
tain Root Gap While Welding Inconel Hoops.

UNCLASSIFIED
PHOTO 17247

 

Fig. 3.4.18. Hoop Welding Assembly Showing
‘Dial Gages Used for Measuring Shrinkage.

- variations in shrinkage within a hoop ihan the
~variations observed from hoop to hoop.

-4, The over-dll results indicate that for '/-m.
Inconel plafe, _inert-arc welded by two welding
operators under the conditions utilized for these
tests, the transverse shrinkage to be expected
will be from 0.111 to 0.138 in., with an average

189

 
 

 

ANP PROJECT PROGRESS REPORT

0.096

DIAL GAGE A
0.080 = - —— DIAL GAGE B
~—— — DIAL GAGE C
~ese-secs- DIAL GAGE D

0.064

0.048

SHRINKAGE (in.)

0.023

0.016

0 20 40 60 80 100 120 -

UNGLASSIFIED
ORNL-LR~DWG 14727

 

140 160 180 200 220 240 260 280 300 .

" WELDING TIME (min}

Fig. 3.4.19. Typical Plot of Dial-Gage Measurements of Slmnkuge Durmg Fmal Fwe Passes of 1he

Welding of Inconel Hoops.

TABLE 3.4.6. WELDING CONDITIONS AND RESULTS OF SHRINKAGE MEASUREMENTS
ON WELDS OF INCONEL HOOPS

 

 

 

Te;t Arc Time Rod Used Current Total Shrinkage (in.) Dial-Gage Shrinkage* (in.)
" Required - Y. Used
No. Amin) (in.) {amp) Maximum Minimum Average Maximum Minimum  Average
1 314 1218 80 (root) 0.126 0.111 0.1194 0.095 0.080 0.088
o 110-120
-2 310 1188 80 (root) 0.138 0.115 0.1261 . 0.102 0.088 0.095

110-120

 

 

*Measure ments ‘made on -logt five of the six passes; tack shrinkage and root shrinkage not included.

\}aide being 0,120 in. The longitudinal shrinkage
to be expected will be 0.250 to 0.375 in. for a
circumferential length of 138 in.

As previously mentioned, the values for the
transverse and longitudinal shrinkage for the 44-in.
hoops do not correspend to the values obtained
by calculation from the formulas given in the
literature.  Only through the accumulation of
empirical data from actual experience can pre-
dictions be made for future weld shrinkages: Since
each variation in the thickness of the Inconel
plate to be welded will create new problems
solvable only by more actual test results, ad-

190

ditional tests will be conducted on each of the
plate thicknesses of interest. . -

EXAMINATION OF NaK-TO-AIR RADIATOR
PWA NO. 2 AFTER SERVICE

R. J. Gray P. Patriarca

A 500-kw high-conductivity-fin radiator, desig-
nated PWA HCF radiator No. 2, failed on December
23, 1955, as the result of a leak. - This radiator
had been operating in a test rig for a period of -
1199 hr in the temperature range 1000 to 1600°F.
For 546 hr of the operating period a temperature
differential was imposed on the NaK flowing
 

 

 

"

o

o

through the radiater by passing cold air across
the fin surfaces. The radiator is shown in Fig.
3.4.20 as it appeared when recelved from 1he test
site,

The entire radiator was leak checked by pressur-
izing it under water and observing it to locate the
origin of air bubbles. This procedure revealed
the point of failure, which is indicated by the
arrow in Fig. 3.4.20. The radiator was then
sectioned for further examination, as shown in

UNCLASSIFIED
Y-17518

  

p‘”’-A

 

Fig. 3.4.20. NaK-to-Air Radiator PWA No., 2

That Failed in Service. Arrow mdlcotes point of
failure,

UNCLASSIFIED
Y. 17586

 

Fig. 3.4.21. NaK-to-Air Radiater PWA No, 2
After Sectioning for Metallographic Examination.

PERIOD ENDING JUNE 10, 1956

Fig. 3.4.21. The side portions of the support
members were removed by using a rubber-bonded
masonry wheel in a portable, electric handsaw
adjusted . for a shallow cut. Each bank of ‘fins
was separated by slicing the support members and
the bottom flanged plate, as shown, with a fine-
toothed, high-speed, steel hacksaw blade in a
portable electric drill equipped with a portable
power-saw attachment. The failed area was then
carefully removed for metallographic preparation
on a standard wet-cutoff machine eqmpped with
an abrasive wheel.

The area of the failure was again pressurized
under water to locate the exact position of the
leak before additional preparation for metallo-
graphic examination was undertaken. The failure
was found to exist in the comer tube on the
periphery which faced the side support member,
The support member was subsequently removed
for unobstructed observation of the emergence of
the water bubbles. The point of failure may be
seen in Fig. 3.4,22. ' |

A UNCLASSIFIED
‘ Y-$7795

s

TUBE SHOWING
i-FIRE DAMAGE
1 TO FINS

 

 

 

im0

i O

NaK-to-Air Radiator PWA No. 2 as Viewed from
the Support Member Side of the Radictor,

"

 

191

 
 

 

 

 

 

 

ANP PROJECT PROGRESS REPORT

The tube that failed and two adjacent tubes
were mounted intact and carefully ground to permit
examinations of their longitudinal cross sections
as seen against the direction of air flow. The
tube that failed is shown in Fig. 3.4.23. The
neckdown of the tube wall indicates ‘a tensile
fracture similar to that observed? in York radiator
Neo. 1.

Longitudinal cross sechons of the two ad|acent
tubes that were examined are shown in Figs.

3.4.24 ond 3.4.25. Incipient fractures may be seen -

in both these tubes. Three tubes were also taken
from corresponding positions on the air exit face

and prepared for examination in a similar manner.
Only one tube. ‘exhibited evidence of incipient

fracture, as shown in Fig. 3.4.26.

It is concluded that the radiator failed as a
result of the initiation of a fracture in a braze
‘alloy fillet by shear forces and the propagation

 

2P. 'Putriqrco et al., ANP Quar. Prog. Rep. Dec. 10,
1955, ORNL-2012, p 145.

UNCLASSIFIED
- ¥-17882

" SUPPORT ¥
MEMBER

 

of this fracture through the tube wall by tensde
forces or combinations of tensile and shear forces
during periods of cyclic operation. Since the
incidence of incipient. fractures was associated
exclusively with the presence of support members

“or plates, it is recommended that these transverse

restraints be removed entirely.  This can be
accomplished, as suggested previously, by using
a high-conductivity fin to provide transverse

support- at 2< or 4-in. intervals® and modifying

the brazing procedure accordingly.
. The tensile loading contribution to the York
radiator failure was attributed ot the time of

examination ‘to the restraining influence of the

support member, which extended up the side of
the radiator.® In view of this conclusion, the
support members of subsequent radiators, including
PWA No. 2, were sht as shown in Flg. 3.4.20,

 

 

3R. J. Gray and P, Pctrlarca, Metallo rapbzc Ex-
amination of ORNL HCF Radiator No. 1 Failures, ORNL
CF-55-10-129. ‘

UNCLASSIFIED J§
Y.17881 ..

©
2

wehes, Jo

jo.04} -

 

 

 

F:g. 3.4.23. Longitudinul View of Opposing Walls of the Tube That Failed in Nak-io-Air Radlator PWA
No. 2 as Viewed Against the Alr Flow. Note neckdown at fructure. 75X Reduced 22%. ) :

192

L

e
 

 

|
!

 

 

_ . _ : ' R R mcmsnr
= B N mncrunz

 

ERIOD ENDING JUNE

 

10,

 

 

1956

» --Ff:g'z‘:s :. FIRE DAMG cvt:;]\;:gleo S
' - TO FINS -

lo.ot |
P

)§7 .

SUPPORT

- MEMBER 5
loee]

: Fig. 3.4.24. Lengitudinal View of Opposing Wulls of a Tube Aclgucent to the Tube Thet Fclled as

Viewed Against fhe Air Flow. Note mcnplent fracfure. 75X. Reduced 34%.

]
|
|
i
o
|
1
1

. Longitudinal Vlew of Opposing Wulls of a Tube Adlncent to the Tube That Failed as

Fig. 3.4.25.
VYiewed Against the Air Flow.
Reduced 34%.

#y

-

 

 

 

Note fracture within the eutectic structure of the braze metal,

75X.

193

 
 

 

 

 

ANP PROJECT PROGRESS REPORT

UNCLASSIFIED §

& R
%%: SUPPORT
MEMBER |

R

 

 

 

Fig. 3.4.26. Corner Tube from Air Exit Face of
NaK-to-Air Radiator PWA No. 2 as Yiewed in the
Direction of the Air Flow. 75X. Reduced 36%.

prior to installation. [t may be noted, however,
that the radiator as redesigned retained the bottom
flanged plate, a top plate, and the four support
members. During the brazing cycle the NaK tubes
were brazed to these members, and a relatively
rigid condition resulted at each of the five trans-
verse sections across the radiator matrix. Local
differences in rates of heating ond cooling, par-
ticularly between the air inlet face and the
remainder of the radiator during blower startup,
could therefore bring about the development of
tensile loading. This condition could be partially
relieved by slicing the support members and plates
in @ manner similar to that utilized during dis-
section for metallographic examination, as shown
in Fig. 3.4.21. '

The development of tensile forces alone, how-
ever, cannot be assigned the full responsibility
for failure. The incidence of numerous incipient
fractures in this radiator has been related to the
presence of a support’ member or heavy plate.
Over 13,000 tube-to-fin joints have been examined

194

metallographically without the observation of o
single incipient fracture, The differences in mass
and ‘thermal conductivity of the support members
and plates as compared with the high-conductivity

“fins could result in significant differences in
 heating and cooling rates during cyclic operation..
These differences could create lateral forces that

could be resbo_'nsib'le for the initiation and propa-
gation of fractures in brazed joints between tubes
and support members in any portion of the radiator.

EXAMINATION OF FUEL-TO-NaK HEAT
EXCHANGER AFTER SERVICE

G. M. Slaughter

Tests of the fuel-to-NaK heat exéhangér, desig--

nated as |HE-3, were terminated as a result of

the detection of a leak in a tube bundle after a

total of 1794 hr of operation in the temperature
range 1100 to 1500°F. There was o temperature

differential imposed on the heat exchanger for-

1015 hr of this total time, and 21 thermal cycles
were applied over this period.

The NaK inlet and NaK outlet headers of the tube
bundle that lecked were separated from the heat
exchanger to facilitate examination and inspection,
Top and bottom views of the inlet header are
shown in Figs. 3.4.27 and 3.4,28. The general
location of the failure is evident in Fig. 3.4.28,
in that a dark reaction product can be distin-
guished from the lighter solidified fuel mixture.
Forty tubes in the area of the failure were indi-
vidually inspected with a dye penetrant and o
Borescope, and at least five tubes were found to
contain obvicus cracks. The NaK inlet header

after dissection with an abrasive cutoff wheel

 

UNCLASSIFIED
Yages

  

Fig. 3.4.27. Top of NaK Inlet Header of Fuel-to-
NaK Heat Exchanger IHE-3 Showing Tube Welds.

®

9
 

 

i

e

-/

 

to permit the examination of individual tubes is
shown in Fig. 3.4.29. Each tube was numbered
for the subsequent investigation as shown in
Fig. 3.4.30. S S

The frequency and severity of the cracks de-
tected in the initidl_inspection were most pro-
nounced in the forward rows of tubes, that is,
those with short bends.” The distance from the
tube bends to the headers was significantly shorter

UNCLASSIFIED
Y-1m47

   

 

 

Fig. 3.4.28, Bottom of NaK Inlet Header of
Fuel-to-NaK Heat Exchanger 1HE-3, Dark reaction
product indicates crea of tube failure. The light
solidified material is the fuel mixture (No. 30)
NaF-ZrF -UF, (50-46-4 mole %), . (Secret with

caption)

PERIOD ENDING JUNE 10, 1956

for the first row of tubes, being 3% in. as com-
pared with 6 in, for the last row of tubes. For g
given expansion of the 6-ft over-all straight length
of the tubes as a result of heating, ¢ substantial
degree of strain occurs in these locations, As
would be expected, cracking was more pronounced
on the tension sides of the tubes. A crack in the
tension side of tube 2 could be seen upon visual
examination (Fig. 3.4.31). Further evidence of
tube distortion at the headers can be seen.in
Fig. 3.4.32, which shows the NaK outlet hecder.
Several of the tubes of interest were mounted
intact in Castolite and polished to the approximate

UNCLASSIFIED
Y

  

 RWEAR o onccnon §

Fig. 3.4.29. NaK Inlet Header After Dissection
for Further Exomination.

UNCLASSIFIED
ORNL=-LR-DWG 14728

 

 

® 0060
000 © 0
©0 006

® @ 0

 ° 
@ 0.0 00
o.

® O

 

 

 

 

_ DIRECTION OF BENDS

Fig. 3.4.30. Diagram of NaK Inlet Header Sfiowing, Identification Numbers of Tubes Removed for

Examination.

195

 
 

 

ANP PROJECT PROGRESS REPORT

EE  UNCLASSIFIED
Y-1m315

 

Fig. 3.4.31. Crack on Tension Side of Tube 2 of
NaK Inlet Header.

center line for metallographic examination. The
tensile side of tube 3, shown in Fig. 3.4.33,
illustrates ‘typical severe cracking and corrosion
by the fuel mixture which circulated on the outside
of the tubes. A similar condition is evident in
Fig. 3.4.34, which is a panorama of the tension
side of iube 17. The opposite face of tube 17,
o pancrama of which is shown in Fig. 3 4.35, does
not exhibit so serious ¢ condition, It appears that
the corrosion and the stresses combined to form
an abnormally unfavorable condition.

196

UNCLASSIFIED
T Yame

 

Fig. 3.4.32. Bottom Side of NaK Outlet Header
Showing Distortion of Tubes.

The extent of the corrosion was investigated by
examining tube 93- to ensure that the large degree
of attack observed in the previous samples did
not resuft from the reaction of the two fluids at
the locations of 'rhe failures. The results of the
examination (F;g. 3.4.36) indicate that, in general,
severe corrosion was prevalent throughout the
tubes in the inlet header. The inner tube wall of
a typical tube is shown in Fig. 3.4.37; cracks
and corrosion emanating from the outer wall may
also be seen. A white deposit was found in the
cracks in some areas, as shown in Figs. 3.4.38
and 3.4.39. The nature of this deposit, as well
as a detailed investigation of the mass transfer
and corrosion, will be reported later.

 
 

| PERIOD ENDING JUNE 10, 1956

 

 

 

 

§
i
i
i

 

Fig. 3.4.33. Tube 3 of NaK Inlet Header Showing Cracks in Tension Side. Etchant: electrolytic oxalic
acid. 100X,

 

2]

 

Fig. 3.4.34. Panoramc of Tension Side of Tube 17 of NaK Inlet Header Showing Gr-osg Cracks and Cor-
g rosion. Etchant: electrolytic oxalic acid. 33X.

197

 
 

 

 

 

ANP PROJECT PROGRESS REPORT

 

Fig. 3.4.35. rPcnorcmrm of Compressién Side of Tube 17 Showing"Corrosion and Occasional Cracks.
.Etchant: electrolytic oxalic acid. 33X.

’

 

 

 

Fig. 3.4.36. Tension Side of Tube 93 Showing Severe Corrosion, Etchant: electrolytic oxalic acid. 1005(.‘

198

 
 

 

 

 

 

 

PERIOD ENDING JUNE 10, 1956

URCLASSI
WIEEY Y

 

 

 

 

Fig. 3.4.37. Inner Surface of Tube 17, Cracks and corrosion emanating from outer surface may be seen.
Etchant: electrolytic oxalic acid. 200X,

 

Fig. 3.4.38. Cracks and Deposits on Tension Side of Tube 19, Etchant: electrolytic oxalic acid. 100X,

199

 
 

 

 

 

ANP PROJECT PROGRESS REPORT

.

 

Fig. 3.4.39. Crack and Deposit Evident in Fig. 3.4.38 at a Higher Magnification. Etchant: electrolytic
oxalic acid. 500X.

 

200

 
 

 

o/

-3.5. MECHANICAL

D. AI

EFFECT OF ENVIRONMENT ON CREEP-
RUPTURE PROPERTIES
OF HASTELLOY B

‘C R. Kennedy!

Revised desugn data obtained from creep tests

of solution-annealed Hasfelloy B sheet stock in

various envuronments at 1300, 1500, und_'|650°F
ore summarized in Figs. 3.5.1, 3.5.2, and 3.5.3.
The times to 0.5, 1, 2, 5, and 10% total strain at
each temperature in the various environments are
the same, and for stresses for which the rupture
life is more than 300 hr the effect of environment
is shown to be negligible, The creep curves
obtained ot 1500°F in air and in argon, shown in

Fig. 3.5.4, appear to indicate that the better

performance in air than in the other environments
at stresses for which the rupture life is less than

 

_IOn assignment from Prott & Whitney Aircraft.

PERIOD ENDING JUNE 10, 1956

PROPERTIES STUDIES
-Douglas -. '

300 hr is caused by the ability of air to strengthen
Hastelloy B during third-stage creep. - For rupture
lives longer than 300 hr there is considerably
less third-stage creep because of the aging
characteristics of the alloy, and, as seen in
Figs. 3.5.1, 3.5.2, and 3.5.3, the effect of environ-
“ment diminishes, Design curves produced from
limited data for solution-annealed Hastelloy B
- sheet tested at 1800°F in argon and in the fuel
‘mixture (No. 30) NaF-ZrF -UF  (50-46-4 mole %)
are shown in Fig. 3.5.5. At 1800°F Hastelloy B
does not age perceptibly and the amount of third-
stage creep is great in tests at all stress levels,
Thus the fuel mixture strengthens the alloy at all
stress levels, Of the environments tested, only
those which are ‘‘surface active” (produce a
thin, tightly adherent film on the surface of the
metal) ‘affect the creep properties of the metal,
It can be seen from Figs, 3.5.1, 3.5.2, 3.5.3,
" and 3.5.5 that, of the environments tested, only

SEORET
ORNL-LR-DWG 14729

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

50,000 <]
\\-\ h\$(j).
M — P~ \\0,9‘\ ,
40,000 \\. ] - \ ‘\\“;/4/4
e N, -
\‘-.._____ \\ 7 \ \\ \':GCN '9(,&
\\\ \ \ \\ B ;Z."l?é\
N N D N /4/4/
30,000 \ )
o | o \
7’5 L ’ '. '- 0.570 \ ._ o ’ _7 ) . 7 : y
: 20000 — e T N s L
e N N N R
\\ NI | |
A2 s 1020 50 lOO 200 500 1000 - 2000 - 5000 . 10,000
Lol e e e TlME(hr) - e :

Fig. 3.5.1. Design Curves for Hustelloy B Sheet Solution Annealed at 2100°F for 2 hr and Tested
in Argon and in the Fuel Mixture (No. 730) NaF.ZeF, UF, (50-46-4 mole %) at 1300°F.

201

 
 

 

 

ANP PROJECT PROGRESS REPORT

— U

ORNL-LR~-DWG _14730

20,000

 

f'g"-
«
- W
L
&
“ 10,000
9000 -
8000
7000
N
6000 : : ' L
1 2 5 10 20 50 100 200 500 1000 2000_ 5000 {0,000
' TIME (hr) ) '
‘Fig. 3.5.2. Design Curves for Hastelloy B Sheet Solution Annealed at 2100°F for 2 hr and Tested -

in Yarious Environments ot 1500°F.

SEGRENr
ORNL-LR-DWG 14734 ) o
14,000

13,000
12,000

14,000
10,000

9000

8000

STRESS (psi}

7000

6000

5000

 

~ 4000 . L
{ 2 5 10 20 50 100 200 500 {000 2000 . 5000 40,000

TIME (hr)

Fig. 3.5.3. Design Curves for Hastelloy B Sheet Solution Anneuled at 2100°F for 2 hr und Tesfed . o
in Argon and in the Fuel Mixture (No. 30) NaF-ZrF UF , (50-46-4 mole %) at 1650°F. S

202
 

 

 

 

PERIOD ENDING JUNE 10, 1956

‘ ! B _ UNCLASSIFIED

60 ORNL-LR-DWG 14732

 

50

 

 

 

 

H
o

 

 

W
o

 

ELONGATION (%)

 

20

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

v 100 200 300 400 500 600 700 800 900 1000
TIME (hr)

Fig. 3.5.4. Design Curves for Hastelloy B Sheet Solution Annealed ot 2100°F for 2 hr and Tested
in Air and in Argon at 1500°F. -

BSOGRE—
ORNL-LR-DWG 44733

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

- - 20,000 -
10,000 ' /RUPTURE IN ARGON
so%o /11
‘ 8000 ©, o, ) ML
7000 1% |2% 5% 5% F R |
."5 6000 03% | IS wp\\!“““i:\fiewflap
a  PVYVY TN e P ~Zr, )
E ~ o RN ‘--..__- \\:\\\\\‘ \-...:lsio;.qs s
o : ' L L ~hao~¢e :
N 4000 s = -e o [ .’no/ey -
T ] Pl | “-\ B \'::- °)
e L TS TR/ IR
3000 - .
: N \\ N o
SRR NG \\\
. '_.-_“-2_000 _\\\\ i
000 L1 o AL . L ,
S g - .5 - 10 - 20 . 50. 100 200 500 - 4000 2000 5000 10,000
el ) _ ' . . TIME {hr) T - o

- Fig. 3.5.5. Design Curves for Hastelloy B Sheet Solution Annealed at 2100°F for 2 hr and Tested
N in Argon and in the Fuel Mixture (No. 30) NaF-ZrF +-UF, (50-46-4 mole %) at 1800°F. '

203

 
 

 

 

 

 

ANP PROJECT PROGRESS REPORT

air and the fuel mixture seem to be surface active,
Air, of course, produces an oxide film which is
ever present, and the fuel mixture, in effect,
creates a very thin surface film by leaching one
of the alloy constituents and producing another
phase on the surface, as shown in Figs. 3.5.6

and 3,5.7.

SHORT-TIME HIGH-TEMPERATURE TENSILE
PROPERTIES OF HASTELLOY B

C. R. Kennedy

The short-time high-temperature tensile proper-
ties of solution-annealed Hastelloy B are illustrated
in Fig. 3.5.8, which gives the yield and ultimate
strengths and the final elongations in the temper-
ature range 1000 to 1800°F. As may be seen there
is a distinct decrease in the final elongation and
in the ultimote strength ot temperatures around
1200°F. This change in properties occurs in the
temperature range in which the type of frocture
transforms from predominently fransgranular to
intergronylar, -It is interesting to note that the

change in the yield strength with temperature is

relatively smoll.

CREEP-RUPTURE PROPERTIES
OF HASTELLOY W
C. R. Kennedy

Creep testing of Haste”oy W is now in progress,
ond design data are presented in Figs., 3.5.9,
3.5.10, and 3.5.11 for solution-annecaled sheet

tested in argon at 1300, 1500, and 1650°F.

Hastelloy W, which has almost the same compo-

sition as Hastelloy B, except for the addition of -
5% chromium and the deletion of 3% molybdenum,:

has creep properties very similar to those of
Hastelloy B. Although Hastelloy W exhibits less
of a tendency to age than Hastelioy B, as shown
in Figs. 3.5.12 and 3.5.13, a decrease in ductility
occurs at 1300°F.. This is also shown in Fig.
3.5.9, in which the absence of a 10% curve indi-
cates that the total strain at rupture was less

than 10%.

Rupture points obtained from tests with the fuel
mixture (No. 30) NaF-ZrF ,-UF, (50-46-4 mole %)
are also shown in Figs, 3.5.9 and 3,5.10. The
times to 0.5, 1, 2, 5, and 10% total strain are
identical for the same stress and temperature,
and, as shown, only the rupture life is affected by

 

Fig. 3.5.6. Surface Effect on the Unstressed Portion of a Hastelloy B. Specimen After Exposure
to the Fuel Mixture (No. 30) NaF-ZrF ,-UF , (50-46-4 mole %) for 1700 he. 1000X. (Secretwith-eaption)

204

 
 

 

1h

o

 

 

PERIOD ENDING JUNE 10, 1956

 

_Fig. 3.5.7. Surface Effect on the Stressed Portion of a Hastelloy B Spec:men After Exposure to
the Fuel Mixture (No. 30) NoF-ZrF, -UF (50-46-4 mole %) for 1700 hr ot a Stress of 13,500 psi. 1000X.

(Secret-with—caption)

120,000 |

100,000

80,000

Cy

" 'STRESS (psi) .

 

 

Fig. 3.5.8.

UNGLASSIFIED

 

N\

ORNL-LR-DWG 14734

 

 

 

60,000

 

 

 

 

 

 

 

 

 

 

“Short-Time T
Range 1000 tc 1800°F.

 

 

 

 

 

 

 

 

 

 

60

- : ] lozuvewp L ,
I & \\ 40
_— ' 20
: - - “ELONGATION - ‘17
- - 7 0
1200 1400 ‘1600 . 1800
- TEMPERATURE("F)

- ELONGATION (%)

ensile Data for Soluhon-AnneuIed Hastelloy B Tested in the Temperature

205

 
 

 

 

 

ANP PROJECT PROGRESS REPORT

SECRET
ORNL-LR-DWG 44735

 

50,000

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

o
40,000 “\\ NS
\._.. \
P \ ) .
""--____-.--.-- i \ \P\
e —— @
30,000 S =
-'_'—-——-_--_'--'_i-— "'---_“_-.- .......‘\ \
- e Tt I P
2 \\\ 5% RUPTURE
2 N
& \\ "\._\
= : , TN
0 000 ~
20 0.5% 1% 2%
@ RUPTURE POINTS FOR TESTS IN NaF- ZrF4-UF‘ {50-46-4 mole %)
10,000 ‘ v
t 2 5 {0 20 50 100 200 500 {000 2000 5000 10,000

TIME (hr)

Fig. 3.5.9. Design Curves for Hastelloy W Sheet Solution Annealed ot 2100°F for 2 hr and Tested
in Argon at 1300°F. S

DGR
ORNL-LR-DWG 14736
20,000

18,000

16,000

14,000

12,000

STRESS (psi}

10,000

8000

& RUPTURE POINTS FOR TESTS IN NaF—ZrF, —UF,; (60-46 -4 mole %)

 

6000 .
1 2 5 10 20 50 100 200 500 {000 2000 5000 10,000

TIME (hr)

Fig. 3.5.10. Design Curves for Hastelloy W Sheet Solution Annealed at 2100°F for 2 hr and Tested
in Argon at 1500°F.

