ORNL-2833.txt

o Moy {,7 1961 MASTER

ORNL-2833
UC-25 - Metals, Ceramics, and Materials

CORROSION ASSOCIATED WITH

HYDROFLUORINATION_II_\J_THE
OAK RIDGE NATIONAL LABORATORY

FLUORIDE VOLATILITY PROCESS .

A. E. Goldman
A. P. Litman

OAK RIDGE NATIONAL LABORATORY
operated by

UNION CARBIDE CORPORATION
for the
U.S. ATOMIC ENERGY COMMISSION

DISCLAIMER

This report was prepared as an account of work sponsored by an
agency of the United States Government. Neither the United States
Government nor any agency Thereof, nor any of their employees,
makes any warranty, express or implied, or assumes any legal
liability or responsibility for the accuracy, completeness, or
usefulness of any information, apparatus, product, or process
disclosed, or represents that its use would not infringe privately
owned rights. Reference herein to any specific commercial product,
process, or service by trade name, trademark, manufacturer, or
otherwise does not necessarily constitute or imply its endorsement,
recommendation, or favoring by the United States Government or any
agency thereof. The views and opinions of authors expressed herein
do not necessarily state or reflect those of the United States
Government or any agency thereof.
DISCLAIMER

Portions of this document may be illegible In
electronic image products. Images are produced
from the best available original document.
Printed in USA. Price §2. 00 . Available from the

Office of Technical Services
Department of Commerce
Washington 25, D.C.

LEGAL NOTICE

This report was prepared as an account of Government sponsored work. Neither the United States,

nor the Commission, nor any person acting on behalf of the Commission:

A. Makes any warranty or representation, expressed or implied, with respect to the accuracy,
completeness, or usefulness of the information contained in this report, or that the use of
any information, apparatus, method, or process disclosed in this report may not infringe
privately owned rights; or

B. Assumes any liabilities with respect to the use of, or for damages resulting from the use of
any information, apparatus, method, or process disclosed in this report.

As used in the above, ‘‘person acting on behalf of the Commission’' includes any employee or

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

or contractor of the Commission, or employee of such contractor prepares, disseminates, or
provides access to, any information pursuant to his employment or contract with the Commission,

or his employment with such contractor.

ORNL-2833 .

Contract No. W-7LO5-eng-26

METALLURGY DIVISION

CORROSION ASSOCIATED WITH HYDROFLUORINATION IN THE
OAK RIDGE NATIONAL LABORATORY FTIIORIDE VOLATILITY PROCESS

A, E. Goldman and A. P. Litman

DATE TISSUED

wov 15 1950

OAK RIDGE NATIONAL LABORATORY
Oak Ridge, Tennessee
‘ operated by
UNION CARBIDE CORPORATION
for the
U. S. ATOMIC ENERGY COMMISSION
; CONTENTS
E .
SUMMARY == == m e e s e e m e e e e e e e 1
I. Development Work at the Oak Ridge National Laboratory ------------ 5
A. Early Laboratory Work -----------a-- U SOy
B. Vessels Used in Bench-Scale Process Studies ---------cccweu-—- 7
1. Inconel Hot-Facility Hydrofluorinator Dissolver ---------- 7
a. Material Selection -----cccmcmc - 7
b. Operational History —-----cmcmcomm e 7
c. Reaction to Environment -------c-mcmmmmcmccmm e 9
d. Discussion of Results -=------m--cccmmm e 9
2. Inconel Hydrofluorinator Dissolver -—------ceceememmmceaua—- 11
a. Material Selection —-----ccmcmmm e e 11
b. Operational History ---=--c-remcemcrc e rr e e 11
c. Reaction to Environment — Salt-Transfer Line Failure - 1L
d. Discussion of Results ---------mcmecmcmme e 14
3. INOR-8 Hydrofluorinator DisSSOlVer -------eececmccmccme—me 16
| a. - Material SelectiOn =--=-emmmmmo oo 16
b. Corrosion of an INOR-8 Hydrogen Fluoride Entry Tube
and Thermocouple Well —----eeememmcmmmmccccmmmm e 16
C. Semiworks-Size Process Development Vessels -=--eeeemmocmeomaaa- 20
1. Mark I Copper-Lined Hydrofluorinator Dissolver ----------- 20
a. Material Selection --==--mommommm oo S 20
b. Operational History —=---cemmo oo 2l
_c. Reaction to Environment —---cecmemmom oo 2l
d. Discussion of Results ----cemmecmmmcmc e e 33.
2. Mark I INOR-8 Hydrofluorinator Dissolver -—-----eeeeeeeoaao 3k
a. Material Selection —-----mmm e e 34
b. Operational History ------ O 34
) c. Reaction to Environment --e=-----eecmmoocoao-- ——————— 3L
. d. Corrosion of Internal Components -------c-ecece-ee-aceaa- 45
: e. Corrosion of Test COUPONS =~~--=----—ccmeocomcccmcee—- 45
f. Discussion of Results — Conclusions --------———=-c-=e- 51
ITI. Screening Tests at Battelle Memorial Institute ---------c-eemmoeo- 52
A. Material Selection ----mm-mmcmcomo e —_—— 52
B. Experimental Procedure and Results -------c-mecmocmccmmmccauan- 52
C. Discussion of ResSultS —-e-mee o mm e e 64
TITI. Studies at the Argonne National Laboratory ---------ee---e--- ————— 66 |
A. Experimental Procedures and Results ------cmemommmmmmcmccmeaao 66
B. Discussion of Results — Conclusions ===-e--—ccecommmmcmommo 67
IV. Oak Ridge National Laboratory Volatility Pilot Plant
Hydrofluorinator DissOlver —--e - oo Th
CONCLUSIONS -----=--- LT T et 76
ACKNOWLEDGMENT == === = = mmm mm e e e e = 80

ii

" BIBLIOGRAPHY - == = == o m o o o o o e e e 81
CORROSION ASSOCIATED WITH HYDROFLUORINATION IN THE
OAK RIDGE NATIONAL LABORATORY FLUORIDE VOLATILITY PROCESS

A, E. Goldman and A. P. Litman
SUMMARY

This reporl summarizcs studies carried out at the 0Oak Ridge National
Laboratory (ORNL), Battelle Memorial Institute (BMI), and a portion of the
work done at Argonne National Laboratory (ANL) on corrosion associated with
the hydrofluorination-dissolution phase in the fused-salt Fluoride Volatility
Process. The Fluoride Volatility Process is being developed as a nonagueous
method for reprocessing spent heterogeneous or homogeneous nuclear fuels.

The application of this process to reactor fuel elements requires conversion

of the fuel to a fluoride form. This can be accomplished for zirconium-base
elements by bubbling hydrogen fluoride through a fused-fluoride salt bath to
dissolve such fuels. The uranium ic converted to uranium tetrafluoride and
the melt is transferred to & second vessel. A fluorine sparge further oxidizes
the UFM to UF6 which is volatilized, decontaminated, trapped, and returned for
conversion into fresh nuclear fuels. Corrosion associated with the fluorina-
tion phase of the Volatility Process has been reported.l

This document is divided into four sections. Section I deals with con-
ceptive studies of the Fluoride Volatility Process and with corrosion of
hydrofluorination-dissolver vessels used in bench-scale and semiworks-scale
process development by the Chemical Technology Division of the Oak Ridge
National Laboratory. Section II summarizes the results of a study on con-
struction materials for the dissolution phase of the Fluoride Volatility
Process carried out at BMI under ORNL Subcontract No. 988. For comparison
purpocese, Section TTT describes some of the corrosion studies on the Volatil-
ity Process performed at ANL. For reprocessing fuels high in zirconium

content, the ANL approach to reprocessing has been similar to the method at

lA. P. Litman and A. E. Goldman, Corrosion Associated with Fluorination

in the Oak Ridge National Laboratory Fluoride Volatility Process, ORNL-23832

(June 5, 1961).
.ORNL.. Section IV discusses and describes a full-size hydrofluorinator dis-

solver that has been installed in the Volatility Pilot Plant (VPP) at ORNL.

| In this'report, corrosive attack is reported as mils per month, based

on molten-salt residence time, or mils per hour, based on hydrogen fluoride

exposure time. These rates are included specifically for comparison purposes,

are th exact in most cases, and should not be extrapolated into longer time

periods for design work or other applications. 7

a Two Inconel hydrofluorinator dissolvers were used in bench-scale process

development studies at ORNL and subsequently examined for resistance to

corrosion. The first vessel, fabricated from a 13-in. length of 2-in.-diam

tubing, 0.065-in.-wall thickness, contained equimolar NaF-ZrFu to which

5 wt % U plus fission products had been added. The vessel was at 600°C for

a total time of about 180 hr. The fluoride salt bath was sparged' with

) 870 standard liters of hydrogen fluoride over a period of 145 hr. Corrosion

Jlosses on this vessel exceeded 30 mils/month in the salt, vapor-salt interface,

and middle vapor regions. A pitting attack seemed predominate in the salt.

and -interface regions, but was not appafent in samples from the vapor region.
Thé second Inconel bench-scale dissolver was made from an 18-in. length

of 3-in. sched-40 pipe. This vessel was exposed to fused-fluoride salts

which varied from 57 LiF-43 NaF mole % to approx 32 LiF—23 NaF-ui5 ZrF) mole %

'~ as Zircaloy-2 subassembly plate sections were dissolved. The salt exposure

time was 112 hr at temperatures of 550-750°C. Hydrogen fluoride was sparged
through the melt for about 100 hr.- At this point, failure occurred in -the
Ificonel salt-transfer line near the dissolution vessel, Examination indicated
that.the region of failure had been thinned previously by welding repairs.
Corrosion had further weakened the :egion_until it failed under the internal
salt pressure. Metallographic examination of the transfer line revealed a
porous corrosion product layer lining the pipe. The layer was highly ferro-
magnetic and had the characteristic appearance of Inconel from which chromium.
has been selectively leached.

| An additional bench-scale dissplver‘was fabficatéd with an Inconel top
section and an INOR-8 bottom section,:each about 0.220 in. thick. This vessel
was 3.5 in. in diameter and 18 in. high. To augment corrosion data, the INOR-8

internal hydrogen fluoride entry. tubes énd thermocouple_Wells were examined for
-3 -

corrosive attack. The tubes were subjected to salt compositions similar to
those used in the Inconel bench-scale dissolvers; fluoride salts were in
contact with the tubes for 395 hr at 500-700°C; and hydrogen fluoride was
sparged through the melts for 168 hr. On examination, the tubes showed a slight
weight gain, the result of an adherent metallic scale that had formed on the
outside diameter of the tubes. The scale contained the constituents of both
INOR-8 and Zircaloy-2. Dimensional analysis revealed no significant wall-
thickness changes excepl at the interface region of the entry tubes. A rate
loss of 3 mils/month on the bath contact side was found in this region.

Two semiworks-size hydrofluorinator dissolvers were used in larger scale
. honradive engineering process studies at ORNL. The first vessel was a cylin-
der 6 in. in diameter, 30 in. long, and 0.190 in. thick. It was fabricated
from deoxidized copper which, in turn, was supported by a type 347 stainless
steel jacket. Twenty-lour dissolution runs were carried out in this vessel.
The copper lincr wae exposed to salts of 62 NaF-38 Zth mole % composition
and also to salts varying in composition from 33~-LO NaF, H7—55 LiF, 520 Zth
mole %. The temperature ranged from 600-725°C and the salt residence time was
425 hr. Over a period of 265 hr, 30,000 liters of hydrogen [luoride wcre
sparged through the melts. Maximum wall-thickness losses of 45 to 69 mils/month,
based on molten-salt residence time, were found in the middle vapor region.
Considerably lower rates were found for the vapor-salt interface region and
negligible bulk-metal losses were found in the salt region. Metallographic
examination of sections removed from the vessel disclosed various surface and
subsurface layers. The surface layers were 1 to 6 mils in thickness and con-
tained fluoride salt residues, zirconium and tin from the subassemblies,
copper from the liner, and relatively large amounts of oxides. The subsurface
layers had a maximum thickness of 7 mils and an appearance closely resembling
that of internally oxidized copper. X-ray diffraction data indicated the
subgscales contained Cu20/Cu0 in a 7:1 ratio. Subscales of maximum thickness
were found in the region exhibiting the greatest corrosion losses.

The second semiworks-size hydrofluorinator was fabricated from 1/4-in,
INOR-8 plate rolled into two right cylinders, 10 and 6 in. in diameter. The
cylinders were joined by a truncated conical section to torm & 4O-in.-high

vessel. The dissolver was uced for nine Zircaloy-2 dissolution runs and one
-4 o-

run when the alloy was not présefit in the system. Total exposure time for

the vessel was approx 200 hr in molten 43 NaF-57 LiF mole % or

37 NaF-50 LiF-13 ZrF) mole % salts at 650-TLO°C. Hydrogen fluoride was
sparged through the melts for 80.5 hr. After the first four runs, thickness
measurements by ultrasonic techniques indicatéd an a&erage reduction in wall
thickness of 6 mils. No further losses were detected during the next five
runs. Ultrasonic examination after the final run, when Zircaloy-2 was not
present, indicated average metal losses of 18 mils at the vapor-salt interface.
Visual and dimensional examinations of the vessel disclosed a pitting attack
which increased in severity from the bottom of the vessel to the vapor-salt
interface region. Pit depths up to 72 mils were noted in the interface region.
Metallographic examinafiion revealed evidence of intefgranular attack which
resulied in the sloughing of whole grains of INOR-8 and the presence of a
porous surface 1ayef on the interior wall of the dissolver. Chemical analysis
revealed that the layer was deficient in chromium and iron when compared td
‘the base metal. Evidence of inte}granular attack was also found on those
metallographic sections that did not exhibit pitting attack.

