ORNL-2832.txt

MASTFR

ORNL=-2832
UC-25 — Metals, Ceramics, and Materials

CORROSION ASSOCIATED WITH FLUORINATION
IN THE OAK RIDGE NATIONAL LABORATORY

FLUORIDE VOLATILITY PROCESS

A. P. Litman
A. E. Goldman

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 M 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; ur

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-2832

Contract No. W-T7405-eng-26

METALLURGY DIVISION

CORROSION ASSOCIATED WITH FLUORINATION IN THE.
OAK RIDGE NATIONAL LABORATORY FLUORIDE VOLATILITY PROCESS

P

A. P. Litman and A, E. Goldman

DATE ISSUED

. JUK 18 1657

OAK RIDGE NATIONAL LABORATORY
Oak Ridge, Tennessee
operoted by
UNION CARBIDE CORPORATION

for the
U. S. ATOMIC ENERGY COMMISSION
 THIS PAGE
WAS INTENTIONALLY
LEFT BLANK
Summary--

I. Mark
A.
B.
C.

*I1. Mark
A,
B.
.C.

iii

CONTENTS

I Volatility Pilot Plant L Nickel Fluorinator-------------
Métefial Selection and Fabrication----cccccmmcmmcmcmm e
Operational History ---------------------------------------
Reaction to Environment----=--- e —— -, ———————————
1. Chemistry---------=-== - —————— - = me—r——ar———
2. Dimensional Analysis---------aa; ______________________
3. Metallographic Study--==-se--cccec-- e, ————————
4, Summary of Corrosive AttacK-----mmmccemccccocaccmocaa-
Discussion of ResultS-ee-cecmmmmc e e

1. Individual Actions of F2, UF6, and Fused Fluoride
Salts----cermr e -

2. Collective Attack During Volatility Process

Fluorination--c-mercmcmm e e
a. Interior Bulk LosseS----ccccmmmcc e ccec e
b. Interior Intergranular Attack-------e-cecoe---
c. Exterior Intergranular Attacke---eccccmccmcccaa-
d. drain—Size Variations--=---eecacmcao-- e
IT Veolatility Pilol Plant L Nickel Fluorinator------------
Material Selection and Fabrication-Design Changes-«--«---~ -
Operational History---=e-e--- e — e ——————————————————
Reaction tO'Environmenfi -----------------------------------
1. Visual and Vidigage Inspections---=w--we-cccoococomon-
2. Chemistryemmececrmcmrecm e r e e e e ——————————-
3. Dimensional Analysis---=~c-cccomccmnccanmcanen- -
L. Metallographic Study~-==~=mc-cemmmm e
5. Summary of Corrosive Attack----s-cmomcammccmsmn e
- Discussion of RESULE S = mom e oo e
1. Interior Bulk lLosseS---=-—ccemwecu--- D o e e e e i e e e e
2. Interior Intergramular Penetration----e~«eemeccecc—cecna--
3. - Exterior Intergranular Attack-=-m--e—cmcmmccmmcommcmeeae
L. Grain-Size Variations--s--ee-meecmooe- e ———————

N v
iv .

E. Corrosion of Internal Components from the Mark IT

VPP Fluorinator----——-—-c--- o ;-—;--_ 71
IITI. Bench-Scale Fluorlnatlon Corr081on Studles--— ------------- - 7T -
A. TInconel Fluorlnator ---------------------------------------- 79
1. - Test MethoG--=-ccmmmm e e e e e e e 79
2. Discussion of Results----memm-- ;7---5---—--——f ------- - 80

‘B. A Nickel Miniature Fluorinators-----=emeecccccccee—ccc—m~ -- 8k -
1. Test MethoG--rmmmmmmmm oo s oo e 84
2. Discussion of ResultsS-=--—ccmmmmmcm o eem s 89
C. INOR-8 Fluorinators--------——ee--- —_———— e mmmm—=== G5
1. Test Metho@-=-=cmommmmmmmlioommcmeo B L LTI - 95
2, Discussion of Results--; ----- —————— U 1
IV. = Volatility Pilot Plant Scéuting Corrosion Tésts-----------;-;;- 108
A. Material Selection-------w-- -------7------—-r-;-----4 ----- 108
B. Test Met OG- = mm = m e mm o m e e mm e m e mmmmemm e 1173
C. Reactions to Environménté-;4--;¥ ------------------ e 113
D.' Discussion of Results------cc-mmmmmmm e e - 119
E. TFuture Studies---mr---mecmmnone mmmmmmm et oee 123
V. Argonne National Laboratory Fluorination Corrosion Stuéiés-;--; 125
A, Test MeBhO@= = o - == e m o e e e e mmEE A ————————— e e e 125
B. Discussion of Results--=--=-cmmmmmmmomomommmcoan e 126
VI. Supplementary Volatility Pilot Plant Equlpment-j --------------- 129
Acknowledgment-——;f———--——-—-—7———--a ------------- —————— T 129
Bibliography ------ [ e m—————— e 130

Appendix A - Photomicrographs of VPP Scouting Corrosion Test

Spec1mens--——-—---—~-———-—-——-—-e-———-—e ------------------- ——————— 131
Appendlx B - Supplementary VPP Equipment«--------—~- ecmmccmmmemmeme 153
Complexible Radioactive Products Trap------w---emee——cceea— - 155"
Waste-8811 Linee—m e m oo oo e e e e . 159
AbSOTbEerS=n-—mm e e e e e e e e 166
Valves and Fittings--rm=ccmcmcm e - 173
Fluorine Disposal System-~—---ceceoomommmeaoan [ ———— 17k

Process Gas Lines--—-—-———-- ) e ———— 184
CORROSION ASSOCTATED WITH FLUORINATION IN THE
OAK. RIDGE NATTIONAL LABORATORY FLUORIDE VOLATILITY PROCESS

A. P, Litman and A. E. Goldman

SUMMARY

This repdrt evaluates chemical corrosion on reaction vessels and
equipment used during the fluorination of fused-salt fuels and subsequent
associated operations in the Oak Ridge National Laboratory (ORNL) Fluoride
Volatility Process and is & continuation and expansion bf the Metallurgy
Division assistance to the Chemical Technology Division in this regard.

‘ Tfiéjfiuorination phase consists of converting uranium tetrafluoride to
volatile uranium hexafluoride by fluorine sparging of molten fluoride salts
and subsequent decontamination and recovery of the uranifim hexafluoride.

For convenicnce in reporting, this document is‘divided into six sec-
tions. Sections I and II describe the corrosion behavior of full-size
fluorination vessels fabricated from—L nickel and used during Volatility
Pilot Plant (VPP) operations. Section III covers corrosion evaluations of
bench-scale fluorinators made of A nickel, Inconel, and INOR-8, which.were
operated by the Volatility Studies Group, Chemical Development Section A,
of the Chemical Technology Division, Section IV describes scouting tests
of many proprietary and nonproprietary materials exposed to the pilot plant
fluorinator environments and the reactions of the various materials to
those service conditions. Appendix A shows selected photomicrographs of
the corrosion specimens described in Section IV. For comparison, results
of some of the corrosion tests performed by the'Argonne National Laboratory
on metal coupons under simglated fluorination conditions are reported in
'Section V. Section VI and Appendix B deal with results of examinations of
supplementary VPP equipment including a radioacti?e-products.trap, a
waste-salt line, the absorbers, valves and fittings, the fluorine-disposal
system, and process-gas lines.

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

on molten salt residence time or mils per hour based on fluorine exposure
1
N
1

pur)

time. These rates are included specifieally for comperison purposes, are
.not exact, and should not be extrapolated into longer time periods for.
design work or other applications. ;
.~ Two fluorinators were used in the VPP to carry out the fluorination
reactions. These vessels, Mark I and Mark II, were fabricated into right
cylinders, approx firl/E ft in height, from the same heat of t (low carbofi)
nickel. The first vessel .contained equimolaf NaF—ZrFu or NaF-ZrEu—UFu
(L8-48-U4 mole %) for approx 1250 hr at 600-725°C. Over a period of €l hr,

57 500 standard liters of F. were sparged into the sdlts. This constituted

2
a F,:U mole ratio of 3:1 beyond theoretical requirements.  The Mark TIT

fluirinator contained fluoride salts of approximately the same compositions
plus small additions of Pth during three runs. The salts were kept molten

:at 540-730°C for approx 1950 hr and about 60 500 standard liters of F2 were
sparged. into the Mark IT melts in 92 hr,.

Both fluorinators sustained large corrosion losses consisting of exten-

.lgive‘wéll thinning, severe interior intergranular attack, and a mederate
exterior‘oxidation attack. Maximum deterioration on.the Mark I vessel oc-
curred in the middle vapor region at a calculated rate of 1.2 mils/hr, based
.on fluorine sparge time, or 46 mils/month, based on time of exposure to mol-

- ten salts.  The second vessel showedlmaximum attack in the salt-containing
region at similarly calculated rates of 1.1 mils/hr and 60 mils/month., Some
evidence was found to indicate that the -intergranular attack may have resulted
from sulfur in the systems. Bulk metal losses from the vessel's_walls were
believed to be the result of cyclic losses of NiFé "protective" films. The
films were .formed -on the interior walls of the fluorinators during conditioning

'and_fluorination treatments and lost as the result of rupturing, spalling,
fluxing,'washing aetions, afid/or dissolution in highly cerrosive condensates‘
formed during operetions. The shift in maximum corrosion attack geometry
in the two fluorinators is believed to have resulted from differenees in
operating conditions. The Mark II vessel experienced higher temperatures,
longer fluorine exposure times, and'extendea uranium residenee times in its

salt baths.
>

- 3 -

Specimens removed from the wall of the first fluorinator showed a

variation in average ASTM grain-size number of 56 to > 1, the largest grains

- being found in the middle vapor region. The second vessel had a more uniform

grain-size pattern, average ASTM grain-size numbers varying from 3~5 to ol
The variations in grain sizes are believed to have resulted from variable
heating rates during initial usage. Low rates permit more complete internal

stress recovery prior to the start of recrystallization which results in

- fewer nucleation sites and therefore larger grains during recrystallization.

Metallographic examinations did not pfovide evidence ot a causal relationship

between grain size and fluorinator wall corrosion,

.- - s BExaminations of bench-scale reactors, where simulated fluorination

environments were provided to study process variables and corrosion, showed

that A nickel had the highest degree of corrosion resistance as a fluorinator

‘material of construction when compared with Inconel and INOR-8. Intergranular

- penetration and subseqfient sloughing of whole grains seemed to be the pre-

dominant mode of corrosive attack on the Inconel vessel. At the higher test

témperatures, 600°C, INOR-8 miniature fluorinators showed large bulk metal

losses plus selective losses of chromium, molybdenum, and iron from the exposed

alloy surfaces. Evidence of a marked reduction in attack on nickel and INOR-8

. was found during lower .temperature studies at 450-525°C, These lower tempera-

ture operations were made possible by adding lithium fluoride to the sodium
fluoride-zirconium tetrafluoride salt mixtures.