206
 

 

 

 

14,000

12,000

10,000

0.5% N% 2%

8000

STRESS (psi)

6000

4000
{ 2 5 10 20 5C

 

PERIOD ENDING JUNRE 10, 1956

UNCLASSIFIED
ORNL-LR-DWG 14737

RUPTURE

100 200 500 1000 2000 5000 10,000
TIME (hr)

Flg. 3.5.11. Design Curves for Hastelloy W Sheet Solution Annealed at 2100°F for 2 hr und Tested

in Argen at 1650°F.

the environment. The effect of the fuel mixture on
the creep properties of Hastelloy W sheet appears
to be the same as that shown for Hastelloy B.

SHORT-TIME HIGH-TEMPERATURE TENSILE
PROPERTIES OF INCONEL

- J. R Weir, Jr.

The tensule properhes of lnconel sheei huve'fi'
been determmed at temperatures from 78 to 2200°F. -
The yield point, by 0.2% offset, and the ulhmate )
~ strength are shown in Fig. 3. 5.14 for both fines
__gramed and coarse-gromed material, As ‘may be -

the fine-grained material may be the more desirable
structural material because of its better tensile
properties in the temperature range of interest,

CREEP TESTS OF INCONEL IN FUSED SALTS
J. R, Weir, Jr.

Inconel was creep tested at 1500°F and a stress

of 3500 psi in NaF-ZrF (50-50 mole %) in order

to compare the severity of attack with that found
after similar creeptests in the fuel mixture (No. 30)

‘NoF-ZrF +UF, (50-46-4 mole %), The results are
- shown m Flgs. 3.5.16 and 3.5.17." Very little
-:v;'surface void formation is seen in the case of the

- specvmen tested in the nonuramum-beurmg mixture,
" in comparison with the attack by the fuel mixture.,

_,f‘.-These results are 'Further ev:dence that much of
seen,. the fine-grained : material “has the better

" the attack by the fuel mixture on lnconel af ]500°F
strength properties at temperotures up to-1700°F, - y

- ma be attnbuted to the reaction
Transuem‘ loads ‘induced by thermal fluctuchons_’:- Y;

may, -in some " cases, be of greater concern than
~ the stahc loads that result from pressure differ-
" ential. - Therefore, ‘even though the fme-gmmed"_"
,lnconel has less creep resistance in the fused o
“salts than the coarse-grdined Inconel (Flg. 3.5.15),

2UF + Cr (m lnconel) —_— 2UF + CrF

CREEP TESTS OF WELDED INCONEL
“J. R Welr, Jr, o '

Several 0060-m.-fh|ck sheet-type creep speci-
mens were machined from welded /-m. Inconel
sheet stock and a few creep tests were run in the

207

 
80¢

254 UNCLASSIFIED:
5706 4

 

 

 

 

 

 

 

 

 

Fig. 3.5.12, Hastelloy B Sheet After Creep Testing afa - Fig, 3.5.13. Mastelloy W Sheet After Creep Testing at a
Stress of 30,000 psi ot 1300°F in Argon; Ruptured in 185hr, - Stress "of-_35.00_’0” psi at 1300°F in Argon; Ruptured in 450 hr.
' S o © - 100X, Reduced 17.5%. o o -

IOOX. Reduced 17.5%. N

s

LA0dIY $STYI0Yd LD3T0dd dNV

 

 
 

 

 

 

-

»

>

. 1000 L

 

v o PERIOD ENDING JUNE 10, 1956

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

UNGLASSIFIED
ORNL~LR-DWG 14962
TENSILE STRENGTH| |
FINE GRAIN
100,000
| TENSILE g7 \
80,000 T COARSE GF?EPIGTH \
5 . \
% N
< 60,000
o
: \
B \
W \
' Fl °"'ELD \\
40,000. \J SRa’ \
CRoPoRy, TION, '\\\
F'INEG LUMIT N
RAIN \ \\\
0.2
20,000 =52 YIELD, Coarge GRAIN N N
‘ N
| U=
o L : _ ~
o 400 800 1200 1600 2000 2400 2800

TEMPERATURE (°F)

Fig.. »3.5.14._ Cbmpurison of Tensile Properties of Fine- and Coarse-Grained Inconel at Tempera-
tures Between 78 and 2200°F at a Stress Rate of 0.016 (in./in.)/min, :

CEORET
20.000 ORNL-LR-DWG 14963
3 .

10,000 |
8000
6000

4000

" 'STRESS (psi)

OOARSE GRAIN,.
e fiSO' .

2000 |-

b b 1= 1ot | L] {FiNe erA,
e b b L 1e500F —

 

Syt 2 08 70 40 20t 5O L 100 200 - 500 {000 2000 5000 {0,000
o Tl ' - ' "TIME(hr),' ST e T

Flg. 3 5. 15. Compunson of Stress-Rupfure Properfies of Fine- and Coarse-Gramed Inconel in the
Fuel Mixture (No. 30) NaF-ZsF, ~YUF, (50-46-4 mole %) ot 1300, 1500, and 1650°F.

209

 
ole

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

"Fig. 3.5.16. Inconel After a Creep Test in.the Fuel Fig. 3.5.17. [Inconel After a Creep Test in NaF-ZrF,
“Mixture (No, 30) NaF-ZrF -UF , (50-46-4 mole %) at 1500°F - (50-50 mole %) at 1500°F Under a Stress of 3500 psi. 100X.
‘Under a Stress of 3500 psi. 100X. Reduced 18%. (Secret ~ Reduced 18%. ‘

LAOdITY SSTYO0Yd LI3F0dd ANV

 

 
 

 

 

 

 

'\

"

fuel mixture (No, 30) NaF-ZrF 1-UF (50-46-4

mole %) and in argon. The .results obtained to
date are summarized in Table 3.5.1,
The positions of the fractures in"the specimens

tested at 1500°F in the fuel mixture are shown in

PERIOD ENDING JUNE 10, 1956

psi and No. 3 at 4000 psi. As may be seen, the
weld did not deform appreciably, compared with
the base metal, in the gage length. Metallographic
examination of the specimen tested at 1300°F
disclosed that the weld metal was more corrosion

Fig. 3.5.18. Specimen No. 2 was tested at 3000

resistant than the base metal.

TABLE 3.5.1. RESULTS OF CREEP TESTS OF WELDED INCONEL SHEET IN THE FUEL MIXTURE
‘ (No. 30) NaF-Z¢F (-UF, (50-46-4 mole %) AND IN ARGON

 

 

 

Stress - Temperature £ H _ Rupture Life Elongation
(psi) (°F) nvironment eat Treatment . (hr) (%)
12,000 1300 Fuel mixture As received 110* _ 21
4,000 ' 1500 Fuel mixture Coarse grained 380 n
4,000 1500 Argon Coarse grained 750%** 10
3,000 1500 Fuel mixture As received 7 820 5
3,000 1500 L Argon B ' As received e 2000** . 4

*Premuture rupture, probably cuused by confuminaiion of the fuel rmxfure.
*+Still in test, o
ENCLASSIFIED
Yalnes)

 

_Fig. 3.5.18. Welded Inconel Specimens After Creep Tests In the Fuel Mixture (No. 30) NaF-ZtF -
UF, (50-46-4 mole %) ot 1500°F. Specimen No. 2 was tested at a stress of 3000 psi and specimen
No. 3 was tested at a stress of 4000 psi. (Sesretawith-ception)

211

 
 

 

 

 

ANP PROJECT PROGRESS REPORT

3.6, CERAMIC RESEARCH
L. M. Doney

RARE-EARTH-OXIDE COMPACTS FOR ART
CONTROL RODS

J. A, Griffin L. M. Doney

The design of the ART control rods calls for a
porous compact of rare-earth oxides. The pores of
the compact are to be filled with metallic sodium,
and the compacts are to be canned in Inconel. The
following process was developed for the fabrication
of the compacts. The as-received rare-earth-oxide
material (Code 920 from Lindsay Chemical Co.)
was pressed -at 4000 psi into compacts, which
were calcined at 1325°C for 1 hr, The calcination
step was carried out to reduce the shrinkage of the
oxide during the final sintering. These compacts

“were crushed to pass an 80-mesh screen. The
final mixture was compounded from 75% of the

80-mesh calcined -material and 25% of the as-
received Lindsay Code 920 oxide. This combi-

nation was thoroughly mixed and pressed, with no -
binder, at a pressure of 4000 psi. These compacts

were then sintered at 1425°C for 55 min in.air.
Samples of the sintered compacts. are shown in

Fig. 3.6.1.

The compacts were oversize in all dimensions
so that they could be ground accurately to size
and so that no difficulty would be encountered

during canning. The grinding was carried out by a
commercial ceramic firm. The inside and outside

 

Fig. 3.6.1. Sintered Rare-Earth Compacts Before Being Ground to Desired Dimenéibns.

212

 
 

W

 

¥

s

of each compact was ground to the required
dimension, and the ends, as well as being ground
to the required length, were made flat and parallel.
The ground pieces are shown in Fig. 3.6.2.

The Lindsay Chemical Co. supplied the foflowing |

analysis of the Code 920 rare-earth oxides used in
the fabrication of the compacts:

Samarium oxide 45.0-49.5%
Gadolinium oxide 22.5-27%
Neodymium oxide 0.9-4.5%
Prasecdymium oxide - 0.9-3.6%

Cerium oxide 0-0.9%

3.6.2.Rar

 

 

PERIOD ENDING JUNE 10, 1956

s
e

Europium oxide 0.9-1.2%
Other rare-earth oxides 8.1-14.8%

.Tb.i_olfrure-earfh oxides 90% (minimum)
PETROGRAPHIC EXAMINATIONS OF
FLUORIDE FUELS

~G. D. White T. N. McVay, Consultant

Examinations of fluoride fuel samples with the
petrographic microscope were performed at a rate
of about 200 samples per month. The materials
examined .included quenched samples for equi-
librium-diagram determinations, control samples

~from experimental runs, and samples from pro-

duction batches. The results of these exami-

 nations are reported in Part 2, '‘Chemistry.”’

UNCLASSIFIED

PHOTO 26348

213

 
 

 

 

 

 

 

ANP PROJECT PROGRESS REPORT

3.7. NONDESTRUCTIVE TESTING STUDIES
R. B. Oliver

EDDY-CURRENT TESTING OF
SMALL-DIAMETER TUBING

J. W, Allen

Continued study of the application of the cyclo-
graph! to the problem of the inspection of small-
diameter tubing has revealed that, in addition to
its use as a flaw detector, the cyclograph may
also be used to gage adherence to dimensional
tolerances. Although changes in diameter and in
wall thickness are inseparable in the readout of
the instrument, the tolerance limits for both di-
mensions may be established by utilizing two
standards: (1) the minimum acceptable diameter
and wall thickness and (2) the maximum allowable
diameter and ‘wall thickness. Dimensional checks
made by using this method, augmented with me-
chanical measurements, have been used success-
fully to determine the dimensional acceptability
of approximately 6000 ft of % -in.-OD, 0.025-in.-

 

1R. B. Oliver, J. W. Allen, and K. Reber, ANP Quar.
Prog. Rep. March 10, 1956, ORNL-2061, p 164,

they have a length of /16»

wall and 0.229-in.-0D, 0.025-in.-wall CX-900

Inconel tubing.

The 200-kc cyclograph trace of a 0.229-in.-OD,
0.025-in.=wall Inconel tube that has small wall-
thickness variations along its length and no per-
ceptible diameter variations is shown in Fig.3.7.1.
The wall thickness was plofted from Vidigage
(ultrasonic-resonance-type thickness gage) and
mechanical measurements. It may be seen that,
although there is not perfect agreement, there. is
a fairly close correspondence between the two
types of checks. The disagreement is probably
due to minute amounts of eccentricity and inters
granular attack on the inside surface of the tube
having been detected by the cyclograph. _

The confidence level of the mterpretahons of
mdacatnons from this instrument is. _increasing
with. continued use. Because the _se_nsmg coil
observes at cmy one instant a section of tubing
approximately / in. long, it is not possfl:le to
detect pin holes, except when they occur in
clusters. [n general, defects can be resolved if
in. ‘or longer and have

UNGLASSIFIED -

 

 

 

 

 

 

 

 

 

— ORNL -LR-DWG 13970
- £ 00275 I ~

fi 0.0270

Z . ’ TOLERANCE LIMITS

X 0,0265

: r:E ,0.0260‘ f\ /\\ N /\ f\ LN AN f\ N y_ .

- UNWTV Y Y YT N \

L0255 ‘
§ 00258 |
Z 00250

AVERAGE WALL THICKNESS DETERMINED WITH VIDIGAGE
OUTSIDE DIAMETER CONSTANT AT 0.230 in,

 

 

 

 

 

 

 

 

 

 

 

 

PSS

 

e 1

 

 

 

 

 

 

 

 

200-ke¢ CYCLOGRAPH TRACE
0 6 2 18 24 30 36 42 48 54 60 66 72 78
LENGTH (in.)

Fig. 3.7.1.

Cyclograph Record at 200 kc from a 0.229-in.-OD, 0.025-in.-Wall CX-900 Inconel Tube

Compared with Dimensional-Variation Measurements Made Mechanically and with the Vidigage.

214

 
T S e i

 

 

 

"

i

wy

&

 

depths greater than the background dimensional
variations in the tube,

ULTRASON'C |NSPECT|6N OF TUBING
R. W. McClung

Approximately 6000 ft of CX-900 Inconel tubing
has been inspected by the immersed ultrasound
method, Two sizes of tubing 3/16 in. OD, 0.025 in.
wall and 0.229 in. OD, 0.025 in. wall, were in-
spected and the average rejection by this test
was about 2%. Each tube received a double in-
spection, with the vltrasound being beamed around
the tube in two different directions to improve the
chance of detection of unfavorably oriented crack-
like defects. This double inspection of the precut
tubing reduced the inspection rate to below the
original estimate. The current inspection rate is
approximately 500 ft in an S-hr day for lengths
up to 10 ft,

Defects 0.0015 in. deep and I/1 6

in. long on

polished, scratch-free tubing can be detected with

" PERIOD ENDING JUNE 10, 1956

‘an estimated confidence of 80 to 90%. If scratches

are present they will produce signals comparable
to those from the very small defects and effectively
increase the minimum detectable defect size. If
the defect is appreciably deeper than the scratches
and if its length exceeds the maximum dimension
of the transducer, it is not difficult to differentiate
between defects and scratches. Since the ultra-
sonic method is insensitive to dimensions, the
inherent dimensional variations that give trouble
in an eddy-current inspection. have no effect on
ultrasonic inspection. Both the eddy-current and
the ultrasonic methods are capable of detecting
very small defects, with the eddy-current inspec-
tion being confused by dimensional variations and
the ultrasound method being confused by scratches.
A comparison of the ‘resulfs of the two tests aids
in the identification of spurious signals. ,

A fypical small defect found on the inside of
the / 6-in.-0D, 0.025-in.-wall CS$-900 tubing is
shown in Fig. 372 This crack is, 0.0015 in.
deep (6%) and about ¥ ¢ ine long.

 

Fig. 3.7.2.

Small Defect Found by Ultrasonic Inspection on the Inner Surface of a ?'/ -ln.-OD

0.025-in.-Wall Inconel Tube. The defect is 0.0015 in. deep and agpproximately 46 in. long.

215

 
 

 

 

 

 

 

 

ULTRASONIC INSPECTION OF PIPE

No development work on the ultrasonic method
for the inspection of pipe had been planned, be-
cause the currently employed contact method was
reported to be adequate. Preliminary attempts
to employ this method, however, revealed several
problems that detracted’ from the reliability of
contact inspection. 1t was found to be very diffi-
cult to fabricate Lucite shoes that would fit the
contour of the pipe and would, at the same time,
limit the sound beam to the proper spectrum of
incident angles. A sizable part of the sound was
propagated as a surface wave and produced inordi-
nately large signals from surface scratches. For
this reason, defects and scratches could not be
separated. Also, the contact method requires that
a thin film of oil be maintained between the trans-
ducer and the pipe surface, and, thus, any surface
roughness, vibration, ovality, or an inadequate
supply of oil resulted in a loss of signal and

frequent failure to detect defects on the inner
surface, Immersed ultrdSound permitted a solution”

to these difficulties, and the use of the “B" scan
for date- presentahon permitted a high inspection
speed with rapid interpretation of "the signals
produced in the tube wall. '

A scanning tank of maximum simplicity was
designed and fabricated for the ultrasonic inspec-
tion of pipe. The tank is 26 ft long, 14 in. wide,
and 20 in. deep and is equipped with a variable-
speed headstock and chuck and an assembly of

 

2On assignment to Homogeneous Reactor Project.

216

two 3-in.-dia casters, in liev of a tail stock This
assembly contains and rotates the pipe around
its own axis. The transducer mounting is aligned
with reference to the pipe axis by a pair of Micarta
guides, and it is manually translated along the
pipe. Very straight pipe can be rotated at speeds
of 200 rpm, but much of the pipe is too crooked
for this speed. _Even.in the worst case, however,
rotational speeds of 60 to 100 rpm are attainable.
This equipment is belng used for the inspection
of large quantities of pipe in sizes from ?’ to
6 in. IPS for the ETU and the ART.

Satisfactory inspection of tubular shopes is de-
pendent upon obtaining reference notches of the
proper depth (3 to 5% of the wall thickness) on
both the inner and outer surfaces. A known, re-
producnble notch is easily produced on the outer
surface of both pipe and tubing; howe_ver, it is
difficult to produce such notches on the inner
surface of even large pipe. No really satisfactory
notch has yet been produced on the inner surface
of small-diameter tubing, and the inside reference

" notch ‘is necessary’ to prove that the alignment
will reveal defects on.the inner surface.’ Therefore

this method is ot being’ used for. the mspectlon

. of small-digmeter tibing,

Four of the eight nozzle welds on the cell bemg
fabricated to house the ART in Building 7503
were inspected by the vltrasound method. The
observed indications of cracking in the weld and
in the heat affected zone resulted in the removal
of all the welds. The cracks revealed were con-
firmed by arc gouging and by Magnaflux tests with
sufficient frequency to justify the rejection of the
welds on the basis of the ultrasound indications.

i .
 

 

Part 4

HEAT TRANSFER AND PHYSICAL PROPERTIES

H. F. Poppendiek

RADIATION DAMAGE

G. W. Keilholtz

FUEL RECOVERY AND REPROCESSING
H.. K. Jackson
_ CRITICAL EXPERIMENTS

ADcalhhan |

 
 

 

 
 

 

 

 

=f.

»

4.1. HEAT TRANSFER AND PHYSICAL PROPERTIES
H. F. Poppendiek

ART FUEL-TO-NaK HEAT EXCHANGER
S. |. Cohen J. L. Wantland

The ART fuel-to-NaK heat exchanger test system
was modified for further studies of heat transfer
and friction characteristics. Six 60-deg staggered
spacers and six §0-deg inclined spacers were
placed alternately on the tube bundle, and trans-

verse pressure taps were installed across one

spacer of each type. The average transverse
pressure drop across each instrumented spacer
was plotted in terms of the ratio of the transverse
pressure drop to the velocity head times the fluid
density over the Reynolds modulus range of 3000
to 8000. For the inclined spacer this term varied
from 1.4 to 1.1, ond for the staggered spocer the
data fell randomly between 0.01 and 0.03, with no
definite trend established. The analytical investi-
gation, referred to previously,? on a method of

- correlating  fluid-friction data for flow parallel

to a square array of cylindrical tubes in rectangular
cross-section flow channels was completed.3

An experimental study is being made of the
fuel-side flvid-friction characteristics of a mockup
of the present design of the ART fuel-to-NaK
heat exchanger. The dimensions of this apparatus
and the spacer configuration will simulate the
ART heat exchanger as nearly as possible, except
that there will be no headers and the tube bundle
will not have the curvature specified for the
ART; that is, the friction characteristics for a
straight length of tube bundle will be determined. .

An experimental heat -exchanger has been de- -
signed and is now. bemg fabricated . ‘that will be -

similar to the one on whlch tests were conducfed

- previously,4 wuth the exception that the tubes will
- be .spaced on a triongular array . rather ‘than a
square ar_ray. Thls mveshgahon will provnde g

 

350 ‘Wantland, A’ Metbod of Correlatmg Experz-

- ;,- .‘mental Fluid Frzct:on Data  for Tube Bundles of
. Different Size and Tube Bundle -to Sbell Wall Spacmg, _‘

ORNL- CF-56-4-162 {Aprit 5, 1956). .

4, L. Wantland Thermal Cbaracter:stzcs of tbe ART

Fuel-tosNaK Heat Exchanger, ORNL CF=55-12-120
(Dec. 22, ]955)0

data for direct comparison of the heat-transfer
and fluid-friction characteristics of the two differ-
ent configurations in the transitional Reynolds
modulus range.

ART HYDRODYNAMICS

C. M. Copenha\)er F. E. Lynch
G. L. Muller®

Sodium Flow in Reflector Cooling System

A 5/22-scale mode! of the reflector—core shell
cooling annulus and inlet system was designed and
fabricated (Fig. 4.1.1) in order to determine quanti-
tatively the flow distribution-in the annulus and
the various pressure drops in the sodium system.
The fluid used in the model was water at temper-
atures between 20 and 50°C, Although the model
does not incorporate the reflector cooling holes,
12 nozzles were spaced radially around the inlet
header to take off fluid at various sectors of the
inlet header to mock up flow through the cooling
holes.

The flow distribution was studied by three
methods: (1) the static pressure measured axially
across the annulus in conjunction with the tfotal
annulus flow rate measured by a rotameter gave

" the average velocities through the annulus for the

two positions nearest the inlets and the two
positions farthest away, (2) pitot-static combi-
nation measurements were obtained at positions

1-3 and 3-3 (Fig. 4.1.1), and (3) high-speed motion

‘pictures were made of periodic ink injections.

A quantitative measurement of the flow distri-
bution was defined as -

Vo, + Vy

!

;01@
R
<

1+ V,

» — — R ' _'where
oY Wuntland ANP Quar. Prog. Rep ]une 10. 1955.? S
. ?,ORNL-IB%. P 149. L S

2J. L. Wonfland ANP Quar. Prog._Rep. Marcb 10. o
_’,1956; ORNL-2061, p 173. . _

0, = volumemc flow rate in annulus farthest
from inlet,

-0, = volumetric flow rate in annulus nearesf to

~inlet, : :
V- ‘= average llnear velocnty ut station indi-
-cated by subscript number.

 

30n assignment from Pratt & Whitney Aircraft,

219

 
 

 

 

ANP PROJECT PROGRESS REPORT

SCTRET™
ORNL-LR-DWG 14738

PUMP

  
 
  
   
 
  
 
  
  
   
 
   

   
     

" STATION 4

STATION 3\

  

PUMP B PUMP A

 

INLET HEADER
(PLEXIGLAS)

Pg3-5

  
   
 

  
   

Fle S,
G S
2/ g
-

’,

 

7
2SS
277

  

7,

  

7,

TC

TO :
ROTAMETER

' ROTAMETER

\\\\\

ORIGINAL GAP = 0.031 in. { x 22

=0.135in.)
PRESENT GAP = 0.043 in. ( x =0.

1875 in.)

3.14in.

5
22
5

SHELL TO REPRESENT
REFLECTOR (PLEXIGLAS

io.91 in.|0.91 in|

 

Pg3

-3
Py3-3

OQUTER CORE SHELL
{14 - S-T-41 ALUMINUM ALLOY)

344 in,

 

 

ROTAMETER ——TO PUMP

Fig. 4.1.1. Vertical Section Through Reflector—Core Shell Annulus Medel.

v

220

TATION 1

 
 

 

 

 

-

This relationship gives the ratio ‘of the averages
of the velocities at the two stations farthest from
the inlets to the averages of the velocities at the
two stations nearest the inlets. The relationship
for the flow distribution when only one pump was
operating was defined as

 

v v,

Y_ pump A 4+ ~—n | PUmP B
0 /o V, | operating Va operating
Q,, 2

Some of the experimental results are presented
in Fig. 4.1.2 in terms of these relationships.
The predicted flow distribution in the ART system

SR
- ORNL-LR-DWG 14739

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

110
ANNULUS THICKNESS = 0.4875in. [
RATIO OF ANNULUS FLOW TO COOLING - HOLE
— FLOW =0.609
A
1.08
A -
. ] . N ] -
i—r""—’
1.06
F A& METHOD ¢
} | ONE PUMP OPERATING {‘ METHOD 2
1.04 0 METHOD {
TWO PUMPS OPERATING {. METHOD 2
. e e
1.02 o 2 ot = = =]
%o
1.00 ‘
Caxi0® 0 2 - 5 a0

' REYNOLDS MODULUS AT MIDPLANE

 Fig. - 4.1 .2: Flow Dismbution in _Céné'efiirié
Reflecfor-—Core Shell Annulus. | S

PERIOD ENDING JUNE 10, 1956

was obtained by extrapolating the data to the
Reynolds numbers for the actual system at the
mid-plane of the annulus. A summary of the
resuylts is given in Table 4.1.1.