Molybdenum and INOR-8 test coupons and INOR-8 internal components exposed
to0 the enviromment.of the INOR-8 dissolver were evaluated for corrosive attack.
“Corrosion losses and modes of attack similar to those noted for the vessel
“were found. However, fhe molybdenum test coupons showed less than one half the
weight losses of the INOR-8 speéimEns. Comparison of corrosion losses occurring
with and without dissolution of bulk zirconium indicated that the presence of
zirconium greatly inhibited attack.

A screening program for potential hydrofluorinator contaifier materials
was carried out at BMI. The materials studied were Inconel, INOR-1, INOR-8,

A nickel, copper, silver, Mbnel, Hastelloy B, and Hastelloy W. Corrosion
rates during simulated hydrofluorination-dissolution conditiohs were fofind to
be directly related to the following factors: (1) the alkali metal content
of the fluoride salts, (2) higher operating temperatures, and (3) increased
hydrogen fluoride flow rates. The rates seemed to be retarded by increases
in the zirconium content of all the fluoride salt systems and by higher over-

pressures of hydrogen in the sodium-zirconium systems. For all materials
S

Lal¥

-5 -

tested, the highest corrosion rates were noted at the vapor-salt interface
regions. Most of the corrosion rates reported were lower than for rates
determined on vessels used in dissolution runs at ORNL., When all factors
were considered, the most promising material studied was INOR-8.

Laboratory-scale corrosion tests have been carried out at ANL to select
construction materials for a dissolution process similar to the ORNL Fluoride
Volatility Process. The materials tested were L and A nickel, Inconel, Monel,
copper, Hastelloy B, molybdenum, silver, gold, platinum, tantalum, niobium,
and several grades of graphite. Graphite, molybdenum, silver, gold, and
platinum showed promise, but because of cost and fabrication considerations,
graphite was chosen for a pilot-scale hydrofluorinator dissolver. A vessel
was built with 1.5-in.-thick walls so that a temperature gradient developed
in the wall during operation. This allowed salt penetrating the graphite or
leaking through mating surfaces to solidify and be immobilized. The vessel
has handled a number of dissolution runs.

The construction material selected at ORNL for a full-scale Volatility
Pilot Plant Hydrofluorinalor was INOR-8. Graphite was rejected because of its
poor structural characteristics and porous nature. The latter wae thought to
present serious difficulties during decontamination and uranium recovery.- To
compensate for expected INOR-8 corrosion, the Volatility Process flowsheet
has been modified to (1) use a lower melting LiF-NeF-ZrF) salt bath, (2) retain
bulk-zirconium metal in the fluoride meltf whenever hydrogen fluoride is present,
and (3) avoid fixed salt-vapor interface levels.

A full-scale process hydrofluorihator, approx 17 ft in height, was con-
structed from 3/8- and 1/k-in.-thick plate rolled into right cylinders of
24- and 5.5-in. diam. The cylinders were joined by a l/2-in.-thick truncated
conical section. Reprocessing of naval reactof fuel subassemblies is expected

to begin the last half of 1961,

I. Development Work at the Oak Ridge National Laboratory

A. Early Laboratory Work

In 1954 scouting tests were performed by personnel of the Volatility

Studies Group, Chemical Developmecnt Section, Chemical Technology Difiision, to
-6 -

determine whether practical dissolution- rates forvmaterials used in nuclear
reactor fuel elements could be obtained in molten-fluoride salt baths by
sparging with hydrogen fluoride.2 The dissolution bath, composed of

ZxF) -KF-NaF (4L ,5-48.5-7.0 mole %), was held at 675°c while 0.050 liters/min
of hydrogen fluoride were bubbled through the melt. The test materials, each
having & surface area of 2-6 cm?, were exposed to the bath for 0.5 to 1 hr
and the.diséolution rates obtained by weight differences. Table I presents

a summary of the results.

Table I. Summary of Early Scouting Tests on Dissolution
of Materials in Fused Fluorides

Penetration Rate . .

Material | | (mils/hr)
Vanadium shot Not detected
Silicon powder : Not detected
. Nickel 0.001
K Monel 0.02
- Molybdenum 0.03
Tungsten , 0.06
Silicon Carbide* 2
Type 304 stainless steel L
Type 347 stainless steel 7
Niobium : T
Tantalum h 8
Manganese 10
Mild steel (Unistrut) : 13 A
Thorium, 1/8-in. plate 1k )
Uranium 17 _ ’
Zirconium** 22—-25 : .
Chromium ' | 31 -
. Titanium - : ' 31
Zircaloy-2%* , 7 22-l6
95 wt % Uranium-5 wt % Zirconium alloy 50 .
Tin - ‘ Sample melted and
dissolved instantly
Zinc E Sample melted and

dissolved instantly

¥

Disintegrated leaving suspended material.
*%
Range believed due to chemical and metallurgical differences

in individual specimens.

2R. E. Lueze and C. E. Schilling, Dissolution of Metals in Fused Fluoride
Baths, ORNL CF-54-7-59 (July 1954).

_.7_

The demonstrated practical dissolution rates for uranium and zirconium
encouraged more intensiQe work on the nonaqueous reprocessing scheme later
Lermed the Fluoride Volatility Process. These early studies also pointed up
certain elements and alloys, in particular vanadium, silicon, nickel, Monel,
molybdenum, and tungsten, which possiblyvcould serve as materials of construc-

tion for dissolution-reaction vessels.

B. Vessels Used in Bench-Scale Process Studies

1. Inconel Hot-Facility Hydrofluorinator Dissolver

a, Material Selection

Predicated on the early laboratory dissolution studies, as
well as availability and fabricability, nickel and nickel-base alloys lenta-
tively were selected as candidate construction materials for the first bench-
scale hydrofluorinator. Inconel, an oxidation-resistant nickel;rich alloy, was
the final choice because of the immediate availability of the necessary stock
sizes. The physical and mechanical properties of the alloy at the projected
operating temperatures and its resistance to molten-fluoride salts were attrac-
tive for the service anticipated.3 The vessel was made from a 2-in.-diam,
0.065-in.-wall Inconel tube and was about 14 in. in height. Figure 1 illustrates

‘a cross section of the dissolver.

b. Operational History

The vessel was exposed to molten-fluoride salts at approx 600°C
for a total of 185 hr and sparged with hydrogen fluoride for the last 145 hr.
The process conditions for the vessel, termed the Inconel Hot-Facility
Hydrofluorinator Dissolver,are given in Table II. After service, the vessel

was decontaminated using the schedule shown in Table III.

3W. D. Manly et al., "Metallurgical Problems in Molten Fluoride Systems,"
p 164 in Progress in Nuclear Energy, Series IV, Vol 2-Technology, Engineering
and Safety, Pergamon Press, London, 1960.

UNCLASSIFIED
ORNL-LR-DWG 55807R

,/-SAMPLE PORT

l

0.065-in.-WALL
INCONEL TUBE

uLer

Vi

o N N\=
I

N\

D

SLUG CHUTE

S

/ u CIIIIIIIIIIIIIEIIEIT > NS

\
\
N
N
N
N
N
N
N
N

THERMOCOUPLE

WELL

INCONEL BOTTOM

U176

Cross Section of Inconel Hot-Facility Hydrofluorinator Dissolver.

Fig. 1.

. gmetms P

A .,
_9_,

Table II. Process Conditions for Inconel
Hot-Facility Hydrofluorinator Dissolver

Salt Composition (mole %) 50 NaF-50 ZrF) + 5 wt % U
and fission products

Temperature (°C) Approx 600

Time of Exposure (hr) 185

Thermal Cycles (RT to 600°C) 70

145

Hr
HF Exposure {; . 870 (0.1 liters/min)

Table I1II. Decontamination Schedfile for Inconel
Hot-Facility Hydrofluorinator Dissolver

Solution

Temp. Time Corrosion Rate

Solution Muke-up (°c) (hr) (mils/hr max)
0.5 M (NH) ), C,0,-H0 90 10 6 x 107
0.5 M (NHu)E cgou-Hzo Boiling 41 6 x 1o‘LL
0.5 M (NH,), €,0,-E,0 20 . 765 6% 1077
10% NeOH-10% Na,C)H, O-H,0 20-40 285 6 x 1077
2% HéOQ-S% HN03-5% Al (NO3)3 Boiling 8.5 8 x 1o'LL

¢c. Reaction to Environment

Following decontamination, the vessel was sectioned and five
areas removed for metallographic study. Figure 2 illustrates the sections
taken and the results of the metallographic examination. Since the corrosion
rates attributable to the decontamination solutions were negligible, the

corrosion noted was assigned solely to the dissolution studies.

d. Discussion of Results

The Inconel Hot-Facility Hydrofluorinator Dissolver demonstrated
poor resistance to the Volatility Process dissolution environment. Both the

salt bath region and the salt-vapor interface region had corrosion rate losses
11 HF

DIMENSIONS ARE

IN INCHES

(L

L

A

5 HF

APPROXIMATE
SALT LEVEL~

2 HF
I HF

(R

N

|
[
!
|
|
1
|

L

Va

1%

i i S o

7

CROSS SECTION

9%,

Fig. 2.
Inconel Hot-Facility Hydrofluorinator Dissolver.

UNCLASSIFIED
ORNL-LR-DWG 58753

Estimated Maximum
_ . Operating Corrosion _ ’
.Sectlon Region Temperature Rates Metallographic Observations
(°C) (mils/month)®-
11 HF Vapor © 150 19 Smooth interior surfaces
9 HF Vapor 250 32 Smooth interior surfaces .
SHF  Vapor 500 19 Slightly roughened interior
2 HF Interface 600 32 Severe pitting ottack — maximum depth of pits
. 2 mils
1HF  Salt 600 38 iregular interior surfaces — appearance of pitting-type
attack and subsequent washing action
°Based on molten salt exposure.
¥
!_l
O
!

Regional Corrosion Losses and Metallographic Observations for the
B .

greater than 30 mils/month. In addition, a portion of the middle vapor region
also had losses of the same magnitude. Of particular interest are the metallo-
graphic sections removed from these three major loss regions which are
pictured in Fig. 3. The salt and salt-vapor interface samples had extremely
rough surfaces and showed evidence of pitting attack. However, the sample
removed from the middle vapor region had a smooth surface and no indication

of preferential corrosion was noted.

2. Inconel Hydrofluorinator Dissolver

a. Material Selection

A second dissolver vessel, slightly larger than the first, was
fabricated from the same nickel-base alloy, Inconel, for use in further process
studies. The body of this vessel was a 3-in. sched-40 pipe section, 18 in.

in height. Figure 4 illustrates a cross section of the dissolver as built.

b. Operational History

The larger dissolver was exposed to molten-fluoride salts for
approx 112 hr at temperatures of 550-750°C and sparged with hydrogen fluoride
for 100 hr of that time. The salts varied in composition from 57 LiF-43 NaF
to 32 LiF—23 NaF-45 ZrFu mole % as Zircaloy-2 subassembly plate sections were
dissolved.

The process conditions for the Inconel Hydrofluorinator Dissolver

are given in Table IV.

Table IV. Summary of Process Conditions for
Inconel Hydrofluorinator Dissolver

Initial Salt Composition (mole %) 57 LiF—43 NaF

Final Salt Composition (mole %) 32 LiF—23 NaF-45 ZrF),

Temperature (°C) 550—700

Time of Exposure (hr) 112

Thermal Cycles (RT to 550-700°C) 6
Hr 100

L {Liters 1540 (0.2-0.4 liters/min)
Number L

R Hr, total 1.5
'ransfers ¢

Temperature (°C) 700-750

s 38 =

UNCLASSIFIED
ORNL—LR—DWG 32070R2

Y-26545

| | INNER SURFACE,
: MIDDLE VAPOR AREA

INNER SURFACE
SALT-VAPOR INTERFACE

Y-26546

DIMENSIONS ARE IN INCHES

200
100

INNER SURFACE,
SALT AREA

Fig. 3. Chemical Development Section A Inconel Hydrofluorinator.
Etchant: glyceria regia. Reduced 24%. 250X. )
o T

UNCLASSIFIED
ORNL-LR-DWG 55808

EFFLUENT "
oh3 *INLET

WIAIIIWIA

THERMOWELL —tad
CHARGING CHUTE
3-in. SCHED. 40 N .
INCONEL PIPE
£
2
%-in. 0D, 0.035-in. WALL
INCONEL TUBE — o |
N
|
SOCKET WELD —a_
\ 0
Yg-in. SCHED. 40
INCONEL PIPE
SIEVE PLATE—N
e
AV ALZIV A
- SALT TRANSFER LINE
Y4-in. INCONEL
PLATE

l‘_' 34/2in. 4,_1

Fig. 4. Cross Section of Inconel Hydrotluorinator Dissolver.
w T =

c. Reaction to Environment — Salt-Transfer Line Failure

Following the exposure listed above, failure occurred in the
salt-transfer line at a point 1/8 in. above the tube-to-pipe weld shown in
Fig. 4. The failure permitted molten salt to drain out, filling the recess
between the vessel and the external furnace wall and reacting vigorously with
the vessel's external surfaces. Figure 5 illustrates the point of failure
and corroded exterior surfaces.

Visual and metallographic examination of the failure area
revealed that the cross section of the transfer line had been reduced prior
to service because of field repairs made on the adjacent socket weld (Fig. 5)
and that a porous corrosion product layer was visible on the interior of the

pipe close to the point of failure.

d. Discussion of Results

The corrosion product layer on the interior wall of the Inconel
salt-transfer line was 0.004 in. thick and had the characteristic appearance
of Inconel from which chromium has been selectively leached. The corrosion
product and the Inconel line in proximity to the failure were highly ferro-
magnetic. It has been reported previously that hydrogen fluoride, produced
from the contact of fluoride with moist air in a fluoride salt system, prefer-

b5

entially removes chromium from Inconel. No analysis of the product layer
was obtained because of the small amounts available.