Scouting corrosion tests were performed in the VPP's fluorinators using

rod, sheet, or wire specimens of commercial and developmental alloys. These

tests were subjected to serious limitations due to the lack of control over

" operating conditions and thus considerable variation in the corrosion of L

nickel control specimens resulted. Those nickel-rich alloys containing iron
and cobalt showed some superiority in corrosion resistance when compared -with
L nickel specimens. This was probably because of the low volatility of iron
and cobalt fluorides. Nickel-rich alloys containing molybdenum additions
showed variable behavior in the fluorination environment. Some of the data

suggested improved resistance over L nickel while other tests showed the reverse.
S M

Since both of the known molybdenum fluorides that could be formed during
fluorination have very high volatility, one would not expect improved resistance
from molybdenum additions. The experiments emphasize that the present method
of selection of test materlals based on the low volatility of netal-fluorides

" that may form duriné fluorination continues to have merit. Additional ex-
perimental nickel-base alloy corrosion specimens, containing magnesium, alumi-
num, iron, cobalt, or. manganese, have been fabricated and will be used in
future screening tests in a subsequent pilot plant fluorinator.

A review of one fluorination test series conducted by tne Argonne

VNational Laboratory gave general agreement with ORNL scouting corrosion test

**specimen:reSults;-althoughweomparisons:were,hampered by different test con-
ditions. . The Argonne National Laboratory has suggested that the corrosion

. problem be attacked by further studies on the use of cold wall vessels, spray
towers, or low-melting salts for volatility processes.

Visual and metallographic examinations plus ultrasonic measurements of

other VPP vessels and equipment: fabricated generally from Monel or Inconel
showed a wide variation in re51stance to those various local serv1ce condl-
tions. The studles suggest that Inconel can.continue. to be used ‘as a material
of constrdctionifor some components.but frequent inspectionS'are indicated.
Monel appears generally-satisfactory-for the applications to. date.

From a corrosion standpoint,'the fluorination vessel in the VPP continues
to. be the most vulnerable to attack due to the nature of the contained en-
vironment and the high temperature neoessary for fluorination. The continued
use of L 'nickel for the fluorination vessel does not appear prohibitive for
batch operatlons only due to the present high value of the pilot plant's
product. At present, the only guarantee for improved service -life for nlckel
fluorinators seems to be utilization of the lowest practical temperature.
Although not conclusively proven for the fluorination vessels, reduction of
sulfur contamination and the ensurlng of a uniform, small- graln 51ae in the

fff vessels may improve vessel performance. For long-time fluorlnator integrity,
selectlon or development of & new material of construction, the use of salts

with lower melting points, or the'use of a cold wall vessel seems necessary.
8

_5..

The evaluation of  process ‘corrosion that occurred during the develop-
ment studies of hydrogen fluoride dissolution of uranium-bearing fuel elements,
the head-end cycle of the volatility process, will be covered in a separate

report.l

I. Mark I Volatility Pilot Plant L Nickel Fluorinator

A. Material Selection and Fabrication

The selection of material for the first pilot plant fluorinator was

made by members of the Chemical Technology Division after a study of the avail-

2,3,k

ablecorrosion literature and the ASME Boiler and Pressure Vessel Code.

Nickel seemed to be the most likely candidate material of constfuction,'although

at 600—TOO°C, the anticipated operating temperature range of the fluorinator,
Myers and Delong reported penetration rates of fluorine on nickel of 16-34
mils/month. The ASME Code allowable design stress data above approx 315°C
were not available for commercial purity A nickel (0.05—0.15 wt % C). This
was because of the known effects of embrittlement through intercrystalline
precipitation of graphite in nickel containing carbon after long-time exposure
to high témperatures.5 However, satisfactory design data were available for
low-carbon L nickel at approx 650°C,so this material was selected for the
first pilot plant fluorinator, '

The Mark I fluorinator was fabricated at ORNL from L nickel using a

heat with the vendor's analysis of 99.36% Ni—0.02% C—0.23% Fe—0.06% Cu—0.26% Mn—

0.0k% Si—0.005% S. Annealed plate stock of 1/L-in. thickness was rolled into

1

A, E. Goldman and A, P. Litman, Corrosion Associated with Hydrogen Fluoride

Dissolution in the Fluoride Volatility Process, ORNL-2833 (to be published).

dw. R. Myers and W. B. Delong, "Fluorine Corrosion," Chem. Engr. Prog.
(May, 1948).

3"Engineering Properties of Nickel," Tech. Bull. T-15, The International
Nickel Company, Inc., New York, Revised, p. 21, July, 1949.

Rules for Construction of Unfired Pressure Vessels, ASME Boiler and
Pressure Vessel Code, Section VIII, Am. Soc. Mech. Eng., 1956 Edition.

[

“W. A. Mudge, "Nickel and Nickel-Copper, Nickel-Manganese, and Related
High-Nickel Alloys," The Corrosion Handbook (ed. by H.- H. Uhlig), p. 683,
John Wiley and Sons, Inc., New York, 1943.

T

i
6.
a lh-in.-diam cylinder, S5k in. in height, for the vessel shell and longitudi-
nally seam welded uSing an inert-gas metal-arc (nonconsumable) process. The
filler material used was INCO- 61 welding wire and ORNL Reactor Material
Specification RMW3-5 was used as the baSis of the Jjoining procedure.6 A
nominal 3/8-in.-thick L nickel flanged and dished head was welded to the shell
to form the bottom of the vessel (Fig. 1). _

B. Operational History

The Mark T fluorinator was ‘used by the Unit Operations Section of
the Chemical Technology_Division during preliminary fluorination equipmént
studies for a period of about three months. During that time, no fluorine or
' uranium-containing molten salts were in.contact with the véssel. Table T
cites the process conditions in detail for those studies and for the more
+, eXtensive "M" equipment shakedown and "C" process demonstration runs performed
{ later in the VPP.
| Figure 2 shows'the-position of the Mark I fluorinator.during the VPP
-runs.while Fig. 3 shows .the interior piping, gas dispersion assembly, and the

‘placement of an early gronp of corrosion test specimens. The lower half oft
:3the'fluorinatiop vessel was surrounded by a vertical tube-type electric-

“ resistance furnace of 30-kw rating to provide the necessary heat (600-725°C)
. for operations. During the pilot plant runs, rod-type electric resistance
heating elements with a total rating of 9 kw were installed on the upper ex-
terior walls of the fluorinator. \ _

Prior to exposing the fluorinator to elementai fluorine during actual
fluorination of the fused salts, a "conditioning" cycle-was perforned wherein
fluorine was introduced into the vessel which was heated to 20-150°C to
induce the formation of Nng protective films. Fluorine used in the VPP
was obtained in steel tank trailers from the Oak Ridge Gaseous Diffusion

Plant (ORGDP) fluorine generating station. A flowing stream sample analyzed

by ORGDP personnel indicated the analysis of the fluorine was 95% Fy, < 5% HF

R. M. Evans (ed.) Oak Ridge National Laboratory, Reactor Materials
Specifications, TID-7T0l7, pp. 117—128 (October 29, 1958).

UNCLASSIFIED
PHOTO 52707

INCONEL TOP FLANGE

INCONCL SLIP-OMN FLANGE

A

-«—— |[NCONEL
ZI'F4 (SNOW)-

Ya-in. L NICKEL
COMPLEXIBLE

CSHELL ey P

RADIOACTIVE
PRODUCTS
TRAP
=

E 5 -

- 54 1/4 in. o ,
. ’ *‘
FURNACE
SEAL
23 in.
: —=—GIRTH WELD
Y ~w—3/_in. L NICKEL

DISHED HEAD

Fig. 1. Mark T Volatility Pilot Plant Fluorinator.
Table I. Process Conditions for Mark I Volatility Pilot Plant Fluorina.
(Unit Operations, Volatility Pilot Plant "M" and "C" Runs)

Phase T Phage II Phase IIT
Unit Operations Runs "M" Runs (1-48) "C" Runs (1-15) Total
Temperature; max 600700 600-T725 600725 600-725
(°c)

Thermal cycles ~ 20 ~ 20 10 ~ 50
(room temperature
to 600-725°C)

Time of exposure ~ 90 445 715 ~ 1250
at terperature (~ 30 with N_.sparge)®
(salts molten-hr) (~ 60 withou N, sparge)

Salt composition NaF-ZrF, (50-50) NaF-ZrF, (50-50) NaF-ZrF,-UF (4)

L L L Tl
(nominal mole %) (48-L8-L)

Conditioning None 35 in 14 hr 530 in 0.5 hr 565 in 14.5 hr
fluorineainput
(liters)

Operations None 16 775 in 40 830 in 51 hr 57 500 in 61 hr
fluorine, input 10 hr (7-33 liters/min)
(liters)

UF, exposure (hr) None None ~ 20 hr ~ 20 hr

SThese operations were done at 20-150°C for the purpose of inducing an initial "protective"
film of nickel fluoride on the walls of the fluorinator.

bAn average of 3:1 mole ratio (FE:U) beyond theoretical requirements was used in order to
reduce the final uranium concentration in the salt to a few parts per million.

CTop flange removed, ~ 5 hr.

dSalts were used previously in unirradiated loop studies and therefore contained significant
amounts of corrosion products as shown below. Ref: C. L. Whitmarsh, A Series of Seven Flowsheet
Studies with Nonradive Salt, Volatility Pilot Plant Runs, C-9 Through C-15, p. 10, CF-58-5-113
(May 12, 19508).

Component: 0.08-0,18 wt % Ni, 0.06-0.10 wt % Cr, 0.01-0.02 wt % Fe, 0,01-0,60 wt % Ti,
0.002-3.4 wt % Si.

UNCLASSIFIED
ORNL-LR—DWG 30402A

N2
v F2
DISPOSAL
J UNIT
CHARGE MELT TANK SPRAY NCZZLES~ (MONEL)
(347 STAINLESS STEE.) — >~ -
- —J =T0
| OFF GAS
1 $
’ CHEMICAL
PUMP - CAUSTIC SURGE TANK WASTE
(MONEL)
4 Ny Fyp
5 A ZrFa—CRP MONEL
S TRAP 2
= (INCONEL) COLD TRAP c
HEAT EXCHANGER SHELL (MONEL) <
’ MONEL HZAT EXCHANGER BAFFLES (COPPER) —a T s
> 4
FREEZE
VALVE FREEZE
VALVE il
o, || X
=
S
18
=
FLUORINATOR ( L NICKEL) §|{ > l
\ CHEMICAL TRAP
TO PRODUCT (MONEL)

WASTE CAN RECEIVER
(LOW-CARBON STEEL)

ABSORBERS
(INCONEL)

Fig. 2. Volatility P_lot Plant Flowsheet.
1O -

THERMOCOUPLE WELL

NITROGEN INLET LINE

AVERAGE OF
VAPOR-SALT
INTERFACES
(35 in. BELOW
SLIP-ON FLANGE) —a

UNCLASSIFIED
PHOTO 52706

FLUORINE INLET LINE

/DIFFUSER CONE

_____— DRAFT TUBE

7 CORROSION SPECIMENS

e

CORROSION SPECIMENS

Fig. 3. Interior Piping, Gas Dispersion Assembly, and Placement of
an Early Group of Corrosion Specimens in the Mark I VPP Fluorinator.
O s,

and 1-2% N, and/or 0,.
passed through a fixed NaF pellet bed at approximately ambient temperatures.