The minimum flow rate was found in the portions
of the annulus nearest the inlets, while the maxi-
mum flow rate was observed in the portions
farthest away. These conditions cre the reverse
of those anticipated, and a possible explanation
is that the fluid entering the annulus from the
inlet header must abruptly change direction and
it probably undergoes high velocity losses, particu-
larly near the inlet to the annulus where the
momentum is the greatest. After the model inlet
header used in these experiments had been fabri-
cated, the design of the actual ART sodium system
was modified so that uniform distribution of sodium
flow would exist in the annulus if there were no
eccentricities, The flow in such a concentric
annulus will be axial; that is, there will be no
spiraling. Studies of the motion pictures of the
flow in the model indicated that, for a constant
pump pressure, the average velocity fluctuation at
any point in the annulus would probably be Iess
than £5%.

The flow distributions in the annulus for two
possible conditions of eccentricity of the core
shell were also studied. Buckling of the core
shell between the spacers was simulated to
produce a local eccentricity. Since the reduction
in flow as a function of the reduction in flow area

‘was the prime concern, the buckling of the core

shell was simulated by adding pieces of tape,
as shown in Fig. 4.1.3. For a locql reduction in
annulus thickness of 50% (¢ = 0.5), the flow

 from _inlet to outlet through that sector of the

annulus was 73% of the average flow.
- Radial eccentricity was also. simulated to study_

L _the.effect_of a _shght canhng of the core sherll on

] ;. TABLE 4.1.1. SUMMARY OF CONCENTRIC ANNULUS FLOW DISTRIBUTION STUDIES

 

: Annulus width ln.

,VT,ReynoIds modulus at mid-plone |

" Ratio of cmnulus flow to cooling hole flow for fwo

S condltlons

For two pumps operotmg

For one pump operating (70% normal flow assumed)

| 0.135  e 0,1875
072 x10° 101 x 10°

1025 1.02
1.075 1407

1.045 1.025
1.08  1.06

 

221

 
 

 

 

‘ANP PROJECT PROGRESS REPORT

| SEOREP
ORNL-ELR-DWG 4740

000000

/-CORESHELL
oo 0

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

TaPE(
' ! CORE SHELL
;‘,:é:a&i___ab\z::mws
o ] REFLEGTOR
A a, |
1.0 ‘
" o
0.8 —
s 0.6 — - - ’/
g /
< 04 - / '
REYNOLDS MODULUS AT
Py MIDPLANE = 1.4 x10*
© ,=0.4875n.
T o
o 0.2 0.4 - 0.6 0.8 1.0
/%

Fig. 4.1.3. Effect of Local Eccentricity on Flow
Distribution.

the flow distribution in the annulus.
displacement of the shell was found to cause a
slight spiraling of the flow through the annulus.
The configuration studied and the experimental
data obtained are shown in Fig. 4.1.4, The data
are in good agreement with a simple, derived
expression based on parallel and equal system
flow resistances. For a mean radial eccentricity
(¢, /1) of 0.8 the maximum deviation of the ratio
Q /Q3 is about 0.8. The static pressures as a
funchon of position in the annulus and in the inlet
header were also obtained for the concentric and
eccentric cases and are being analyzed.

Fuel Flow in Core

Further studies of the flow through the 21-in.

ART core model with the Pratt & Whitney vortex
generators installed in the entrance region were
delayed pending the construction of a second test
stand for the core-model experiments. Studies of
other core configurations that may give stable
flow were, however, initiated. In the preliminary
study the following three configurations are being
considered: a core with an area expansion rate
obtained from Nikuradse'’s gamma function, a

222

Radial

ontrr™
ORNL-LR-DWG t4984.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

b %
N _
° 0 [~
REYNOLDS MODULUS AT s
MIDPLANE = 0.8 x40*% ,,/’
Qn . =0.435in. . // !
.0 —
S ,< —
A e @, £
L~ —= | —
+ 3
0.8 1
0.8 0.9 1.0
W5

an.4.1 4, Effect of Radial Eccenmcuy on Flow
Distribution,

cylindrical annular core with a large area ex-
pansion at the entrance and screens of perforated
plates placed in the region of expansion to obtain
good flow characteristics, and the present 21-in.
core with screens or perforated plates in the
expansion region,

ART CORE HEAT TRANSFER EXPERIMENT

N. D. Greene F. E. Lyf‘lchi
G. W. Greene G. L. Muller
L. D. Palmer H. F. Poppendiek

The instrumentation and operating character-
istics of the ART volume-heot-source experiment,
described previously,®

couples was calibrated. Fifteen complete power
runs were made for the swirl entrance system.

 

N. D. Greene et al,, ANP Quar. Prog. Rep. Dec. - 10,
1955, ORNL-2012, p 174. .

were checked, and the’
- galvanometer ‘used with -the transient thermo-

w
 

*f

"

At

 

The parameters of the experiment were the follow-
ing:
Helical Reynolds modu|us 66,000 to 256,000
Prandtl numbe_r ‘ 4105
Axial temperature rise’ S to 10°F

Total pr'cwv'er generated ' 0.08 to 0.12 Mw

The heat,.bqlan_ces_ obtained for the system were
within £5% of being perfect. Two power runs were

also made with the swirl entrance system for

the simulated case of *‘one pump off.”

A photograph of the one-half-scale model used
for these experiments is presented in Fig. 4.1.5;
it is made almost entirely of Micarta and platinum.
The mixing chamber, pump-scroll head, core,
power leads, and thermocouple leads for the outer
core shell can be identified in the photograph.

The panel board containing recording and power

control equipment for the experiments is shown

in Fig. 4.1.6, and @ view of the fuel annulus

after the pump-scroll head was removed at the
end of the experiments is shown in Fig. 4.1.7.
The platinum—platinum rhodium thermocouples,
as well as the platinum electrodes, which are

visible in Fig. 4.1.7, were in good condition at

the end of the experimental study.

A two-dimensional plot of the electric po-
tential field for the 24-electrode power circuit is
presented in Fig. 4.1.8. Except for the very ends
of the system where some flux distortion exists,

the axial voltage gradient was within dbou} 5.5% ..

of being uniform; consequenfly, the power densfiy
was within about 11% of being wniform. In the

actual ART system in which the heat sources will -
be generated by fission, the volume heat sources
will not ‘be uniform; in fact, near the wall they,-f
will be from 2 to 3 times as great as they will be
near the center of .the channel. The mean, un- -
cooled wail und fluid temperature profiles obtained

in these experlmenfs ‘with @ uniform volume heat
source are .presented in Fig. 4.1.9 in normalized

form, .The asymmetries in the outer core shell and
island core shell wall temperatures can be ex-

plained on'the basis of hydrodynamic flow asymme-

tries. For example, the high island core shell
‘wali temperofure in the northern hemisphere exists-
because @ separation . region completely encom-
- passes the island ‘shell in that  region. “The

solution for an idealized ART (paraliel-plates
system)? is also plotted on Fig. 4.1.9; this
predicted uncooled wall temperature profile lies

PERIOD ENDING JUNE 10, 1956

  
 
   

Poo.o® s
UNCLASSIFIED
‘'PHOTO 25940

x0T ’

¥

i

  

Fig. 4.1.5. One-Half:Scale Model of ART Core

for Yolume-Heat-Source Exp'eriments‘.

between the is!and ond outer core shell wall
temperature measurements.

 

7H, F. Poppendiek and L. D. Palmer, Forced Con-

vection Heat Transfer Between Paralle! Plates and in
Annuli with Volume Heat Sources Within the Fluids,
ORNL-1701 (May 11, 1954).

223

 
 

 

ANP PROJECT PROGRESS REPORT

 

 

 

 

 

 

 

4
a
wmm
‘i
g3 | :
_AT :
‘Ao
M.MH
UP .

- ®
£
=
o

- E
=
®
o
- v
v
s
=2
o
vy =
<
. . -
, %
£
Q
>
|-
O
Yo
- -
-
&
o
T
o
fim r
2
:
o Ba
-
e
o
.Ch
um -
-
S
Q
o
e
- B =
0
-
g
.
. -t
-
2
- F -
£

224
 

 

 

"

)

 

Figo .401070

 

 

Fuel Annblus of the Volume-Heat-

Source Apparatus.

 

 

 

 

 

 

 

 

 

 

 

PERIOD ENDING JUNE 10, 1956

 

 

 

      

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

e . ORNL-LR-DWG 1474{
#= WALL OR FLUID TEMPERATURE
| "%, = INLET MIXED MEAN FLUID TEMPERATURE
r,,, = QUTLET MIXED MEAN FLUID TEMPERATURE )
"2~ 4 « POWER DENSITY 1]
_ oA
ol i
- . - Lt
1 o . [ f———.._;/
8 ISLAND CORE SHELL b |~
S . » / -
< | 2~ 7
0.8 : i
- IDEALIZED ART / L .
K (THEORETICAL ) ;f”>'\
! \ - /7| .~ OUTER CORE SHELL
~ 0.6 : Vs /7 | |
' // ,r’ 7 i I
- » A 7 MIXED MEAN FLUID
” ’
7’
£z
0.4 y
L.
1’ -
0.2 1 A HELICAL Re = 256,000
‘ A /,-/ Pr=4.8
] W =% (UNIFORM)
0 -] | |
0 2 4 6 8 10 12 t4 6 18

AXIAL DISTANCE FROM INLET (in.)

Fig. 4.1.9. Temperature Profiles of the Mean,

‘Uncooled Outer Core Shell, Island Core Shell Wall,

and Fluid of the One-Half-Scale ART Core Model
‘with a Umform Volume Heat Source.

UNCLASSIFIED
PHOTO 26304

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

225

 
 

 

 

ANP PROJECT PROGRESS REPORT

Experimental transient wall and fluid temper-
ature data are shown in Fig. 4.1.10 in terms of the
total temperature fluctuation divided by the axial
temperature rise of the fluid going through the

core. The frequencies of the temperature fluctu-

ations for this one-half-scale volume-heat-source
system vary from about é to 4 cps. The frequency
range of temperature fluctuations in the full-scale
ART system operating with fluoride fuel should be
similar to that in the one-half-scale volume-heat-

 

 

 

 

 

JA V'I \ | ', At
]\/ LJ \J"' { Afm

 

 

ISLAND SHELL AT EQUATOR

 

 

honstii—— § S8 C =]

{.
WS

| QUTER CORE SHELL AT EQUATOR ——

 

 

 

 

 

 

A~ /L\ o

 

 

 

MID-STREAM BEYOND EXIT

 

 

 

 

 

 

 

 

 

0 10 2.0 3.0
TIME (sec)

source system in which.an acid was circulated.
The frequencies of the temperature fluctuations
were found to be in agreement with the frequencies
of the velocity fluctuations observed in the full-
scale core model used for flow studies..

The results of a study of the temperature

‘structure in an idealized ART core were presented

in the previous report.® The increase in 'the

 

8H. F. Poppendiek and L., D. Palmer, ANP Quar '

Prog. Rep. March 10, 1956, p 176.

SENET
ORNL-LR-~DWG 14742

HELICAL Re = 256,000

W=W (UNIFORM)

Afg = TOTAL CHANGE IN WALL TEMPERATURE

Aty = AXIAL FLUID TEMPERATURE RISE IN
GOING THROUGH CORE

-
LS NN 1 B ot

Flg. 4.1.10. Transient Surface and Fluid Temperatures Obta ined fnr ART Core Model wuh a Umform

Yolume Heuf Source.

226

 
 

 

 

 

»

[}

»

-

uncooled wall temperature that would arise if a
hyperbolic cosine power density distribution
existed rather than a uniform one was determined.
The uncooled wall temperature increment, above
the fluid temperatures, for this nonuniform power
density case was found to be more than twice
as great as the corresponding increment for o
uniform power density case. The experimental
measurements obtained for the uniform power
density cose have been modified by this factor
and are plotted in Fig. 4.1.11, along with the
mathematical prediction for an idealized ART
core with a hyperbolic cosine power distribution.
As may be seen, temperatures as high as 1800°F
might occur if the walls are not cooled properly
for this very high Reynolds number case.

Elastic thermal stress calculations were also
made for the core shell and heat exchanger tubes
on the basis of the experimental temperature
fluctuations that were observed in this experiment

SeCRT™
ORNL-LR-DWG 15045

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2.4
2.2
W = POWER DENSITY ANYWHERE IN FUEL CHAMBER
W, = POWER DENSITY IN CENTER OF FUEL CHAMBER
20 Ko = RECIPROCAL OF THERMAL DIFFUSION LENGTH IN FUEL
[T 7= RADIAL DISTANCE FROM CENTERLINE OF CHANNEL,
48 1™ 7= WALL OR FLUID TEMPERATURE
[ fp= INLET MIXED MEAN FLUID TEMPERATURE
1.6 ™ 4, = OUTLET MIXED MEAN FLUID TEMPERATURE
4 ,,/
& . //
1 o
RARE: e
= ISLAND CORE SHELL — "‘\54"' 1
TE w0 . /:/ f~—]
L IDEAUZED ART (THEORETlCAL)-a 77 T _F
oun-:a com-: sneu.---_w T [
0.8 MIXED MEAN FUEL ST
| ""”"?4,( A
: 17
0.6 A
. . /, g /| , _

o4 b 17 AT |/ HELICAL Re = 256,000
v A A - Prw48
T N W = W, cosh K r |
LT =t 1
0.2 ——71P 7 - 1

L~

 

 

 

 

 

 

 

 

 

AXIAL D!STANCE FROM lNLET !il'l)

 Fig. 4111, fefn;pg;u-scre ,mme's of the Mean,
‘Uncooled Outer Core Shell, Island Core Shell Wall,

and Fluid of the One-Half-Scale ART Core Model
Adjusted for the Nonuniform Yolume Heat Source.

o -2 4. 6 -8B 10 12 4 6 18

PERIOD ENDING JUNE 10, 1956

for the uniform—power-density case. For the
uniform-power-density system the results indicated
that cyclic stresses exist that are similar in
magnitude to the endurance limit or fatigue stress
of Inconel. It is believed that the stresses in the
actual ART system may be higher than those
caleulated for the uniform-power-density case.

A third set of volume-heat-source experiments
is under way. For these experiments entrance
vanes that will reduce the rotational velocity
component are located in the core throat.

THERMAL-CYCLING EXPERIMENT
H. W. Hoffman D. P. Gregory?

The volume-heat-source experiments with the
one-half-scale model of the ART core have verified
that the hydrodynamic instabilities that exist in
some regions of the ART core, as presently
designed, would result in rapid, high-temperature-
differential cycling of the Inconel core-shell
surfaces. These studies have also indicated that
the tube bends at the fuel inlet end of the fuel-to-

‘NaK heat exchanger will be subjected to this

temperature cycling. The volume-heat-source
experiments indicate that the temperature fluctu-
ation at the Inconel surface may be of the order of
+60°F. The thermal diffusivity of Inconel is
poor, ond therefore these large temperature fluctu-
ations will occur, chiefly, in a region close to the
metal-fuel interface. For example, the amplitude
of a temperature fluctuation with a Y-sec period
will be reduced to 40% of its origino? value at a

position 35 mils below the metal surface., A
~ possible result of this cyclic thermal expansion
~within the metal will be cracking of the surface
~because of metal fatigue. The possuballtles of
accelerated creep  and corrosxon * under these
" gircumstances also arise. :

_ The effect of thermal cycling on materlul strengfh

_must be determined experimentally, and a system

 for accomplishing this study -at reactor temper-
- .gtures with a fuel environment is currently being -
o consfructed and tested. The ‘apparatus is shown
schemohcally in Fig. 4.1.12.
-~ NaF+ZrF -UF (50-46-4 mole %) will flow through
- the heater under gas pressure and will be subjected
‘to eyclic heating. ‘The surface of the unheated
test section will experience the resultmg periodic

“The fuel mixture -

 

%0n assignment from Pratt & Whitney Aircraft.

227

 
 

 

 

 

ANP PROJECT PROGRESS REPORT

 

  

SCORES
CRNL—LR—DWG 14743

INCONEL TEST SECTION

ey
------
------

 

EXHAUST -

 

 

 

 

 

 
  

 
 

    
 

  
 
 

EXHAUST - :-:‘ .:‘.
ELECTRODES
IN LIQUID TIN B Cswny
b _SOLENOID VALVE -
. POWER IN . g :
SUPPLY \ BELLOWS SUPPLY
TANK NO. 1 TANK NO. 2

 

 

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

 

1. FUEL MIXTURE

 

 

 

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

   

 

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

 
 

 

 

 

 

 

 

 

 

 

 

BEAM
SUPPORT

 

 

 

WEIGH BEAM

 

MICROFORMER

Fig. 4.1.12. System for Thermal Cycling of Inconel Tubing in a Fuel Environment.

temperature fluctuations as the fluid moves down-
stream. Effects of the thermal cycling will be
determined by metallographic examination of the
Inconel surface subsequent to the test.

. The heater and the test section are fabricated
from a single 10-in, length of Inconel tubing
(/-ln. OD, 0.035-in. wall). The heater comprises
the first 4 in. of this tube and will be supplied
with a pulsed current through a device designed to
provide a range of pulse frequencies. In order
that the stresses imposed on the test section will
result only from the thermal cycling of the tube
surface, the heater electrodes are floated in
pools of molten tin to eliminate tube bending, and

228

a bellows is provided at the outlet end of the test
section to take up the over-all thermal expansion,
Preliminary tests with NaNO,-NaNO,-KNO
(40-7-53 wt %) as the heated fluid were termmatej
when oxidation at the electrodes resulted in loss
of power. This situation is now being corrected
The results of these first experiments for a 11 /-sec
power-on — 1V-sec power-off cycle are glven in
Fig. 4.1.13. 4T|1e amplitude of the temperature
fluctuation at two points in the test section is
shown as a function of the heater power., The
““tailing-off’’ of the data at high heater powers is
caused by the slow response of the measuring
equipment. The cycle perlod will be varled in

future experiments,
 

 

 

 

Ty

"

-

TEMPERATURE AMPLITUDE AT OUTER WALL OF TEST SECTION {°F)
. \
o

 

SHIELD MOCKUP REA.CI:OR STUDY
L. C. Palmer '

A general study of the heat tronsfer and fluid
flow characteristics of the proposed Shield Mockup
Reactor (see Chap. 5.5, ‘‘Shield Mockup Reactor
Design and Construction’’) has been initiated. An
anclysis of the temperature structure within the
fuel elements and the coolant and a study of the
pressure distribution within the core were com-

UNCLASSIFIED

 

 

 

 

 

 

 

 

 

ORNL~LR—DWG 14744
90
/
80 / iR
® %4 in. FROM TEST SECTION ENTRANCE // o
o 3% in. FROM TEST SECTION ENTRANCE ;/ ¢ /
TO /ll [

A/
/ /]
/| £

 

o
o

 

8

 

 

N

%
T

10 20 30 40 50 60 7O 80 90 {00
HEATER POWER (%) : :

 

8

 

i
o

 

 

 

 

 

 

 

 

 

 

 

 

o
o

Fig. 4.1,13. Temperature Amplitudes Observed
in Preliminary Thermal-Cycling Tests of Inconel
Tubing.

PERIOD ENDING JUNE 10, 1956

pleted. .The studies are being extended to other
regions of the reactor,

TEMPERATUR_E smucfuns IN THE REGION
BEYOND THE ART REFLECTOR

H. W. Hoffman

The ‘previous calculations19 on the temperature
structure in the region beyond the reflector in the
ART were extended to include a recent system
modification. The new system, illustrated in
Fig. 4.1.14 and designated System C, differs from
those previously studied (Systems A and B)!? in
that a portion of the boron carbide has been
replaced by a stainless-steel-clad copper~boron
carbide matrix. This layer, located between the
reflector and the remaining B C, will absorb the
leakage neutrons from the core and thus prevent

-radiation damage of the more brittle boron carbide.

A detailed description of the idealized system
used in this study has been presented in a separate
report.1! Based on these idealizations the system
can be analyzed by an iterative procedure to
obtain both the temperature rise of the sedium and
the temperature profile for the region between
the sodium and fuel return streams. The heat
deposition rates used in this analysis were tenta-
tive upper limits. 12

 

104, w, Hoffman, J. L. Wantland, and C. M. Copen-
haven, ANP Quar, Prog, Rep., March 10, 1956, ORNL-
2061. p 174.

Ty, W, Hoffman, Thermal Structure for the Region
Beyond the ART Reflector — Supplement I, ORNL
CF-56-4-I29 (April 17, 1956).

2H. W, Bertini, personal communication to H, W,
Hoffman,

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

SEoREeT.
ORNL—LR—DWG 13335A
E 4. |8 c D E Fle|H I Jikl] ¢ M
I 11 =z o 12 o
l s wl |8 |y |k x
!5 |z bl o ol sk = alsl & | |
3 1= INCONEL  |@la] S o |o|812(8 BORON CARBIDE'. |&|2] 2 ]
e |8 a9l wa Qi Wil O T
ow | @ z[° &°© |9ziT|=Z zZ| Tl £ |
1 = 8 ge =
| v w wn v
|
X % X% X% X % X Xg X X R A T &
SYSTEM C

Fig. 4.1.14. Schematic Diagram of Region Beyond the ART Reflector.

229

 
 

 

 

 

 

ANP PROJECT PROGRESS REPORT

The recent system modification resulted in a
temperature rise of 33°F for the sodium coolant
stream and a maximum boron carbide temperature
of 1526°F for a sodiuminlet temperature of 1200°F.
These values may be compared with the previous
results for a system without the copper—boron
carbide matrix in which the sodium femperature
rise was 24°F and the boron carbide attained a
maximum temperature of 1586°F. The sodium
temperature rise was found to vary 10% (decreasing
from 36 to 33°F) when the sodium inlet temper-
ature was changed from 1150 to 1200°F. Typical
temperature profiles at the inlet and outlet ends
of the system for a sodium inlet temperajure of
1175°F are presented in Fig. 4.1.15, and the
temperature profiles for the system without the
Cu-B ,C (System B) are alsc shown for comparison.

 

Be, 4. -Barton, Fused Salt Compositions, ORNL
CF+55-9-78 (Sept. 16, 1955).

 

NaF-LiF-ZrF  (22-55-23 mole %)

Beta form (106 to 368°C)

Hy = Hygoe

The system medification increased the -temper-
ature drop across the inner Inconel shell by 30 to
35°F. The temperature gradient in the outer
inconel shell was essentially unchanged. -

- HEAT CAPACITY
W. D. Powers

The enthalpies and heat .capacities of two
mixtures were determined in the liquid and solid
states. For one mixture, NaF-LiF-ZrF 4 (22-55-23
mole %), two distinct discontinuities in the temper-
ature-enthalpy relationship were found; one was at
370 to 380°C and the other was at 565 to 585°C.
The reported!® liquidus temperature is 570°C.
The lower discontinuity is assumed to be a phase
transition from the alpha form to the beta form, the
beta form being stable below 375°C. The dis-
continuity at the higher temperature is the fusion
of the alpha form to the liquid at 570°C. Enthalpy -
and heat capacity measurements of a second salt
mixture, LiF-NaF (60-40 mole %), showed no
unusual characteristics. The enthalpy and heat
capacity equations for these two salts follow:

= ~8.7 + 0.2392T + (7.20 x 10~5)12

c, = 0.2392 + (14.39 x 10=3)T

Alpha form (407 to 556°C)

H., - H30°

T
“p

Liquid (603 to 897°C)
H,. -~ H

30°C

€p

T

At 375°C
H, -~ Hg = 28

At 570°C
Hyo = Hy = 24
LiF-NaF (60-40 mole %)

Solid (112 to 572°C)

H - H

T 30°

c = 642 — 0.009165T + (41.82 x 10~
= ~0.,009165 + (83.64 x 10~3)T

~20.2 + 0.4526T - (5.95 x 10~-5)12?
0.4526 - (11.89 x 10=5)T

c = =9.8 + 03191T + (9.94 x 10~5)T2

C, = 0.3191 - (19.87 x 10~3)T

230

 
 

 

 

 

.}

M

Liquid (688 to 896°C)

PERIOD ENDING JUNE 10, 1956

Hy = Hygoo = —882 + 0.9249T — (24,62 x 10-5)72
c, = 0.9249 - (49.23 x 10°°)T

At 652°C

Hyjq = Hyg = 170
In these expressions
H = enthalpy in cal/g,
= heat capacity in cal/g+°C,
= temperature in °C,

VISCOSITY AND DENSITY
'S. . Cohen |

A study was made of the effect of fission
products on the viscosity of a fused fluoride
reactor fuel. The increase in viscosity as a
result of the presence of dissolved fission-product
additives was found to be less than 5%. The
formula for the fission-product additives was,
however, based on a much higher burnup rate than
would actually be encountered in the ART. The
formula called for an amount of LaF, (the com-
ponent used to representall the rare-earth fluorides
formed by fission) that exceeded the solubility
limit in the fuel ot temperatures below 800°C. ‘It

is felt that the only serious possible problem that -
might arise would be the precipitation of solids.
This problem is being investigated in solubility |

studies on fission products in the reactor fuel
(see Chap. 2.3, ‘‘Chemical Reactions in Molten

Salts’’).
Interest in the fuel mixture NaF-KF-LiF-UF

(H 2-4]-45.3-2 5 mole- %) prompted a redetermi- .
nation of its viscosity, since the previous measure-
ments 14 were made on a relahvely impure sample S
- plummet. The ‘density varied from 1.675 g/cm3
“ at 425°C to 1.50 g/cm .at 770°C and may be

- expressed as

and over o' narrow temperature range. The vise-

- f'cosnty of a h:ghly pure sumple prepared for . these
Lol measurements varied from 8.1 centipoises at about
o 525°C fo 2.7 centiponses at about 725°C and moy— ,
be represented by the equehon o , '

# - 00292 e4.‘507/'1" '

o :menis are about 10% lower than the prewous ones.

 

V45, |, Cohen, ANP Quar. Prog. Rep. Sept. 10, 1955,
ORNL-1947, p 156.