The failure described apparently occurred when the previously
thinned weld repair area was weakened further by corrosion to the point where
it could not withstand the internal salt pressure of a regular salt transfer.
The failure allowed molten salt to flow outward, greatly enlarging the original
hole. No further study was carried out on the vessel. Presumably, similar

corrosive attack had occurred on the parts of the vessel in contact with salt

and hydrogen fluoride.

hL. R. Trotter and E. E. Hoffman, Progress Report on Volatility Pilot Plant

Corrosion Problems to April 21, 1957, ORNL-2495, pp 1416 (Sept. 30, 1958).

5A. P. Litman and A. E. Goldman, Corrosion Associated with Fluorination
in the Oak Ridge National Laboratory Fluoride Volatility Process, ORNL-20532,
pp 153186 (June 5, 1961).

UNCLASSIFIED
Y 26838

POINT OF FAILURE

o poip ol g e ]

0.1C IN/DIV.

UNCLASSIFIED
Y-27032

N

SPONGY LAYER

COPPER REGIA ETCH

15X COPPER REGIA ETCH

(NOTE SPONGY LAYER ON INTERICR OF TU3E CLOSE TO POINT OF FAILURE)

Fig. 5. ©Salt-Transfer Line Failure from Inconel Hydrofluorinator Dissolver.

UNCLASSIFIED
Y-27029

100X

-g'[_.
o 16 =

3. INOR-8 Hydrofluorinator Dissolver

a. Material Selection

A nickel-molybdenum-iron-chromium alloy, INOR-8, developed at
ORNL for use with fused-fluoride salts,3 was selected as the replacement
material for the Inconel vessel discussed above. The alloy possesses excellent
oxidation resistance, has good mechanical properties, and has had extensive
corrosion testing in fused-salt experiments.

Only the bottom half of the original dissolver was replaced
with INOR-8, since only the bottom portion is exposed to molten salts. A
1/4-in. plate was rolled and seam welded into a 3-1/2-in.-diam cylinder,
approx 10 in. long. The cylinder was welded to the upper half of the original
Tnconel dissolver and an INOR-8 bottom cover plate and salt outlet nozzle
were attached. Figure 6 illustrates the modified dissolver vessel.

b. Corrosion of an INOR-8 Hydrogen Fluoride Entry Tube and
Thermocouple Well

A hydrogen fluoride entry (sparge) tube and thermocouple well,
both fabricated from 1/4-in.-diam, 0.025-in.-wall thickness INOR-8 tubing, were
used in the above vessel during three successive hydrofluorination-dissolution
runs using a LiF-NaF salt and sections from a Zircaloy-2 dummy fuel element
subassembly.

A summary of service conditions for the entry tube and thermo-
couple well used in the INOR-8 Hydrofluorinator Dissolver is given in Table V.

Table V. Process Conditions for Bench-Scale
INOR-8 Hydrofluorinator Dissolver

Initial Salt Composition (mole %) 58 LiF—L42 NaF
Final Salt Composition (mole %) 31 LiF—Q: NaF—L45 szu
Temperature (°C) 500-700
Time of Exposure (hr) 395
Thermal Cycles (RT to 500-700°C) L

Hr 168

HF Exposure {Liters ~3000 (0.2-1 liters/min)

L

UNCLASSIFIED
ORNL-LR-DWG 55809R

EFFLUENT
GAS
[l 1)
JrT T T T
i JID (o
Tor . TR
oo 1 w1
Ly N W 5
Vil [[Z/Z4 1271 ¥ 4 A
1 5 N 4 )
N b N 4/ 47
\ |_| N 4 L/
N L7 L7
N O 4
/A
N /7
\ V-4
A\
a
PN I,
THERMOWELL ; S 47 ~=-CHARGING CHUTE
by /
N 4
N 4
"REREL B
s /] ~=—NOZZLE SHOWN
§ A ROTATED 90°
N Q
\ \
\ N
\ A
\ N
] / ;
/
. /
o /
/
/ A
% %
|/ L/
Y.-in. INOR-8 /
‘ .
ROLLED PLATE ; /
/ Y
/ %
: % /
.E / //
o /
= v
/
L/ /
SIEVE PLATE o 7 T4
/ /Bl
n W
/) ]
] |
= J¥] [ J==SALT TRANSFER LINE
Y.-in, INOR-8 = i B
/4'[". = — i '
PLATE D s )
‘ V -
~_ 4

Fig. 6. Cross Section of Bench-Scale INOR-8 Hydrofluorinator Dissolver.
i Y

After exposure, the tubing had a weight increase of approx 0.3 g.

Dimensional analyses on the tubes disclosed the changes noted in Table VI.

Table VI. Summary of Dimensional Analyses on INOR-8
Entry Tube and Thermocouple Well

Wall-Thickness

Change Diameter Change

Description Region (mils)* (mils)**

HF Entry Tube Vapor 0.0 6 «0.5 =15 to 0.3
Vapor-Salt Interface -1.5 2.1

Salt -0.5 +.9 o +5.2

Thermocouple Vapor ~ 0 -2.9 to +1.0
Tube Vapor-Salt Interface -0.5 +0.6

Salt ~ 0 +0.1 to +6,6

*
By micrometer measurement.

By microscopic examination.

Except for the interface region of the hydrogen fluoride entry tubes, no
significant wall-thickness changes were noted. However, definite increases
in diameter were apparent. Metallographic examination performed at BMI of
sections removed from the tubes disclosed thin scales on the surfaces which
had been in contact with the salt baths.

Figure 7 shows the interface and salt region samples from the
thermocouple well. Qualitative spectrographic analysis of the metallic-appearing
scale indicated that it was composed of the constituents of both INOR-8 and
Zircaloy-2. Some surface roughening was apparent in the samples but no evidence
of intergranular attack was noted.

Maximum corrosive attack occurred on the hydrogen fluoride
entry tube in the region of the vapor-salt interface. If the assumption is
made that all corrosion occurred during the hydrogen fluoride sparge, the
rate loss for this area would be 6.5 mils/month for the combined inside and
outside-diameter losses or 3 mils/month for the outside diameter in contact
with the hydrofluorination environment. These rates were unusually low and
lent impetus toward building a semiworks-size INOR-8 hydrofluorinator dissolver

discussed below.
w15

UNCLASSIFIED
PHOTO 54068

g—— DEPOSIT

VAPOR-SALT INTERFACE SAMPLE

a— DEPOSIT

SALT PHASE SAMPLE

Fig. T. Deposits and Microstructure of INOR-8 Thermocouple
Well used in the INOR-8 Inconel Hydrofluorinator Dissolver.
Etchant: Chromic-hydrochloric acid. 100X.
w B =

C. Semiworks-Size Process Development Vessels

In conjunction with the chemical development studies at ORNL, additional
phases of the Fluoride Volatility Process were explored by the Unit Operations
(UNOP) Section of the Chemical Technology Division. Using semiworks-size
equipment in a cold (nonradive) engineering system, the UNOP program included
studies on dissolution rates and the effects of process scale-up on heat re-
moval, fission product entrainment, hydrogen fluoride utilization, process
control, and the collection of corrosion data. Figure 8 is a schematic
process flowsheet and Fig. 9 is a photograph of the UNOP installation in which

the following hydrofluorinator-dissolver vessels were used.

1l. Mark I Copper-Lined Hydrofluorinator Dissolver

a. Material Selection

The ORNL fluoride salt-purification facility (Y-12 Plant)
originally had used nickel as a material of construction. Sulfur contamination
and subsequent embrittlement led to the substitution of copper-lined stainless
steel vessels. Since copper was known to have adequate resistance to hydrogen
fluoride at the temperatures being considered6’7 and was in use in the salt-
purification facility, it was selected as the liner for the initial semiworks
dissolver of the Unit Operation's Development Program.

The liner was fabricated from a 30-in. length of type K, de-
oxidized, high-residual phosphorus (DHP) 6-in.-diam copper water tubing
(0.186-0.192-in.-wall thickness) which was supported by a section of 6-in.
sched-40 type 347 stainless steel pipe. An 0.192-in. copper plate was welded
to the liner to serve as a bottom and the liner top was sealed to the type 347
stainless steel sleeve by silver brazing with ASTM B-Ag-7 alloy. Considerable
difficulty was experienced in achieving acceptable welds in the copper. Figure 10

is a cross section of the as-built vessel showing the disposition of the interior

=

W. R. Myers and W. B. DeLong, "Fluorine Corrosion, High Temperature
Attack on Metals by Fluorine and Hydrogen Fluoride,”" Chem. Eng. Progr._&&(S)
359 (1948). —

7"Hydrogen Fluoride," p 248 in Fluoride Chemistry, Vol 1, ed. by J. H. Simons,
Academic Press, New York, 1950,

UNCLASSIFIED
ORNL-LR-DWG 35261AR

HF COOLER
(COPPER)
B 75 BUILDING
> <— H,0 IN L v// T OFF GAS
SYSTEM
WET-TEST
METER
— H,0 OUT
L I . HF CONDENSER (COPPER)
(COOLEDBYDRY ICE IN
TRICHLORETHYLENE) OFF GAS
Ny IN =
| N, IN
FREEZE
] ] HF IN VALVE
€ ‘ ( INCONEL) .
FL' C | o
L ' :
ROTAMETER |
| TT.
ACID | I
3 RECEIVER | |
(MILD qmflfi |
« STEEL) | hMflm
i | |'I||||||' |
LAPP REMOTE | NNMN |
PUMP HEAD
MON
(MONEL) HYDROFLUORINATOR SALT PREPARATION TANK
DISSOLVER (DHP COPPER LINER
(DHP COPPER LINER,TYPE TYPE 347 S.S. JACKET)

347 5S.S. JACKET; OR INOR-8)

Fig. 8. Unit Operation's Hydrofluorination Process Flowsheet.
JUNICLASS IFIED
PHOTO 41749

— 1]

HF | I
Condenser! : ‘gJ

>,

4 Yy - P Control
“ Valve

P

ik
s i "

Dissolver.

~ * Preparation Tank

Fig. 9. Unit Operation's Installation of the Ccpper-Linsd Dissolver and Related Equipment.

_88—
INCONEL

FLANGES
=~ SALIT
TRANSFER

UNCLASSIFIED
ORNL-LR-DWG 55810R

COPPER
SEAL RING

A
N

AATERRTREET USRS UL O OO OO OSORSOOUUOURO

///////[/////f/////////////’////////////////I

IQZC%%ZJW%ZIQZIU%\

SILVER
RAZE JOINT:

DENSITY AND
LEVEL PROBE

B

;

DISSOLUTION

: ——PLATFORM\——

w B

HF FEED

ANTURRRRRRRRRRN
Ny,

LD 77 7 77 7 7 7

N
\
\
”
'
\

65/8 in.

o~
S
'//////’//I///////’/”////1’//’/////////

///////‘I//////////’///I/////’/////I//”I’

iy N\
N
N
‘é

SWAGELOCK
COPPER —

l
|
Il

TYPE 347
STAINLESS STEEL

B E——

3OV4H3LNI
40dVA-11VS
318VIYVA

\/\

S
4

7

m

m

¢

/

7227

SIS IS
RECEEKEEELELLEE

I T W— -

Cross Section of Mark ] Copper-Lined Hydrofluorinator Dissolver.

Il)\|||.c_ 9y —

Mg, 10,

- Bl

piping, while Fig. 11 is a photograph of the vessel after service illustrating

the general configuration.

b. Operational History

Twenty-four dissolution runs were carried out in this vessel.
The liner was exposed to molten salts for 425 hr and approx 30,000 liters of
hydrogen fluoride were sparged during 265 hr of this time. Pertinent data
for all these runs and exposure conditions for the Mark I UNOP Copper-Lined
Hydrofluorinator Dissolver are given in Table VII. The wide variation in
exposure conditions for the vessel is obvious.

Table VII. Summary of Process Conditions for
Copper-Lined Hydrofluorinator Dissolver

Temperature of Molten Salts (°C) 600-725
Time of Exposure to Molten-Fluoride Losx
Salts (Hr)
Salt Compositions (mole %) 62 NaF—38 ZrF, or
33-40 NaF—47—55 LiF—5-20 ZxF),
Thermal Cycles (RT to Molten-Salt 2L
Temperature )
HF Flow Exposure 30 x 103 liters in ~ 265 hr;
6.3 to 8.8 liters/min
Zirconium Dissolved/run (1b) 0.13 to 6.86 1b; average rate

of dissolution = ~ 1 1b/run

*
Vessel operated for 8 hr at temperature in an argon blanket
with the top off.

c. Reaction to Environment

(1) Visual Examination. After removal from service, the

vapor region of the vessel was examined and found to be covered with a reddish
copper-colored scale which had the appearance of CuQO. The color became
darker toward the lower vapor region. The salt-vapor interface area was covered
with a thin black scale which thickened toward the bottom of the vessel.

Varying amounts of surface roughening, indicative of
corrosive attack, were noted on the interior walls of the vessel. The rough-
ening appeared to increase in the lower regions and some pitting attack was

evident from the middle interface area down through the salt area.
- D

UNCLASSIFIED
PHOTO 48124

INCONEL SALT
TRANSFER LINE

31 Y% in.

Fig. 11. Mark I Copper-Lined Hydrofluorinator
Dissolver after 24 Dissolulion Runs.
= D6 =

(2) Dimensional Analyses. Wall-thickness determinations

were first attempted using an ultrasonic-thickness device. However, satis-
factory readings could be obtained only in the vapor region and a small portion
of the bath level region. Table VIIT shows the results of this examination.

Table VIII. Summary of Ultrasonic-Thickness Readings on the
UNOP Copper-Lined Hydrofluorinator Dissolver*

Wall-Thickness Readings (in.)**

Location Quadrant
(in. down
from top flange) Region E W N S
2 Vapor 0.164 0.159 0.162 0.178
6 Vapor - 0.158 0.166 -
8 Vapor - - - 0.167
10 Vapor 0.166 - - -
12 Vapor 0.165 0.162 0.167 -
16 Vapor - - - 0.162
22 Vapor-Salt 0.165 - 0.168 0.163
Interface

"Original wall thickness was 0.186-0.192 in.
Satisfactory readings could not be obtained on the
remaining interface or salt regions.
Maximum wall-thickness losses were found to be highest in the upper part of
the vapor region, west quadrant, in vertical alignment with the salt-transfer
line. The dissolver then was sectioned into quadrants and micrometer measure-
ments taken. Again, maximum losses were noted for the west quadrant. These
maximum bulk-metal losses, converted into rates for comparison with previous
data, are shown in Table IX. Negligible losses were found in the salt phase

of the vessel.