Prior to use of the fluorine in the VPP, the gas was

Under these conditions, the hydrogen fluoride content in the fluorine was
lowered to approx 20 ppm.(ref 7)

After conditioning, the system was purged with commercial grade
nitrogen dried to < 1 ppm H20. The nitrogen contained approx 100 ppm O2
which was not removed. The fluorinator was heated to approx 600°C along with
the salt freeze valve and salt inlet line. The latter two components were
heated by autoresistance. Then a batch of fluoride salt was meclted in the
charge melt tank and drained by gravity flow into the fluorinator.

Fluorine was bubbled through the molten salt to convert any UFA in
the salt to volatile UF6. During fluorination, the vessel operated with
approx 25% of its volume filled with about 50 liters of fused salts. The re-
maining 75% of the volume contained variable quantities of fluorine, uranium
hexafluoride, nitrogen, and various metal fluorides of high or intermediate
volatility. During the process demonstration "C" runs, an average mole ratio
of 3:1 (F2:U) beyond theoretical requirements was used in order to reduce the
final uranium concentration in the salt to a few parts per million.

While the vessel wall in the salt-containing region of the fluorina-
tion vessel reached temperatures of 600-725°C, the upper vapor region remained
at lower temperatures. ‘T'he maximum temperature recorded on a thermocouple
attached to the exterior wall of the fluorinator 12 in. down from the slip-on
flange was 500°C. The average temperature in this same region was about 400°C.

After completion of fluorination, the waste salt left in the fluori-
nator was pressure transferred through a freeze valve into a waste container;
and the gas from the fluorinator was passed through an Inconel trap, containing
either nickel mesh or NaF pellets, which was maintained at approx 4O0°C. Prior
to Run C-9, the trap contained nickel mesh for the purpose of collecting ZrFu,

Hsnow, "

and thereafter the unit contained NaF pellets to trap entrained salt,
chromium, and zirconium fluorides. During and after Run C-9, the trap was

termed a "CRP" or complexible radioactive products trap.

7F. W. Miles and W. H. Carr, Engineering Evaluation of Volatility Pilot
Plant Equipmenl, CF-60-7-65, Section 15, p. 228.

= 1B

Downstream from the Snow-CRP trap, the product stream was diverted
through absorbers containing NaF at 65-150°C to absorb the UFg. The un-
absorbed gas, mostly fluorine, was routed through a chemical trap (a NaF bed
at ambient temperature) to retain any residual UF6 and subsequently through
a KOH gas disposal unit to neutralize the fluorine before being exhausted to
the atmosphere. The product, UF6, was desorbed from the absorber bed by
heating it to approx 400°C in a fluorine atmosphere and then passed through
two cold traps maintained at -40°C and -55°C where the UF6 condensed. The
cold traps were isolated from the rest of the fluorination system and heated

to approx 80°C to liquate the UF6 which drained into a heated product cylinder.

C. Reaction to Environment

Ultrasonic-thickness measurements of the fluorinator were made with

an "Audigage," an ultrasonic-thickness measurement device, after the Unit
Operation's preliminary fluorination equipment studies. No detectable metal
losses could be found in either the shell of the vessel or in the bottom head;
this could be expected because no fluorine, uranium-bearing salts, or UF6 was
present during the short period of operation at elevated temperature and what-
ever attack occurred was so slight as to be undetected by the measuring

equipment.

1. Chemistry

During VPP Run C-6, a study was made of the interior deposits
which formed on the wall of the fluorinator. Figure 4 shows the location and
subsequent chemical analyses of these deposits. These data indicate a tendency
for chromium, presumably from impure feed salts, and uranium to collect in the
middle vapor region of the vessel. The values shown for nickel indicate that
extensive corrosive attack had occurred in the system during operations.

After completion of the "M" and "C" runs described in Table I,
the Mark I fluorinator was turned over to Metallurgy for corrosion evaluation.
Figure 5 shows the interior of the fluorinator after retirement. Most of the

interior walls of the vessel below the molten salt levels were free of surface
SLIP-ON

UNCLASSIFIED

ORNL-LR-DWG 49156

ANALYSES OF DEPOSITS FROM VPP MARK-1 FLUORINATOR AFTER RUN C-6

FLANGE
\l

35 in.

55 in.

!

AVERAGE
OF
VAPOR-SALT
INTERFACE

LEVELS

Component (wt %)(1)

Sample Location Region Description
Na Zr Ni Cr F
Underside of Inconel Top vapor Pale blue-green 2.45 1.59 49.7 1.6 0.98 40.4
top flange scale
Interior wall-7 in. Uppe- vapor  Bright biue- 200 1.28 278 151 123 33.0
below slip-on green scale
flange
Interior wall-g’w in. Upper vapor  Dirty yellow- 8.19 3.3 144 34.2 365 37.5
below slip-on green scale
flange
Interior wc:ll-"18 in. Midd e vapor Bright yellow- 1.66 3.66 7.2 48.2 0.75 38.0
below slip-on green scale
flange
Interior woll-z'r’/“ in. Lower vapor  Dirty yellow- 230 3.36 10.2 43.3 0.51 38.3
below slip-on brown scale
flange
Underside of dif- Vapeor-salt Yellow-green 0.15 3.50 23.2 45.0 0.87 41.0
fuser cone interface scale
Outside of draft Salt Pale yellow- 0.26 3.50 33.3 16.3 0.08 40.5
tube green de-
posit

(DORNL Analyses.

Fig. 4. Analyses of Deposits from Mark I VPP Fluorinator After Run c-6.

_E-[_
Unclassified
(ORNL. Photo 41392

L e Ay . ; ':w
% R o b B LS
gt L

Spare Salt} 4

&N

@ Droin Line i s 8
SR
|
3

" Salt
Inl=t Line

e

k Regular Salt 'l
Drain Line (%8

]
ki

—1—('[-

Corrosion}
Specimens

P, %>ffl:g £
& : 5 . V ‘!/’ &

Fig. 5. Interior of the Mark I Volatility Pilot P_ant Fl_crinator After Run C-15.
- 15 -

deposits but the regions above the interfaces were covered with heavy scale
and corrosion products. A solid ring of material, about 1 in. in cross sec-
tion, was present.on the interior of the vessel wall at about the same
elevation as the exterior furnace seal. This was a few inches above the
average elevation of the vapor-salt interfaces. Samples of some of these
interior deposits were submitted for chemical analyses and identification by

x-ray diffraction. The results are shown in Table TT.

Table II. 'The Oak Ridge National Laboratory Analyses of Scale from the
Volatility Pilot Plant Mark I Fluorinator aftcr Run C—lS'a

Approx Composition

Component, wt % Indicated by X-ray

Origin of Sample U Na  Ni Cr Zr F Diffraction Intensities

Underside of Inconel 1.95 0.78 45.54 0.79 0.72 39.60 90% NiF,
slip-on flange 10% NaF-NiF,-2ZrF)

From A Nickel F 0.98 5.10 33.76 0.09.0.64 L41.15 60% NiF,

- inlet tube, NS T - :
approx 21 in. below 30k Ner NiF,-2ZrF),
slip-on flange 10% B, 2NaF - ZxF)

From A Nickel F 0.13 6.64 8.36 0.02 1.18 43.30 -

e 2 .
o inlet tube, S
at vapor-salt interface B s ‘ e

aC. L. Whitmarsh, A Series of Seven Flowsheet Studies with Nonradive Salt,
Volatility Pilot Plant Runs, C-9 Through C-15, p, 1k, CF-58-5-113 (May 12, 1958).

Most of the salt deposits were removed by washing the interior
of the vessel with a mixture of 0.7 M H,0,, 1.8 M KOH, and O.L M Na C)H O, at
room temperature, aided by hand chipping. After cleaning, another visual

inspection was made and the results are given as follows:
- 16 - .

~ Region . Results

Vapor Smooth, etched appearance near the top of the vessel with iso-
lated, shallow pits. A yellow-to-green deposit encircled the
vessel from a point approx 10 in. down to a point approx 20 in.
from the top. Several small. areas of flaking and scallng were

" noted at approx 16 in. from the top in the deposit zone. " The
area from 20 in. down to approx 24 in. from the.top had a bluish
cast and was -rougher in texture than the top section.

Vapor-salt Smooth metallic apfiearance with distinet indentations encircling

interface the vessel at several levels.
Salt Smooth metallic appearance., Flange-to-vessel weld not noticeably
corroded.

In the middle Vapbr region,la tightly adherent, yellow-to-green deposit re-
mained on the wall of the fluorinatof. ~Samples of this deposit, sfirfece, and
subsurface millings, were. removed .and submitted for chemical analyses. Figure 6
details the results which indicate that chromium which had previously been
found to -collect in the upper vapor region of the fluorinator had penetrated
-into the véssel wall to eome depth greater than 10 mils. This chromium con-
»centfetion éredient was.feund both in the upper and middle vapor regione al-

though higher concentrations were found in the former region. No excessive

gquantities of sulfur over that preseht_in'the base'material were found.

2. Dimensional. Analysis

Micrometer measurements were taken in the.three major regions of
the fluorinator in all‘quadrants end show the greatest wall-thickness losses
to be.concentrated in the vapor region of the vessel shell. Figure 7 shows
.a schematic drawing of the vessel ‘and denotes the sections that were removed
from the vessel for these measurements and for metallographic study The loss
data are given in Table III. A full- length vessel section was removed from
the northeast-by-east quadrant and micrometer measurements taken every vertical
inch to establish a corrosion wall-thickness-1loss profile. Flgure 8 shows this
plot and pinpoints the maximum metal loss of 47 mils as approx 12 in. below

the .bottom of the slip-on flange.
SLIP-ON

551

FLANGE ~_
| ].
.‘_l_ ] g
.
<
|
LL.
=z
o
a
. . 2
| <j_ 7z
o
|
35in o
¥ Z
in. g
prd
<{
l' d
=
AVERAGE o
oF — =
VAPOR-SALT 2
INTERFACE -9
LEVELS @
z
o
~

Fig. 6. Analyses of Scale and Millings From VPP Mark I Nickel Fluorinator After
Run C-15 and Vessel Decontamination.

UNCLASSIFIED
ORNL—LR—DWG 49157

Component (wt %){(V)

Sample Location

Cr Fe Mn Zr Na  S(ppm)
Exterior woll sub-surface mlllmgs 0.005 0.17 0.17 10
at lo/20 mils below surface '
Interior wcll surface scale 0.40 043 0.17 0.30 0.7)
Interior wcll surface millings ot 0.12 0.34 0.15 0.007 0.5 5
o/s mils below surface
tnterior well sub-surface millings 0.06 0.21 0.14 0.006 0.07 5
at 5/10 mils below surface ' '
Exterior wall sub-surface millings 0.006 0.41 0.15
ot '0/20 wrils below surface
Interior wcll surface scale 026 0.2 009 025 1.09
Interior well surface millings at 0.09 034 0.13 002 0.28
0/5 mils below surface :
Interior wcll sub-surface millings 0.02 0.18 (.13 0.01 0.07

at 5/]0 mils below surface

(NORNZL spectrographic determinaticns except sulfur which was done by indi-

rect polarographic method.

s

"_)_I"
. UNCLASSIFIED :
ORNL—LR—DWG 49158

—
N A
N T

{2 in.