 

Measurements were also made'3 on NaF-LiF-KF-
UF, (10.9-44.5-43.5-1.1 mole %). The viscosity
vaned from 8.8 centipoises at about 500°C to
1.8 centipoises at about 800°C and may be
represented throughout this range by the equation

p o= 0.0348 e4265/T

where T is in °K,

A modest program has been initiated to measure
the physical properties of some fused chloride
mixtures. Two instruments have been designed
to study the viscosities of these materials. One
is a special Brookfield viscometer, which has a

spindle assembly that contains two large shear

surfaces.  This increased shear is necessary
because of the low viscosity range in which this
class of materials falls. The other instrument is an

‘adaptation of the capillary viscometer currently
being used for fluorides. The principal modifi-

cation is the use of commercial hypodermic needles
for the capillary tubes. This affords a number of
advantages, such as ease of replocement and
interchangeability of needle sizes to cope with
different viscosity ranges.

‘A density measurement was made on one chloride

mixture, LiCl-KCI-NaCl (56-4]-3 mole %), by using

the buoyancy principle with ‘a balance and «

p = 1.885 ~ o.oooso:r .

~ where T is in °,_C. These values are in excellent
. agreement with values obtained on very similar
S : " mixtures by Van Artsdalen and Yaffe.16
o where T is in °K These vclues are: m sahsfoctory-- o T T Ty -
’ A’-_ngreement with the previous data; the new. measure-

 

'-.7155. l. Cohen and T. N. JoneS,r Measurement of the

- Viscosity of Compositions 12, 14 and 107, ORNL

CF+56-5-33 (May 9, 1956).
16E, R. Van Artsdalen and 1. S. Yaffe, J. Phys,
Chem, 59, 118-27 (1955).

231

 
 

et

1650

1600

1550

1500

1450

1400

TEMPERATURE (°F)

1350

1300

1250

i2o0

1150

€0

SYSTEM | :
INCONEL - CARBIDE

SYSTEM C, ExiT NORTHERN HEMISPK

-
-

NLESS STEEL STIAINLESS STEEL

COPPER-- YSTEM €
STAINLESS STEEL BORON CARBIDE S -
INCONEL CARBIDE
HELIUM |

50 40 30 . - 20 10

DISTANCE FROM INCONEL—FUEL INTERFACE (ft x 10™%)

 

=LEORES
QORNL—LR~-DWG 13336A

INCO

INCONEL

 

1¥0d3Y SSIYD0Yd LDIrONd dNV

Fig. 4.1.15, Compurison of Temperature Proflle? in Region Beyond the ART Reflector for Various Systems and a Sodium inlet Tempera-
ture of 1175°F, ' . ' .

(\

 

 
 

 

 

o

THERMAL CONDUCTIVITY
W. D. Powers

Two new devices have been built that are being
used to determine ‘the conductivities of liquids.
in both systems a measured amount of heat flows
down through a horizontal sample of known thick-
ness and area. The thickness of the sample may

be varied so that contact resistances at the

surfaces of the sample can be eliminated.’

One device has been fabricated in which the
sample is contained within glass. This makes it
possible to observe whether any gas bubbles
or films are present in the cell at the time measure-
ments are being made. Good agreement with the
known thermaleconductivity data for water was
obtained. Initial experiments with heat-transfer
salts have shown the formation of gas bubbles on
the surfaces. The effect of this gas on the thermal
conductivity is currently being studied..

One - of the difficulties associated . with the
measurement of thermal conductivity is in knowing

PERIOD ENDING JUNE 10, 1956

-exactly where the heat is flowing. The second

device has two heat meters, one directly above
and the other directly below the sample. It was
assumed that if the heat flow were the same
in both meters, the heat flow through the sample
would be well established. The original design
did not meet this specification and therefore o
modification of the system is now being made.

ELECTRICAL CONDUCTIVITY
N. D. Greene

An electrical conductivity cell has been standard-
ized at high temperatures with molten potassium
chloride, and the cell constant so obtained agreed
within experimental etror with those determined
by means of standard, aqueous solutions. Further
cell standardization with molten NaCl and ZnCIz,
as well as refinements of existing measuring
techniques, will precede measurements of the
conductivities of fluoride melts.

233

 
 

 

 

 

 

ANP PROJECT PROGRESS REPORT

42, RADIATION DAMAGE S U
G, W. Keilholtz ' S . : o

EXAMINATION OF THE DISASSEMBLED MTR 42.1 and 4.2, 2. The ;ear"beqrmgs,i'Fag: 423,

IN.PILE LOOF NO. 3 appeared to be i |n excellent condition, ¢
A. E. Richt The neck of fhe pump, between fhe pump sumpri
C. Ellis W. B. Parsley and the heat exchonger, was parhully filled with
E. J. Manthes R. N. Ramsey ' oil; however, no- ewdences of fuel or ZtF =vopor ..
E. D. Sims deposits were found. in this region. The slmger___

assembly was “removed | cnd ‘examined, and the

- . slingers were found to be coated with a black,
No. 3 has been completed except for the straight shiny deposnt th. . 24 No ewdences of fuel

sections of the fuel line. Disassembly of the fuel- Z F de ' f d'on the slingers.
circulating pump. was found to be more difficult or &r -vqpor eposu s were foun on .e 'gers. S e

than had been anticipated. The first operation
was the removal of the water jacket from the pump
bulkhead. The pump was then sectioned by sawing
through the bulkhead. During this cutting operation
the oil finger below the forward bellows was
exposed, and no evidence of oil was seen in the
finger, The pump shaft rotated freely and showed
no signs of binding after it was freed from the
impeller. The pump shaft, forward bellows, bear-
ings, and seals were removed from the bearing
housing. The forward bellows was partially filled
with oil. The bellows, bearings, and seals were
cleaned ultrasonically in perchloroethylene to re-
move any traces of oil. The face and inside of
the forward bellows contained a brittle, amber- o o R
colored, amorphous deposit, as shown in Figs. Fig. 4.2.2. Enlorgement of Depesit Shown in

Fig. 4.2.1. 2X. Reduced 25.5%,
RMG 1380 _

Metallographic examination .of MTR in-pile loop

 

 

x

Flg. 4.2.1. Front Face of Forward Bellows of Fig. 4.2.3. Rear Bearmg from MTR In-Plle Loop B _
Fuel Pump from MTR In.Pile Loop No. 3. 1/2X.  No. 3 Fuel Pump. 2X Reduced 14%. (Secref w:ih Lt
(Secretwith-captian) caption) | R T u

234

 
 

 

9

»

B UNCLASSIFIED
RMG 1383

 

Fig.4.2.4. Slingers from MTR In-Pile Loop No. 3
Fuel Pump. Note shiny, black deposit. 2X. Re-
duced 24.5%. (Seere-f—m-fh—ea-p-h-on-)—

UNCL ASSIFIED)
- RMG 1384

 

.F‘\ig. .2 5 Impeller (Top) und Shnger (Bofl'om)
Halves of Pump Sump from MTR ln-Pile Loop No. 3
Fuel Pump. 1/2X. Reduced 23.5%. (Seeret.with

~ssption}

"PERIOD ENDING JUNE 10, 1956

The fuel in the pump appeared to be inhomo-
geneous -in that it was made up of two phases.
One phase was green and had the normal appear-
ance of the ‘fuel mixture used, and the other phase
was black. Samples of both phases were obtained
for chemical analyses. The two halves of the
pump sump, with the fuel still present, are shown
in Fig. 4.2.5. The impeller showed no evidence
of erosion, Figs. 4.2,6 and 4.2.7.

Charcoal samples were obtained from two of the
fission-gas adsorption traps in the purge system,
A black deposit that was found around the inlet

UNCLASSIFIED
-RMG 1384

 

Fi'g'. 4.2.6. Front Face of Impeller from MTR In-
Pile Loop No, 3 Fuel Pump, 1/2X. Reduced 15%.

 

Fig. 4.2.7. Side View of |mpe||ef from MTR In-
Pile Loop No. 3 Fuel Pump. 2X. Reduced 23,5%.
(s Ly trom}

235

 
 

 

 

ANP PROJECT PROGRESS REPORT

hole to one of the traps is shown in Figs. 4.2.8 -

and 4,2.9. The walls of the trap on the inlet side
also showed a black deposit around the sides for
a depth of approximately 1 in., as shown in Fig.

4,2,10.

MTR in-pile loop No. 4 has been sectioned at-

NRTS, and the active portion is being shipped to
Oak Ridge. Disassembly of the loop will begin
as soon as the hot cell equipment has been re-
paired and checked out. Priorities have been
established for the examination of sections that
were removed from the ARE, and these exami-

UNCLASSIFIED
RMG 1388

 

Fig. 4.2.8. Inlet Side of Fission-Gas Adsorption

Trap from MTR In-Pile Loop No. 3. 1/2X. Re-

duced 15%. (S=cret-withocgption)

 

Fig- 4,29, Enlargement of Deposit Shown in
Flgo 4. 080 2x Reduced 28 5%.

236

previously,?

 

‘nations will proceed as rapldly as cell tlme_‘

ullows.. _ _ _
'mvssflénldfis OF MATERIALS REMOVED
~ FROMMTR IN-PILE LOOPS NOS. 3 AND 4
W E. Browmng
C C. Bolta R P Shlelds

Operctlon of MTR |n-p||e |ooP No. 3 descrlbed. ,
was terminated beCQUSe ~a plug

formed somewhere in the purge system. Therefore
a postoperational inspection of the fission-gas
adsorption traps was undertaken. A ‘gas sample
was obtained by purging the traps for 2 hr with

helium and collecting the effluent fission gas in

two refrigerated charcoal trops. A radioassay of
the gas is to be made. The purged traps were
sectioned, and samples of the charcoal adsorber
and tubing were collected for radicassay. An

~extensive black deposit was observed around the
“inlet to one trap, and a thin brownish film was -

present around the inlet to the second trap. These
deposits were photographed (Figs. 4.2,8, 4.2.9,
and 4.2.10) and then collected for analysis. The
resylts of the analyses of the samples are not yet

available, but it is suspected that oil or decompo-

sition products of oil entered the traps from the
fuel-pump bearing housmg ‘ond sump ond caused

 

lOn assignment from Pratt & Whmiey Alrcroft

2. P, Carpenter et al.,, ANP Quar. Prog. Rep. Dec.
10, 1955, ORNL-2012, p 27.

  

 

Fig. 4.2.10. Inner Walls of Inlet Side of Fission-
Gas Adsorption Trap from MTR In-Pile Loop No. 3

Showing Depth of Formation of Black: Deposlt.

1/2X. Reduced 25.5%. (bevret-witircopticnl
 

 

 

 

1

al

.

»

"

 

a plug when they came in contact with the refr:ger-
ated trap inlet.

A study of an apparent chemical change brought
about in the fuel during operation of MTR in-pile
loop No. 3 is also being conducted. Determinations
ore being made of the nickel, chromium, and iron
content of the fuel, and radiochemical analyses
of the fuel are nearly complete. The analysis of
the foreign materials taken from the pump is
partially completed.

The fission-gas adsorption traps from MTR in-
pile loop No. 4 have been received, and analyses
of copper tubing sections and of acetone washes
of the carbon adsorber are under way.

CREEP AND STRESS.CORROSION TESTS

J. C. Wilson |
W. W. Davis A. S, Olson
N. E. Hinkle J. C. Zukas

A heat transfer test of the stress-corrosion appa-
ratus designed® for operation in HB-3 of the LITR
was made in the LITR to check the heating to be
expected from the fuel, It was not possible to
reach the desired temperature of 1500°F with the

2.5 g of the fuel mixture (No. 41) NaF-ZrF -UF,

(63-25-12 mole %) used in this test. A bench
mockup which reached the same temperature as
that attained in the reactor indicated that only
about 70 w of heat was generated by fission.
About 200 w of heating was expected from this
fuel at the depressed flux measured in -HB-3 of
the LITR. For the next test a fused salt with
more uranium will be used, or the number of con-
ductive fins will be reduced. - ' '

Three more’ fube—burst stress-corrosnon speci-
mens were exposed to radiation in the LITR, and

two more bench tests were completed. The irrodi-

‘ated specimens_ fmled after 164 1260, and’ 285 hr

‘at 1500°F with: ‘a maximum stress of 2000 psi, .

- whereas one specimen bench-tested under similar

. test’ condmons ‘had not .failed in about. 935 hr.

" The other bench test was made at 1450°F, 50 deg

‘_'—-:iower, and the speccmen rupfured ofter only 500 he

. under stress, . _

- “Tubing stock’ of the type fo be used in ART' B

'NaK-to-cnr radmtors is- expected to be available
" 'soon ‘for “use in tube-burst and stress-corrosion

 

3J. C. Wilson et al.,, ANP Quar. Prog Rep. Sept. 10,
1955, ORNL-1947, p 165.

PERIOD ENDING JUNE 10, 1956

experiments. Components for the MTR tube-burst
test apparatus are being fabricated.

One of the specimens tested previously in the
MTR tensile-creep-test apparatus? ruptured outside
the gage length, and therefore the postirradiation
elongation measurement was made on this speci-
men. The elongation was 1.7% after 760 hr ot
1500°F in helium; this value is higher than was
expected. The corresponding bench tests have
still not been completed, because experimental
difficulties (mainly lecks in the bellows) required
that the apparatus be rebuilt.

A pneumatic stressing device for bench and in-
pile tensile-creep tests was built and is being
leok tested. Two pneumatic extensometers for
use with this device have been tested. The
pneumatic device will provide a powerful, compact
means for stressing creep specimens with loads
as great as 10,000 Ib,

Parts for a stress-relaxation machine are being
built, A pneumatic valve operator has been adapted
for stressing the specimen, and a pneumatic strain-
controlling system is being developed.

MTR STRESS-CORROSION APPARATUS
J. C. Wilson
C. D. Baumann W. E. Brundage

The stress-corrosion equipment formerly desig-
noted as the alternate stress-corrosion apparatys
was modified to change the platinum heater to a
simpler, helically wound heater made from molyb-
denum wire. . The design, described previously,’
is completed, and fabrication and assembly have
been 'parnally completed on a model with which
heat fransfer tests w:ll be mcde in the LITR,

'f DUCT’LITY OF NICKEL BASE ALLOYS -
J. C. Wllson "T.C. Prtce

. An mveshgaflon is . belng made of the factors

\?that govern the ductility of mckel-bcse alloys.
- As a part of this study, tensxle tests were run on

Inconel .and Nuchrome V at strain rates of 0,002
and 0.2 in./min ot temperatures ranging from 500

to 1400°F At. no _temperature was any | obrupt

 

‘w W. Dovns, N. E. Hmkle, and J C. WI|son, ANP.

'Quar. Prog. Rep. March 10, 1956, DRNL-2061, p 190,

Sw. W, Davns, N. E, Hinkle, and J. C, Wilson, ANP
Quar. Prog. Rep March 10, 1956, ORNL.2061, Fig.
8.12, p 191.

50n assignment from Prufl & Whutney Aircraft,

237

 
 

 

 

 

ANP PROJECT PROGRESS REPORT

‘decrease in ductility observed in either alloy.
The ductility of Inconel remained constant between
800 and 1100°F at the slower strain rate and
between 800 and 1200°F at the faster strain rate.
- At higher temperatures the ductility steadily in-
creased. The ductility of Nichrome V decreased
steadily as the temperature was increased from
800 to 1400°F. _

Both alloys showed stress fluctuations at both

strain rates, the fluctuations being more noticeable
at the slower strain rate, but this may have been
the result of the inadequacy of the recording
system at.-high speeds. In the tests on Inconel
the stress fluctuations were present at tempera-
tures from 500 to 1200°F and were largest in
number and amplitude between 600 and 950°F.
The lower temperature limit of these fluctuations
has not yet been determined for Nichrome V, but
the upper limit seems to coincide with that of
Inconel., _ -
. Specimens of Monel, Inconel, and Nichrome V
have been prepared for testing in the LITR (and
later in the MTR) to determine the effect of irradi-
ation on ductility. Out-of-plle tests will be made
on the Hastelloys.

EFFECT OF RADIATION ON STATIC
CORROSION OF STRUCTURAL MATERIALS
BY FUSED-SALT FUELS

H. L. Hemphill

The study of the effect of radiation on the cor-
rosion of Inconel capsules containing static fused-
salt fuels was continued.” Two additional Inconel
capsules containing the fuel mixture (No. 30)
NaF-ZrF -UF, (50-46-4 mole %) were irradiated
in the MTR for six weeks each at ]500°F at a
power density of approximately 3.5 kw/ cm3. Two
other Inconel capsules containing the fuel mixture
(No. 44) NaF-ZfF -UF, (53.5-40-6.5 mole %) are
being irradiated at 1500°F at a power density of
approximately 6 kw/em3.

Further mockup tests were run on Hastelloy B
capsules under simulated MTR conditions. A
method was developed for mounting thermocouples
on the capsules without domaging the chromium-
nickel plating required to protect the capsule from
the atmosphere. Hastelloy B capsules are being
prepared for bench tests and for MTR irradiation.

W. E. Browning -

 

W, E. Browning and H. L. Hemphill, ANP Quar. Prog.
Rep. March 10, 1956, ORNL.2061, p 192.

238

They will be filled with an NoF-KF-LiF-UF, fuel
mixture containing approximately 25 wt % enriched
vranium and LiF enriched in Li”.

LITR VERTICAL IN- PILE LQOP
w. E. Browmng

D. E. Guss® H. E. Robertson
M. F. Osborne R. P. Shields

A vertical in-pile loop, which was eSsenhully
the same as the one operated prevuously, was
installed in the LITR and put into operation; how-
ever, the pump stalled when the reactor was
started. By reducing the pump temperature it was
possible to operate the loop with the reactor at
500 kw; but at higher power levels the temperature
drop in the loop was great enough to freeze the
fuel in the cold leg. The pump stalled permanently
during an effort to increase the pump temperature

‘gradually so that the reactor could be brought to

The specifications for the power
total power,

full power.
characteristics of the loop were:
1.5 kw; maximum specific power, 577 w/em’;
dilution factor, 7.64.

While the cause of the pump failure cannot be
precisely determined until the loop has been
examined, it is known, from the sounds emanating
from the pump and heard through a microphone,
and from varigtions in electric power demands of

the motor, that contact of moving solid parts

overloaded the motor,

The parts that had already been fabricated for
three additional loops will be changed to eliminate
most of the possible couses of pump failure. The
changes will include a shorter, more rigid impeller
shaft and improved bearings. Also, pump models
are being tested to .determine the effect of in-
creasing the cleorance around the pump impeller.
When these tests are. completed, existing parts
will be modified and assembled for bench and in-
pile tests.

Although failure of the pump prevented operatlon
of the loop, the experiment provided a test for the
reliability of other components. Thermocouples,
heaters, safety circuits, the temperature control
system, the tachometer, the lead wires, aond the
entire instrumentation system performed satis-
tactorily. Piatinum—platinum-rhodium thermo-

 

80n assignment from United States Air Force.

9G. W. Keilholtz et al., ANP Quar, Prog.. Rep. Sept.
10, 1955, ORNL-1947, p 164.

 
 

o).

.

"

<

 

couples were used on this loop to eliminate the

severe oxidation that occurred in previous loop
assemblies where the Chromel-Alumel thermocouple
leads passed through heaters.. The change in
thermocouple materials necessitated a redetermi-
nation of thermocouple corrections.

Three conditions contributing to errors in temper-
ature measurement were investigated for the
platinum—platinum-rhodium thermocouples used.
First, the thermoelectric output of the couples
might be changed because of incompatibility with
Inconel; second, the junction of the thermocouple
is cooler than the Inconel surface to which it is
attached because of air cooling; third, the outside
surface of the Inconel is cooler than the surface
in contact with fuel because of the heat flux
through the wall.

A compatibility test was run in which the plati«
num—platinum-rhodium thermocouples were welded
to Inconel and heated for 1000 hr ot 900°C. The
thermocouples were found to be mechanically
sound after this treatment, and their absolute
calibration has changed not more than 1,2°C,

Tests were also run to determine the temperature
correction necessary because of air cooling of the
thermocouple kead. A segment of the loop was
mocked up in full scale, with electrical heating

. TEMPERATURE DIFFERENCE BETWEEN
THERMOCOUPLE JUNCTION AND FUEL TUBE SURFACE (°C)

PERIOD ENDING JUNE 10, 1956

instead of fission heating, and the thermocouples
were mounted on the fuel tube in the same way
that they ore mounted on the loop. Surface temper-
atures were observed, with an optical pyrometer,
through a quartz window ond compared with the
thermocouple readings. These tests were similar
to those reported previously,? which yielded
different results for different thermocouple ma-
terials. The differences are no doubt due to bead
geometry differences that result from melting-point
differences during thermocouple fabrication. The
results of these tests are shown for six thermo-
couple assemblies in Fig. 4.2.11. For the oper-
ating conditions of the LITR vertical in-pile loop
the correction amounts to 15 % 10°C,

The correction for the differential temperature
through the wall of the loop was calculated by
vsing theoretical heat tronsfer relationships in
the loop and experimental flux data obtained from
previous loop operation in the LITR.'! The heat

 

 

'Ala FLOW (efm)

‘°w E. Browning, G. W. Keilholtz, and H. L.
br!“" ANP Quar. Prog. Rep. March 10, 1955,
L-1864, p 146.

"G W. Keilholtz et al., ANP Quar. Prog. Rep.
Sept. 10, 1955, ORNL-1947, p 164.

UNCLASSIFIED
CRNL~LR-DWG {4746

66 S 88 n Lo 40 .

.Fig. 42,11, Temperature Diflerences Between Six Thermocouple Junctions and Outer Surface of
Inconel Fuel Tube as a Result of Air Flow Around the Thermocouple Junction,

239

 
 

 

 

ANP PROJECT PROGRESS REPORT

transfer calculations of Robinson and Weekes!2
were modified to apply to the presently planned
position for the loop in the LITR by using certain
simplifying assumptions. The tip of the loop is
to be 1.5 in. below the center line of the reactor.
‘This is more than 4.5 in. below the position of
maximum flux. The flux measured near the fuel
tube in the previously operated loop'! was taken
as the depressed flux for that loop. 1t was assumed
that the flux depression was the same for corre-
sponding parts of the present loop. The difference
in shape of the flux profile for the two loop po-

‘sitions’ was ignored in calculating the fuel tube

wall temperature - differential.  Fission-product
poison contributions to the macroscopic cross
section of the fuel were assumed to be negligible.

 

' 12M T. Robmson and D W, Weekes, Design Calcu-
lations for a Miniature Hng-Temperatw'e In.Pile Circu.

'latmg Fuel Loop. ORNL-1808 (Sept. 19, 1955).

The transmission coefficient for neutrons in the
fuel was determined from work reported by Holmes, 13
and it was used in conjunction with the flux distri-
bution along the loop, as shown in Fig. 42 12
to determine the total loop power.

Finally, to find the temperature dn‘ferenhal
through the Inconel fuel tube wall, the curves of
heat transfer variations along the loop designated
Mark VIIi in the work of Robinson and Weekes'?
were modified for the higher total power of the
loop being investigated. A plot of the calculated
temperature differential through the tube wall vs
the position on the loop for proposed operating
conditions is presented in Fig, 4.2.13. The maxi-
mum combined temperature corrections for all three
errors totaled 35°C.

>

 

13p, k. Holmes, Problems of Neutron Population in
Localized Absorbers, ORNL CF-56-1-141 (Jdn. 27,

- 1956).

 

 

 

 

 

 

 

 

 

 

 

SL bt BN Lldre
ORNL-LR-DWG 14747

3
n
o
= 2
%
-J
-
<L
2z
w
|
=
-
-
|
2
w
- 4
I
x|
2
|
w

o - o

0 50 100 150 200 ' 250

DISTANCE ALONG LOOP FROM INLET TO COOLED SECTION {cm)

Fig. 4.2.12. Calculated Profile of Flux at Fuel Tube Surface Along the LITR Vertical In-Pile qup.

240
 

 

 

 

]

)

"

 

KEaETT™
ORNL-LR~DWG 14748

 

30

REYNOI|_DS I'NIO. OI" COOLINGIAIR: 20,000
REYNOLDS NO.OF FUEL: 3000

 

/
S

N

I- f02.5cm (TiIP OF LOCP)

 

TEMPERATURE DIFFERENTIAL THROUGH FUEL TUBE WALL {°C)
o

 

 

 

 

 

 

 

 

 

 

 

 

0

 

O 20 40 60 80 100 120 440 160 {80 200 220
DISTANCE ALONG LOOP FROM INLET TO COOLED SECTION {cm)

Fig. 4.2.13. Plot of Calculated Temperature
Differential Through the Fuel Tube Wall as a
Function of the Distance from the Fuel Inlet of
LITR Vertical In-Pile Loop. ‘

FURTHER DESIGN CALCULATIONS FOR
LITR VERTICAL IN-PILE LOOP

M T. Robmson T
F P Green J. F. Krouse

The method descnbed ;:orevlously12 hcls been_t ?
- “used fo estimate coolmg—alr requirements’ and fuel-
: _tempemiure profiles for a new conflgurutlon of fhe :

miniature - vertical in-pile “loop. ~ The Ioop is

- physically |denhcai to the Mark Vill loop described
s "prewously, '2 but it is to be inserted to the bottom
.. ‘of the core. of posmon C-46 of the LITR instead
o of reochmg ‘only ‘to "the" “position of ‘maximum |
__fthermal-neufron Hux.. - The thermai-neufron flux
“used ‘was’ that measured's ‘in position C-46 and -

_'extrapolated to the bottom of fhe lotflce by re-
A S , L EE Roblnson, March 16, 1956, and May 4, 1956.

 

o “On as#nghrhenf >from Pl‘flff&VWh‘l.fne.i Alrc}uff.

By T Robinson, Solid State Semiann. Prog. Rep.
Feb, 28, 1954, ORNL-1677, p 27.