(3) Metallographic Examination. Photomicrographs of

selected specimens removed from the hydrofluorinator dissolver are shown in
Fig. 12. Metallographic examination of the specimens disclosed various sur-
face and subsurface layers which varied in thickness from 1 to 6 mils. An
especially predominant subsurface layer found in the middle vapor region had
the typical appearance of a finely dispersed phase in internally oxidized

copper. The subscale had a maximum thickness of 7 mils. Vapor-salt interface
= 7 =

Table IX. Maximum Wall-Thickness Losses from the West Quadrant of the UNOP
Mark I Copper-Lined Hydrofluorinator Dissolver

(By Micrometer Measurement)

Location Wall-Thickness Wall-Thickness
(in. down Losses Losses
from top flange) Region (mils/month)* (mils/hr)**
3.5 Vapor & 0.1k4
5o Vapor 5§ 0. 125
Te5 Vapor 69 0.15
10 Vapor 65.5 0.1kL
12 Vapor 50 0.11
15 Vapor L5 0.10
17 Vapor-Salt 31 0. 17
Interface
20 Vapor-Salt 14 0.03
Interface
21 Vapor-Salt 20 0.045
Interface
22 Vapor-Salt 14 0.03
Interface
2L Vapor-Salt 3.5 ~ 0.01
Interface
25 Salt negligible negligible
27 Salt negligible negligible

*
Based on molten-salt residence time.

KK
Based on hydrogen fluoride sparge time.
S 4
%
7

N— DHP COPPER
LINER

N
P

ARATARIRA RV IR AR AR IR ARER RN AR RN

(LI Tl L L

29in.

FLLRLL 7 IR L7 IR

“-SALT LEVEL

; RANGES;

,,,,,,,,,,,, LTS
\\A NAANNANNNNNARNNNNNN

TYPE 347 STAINLEy
STEEL JACKET

{6 in,
I

1111111111711 IILLLLIILLLL 111117 ////2

—e— 3 n, =
=
m
=
o
AT

ETCHANT: HPO,~HNO;~-CH3COOH (55-20-25)

Fig. 12.

acid. Reduced 62%.

INNER SURFACE
UPPER VAPOR AREA

'
55

INNER SURFACE
MIDDLE VAPOR AREA

INNER SURFACE
SALT AREA

Etchant:

INNER SURFACE
LOWER INTERFACE AREA

UNCLASSIFIED
ORNL-LR-DWG 56146

UNCLASSIFIED
¥-33%0)

e e T

INNER SURFACE
UPPER INTERFACE AREA

INNER SURFACE
MIDDLE INTERFACE AREA

é]@“Jg
of | | of

~
o
=
o o|

014 |
016
n

g
o

Typical Microstructures of Samples from the Copper Liner of
the Mark I Hydrofluorinator Dissolver.
200X.

Phosphoric-acetic-nitric
= 89 =

specimens showed duplex layers and at various distances below the surface the
layers appeared to be separated by alignment of voids. Of particular interest
were the particles in the subscale of the upper interface specimen in that they
seemed to be deposited in subgrain boundaries.

Various degrees of roughening were apparent in these
specimens. The lower portion of the interface and the salt region showed the
pitting attack previously mentioned.

Metallographic examination of the circumferential copper
cap-to-liner weld on the dissolver disclosed severe porosity and cracking, as
seen in Fig. 13. The fissures, in some instances, extended completely through
the corner weld. Figure 13 also indicates the extent and magnitude of the
pitting attack on the liner and bottom head.

Molten salts were found to have penetrated and filled the
space between the liner and the stainless steel jacket, attacking the stainless
steel. Figure 1h4 illustrates this attack. The corrosion of the stainless
steel seemed similar to the attack on Inconel vessels in contact with fused
fluorides and hydrogen fluoride at elevated temperatures reported upon earlier
in this document. This is not unexpected since the attack mechanisms on stain-
less steel in contact with the above reactants should be similar to that on

Inconel.

(4) Chemical Analyses. Spectrographic and wet chemistry

analyses and x-ray diffraction patterns were taken on the surface deposils, sub-
scales, and base metal removed from the copper-lined dissolver. Electron
diffraction studies also were attempted, but the results were inconclusive.
X-ray diffraction data disclosed that the subscales contained CuEO/CuO in a8 T3l
ratio. Table X summarizes the results of the two chemical analyses methods
which indicate that the surface deposits contained fluoride salt residues,
elements from the subassemblies and copper liner, and relatively large amounts

of oxygen.

8F. F. Blankenship, The Effect of Strong Oxidants on Corrosion of Nickel
Alloys by Fluoride Melts, ORNL TM-2 (Sept. 22, 1961).

- B0 =

UNCLASSIFIED
Y 34393

UNCLASSIFIED
Y 34395

347 STAINLESS
STEEL JACKET

'COPPER LINER

347 STAINLESS

STEEL JACKET COPPER LINER

WELD POROSITY

COPPER
BASE PLATE

347 STAINLESS STEEL BOTTOM 347 STAINLESS STEEL BOTTOM

= UNCLASSIFIED
' Y 34394
! 58

COPPER
LINER

H,0,- NH40H ETCH

COPPER BOTTOM

Py -

\,

Fig. 13. Typical Microstructures of Copper Liner-to-Bottom Head Weld
from the Mark I Hydrofluorinator Dissolver. Etchant: Hydrogen peroxide-
ammonium hydroxide. 5X.
- 3]

Unclassified

Y-37154

Fig. 14. Photomicrograph of Type 347 Stainless Steel Jacket from Copper-

Lined Hydrofluorinator Dissolver. (Note spongy surface layer and inter-
granular modifications on side unintentionally exposed to the hydrofluorination

environment.) As polished. 250X.
Table X.

Chemical Analyses Summary for UNOP's Mark T
Copper-Lined Hydrofluorinator Dissolver

Element wt %

Sample Description Cu 4 0 Si Mn Zr on Na Li F Ni

Vapor Region Surface 60.37 0.05 13.4b - == <2,0 1.01 7.02 2,27 10k =«
Deposit

Vapor-Salt Interface 32,05 0.06 2.55 == == 10.0 0.47 14.30 5.38 19.0 --

Region Surface Deposit

Salt Region Surface 4,39 0.13 2,53 == == 8.3 0.86 23.046 8.3k 38,0 -~
Deposit

Vapor Region Salt 98.95 0.07 0.06 0,05 <0,01 € 0.5< 0,15 - ope == 0,03

Subscale
Vapor-Salt Interface 97.15 0,06 0.20 .06 0.02 0.7 0,55 =~ - -- 0.29
Region Subscale
Salt Region Subscale 98.73 0.04 0,10 0.05<0.,01 0O.4% 1.4l - - -- 0,36
Base Metal 99,44 0,07 0.07 0.04 0.01 Wl o T | — - PR & i

-EE-
- B -

d. Discussion of Results

The subsurface layers present in all regions of the dissolver
had an appearance closely resembling that of internally oxidized copper. In-
ternal oxidation is generally regarded as a process wherein the most reactive
components in an alloy (in this case silicon and phosphorus) are oxidized below

9

the original surface of the material by inward diffusion of oxygen. Circum-
stances were such that internal oxidation could occur in the copper liner of
the UNOP dissolver, i.e., (1) The concentration levels of silicon and phosphorus
JO;11

(2) Oxygen

from air contamination was present at various times during the operation of

in the base metal were sufficiently high to produce subscales.

the dissolver, especially at the end of the dissolution study series. (3) There

is a suitable degree of oxygen solubility in copper (7.1 x 1073 wt % 0, at
600°c).(ref 12)

than copper oxides.

and (4) Silicon and phosphorus form oxides which are more stable

The most prominent subscales in the copper liner were found
in the region of maximum wall thinning. This may or may not be coincidental,
but the additional fact that the salt region of the vessel sustained negli-
gible corrosive losses and exhibited only a thin, spotty subscale indicates
that additional study on copper as a hydrofluorinator material of construction
may be profitable.

As evidenced by the severe weld metal cracking and porosity,
considerable difficulty was encountered in making the copper welds. Porosity
was noted during fabrication, but the welds were accepted "as is" for expedi-
ency. The weld cracking is believed to have been engendered by pressures
exerted during the thermal expansion of frozen salt during melting. The plane
of weakness resulting from the columnar structure of the weld metal was

weakened further by the presence of porosity.

9C. R. Cupp, "Gases in Metals," p 151 in Progress in Metal Physics Vol L,
ed. B. Chalmers, Interscience, New York, 1953.

lOF. N. Rhines, W. A. Johnson, and W. A. Anderson, Trans. Met. Soc. AIME,

ifil’ 205-221 (1942).

e, w Rhines, "Internal Oxidation," Corrosion and Material Protect. 4(2)
1521 (1947). =

12F. N. Rhines and C. H. Mathewson, Trans. Met. Soc. AIME 111, 337—353 (l93h).

= 3 »

After the process salts had penetrated the vessel liner,
capillary action forced the molten salt up the liner-jacket annulus to within
6 in. of the vessel top. The type 347 stainless steel jacket experienced cor-
rosive attack but did not fail even though the jacket-to-base welds were not

full-penetration welds (Fig. 13).

2. Mark I INOR-8 Hydrofluorinator Dissolver

a. Material Selection

The second semiworks-size dissolver used in the UNOP's program
was fabricated from INOR-8. This alloy was chosen because of superior perform-
ance demonstrated in previous ORNL bench-scale tests and concurrent BMI tests.
The difficulty encountered in obtaining sound copper welds in the original
semiworks dissolver was a further reason for utilizing INOR-8.

The vessel was composed of two right cylinders of l/h-in.—thick
INOR-8 plate, joined by a truncated conical section also fabricated from the
same material. The top cylinder was approximately a 10-in.-diam cylinder
21 in. in height, while the lower cylinder was 1l in. high and approx 6 in. in
diameter. Figure 15 illustrates a cross section of the dissolver vessel;
Figure 16 shows a photograph of the dissolver in operating position and Fig. 17

shows the internal components used with the vessel.

b. Operational History

The INOR-8 dissolver vessel was used for nine dissolution runs
on simulated fuel element subassemblies fabricated from Zircaloy-2. One
additional run was made without Zircaloy-2 present in the vessel. Molten
NeF-LiF or NaF-LiF-ZrF) salts were used at temperatures of 650-740°C for all
runs. Total exposure time of the vessel to the molten salts was approx 200 hr.
During this time, hydrogen fluoride was sparged at a rate of 17—100 liters/min
for 80.5 hr. Table XI presents a condensed operating log for the dissolver.

c. Reaction to Environment

The INOR-8 Hydrofluorinator Dissolver operated satisfactorily
with the exception of a salt-outlet line failure during run No. 8. The line
was attached to the bottom of the dissolver and failed at a point above a weld

connecting the ground electrode to the line. The electrode carried current
- 35 -

UNCLASSIFIED
ORNL-LR-DWG 55844R

OFF GAS

.
FLAT
e N
£
~
~AO in. DIA—w] | —CONICAL
L//fi\ e
Jh d 5

21/2 in, —s—

~ 6-in. DIA l

SLIESSONT

—————— | in, ——————

HF INLET J fl SALT OUTLET

Fig. 15. Cross Section of Mark I INOR-8
Iydrofluorinator Dissolver Illustrating Nozzles
and Placement of Interior Piping.
Merk

I INOR-8 Hydrofluorinator Dissolver in Operating Position.

Unclassified
ORNL Photo 45321

- e a—

”"\.’/

_9€-
HF DISTRIBUTOR
PLATES

Fig. 17.

DRAFT TUBE

Internal Components of the

T T UNCLASSIFIED
PHOTO 45622

LOWER BAFFLE PLATE

UPPER BAFFLE PLATE _

TOP FLANGE

AL LABORATORY

Mark I INOR-8 Hydrofluorinator Dissolver.

_LE -
Table XI. Condensed Operating Log for the Mark I INOR-8 Hydrofluorinator Dissolver

Exposure Conditions*

Time Temperature Initial Salt HF Flow HF Time
Operation (hr) (* ] Composition (mole %) (1b/hr) (hr)
Check-out 2 680 - i -
Check-out L.5 750 L3 NaF-57 LiF - -
Check-out 15 750 43 NaF-57 LiF k.o 8.0
Run 1 7 748-812 43 NaF—57 LiF 4.0 50
(792 av)
Run 2 10 700 43 NaF-57 LiF o) 8.0
Run 3 13 675—-700 43 NaF-57 LiF 2.0 11,0
Run 4 i 675~700 43 NaF—57 LiF - e
Cleaning Cycle 16 100 0.5 M NH)C,0) - s
Run 5 8 700 37 NaF-50 LiF-13 ZrF o o
Run 6 10 700 37 NaF—-50 LiF-13 ZrF 3.0 1.6
4 2,0 0.5
Run 7 19 > 650 43 NaF-57 LiF 5.5 14,0
LT 695715
Run 8 16 > 650 43 NaF-57 LiF 6.0 9.5
14 703722
Run 9 17 > 650 43 NaF—57 LiF 6.0 125
15 TOT-T22
Cleaning Cycle 68 85 0.5 M NH, C,0), - -
Run 10 (no Zircaloy-2 15 > 650 43 NaF-57 LiF 4.0 11,6
dissolution) 13 687—703 4.0 166
Cleaning Cycle 22 85 0.5 M NHMCEOM s .