1 G ———————————

2

i

T

er— FULL LENGTH

~ SECTION
REMOVED HERE

oS

JNEH

:

o p—
-

] NS /7 S ; © S

g
N

TI/77/777
ZH%

7/ 7777777777

777
47

7]
w
W

Sy S s SsSSS S

Ll L Ll LLL L ElLLLLLL

/

/77,
A

l

<

JRDBBNEREI )

Y

e b o

E2

[~ AVERAGE OF

VAPOR SALT INTERFACES

® CIRCULAR COUPONS
TREPANNED FROM
VESSEL WALL

\

LY

4

‘Fig. 7. Schematic of Mark I VPP, Fluorinator Showing Areas Removed
for Metallographic Examination and Micrometer Measurements. ‘
- 19 -

Table IITI. Wall-Thickness Lossesa on Coupons Removed from the Mark T
Volatility Pilot Plant L Nickel FluorinatorP

Location
Elevation Wall-Thickness
Sample (Inches below Loss
Number slip-on flango) Quadrant Region (nils)
N-1 12 North Vapor 38
E-1 12 East Vapor 29
S-1 12 South Vapor 33
W-1 12 West Vapor L7
N-2 35 North Vapor-salt interface 6
E-2 35 East Vapor-salt interface I
8-2 35 South Vapor-salt interface 8
W-2 35 West Vapor-salt interface 10
N-3 L5 North Salt 10
E-3 L5 Kast Salt 8
S-3 L5 South Salt 11
W-3 45 West Salt 15

a A
By micrometer measurement.

~

bOriginal wall thickness = 250 mils.
e | ETCH: ACETIC-NITRIC-HYDROCHLORIC ACID
MATERIAL: L NICKEL

4 ~AVERAGE OF
S VAPOR-SALT
RS | INTERFACES,
{»1 ) A ] N J .|
,:\ ALK ;\\ - GIRTH WELD
P -
L

W
(6]

20

25 30
WALL THICKNESS LOSS (mils)

n
(S ]

ol
(@)
DISTANCE DOWN FROM SLIP-ON FLANGE (in.)

55

UNCLASSIFIED
ORNL-LR-DWG 40728

SLIP-ON FLANGE
T

0.4
g
>
|
gl |
LJ
I
w
L
2
3 ~ g
8ok
e
i g
BOTTOM

Fig. 8. Corrosion Profile and Typical Microstructures from the Mark T
VPP Fluorinator. (Profile based on micrometer measurements from northeast

quadrant full-length shell section.)
oy B

3. Metallographic Study

Based on the corrosion wall-thickness-loss profile plot, areas
were selected from the full-length section of the fluorinator and from previ-
ously trepanned samples for metallographic study. The location of these
areas is shown in Fig. 8. In the as-polished condition, some slight roughening
of the surfaces of the samples was noted. No grain-boundary modifications
such as is common to intergranular corrosive attack could be found.

After etching with an 0.5% HCl, approx 60% HC2H302 and approx 4O0%
HNO3 mixture, the grain boundaries of the interior surtace samples appeared
darkened below the exposed surfaces to depths varying trom 4 to 25 mils.

Figure 8 illustrates the typical structures found at varying elevations on
the interior surfaces of the fluorinator wall. The deepest penetrations were
found on samples from the middle vapor region of the vessel, the vapor-salt
interface, and the salt region of the fluorinator. The exterior surface sam-
ples also showed intergranular attack varying from 3 to 8 mils in depth.

Widely variant grain sizes were found in the metallographic sam-
ples examined. A summary of these sizes, converted to average ASTM grain-size
numbers, is shown in Table IV.

Table IV. Summary of Grain Sizes in Samples Removed from the Mark T
Volatility Pilot Plant Fluorinator®

Distance
down from
slip-on flange ASTM
(1ns) Region Grain size number
1 Vapor 56
9 Vapor Sl
12-1/2 Vapor 0 |
20 Vapor 3
2l Vapor 2
29-1/2 Vapor 5
35 Vapor-salt interface 5
L3 Salt 56

aGrain sizes from interior wall and exterior wall samples were approxi-
mately equal.
= D9

At a later date, in an attempt to determine the reasons for the
variant grain size, diffractometer traces were obtained on selected samples
removed from the wall of the Mark I fluorinator. ©Samples were taken from
the l=, 12-1/2-, and 43-in. levels, respectively, below the bottom of the
vessel slip-on flange. The maximum amount of residual strain was noted for
the l-in. sample with little, if any, evidence of recrystallization having
taken place during the Mark I operations. A very limited amount of recrystal-
lization appeared to have occurred in the lQ-l/Q-in. sample since there seemed
only slightly less strain present in this sample when compared to the one
above. The 43-in.-sample traces showed considerably less residual strain
present than either of the other samples. The indications were that partial,
if not complete, recrystallization had taken place in the lower portion of
the fluorinator wall.

Figure 9 shows a section through the girth weld which was
located in the salt phase of the fluorinator. After etching, the specimen
showed no corrosive attack at low magnification. At higher magnification,
the weld deposit showed a grain-boundary attack similar to that found in the
base metal, but to a lesser degree of severity.

Early work on L nickel corrosion rods placed in the Mark I
fluorinator indicated sulfur contamination was probably a factor in the cor-
rosive attack of the vessel.8 In view of the lack of evidence from chemical
analyses, other attempts were made to prove the presence of sulfur at the
grain boundaries of interior wall specimens from the Mark I fluorinator.

The use of sulfur print papers did not provide positive evidence
of the existence of sulfur compounds at the mating surfaces of the grains.

One sample from the middle vapor region of the fluorinator did present the
colors described as needed for the identification of nickel sulfide.

Figure 10 shows the grain-boundary deposit at the 12-in. level below the
bottom of the slip-on flange after etching with cyanide-persulfate and partial
repolishing to show the deposit to its best advantage. However, duplicate

results could not be obtained with the method.

8L. R. Trotter and E. E. Hoffman, Progress Report on Volatility Pilot

Plant Corrosion Problems to April 21, 1957, ORNL-2495 (September 30, 1953).

w PR =

Exterior Interior
(contact with (contact with
air) salt phase)

Fig. 9. Macrostructure of Section Through Girth Weld 47 in. Below Top
Flange of Mark I VPP Fluorinator. Etchant: Acetic, nitric, hydrochloric
acid. 10 X.
= Bli =

Unclassified

1-35150

.003

1000 X

=)
o
e

.‘3

Fig. 10. Photomicrograph of Sample from Interior Surface of Mark I VPP
L Nickel Fluorinator 12 in. Below Slip-on Flange (Vapor Phase) Showing Grain
Boundary Deposit. Deposit was pale yellow under nonpolarized illumination
and black under polarized light in accordance with Hall's method for
identifying nickel sulfide deposits. Etchant: Potassium cyanide-ammonium
persulfate, partially repolished. 1000X. Reference: A. M. Hall, "Sulfides
in Nickel and Nickel Alloys," Trans. Met. Soc. AIME 152, 283, 1943.

- 25 -

Sulfur in slmost any chemical form in contact with nickel at
high temperatures will result in the formation of a low-melting nickel-nickel
sulfidé eutectic primarily along grain boundaries, leading to embrittlement
of the material.9 Consequently, samples of the fluorination vessel wall were
bend tested? The samples did not show brittle behavior.

Two other techniques were considered as identification methods
for the grain-boundary deposits described. One was the use of a microchisel,
presently under development by the Metallography Group of the Metallurgy
Division, by which some of the grain-boundary deposit could be removed and
subsequently submitted for x-ray diffraction analysis. The very small size
of the'grain—boundary deposits prevented the microchisel's usage in this
éituation. The second technique considered was the use of an electron probe
microanalyzer which could péssibly identify a small portion of the deposit by
X-ray diffraétion analysis, in situ. Such an instrument is not yet available
at ORNL and it was not possible to obtain service time on the few instruments
currently in operation in fihis country. Consequently, the nature of the Mark

fluorinator grain-boundary deposits was left in doubt.

4, . Summary of Corrosive Attack

Table V summarizes the maximum corrosion losses of all types
found in the three major regions of the VPP Mark I fluorinator. The maximum
attack was calculated to be 46 mils/month based on exposure time to molten
salts during the VPP "M" runs (1-48) and "C" runs (1-15) or 1.2 mils/hr based
on fluorine sparge time during’fluorination of molten éalts. The maximum

attack occurred in the middle vapor region.

D, Discussion of Results

A

1. Tndividual Actions of F., UF,, and Fused Fluoride Salts

Major corrosive agents in contact with the VPP L nickel fluori-
nation vessel, Mark I, were elemental fluorine, uranium hexafluoride, and
9W. A. Mudge, "Nickel and Nickel-Copper, Nickel-Manganese, and Related

High-Nickel Alloys," The Corrosion Handbook, (ed. by H. H. Uhlig), p. 679,
John Wiley and Sons, Inc., New York, 1948,

Table V.’ Summary.of Maximum Corrosive Attack in Eack Major Region and Quadrant
" . of the Mark I Voletility Pilot Plant L Nickel Fluorinator

'

Locatlon » ' .-  Intergranular . " Total Rate LossesP
Elevation . L wall Penetration .. Total mils/ ,
(inches below ‘ . thickness Interior Exterior Corrosive  month mils/hr
slip-on . ' loss? wall wall - Attack (Molten (F, sparge
flange ) Quadrant Region (mils) (mils) (mils) (mils) salt time)€ time)d
12 North Vapor 38 2l 3 - 65 b1 Ll
12 East Vapor 29 20 I 53 33 0.9
12 South  Vapor - 33 23 3 - .59 37 1.0
12 West Vapor = L7 21 s 73 L6 1.2
35 - North Vapor-salt 6 18 7 31 20 © 0.5
) o ~ interface ' B i ’
35 East Vapor-salt b 17 6 27 17 O.h
‘ interface o « :
35 South Vapor-salt - 8 23 5 36 23 ) 0.6
‘ interface '
35 West Vapor-=salt 10 17 8 35 22 0.6
' ' ~interface )
L5 ' North Salt - - 10 25 3 38 ol 0.6
45 "~ East | Salt 8 21 4 33 21 0.5
L5 South  Salt 11 25 L 4O 25° 0.7
45 West Salt 15 22 5 b2 - 26 0.7

aBy micrometer measurement.
bIncludes exterior 1ntergranular penetratlon
®Based on molten salt residence time during VPP "M" (1-48) runs and "C" (1-15) runs.

dBa_sed on fluorine sparge time during fluorination of molten salts.