PERIOD ENDING JUNE 10, 1956

flection of the measured flux in the plane of its
maximum value.  The computations were carried
out on the Reactor Controls Computer according

to the technique previously described,® except

that the flux function was generated electronicully
instead of mechanically. :
Calculations were made for the fuel mixture

(No. 44) NaF-ZrFl-UF‘ (53.5-40-6.5 mole %) con-

“tained in Inconel in the new position and for the

fuel mixture (No. 107) NaF.KF-LiF- UF (11.2-
41-45.3-2,5 mole %) contained in Hasfelloy B in
both positions. The nhysical property data used

for air, for Inconel, and for the zirconium-bearing

fuel mixture were the same as those used previ-
ously. The data for the alkali-metal fuel and
Hastelloy B are given below:

Density of fuel,]7 g/cm3 2.09
Specific heat of fuel, '8 wesec/g:°C 2.28
" Yiscosity of fuel, 19 g/cmesec 0.022
Thermal conductivity of fuel,zo 0.035
w/em°C
Thermal conductivity of Hasteiloy B, 21 0.113
w/em«°C '

As before, the depression of the thermal-neutron
flux because of the presence of the loop was

 neglected.

The results of the calculations are summarized
in Table 4.2.1. It is shown that, in spite of the
greater length of the irradiated loop in the ‘“‘deep’’
position, the fuel temperature differential is less
than that in the position of maximum flux. The
principal result of the change in position of the

loop is ' a substantidl -'decreGSe in the dilution

 

g R, Mcmn, F. P, Green and R. s Storie, **An

- Appendix on Anglog Simulation,’* in Design Calculations
. for a Miniature - Hng-’l‘emperature In.Pile Circulating
- . Fuel Loop, by M.
- ORNL-1808 (Sept. 19, 1955)

Robmson and D, F, Weekes, 7

. 17Es!'imctted by method of $. I. Cohen and T N

"~Jones, A Summary of Density Measurements on Molten

Fluoride Mixtures and a Correlation - Useful for Pre-

dicting Densities of Fluoride Mixtures. of Knoun
'i'Composztzons, ORNL 1702 (Moy 14, 1954).

18Eshm«:n'ed on assumption that the alkalismetal fue!

" has ‘the same molar heaf capacny as that of the zir-
. conium-bearing fuel. : .

‘95 l.: Cohen, personal communi cations. to M, T

2O stimated by W. D, Powers, March 16, 1956.

21 Haynes Stellite Company, Hastelloy High-Strength,
Nickel-Base, Corrosion-Resistant Alloys, p 15, Sept. 'I
1951.

241

 
 

 

s v e e

 

ANP PROJECT PROGRESS REPORT

TABLE 4.2.1. SOME CALCULATED DESIGN PARAMETERS OF THE VERTICAL IN.PILE LOOP DESIGNED
FOR OPERATION IN POSITION C-46 OF THE LITR

 

 

. Fuel Temperature Maximum Air
Fuel Loop Reynolds Numbers Differential Temperature,. Air Flow
Position Fuel Air ©C) ©C) Rate (gfm)
NaF-ZrF -UF, Shallow®® 1,500 32,000 135 380 4
(53.5-40-6.5 _ 3,000 22,000 65 410 30
mole %) 6,000 20,000 33 510 28
' 10,000 20,000 33 520 28
Deep® 1,500 74,000 103 250 100
3,000 50,000 56 450 70
6,000 44,000 30 530 60
10,000 42,000 17 550 59
NaF-KF-LiF-UF Shallow? 1,500 19,000 125 450 27
(11.2-41-45.3 3,000 16,000 65 550 23
mole %) 6,000 15,000 K 580 21
' ‘ 10,000 15,000 20 600 21
Deep® 1,500 32,000 94 380 45
3,000 26,000 50 410 36
6,000 24,000 27 460 33
10,000 23,000 16 480 32

 

%Tip of loop at position of maximum thermal-neutron flux,
bBosed on earlier calculations. 12

“Tip of loop at bottem of active lattice.

factor from about 11 in the *“‘shallow’’ position to
about 6 in the ‘‘deep’’ position.

FAST-NEUTRODON DETECTORS
D. Binder

A great variety of radiation-damage studies re-
quire a knowledge of the fast-neutron flux of the
reactor being used for the irradiations, For
instance, changes in the mechanical and electrical
properties of metals and insulators are dependent
on the shape of the neutron-energy spectrum. As
the neutron energy increases, the number of atoms
displaced increases until a critical energy is
reached at which the effect saturates.??2 There-
fore, it is essential to know the number of neutrons
per unit energy up to about five times the critical
energy and the integrated flux above this energy.

The energy spectra in two types of reactors from
thermal to 100 ev or 1 kev and from 0.7 to 8 Mev

 

22G, H. Kinchin and R. S. Pease, Reps. Progr. in
Phys. 18, 152 (1955).

242

have been investigated by Trice.23:24 For the
remaining investigation, that is, the l-kev to
0.7-Mev region, a detector must be developed with
which to study the variation of displacements with
neutron energy. The development of useful de-
tectors is dependent on a knowledge of this vari-
ation, and vice versa. The first step therefore is
to calculate a known case based on simple as-
sumptions.

Germanium was selected as the moterml to be
studied because it is sensitive to neutrons and
the relationship between the solid-state effect and
the number of displacements is known. From the
work of Cleland et a.2% at low temperatures, it is
known that five electron traps are introduced per

 

23;, , Tnce, Fast Neutron Flux Measurements in
E.25 of the Brookbaven Graphite Reactor, ORNL CF-
55-7-130 (July 27, 1955).

24; B, Trice, A Series of Thermal, Epithermal and
Fast Neutron Measurements in the MTR, ORNL CF-
55-10-140 (Oct. 28, 1955).

255, W, Cleland, J. H. Crawford, Jf, and J. C. Pigg,
Pbys. Rev. 98, 1742 (1955).
 

 

"

o)

"

n

 

‘*fast’’ neutron in a graphite reactor and that there
are two traps per displacement.: The question is
whether it is possible to calculate the number of
displacements on the basis of a reasonable
graphite reactor spectrum and to achieve approxi-
mate agreemenf with the experimental result of
2,5 displacements per fast neutron. (A fast neutron
is defined?’ by Cleland et al. to be a neutron with
energy greater than that required to displace an
atom.)

‘For the calculation the method of Kinchin and
Pease?? was modified to include a spectrum which
is the sum of 1/E distributions extending to each

‘element of a fission spectrum,2® The agreement

between the calculation and the experimental re-
sult was within a factor of 4. The spectrum and
the variafion of displacements with energy are
then roughly correct, or at least there are com-
pensating errors, '

In order to test the variation with neutron energy
alone, it is plonned to use neutron sources of
known energy. The flux from these sources is
reduced by factors of 10° to 105 from graphite
reactor fluxes, but the sensitivity of pure germanium
allows for experiments even ot these low levels.
Two sources will be used: a uranium converter
plate in a thermal column and an antimony-beryl-
lium source. The converter plate gives a pure
fission spectrum, and, when used in the slanting
animal tunnel of the ORNL Grophlte Reactor, emits
2 x 108 fission neutrons/cmZ.sec near the center
of the plute. A first irradiation of 5.8 x 10'3
neutrons/cm? introduced 1.9 x 1075 acceptors/cm?

in a germanium crystal. This experimental value
agrees to within a factor of 2 wuth the coiculufed |

value of 3.4 x '!0I

‘The. result of the firsf expenment mdlcafes fhot'“
‘the present ossumphons on ‘the variation ‘of d:s—'.:_’.
 placements with neutron energy are roughlymrrecf. L
- This will be checked by &u_’the_r irradiations with
fission neutrons and with 30-kev neutrons'fr'on"l:-«-"'
~the antimony-beryllium photoneutron source. More -

- adequcte neutron monitoring is to- be used for -

~ future germanium irradiations. This ‘may lead to a
_ complete unclerstcmdmg of dlsplocements |n ot .

leost one solld-state reaction. S

 

1'6D Bmder, Solid State Semiann, Prog. Rep Feb,
29, 1956, ORNL-2051, p 48.

- PERIOD ENDING JUNE 10, 1956

. VISCOMETER FOR REMOTE MEASUREMENTS
OF THE VISCOSITY OF IRRADIATED
FUSED-SALT FUELS '

C.C. Webster

A viscometer was designed for use in a hot cell
which will be capable of measuring the viscosities
of irradiated fused-salt fuels at temperatures up
to about 750°C in an inert (helium) atmosphere
within a shielded cell. The apparatus consists of
two vessels, 6 in, long, 1.25 in. OD, mounted in
a vertical position, with 2 in. between centers. An
Inconel tube 3 ft fong and bent into a U-shape

‘connects the bottoms of the vessels. A small

ferromagnetic ball travels with the liquid through
the tube, and the position of the ball is detected
at two points on each |eg of the U-tube by elec-
tronic equipment.

Swagelok connections are provided at the top
for inserting liquid-level probes and for allowing
helium pressure to be applied on the surface of

_the liquid in either vessel. The whole apparatus
‘is inserted into a 4-in.-dia vertical furnace with

an 18-in. controlled-temperature heated zone. End

~heaters are provided to control the heat loss out

the ends of the furnace.

EFFECTS OF RADIATION ON
ELECTRICAL COMPONENTS

J. C. Pigg C. C. Robinson?7

Insulation

28,29

In previous experiments on the effect of

_radiation on electrical insulation the insulated
_. wire to be irradiated was placed in the reactor as
" if it were a sample lead. Care was taken to ensure -

~ that possible feakage paths from the central con-
- ‘ductor to ‘the shield were as long .as posmble.'
 Whenever practicable, the end of the insulation .
- that was in the reactor was sealed over at the end
‘of the central conductor. However, the possibility

of conductlon by end surfoce contamination and
ionization of the air still existed when there was
no seal or when the seal was defective.
In order to determine whether such leakage was
important in the. expemnent:s prev:ously conducted,

 

g 27On uss:gnment from Wright Alr Development Center.

28 C.Pigg and C, C, Robmson, Solid State Semzarm.

Prog, Rep. Aug. 30, 1955, ORNL.-1945, p 6.

2%, c. Pigg et al., Communication and Electromcs
22, 717 (Jaonuary 1956). ,

243

 
 

i e s b

 

 

 

ANP PROJECT PROGRESS REPORT

two samples were taken from the same spool of
Teflon-insulated wire, One sample was the usual
length, 25 ft long, and the second was 50 ft long.
The two samples were installed at the same time
in hole 50-N of the ORNL Graphite Reactor. The
25-ft Teflon-insulated sample was installed in the
same manner as in the previous experiments, with
no seal at the end. The 50-ft sample was doubled
back on itself so that both ends of the wire were
outside the reactor. The leakage between the
shield ‘and the central conductor was measured to
be twice as large for the 50-ft sample as for the
25-ft sample. The photo-emf measurements on the
two wires were identical. It can therefore be
concluded that the radiation effects described
previouslyzs‘ were due to bulk properties rather
than to end surface conduction or to air ionization.

Barriers

A 1IN3BA germanium point-contact rectifier was
exposed in a 2 x 10%-r/hr Co®® gamma-ray source,
and the forward and reverse currents were measured
at l-v bias. The behavior of the current is shown
in Fig. 4.2.14. These data are in qualitative
agreement with those obtained by Young.so, The
increase in reverse current, followed by a return

toward its initial value, is an effect not found in .

reactor irradiation experiments because it occurs
at low total damage. As may be seen from the
change in forward current, the effect illustrated in
Fig. 4.2.14 would have occurred before the sample

 

30R, c. Young, Gamma Radiation of Crystal Diodes,
;Jrlglht A)Il‘ Development Center, WCRT-TN-54-255 (Dec.
8, 1954).

UNCLASSIFIED
ORNL~LR=—DWG 13755

 

 

 

 

CURRENT AT 1 v BIAS

 

IN38A-B
o REVERSE CURRENT, uo
® FORWARD CURRENT, mo X107

{N3BA-A

A REVERSE BIAS; NORMALIZED TO
COMPARE WITH {N38A—-B

 

 

TS

 

Ad

 

 

 

 

 

 

 

 

 

o 1 2 3

GAMMA EXPOSURE {r)

4 5 6 : T 7{xt0")

Fig. 4.2.14, Effect of Gamma Radiation on Conduction Through a Barrier. Germanium point?co;-:tfict
rectifier IN38A-B exposed to 2 x 108 t/hr from Cob? source; 1N38A-A exposed to 2.5 x 105 r/hr 1from

Cob9 source.

244

i

 
 

 

 

"

b

 

reached thermal equilibrium in .an-in-pile experi-
ment. Whether or not the effect does result from
neutron irradiation has not yet been determined.

A series of exposures were also made in a
5 x 10%-r/hr Co®? source, and the voltage-current
characteristics of the rectifier were measured
after the samples were removed from the source.
The change in the characteristic curve can be
seen in Fig. 4.2.15. There is little change in the
forward curve, which is in agreement with the data
in Fig. 4.2,14. The current at l-v reverse bias
was normalized to make the preirradiation value
coincide with that of the sample exposed in the
2 x 10%r/hr Co®® source. The normalized data
are shown in Fig. 4.2.14.

UNCLASSIFIED
ORNL-LR-DWG 13756 2

—— BEFORE EXPOSURE
—— AFTER FIRST EXPOSURE TO # x 40° r

—— AFTER SECOND EXPOSURE TO 242 x1C° r
—-— AFTER THIRD EXPOSURE TO 7.77 1408 ¢

[: ]

M=967 21077

H o
REVERSE CURRENT ()

FORWARD CURRENT {(ma)
o

B

 

M=a4x1077
2 Me=3ax07 2
=18 %407
0 . — — 0
o oz 0.4'_ 06 08 W0 24

VOLTAGE (v) S

Fig. 4 '.2 A5, Effect cf Gclmmu Rcdinhcn on

’-,'-Vohage-Currenr Churucierishc of Germnnium Pointe.
: _""Contaci Rectifier IN38A-A at 20°C.; 11\e M vclues_‘ '
' fj:nre the slopes of the Imes. L - '

  

f The forwurd chcracterrsflc corresponds to fhe B
S ;_'?volfage-current behcvuor of the emitter of a tran-
0 sisters . The “reverse ‘curve . corresponds o fhe:fi”
L :collectcr chcrcctertsflc of a transistor. It may be
- seenin_Fig.4.2.15 that, ‘olthough the emitter:
" characteristic s little” changed,”. rhe collector
~ characteristic changes drastically both in mugmrude :

and slope. The magnitude of the current at l-v
bias changed by a factor of about 6, and the slope

 

PERIOD ENDING JUNE 10, 1956

changed by a factor of about 5.3. The changes in
the characteristics are similar to those caused by
neutron irradiation,3!

A series of transistors were irradiated in the
circuit shown in Fig. 4,2.16 to measure the ampli-
fication of the unit during, as well as before and
after, irradiation. The series resistances in the
emitter and in the collector circuit were changed
to match the unit being considered in each experi-
ment.  Provisions were made to switch from
grounded-emitter to grounded-base circuitry so that

“both types of circuitry could be studied simul-

taneously with the same unit. Exposures were
made both in Co%® gammoa-ray sources and in the
ORNL Graphite Reactor,

The changes in the collector current at a 2-v
bias and in the amplification of a Minneapolis-
Honeywell H-2 power transistor as a function of
gomma irradiation may be seen in Fig. 4.2.17. The
collector, base, and emitter voltages were adjusted
to the original values before each reading. Hence,
the changes in amplification were due to changes

‘in characteristics and not to changes in operating
~ voltages, It may be seen that the collector current

and the amplification were essentially the inverse
of each other. Whether the change in amplification
was due to a change in the slope of the collector
characteristic, a change in separation of the
collector curves, or a change in impedance match-
ing has not yet been determined.

 The changes in amplification of two Minneapolis-
Honeywell - power transistors as a function of
irradiation .in hole 51-N of the ORNL Graphite
Reactor are shown in Fig. 4.2.18. These curves
illustrate the same type of behavior as that ob-

o served for- RCA transistor No. 1241.32 The tempo-
<~ - rary increase in umpllfucanon observed previously -

in RCA transistor No. 1266 was not noted. - Again,

'-_.'the flux: Ievel was - such that this region of the
‘eurve would have been traversed before thermal -
ethbrwm was obtalned in the m-plle experlmenf. =

lnfluence of an Elecmcul Field on Drffusion
' oflntersimul Aioms L

When germcnwm is bomborded by hrgh-energy

_,,neutrons the tecoils from the neutron-atom col- -
*;';.iltswn cnd secondary collrsrons cf fhe orngmal

 

3!J C. P:gg, Solid State Semzann. Prog. Rep. Aug. '

'31 1953, ORNL-1606, p 81.

3 c. Pigg and J, W, Clelund Solzd State Semzann.

' Prog. Rep. Feb. 10, 1953, ORNL-1506, p 47.

245

 
 

 

 

 

ANP PROJECT PROGRESS REPORT

 

UNCLASSIFIED
© ORNL-LR-DWG 13757

Q

 

 

Vs I _
1.0 ]
t i
it AW
3 Re
- Z1000

|K {0-TURN
HELIPOT

 

 

 

 

 

 

 

o

@
| L 0-50pa

 

 

 

 

 

 

 

 

 

 

 

_ Jfi‘éfl: = 0-10v DC
T T 5K 10-TURN
- . HELIPOT .
. b GROUNDED
GROUNDED EMITTER
BASE
5K 10-TURN :I____
HELIPOT ) =
0-150v DC

 

-L%J

 

 

 

 

HOv-60~

o—4
]
9)

Fig. 4.2,16. Transistor Test Circuit.

atom as it travels through the lattice result in the
formation of lattice vacencies and interstitial
atoms. Since the germanium lottice is fairly open,
the interstitial atoms can diffuse readily through
the crystal, Because of the high dielectric con-
stant of germanium, the interstitial atom is ionized
at room temperature, and hence its diffusion should
be influenced by the presence of an electric field.
Fields of the order of tenths of a volt per micron
are available in the barrier region of o diode, and
the conductivity of the dicde is extremely sensitive
to changes in carrier concentration in the region
of the barrier.

A 1IN38A point-contact germanium rectifier was
exposed for 5 min in hole 51-N of the ORNL
Graphite Reactor. After removal from the reactor
the rectifier was biased in the forward direction

246

at 0.7 v in order to remove the potential gradient
in the borrier region. The measured current was
low because of radiation damage. The unit was
then placed in an oil bath to maintain a constant -
temperature of about 25°C for six days. At the
end of the six-day period there had been no
observable change in the current passing through
the sample at 0.7-v bias. The bias was removed
for 2 hr, then restored. When the bias had been
restored, it was noted that the current had in-
creased by about an order of magnitude. After
eight days in the biased condition, the current had
not changed.

When the bias was removed, an electric'field
was established at the barrier to cause the relative
interstitial atoms to diffuse from the barrier region,
This removal of interstitial atoms from the p-type

 
 

"

-}

&)

Ll

 

UNCLASSIFIED
ORNL—LR-DWG 44749

 

 

 

 

 

 

05 ' | 100
COLLECTOR CURRENT: DC DROP -
Y ACROSS 1000 ohms AT 2 volts BIAS
04 & o> P00 O - 80
ho
| .
2 03 _ ~ 60 _
b— i b
E . /"’ ] E
= . -~ t
e ~T 1 ~
oz [ - 40
: AMPLIFICATION: AC DROP ACROSS
RESISTOR IN COLLECTOR LEG
0.4 20
0 0

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

0 1 23 456 7 8 910 #1213 14 15(x10%)
' RADIATION DOSAGE (r/hr)

Fig. 4.2.17. Effect of Radiation on Collector
Cutrent and Amplification of a Minnefi:polisr-Honey-
well H-2 Power Transistor kradiated in Gamma-
Ray Source of 2,57 x 105 r/hr.

PERIOD ENDING JUNE 10, 1956

side of the barrier results in annealing out the
type of damage which tends to make the p-type
side less p-type and transporting it to the n-type
side. Hence both sides of the barrier are annealed
toward their initial condition and the forward
current increases toward its initial value.

RADIATION DAMAGE TO BOR{ON CARBIDE
J. G. Morgan

The study of radiation effects on B,C, initiated
previously,®3 was completed. The apparatus used
for the irradiation of samples in the LITR was
described in the previous report. The first samples
irradioted were hot-pressed, high-density 8,C,

0. Sisman

_ supplied by the Norton Company, and slsp-casf

B 4C bonded with SiC, supplied by The Carborundum

- Company. Four other types of B, C have now been
_irradiated. The results of spectrographic analyses

 

330, Sisman et al., ANP Quar. Prog. Rep. March 10,
1956, ORNL-2061, p 204.

UNCLASSIFIED
ORNL—LR—DWG {13759

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1.0
0.9
0.8
AMPLIFICATION: A-C DROP ACROSS RESISTOR IN COLLECTOR LEG
0.7
06
2
- Lo
"2 0.5
& ] . . -
E\r 4 1 e 1 - _ " ,
204 N(?.A;_fLU* _ltBGx 0 nv —
03— . — _
02 T . T
R b - - ..“'u-. o ‘“*o--,._____ :
e L b ‘ 7"“"-3-- — —
o NO.G;"-FLJX186x!0‘°nv Sl f Sttt e e
S0 b— L : - - ' - : - |
e o4tz o3 4 "B 8 7T B 9 .0 . 92 3. 14 .45 16 17(x10%

INTEGRATED NEUTRON FLUX (avt,)

Flg. 42 18 Chunges in Amplificuiion of Two aneapohs-Honeywe" H-2 Power Trcnsustors as a

Function of Integrated Fast-Neutron Exposure.

247

 
 

 

ANP PROJECT PROGRESS REPORT

of the six types of B,C are presented in Table
4,2.2, and the resvlts of examinations of 20 irradi-
ated samples of the six types are given in Table
4.2,3.

The four types of B,C irradiated durlng this
quarter are described below‘

1. The Norton Company supplied low-density,
high-purity samples that had been hot-pressed from
325 mesh, and finer, powder and fired at above
2000°C. The boron content of this material was
62,4 + 0.4 wt %. These bodies were hard, strong,
and impervious. :

2. The Norton Company also supplied low-
density, technical-grade samples that had been
hot-pressed from 60-mesh, and finer, material aond
fired at above 2000°C. These bodies were hard,
strong, and nearly impervious, and they contained
62.4 + 0.4 wt % boron.

3. The Carborundum Company supplied tiles
(No. 1) cast from a basic mixture of 90% B,C
grains and 10% 200-mesh silicon metal. The cast
bodies were fired at 1350°C and refired at 2000°C,
They were porous and friable, ond they contained
64.7 + 0.5 wt % boron,

4, The Carborundum Company also supplied
tiles (No. 2) cast from a basic mixture of 80%

B,C, 10% 200-mesh, and finer, boron-metal powder,
and 10% 200-mesh, and finer, silicon-metal powder.
The cast bodies were fired at 1350°C and refired
at 2000°C. They were porous and friable, and they
contained 67.8 % 0.4 wt % boron. - |

The samples were exposed for 800 hr in the

"LITR in a thermal flux of 2 x 10'3 nv. The sample

temperature was less than 200°C, and the total
dose received by each sample was 6 x 10'° nvt
(thermal). These conditions are to be compored
with those expected in the ART in which the
temperature of the B,C layer will be 1450°F and .
the integrated neutron flux dosage for 500 hr of
operation will be 1.2 x 10'? nvt (exposure behind -
a 0,030-in. gap in the cermet layer). The average
burnup of these samples was 2.9%, and most of
the burnup occurred in the first few thousandths
of an inch of the surface layer. '
A much greater exposure will be requnred to
simulate the ART operating conditions for the
copper-B,C cermet layer and the boron steel, and
fherefore these materials will be irradiated at the
MTR at a temperature of 1600°F. Calculations
indicate that these samples should receive total .
doses of 1.5 x 1020 nvt for the cermet ond 8 x 1017
nvt for the boron steel, which are equivalent to

TABLE 4.2,2. RESULTS OF SPECTROGRAPHIC ANALYSES OF B,C SAMPLES

 

 

 

Composition (%)

 

 

Norton Norton Norton Carborundum  Coarborundum  Cerborundum
Element*  High-Density Low-Density Low-Density Sample Cast Tile No, 1 Tile No. 2
Hot-Pressed Technical-Grade High-Purity and Bonded Refired at Refired at
Sample Sample Sample with SiC 2000°C 2000°C
Al 0.01-0.1 1-10 0.1-1 0.1-1 0.1-1 0.1-1
Ca 0.001-0.01 0.001-0.01 0.001-0.01 0.001-0.01 0.001-0.01 0.001-0.01
Cr ek *x b 0.01-0.1 0.01-0.1 0.01-0.1
Cu *x 0.01-0.1 *h *k >k '
Fe 0.01-0.1 0.1-1 0.1-1 1-10 0.1-1 1-10
Mg -0.0001-0.001 0.00t-0.01 0.0001-0.001 0.001-0.01 0.001-0.01 0.001-0.01
‘Mn *k 0.01-0.1 >k 0.01--0.1 0.01-0.1 0.01-0.1
Ni *k *x ** 0.1-1 0.01-0.1 0.01-0.1
Ti rx 0.1-1 0.1-1 0.1-1 0.1-1 0.1-1
Zr *x 0.1-1 0.1-1 1-10 1-10 1-10

 

**Sought but not found.

248

*Limit of detection: Cr, 0.004%; Cuv, 0.0bOI%: Mn, 0.001%; Ni, 0.02%; Ti, 0.04%; Zr, 0.03%.
 

 

N

o

®)

PERIOD ENDING JUNE 10, 1956

TABLE 4.2.3. RESULTS OF EXAMINATION OF IRRADIATED B,C SAMPLES

 

 

 

Bulk Density Dimensional
Specimens Gas ~Appecrance Change ‘ Change
fan*
Evolution Change (g/cm3) (in./in.)
Three Norton Co. high-density 0 £0.4 em? None 0 * 0.01 0 1 0.001
(2.49 g/cms) hot-pressed samples
Three Norton Co. low-density 0 * 0.1 emS None 0 1 0 * 0.007
(2.02 g/cma)** technical-grade '
samples
Three Norton Co, low-density 0 £0.1 cm® None 0 £0.05 0 *0.001
(2.17_§/cm3) high-purity samples ’
Three Carborundum Co. samples 8.1 cm3/g (av) None, black
cast and bonded with S$iC deposit on
container Carborundum samples too
Two samples of Carborundum Co. 0 0.1 em® None porot.ts and f':i able for
tile No. 1 refired to 2000°C | precise density or
dimensional measurements
Six samples of Carborundum Co. None None
tile No. 2 refired to 2000°C
3

*Theoretical production, approximately 1 em”.