*
Total exposures:

106 hr at 80-100°C in 0.5 M NH)C,0,

119.5 hr at > 650°C in NaF-LiF

18.0 hr at > 650°C in NaF-LiF-ZrFl

80.5 hr at > 650°C in NaF-LiF or NaF-LiF-ZrF) with hydrogen fluoride,

_82_
- 39 -

furnishing autoresistance heat for the line., The failure itself was found to
be a crack, intergranular in nature, which started on the inside of the pipe.
Excessively large grain sizes were found in the region of failure which suggested
that excessive local heating had occurred. The conclusion was reached that the
attendant high temperatures accelerated corrosion at the point of failure and
that, because of the geometry of the ground electrode, stresses may have been
present which precipitated the crack described.

Salt that leaked from the point of failure solidified and
formed a self-sealing repair,l3 after which run No. 8 was completed without diffi-
culty. The salt linec was replaced and the ground electrode was moved to the top
flange of the vessel to reduce stresses. After the two additional runs described
in Table XI were completed, a 20-in. bottom section of the hydrofluorinator was
removed and sent to BMT for corrosion analyses.lu Their results are summarized

below.

(1) Visual Examination. The appearance of the vapor-salt

interface and salt regions of the dissolver after the ten experimental runs is
shown in Figs. 18 and 19. The severe pitting attack apparent in the photographs
increased progressively from the bottom of the vessel to the interface region.
Pitting near the vessel bottom was negligible probably because the draft tube

prevented hydrogen fluoride impingement on the wall of the dissolver.

(2) Dimensional Analyses. Loss of bulk metal from the walls

of the hydrofluorinator dissolver was determined by ultrasonic-thickness
measurcmente at prescheduled times during operations. After the first four
runs, the average reduction in wall thickness was found to be 6 mils. However,
no significant losses could be detected after the next five runs, Nos. 5, 6,

7, 8, and 9. The last run, No. 10, when zirconium was not being dissolved in
the melt, produced a severe pitting attack and general wall-thickness losses.
Average metal losses of 18 mils at the vessel interface were detected by ultra-

sonic measurement after run No. 10.

13R. W. Horton, Chemical Technology Division, ORNL, private communication,

Aug. 25, 1959.

th. W. Fink, Corrosion of INOR-8 and Inconel Dissolver Components of the

Fluoride Volatility Process, AEC-U-4633 (Dec. 30, 1959).

Unclassified

BMI 61958

Fig. 18. Appearance of the INOR-8 Hydrofluorinator Dissolver Vapor-Salt
Interface Region after Ten Experimental Runs. (Most of the corrosive attack shown
occurred during run No. 10 during which time zirconium metal was not dissolving.)
N

' Unclassified
BMI 61959

Fig. 19. Appearance of the INOR-8 Hydrofluorinator Dissolver Salt Region
after Ten Experimental Runs. (Most of the pitting attack occurred during the
last run when bulk zirconium was not present.)
D) 3.

After the ten runs described, BMI personnel conducted a
survey of the pitting attack and bulk-metal losses on the Mark I INOR-8 vessel.
A summary of these data, shown in Table XIT,indicates that the most severe .
attack occurred at the vapor-salt interface region. Pit depths up to 72 mils
were noted in the interface region.

Table XII. Summary of Wall-Thickness Losses and Pitting Attack
on the Mark I INOR-8 Hydrofluorinator Dissolver

Wall-Thickness Reduction Pit Depth

(mils) (mils)
Micrometer and Depth

Ultrasonic Metallographic Mierometer

Region Measurements Measurements Measurements
Lower Vapor L—10 16-64* 317
Interface 1627 3272 2372
Salt =10 F27 I=15

*
After subtracting scale thickness varying from 8-16 mils.

(3) Metallographic Examination. Representative photomicrographs

of samples removed from the vapor-salt interface and salt regions of the dissolver
are shown in Figs. 20 and 21. Figure 20 illustrates a porous surface layer

found on "as polished" salt-interface specimens. This layer subsequently dis-
appeared upon etching. As shown, etching also uncovered evidence of inter-
granular composition changes in the INOR-8. This same figure shows a typical
cross section of a pit found in the interface region and indicates that

sloughing of entire metal grains occurred. Figure 21 shows that less corrosive
attack occurred in the salt region of the vessel, although some pitting attack,

particularly on the weld metal, was evident.

(4) Chemical Analyses. The surface layer shown in Fig. 21(a)

was analyzed at BMI by spectrographic and wet analysis methods. Table XTIIT
summarizes the information obtained on the major elements and indicates the
porous layer was severely depleted in chromium and moderately depleted in iron
content as the result of contact with the Volatility Process hydrofluorination

environment.
= 43 =

UNCLASSIFIED
BMI C-780

UNCLASSIFIED
BMI C-724

(a) o
100X UNETCHED

CROSS SECTION OF WALL SHOWING POROUS SURFACE
LAYER

UNCLASSIFIED
BMIC-734

(c)
70X ETCH H,CrO4-HCI
CROSS SECTION OF TYPICAL PIT SHOWING SLOUGH-
ING OF GRAINS AS THE RESULT OF INTERGRANULAR
ATTACK

(b) ¢ “hi b { R
100 X ETCH H,CrO4-HCI

CROSS SECTION OF WALL SHOWING SAME REGION AS
ABOVE IN ETCHED CONDITION AND ILLUSTRATING INTER-
GRANULAR ATTACK

Fig. 20. Microstructures from the Mark I INOR-8 Hydrofluorinator
Discolver Vapor-Salt Interface Region Showing Typical (a) Corrosion
Product, (b) Intergranular Attack, and (¢) Pit Cross Section.
(a)
100 X

. . UNCLASSIFIED
: BMIC-736

CROSS SECTION OF WALL FROM LOWER PORTION OF SALT

REGION IN PROXIMITY TO DRAFT TUBE

Fig. 21.

UNCLASSIFIED

g Lp , BMI C-735
; 'aj { : *
B R » -
A o ! g . L | ¥ ¥ / -
SRR i .
Y ; 8 ) £
v (.' 5 y ‘
=4 ’
} l{ a4 {g: *; '] . ¢
\ (\ ( d 3 2 ¢ 4
\ ) % T
v M3 #y
\lA '\__‘ .‘g i\ f 1 ‘I Y 4 ’\l{ <P
L A\ el A7 F Ao .
'\\d g i L
ol Chat by A 1 3 ' Y
\ s V) f R
W b )
?\ R L) \\ .
-t \ ~ \
. %‘w v EY. Lf\\ A Y /c ©
(X L) ('If? ¥ fi‘z,
\;'}\ “%3,:'-}\_;{'1"'
i * S~
‘F}Qtlk } TR ypot Foa
/ ¢ . \ "l y '\ X ' { \/
73 0 0, BN B s
& h ol "I .
’:f\ '/ = ‘;" ' "\"h .
P oy
> 4 .&,\%
’? () _& \ A @
A r& " )l :g /
¢ : ! 'S
' ;" y KOS
7 »./ 1 , ;
rd o )’ 3 { R s
1 R é ¢ i g ¢
T -
(b) § & ¢ K4
100 X ETCH H,CrO4—HCl

CROSS SECTION OF CIRCUMFERENTIAL WELD
SHOWING EVIDENCE OF PITTING ATTACK

Microstructures from the Mark I INOR-8 Hydrofluorinator Dissolver

Salt Region Showing Typical (a) Base Metal, and (b) Weld Metal Pitting Attack.

0.02

_-|7<|—(—
=~ U5 w

Table XIII. Summary of Chemical Analyses on the Porous Layer and
Base Metal from Interface Region of INOR-8 Hydrofluorinator Dissolver

Element wt %

Sample Ni Mo Cr Fe Analysis Method
Porous Major 1520 23 ol Spectrographic (BMI)
Layer

Porous Major 15.3 1.3 - Wet Chemistry (BMI)
Layer

Base Major 1415 710 5-10 Spectrographic (BMI)
Metal

Base 70 16.65 7.43  4.83 Wet Chemistry

Metal (INOR-8 supplier)

d. Corrosion of Internal Components

Corrosive attack on the INOR-8 draft tube, baffle assembly,
hydrogen fluoride distributor plates, and thermocouple wells used in the INOR-8
hydrofluorinator dissolver closely paralleled the attack found on the walls of
the vessel itself with respect to severity and types of attack. Details of Lhe
corrosion losses have been published.15 Figures 22, 23, and 24 show the draft
tube, baffle assembly, and a distributor plate after the ten process runs

previously described..

e. Corrosion of Test Coupons

A few INOR-8 and molybdenum coupons were exposed to the environ-
ments of runs Nos. 9 and 10 in the semiworks-scale dissolver. The coupons were
suspended from the conical baffle, as shown in Fig. 23, at the vapor-salt inter-
face. Removal of the coupons after run No. 9 revealed no noticeable attack,
based on weight change or visual observation. The coupons were reinserted for
run No. 10 which proceeded without the presence of zirconium metal in the
system. After run No. 10, it was found that the molybdenum coupon had lost
about 10% of its weight and the INOR-8 coupons had lost 22 and 29 wt %, respec-
tively. Both materials also suffered a severe pitting attack. Figure 25 shows
the visual appearance of the test specimens while Fig. 26 illustrates typical

microstructures of the coupons.

lSTbid., pp 16-33.
w A -

Unclassified
BMI N-60572

Fig. 22. Several Views of INOR-8 Draft Tube used in Mark I INOR-8 Hydro-

fluorinator Dissolver. White numbers indicate pit depth in mils. Approximately
one fourth actual size.
~ U =

Unclassified
ORNL Photo L6668

Fig. 23. Baffle Assembly used in Mark I INOR-8 Hydrofluorinator Dissolver.
Corrosion specimens are suspended from conical baffle. Approximately one third

actual size.
- 48 -

UNCLASSIFIED
PHOTO 46663

Q

GAS INLET SIDE

INCH

UNCLASSIFIED
PHOTO 46669

Q

GAS OUTLET SIDE

Corrosive Attack on Hydrogen Fluoride Distributor

Fig. 24.
Plate used in Mark I INOR-8 Hydrofluorinator Dissolver.

(a) Gas

inlet side, and (b) gas outlet side. Reduced 12.5% from actual size.
UNCLASSIFIED
PHOTO 46675A

INOR-8 SPECIMENS MOLYBDENUM SPECIMENS

AS RECEIVED AFTER EXPOSURE AS RECEIVED AFTER EXPOSURE

WEIGHT
LOSS (%) = 2e 29 o 10

Fig. 25. Vapor-Salt Interface Corrosion Specimens Expcsed during Runs No. 9 and 10
in the Mark I INOR-8 Dissolver. (Weight-loss data indicated all corrosion occurred
during run No. 10.)

_6.17_
(b)

Fig. 26,

- 50 -

UNCLASSIFIED
BMI| C-743

ETCHANT: MURAKAMI'S REAGENT

LINCLASSIFIED
BMIC-742

0.02

0.02

0.03

T
100X
|

Typical Microstructures of (a) Molybdenum, and (b)

INOR-8 Corrosion Specimens Exposed in the Mark I INOR-8 Hydro-

fluorinator Dissolver.

100X.
- 51 -

f. Discussion of Results — Conclusions

‘Based on the preceding informatidn, two important conclusions
were reached., First, the most severe corrosive attack during hydrofluorination
in INOR-8 vessels can be expected at the vapor-salt interface level and, second,
accelerated corrosive attack can be expected on INOR-8 when dissolution of bulk
zirconium is not taking place.

The original hypothesis by BMI was that cathodic protection
is the mechanism whereby dissolving zirconium metal retards hydrofluorination
process corrosion.l6 The galvanic couple was attributed to INOR-8 being
cathodic and zirconium anodic when they are both in the fluoride-salt bath.
However, other studies by the same group, reported in Section III of this
report, do not substantiate the claim that galvanic protection was responsible.

It was later suggested’by BMI that other factors, such as the
formation of fluoride-salt complexes and hydrogen (from the dissolution reaction
of zirconium with hydrogen fluoride), might be responsible for the corrosion
retardation described.

As noted in a subsequent section (II), the attack of INOR-8 by
hydrogen fluoride was found to be less in NaF-Zth mixtures than in LiF-NaF
mixtures. Thus, there is evidence that the presence of Zth in alkali-metal
fluoride melts also will retard corrosion by hydrogen fluoride.” It is inter-
esting in this respect that work at ORNL has shown that corrosion resulting
from UF, contained in alkali-fluoride melts likewise decreases as a function
of Zth content. L7 A second effect resulting from the presence of zirconium
metal in the hydrofluorination system is undoubtedly associated with the
buildup of hydrogen resulting from the zirconium-hydrogen fiuoride reaction.
The concentration of hydrogen produced by this extremely reactive system should
serve to inhibit hydrogen fluoride reactions with less reactive metals,

e.g., iron, chromium, molybdenum, and nickel, since these attain equilibrium

at much lower hydrogen fluoride-to-hydrogen ratios than zirconium.

l6?aul D. Miller et al., Construction Materials for Hydrofluorinator of

the Fluoride-Volatility Process, BMI-1348 (June 3, 1959).

17

l8J H. DeVan, Effect of Alloying Additions on Corrosion Behavior of Nickel-
Molybdenum Alloys in Fused Fluoride Mixtures, (unpublished M.S. Thesis, University
of Tennessee, 1960). |

J. H. DeVan, private communication, April 13, 1961..

- 52 -

Because of the'naturé of the corrosive attack and the un-
usual combination of operating conditions, it is.difficult to assign corrosion
~rate losses to the INOR-8 vessel. However, if ruh No. 10 is ignored, based only
on ultrasonically méasured-thickness changes, the_dissolver sustained approxi-
mate rate losses of 0.1 mils/hr, based on hydrbgen fluoride exposure time, or

23 mils/month, based on molten-salt residence time.

II. Screening Tests at Battelle Memorial Institute

A. Material Selection

A screening program for container materials capable of withstanding'.
the corrosive attack of volatility process hydrofluorination conditions was
done at BMI. This prografi was carried out under ORNL Subcontract No. 988.(ref 16)

The candidate materials, Inconel, A nickel, copper, silver, Monel,
‘IHastelle B, Hastelloy W, INOR-1, and INOR-8, were selected on the basis of a
literature search and previous ORNL corrosion studies. Table XIV gives the

nominal analyses of the alloys tested.