- 92 -
" nickelous fluoride, NiF

_27_

molten fluoride salts, generally of the NaF-ZthgUFu type. The compatibility
of each of these agents in contact with metals has received increased atten-
tion during the past decade, but a composite system has not been studied
extensi&ely.

Fluorine, a most active element, wés known to react with virtual-
ly every‘mefal under suitable conditions. Resistance to ffirther attack was
felt to be imparted by passive fluoride films which form on materials.lo’ll’lg

For nickel, the only known binary compound with fluorine was found to be
(ref 13)

2°

to be about 1000°C, well in excess of the operating temperatures (600-725°C)

The melting point of NiF,, had been reported

. _r-
of the fluorinaw:,or.l’+ The vapor pressure of NiF approx 1 x 10 ? mm Hg at

’
650°C, appeared suffiéiently low so that little iolatilization of the protec-
tive film would occur.lS

| RecentAexperiments at the Argonne National Laboratory had in-
dicated that the relative amounts of fluorine consumed by a nickel vessel and
the change in rate-law behavior with temperature can be represented as shown
ih Pig. 11. At'lower temperatures, 300 to 400°C, a logarithmic rate law ap-
peared to hold; but at higher temperatures, 500 to 600°C, a parabolic behavior
seemed prevalent. | ' |

10 : :
M. J. Steindler and R. C. Vogel, Corrosion of Materials in the Presence
of Fluorine at Elevated Temperatures, Argonne National Laboratory Report,

ANL-5662 (January, .1957).

llC' Slesser and S. R. Schram, Preparation, Properties, and Technology of
Fluorine and Organic Fluoro Compounds, National Nuclear Energy Series, Div. VII,
Vol. 1, pp. 157, 173, McGraw-Hill, New York, 1951.

125, 7. Barber and H. A. Bernhardt, K-1421 (April 9, l959)(classified).

3H J. Emeleus, Fluorine Chemistry (ed. by J. H. Slmons), l T,
Academic Press, Inc., New York, 1950.

luLaurence L. Qulll (ed.), Chemistry and Metallurgy of Miscellaneous
Materials: Thermodynamics, National Nuclear Energy Series, Div. IV, Vol. 19b,
p. 202, McGraw-Hill, New York, 1950.

15M Feber, R. T. Meyer, and J. L. Margrave, "The Vapor Pressure of Nickel
Fluoride," J. Phys. Chem. 62, 883 (1948).

M B

FLUORINE CONSUMPTION
(millimoles/sq cm x103)
o

- 28 -

UNCLASSIFIED"
ORNL-LR-DWG 55786

600°C

500°C

-——""_—i;?—

400°C

-300°C

0 1000 | 2000 3000
TIME (min)

Fig. 11. Consumption of Fluorine by a Nickel Vessel. Reference:
R. K. Steunenberg, L. Seiden, and H, E. Griffin, "The Reaction of Fluorine
with Nickel Surfaces," Argonne National Laboratory Chemical Engineering
Division Summary Report, July, August, September, 1958 ANL-5924, pp. 42-L43.

= BY =

1
More detailed work in this field has been reported by ORGDP. 6

Nickel was found to form a continuous, adherent fluoride film with an un-
diluted fluorine atmosphere at temperatures up to about 980°C. Electron
microscopy indicated that the nickel fluoride films had few flaws in the in-
dividual crystals which would permit direct access of the fluoride to the
nickel surface .underneath. The studies also indicated that, as the nickel
fluoride film thickened, more resistance to attack was obtained. However,
considerable intergranular attack of the metallic nickel was noted at tempera-
tures of approx 815 and 980°C (Fig. 12). The depth of intergranular penetra-
tion was estimated to be 5 to 8 times the average attack as calculated from
scale formation. The report indicated that the primary mechanism of attack
appeared to be diffusion along the crystal boundaries.

The second major corrosive agent placed in contact with the
Mark I fluorination vessel during operations was UF6' The effect of this

compound on metals had undergone investigation in connection with the design,

L7

construction, and operation of the ORGDP. Some recent work at the same site

indicated that on A nickel samples which were coated with nickel fluoride films
of 37 000 and 74 000 A, the average penetration of the nickel by UF6 at about

815°C, calculated from the average nickel fluoride scale formation, appeared
-to be about one third of that experienced with fluorine at about 700°C.(ref 16)
Later work by the same group indicated that at times one order of magnitude
greater (hundreds of hours versus tens of hours ) NiF2 is lost by vaporization

and/or a reaction process so that catastrophic attack can occur by additional
UF6 contact.18
The nickel exposed to the UF6 at those elevated temperatures ex-
hibited a grain-boundary attack beneath the fluoride scale quite similar to
16C. F. Hale, E. J. Barber, H. A. Bernhardt, and Karl E. Rapp, High
Temperature Corrosion Study, Interim Report for the Period November, 1958,
Through May, 1959, KL-498 (July 28, 1959).

17J. J. Katz and E. Rabinowitch, The Chemistry of Uranium, National Nuclear
Energy Series, Div. VIII, Vol. 5, pp. 44546, McGraw-Hill, New York, 1951.

laE. J. Barber, Technical Division, ORGDP, Jan. 7, 1960, Private

communication.

- 30 =

Fig. 12,

Exposure to Fp at 815°C for 4 hr and for 32 hr at Lower Temperatures; (c) After

Exposure to Fp at 980°C for 35 hr and for 350 hr at Lower Temperatures.
Etchant: Soldium cyanide, ammonium persulfate. 100X.

E. J. Barber, H. A. Bernhardt, and Karl E. Rapp, High Temperature Corrosion

Microstructures of A Nickel Coupons (a) As Received; (b) After

Reference: C. F. Hale,

Study, Interim Report for the Period, November, 19508, Through May, 1959, KL-498
(July 28, 1959).

- 31 -

that experienced in the exposures to fluorine. It was also observed that
both nickel fluoride and UF6 corrosion products had appreciable vapor pres-
sures at those test temperatures and were observed to migrate to the cooler
portions of the reactor by vapor-phase transfer,

In addition to F2
contact with the Mark I VPP fluorinator. At ORNL, considerable progress has

and UF6, fused-fluoride salts were also in

been made in determining material compatibility in various fused-fluoride-
salt systems through the Aircraft Reactor Experiment and Molten-Salt Reactor
19

Projects. Nickel-base alloys were found to be, in general, superior to other
commercial alloys for the containment of fluoride salt mixtures under dynamic

flow conditions.

2. Collective Attack During Fluorination

The fluorination cycle of the fluoride volatility process pro-

‘‘duced greater corrosive attack by the collective system of FE’ UF6, and fused

fluoride salts on nickel than had been reported for the individual constituents.
During VPP operations, the Mark I L-nickel fluorinator displayed maximum cor-

rosion rate losses of 1.2 mils/hr, based on F_, sparge time, or 46 mils/month,

2
based on molten-salt residence time during VPP "M" runs (1-48) and "C" runs

. (1-15). These rates include wall thickness or metal losses as determined by

dimensional analysis plus intergranular penetration as determined by metallo-
graphic examination. The rates are generally consistent with early bench-scale
work on the volatility process.

For convenience in repo}ting, the proposed reasons for the high
corrosive attack on the fluorination vessel will be discussed under four major
headings: (a) Interior Bulk Losses, (b) Interior Intergranular Attack,

(c) Exterior Intergramilar Attack, and (d) Grain-Size Variations.

_l9w. D. Manly et al., "Construction Materials for Molten-Salt Réactors,"
Flnid Piel Reactors, (ed. by Lane, MacPherson, and Maslan) Chap. 13, Addison-

Wesley, Reading, Mass., 1958.

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

May 20-25, 1957, TID-7534, pp. 571-572.

- 32 -

a., Interior Bulk Losses

- The first portion of this section described "conditioning"
treatments whereby nickel fluoride was induced on the walls of the Mark I
fluorinator prior to actual high-temperature operations. This was in harmony
with prevailing generalizations concerning\pdssive,fluoride films, which, when
formed on exposed metal surfaces may inhibit further attack by elemental
fluorine. However, based upon ORGDP and Argonne National'Lafioratory W'ork,l6’2l |
it appears that fiaésivation temperature should have been edual to, or greater
than, the operating temperature, rather than 20-150°C, in order to induce
greater film thicknesses. The work by Hale et al., indicated that while cor-
rosive attack on nickel'by fluorine did not cease after the nickel fluoride
film thickened,'additional resistance to attack was obtained at test tempera-
tures up to 980°C. Considering the very high rate losses found on samples
removed from the wall of the first VPP fluorinator, it would appear that
conditions We;e presenf’in the vessel which (1) did not allow sufficient film
tpickening to occur, (2) reduced'the protectivity of thickened nickel fluoride
~~f£ilms, or (3) permitted catastrophic losses of the nickel fluoride films.

“‘ Free fluorine was present, periodically, during all of the
‘VPP operations in quantities above those amounts necessary to oxidize ény
flFh in the fluoride salt mixtures to UF6. Thus, even though catastrophic
losses of o0ld nickel fluoride films might occur, new films would have oppor-
tunity tp form. Therefore, a continuous cycle of initial loss, reformation,
:and secdndary'loss of the nickel fluoride films forming on the walls of the
fluorinator is proposed as the method whereby the large losses of bulk metal
occurred.

This proposed loss cycle could be initiated and maintained
in several ways depending on the region of the fluorination vessel under con-

. sideration. In the salt region, fused fluoride salt baths could dissolve the

2lR. K. Steunenberg, L. Seiden, and H, E. Griffin, "The Reaction of
Fluorine with Nickel Surfaces," Argonne National Laboratory Chemical Engineering
Division, Summary Report, July, August, September, 1958, ANL-5924, pp. 42-L3.

- 33 -

nickel fluoride films until saturation of the baths with nickel fluoride
occurréd. It has been reported that NiFo is soluble in equimolar NaF--Zr‘Fl+
to the extent of 1.8 wt % at 700°¢c, (¥ef 22)

At the vapor-salt interface, similar dissolution could have
occurred to remove protective wall films. Also,. a washing action caused by the
fluorine sparge agitation and the rise andrfall of the bath level could aid
film removal.

In the vapor region, where maximum metal losses occurred,
nickel fluoride films that have limited plasticity at the lower operating
temperatures might be lost throfigh cracking or rupturing. Also, the dif-
ference in linear coefficients of thermal expansion of nickel and nickel
fluoride would be sufficient to exaggerate the spalling tendency. These
actions may also occur in the other two major regions of the fluorination
vessel, but the other loss mechanisms described for those areas would proba-
bly predominate.

Cathers has suggested an additional mode of film loss for
the vapor region, i.e., dissolution of nickel fluoride in very highly corrosive =

23

liduids which condense in the cooler zones of the fluorinator. Fluoride
compounds of intermediate volatility such as those containing high-valent
chromium, sulfur, titanium, and/or silicon, might be responsible. This con-
densate is pictured as being extremely variable in composition (possibly
including HF and water vapor, if they were unintentionally admitted into the
system) and in dissolution ability. After the condensate forms in a rela-
tively cold region of the vessel, it would run down the wall toward a higher
temperature zone, where it could then dissolve the nickel fluoride surface
films. Progressing further down the walls of the vessel, it would eventually
reach a temperature zone where it could reflux, and leave with the product

stream or re-enter into the vessel corrosion environment.