**Bulk density; sample slightly porous,

two weeks aond one week, respectively, in the
MTR at a thermal flux at 1.5 x 1074 nv.

Irradiations have been initiated on a 7.6%
CaB,-92.4% Fe plate clad with stainless steel.
Studies of a 10,3% BN-89.7% Ni plate clad with
type 304 stainless steel are planned.

IRRADJIATIONS OF STRESSED SHIELDING
MATERIALS ‘

4. C. Wilson

. W. E. Brundage | W, W Daws B

A test apparatus has been demgned for |rrad|ot|on
of a 1 wt % boron (Bw)—stamless steel alloy under

 stress in the LITR ot 1300 and 1600°F, "In these :
tests one sample will be irradiated while stressed_
in compression at 500 psi ot each temperoture, and- -

one sample will be stressed similarly at the lower -

temperature out ‘of .the reactor.” One unstressed -
specimen will also be irradiated ot each tempero-"..’;"
ture, ‘The irradiation periods will be three to six
weeks., The data obmmed will be compared with .
available data on” MTR irradiations of boron—

stainless steel alloys for Westinghouse (WAPD). -
Specimens of an austenitic-stainless-steel-clad
copper-B4C cermet (6.6 wt % B,C) are also to be

irradiated in the LITR at 1300, 1600, and 1700°F.
The samples will be checked for dimensional
stability and hardness.

THE EFFECT OF RADIATlON ON POLYMERS

et

0. Sisman
W. W. Parkinson W. C. Sears

There has been a continuing program in the
Solid State Division to study the effects of high-

‘energy radiation on plastics and elastomers. This
- work has heretofore been reported only in the
~ Solid State Division progress reports, but because

of increasing interest in this work in connection
with allied ANP projects and because there is an
increasing - need . for - radaahon-damoge data on

_organic- materials, some of the current polymer
- work “will now also be presented in the ANP
- progress reports.

The plastics ancj elostomers have been studied

‘quite extensively, _and some of the chemical re-
actions have been identified which. cause the
. changes that' have been observed in tfie physical
‘properties of irradiated polymers. A study is

currently under way of the reaction products of
irradiated polymers. A summary of this work, for

249

 
 

 

 

 

ANP PROJECT PROGRESS REPORT

which the infrared spectrometer is used, is presented
here.

The infrared spectra of polystyrene, polyethylene,
polybutadiene, GR-S, natural rubber, deproteinized
rubber, polyvinyl chloride, and Teflon have been
measured before and after irradiation. The con-
ditions of irradiation of these materials and some
others which were not studied so thoroughly after
irradiation are given in Table 4,2.4.

The polymer films were cemented on 0 2.9 x 1.1
cm rectangular aluminum wire frame, evacuated
for 2 to 5 hr, and then measured to provide a
preirradiation spectrum. The films to be irradiated
in vacuum were evacuated for three to five days
at 0.2 ¢ ond sealed in vacuum in pyrex or quartz
tubes prior to irradiation. These sample tubes

were packed with aluminum foil in aluminum cans
for irradiation. The polymers were irradiated in
one or more of the following facilities: (1) a Cob?
gamme-ray source, (2) water-cooled hole No. 19
in the ORNL Graphite Reactor, and (3) lattice
position C-46 in the LITR, The maximum permis-
sible dosage was limited by the formation of open
slits in some polymer films (caused by shrinkage
and loss in film strength). It was found that the
dosage on polystyrene, polyvinyl chloride, mylar,
and polymethyl methacrylate could be increased
by irradiating unmounted films.

The discovery that postirradiation oxidation had
occurred in irradiated polymers during exposure to
air necessitated opening the evacuated somple
tubes in a helium atmosphere. A glove box that

TABLE 4.2.4. IRRADIATION DOSAGES OF POLYMERS

 

Irradiation Exposure (X IDB rads)

 

 

 

Polymer Cob0 Gomma, .\,]06 o/ he sz,h"z Reactor, LIBTR, Averc'.ge Sample
105 «/hr, ~108 ¢/hr, Thickness
In Oxygen  In Vacuum in Vacuum in Vacuum (in.)
Polystyrene® '} | 2.3, 11 1.5, 16, 23, 31, 35 1000, 1000, 1500  0.0018
Polyalphamethyl styrene® 0.39, 6.9 6.6° 7.0 0,002
Polyethylene® 0.39, 3.9 1.2, 2.3, 6.2 3.4 0.0012¢
Polybutadiene® 3.8, 6.1, 8.9 4.0, 14, 18 0.003
GR-S¢ 0.72, 6.5 3.8, 6.1, 8.9 3.8, 13, 20 0.002
Natural (Hevea) rubber® 1.4, 8.9 4.8,/ 16 0.004
Extracted Hevea® 1.4, 8.9 16, 22 0.004
Polyvinyl chloride® 0.39, 0.72, 3.9 3.9, 5.2, 8.9 12, 40° 0.002
Polymethyl methacrylate® 0.39, 6.9 6.5% 2.0, 6.5 0.0004
Tefion® 0.72 3.8, 6.1 16, 22 | 0.0006
Mylar® 0.39, 6.9 2.3, 6.2 16, 17 0.00025
Nylon€ resin No. 63 0.72 1.4, 6.1 22 C 0.0015

 

“Supplied by Dow Chemical Company.

Sample crumbled; measurement after irradiation impossible.

“Commercial film,
9A1s0 0,008 and 0.030 in. exposed to 0.12 X 107 rads.
€Supplied by B. F. Goodrich Co.
Tube cracked during irradiation,
8Supplied by E. I. du Pont de Nemours & Co., Inc,

250
 

 

 

 

o)

»

could be evacuated was used for this operation
ond for transferring the specimens to a gas-tight
infrared absorption cell for spectral measurement.
The absorption cell was then opened in air, and
measurements of the spectrum were repeated after
various periods of exposure to the atmosphere.

As an example of the spectral changes observed
in polymers, the spectrum of a polystyrene sample
is shown in Fig. 4.2,19. While polystyrene is more

\ | UNCLASSFIED

 

 

 

  

 
   
  

 

 

 
 
    

 

 

ORNL-LA-OWS 10939
®o T T T T 1 I
gr
. FEe0 —
wl POLYSTYRENE N
N BEFORE IRRADIATION ]
= 20 |- —
— —
o i ] 1 to
00 T T T T ]
g -
] = 60 - . ]
- 100 x 10° RADS =]
or MEASURED N HELIM 7]
b .
o . ! .
00 L ot
gw
- &0 ]
5 SAME SAMPLE
40 AFTER AR EXPOSURE -1
FOR 1T HOURS -
+ 20 - -
°
00
y .
& 80 -
§‘° ;
20 —
0 L ]
+ 4000 3000 2000 . 800 600 1400 1200 K00 800 s00

 

 

 

 

Fig.r4.2.l 9. Infrared Spectra of Polyflyrene. ;

PERIOD ENDING JUNE 10, 1956

resistant to change by radiation than most polymers
are, the changes that may be seen in Fig. 4.2.19
are typical of the alterations of the infrared spectra
by irradiation. Dosages of the order of 108 10 1010
rads were required to produce significant changes
in the infrared spectra of most polymers. -

A large postirradiation effect was observed in
the spectra of polystyrene, GR-S, and natural
rubber, Upon exposure to air following irradiation
in vacuum, oxidation products continued to form

for periods of up to 95 days. These reactions

were indicated by the growth of strong hydroxyl
and carbonyl bonds. The oxidation products
formed in postirradiation oxidation were different
from those produced by irradiation in oxygen.

Irradiation of polystyrene in vacuum to high
doses produced a wholesale ‘disruption in which
both the aromatic and aliphatic components were
equally affected. On the other hand, the aliphatic
component of GR-S (a styrene-butadiene copolymer)
showed a greater percentage change than did
polybutadiene at equal dosages.

All polymers showed significant changes in the
double-bond regions as a result of irradiation. In
polyethylene, RR,C=CH, groups disappeared as
trans RCH=CHR, groups formed. In GR-S and

-polybutadiene the number of terminal vinyl groups

decreased and the unsaturation in the hydrocarbon
chains of GR-S ond of natural rubber was de-
creased. lrradiation increased the number of trans

‘RCH=CHR, groups in natural rubber, as it did in

polyethylene. Conjugated and unconjugated un-

saturation was produced in polyvinyl chloride.

251

 
 

 

 

ANP PROJECT PROGRESS REPORT

4.3. FUEL RECOVERY AND REPROCESSING
7 | H. K. Jackson ' .
D. E. Ferguson W. K. Eister H. E. Goeller

-VOLAT!LI;I'fY PILOT PLANT DESIGN
AND CONSTRUCTION

R. P. Milford F. N. Browder

The design of the volatility pilot plant for re-
covering fused-salt fuels is complete except for the
molten-salt sampling device for the fluorinator and
-the trap-door closing device for the waste-salt car-
rier. - All major equipment items were received, as
well as the electrical power and control center and
the instrument panelboards. The equipment, with
the exception of the ARE fuel hold tank, which is
not required until later, is in place; and piping,
electrical work, ond instrument installation are
proceeding rapidly. It is expected that the plant
will be completed by June 30.

NICKEL FLUORIDE SLUDGE
FORMATION STUDIES

G. |. Cathers M. R. Benneit

The behavior of NiF, in molten NaF-ZrF, and
NaF-ZrF ,-UF, systems was studied to determine
whether the presence of this corrosion product
would cause the formation of sludges which would
interfere with salt transfer in the fluoride-volatility
process. In order to determine the solubility of
NiF, in the salt mixtures, NiF, was added to mol-
ten NaF-ZrF , (50-50 mole %), and the mixture was
heated until a clear solution was obtained. it was
then cooled until turbidity reappeared. Solubility
values estimated by the disappearance of turbidity
were in fairly good agreement with solubility values
determined electrochemically?! for the solvent NaF-
Zrf , (53-47 mole %). Based on the visual deter-
minations, the estimates of the solubility of NiF
in NaF-ZeF ; (50-50 mole %) were the following:

Temperature Solubility of NiF2
(°C) (wt % NiF )
640 0.7
670 1.0
685 1.3

 

. E Topol, ANP Quar. Prog. Rep. March 10, 1956,
ORNL-2061, p 89.

252

These values show the very definite increase in
solubility with temperature when compared with the
reported solubility of 0.2 wt % Ni at 600°C.

The addition ofas much as 6 wt % NiF, to molten
NaF-ZrF4 (50-50 mole %) at 600°C resulted in the
formation of a viscous dispersion which was fairly
stable, and thus other tests were made to determine
the effect of concentration on sedimentation. These
tests were made by dry mixing the required amount
of salt (~30 g total) and melting it in a %.in.-ID
nickel tube. Nitrogen was used initiolly for agita-
tion and then as a blanket while the material was
kept at 600°C for various times. . The tube was
quickly quenched with cold water at the end of the
test to fix the NiF, concentration at various
heights in the tube. The tube was then cut into
'/2-in.-|ong sections, and the salt was analyzed for
nickel. Although some settling of NiF, was evi-
dent after only 0.5 hr when 2 wt % Ni was added
(as NiF,), complete settling had not occurred even
after 72 hr (Table 4.3.1). When only 1 wt % Ni was
added, settling was more nearly complete at 72 hr;
since the NiF, concentration was lower, the in-
crease in viscosity ot the bottom was not so great
and thus was not so much of a deterrent to settling.

In further experiments the sedimentation of NiF
in molten NaF-ZrF -UF , (48-48-4 mole %) at 600°C
(Table 4.3.2) was similar to that found in the
vranium-free system. However, with 2 wt % Ni and
with uranium present, the settling was less after
2 hr than in the test with no uranium. Since the
solubility of NiF, in NaF-ZrF# is near the lower of
the nickel concentrations found in these settling
tests, it appears that the solubility of NiF, is ap-
proximately the some in uranium-bearing and ura-
nium-free mixtures. These experiments have dem-
onstrated that NiF, concentrations of up to 2 wt %,
more by a factor of 10 than is expected in qircraft
reactor fuel reprocessing, would not interfere with
salt transfers unless the molten salt were permitted
to stand unagitated for long periods of time.

DECOMPOSITION OF UF6-3N0F COMPLEX
G. |. Cathers R. L. Jolley

Some exploratory work was carried out on the
decomposition of the UF,.3NaF complex at high
 

 

"

»

»

 

". PERIOD ENDING JUNE 10, 1956

TABLE 4.3.1. SEDIMENTATION OF NiF, IN MOLTEN NuF-ZrF4 (50-50 mole %)-AT 600°C

 

Nicke! Concentration (wt %)

 

~ Relative Position

After : " After - After - After

 

bf Sampling '
S TRETE Initially 0.5 hr 2k 8 hr 72 hr
Inttial Nickel Content* — V2 wt %
1 (top) 1.60 0.28 0.20
2 1.74 0.73 0.30 0.22 0.20
3 1.72 1.72 2.20 1.40 0.21
4 178 1.86 2.62 © 3.48 1.23
5 (bottom) 2.13 2.02 3.02 3,54 2.99
Initial Nickel Content* — 1 wt %

1 (top) 0.22
2 0.30 0.16
3 0.31 0.17
4 1.71 0.18
5 (bottom) 3.06

 

*The nickel was added as Nin.

TABLE 4.3.2. SEDIMENTATION OF NiF, IN MOLTEN NaF-ZrF ~UF , (48-48-4 mole %} AT 600°C

 

Nickel Concentration (wt %)

 

Relative Position Initial Ni Content — 2 wt %

" Initial Ni Content = 1 wt %

Initial Ni Content — 0.5 wt %

 

 

 

of Sampling
After 2 hr After 48 hr After 2 hr After 48 hr After 48 hr
1 (top) 1.26
2 123 0.25 0.36 0.18 0.24
3 2.12 034 040 0.20 0.33
4 278 275 283 0.47 0.22
281 6.94 308 0.85

55'(b6fiom) '

 

| ,_’temperatures in an effort to develop a UF 6 desorp-.
- tion procedure ‘which avmds UF decomposmon. '

The decomposmon reaction would lead to uranium

| “being- held - up on the NaF bed ond wouid thus -

necessntofe a subsequenf recovery step. Tests

'were conducted by saturating an NoF bed with UF
- at 100°C ond then heating the bed, as @ closeg
- sysfem, at the desired temperature for 1'to 4 hr. In

- some of the tests the bed was subjected to as much

as 55 psia UF pressure during the high-tempera-

ture cycle. HoWever, no significant correlations
‘were apparent as to the effect of UF, pressure or

length of time of treatment., The residual uranium

- content in several tests at 400°C varied in the
range 18 to 26%. At 300°C the decomposition ef-
fect was much less, a residual ‘uranium content of

2% being produced. Use of excess F, (10 psia) in

‘tests at 400°C also resulted in less decomposition,

with residual uranium contents of 8 to 14%.
The residual uranium present in the product of all

253

 
 

 

 

 

A

ANP PROJECT PROGRESS REPORT

runs was pentavalent. ~ As a result of dispropor~
tionation of the pentavalent uranium in the analyti-
cal procedure, however, the tetravalent uranium
content in each case was approximately equal to

- the hexavalent content. The decompasition reac-

tion is therefore believed to be

UF ;-3NaF —> UF ;-xNaF + %F,

254

These results indicate that, if a significont par-
tial pressure of UF, is retained in the NaF bed
because of a plugged line or cold trap during UF
desorption, excessive UF, decomposition will oc-
cur when the temperature reaches the 300 to 400°C
range. Therefore precautions must be taken to
ensure that full sweep-gas flow through the NaF
bed is maintained during UF ; desorption.
 

 

"

o

 

PERIOD ENDING JUNE 10, 1956

4.4. CRITICAL EXPERIMENTS
A, D, Callihan

CRITICAL EXPERIMENTS FOR THE \
COMPACT-CORE REFLECTOR-MODERATED

REACTOR
E. Demski' J. J. Lynn
W. J. Fader! D. E. McCarty
D. A. Harvey! E. V. Sandin!

D. Sc ofi

The study of the NDA2-proposed, sodium-cooled,
reflector-moderated reactor with solid fuel ele-
ments® has been completed.  In the critical as-
sembly the fuel region contained 0.004-in.-thick
vranium sheets interleaved between aluminum and
stainless steel sheets, This fuel region was
separated from the beryllium of the island and
of the reflector by stainless steel shells. As
described previously,* additional uranium, in the
form of 0.01-in.-thick disks, was added -in one
section of the fuel region to provide excess re-
activity for other measurements. An evaluation
of the effect of this local nonuniform fuel distribu-
tion on the reactivity was made by replacing
2636 g of U235 in 0.004«in.-thick sheets with
2678 g of U235 in 0.01-in.-thick disks in another
section of the core. With the other materials
unchanged, there was a reactivity loss of only
19 cents, |

Evaluation of Stainless Steel Shell

The loss in reactivity caused by the stainless
steel shells was determined by substituting alumi-

num shells of the same dimensions.  The exchange
of 4.4 kg of atuminum for ll 9 kg of stainless
steel resuhed in o gain in reactivity of $4.20,
estimated to be equwcient to o 14% decrease in
~the critical ‘mass.  The excess reachvn‘y was
 partly compensafed for by the .removal of some -
" of the outer ]cyer of the “/ einethick reflector. -
e Removal of a 2/8-|n.-th|ck sectuon that extended"

 

!On nssignmenf from Prctt & Whitney Aircrnft. . o
2Nuclear Developmant Corporoflon of Amer!co.

3CCR-2 A Compact Core Reactor for Azrcraft Pro-g- |

pulszon. NY0-3080 (July 30, 1954), Qudrterly Progress

" "Report, - ANP Developmeflt- Oct. -1°w"Dec.” 31, 1953,
"NDA-20 (Jan, 23, 1956). |

4A. D. Callihan et al., ANP Quar, ng. Rep. March 10,
197536 ORNL-206), p 64; Dec. 10, 1955, RNL 2012,
P

over the 283/-|n. length of the outer reflector and
comprised 70% of the outer cylindrical layer re-
sulted in o loss in reactivity of $2.39.

Measurements of Gamma-Roy Heating in Beryllivm

Capacitive ionization chambers were used to
measure the distribution of gamma-ray heating in
the beryllium of the island and the reflector of the
critical assembly by a method developed at the
Knolls Atomic Power Laboratory.> The chambers
are constructed of beryllium and have a 10-mil-
thick annular cavity 5’16 in. OD and '78 in, deep,
The results are expressed as power dissipated as
heat in a unit volume per unit reactor power. The
reactor power was determined from a calibration
based on the intensity of a fission-product gamma
ray from an irradiated uranium foil.%

A plot of the data obtained from radial traverses
3/4 and 109/‘6 in. from the mid-plane of the reactor is
given in Fig. 4,4.1. Three longitudinal traverses,
one along the axis of the reactor and the others
4'5’ and 7/ in. from the axis, are plotted in
th. 4.4.2, These data have not been corrected
for the ionization resulting from the (n,p) reaction
in the air-filled chambers and may overestimate

‘the heating adjacent to the fuel by as much as

25%, an estimate based on the work at KAPL.
An attempt was made to measure this error by
fitling the chambers with CO,, but it was apparent
that they were not gas-tight during these experi-

‘__ments._ It may be possible to maoke a correction
“to these data by using the results of similar
“measurements in ancther assembly.

-A layer of beryllium, 2/8 in. thick, was removed

from the top of the reactor, and one traverse was
‘repeated with this thinner reflector, = The heating
- was observed to be unaffected in the region be-
~tween the fuel and a point 3 in. from the surface
- of the modified reflector, - In this outer 3-in. layer
" the heating decreased to a value, ‘ot the surface,
- about 35% less than it was at the same,d:stonce
B from fhe fuel in the thrcker reflector. -

 

' SC ‘A. Rich and R. E. Slovocek Gamma-Ray Heatmg
gd;sas.s')urements in the .S‘IR PPA- 18 KAPL-866 (Jan. ,

6S. Snyder, Absolute Determination of Power Produced
;rg a]hslg;nmally Zero Power Reactor, ORNL-2068 {May

255

 
 

 

 

ANP PROJECT PROGRESS REPORT

BTN T
ORNL-LR-DWG 44448

 

 20x1077 /-

 

 

 

1.8x 1077
34‘ in. .
FROM MIDPLANE
1621077
1451077

L/

 

1.2x40°7

 

 

est— BERYLLIUM ISLAND —weteest— FUEL REGION —— it \ BERYLLIUM REFLECTOR

 

1.0x 40°7

 

8.0x {078

GAMMA-RAY HEATING (w/emS-w)

AT | A
/

 

 

6.0x 4078 e

 

10%¢ in. FROM
MID PLANE (% in.
FROM FUEL ANNULUS)

N\

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

401078
\
20x 1078 - \\
—->
0
0 2 4 6 8 2 1 6 18 20 22

DISTANCE FROM REACTOR AXiS {in.)

Fig. 4.4.1. Radial Distribution of Gamma-Ray Heating in the Compact-Cu;e Reflector-Moderated-

Reactor Critical Assembly,

Fa stheufion Leakage |

Relative measurements of the fast-neutron leak-
age at points on the outer surface of the reflector
were made with a Hornyak button, 2 in. in diameter
and ‘/4 in. thick, mounted on @ Du Mont 6292 photo-
multiplier tube. The button had the same compo-
sition, 0.15 g of powdered ZnS in 1,0 g of Lucite,
as a similar button described by Hornyak.” By
proper choice of a discriminator bias it was
possible to detect neutrons with energies above

 

7W. F. Hornqu, Rev, Sci, Instr. 23, 264 (1952).

256

0.5 Mev against the gamma-ray background of the
reactor, : : . S

A polar plot of the fast-neutron leakage distribu-
tion observed in a fongitudinal traverse over the
top of one end of the reactor is presented in
Fig. 4.4.3. The observed counting rates, nor-

malized to the counting rate at point 19, near the -

axis of the assembly, have be’e'rjl“ plotted on the
polar radii drawn through points on the reflector

surface where measurements were made.. For this

traverse, the reflector thickness was approximately
”'/2 in., and the top layer of beryllium wczs,2834

.
 

 

 

. (x108)

PERIOD ENDING JUNE 10, 1956

seCRET™
ORNL~LR—DWG 14419

 

20 g

 

 

' \5/@1. FROM AXIS
16 _ _

 

"

 

 

 

 

  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 DISTANCE FROM REACTOR MIDPLANE (in)

= '2 ¢

nfi “—.—:._‘:\- AXIAL

3 N

g .

=
- g 10 o ‘
. I

X .

g \,

P ‘\

% 8 \\.
" g \. .

. Y\
N,
N, |
. N~.
. \‘ ]
| 73/,6m FROM AXISN \ \.\
|t CORE REGION —————————— ~ "'-:-.\[’\
| R AR N TS
S 0. 0 2 - 4 6 B 10 g2 14 16 18 20

Fig. 4.4.2. Longitudinal Disfribuhon of Gammu-RcyHeafing in the Compact-Core Reflector-Moderuted-

Reacfor Crificcl Assembly.

257

 
 

 

 

ANP PROJECT PROGRESS REPORT

 

 

 

 

 

 

 

 

 

 

FRcRETe
ORNL-LR-DWG 14420
10
9 |— ]
8 MEASURED WiTH 2-in.-DIA HORNYAK BUTTON, —
TOP BERYLLIUM LAYER: 48%¢ x 28%, in.
7 ]
"
€6 — o 1 2 3 a4 \ 5 6 7 —
3 4 3 1 5 ) y 5 )
ey ' T ¥ ¥ T T T
5 | |
5
2 I 5 8
> 5 — t 7
-
c |
o
- l BERYLLIUM 5
= REFLECTOR :
w ¥
Z 4 b l -
5
- I <+ 10
&
2% in. 4+ i
6
3 Iill + 2 —
3
oy 44
Q [} -:_Lé—_——
EI 3.___ o\.
| 7 {2./ | 15 \.
' 9
2 |
B FUEL REGION ¢ .’/ 1. ©
. 10
I T {6 . 47.
|
i1 — 4147 _
BERYLLIUM ISLAND " 18
AXIS T 19
0 —_ e e e e e e
L 20 20

 

 

 

 

 

Fig. 4.4.3. Polar Distribution of Fast-Neutron Leakage at Surface of llé-m.-Thlck Reflecfor of
Compact-Core Reflector-Moderated-Reactor Critical Assembly.

258

.
 

 

 

i

&

in. long, 18! ]ffi in. wide, and 2/ in. thick. The
beryllium layer immediately below it was 34 R
long and 24/16 in. wide,

The top layer and the corners of the second
beryllium layer were subsequently removed, and
the resulting reflector was 8% in, thick. The
new top layer was ]8”/6 in. wide and 3434 in,
long, A second longitudinal traverse was then
made. The results are shown in Fig. 4.4.4, in
which the top lobe is plotted to a scale one-half
that used to plot the end lobe.

Two series of lateral traverses were also made
at the side of the reactor at several distances
from the mid-plane. For reflector thicknesses of
'Hl/2 and 8% in, the outer slab dimensions for
each side reflector thickness were the same as
those given above. The latitudinal variations of
the fast-neutron leckage for the two reflector
thicknesses are shown in Figs. 4.4.5 and 4.4.6.
The curves show a slight asymmetry about the
center of the reactor, which is attributed to the
reflection of neutrons by the aluminum and steel
structure on which the critical assembly rests.
The area under each curve of Figs, 4.4.5 and
4.4.6 is proportional to the counting rate observed
at the corresponding distance from the mid-plane
in the longitudinal traverses made ot the top of
the reactor., By integrating the results of the two
longitudinal traverses over the distance from the
mid-plane, it was possible to determine that the

removal of the 27/ in. layer of beryllium from the

top of the reactor mcreased the fast-neutron leak-
age there by o factor of 3.7. '

RELATIVE INTENSITY (arbitrary units)
o

PERIOD ENDING JUNE 10, 1956

SECREY.
ORNL-LR-DWG 14424

 

   
   
   
 
   
  
  

10
I,
9 [— ]
8 — MEASURED WITH 2-in.-DIA |
" HORNYAK BUTTON
TOP BERYLLIUM LAYER:
18 X 3434 in.
7 pom —_

NOTE:
THE DATA AT THE TOP
OF THE REACTOR ARE
PLOTTED TO A SCALE
ONE-HALF THAT USED
FOR PLOT OF THE END
I

3 |- €,
' o (2345

    

6/7g o

  
  

!