Table XIV. Composition of Alloys used in BMI Screening Tests

.Nominal Composition, wt %

Alloy Ni Cr Fe Mo Co Cu Mn Si C S \'
INOR-8 70° 7 5 16_ 0.2 0.35 0.8 0.3 0.06 0.01
INOR-1 7 - 0.3 200 -- -- 0.5 0.5 0.01L 0.01
Hastelloy B 62 1 6 28 - 1 1 0.12
Hastelloy W 60 5 6 25 2.5 - 0.12 0.6
Inconel Tex 16 8 - 0.5 1 0.5 0.15 0.015

Monel 67¥ - 1.5 ==  -= 30 1 0.1 0.15 0.01

*
Includes cobalt.

B. Experimental Procedure and Results

The test materials were evaluated under conditions which closely
simulated those that might exist in an actual hydroflfiorinator, i.é.,,co-
existing hydrogen fluoride and molten-fluoride salts at elevated temperature.

Figure 27 is a flow diagram of the apparatus used. The specimen container
MOLECULAR
SIEVE DRIER

NITROGEN

THERMAL FLOW METER
THERMOSTATED AT 82°C

ANHYDROUS HF '
THERMOSTATED AT 38°C

COPPER TURNINGS
(25°C) -
ZrF, (SNOW) TRAP

UNCLASSIFIED

ORNL —LR—DWG 55812

MINERAL OIL
BUBBLE BOTTLE

7

NICKELTURNINGS(GOO°€L//
FOR SULFUR REMOVAL

U

EXPOSURE VESSEL
(600° - 700°C)

Fig. 27. Flow Diagram of BMI Corrosion Test Assembly.

"OFF GAS

HF - H,0
AZEOTROPE
TRAF

- -KOH ~
SOLUTION

MINERAL OIL
BUBBLE BOTTLES

_gg-
- 54 -

and auxiliary equipment were fabricated of Inconel. Figure 28 illustrates a
representative cross section of théAtest container and the method used for
holding the specimens in place. The specimens were 2-in. lengths cut from
welded tubing made from the méterial to be tested. The specimens were fixed
over the sparge tube end so that the hydrogen fluoride swept the inner surface
of. the tube. ,

The initial studies were made on Inconel, INOR-1l, INOR-8, Hastelloy W,
A nickel, and Monel specimens. Exposure conditions and results for these
initial runs are shown in Table XV. These runs resulted in serious leaching

Table XV. Summary of BMI Corrosion Tests in Equimolar NaF-ZrF
and Hydrogen Fluoride at 650°C using Inconel Contdiners*

n - o . S e U Y Y S

Test Periods.24—1000 hr
: HF Rate 10 g/hr
B ‘ Alternate HF and Salt Phase Contact on Impingement Tubes
‘ Corrosion Rate

By Weight By Metallographic By Metallographic

Loss Examination Examination
Material ~ (mils/month)*¥ (mils/month)** (mils/hr)*** Remarks
Inconel ' - 14-39 85—-120 0.1-0.17 Selective leaching
Hastelloy W weight gain 3.6 0.005 Some intergranular
' - L _ attack
. A Nickel 8.9 6.8 0.01 Intergranular attack
: _ Monel 5.2 4.6-9.2 0.006-0.012 Selective attack
= - INOR-1 ‘ - none - Crystal deposition
. " INOR-8 0.09-1.5 0.6-1.8 0.001-0.003 Some intergranular
attack and crystal
deposition

r—
‘Based on data from BMI-13.48.
Based on molten-salt residence time.

*Nke :
Based on hydrogen fluoride sparge time.

of chromium along the outer layers of the Inconel coubon specimens and interior
piping. A typical analysis of exposed Inconel surfaces indicated 7.6 wt % Cr
remaining compared with an initial chromium content of 16 wt %. Also, metal
crystals were found deposited on Inconel coupons, interior piping, and on the

. walls of the Inconel containers. These crystals were primarily nickel and

contained\less than 0.1 wt % Cr. Figure 29 shows a section from an Inconel
- 55 -

UNCLASSIFIED
ORNL-LR-DWG 55813R

HF INLET

!

HF OUTLET

THERMOWELL

4-in.-SCHED 40
INCONEL PIPE ——

SALT LEVEL
(~ 8 in. FROM BOTTOM)——

/

INCONEL TUBE
'/4 -in. OD x 0.035-in. THICK WALL—

|

IMPINGEMENT SPECIMEN —

|

AN AR AR TR AR AR RN

\\\\\\\\\\‘\\\\\\\\\\\\\\\\\\\\L\\\\\\\\\\\\\\\ NANNANNANCNEN
14 in

!/, —in. INCONEL PLATE —

il

4‘/2 in.

Fig. 28. 1Inconel Container for Initial BMI Screening Tests.
o T SN A R,
Sy UNCLASSIFIED;
NTRE  BMI C-272

INCHES

NICKEL
CRYSTALS

0.05

UNATTACKED
METAL

SPONGY
ATTACKED AREA

Fig. 29. PzotomZcrograph of Iaconel Sperge Tube which Failed during Early BMI Tests.

Etchant: Nitrie-hydrochloric acid. 100X.

_9g_
- 57 -

sparge tube indicating the spongy areas where leaching of chromium occurred
and where the metal crystals were deposited. Similar results were found

during operations of a semiworks-scale vessel at ORNL when an Inconel salt line
and hydrogen fluoride carrier line failed in service (see Section I of this
report).

The A nickel specimens used in the early studies exhibited pronounced
intergranular attack. This attack was similar in appearance to the inter-
granular attack on A nickel subjected to the Volatility Process fluorination
environment.l9

Although the Monel coupon specimens tested in Inconel containers
showed only moderate rates of corrosive attack, a sparge tube, fabricated from
Monel, was severed after about 1000 hr of exposure to the hydrofluorination
environment. The penetration rate based on a 1000-hr exposure for this tube
was 25 mils/month.

Hastelloy W and the two INOR alloys, 1 and 8, demonstrated the best
resistance to corrosive attack during these early BMI studies. While some
evidence of intergranular attack was found on these nickel-rich molybdenum
alloys, their over-all resistance presented a strong case for continued screening
studies. Accordingly, these alloys (Hastelloy B was used in place of Hastelloy W)
were subjected to additional testing in detail with emphasis on the effect of
the following variables: (1) fluoride salt bath composition; (2) temperature;
(3) hydrogen fluoride flow rate; and (4) hydrogen content of environment.

The beginning salt compositions were divided into two major categories,
N&F-ZrFu and NaF-LiF. 1In addition, the effect of a final salt composition con-
taining 0.1 to 0.2 mole % UFH was studied. Temperatures were varied from 600
to 700°C and hydrogen fluoride flow rates of 0.2 to 0.5 liters/min were used.

The later BMI corrosion tests were carried out in containers fabricated
from Hastelloy B. Figure 30 shows a cross section of a Hastelloy B container
and illustrates the method of mounting test specimens. Figure 31 is a
photograph of an empty container, the internal piping, and the top closure.
Ixposure conditions and results for these later runs are shown in Tables XVI

and XVIT.

l9A. P. Litman and A. E. Goldman, Corrosion Associated with Fluorination
in the Oak Ridge National Laboratory Fluoride Volatility Process, Sections I
and IT, ORNL-2832 (June 5, 1961).

PRESSURE TRANSMITTER
AND EMERGENCY EXHAUST\m

- 58 -

UNCLASSIFIED
ORNL-LR-DWG 55814R

HF INLET

ZrF, (SNOW) TRAP

0

THERMOCOUPLE WELL

@B
162Y6% £ g Cy8'a LR 5(6 D6l
%% COPPER 3457
2526 TURNINGS (o408

$9c°6%0 g €0 939 atia'dy ]

iy

WL e
OFF GAS

Seey——fccceoeos M 0 s =

NI

COPPER COOLING COILS

ALK

(Ll

N
\
s 4-in. SCHED 10 HASTELLOY B
\ PIPE
& \
= \
Q BAFFLE PLATE Q
\
N
N
\
N
\
N
\
N
__— TEST SPECIMEN
SA SUPPORTS
3/8-in. - - .'-',
HASTELLOY B ///
PLATE — // IMPINGEMENT SPECIMEN
! 7 /
7//// ////////
/FURNACE /
Y 7
Fig. 30. Hastelloy B Container used in later BMI Screening Tests.
UNCLASSIFIED
N48626

HF - H,0

" AZEOTROPE
TRAP

Fig. 31. Hastelloy B Container and Closure Showing Gas
Impingement Specimen used in BMI Tests.

Table XVI.
and Hydrogen Fluoride using Hastelloy B Containers

Summary of BMI Corrosion Tests in NaF-ZrF, Salts

Corrosion Rates (mils/month)
By Weight By Micrometer By Metallographic
Loss Measurement Examination

Specimen Exposure
Description Position

Specimen
Material

Remarks

Group A — NaF (50-59 mole %)/ZrFLL (41-50 mole %) 650-700°C
Test Periods — 200-510 hr
Hydrogen Fluoride Rate — 0.2-0.5 liters/min

Hastelloy B coupons interface  0.8-14(6) 6.2-8.7(2) 12(1)
INOR-1 coupons vapor 0.1-0.76(5) 0.0 -
coupons interface 1.3-21(8) 6.7-31(5) 2.2-15(2)
and tubes
coupons salt 1.9—17(5) - -
impingement alternate 0.003-3.4(8) -- 0.86-5.2(3)
tube vapor-salt
contact
INOR-8 coupons vapor 0.25-1.5(7) 0.0(1) 1.4-3.9(2)
coupons interface 0,18-29(8)  14-L5(L4) 0.86-23(2)
and tubes
coupons salt 0.35-17(5) - _—
impingement alternate 0.09-5.7(7) - 1.7-2.9(3)
tube vapor-salt

contact

Necked down or severed
at interface — etched

Light-to-moderate etch

Necked down or severed
at interface — etched

Moderate etch, some
crystal deposits

Light-to-moderate etch

Light etch
Light-to-heavy etch —

necked down at interface

Moderate-to-heavy etch
Light-to-heavy etch

- 09_
Table XVI. (continued)

Specimen Specimen
Material Description

Ekposure

Position

Corrosion Rates (mils/month)

By Weight By Micrometer
Loss Measurement

By Metallographic
Examination

Remarks

Hastelloy B  coupons
INOR~1 coupons

coupons
and tubes

coupons

impingement
tube

INOR-8 coupons
coupons
coupons

impingement
tube

interface

vapor
interface
salt

alternate
vapor-salt
contact

vapor
interface
salt

alternate
vapor-salt
contact

0.26—2.6(2) 5.1(1)
0.08-0.3k4(2)

0.32-3.6(3) 4.0(1)
0.61-5,0(2)
0.63(1)

0.16-0.44(2)
0.39-4.5(2) h,0(1)

0.84-3,7(2)

0.01(1)

Group B — NaF-ZrF) -UF) (4b—56-0.1 mole %) 650°C
Test Periods — 200 hr
Hydrogen Fluoride Rate — 0.2 liters/min

2.9(1)

Necked down

Silver deposits —

.dark film

"Tarnish — dark

powder

Dark film - light
etch

Crystal deposits

( ) Numbers in parentheses refer to test specimens from which data were taken.

—'[9—
Table XVITI. Summary of BMI Corrosion Tests in NaF-LiF or NaF-LiF—ZrFu

Salts end Hydrogen Fluoride using Hastelloy B Containers

Corrosion Rates (mils/month)

Specimen Specimen Exposure By Weight By Micrometer By Metallographic
Material Description  Position Loss - Measurement Examination Remarks
“Group C — NaF (43-47 mole %)/LiF (53-57 mole %) T00°C |
' : Test Periods — 93-263 hr | :
Hydrogen Fluoride Rate — 0.2 liters/min
Hastelloy B  coupons interface -- -- -- Severed
INOR-1 coupons - vapor 0.02-13(3) -- - Etched and mottled
A ‘ finish '
coupons interface n -- 120(1) Two specimens
severed !
“coupons salt 17-38(2) .- -- Moderate-to-heavy oA
' ' etch, some crystal \
_ deposits
impingement alternate 2.9-18(2) -- -- Moderate etch, some
tube vapor-salt _ crystal deposits
contact o
INOR-8 coupons vapor 0.11-7.6(3) - - Light-to-heavy etch
coupons interface L48-60(2) - 100(1) One specimen severed
and tube . N . _ .
coupons movable  0.15-0.73(2) 0.68-31(5) ..  2.7-31(3) All samples necked
and tube¥* interface¥** a ‘ : : down ~
coupons salt 20-43(3) -- 28(1) Heavy etch
impingement - ~alternate - 9-18(2) | - C e o ‘Heavy etch, crystal
tube vapor-salt deposits

- contact

Table XVII. (continued)

Corrosion Retes (mils/month)

Specimen Specimen Exposure By Weight By Micrometer By Metallographic
Material Description Position Loss Measurement Examination Remarks
Group D — NaF—LiF-ZrFu-UFlL (19-26-55-0.2 mole %) 600~700°C
- Test Periods — 160-200 hr
Hydrogen Fluoride Rate — 0.2 liters/min
Hastelloy B  coupons interface 1.7(1) - -- - 600°C, Light etch
interface 2.1(1) -- -- - 650°C, Light etch
interface  3.9(1) 15(1) -- 700°C, Necked down
INOR-1 coupons vapor 0.24(1) -- - 600°C, Light etch
vapor 0.55(1) -- -- 650°C, Light etch
vapor 1.3(1) -- - 700°C, Light etch
TINOR-1 coupons interface 0.76(1) - - 600°C, Light etch
interface  3.3(1) -- - 650°C, Moderate
. etch : ‘
) interface. 13(1) 22(1) -- 700°C, Necked dowr
INOR-1 coupons salt 1.7(1) - - 600°C, Light etch
salt b, 4(1) -- - 650°C, Moderate etch
salt 11(1) -- -- 700°C, Heavy etch
INOR-8 coupons vapor 0.43(1) -— -- 600°C, Light etch
vapor 1.1(1) -- -- 650°C, Light etch
vapor 2.4 - - 700°C, Light etch
INOR-8 coupons interface = 1.0(1) -- - 600°C, Light etch
interface 4.6(1) -- -- 650°C, Moderate etch
interface - 13(1) 2c(1) - 700°C, Heavy etch
INOR-8 couporns salt 2.5(1) - - 600°C, Light etch
salt Lh.2(1) -- -- 650°C, Moderate etch
salt 15(1) - -- 700°C, Heavy etch

) :
This run was conducted in the presence of dissolving Zircaloy-2.