220. L. Whitmarsh, Uranium Recovery from Sodium-Zirconium Fluoride-Salt
Mixtures, Veolatility Pilot Plant Runs L-1 Through L-9, CF-59-9-2

(September 30, 1959).

23G. I. Cathers, ORNL Chem. Tech. Div., Private communication.
- 34 -

Support is given to the Cathers' suggestion in view of
the design used for the VPP fluorination vessel and the results of the chemi-
cal studies‘on the interior wall deposits from the vapor region of the vessel.
The Mark I fluorinator heating system was designed to produce cooler tempera-
tures in the upper regions of the vessel. The maximum temperature recorded
on a thermocouple attached to the exterior vessel wall 12 in. below the
slip-on flange was 500°C, while the average temperature in the same region
was 400°C. The thought wa% that a cooler top vessel zone would permit
deposition. of low volatility compounds that might entrap'uraniumlproducts.
After some indefinite time of deposition, it was felt that these layers
would gradually fall back into the salt baths and become available for fur-
ther fluorination. '

In support of the condensable corrosive liquid theory,
chemical analyses of residual wall deposité and millings from the fluori-
nator vessel walls in the middle vapor region (Figs. 4 and 6) showed high
concentrations of uranium and chromium. Thus, it would seem that entrain-
ment of uranium and redeposition of materials containing uranium did occur
éhd_that chromium was a part of the entrapping agent or agents.

A Rather than describing the agents which seem to have aided

- the cofrosion progress in the middle vapor region as condensable liquids
which dissolve nickel fluoride, favor is given to the idea that the nickel
fluoride complexed with other more volatile fluorides, the combination of
which have lower melting points than nickel fluoride. It has been reported
that nickel fluoride complexes easily and that many combinations are known.gu

A complex of NiF, and high valent chromium fluoride appears to be a parti-

cularly good posiibility.

Of interest in discussing the wall-thickness losses on the
Mark I fluorinator is the region slightly above the salt-vapor interface where
very low losses were found. In this same region, a ring of deposited material

2hH. J. Emeleus, Fluorine Chemistry (ed. by J. H. Simons), 1, T,
Academic Press, Inc., New York, 1950. -

_35_

was found after operations. Unfortunately, chemical analyses were not ob-
tained on the deposit. However, the fact that this thick ring was present,
probably continuously during operations, is believed to have prevented inter-
actions with the nickel fluoride "protective" surface films underneath and
thereby prevented high losses in that region. The reason for the formation
and continued presence of the ring is associated with the low temperatures

that predominated in this region. The location is between the furnace windings
and the tubular-type heating elements attached to the upper portion of the
vessel. Also, the presence of the furnace seal on the exterior ot the fluori-
nator at this same elevation probably affected temperature by acting as a

sink .for conducted heat in this region.

b. Interior Intergranular Attack

In the "as-polished" condition, interior surface samples from
the Mark I fluorinator showed no signrof intergranular corrosive attack." After
etching with a strong acid solution, a mixture of 0.5% HCl, approx 40% HNO ,,
and. approx 60% HC, H

were attacked more severely and rapidly than grain boundaries closer to the

302, the surface grain boundaries to depths of from 4—25 mils
| interior of the specimens. Using a mild etchant, KCN—(NHM)ESEOB’ and partial
repolishing, a pale yellow deposit was observed at the mating surfaces of the
-grains. The appearance of the intergranular attack on the fluorinator walls
was distinctly different than that produced in nickel by fluorine or UF6 (with
an initial nickel fluoride film) and reported by Hale et Ei.eB
| The formation of grain-boundary compounds or other changes in
grain-boundary regions, such as those observed, and which are generally catego-
rized as intergranular corrosion attack, often result in the loss or "sloughing"
of entire grains from the exposed surface of the material. However, metallo-

graphic examinations of specimens from the fluorinator revealed relatively

smooth surfaces.

220, F. Hale, E. J. Barber, H. A. Bernhardt, and Karl E. Rapp, High
Temperature Corrosion Study, Interim Report for the Period November, 1958,
Through May, 1959, KL-498 (July 28, 1959).

- 36 -

Some circumstantial évidence was available to indicate that
sulfur contamination produced the grain-boundary modifications observed. In
addition to previously reported work26 where microstructures of nickel speci-
mens purposely embrittled with sulfur were compared with early VPP L nickel
corrosion specimens, a communication from ORGDP, which supplied the VPP with

process fluorine, indicated that as much as 200 ppm of sulfur as-sulfuryl

27

fluoride had occasionally been detected in their manufactured fluorine.
Contamination from the commercial sodium fluoride used to strip hydrogen
fluoride and water from the.process fluorine could also be a source‘of sulfur.
Sulfuf could also be introduced into the system from impure feed salts or from
trace quantities contained in corrosion test specimens.

Attempts to prove the presence of sulfur in the fluorinator
wall Samples in quantities greater than that present in the base metal by
. using sulfur print papers, chemical analyses of surface millings and wall
deposits, bend tests, and various metallographic techniques either met with
failure or produced inconclusive results. This was particularly frustrating

in view of early reports that free Ni_S, had been identified in nickel micro-

32 29

; scopically when 0.05 and 0.005 wt % 'S, respectively, were present.28’ Exam-

ination of the nickel-sulfur constitution diagram indicated that‘Ni3S2 would
. be the compound to seek in identifying low percentages of sulfur contamination
30

in nickel. It was considered that chromium and/or other system contaminants

may have complexed the Ni and thus prevented identification by some of the

S
“3¥e
described methods. However, considering the extensive efforts employed on the-
fluorinator samples, if sulfur were the major agent responsible for the inter-

granular attack, more fiositive indications should have been found.

.261. R. ‘Trotter and E. E. Hoffman, Progress Report on Volatility Pilot

* Plant Corrosion Problems to April 21, 1957, ORNL-2L95 (September 30, 1958).
27
28 _ |
“G. Masing and L. Koch, Z. Metallkunde 19, 278-279 (1927).
29P. D. Merica and R. G. Waltenberg, Trans. Met. Soc. AIME 71 709716 (1925);
National Bureau of Standards Technical Paper 281, 155-182 (1925).

3OM. Hansen and K. Anderko, Constitution of Binary Alloys, p..l1l035,

McGraw-Hill, New York, 1958.

H. J. Culbeft,'Process Engineering Division, ORGDP, Private communication.

_37_

The nature of the grain-boundary deposits may not be deter-
mined until they are studied by an electron probe microanalyzer or some other
pfecise tool.  Nevertheless, because of the potential effeét on corrosion and
the cumulative and irreversible embrittling tendencies of the nickel-nickel
sulfide eutectic, reduction of sulfur in VPP nickel fluorinator's environ-

ments to the lowest possible levels is essential,

c., Bxterior Intergranular Attack

~

As described, the exterior of the Mark I fluorinator shell
also suffered intergranular attack, but to a lesser degree than the interior.
Penctration on the exlerior surfages varied from 3-8 mils. The appearance
of this attack at high magnification, the exterior environment, énd the

presence of NiO, found on the second VPP fluorinator operated under similar

2
conditions, indicate typical high-temperature intergranular oxidation of

nickel.

d, Grain-Size Variations

,MEtallographic examination of samples removed from the shell
- of the Mark I fluorinator disclosed widely variant grain sizes. ©OSmall, uniform
grains of 5-6 average ASTM grain-size number prevailed in the salt and salt-
vapor interface regions, while much larger grains were observed in most of the
vapor region., ‘The largest grains found in the vapor region were at an eleva-
tion of approx 12 in. below the bottom of 'the slip-on flange. The average
ASTM grain-size number at the 12 in. elevation was greater than 1. Diffractom-
eter traces were obtained on vessel samples removed from the 1-, 12-1/2-, and
43-in. levels in order to compare the residual strain remaining in the samples.
The results indicated approximately no recrystallization had occurred at the
1-in. level, only a very limited amount of recrystallization took place at
the 12-1/2—in. level, and partial, if not complete, recrystallization occurred
in the salt-phase sample. . i

The classical factors which determine crystallite growth upon
heating are afiount of prior cold deformation, annealing temperature, and

annealing time. Generally, larger amounts of cold deformation provide for
.38 -

smaller resultant grain sizes after isothermal and constant time annealing.
Higner temneratures and- to a lesser degree, longer times result in larger
graln 51zes on material containing some fixed amount of cold work.

In meany materials a small amount of prior cold deformation
results in the production of exaggerated grains after annealing at ordinary
times and temperatures. In these cases, the term "critical strain" has been
given to the quantity of cold work originally introducedu The mechanisms
involved here have been studied and reported in detail.31 ~

‘ Thé VPP Mark I fluorinator was fabricated from L nickel by
converting a flat, annealed, i/h—in. plate into a lh-in.-diam right cylinder
by roll forming. During forming, cold deformation was induced in the outer-
most fibers of the shell to a calculated_maximum of approx 2%. This amount
‘of cold work is at the right level tO be termed "critical strain" for most
.metals:and‘presumafiiy was present along the entire length of the fluorinator
shell. However, exaggerated grain sizes were observed only in a portion of
the vapor phase of the vessel after pilot plant operatlons 'Therefore an
,addltlonal variable, rate of heatlng after prior deformatlon, has been sug-
gested as' the most influential factor in producing the grain sizes found in
the Mark I fluorinator.32

- The first time the fluorinator was heated was during pre-
llmlnary fluorination. equlpment studles. During those cycles, the lower

23 in. of the vessel was surrounded by a high heat-flux 30-kw’ electrlc— _
resistance furnace which raised the temperature of the enclosed portion of
the vessel to approx '450°C. No external heat or insulation was in ciose
prox1m1ty with the upper portion of the fluorlnator. Later cycles during
the same prellmlnary studles ‘were done w1th a glass-lined heating mantle
coverlng the top portlons of the fluorinator. Temperatures in the lower re-
gions of ‘the vessel reached 600-700°C, while temperatures in the upper regions

~

- of the vessel were qmuch lower.

31J E. Burke, "The Fundamentals of Recrystallization and Grain Growth,"
Grain Control in Industrial Metallurgy, Amerlcan Society for Metals, Cleveland
Ohlo, 1949, . : I

32L -K. Jetter and €. J. McHargue, ORNL Metallurgy Division, Private
communication, - ~

-
- 39 -

In view of the initial Mark I fleating cycles, and confirmed
by the diffractometer traces, the salt region of the fluorinator appears to
have experienced thermal levels where recovery and at least partial recrystal-
lization occurred. The recrystallization temperature for nickel varies from
approx 600°C for A nickel, 99.4 wt % (Ni + Co) to 370-470°C for various
purities of electrolytic nickel, 99.9" wt % (Ni + Co) after 50% reduction by
cold rolling and anncaling for 1 hr at lhe temperature indicated.33 Jetter
'and McHargue have suggested that during the initial heating the rapid rate
of heating did not allow complete relief of internal stresses which at higher
temperatures served as nucleation eitec for recrystallization. The rate of
.. .appearance of these nuclei is known to increase with time, exponentially with
temperature, and with increasing amounts of prior cold deformation.3l After
nucleation, stable nucleus growth occurred and the presence of so many grains
growing in a fixed volume limited the salt region grain size. Later heat
"cycling of the fluorinator was done at approkimately the same temperapure
levels as the highest original heating (600-725°C) so that probably little, if
any, grain-boundary migration occurred after reérystallization to increase the
grain size in the lower portions of the vessel.