BERYLLIUM ]
REFLECTOR o4

I
FUEL
/ REGION 43\'

       

 

  

MIDPLANE

 

 

7 1
2.'73 in.——-l I--

-BERYLLIUM ISLAND *20 0%

AXIS

 

 

 

 

 

 

Fig. 4.4.4. Polar Distribution of Fast-Neutron
Leakage at Surface of _8"‘/a-in.-Thick Reflector of
Compact-Core Reflector-Moderated-Reactor Critical

Assembly,

259

 
 

 

ANP PROJECT PROGRESS REPORT

 ORNL—LR~DWG 14422,

 

1.5

 

. .
MEASURED WITH 2-in.-DIA HORNYAK BUTTON »'/ \

OUTER BERYLLIUM LAYER: 18'/16 x 28 %4 in. / \ '

- d = AXIAL DISTANCE

FROM MIDPLANE OF REACTOR / Cin. ‘ \

S

N

 

 

{
<

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

g o //<d=4.69 in. ) ‘ \\\ \
g T ‘ |
: N \.,/ / /4d= 8.25 in. ' \\
] % | )
e -
v
~ /// J=1294in T
o181 Y CENTER LINE
"1 OF REACTOR
o _ :
-4 -2 0 2 4 6 8 10 12 14 16 18 20

DISTANCE FROM ONE EDGE OF QUTER BERYLLIUM SLAB (in.)

Fig. 4.4.5. Latercl Distribution of Fast-Neuiron Ledakage at Surface of llg-in.-'l'hick Reflector of
Compact-Core Reflector-Moderated-Reactor Critical Assembly,

260

o
 

 

(1)

 

PERIOD ENDING JUNE 10, 1956

SEGRET
ORNL-LR-DWG 14423

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

7
1+ T 1 T 7 |
MEASURED WITH 2-in.-DIA HORNYAK BUTTON
OUTER BERYLLIUM LAYER: 18'Y,¢ x 343 in. '
6
_ - d=0in.
o= AXIAL DISTANCE FROM MIDPLANE _ e
5 OF REACTOR A
w T
E ) '
d=4.75in
4 : Ty | \\
§ 4 ' i \'\\‘ \
§ / ' _‘,/ ' ™\ \
N W
s - d=8.25i \\
& / .29in.
E 3 A /‘: ] ~—
& / '/ /"/ « NS
~ | / . o
2 - v ‘/ § \"\.
N w |
&
1 — d=14.63in
_ . v |
| -‘-_-._'_'l'-fl
o I

 

 

 

 

 

 

 

 

 

-2 -t 0 1 2 3 & 5 6 7 8 9 40 W 12 13 44 5 16 17
| DISTANGE FROM ONE EDGE OF OUTER BERYLLIUM sLaB (in.) |

Fig. 4-4.6. Luferul Dnsiribuflon of Fust-Neuiron Lenlu:ge at Surl"ace of 85’ -ln.-Thicl: Reflector of
CompacioCcre Reflector-Modemted-Reactor Crmcul Assembly,

261

 
 

 

 

 
 

| e b et it

 

»

s

 

Part 5

REACTOR SHIELDING

E. P. Blizard

 
 

 

 

 

e ancam, ool

e bt
 

 

 

A

 

5.1. SHIELDING THEOQRY

C. D. Zerby

DOSE RATE IN A CYLINDRICAL CREW
COMPARTMENT RESULTING FROM
AIRSCATTERED GAMMA RAYS

C. D, Zerby

The problem of determining the contribution of
air-scattered gamma rays to the dose rate in a
cylindrical crew compartment has been separated
into two parts. In the first part the gamma-ray flux
in air ot various distances from an idealized point
source is to be obtained. In the second part the
flux is considered to be the source at the ocutside
surface of the crew compartment, and the dose rate
at several arbitrary positions inside the crew com-
partment is to be determined. The results ob-

tained for both parts of the problem will then be

integrated, The complete problem is a joint effort
of the Wright Air Development Center (WADC) and
the Oak Ridge National Laboratory, with the bulk

~of the computation to be done by the Monte Carlo

method and to be performed at WADC,

The first part of the problem has been programed
and coded for calculation on a type 1103 automatic
computing machine, and test cases are being run
in preparation for a complete parameter study. In
this part of the problem the angular distribution and
the energy spectrum of air-scattered gamma rays
will be determined at several separation distances
from a point source. The source will be considered

to be monoenergetic and to be emitfing gamma rays -

symmetrically about the source-detector axis in a
conical shell. ‘The parameter study will, therefore,
include a survey of various separation distances,
source energies, and apex angles of the conical
shell beam. With the results of the parameter study

it will be possible to obtain the radiation current -
ot eaéh ‘separation distance for any point source
which emits gamma rays. symmetrically about the
' source-detectcr -axis with any energy- spectrum,
‘Although the density of air at sea-level conditions
will be used in the calculations, it will be possible

to obtain the results at any altitude (different

density of the ulr) by using the transformations
'developed for this purpose and reported below, -

- The anoly5|s and procedure for the second part
of the problem are complete and have been re-
ported.! Results of the calculation will include
the detailed angular distribution and energy spec-

trum of the radiation entering the crew-compartment
cavity and the dose rate at various positions in
the cavity, A parameter study will be made for
various thicknesses of lead ond polyethylene in
the crew-compartment walis.

RADIATION FLUX TRANSFORMATION AS A
FUNCTION OF DENSITY OF AN INFINITE
MEDIUM WITH ANISOTROPIC
POINT SOURCES

C. D. Zerby

The transformation of flux, current, or dose rate
as a function of density of an infinite homogeneous
medium with anisotropic point sources can be
derived directly from the Boltzmann equation.?
This transformation is particularly of interest be-

-cause of the many calculations which use this

geometry, and, in addition, it provides a means of
transforming the Tower Shielding Facility (TSF)
sea-level dose-rate data to data at any altitude.

The transformation is obtained by writing the

Boltzmann equation for an anisotropic point source

at the origin in nondimensional form:

og(E) -+
G(A,E,Q)

 

1) 2V GLEQ) +
9L,

 

as(E) s d -~ e d
- f.]:’ /(ElQlE"fl') G(A'E"Q') dE’ dfl'
JpJQ) 9(Eg)

S(E,Q)
+
27A2

500 5(1 = ) |

 

where

unit vector,

20
il

= energy,

Ey = arbitrary fixed energy,
o (E)} = microscopic scatteringcross sec-
tion atrenergy'_E,_cm_z,

 

ic, o, Zerby, A Monte Carlo Method of Calculalmg
the Response of a Point Detector at an Arbitrary Position
Inside a Cylindrical Shield, ORNL-2105 (June 12, 1956).

2¢, p. Zerby, Radiation Flux Transformations as a
Function of Density of an Infinite Medium with Aniso-
tropic Point Sources, ORNL-2100,

265

 
 

 

 

(5)

 

ANP PROJECT PROGRESS REPORT

ot(E) = microscopic total cross section
at energy E, cm?,
o(Eq) = microscopic total cross section

ot energy, EO' cm?,
A = scalar distance from source
point measured in units of mean
free path at energy E,
- . -
A = position vector from source
. point,
e -
f(ElQ'E "Q ')
- -
G(ME Q)

scottermg kernel

' parflcle current particles per

unit energy at energy E per unit

‘solid angle in the duechonfiper

square mean free path (ot energy
E ;) at position A,

o

- -_’_-A - '
S(E, ) = source  strength, "particles per
""" unit energy at energy E per unit
+ solid angle in direction {}.

Equahon 1 is mdependent of the nuclear density

" of the medlum, and thus G(/\,E fi) is also inde-
"pendent of ‘density.  This means that if two ex-

periments are set up with the same point source

“but with different densities p, and p, and the

current is measured in each experlmenf at positions
such that

- >
(2) A’i = Ag ?
then
b - -5 -5
(3) Gpl()\l,E,Q) = sz(Az,E,Q) .

The significance of Eq. 3 can be realized if A and
G are reduced to conventional units. The relation
between distance r (measured in centimeters) and
A is given by

(4) X = No(Eg)r

where N is the nuclear density of the medium. The

‘conventional particle current can be defined as

F(7, Efl) given in particles per unit energy at
energy E per unit solid angle in direction ( per
cm? at position 7. The relation between F and G
is then

F(r,E,Q)

2.2
Noy(E o)

= GIA(),E,Q -

266

By using Eq. 4, Eq. 2 becomes

-» Nl-& Py o
r ——— ———

= r :.-.——T
2 N2 ]7 p2 1

(6)

and by using Eq. 5, Eq. 3 becomes

(7) F- - Py o Efi

 

N2 T .1’2 .
2 F GLED) =——F, (s ED)
N2 Fo,(r B = 2 PR
i | 2

which is the desired transformation.” Proper inte-
gration of Eq. 7 gives the flux transformation .as

: ’ 2
(8) ‘i’p 2 (-;2 =

Pi ., o " o S
ruE) =— ¢’p (rp_E)
and the _dose rate tronsformati.on as

"2
(9 D - [ TS r -
) ps ry = —p—z- )= Dp](r,) .

  

These transformations can be applied to much of

the TSF data, but the application is not -entirely
general,?2 To make full use of the transformations
it will be necessary to obtain additional data at
the TSF at several separation distances so that
the measurements can be interpolated and applled
at any altitude. :

ENERGY ABSORPTION RESULTING FROM
GAMMA RADIATION INCIDENT ON A
MULTIREG[ON SHIELD WITH '
- SLAB GEOMETRY '

S, Auslender3

The code for a Monte Carlo calculation of energy
deposition in a multiregion shield with slab
geometry“ has been used to obtain the results for
1-Mev gamma rays incident on a slab consisting of
regions of fuel, Inconel, sodium, and Inconel again.
A diagram of the composite slab is shown in Fig.
5.1.1, which ‘gives the normal thicknesses in

 

3On assignment from Pratt & Whitney Aircraft.

43, Auslender, ANP Quar, Prog. Rep, March 10, 1956

-)

h

 
 

 

 

 

L

&

 

 

 

seencr
2-01-059-67
4
-Mev
GAMMA RAYS FUEL\!NCONEL SODIUM INCONEL
7 7
NaF~ ZrFy-UF, / /
(52.5-42.5-
5.0 mote %) ) % %
J
NN 2
N\ 379\ \V05767 2149 0488
™~ {cm)

NNV o
NN\

Fig. 5.1.1. Geometry of Fuel-lnconel-Sodium-
Inconel Slab, '

 

st (mfp)

 

 

 

 

 

centimeters ond in mean free paths. The fuel is
NoF-ZrF .UF, (52.5-42.5-5 mole %). The per-
centages of energy absorption throughout the slab
for various angles of incidence of gamma rays are
given in Fig, 5.1.2, The percentage of the total
energy incident on the slab that is reflected, ab-
sorbed, or transmitted for each angle of iricidence
is shown in Table 5.1.1, and the percentage of the
energy thot is absorbed in each of the four regions
of the slab is shown in Table 5.1.2,

A plot of the dose buildup factor as e function
of the distance through the slab, in mean free
paths (mfp), is presented in Fig. 5.1.3. The sharp
break in the curves at 2,75 mfp is due to the rapid
attenuation of low-energy gamma rays (degraded
principally in the sodium) by the final Inconel slab,

PERIOD ENDING JUNE 10, 1956

SESREP
2-01-059-88

ISCTROPIC

0.5

ENERSY ABSORBED (% OF TOTAL ENERGY)

0.2
o4

- 005

002

 

0.01
0 5 10 15 20 25 30

#, NORMAL DISTANCE INTO SLAB {¢m)

" Fig. 51.2. Percentage of Total Energy from

Incident leMev Gamma Rays That Is Absorbed in
a Fuel-lnconel-Sodium-Inconel Slab as o Function
of Normal Distance into the Slab,

267

 
 

 

ANP PROJECT PROGRESS REPORT

TABLE 5.1,1. PERCENTAGE OF TOTAL ENERGY
FROM INCIDENT 1-Mev GAMMA RAYS THAT IS
REFLECTED, ABSORBED, OR TRANSMITTED
IN A FUEL-INCONEL-SODIUM-INCONEL SLAB

 

 

 

0, Angl.e of Percentage of Total Energy
Incidence
(deg) ' Reflected Absorbed Transmitted
0 1.151 88.01 10.79
45 2.45 93.36 4.19
60 - 4.10 94,30 1.59
70} - 8.76 90.60 0.635
75!5 ' 13.22 86.05 0.732

 

TABLE 5.1.2. PERCENTAGE OF TOTAL ENERGY
FROM INCIDENT l-Mev GAMMA RAYS THAT
IS ABSORBED IN EACH REGION (R;, R,,
Ry, AND R,) OF THE FUEL-INCONEL-
SODIUM-INCONEL SLAB

 

Percentage of Total

 

 

sec 0
{0 = Angle Energy Absorbed
of incidence) Rl R,y R, R,

] 37.86 29.40 15.04 5.752

V2 48.39  30.97  10.65 3.423
2 ' 60.23  26.10 6.14 1.895
3 69.7 17.32 2913  0.697
4

726 10,52 2,144 0.692

 

 

268

€TCRE o
2-01-059-89

: 8 = 70% deg

DOSE BUILDUP FACTOR

. §=60
=45 deg

8 =0 deg

 

0 4 2 3
#, NORMAL THICKNESS THROUGH SLAB {mfp)

Fig. 5.1.3. Gamma-Ray Dose Bui'ldup Factor as
o Function of Normal Thickness Through the Fuel-
Inconel-Sodium-Inconel Slab. S
 

i

 

-PERIOD ENDING JUNE 10, 1956

5.2 LID TANK SHIEL DING FAClLlTY
R. W. Peelle

STUDY OF ADVANCED SHIELDING MATERIALS

W. R. Burrus! J. M. Miller
W. J. McCool?2 ~ D.R. Otis?
- J. Smolen?

An extensive mockup experiment was initiated
in which combinations of advanced shielding ma-
terials, such as lithium hydride, depleted uranium,
zirconium, and tungsten, are being investigated.
These tests are important to the ANP program

because lead and water, the prototype shielding

matérials‘ used in most of the previous Lid Tank
Shielding Facility (LTSF) mockup tests, are not
optimum -aircraft construction moterials. In the
initial tests the shielding materials being studied

are combinations of lithium hydride and other
shielding matericls immersed in transformer oil.

The particular configurations used have been those
of immediate interest in the GE-ANPD program, but

such general interest surrounds the use of these

materials that the more reliable data obtained thus
far are presented here. This report is preliminary,

however, in the sense thct little analysrs work has
-been performed.

 

: _]_On assignment from U. S. Air Force,
20n assignment from Pratt & Whitney Aircraft,
On assignment from Convair, San Diego.

'/4 in~THICK AI PRESSURE PLATE o
- Mgein~THICK Al PLATE .~ e
|/ Yhe-in~THICK; 21% ENRICHED um sounce
vs-mqmcx BORAL SHEET -~ . -

      
 

- / ~in. -mncx Al SOURCE~PLATE COVER R ,
f 'fwn -Tmcx ou. LAYER ™ REc:-:ss OF TANK WALL

\

Thermal-neutron flux and gamma-rcy tissue dose-
rote measurements have been made for conf:gura-
tions involving combinations of zirconium, lead,
and depleted uranium with lithium hydride and
transformer oil. All the measurements were made
in the oil along the axis of symmetry of the LTSF
source plate. (Fast-neutron dose-rate traverses
were also made, but experimental difficulties pre-
vent the publication of reliable data at this time.)

A typical configuration (No. 69-6a) is shown in
Fig. 5.2.1. The various material combinations
studied differed only in the region to the right of
the line marked ‘'beginning of configuration.” The
parameters of all the configurations are given in
Tableé 5,2.1, and the known properties of the
various materials are given in Table 5.2.2. The
lithium hydride is the only material that was

- ¢anned in an extraneous material.

- All the radiation levels indicated in plots of the
data (Figs. 5.2.2 through 5.2,9) are quoted per watt
of effective source plate power, based on a total
power of 5.5 £ 0.5 w for the 28+in.-dia source plate.
A final evaluation of the effective power may ne-
cessitate a small change in.the results presented
here. All the thermal-neutron fluxes reported are
equal to the neutron density times 2200 m/sec.

The decreasing magnitude of the slope of the
neufron traverses shown in Fig. 5.2.2 at large

GEGRETF
2=01~QE7T—-69—272

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

S

TR SN N %S L T e
N. ¥ -"§§ 2 sl ==
T : L N - T 4 1 4 - —_——
NS N s (58 IS 8 / T T
\ 'S : A 3 =8 SRHS %

Nz ¥ 2p8 152 L X DETECTOR
NS N 2Eg =2 08 POSITION
NN e iEE Z _ posiTion

- NE W "g’g- shTRR T "
SOy “Ehg {2 '
- §5 Q;‘ EEa 1R TR £ -]
ENER\ B TR n TR
S8 NN TS LEOT XIS - - - - —
St \ e NVA e Pl c RHNR '
\ N .//'k TE F ol o"‘:“’":’:’o’ooo
R NN - TN RSRRRRIRRRS ]
U TTBOURCE | .. 07 .+ Fe-BEGINNING OF CONFIGURATION . .~ ...~ - "
PATE e g -
_ . - - . Lo T~ K . T L. ’ - ,,'- . e
: -7,

 

~ Fig. 5.2.1. Typical Configurution for LTSF Mockup Tests of Advanced Shielding Materials.

269

 
 

e e e

 

 

ANP PROJECT PROGRESS REPORT

TABLE 5,21, SUMMARY OF THE CONFIGURATIONS
USED FOR LTSF MOCKUP TESTS OF ADVANCED
SHIELDING MATERIALS

 

 

- Configuration
" Ne. ' . Composition
69-0 Pure water
Transformer oil
69-1 1 #t of LiH in oil

2 £+ of LiH in oil
-3 ftof Li_H in oil

69-2 4 in. of Zr in oil :
4 in.of Zr+ 1 £t of LiH in oil -
- 4 in..of Zr + 2 ft of LiH in oil
4. in. of Zr + 3 ft of LiH in oil

69-6 * 3 in. of Pb in oil
. " 3 in.of Pb+ 1 ft of LiH in oil
'3 in. of Pb+2 ft of LiH in oil
3'in. of Pb +3 ft of LiH in oil

69-7 3 in. of U in oil
: S 3m. of U+1 ftof LiH monl
3in. ofU+2 ft of LiH in oil
3 in. of U +3 ft of LiH in oil

 

distances from the source is interpreted as an
effect of photoneutron production in C1? and deute-
rium, although no quantitative analysis has been
made. At smaller distances from the source, these
data may be used to estimate the macroscopic re-
moval cross section of lithium hydride at room
temperature in an oil medium. A preliminary value
of 0.12 em~! was obtained, which is in good agree-
ment with the value expected on the basis of pre-
vious removal-cross-section measurements on a
lithium-metal slab.4

The observed gomma-ray tissue doses are given

in Figs. 5.2.6 through 5.2,9. It can be noted that

270

TABLE 5.22. PHYSICAL PROPERTIES OF
THE SHIELDING MATERIALS TESTED

 

_ Material ' Description

 

Density, 0.87 g/cm3 at 20°_C;
onalysis, 86.7 wt % C and
12,7 wt % H

5x5x%1 ft slabs encosed in
Al cans (/-m.-thu:k wulls),
density, obout 0.75 g/crn ;
purity, about 95% -

Transformer oil

Lithium hydride*

Zirconjum* 52 x 55 x 2 in. metallic slabs

. Lead 55 X 60 X 1.5 in. me-tallic slobrs_
Uranium 52 X 55 X 1.5 in. depleted me-
tallic slabs contolnmg 0.24
wt % Y233 '

 

*Raw material s furnished by GE-ANPD.

one slab of lithium hydride behind a heavy shield-
ing material reduces the dose at larger distances
from the source. This is interpreted as a reduction
of secondary gamma-ray production caused by the
presence of Li® in the lithium hydride. The data
obtained upon subsequent additions of LiH indi-
cate, as expected, that LiH does not have so large
a macroscopic gamma-ray absorpflon coefficient as
that of the oil.

It is planned to continue these studies of combi-
nations of lithium hydride with other shielding
materials. Attempts will also be made to obtain
reliable fast-neutron measurements, where pos-
sible, for the configurations already studied.

 

4G. T. Chapman and C. L. Storrs, Effective Neutron
g;m?;gé)Cross .S‘ectums for Shielding, ORNL-1843 (Aug.

-
 

 

 

;TR
s 2-04-057-69- 264
10

BASED ON SOURCE PLATE
POWER OF S5 w

nv,, THERMAL—~NEUTRON FLUX (neutrons/cm? . sec.w)

PURE H,0
PURE OIL

CONFIGURATION 69-~0

faft OF LIH
CONFIGURATION 69- -4 {2 ft OF LIH
3t OF UH

10~2

5 DISTANCE, FROM SOURCE PLATE: (cm)

Fig. 5.2.2 Thermul-Neutron Flux Truverses for

Conf:guruhons 69-0 cnd 69-1.

n v, THERMAL-NEUTRON FLUX ( neutrons /cm?- sec - w)

 

. 40 60 80 100 1220 440 160

PERIOD ENDING JUNE 10, 1956

oEenTY
2~-01-057—-69-268

 

in, OF Zr

PURE OIL

2 4in. OF Zr AND 1 ft OF

4in. OF Zr AND 2t OF

4in. OF Zr AND 3 ft OF

O 20 40 60 - 80 100 120 140 160
23, DISTANCE FROM SOURCE PLATE (cm)

“Fig. 5.2.3. Thermal-Neutron Flux Traverses for
Configuration 69-2.

271

 
 

 

 

ANP PROJECT PROGRESS REPORT

SEeRp™
2-04-057-69-271

2 BASED ON SOURCE PLATE
POWER OF 55 w

v, THERMAL~NEUTRON FLUX (neutrons/cm2-sec:w) .
N

0!

3 in. OF Pb AND 4 ft OF LiH
2 3in, OF Pb AND 2 ft OF LiH
-2 3in. OF Pb AND 3 ft OF LiH

 

103
0 20 40 60 80 400 120 440 160
2, DISTANCE FROM SOURCE PLATE {em)
o

Fig. 5.2.4. Thermal-Neutron Flux Traverses for
Configuration 69-6. . ' :

272

omeRET .
2-01- 05769266

BASED ON SOURCE PLATE POWER OF 55 w

3in.OF U

3in.OF U AND 1t OF LiH

A Vo, THERMAL~NEUTRON FLUX (neutrons/cm?+sec * w)

3in.OF U AND 2t OF LiH

3in, OF U AND 3 ft OF

 

o 20 40 60 80 100 120 140 160
Zsyr DISTANCE FROM SOURCE PLATE (cm)

Fig. 5.2.5. Thermal-Neutron Flux Traverses for |

Configuration 69-7.

cy
 

 

 

c

SEONGT
2-01-05T-69-265

BASED ON SOQURCE PLATE
POWER OF 5.5 w

3 £ OF LiH
2 ft OF LiH

69-4
69-0, PURE OIL
€9-1, ¢ 1t OF LIH

2 69-0, PURE WATER

GAMMA-RAY TISSUE DOSE RATE (ergs/g-hr-w)
5

2

 

w0
20 40 €0 80 400 120 140 160

%, DISTANCE FROM SOURCE PLATE {cm)

Fig. 5.2.6. Gamma-Ray Tissue Dose<«Rate Trav-
erses for Configurations 69-0 and 69-1.

wpreney
2-01-057-69-269

'03

PURE OIL

4 in. OF Ir

4 In. OF Zr ond § f1 OF LIH -
4in. OF Zr and 2 ft OF LIH.
= ~4in. OF Zr and 3 ft OF LIH

gl " BASED ON SOURCE PLATE-.
o ~ -~ POWER OF 5.5w

.'\"GAMMA'-R'AY TISSUE DOSE RATE (ergs/g-hr-w)

 

20 - 40 80 - 80 - 400 . 420 - 440 460
jlso; DISTANCE FROM SOURCE PLATE (cm) S

Fig. 5:2.7. Gamma«Ray Tissue Dose<Rate Trave
erses for Configuration £9.2.

PERIOD ENDING JUNE 10, 1956

SECREF
2—-01-057-69-270

o PURE OIL

3in. OF Pb AND 1t OF

GAMMA-RAY TISSUE DOSE RATE (ergs/g-hr-w)

BASED ON SOURCE PLATE

 

SRR
2—-01—037—69-267

5 POWER OF 5.5 w

Bin. OF U AND it OF

. GAMMA-RAY TISSUE DOSE. RATE (ergs/g-hr-w)

3 U 2%t OF

 

2 f— ————3in. OF U AND 31t OF
20 40 60 - 80 100 - 120 140 160
2, DISTANCE FROM SOURCE PLATE (cm) -

Fig. 5.2.9. Gamma.Ray Tissue Dose-Rate Trav-
erses for Configuration 69.7.