*¥
For constant interface position. _
NOTE: Numbers in parentheses refer to test specimens from which data were taken,

_€9_
_ 6l -

The daté resulting from the test series described indicated that INOR-8,
INOR-1,  and Hastelioy B showed approximately the same degree of resistance to
various simulated Fluoride Volatility hydrofluorination environments. In all
cases, the highest corfosion rates were experienced at the vapor-salt interface,
At the BMI test temperatures, the sodium-lithium systems seemed more corrosive
than the sodium-zirconium systems and the general conclusion was reached that
the corrosiveness of a'given salt was directly related to its alkali-fluoride
content, or conversely, inversely proportional to the zirconium-fluoride content.
Higher temperatures increased the corrosion rates significantly as did increases
in the hydrogen fluoride flow rate. Figure 32 shows a plot of corrosion rates
on INOR-8 at three temperatures indicating the strong temperature effect. For -
the sodium-zirconium system, the presence of hydrogen seemed to slow down the

rate of material attack but the effects of hydrogen in sodium-lithium systems

were not fully determined.

»

C. Discussion of Results

The BMI screening tests described added considerable insight .into the

corrosion aspects of the hydrofluorination phase of the Fluoride Volatility

. Process. However, most of the corrosion rates reported were lower fihan rates
A?found on vessels used in dissolution runs at ORNL. The observation that the
s.corrosion rate of the materials studied was proportional to the alkali-salt
vcontent'paféllels results that were obtained at ORNL during bench—écale cor-

rosion studies on the fluorination phase of the Volatility Process.eo Also,

their finding that the presence of zirconium fluoride and, to an even greater

degree, the presence of bulk-zirconium metal serves to retard the corrosion

rate and further strengthens the discussion of zirconium effects presented in
Section I of this report.

It was noted by BMI that maximum corrosive attack occurred at the
vapor-salt interface levels. They suggested that actual localized attack at the

liquid line is electrochemical in nature, arising from potential differences

20Ibid., Section III.
- 65 -

UNCLASSIFIED
ORNL-LR-DWG 39517

16
| | 10 g/hr HF
14 ——— 19.3 NaF, 25.6 LiF, 54.9 ZrF,, 0.2 UF,
(mole %)

10

= —

¢
4 | /

CORROSION RATE (mils/month)

600 650 700
TEMPERATURE (°C)

Fig. 32, Vapor-Salt Interface Corrosion Rates on INOR-8 at Three Tempera-
tures. Ref. Paul D. Miller et al., Construction Materials for Hydrofluorinator
of the Fluoride-Volatility Process, BMI-1348 (June 3, 1959).

- 66 -

between that portion of the specimen submerged and the portion at the interface
line. " This pofential difference is easily influenced by the relative solubility
of hydrogen fluoride in the salt and the partial pressure of hydrogen fluoride
ifi the vapor. The increased corrosion rate found with increased hydrogen
-fluoride flow rates is cited to support this theory.

Studies on fluorination corrosion at ORNL(ref 21)

precipitated the
hypothesis that corrosion of metal reaction vessels essentially results from a
continuous cycle of: (1) formation of protective fluoride films; (2) loss
of these films by rupturing, spalling, fluxing, washing actions, and dissolution
in highly corrosive condensates; (3) reformation of the protective films due
to an excess amount of fluorine in the system over that quantity necessary to
fluorinate UFM to UF6; and (4) additional loss of the films via the means
stated under 2. At the vapor-salt interface, this cycle could be accelerated
due to the vigorous bubbling action of the sparging hydrogen fluoride and the
rise and fall of the bath level. Regardless of the mechanism by which interface
-attack proceeds, all observations to date have emphasized that fiaximum attack
in these systems occurs in this area. This was an important éonsideration of
the design of the full-scale pilot plant vessel (see Section IV of this report).
The BMI studies concluded that INOR-8 was the most promising candidate

-rconstruction material for Volatility Process dissolver vessels.

TII. Studies at the Argonne National Laboratory

A. Experimental Procedures and Results

One of the approaches to reprocessing zirconium-uranium fuel alloys at
ANL has been utilization of fused-fluoride systems in a manner similar to that
employed by ORNL. Studies relating to the selection of a construction material
for containment of the dissolution hydrofluorinafiion environment involved
laboratory-scale corrosion tests on L and A nickel, Inconel, Monel, copper,
Hastelloy B, molybdenum, silver, gold, platinum, tantalum, niobium, and several

grades of graphite.

2lrpid., pp 32-3L.
- 67 -

Barly work at ANL permitted alternate contacting of specimens in the
liquid and gas phases of the dissolution system ("impingement conditions') as

(ref 22’23) The test material usually was used to fabricate

shown in Fig. 33.
the liner, impingement plate, and often the sparge tube. In addition, coupons
sometimes were positioned within the liner. Summaries of these impingement
corrosion test results are detailed in Tables XVIII and XIX. At a later date,
nonimpingement tests, limited to salt-phase contact, were carried out in an
apparatus consisting of a vessel liner, a sparge tube, and a draft tube made
of the test ma.te:r:ial.glL

Figure 34 shows the single-phase test configuration, while the test
results are detailed in Table XX. Of particular importance during these early
studies was the ANL observation that the dissolution rate of zirconium was
directly proportional to the hydrogen fluoride velocity. This information was
later confirmed in studies at ORNL and was used in designing the ORNL Volatility
Pilot Plant Hydrofluorinator Dissolver,

Additional tests at ANL were carried out on several materials using
strips of the test material wired together to form a simulated draft tube of

25

rectangular cross section. The strips were 6 in. long and 1/2 in. wide. The
draft tube assembly was attached to an A nickel sparge tube and partially immersed
in sodium-fluoride, zirconium-fluoride salts. Thus, when hydrogen fluoride was
introduced below the melt, alternate liquid-gas contacting of the internal sur-
faces occurred. The container vessel liner, in all cases, was A nickel. Table XXI

shows the results of this test series.

B. Discussion of Results — Conclusions

The results of the ANL studies were variant depending on the process
conditions, but limitations were noted for all of the metallic test materials.

Both L and A nickel suffered high corrosion rates with alternate vapor and

22W. B. Seefeldt et al., Chemical Engineering Division Quarterly Report,
April, May, and June, 1956, ANL-5602, pp.16-30.

23w. B. Seefeldt et al., Chemical Engineering Division Summary Report,
July, August, and September, 1956, ANL-5633, pp 15-21.

ehw. B. Seefeldt et al., Chemical Engineérihg Division Summary Report,
Qctober, November, and December, 1956, ANL-5668, pp 15-18.

QSW} B. Seefeldt et al., Chemical Engineering Division Summary Report,
January, February, and March, 1958, ANL-5858, pp 31-32.

- 68 -

UNCLASSIFIED
ORNL-LR-DWG 55815

FLUORIDE SALT
-MELT LEVEL

VESSEL LINER

SPARGE TUBE

’.’”””""""/"”””’/”””A”””/”’””””’,””””’
A 1

AAAAETITHTHAIRITHTTAGRTE T I HE I A GG AT TTE ; T E i TEE T HT TR A I HTH I T T AT A TR TARARRARRERY AN

IMPINGEMENT
PLATE

SONOONNONNNNNSNANNNNNY
) .

SONNNANNNNNNNNNNNNNY

Argonne National Laboratory Hydrofluorination Corrosion

Test Apparatus for Alternate Vapor-Salt Contact (Impingement Conditions).

33.

Fig.
- 69 -

Table XVIII.  Corrosion Tests at ANL on Selected Alloys in Equimolar

NaF—ZrFu and Hydrogen Fluoride at 600°C

Material

Test Periods — ~50 to > 1000 hr

Hydrogen Fluoride Rate — ~20-57 g/hr
Alternate Vapor and Salt-Phase Contact

Corrosion Rate (mils/month)¥

Liner Impingement
Wall Plate Coupon _ Remarks

L nickel

A nickel

Copper

Inconel

Monel

Hastelloy
B

36-57 - . 5,1-8.7 After 410-630 hr exposure, liners
were embrittled. Coupons showed
2—7 mils intergranular attack.

9,0-60 12-45 5.4-36 After 50-120 hr exposure, liners
were embrittled severely due to
formation of Ni-Ni S, eutectic.
Sulfur source was identified as
contaminated hydrogen fluoride.

Rapid -- 4. 2—29 Excessive external oxidation. Mass
Failure transfer of copper occurred believed
' due to excessive temperature gradients.

5487 - 19-25 Chromium depletion noted; at later
times, resulted in stratification
and spalling of depleted layers.

16.5 6.0-18 -- Copper depletion noted plus
' : stratification of depleted layer.

-- -—- 0.9-1.5 Embrittled-intergranular fracture
on bend tcst; age hardening
believed to be occurring.

*
Based on dimensional changes.
Table XIX.

_'70 -

Corrosion Tests at ANL on Graphite in Equimolar

NaF-ZrF) and Hydrogen Fluoride at 600°C

Test Periods —

~170-340 hr

Hydrogen Fluoride Velocity — 2-3.5 ft/sec
Alternate Vapor and Salt-Phase Contact

Corrosion Rate (mils/month)*

Titanium
Carbides¥*¥*

Liner Impingement HF Entry
Material Wall Plate Tube o Remarks
CS Graphite - 0.75-5.1 0.0-5.1 1.1-4,3  Very little salt permeation
AGR Graphite 1.1-1.7 0.4 1.5
ATZ Graphite 0.6 k.3 4.3 Little salt permeation during runs
: but liner cracked after allowing
melt to solidify
HC Graphite 6.4 1.1 0.4
-HIM Graphite 2.1 1.3 0.9
"HPC Graphite 2.1 0.2 0.6

After approx 40 hr, insufficient
material was left for evaluation

- %
Based on dimensional changes.

*3
Approximately 70% TiC plus 10% NbC or TaC bonded with nickel or cobalt.
- 71 -

B ) UNCLASSIFIED
ORNL-LR-DWG 55846

VESSEL LINER ——ma_

SPARGE TUBE

i——— FLUORIDE SALT
MELT LEVEL

DRAFT TUBE

L8
"

HF ENTRY HOLES

R TIIIIETE IS IS ST T ERERI IR OT I T EE ST EI TSI,

i:'le:

Fig. 34. Argonne National Laboratory Hydrofluorination Test
Apparatus for Salt-Phase Corrosion Tests (Nonimpingement Conditions).
- 72 -

Table XX. Corrosion Tests at ANL on Some Materials Used
as Limers to Contain Equimolar NaF-ZrFu and
Hydrogen Fluoride at 600°C

Test Periods — ~L00-1000 hr
Hydrogen Fluoride Rate — ~20-52 g/hr
Salt-Phase Contact

Corrosion Rate

Material (mils/month)* - Remarks

L nickel 1.8-12 .

A nickel 1.5-18 - Approx 1/2 of the samples examined
were embrittled severely

Monel 3 _

Graphite -- Reasonably ductile, some microscopic

cracks

* .
Based on dimensional changes.

Table XXI. Corrosion Tests at ANL in Equimolar I\TaF-ZrF)+
and Hydrogen Fluoride at 600°C

Test Periods — T2—667 hr
' Hydrogen Fluoride Rate — ~21-27 g/hr

Alternate Vapor and Salt Contact
Corrosion Rate ‘
Material (mils/month)* | Remarks

Molybdenum 0.2-<2

Silver . 0.3 < 10 Formation of subsurface voids and sur-
face blisters noted. Conjectured as
being the result of reaction between
‘hydrogen and internal metal oxide

Gold <.10

Platinum 2 ~ Intergranular attack at interface and
at lower end of specimen exposed to
salt phase

Tantalum - - Dissolved readily

Niobium -- Dissolved readily

%
Based on dimensional changes.
_73_

salt-phase contact indicating a severe interface corrosion problem. Also,
serious embrittlement of both types of nickel occurred which was traced ulti-

mately to the formation of the Ni-Ni eutectic at the grain boundaries. The

S
source of the sulfur was found to be3tie hydrogen fluoride process gas. Some
evidence was found that Monel also was sensitive to sulfur embrittlement. Al-
though several methods of sulfur removal were considered, ANL adopted the
philosophy that adequate removal of sulfur from all reagents used in the fused-
salt volatility procese wae impossible. '

Selective losses of chromium and copper from Inconel and Monel,
respectively, were noted during the tests, making the use of those materials
of construction undesirable. Although Hastelloy B showed low corrosion rate
losses, it was discarded due to its embrittling tendency during exposure to
the serviqe temperatures. Copper showed rather high rates of attack in the
few exposures made and evidence of mass transfer»(i.e., metal deposits) was
found. |

Scouting tests on molybdenum, silver, gold, and platinum.showed
promising results, especially in the case of molybdenum. However, because of
the cost and fabrication disadvantages of these materials, they were eliminated
in favor of graphite which also had experienced generally low corrosion rates
during simulated hydrofluorination tests.

Following the decision to use graphite as the construction material
for a pilot-scale hydrofluorinator dissolver, ANL initiated extensive studies
on carburization of metals which would serve as back-up materials or Jjackets

26,27

for a graphite vessel. The presence of the fused-fluoride melt seemed
conducive to carburization, but several materialé appeared suitably resistant
to attack. These materials included nickel and the austenitic stainless
steels,

Permeability of graphite by the process media also was studied in some

detail but the test data did not correlate this characteristic with known

égfi. B. Seefeldt et al., Chemical Engineering Division Summary Report,

July, August, and September, 1956, ANL-5633, pp 21-22.