The top half of the Mark I fluorinator sustained lower tempera-
tures during the first heat cycles as the result of heat losses by radiation
| and convection frém the upper portions of the vessel wall., A steady state
temperature of < 200°C at the 12-in. level has been estimated.3u The upper
portions of the vessel also experienced éluggish heating rates when compared
to that portion of the vessel enclosed by the heating furnace. It has been
suggested that the latter influenced the grain size in the vapor region by
permitting more complete recovery and resulting in few nucleation sites.

, During later operations in the pilot plant, rod-type electric

heating elements.with a total rating of 9 kw were in contact with the vapor

33g. M. Wise and R. H. Shafer, "The Properties of Pure Nickel - I, II, III,"
Metals and Alloys, 16 Sept., p. 424, Nov., p. 891, Dec., p. 1067, 1942,
3k o

S. H. Stainker, ORNL Chemical Technology Division, Private communication,
- 4o -

region exterior surfaces of the fluorinator. A thermooouple at approximately
,the 12-in. level indicated average temperatures of 400°C and a maximum of
.500 C. Such thermal levels permitted only, a small amount of recrystallization
to occur, as indicated by the diffractometer studies. An extremely coarse
grain size resulted from the described conditions. .
o ~ Examination of samples removed from the Mark I fluorinator
revealed that the metal-loss profile and grain-size profile closely paral-
leled one another except for the region Just below the top flange of the
vessel. Because of the evidence of corrosion by intergranular attack_during
exposure to ohevvolatility fluorihation'environment, arcausal relationship
is suggested. Thus, loss of,a large grain by effectively losing grain-boundary
material until the grain'slogghs off would mean greater losses than for the
same.proéedure happening in finngrain regions. However, sloughing of grains
was rotAObserved and only a'slight widening of the boundaries was noted at
the.junction of the boundaries'and the exposed metal -surfaces. Also, the
depth.of intergranular penetration in coarse-grain regions was ébout,the same
as in most of the fine-grain regions. The conclusion is that mefal loss and
- .grain size do not seem to be interrelated in -the corrosion of the Mark T
fluorinator. - |
& ‘However, for two reasons, the authors recommend using fine-
grained material for fluorination vessels. One, a higher strength level will
be achieved in nickel, a material not noted for superior strength; and two,
*’ one variable will be removed-in a system replete with variables. Fine—grained
material can be provided by modification of ihe specifications of purchased

stock and by revising the'annealing cycle after fabrication.35

II. Mark IT Volatility.Pilot Plant L Nickel Fluorinator

A, Material Selection and Fabrication-Design Changes

While L nickel seemed deficient in its corrosion resistance for long-

time service, no other commercially available material had, at that time, proved

35"Annealing of Nickel, Monel, and Inconel," Tech. -Bull. T-20, The
International- Nickel Company, Inc., New York, April, 1953

S -1y -

itself on the basis of scouting or other tests to be worthy of immediate sub-
stitution as a fluorinator material of construction. Moreover, the heavy
vapor-phase attack on the Mark I fluorinator was still under investigation,
so that the use of the same construction material seemed especially appropriate
to ascertain more about this corrosion problem. Consequently, the Mark IT
fluorinator was fabricated in the ORNL shops from the same heat of 1/4-in.-
thick L nickel as was used in the previous pilot plant vessel. Analogous
fabrication techniques were used. Figure 13 shows the second VPP fluorinator,
the vessel furnace, and the complexible radioactive products (CRP) trap.
Certain design changes were incorporated in the second pilot plant
fluorinator. In order to provide for remote handling and to allow desirable
upward flow through the CRP trap, its location was changed from the original
side position shown in Fig. 1 to an elevation completely above the Mark IT
fluorinator. Other process design changes included some modification of the
draft tube assembly, repositioning of the off-gas line, and installation of a

small port in the top blind flange of the fluorinator for observation and entry.3

B. Operational History

The Mark II fluorinator was placed in service at the beginning of
the "E" runs during which time fuel used in the Aircraft Reactor Experiment
was reprocessed.37 The use of the vessel was continued during the "L" (spiked)
runs and during the gas-entrainment studies, M-62 through M-64. The process
history, summarized in Table VI, has been divided into three phases for con-
venience in reporting the Vidigage inspections on the vessel at the completion
of Runs L-4 and L-9.

The "L" or spiked runs were made in the VPP to extend process develop-

36

ment data and used a nonirradiated salt with fully enriched uranium. Some

36F. W. Miles and W. H. Carr, Engineering Evaluation of Volatility Pilot
Plant Equipment, CF-60-7-65, pp. 43-45.

3y, u. Carr, Volatility Processing of the ARE Fuel, CF-58-11-60 (November
14, 1958).

380. L. Whitmarsh, Uranium Recovery from Sodium Zirconium Fluoride Salt
Mixtures, Volatility Pilot Plant Runs L-1 Through L-9, CF-59-9-2 (September
30, 1959).

- 42

Unclassified
ORNL Photo 45554

Py
-, 3
» J %

-

Fluorinator l:_ fl

(L Nickel)

»

Fig. 13. Mark IT VPP Fluorinator and CRP Trap.
Table VI. Summary of Process Conditions for Mark II Volatility Pilot Plant Fluorinator
VPP "E," "L," and "M" (62—64) Runs
Phase I Phace IT Phase IIT
Runs E-1 Through E-6; Rune L-5 Rurns M-562
L-1 Through L-4 Throigh L-9 Through M-6L4 Total
Temperature, °C 540-730 550—700 600—690 540-730
Thermal cycles 10 L 17 31
Time of expcsure at temper- ~ 900 ~ 310 ~ T25 ~ 1935
ature, salts molten-hr )
Fused salt, E1-E6 NaF-ZrFu-UFu(C’d) NaP-ZrFu-UFu(g NaF-ZrFu
nominal mole % (4L8-49.5-2.5) (52-L45.5-2.5) (50-50)
L1-Lk4 NaF—ZrFu-UFu(e’ )
(54—L1-5)

Conditioning fluorine " 4535 in 18 hr 756 in 3 hr none 5285 in 21 hr
input, standard liters (2.8-6.9 standard (2.57 standard (2.5~7 standard

liters/min) liters/min) liters/min)
Cperating fluorine input, 44 470 in 74 hr 16 000 in 18 hr none 60 470 in 92 hr
standard litersP (1.6-20.2 standard (1.6-20.2 standard

liters/min) liters/min)
UF6 exposure, hr ~ 25 ~ 10 none ~ 135

@These operations were done at 20-150°C for the purpose of inducing an initial "protective" film of

nickel fluoride on the walls of the fluorinator.

ban average of approx 3:1 excess Fp over that quantity necessary for theoretical requirements was
used in order to reduce the final uranium concentration in the salt to a few parts per million.

CUranium was fully enriched. Feed salts contained the following ranges of impurities: Component:
0.031-0.056 wt % Cr, 0.017—0.138 wt % Ni, 0.033-0.078 wt % Fe, 0.0015-0.080 wt % Si, < 0.001- < 0.005 wt % Mo.
d¢, L. Whitmarsh, Reprocessing of ARE Fuel, Volatility Pilot Plant,Runs E-1 and E-2, CF-59-5-108,

-E"T—

(May, 1959) and Reprocessing of ARE Fuel, Volatility Pilot Plant, Runs E-3 end E-6, CF-59-8-73 (August, 1959).

©Ten milligrams of PuFL added in Run L-3-4. Feed salts contained the following ranges of impurities:

Component: 0,0183-0.0345 wt % Cr, 0.035-0,236 wt % Ni, 0.0133-0.030 wt % Fe, < 0.001-0.005 wt % Si,

< 0.001- < ©.0052 wt % Mo, and ~ 0.010 wt % Ti for L-1,-3, and -5 only.

£, L Whitmarsh, Uranium Recovery from Sodium-Zirconium Fluoride Salt Mixtures, Volatility Pilot

Flant, Runs L-1 Through L-9, CF-59-9-2 {September 30, 1959).

8Uranium was fully enriched. One gram of PuF), added in Run L-6; 10 grems of PuF), added in Run L-8.

Same feed salt impurities and reference as detailed in (d) above,
s’ b

of the runs were spiked with high burnup salt, thus the name of this series.
During three of the "L" runs, selective additions of PuFu were added to the
feed salt so as to gather information on the behavior of plutonium in the

39

volatility process. Previous experiments had indicated that some volatili-
zation of PuF6 would occur. However, greater than 99% of the plutonium
remained in the salt in the fluorinator during the "L" runs.

Several months after the "L" runs were completed, the Mark II fluori-
nator was used to study gas-phase entrainment problems in the VPP.LLO These
runs, M-62 through M-6k4, investigated ZrF) solid condensation in the vapor
phase. Barren equimolar NaF-ZrFu salt was used in these studies and fluorine

was not present.

C. Reaction to Environment

l. Visual and Vidigage Inspections

During the first portion of the "L" runs, 1 through 4, persistent
line plugging occurred. Presumably this was due to nickel contamination in
the feed salts plus corrosion losses which allowed the concentration of nickel
fluoride in the salt to exceed the solubility limit. After run L-4, the plant
was shut down and approx 200 1b of a high-melting salt complex was found in
the bottom of the Mark II fluorinator. See Fig. 1lbk. Petrographic and x-ray
diffraction analyses showed the mass to contain 20-30 wt % NiF, and

2
50-60 wt % NaF-NiF —QZrFu. In order to continue using the vessel, the frozen

2
salt was chipped out with a pneumatic hammer.

Results of a visual inspection of the Mark II VPP fluorinator
after exposure to "E" and L-1 through L-4 runs and cleanup operations are
gclited:

Vessel walls Free of salt on the interior. Exterior surfaces were
covered with a dull gray film. Those portions of the
wall near the girth weld had scars on the interior sur-

faces presumably caused by glancing blows of the pneumatic-
hammer blade.

39R. A. Cross and C. L. Whitmarsh, Plutonium Behavior in the Fluoride
Volatility Process, CF-59-9-5.

AhOJ. B. Ruch, Volatility: Fluorinator Design FV-100, Zr-U Fuel Element

Processing Phase, CF-59-5-89.