273

 
 

 

 

ANP PROJECT PROGRESS REPORT

5.3. BULK SHIELDING FACILITY

F. C. Maienschein

GAMMA-RAY STREAMING THROUGH THE NaK
PIPES THAT PENETRATE THE ART SHIELD

T. V. Blosser D. K. Trubey

As reported previously,' o mockup experiment
hes been initiated to investigate the increase in
the dose rate outside the ART lead shield caused
by gamma-ray streaming through the NaK-filied
pipes that penetrate the shield. Thus far gamma-
ray dose-rate measurements have been made in the
water beyond straight-through penetrations, that
mock up portions of the north-head ducts. In future
experiments, measurements will be made beyond
similar penetrations that mock up portions of the

 

T, V. Blosser, ANP Quar. Prog. Rep. March 10, 1956,
ORNL-2061, p 249.

south-head ducts. Ducts placed through the shield
at a 45-deg angle will also be investigated, and
final measurements will be made on actual NaK
pipe mockups. ,

The mockup of the north-head straight-duct pene-
tration is shown in Fig. 5.3.1. It consisted of a
3‘/8-in.-|D Incone! ‘pipe filled with air or aluminum
turnings and surrounded by an air annulus that
simulated the insulation of the ART. = The duct

was housed in a watertight fhin-walled_(_}’w-in.)-

aluminum case. The aluminum turnings, which were
pressed into the pipe to a density of 0.85 g/em?,
simulated the NaK of the ART. A tungsten collar
was fitted around the top of the duct to form a step
which would reduce the streaming of gamma rays
through the insulation. |

StererT )
ORNL~-L.R-DWG 14895

'/—TUNGSTEN COLLAR

 

 

.

§

 

 

——— 2T in——

 

3% in.
5)Z =0,

 

 

35/3 in.

 

 

5% in.

 

4% in.
INCONEL PIPE
AIR

    

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

/—AIR- OR ALUMINUM-FILLED DUCT

Yi6-in~THICK ALUMINUM LINER

 

4.254in. ~THICK LEAD -

AIR

Yerin.~THICK BORAL

0.040-in.~THICK CADMIUM
DISK (42-in. DIA)

M %-in.~THICK ALUMINUM
. (BOTTOM OF TANK)

Fig. 5.3.1. Mockup of the North-Head Straight-Duct Penetration of the 4.3-in,-Thick ART Lead

ShiEIdo '

274
 

 

 

o

M

M

Ny

 

In order to determine the magnitude. of the gamma-
ray dose rate produced by streaming through the
duct, it was necessary to know the dose transmitted
through the lead shield, as well as the effect of
the background produced by adjacent experimental
facilities, Measurements of the gamma-ray dose
rates were therefore made beyond a solid lead
shield in the x,y plane of the source ot a distance
z above the top of the lead (see Fig. 5.3.1). The
experimental data (Fig. 5.3.2) are presented in
terms of the ratio of the dose rate with the duct in
position to the dose rate with the duct replaced by
a lead plug as a function of distance from the
duct center line. The ratio of the dose rate beyond
a water-filled hole to that beyond the solid lead is
also shown in Fig. 5.3.2,

The actual ART duct mockup will consist of two
concentric pipes. The inner one, which will con-
tain the NaK, is not present in these straight-duct
mockups, but it would reduce the dose by about
30%. It should also be pointed out that the Inconel
pressure shell is not part of the mockup shield. |f
it were, the dose-rate ratios would.be higher by a
factor of approximately 2 for a 1l-in.-thick pressure
shell.

A ratio of the dose raote with the aluminum-filled
pipe to that with the air-filled pipe was 0.86 on the
center line, which is in reasonable agreement with
a calculated value. In the calculation the average
energy of the gamma rays was assumed to be 3.5
Mev from the composite source, The ratio was
determined from the following relation:

where

D, = dose| rate wnh plpefnlled with aluminum .
' ~ as compared with the experlmental vcllue of 0.86

~ turnings (w:th fungsfen collar), -

x = length of ducf(12 8 cm), .

po= alummum linear absorphcn coefflcuent-_'

= #mPJ

jzm = mass absorpflon coeff:cnent (0. 034) for "

3.5-Mev gamma rays in alummum, o

p Qdensnty of aluminum = 0.85 g/cm " o

B(ux)

= dose rate bunldup factor (1.2)."
Thus | o
_ D,
— = 0.83
D2

2 = dose rate through empty ptpe (no collor),'

- PERIOD ENDING JUNE 10, 1956

SECRET
ORNL~LR—DWGC 14896
500
| AIR-FILLED DUCTS

© x TRAVERSE, z = 0.5 cm
A yTRAVERSE z=105 cm

ALUMINUM-FILLED DUCTS
A y TRAVERSE, TUNGSTEN COLLAR, z = 10.5 cm
100 |y CENTERLINE MEASUREMENT, NO COLLAR, z = 10.5 cm

WATER-FILLED HOLE
50 | ® ¥ TRAVERSE, z = 10.5 cm

x=0

20

RATIO OF DOSE RATE S8EYOND DUCT
TO DOSE RATE BEYOND SOLID LEAD

05

0.2

DOSE RATE BEYOND
SOLID LEAD SHIELD (r/hr)

 

-60 —40 —20 0 20 40 60
x OR y, DISTANCE FROM DUCT CENTERLINE (cm)

Fig. 5.3.2. Gamma-Ray Dose Rates Beyond
Straight-Through ART North-Head Duct Mockups

Filled with Air or Aluminum.

mentioned above. "
It may be seen in Fig. 5 3.2 that the tungsten
collar offers little attenuation to: the dose along

. the center line of the pipe. However, the ratio of
‘the dose rates. with and without the tungsten collar
“and aluminum turnlngs, _uveraged over the emitted

beam, was 1.76. This ratio indicates the effective-
~ ness of the tungsten collar in this- partlcular con-
| flgurahon only. .

In the ART shield deslgn the !ead shleld is fol-

~ lowed by 31.5 in. (80.0 cm) of water.- The water

attenuation of the gamma-ray beam emerging from
the duct mockup (with aluminum turnings and tung-
sten collar) along the z axis is shown in Fig.

275

 
 

 

 

 

 

 

ANP PROJECT PROGRESS REPORT

5.3.3, and the broadening of the beam at 30 and 80 ORNL—LR-DWG 14898
cm is shown in Figs. 5.3.4 and 5.3.5, respectively, :

8

o

x=0,z=30 cm

The calculated ratio of the dose rate with the
duct in place to the dose rate with a solid lead
shield can be determined if an angular distribution
(cos™ 0) is assumed. If o is the number of photons
per unit area emitted in the positive direction from
a plane source and o’ cos™ 0 is the number emitted

M

10~

RATIO OF DOSE RATE BEYOND DUCT
TO DOSE RATE BEYOND SOLID LEAD

 

1 -
per unit solid angle at an angle @ from the normal, 5
then the dose rate (neglecting the buildup factor) 89
! ) , s 22
on the center line at a distance z from the plane of BT
i
9
.y
2 Yo
e
oEeRET »
ORNL~-LR~DWG 44897 10~2

4%10 -60 ~40 ~20 0O 20 40 60 .

7, DISTANCE FROM DUCT CENTERLINE (¢m)

Fig. 5.3.4. Gamma-Ray Dose Rates 30 cm
Beyond Straight-Through ART North-Head Duct
Mockups Filled with Air or Aluminum.

{0

SEGREY
ORNL-LR-DWG 44899

 

ALUMINUM-FILLED DUCT WITH TUNGSTEN COLLAR
x=0,z=80cm

RATIO OF DOSE RATE BEYOND DUCT
TC DOSE RATE BEYOND SOLID LEAD

w0t

RATIO OF DOSE RATE BEYOND DUCT TO DOSE RATE BEYOND SOLID LEAD
n

 

5
. of
Sq
Wi
ol
w X
=W
2 = Q
w
nJd
2 ge
-
. Q
107! @
20 40 60 80 100
z, DISTANCE FROM LEAD SHIELD {cm) ©-2
-20 =40 ~20 o 20
, ¥ DISTANCE FROM DUCT CENTERLINE (cm)
Fig. 5.3.3. Water Attenuation of Gamma-Ray ' o
Beam from the Straight-Through, Aluminum-Filled, Fig. 5.3.5. Gamma-Ray Dose Rate 80 ¢cm Beyond -

ART North-Head Duct Mockup with a Tungsten  Straight-Through ART North-Head Ducf Mockup
Collar, Filled with Aluminum,

276

1

3

tr
 

 

D(zr)

radius a (Fig. 5.3.6) is:

o’ cos” Ge~FR 2pp dp

fa ” m
=0 R2c

P

 

2n0’ R dR

PERIOD ENDING JUNE 10, 1956

]

_ (m + V)o

 

c

2 -
= [ ()
R R2

R=z

o uzl 14(a2/22)] 12
()™ f

Bz

 

e MR d(uR)

(uR)m'l']

Xx»

N

_(m+ o

where

o - -y
m f e™ dy
x ’
m+]

x Yy

/2 2
-affl f i cos™ @ sin 0d0d¢

Epai(x)

2no’

 

m+1 '

p = linear absorption coefficient,

flux-to-dose conversion factor,

H

c

These functions have been tabulated by Placzekz
for values of m up to 19,
If ais the radius of the duct and & is the radaus

of the source below the lead, then the ratio of the -
center-line dose rate with the duct in place to that

with the solid lead shleld in place is

m+l(l“z) -

1/2
a2

+ — ,
z2

Em+](’-lz) <]

 

distance from source to dose-measurement
position

C -z (distance from lead) plus thickness of

lead (4.3 in.),

gy = average linear absorption coefficient along
the center line with duct in place,
p, = average linear absorption coefficient along

the center line with lead in place.

The choice of the buildup factors B, and B, is
very uncertain. |t does not matter whether B, is
taken for aluminum or water, but, since the water

 

V 2G- Pluczek' .‘Thg Functiohs En(x) =I e-—x,U- ‘L—fl
' 1
dp," MT-1 '(n.d.). .

 

 

 

 

 

B | oy : | 2 |/2
Epnlimt) = ——————— E_ (1) (' t— By(pq)
. o 2 m/2 z2
. a
Lo | (1 +—-)
T Dyl LY F ~
Dy, a(t)__‘, . . ' ,
.,‘__70__ . oy | _ 52 1/2
Eulpngt) - - . , mfl(Fz‘) T+ — 1 |Byluyt)
o 2 m/2 z2 '
b
| z2 ]

277

 
 

 

 

 

ANP PROJECT PROGRESS REPORT

UNCLASSIFIED
ORNL—LR—DWG 14900

 

 

 

Fig. 5.3.6. Geo'métry for Calculation of Dose
Rate at Distance z from Source.

thickness following the lead is thin (~0.4 mfp),
B, was assumed to be the buildup factor for lead.

The calculated ratio is plotted as a function of
the angular distribution index m in Fig. 5.3.7.
There is excellent agreement with the experimental
value if m is taken to be 2. This is what might be
expected from o combination of an isotropic com-
ponent (cadmium capture gamma rays) and a highly
directional component (carbon capture and reactor

278

. SEGHES
ORNL—-LR—-DWG 14901

8

n
o

WATER-FILLED HOLE (z=6.6 cm)

3

6 .

ALUMINUM~-FILLED DUCT WITH
TUNGSTEN COLLAR {z=10.5cm)

o

M

 

-

4 6 8 10 12 14
m, ANGULAR DISTRIBUTION INDEX

RATIO OF DOSE RATE BEYOND DUCT TO DOSE RATE BEYOND SOLID LEAD
L+
r

Fig. 5.3.7. Calculated Ratio of Center-Line
Gamma-Ray Dose Rate Beyond Straight-Through
ART North-Head Duct Mockup to That Beyond
Solid Lead Shield as a Function of Angular Distri-
bution (cos™ &) of the Source, '

gamma rays). Of course, the buildup factor and the
monoenergetic source assumptions introduce some
uncertainty, Also, both the energy spectrum and
the angular distribution for the ART reactor are
undoubtedly somewhat different than those for the
mockup.

N

(4

i

e
 

2

 

 

 

PERIOD ENDING JUNE 10, 1956

5.4, SHIELD MOCKUP CORE

C. E. Clifford

Tests of optimized reactor shield designs are
made, currently, with a full-scale mockup of the
shield and @ known source. The effectiveness of
the shield in its final application is then predicted
by making source corrections., To date the source
for the shielding tests has often been a swimming-
pool type of reactor, but with the circulating-fuel
reflector-moderated reactor (CFRMR) now being
considered for aircraft propulsion it is considerably
more difficult to correct for the differences between
the two sources. In recent months it has become
apparent that a shield test with a reactor that more
nearly mocks up a CFRMR is required.

- The . construction of a 5-Mw reactor 1hat will use.

fixed fuel elements which will correctly mock up all
the radiation” expected from the CFRMR desngned
by -Pratt & Whitney, except the decay gamma rays

and neutrons from the circulating fuel in the heat" '_
. ‘quired by the CFRMR.

exchanger, has’ been proposed. A preliminary de-
sign ond a cost estimate are being. prepared on
which to base a formol consfruction proposal. The

and will be used with the latest optimized shield

at the Tower Shielding Facility (TSF) to determine

whether the calculated dose in the crew compart-

 _ment agrees with the measured dose. It will also
be used ot the Bulk Shielding Facility (BSF), -
where measwements will be made of the spectra
of neutrons and gamma rays emergmg from the_*

reactor shield.

Calculations for the SMC have thus for been
directed toward obtaining a good nuclear -mockup; -
the design effort has been directed toward making
a simple yet flexible over-all de5|gn.‘ ‘The SMC,
as it is now planned, is described below, and the o
differences between it and the present CFRMR. de-
~signs are noted. The SMC sources are then come f;fij

- pared with the CFRMR to show. how closely they Lo

- agree. The SMC- construction ‘program _ is -being
. scheduled so that crmcohty w1I| be reached obout_}?

THE REACTOR

_ The source geometry becomes on lmportont foctor_}_')'-'
" when a shoped reactor ‘shield is used, and there-
' fore -every attempt has been made to preserve the

geometry of the CFRMR in the SMC (see Fig. 5.4.1),

The central beryllium island will be cylindrical,
and its center will be removed to provide space for

L. B. Holland

the inlet cooling line. The CFRMR has a control
rod of slightly smcller diameter in this position,

" but calculations have shown that the effect of the

cooling line will not be significant. The SMC will
be controlled by using thin, curved plates con-
taining B19 at the outer edge of the island between
the fuel plates and the beryllium. In the core region

- it is proposed to use fixed fuel plates of UO, and

stainless steel, which will be similar in cross
section to the Army Package Power Reactor fuel

- plates, The shape of the plates will be peculiar
“to the SMC,
 already been determined with test fuel plates.
- There will be a total of 200 vertical fuel plates
* placed radially around the beryllium island, and
normal water will be used as the coolant. The
~"core region will be separated from the beryllium

The feasibility of the shape has

island and the reflector by Inconel shells, as re-

~ The region above the CFRMR reactor core will

_"be difficult to mock up. Since it constitutes a

reactor will be colled the Shield Mockup Core (SMC)

small total angle,- it seems best, in the SMC, to

“attempt to black out radiation from this region,
- Another feature in favor of this idea is that the
: control-rod drives and control-chomber leads will

be in this region,
‘The _reactor is being designed so that it can be
trensported between the TSF and the BSF with a

“ 'minimum amount of dismantling, The fuel can be
removed without disturbing the control rods of the

control chambers. This will be accomplished by

__placing the elements in a rotating rig that can be

locked “into position. A removable plug in the rig

o will o"ow removol of the elements.

. THE REFLECTOR

The berylllum reflector of the SMC will be similar
in- shcpe to that. of the CFRMR, except at the top

of. the’ berylllum where the question of hydraulics
_ greatly influences the CFRMR design. The sodium
~.-  coolant of the CFRMR reflector will be’ simulated
. -with aluminum in the beryllium. * If cooling is re-

" ~"quired in the berylhum of the SMC, normal water
fw:ll be used ‘and it wnll be kept at a minimum.

REGIONS 0UTS|DE THE REFLECTOR
The first region outside the berylhum reflector

"will be the first boron curtain, In the CFRMR this

region consists of a layer of sodium, followed by

279

 
 

ANP PROJECT PROGRESS REPORT

2-01-0%9-7T9

       

 

!

AT
AN
WY,
NS
)

»

 

 

 

 

 

             
  

 

 

 

 

 

 

 

 

 

 

 

 

 

TN Y
N
LA

 

 

 

 

~

N

R

-
N !
W

NN
W

D

 

 

 

 

 

SECTION A-A

P

722

P A e d

to01234567
SCALE IN INCHES

 

 

 

 

 

 

 

SUPPORT ASSEMBLY'—\

 

 

  
    

REMOVABLE PLUG—-

SECTION 8-8

101234567

SCALE IN INCHES

 

LEAD LIFTING HOOK

ad »

SALT REGION

8e ISLAND
INCONEL (CFRMR PRESSURE SHELL)

INCONEL (SMC PRESSURE SHELL)
INCONEL

ALKYL BENZENE

CONTAINER

      
  

 

  

 

SCALE IN FEET

 

    

(3

 

  

s Vot

   

S
L)
SR N

 

       

 

Fig. 5.4.1. The Shield Mockup Core.

280

 

 

L it i
 

 

 

.

an Inconel shell and a complicated boron-copper
layer. In considering the radiation as seen outside
the shield, this region can be adequately simulated
in the SMC with a shell of Inconel followed by
layers of boral to give the proper density in g/cm2,
Calculations are being corried out at Pratt &
Whitney Aircraft to determine the importance of
each region as a source of radiation.

The heat exchanger region just beyond the first
boron curtain affects the crew compartment dose
rate in several ways: (1) it attenuates the radia-
tion from the core and the beryllium; (2) it is a
source of some capture gamma radiation from core
neutrons; and (3) it is a source of delayed neutrons
and fission-product-decay gomma rays. In order
to account for the first two effects, the heat
exchanger region will be mocked up with fused
salts (in the form of NaF and KF) and NaK, as a
homogenized mixture. - This will be accomplished
by heating the salt, the NaK, and the can to about
1000°C in an evacuated furnace. The cans will
be placed in two layers to eliminate leakage paths
between the cans. The Inconel in the heat ex-
changer will be split to form two shells around the
salt region. The inner shell will act as the
pressure shell for the SMC. Outside the outer heat
exchanger shell will be the second sodium-cooled
boren curtain, which, like the first curtain, will
be mocked up with boral.

The regular CFRMR pressure shell will follow
the heat exchanger region. It has been proposed
that part of this shell be split off for use in mount-
ing the lead shielding. This section is to be re-
movable to aliow the lead shield to be chonged
without dismantling the reactor, -

The neutron-shielding material will be contamed '
in an aluminum tank. The optimized neutron shield
will be a sphere placed off-center with respect to
“the reactor. It is proposed to permit lateral motion
of the neutron. shield, while using water as shield
material, to check the present optimization. Later.
“the shield container is to be sealed in the optimized

position, and neutron- shielding - materlals other
than water can be used. ' ‘ .

_ In order to facilitate the measurements at the TSF'
the whole reactor -and shield system has been de-

signed so that it can be rofoted about fl1e verhcal
axis. N

'coM'PARisoNOF SMC AND CFRMR
An examination of some of the results qf the cal-
culations performed by Pratt & Whitney indicates

PERIOD ENDING JUNE 10, 1956

how closely the SMC radiation simulates  the
CFRMR, Thermal-neutron captures in the reflector -
and the power distribution within the core have
been considered, and the importance of each region
of the reactor as a gamma-ray source is being
investigated. -

Neutron Captures in Beryllium

Since approximately 20% of the dose rate in the
crew compartment is expected to originate from
neutron captures in the beryllium, this source is
to be accurately simulated. Figure 5,4.2 shows
the captures that can be expected in the SMC
beryllium with normal water as the coolant in the
core. The lower curve shows the absorptions in
the SMC beryllium when the space between the
fuel plates is completely filled with normal water
and the reactor is operated at room temperature,

Rl
7 2-01-059—778

 

S5x10

 

 

 

 

 

CONFIGURATION 160: SPACE BETWEEN FUEL PLATES
FILLED WITH 100 % HZ0 |
CONFIGURATION 167: SPACE BETWEEN FUEL PLATES
7 FILLED WITH 50 % H,0—50% Al
4x10 " I~ CONFIGURATION 168: SPACE BETWEEN FUEL PLATES ]|
FILLED WITH 25% H,0~75% Al
CONFIGURATION 168A: SPACE BETWEEN FUEL PLATES
FILLED WITH 12.5% H,0—87.5% Al

 

310"

/,\owusumnon 168A
-7 ' _\\
\

 

SN
- |so \\\\ |

a5 40 45 S0 55 60 65
SPACE POINTS IN REFLECTOR

NEUTRON CAPTURES IN REFLECTOR (fraction/cm®)

f

 

 

 

 

 

 

 

 

Fig. 5.4.2. ,Co-:'rlpélrisori of Neutron Captures in
Beryllium Reflector of SMC for Various Configu-
rations with Neutron Captures in Reflector of

CFRMR.

281

 
 

 

 

‘Percentuges of Wuter m the Cofe_ £

 

ANP PROJECT PROGRESS REPORT

Comparing this with the calculations for the ab-
sorptions in the CFRMR reflector (this includes
beryllium and sodium) operating at CFRMR tem-
perature it can be seen that the absorptions in the
SMC are low. By reducing the water volume in the
core of the SMC with aluminum spacers (and thereby
reducing the moderation in the core) the absorptions
of neutrons in the beryllium can be increased. With
75% aluminum and 25% water ‘in the spaces, the
tails of the absorption curves aré in good agree-
ment, The shape of the curve for the region near
the core-reflector interface does not match as well
as the rest of the curve, but it may be lowered by
adding the water necessary for cooling and ad-
justing the amount of aluminum simulating the
sodium. In this manner the curves can be matched
throughout.  Both curves are normalized to one
fission per cubic centimeter in the core region.

Power Distribution

Some effort has been made to match the power
distribution in the core of the CFRMR with that in
the SMC for various percentages of water in the
core (Fig. 5.4.3). Since a difference between the
CFRMR and the SMC with 75% aluminum and 25%
water was noted, some effort was made to bring
these into better agreement, A slightly different
core (100 fuel plates rather than 200) was divided

into five regions to determine the variation that

 

 

 

 

 

 

 

 

 

 

 

 

SESRE™
s6 2-01-059-768
’ I l F l ! | T
CONFIGURATION 60: SPACE BETWEEN FUEL PLATES

32 — FILLED WITH 400 % H,0.
- CONFIGURATION 167 : SPACE BETWEEN FUEL PLATES
2 28} — FILLED WITH 50% Ha0-50% Al
5 CONFIGURATION 68: SPACE BETWEEN FUEL PLATES
g 241 FILLED WITH 25% HoO=75% Al.
§ CONFIGURATION {68A: SPACE BETWEEN FUEL PLATES
Z FILLED WITH 12.5% Hp0—-87.5% Al /
> 20 2
g \ (F;ONFI(‘JURATlloN 16!) l //
2 16 RN 167 /
E DN /] /
2 42 s:"‘ e 7 //

"-__-_-
§ o6 \\ 7~ = .
/ T

& 68a | T~

 

 

 

 

 

 

 

 

 

 

 

 

 

 

[=]
H
O
—m
2
=
"’\\

o

 

12 4 16 48 20 22 24 26 28 30 32 34
SPACE POINTS IN CORE

Fig. 5 4.3. Compaflson of. quer Distribution in
Core - of CFRMR with That ‘in __,.SMC for Various

  

282

POWER DISTRIBUTION (NORMALIZED)

can be obtained in the SMC core power distribution
(Fig. 5.4.4) by adjusting the fuel in the five regions
from the center out in the proportions of 2:2:3:3:4
and 1:2:3:4:5. These calculations are not final,
but they mdlcate that the proper ‘mockup of ‘the
CFRMR radlchon can be obtained i in the SMC

 

BESREF

2-01-059-758
4.0 T T T T T T ‘ ;
100 -FUEL-PLATE CORE WITH SPACES
‘BETWEEN FUEL PLATES FILLED WITH
IZSZ H20 ;(57 nlu ]
CONFIGURATION 273: RATIO OF MASSES OF
_ U0z IN FIVE REGIONS OF CORE: 2:2:3:3.4;

 

 

 

 

 

 

 

 

 

 

n
o

3.0
o, TOTAL MASS, 10.5kg Y2a3s,
W CONFIGURATION 275 RATIO OF MASSES OF
UOp:1:2:3:4:5; TOTAL MASSES,
O 1.8 kg Y235,
'9@ .

 

 

 

 

 

P —r——: - -
S % C2 N 273l 275__4La
o st

CONFIGURATION 273

 

o

 

 

 

 

 

 

 

 

 

 

 

 

 

 

% 18 20 22 24 26 28 30 32 34 36 38
SPACE POINTS IN CORE

Fig. 5.4.4, Effect on SMC Power Distribution
of Varying the U0, Mass in Five Regions of the
Core.

Gamma-Ray Sources

Another calculation is being carried out at Pruh‘
& Whltney to determine the importance  of each
region as a gamma-ray source. The reactor is being
divided into shells and the gamma-ray intensity
from sources in each shell is being determined in
a line-of-sight attenuation calculation. This is

being done both for the CFRMR and the SMC as a

basis of comparison of the two reactors.

The remaining region of importance that requires -

some further work is the heat exchanger. As pre-

viously mentioned the source from the circulation _

of the fuel is not present. Previous LTSF data!
indicate that the gamma-ray source resulting from
circulation of the fuel contributes approxlmofely
30% of the gamma-ray dose rate outside the reactor
shield. It is expected that further anolyms of the

 

TH, Woodsum, ANP Quar, Prog. Rep. Marcb 10, 1956,
ORNL-2061, p 237.

B

o

¥

C.
 

 

C

B

e

w .

 

fission-product gamma-ray data réported by Zobel
and Love? will enable a calculation to be made,

2y, Zobel, T. A. Love, and R. W. Peelle, ANP Quar,
Prog. Rep. March 10, 1956, ORNL-2061, p 250.

 

 
 

PERIOD ENDING JUNE 10, 1956

by the Monte Carlo method, to determine the gamma-
ray dose rate resulting from the circulation of the
fuel at least as well as the rest of the radiation
from the SMC can be measured.