27W. B. Seefeldt et al., Chemical Engineering Division Summary Report,
October, November, and December, 1956, ANL-5668, pp 17, 19-20.

.‘7u_

28,29

properties. Pefsonnel at ANL noted that exposed graphite crucibles some-
times cracked after two or three days standing at ambient indoor temperatures.
and reasoned that melt hydration and attendant salt expansion was responsible.

A-hydrofluorinator dissolver; fabricated fram slabs of graphite and

thermally insulated with carbon‘black,’subsequently was installed at ANL for

" use in demonstration process runs. The vessel was desigfiéd so that salt
pehetrating the graphite or leaking through mating surfaces would freeze and
be immobilized. This was accomplished by providiné internal electric resistance
heaters, -thick graphite walls (1.5 in.) and insulation (1.5 in.) so that the
outer-wall temperature is.below the salt melting point. Cross sections of the
graphite dissolver and a heating element are shown in Fig. 35. To date, a

. number of dissalution runs using simulated fuel elements has been completed.
The only reported difficulties with the vessel have been intermittent salt
plugging_of the slug-charging éhute and off-gas lines and replacement of a

- L? graphite heater and the salt-trahsfer‘line. The latter components cracked

. during shutdown periods. A _

b - : The heater failure was attributed to salt entrapment in the annulus

and the salt-transfer line cracking was felt to be the result of damage incurred

durlng a maintenance period. 30,31

IV. Oak Ridge National Laboratory Volatlllty Pilot Plant Hydrofluorinator
. Dissolver

- ' “The "search for a suitable matefiai of construction for the ORNL Fluoride
Volatility Process dissolver vessel has been described in the preceding sections
of this report. The limitations associated with each of the elemental metals
or alloys studied and their relative compatibility with the volatility-dissolution

environment also havebeen discussed. Graphite,'the construction material for

28,

=. W. B. Seefeldt et al., Chemical Engineering Division Summary Report,
July, August and September, 1957, ANL-5789, pp 1620 (confidential).

298 Vogler et al., Chemical Englneerlng Division Summary Report January,
February, and March, 1958*’ANL-5858, pp 17—-21.

3OR W. Kessie et al., Chemical Engineering Summary Report, October, November,
and December, 1959, ANL—6lOl p 107.

3;R W. Kessie et al., Chemical Engineering Summary Report, January, February,
and March, 1960 ANL-fls,pp 120-121.

Bt B s i it e e v A A Ak EAC ot Nl e et e e ke e R oaiee o
- 75 -

] UNCLASSIFIED

ORNL-LR-DWG 55817

|l4 30 in. |

———————— 24 in, ——————

H
N
=)
i
@
3
l!
_‘_
&
E)
3 in, —=—

. 1/4—in.
. NICKEL CLAD n
\&“ o —
I ’f
- .”0
g -in. STEEL ™ o
W FLANGE N
’;252 EN
:.- BRASS HEAD §§
HF SPARGE LINE — 3 gg
:. =N
3 | DIVIDER WATER COOLING ] FEENED
Y PLATE COlL.S 3 =
< %
I 2
™ ¥
SEAL GASKET R
- ooy
000 Y _: RS
GRAPHITE CRUCIBLE # gl S =
1%-in. THICK WALL — GRAPHITE SHELL g ©
. ; (HOLES IN LOWER PART) S
3 (3-in. OD)
- B fi—’r GRAPHITE SUPPORT ROD
— ORAF T ~
Yp-in. STEEL TUBE ACME THREAD %
A SHELL—\ i o
0 ::\.
3 GRAPHITE CENTRAL &
R , - 3
K- — PROTECTIVE S HEATING ROD %
GRID < ;E:
MACHINED SLIDING FIT |
LAMPBLACK
1INSULATION P 6AS
1%,-in, THICK--
- OSTAIBUTOR | GRAPHITE HEATING ELEMENT

GRAPHITE DISSOLVER

Fig. 35. Argonne National ILaboratory Graphite Dissolver Vessel and Heat-
ing Element. Ref: Chemical Engineering Division Summary Report, January,
February, and March, 1958, ANL-5858, p 25.

- 76 -

the ANL hydrofluorinator dissolver, was rejécted by ORNL personnel charged

with the selection of a material for the Volatility Pilot Plant dissolver vessel.
This rejection was based on graphite's poor structural properties and thé con-
cern that problems would arise in decontamination and uranium recovery from
large porous graphite bodies. o

Based on comparative testing, INOR-8 was the logical choice of material
for the full-size hydrofluorinétor dissolver vessel, although INOR-8 had not
demonstrated exceptional corrosion resistance to the hydrofluorination environ-
ment. Modification to the original dissolution flowsheet, to improve the
reliability of INOR-8 for the veséel in question, included:

1. The use of NaF-LiF-Zth as the dissolution salt permitting lower tempera-
tures (approx 650-700°C) at the start of dissolution, and after the commencement
of dissolution (approx 500°C). | “

2. The retention of bulk-zirconium metal in the dissolution bath whenever
~hydrogen fluoride is present for a maximum possible time.

3. 'The avoidance of a fixed vapor-salt interface region through charging
and dpérating techniques;

Based on available data, a full-scale prototype hydrofluorinator dissolver
-approx 17 ft tall was designed by the Process Design Section, Chemical Technology
‘Division. . The final configuration is shown in Figs. 36 and 37. In'éeresting
idesign facets include an assigned corrosion allowance of 125 mils and an un-
fisually small-diameter lower section.. The latter allows subassemblies to be
vertically stacked in this section in order to.achieve high dissolution rates
with maximum metal concentrations. This vessel, the VPP Hydrofluorinator
Dissolver Mark I, has been fabricated from INOR-8 '(Hastelloy N) in the ORNL
shops using reactor-grade materials, welding,and inspection techniques. The
vessel was installed in Bldg. 3019 at ORNL and estimates indicate reprocessing
of naval reactor fuel elements will begin the last half of calendar 1961.

Figure 38 shows the vessel in an early stage of installation in the Volatility

Pilot Plant.
- CONCLUSIONS

1. The corrosion resistance of various materials has been studied in
laboratory-scale and semiworks~size hydrofluorination dissolution tests. The

results indicated that it was feasible to use INOR-8 for construction.of a N

- T7 -

UNCLASSIFIED
ORNL-LR-DWG 39255R

SLUG CHUTE

VAPOR OUTLET\ DIP LINE
LN s
-— 24 —in. OD
<
SALT INLET— ~— g in. o
N\ ~
I ) - 3g-in. NPS &
— 0.091-in. WALL e
\D)]
N
' !
{—in. OD b oin
0.253-in. WALL /fl z N
E )
(QV
~
5'/2-in. 0D — I~
Y4 in. £
o
\®)
[MATERIAL INOR-8] o

HF INLET —» =2 _ o SALT OUTLET
~ 5° SLOPE FROM HORIZONTAL

Fig. 36. Mark I INOR-8 VPP Hydrofluorinator Dissolver.
- T8 -

UNCLASSIFIED
ORNL-LR-DWG 39152

HF OUT
"y 7-CN CALROD
\ ) HEATER

CASTERS
TRACK

FIRE BRICK

AIR ADJUST-
MENT VALVES

Iy \ SALT OUT

Fig. 3(« Mark I INOR-8 VPP Hydrofluorinator Dissolver and Furnace.
- 79 -

Unclassified
ORNL Photo 47611

Ll

I

Fig. 38. Mark I INOR-8 VPP Hydrofluorinator in an Early Stage of
Installation.
- 80 -

large pilot plant dissolution vessel. A Fluoride Volatility Process hydro-
fluorinator has been constructed of this alloy and is presently operating at
ORNL.

2. Disadvantages associated with the use of INOR-8 in the above applica-
tion include the leaching of chromium and susceptibility to pitting attack
under certain process conditions.

3. The presence of zirconium metal or Z.I‘FLL during fluoride-hydrogen
fluoride dissolution significantly retards the corrosion of INOR-8.

4k, Other materials, such as graphite, copper, molybdenum, tungsten, and
the noble metals, deserve further consideration for hydrofluorinator use.
Graphite is presently being used as a pilot-scale construction material for

hydrofluorination processes at ANL.
ACKNOWLEDGMENT

The authors wish to acknowledge the assistance given by the Metallography
Section of the Metallurgy Division, by personnel of the Spectrochemical Group
and the Special Analyses Laboratory of the Analytical Chemistry Division, and
by the Graphic Arts and Photography Departments.

A portion of the corrosion evaluations reported in this document and the
accompanying illustrations are the work of the Corrosion Research Division,
Battelle Memorial Institute, and their contribution deserves special mention
here.

Thanks are due to the personnel of the Chemical Technology Division
working on the ORNL Fluoride Volatility Process who aided in gathering process
data or in reviewing this publication.

Special thanks are due J. H. DeVan, D. A. Douglas, Jr., and E. E. Hoffman
for their constructive appraisal of this report and to the Metallurgy Division
Reports Office for their patient typing and careful preparation of the material

for reproduction.

- 81 -

BIBLIOGRAPHY

G. I. Cathers and R. E. Leuze, "A Volatilization Process for
Uranium Recofiery," Preprint 278, Paper presented at Nuclear Engineering
and Science Congress, Cleveland, Ohio, December 12-15, 1955; also printed
in "Selected Papers," Reactor Operational Problems, Vol 2, Pergamon Press,

London, 1957.

G. I. Cathers, "Fluoride Volatility Process for High-Alloy Fuels,"
Symposium on the Reprocessing of Irradiated Fuels, Held at Brussels, Belgium,

TID-7534%, Book 2, pp 560-573 (May 20-25, 1957).

G. I. Cathers, M. R. Bennett, and R. L. Jolley, The Fused Salt Fluoride
Volulility Process for Recovering Uranium, ORNL-2661 (April, 1959).

~ THIS PAGE
WAS INTENTIONALLY
LEFT BLANK
- 83 -

ORNL-2833
UC-25 — Metals, Ceramics, and Materials
TID-4500 (16th ed.) |

INTERNAL DISTRIBUTION

Biology Library 65. ‘C. S. Harrill
Central Research Library 66~70. M. R. Hill

Reactor Division Library 71. E. E. Hoffman
ORNL — Y-12 Technical Library 72. A, Hollaender
Document Reference Section 73. R. W. Horton
Laboratory Records Department 74. A. S. Householder

Laboratory Records, ORNL-RC 75. R. L. Jolley

G, M. Adamson, Jr. 76. R. G. Jordan (Y-12)

E. J. Barber (K-25) 77. M. T. Kelley

M. R. Bennett 78. R. E. Lampton

H. A Bernhardt (K-25) 79. J. A. Lane

D. S. Billington 80. W. H. Lewis

R. E. Blanco ~ 81. R. B. Lindauer

F. F. Blankenship 8a—84. A. P. Litman

A. L. Boch 85. R. S. Livingston
E. G. Bohlmann 86. J. T. Long

B. S. Borie 87. H. G. MacPherson
J. C. Bresee 88. C. J. McHurgue

R. B. Briggs 89. W. D. Manly

F, N, Browder 90. S. Mann

K. B. Brown 91. J. L. Matherne

W. A, Bush 92, F. W. Miles

D. 0. Campbell 93. R. P. Milford

W. H. Carr, Jr. ok. A. J. Miller

G. I. Cathers 95. E. C. Miller

C. R. Center 96. E. C. Moncriet

W. E. Clark 97. K. Z. Morgan

J. A, Cox 98. J. P. Murray (K-25)
H. J. Culbert (K-25) 99. M. L. Nelson

F. L. Culler, Jr. 100, R. G. Nicol

J. E. Cunningham 101. P. Patriarca

J. H. DeVan 102, D. Phillips

D. A. Douglas, Jr. 103. J. B. Ruch

D. E. Ferguson 10k, H. W. Savage

J. H Frye, Jr. 105. H. E. Seagren

J. H. Gillette 106. L. D. Shaffer

H, E. Goeller 107. E. M. Shank

A. E. Goldman 108. T. Shapiro (K-25)

A, T, Gresky 109. M, J. Skinner

W. R. Grimes - 110. S. H. Smiley (K-25)
C. E. Guthrie 111. C. O. Smith

C.. F. Hale (K-25) 112, S, H. Stainker
.84 -

. Swartout 120,

113. J. A E. L. Youngblood

114. E. H. Taylor o 121. A. A. Burr (consultant)

115. A. M. Weinberg - 122, J. R. Johnson (consultant)
116. M. E. Whatley . 123. C. S. Smith (consultant)
117. C. L. Whitmarsh 12k. R. Smoluchowski (consultant)
118. J. C. Wilson ' 125. -H. A. Wilhelm (consultant)
119. C. E. Winters '

EXTERNAL DISTRIBUTION

126, E. L. Anderson, AEC, Washington

127. D. E. Baker, GE Hanford

128. J. E. Bigelow, AEC, OAD
120-130, David F. Cope, AEC, ORO

131. F. R. Dowling, AEC, Washington

132, O. E. Dwyer, BNL

133. Ersel Evans, GE Hanford

134k, F. W. Fink, BMI

135. dJ. L. Gregg, Cornell Unlver51ty
136. S. Lawroski, ANL

137. P. D, Miller, BMI

138. W. Seefeldt,. ANL

139. J. Simmons, AEC, Washington
140. E. E. Stansbury, University of Tennessee
141, M. J. Steindler, ANL

142. R. K. Steunenberg, ANL

143. D. K. Stevens, AEC, Washington
14k, J. Vanderryn, AEC, ORO

145, R. C. Vogel, ANL

146720, Given distribution as shown in TID-4500 (16th ed.) under
Metals, Ceramics, and Materials Category (75 copies - OTS)

-
[}