Unclassified
QRNL Photo_hh004

. Waste @8
\Salt Line'

Fluorine &g
f9et ¢ Inlet Linefsii 2. -
[ W@ “d ..x(v ¢ .‘ .l ¢ < ): 4 ‘:. \

2
L 4

Fig. 14. 1Interior of Mark II VPP Fluorinator After Run L-4. (Fluorine
inlet cut off prior to taking of photograph.)
_ 45 -

Longitudinal Interior surfaces were sharply defined by corrosive attack.
seam weld Corrosion of the weld metal (INCO 61) appeared to have been
as severe or more severe than that of the wall (L nickel).

Girth weld Interior surfaces were sharply defined by corrosive attack.
Corrosion of the weld metal (INCO 61) appeared to have been
as severe or more severe than that on the wall or the dished
head.

Dished head Interior surfaces exhibited many scars on sides and bottom
of the head, presumably caused by the pneumatic-hammer blade.
Exterior surfaces were bulged approx l/l6 in. in about six
places corresponding to the deeper internal scars.
Vidigage (ultrasonic thickness) measurements were made to assess the damages
during salt removal plus the wall-thickness losses that had occurred during
previous process cycling. A maximum wall-thickness loss of 68 mils was found
in the salt-containing region of the fluorinator. Complete results of the
Vidigage examination are shown in Table VII along with other data.

Despite the high-corrosion losses indicated, a decision was made
to use the vessel for an additional five runs so as to complete the scheduled
process development studies. No plugging difficulties were encountered during
the subsequent "L" runs 5 through 9. This was accomplished by diluting the
feed salt with an equal amount of nickel-free barren salt.

Figure 15 shows the interior of the Mark II fluorinator after
run L-9. The results of a visual inspection after hand chipping and decon-
tamination were as follows:

Vessel walls Free of salt on the interior. Vapor region appeared rough
and pitted and had an adherent green deposit in the region
1020 in. below the slip-on flange. Exterior surfaces were

covered with a dull gray film. The scars caused by the
pneumatic-hammer blows after Run L-U4 were still present.

Longitudinal Interior surfaces were more sharply defined by corrosive at-
seam weld tack than previously noted. Corrosion of the weld metal
appeared to have been more severe than that of the base
metal.
Girth weld Interior surfaces were more sharply defined by corrosive at-

tack than previously noted. Corrosion of the weld metal
appeared more severe than that of the base metal.

Dished head Interior and exterior scars from previous chipping operations
were still present.
Table VII. Comparison of Vidigage Thickness Readings, in Inches, on the Volatility Pilot
Plant Mark IT L Nickel Fluorinator®

S.W. Quadrant N.W. Quadrant N.E. Quadrant S.E. Quadrant
Elevation Area A Area B Area C Aree D Area E
(in. below Reading Reading Reading Reading Feading Reading Reading Reading Reading Reading
slip-on after after after after after after after after after after
flange) Region Run L-9 Run L-4 Run L-¢ Run L-4 ERun -9 Run L-4 Run L-9 Run L-4 Run L-9 Run L-k

X 0.248 0.248 C.248 0.248 0.250 0.248 0.250 0.2.8 - -

3 A 0.249 - - 0.247 - - - 0.250 o =

5 0.250 - - - - 0.250 - o - -

T 0,248 0.24¢ - - 0.250 - 0.248 - - -

9 0.248 - 0.248 0,247 - 0.249 - 0.250 = -
11 5 0.24k 0.243 - 0.242 - 0.248 - 0.246 - -
13 9 0.238 0.232 - - - 0.241 - 0.238 - -
15 9 - 0,230 0.234 0,230 0.234 0.235 - 0.234 = -
17 0,230 0,229 0.234 0,228 - 0.234 - 0.236 - -
19 0.226 - - - - 0.232 - 0.238 - -
21 0.230 0.234 - 0.234 - 0.236 C.238 0.236 - -
23 0,240 - - 0.238 - - - 0.240 - -
25 " 0.236 0,23L 0.232 0.230 - 0.240 - - i -
27 H_$ 9 0.230 - - - 0.230 0.230 - - = -
29 o+ & 0.234 0.240 0.224 0.222 - 0.230 0.236 0.240 - -
31 s a8 ! 0.226 - 0.224 0.224 0.221 0.218 0,228 0.230 0.230 0,229
33 i'fi 0.210 - 0.207 0.206 0,210 0.204 - 0.226 0.222 0,220
35 y e - 0.236 0.196 0.236 0.198 0,196 0,202 0,210 0,210 0.210
37 % 0.198 0,230 0.196 0.230 0.192 0,212 0,188 0.205 0.195 0.194
39 l 0,194 0.227 0.198 0.234 0.189 0.215 0.193 0.194 0.192 0,187
41 B 0.194 0.22k4 0.204 0.242 0.187 0.210 0.1869 0,192 0,192 0.187
43 [ 0,206 0.227 0,206 0,248 0.184 0,202 0.166 0.184 0.190 0.186
45 [ 0.210 0.242 0.208 - 0.188 0.204 0,186 0.184+ 0,200 0,182
L7 \Y% 0.238 0.246 0.240 0.252 0.192 0,204 0.198 0.194% 0.196 0.196

aAccuracy of readings estimated to be * 0,003 in,

_L-W_
4 TUnclassified
ORNL Photo LL802

_ 48 -

9

TI VPP Fluorinator After Run L-

Interior of the Mark

Mg, 19.
- 49 -

Decontamination was done by alternate washings using a mixture of 0.7.M'H202,
1.§ M KOH, and O.4 M Na,C)H O at room temperature and 4 M (NF),),C,0), at 100°C.
At that time Vidigage readings were obtained and are reported in Table VIT,
along with the previous readings obatined after Run L-4. No significant ad-
ditional wall-thickness losses were found. _

The corrosion conditions established during the gas-entrainment
studies, Runs M-62 through M-64, which were done at a later date uéing the
Mark II fluorinator, were not felt to be sufficiently serious to warrant

another complete Vidigage inspection.

2. Chemistry

Samples of the dull gray film present on the exterior wall of
the Mark II fluorinator after Runs L-4 and L-9 were analyzed by x-ray diffrac--
tion techniques to be NiOE. In addifiion, samples of the adherent green
deposit which remained on the interior wall of the vessel in the vapor region
NiO, and

were submiltted for x-ray diffraction analyses. The presence of NiFg,

ZrO2 was noted,

Surface and subsurface millings were taken from the interior wall
of the vessel in the region where the green deposit seemed thickest, 1517 in.
below the bottom of the slip-on flange, and submitted for additional chemiéal
analyses. For comparison purposes, subsurface millings were also taken from
the exterior vessel wall at the same elevation. The results of the metal
chemistry are shown in Fig. 16, and indicate an increase in chromium con-
centration in the millings near the surface of the interior wall of the ves-
sel. No significant increase in sulfur in the vapor region of the fluorinator
over. that quantity present in the base material was found. However, on
specimens removed from the salt region of the Mark IT fluorinator, which were
later analyzed by Battelle Memorial Institute (BMI), an increase in the sulfur

content of the interior wall specimens was noted. Figure 16 also shows these

results.
SLIP-ON”
FLANGE

55in.

29in.

Y

AVERAGE
OF

'VAPOR-SALT
INTERFACE

LEVELS

42 IN. BELOW SLIP-ON FLANGE

w

15-17 IN. BELOW SLIP-ON FLANGE

\

UNCLASSIFIED

ORNL-LR—-DWG 49159

Center of wall cross-section

) ) Cr Fe . Mn ‘Na S
Sample Location Awt %) (wt %) (wt %)  (ppm) (ppm)
Exterior wall sub-surface millings 0.0150 0.37 0.26 65 40
at "’/2-5 mils below surface '
Interior wall surface millings ot 0.0785 0.64 0.29 155 - 20
o/5 mils below surface
Interior wall sub-surface millings  0.0880 0.40 0.26 110 45
at lo/20 mils below surface ‘ .

. Interior wall sub-surface millings  0.0120 0.29 0.26 65 40
at .20/35 mils below surface .
Exterior wall surface millingslat 120

°/s mils below surface
Interior wall surface millings ot 240
°/s mils below surface
Interior wall sub-surface millings 160
ot ‘f‘/]o mils below surface
120

Fig. 16. Analyses of Millings -from Mark Ii VPP L Nickel Fluorinator After Run L-9
and Vessel Decontamination. ' '

-

\

ORNL ANALYSES

BMI ANALYSES

- o6 -
- 51 -

3. Dimensional Analysis

After decontamination and other cleanup operations, s&mfiles
were trepanned from the Mark II fluorinator and sent to BMI for dimensional
analysis and metallographic examination. The location of these coupons is
displayed in Fig. 17. 1In addition, a full-length section from D-E area,
southeast quadrant, was removed and micrometer measurements taken at ORNL to
plot a complete cérrosion profile of wall-thickness loss. This plot is shown
in Fig. 18 and incorporates BMI micrometer measurements on the trepanned
samples from the other quadrants of the fluorinator. For comparison, Vidi-
gage readings of the D-E area after runs L-4 and L-9 are also plotted in
Fig. 18. The maximum wall-thickness losses found were in the salt region
of Mark II fluorinator. In a region 4O-43 in. below the bottom of the slip-
on flange of the vessel, a wall-fihickness loss of 82 mils was recorded. Fair
correlationvbetwéén the Vidigage readings and the miqrometer measurements was

noted.

4. Metallographic Study

A metallographic study was made on the trepanned samples from the
Mark II fluorinator by personnel.of the BMI Corrosion Research Division.LFl At
low magnification, 20X, the surfaces from the top vapor (T) sections showed a
light etch =nd some outline of grain boundaries. The middle vapor region
specimens (la and 1b) were etched to a greater extent and were covered with a
thin green crystalline deposit. Previous chemical analyses of this deposit

at ORNL indicated that the scale contained NiF,, NiO, and ZrO The vapor- .

’ .
salt interface (2a and 2b) and salt region'speiimens.(3a-and gb) seemed to
héve_been severely etched, and, in the latter, areas of intergranu%ar attack
could be seen, . |

Typical photomicrographs of sections from the Mark TIT fluorinator

can be seen in Figs. 19 through 23. Intergranular attack was prevalent in all

ulLetter Report from F. W. Fink, Battelle Memorial Institute, to
R. P. Milford, ORNL Subcontract No. 988 (October 7, 1959).

-

t
- 52 -

- ) UNCLASSIFIED
ORNL-LR-DWG 49160

. CELL NORTH ,

inet=3in:

.

15 in, —————m="
‘-3.

19 in.

-=-VAPOR

29 in,

34 in.

42 in.

47 in.

M INTERFACE

N\, [
R A

INY | =~=-LIQUID

1TSS raat

e O O L N O OO ORI
7/
7
Z

Fig. 17. Location of Specimens Removed from the Mark II VPP
L Nickel Fluorinator for Dimensional Analysis and Metallographic Study.
SLIP-ON FLANGE

UNCLASSIFIED
ORNL-LR—-DWG 49161

| | | |
(200-500°C IN THIS REGION:

* ORNL MICROMETER READINGS ON FULL

Z
Ay,

2
‘////’V//////'

VAPOR PHASE —

s VIDIGAGE READINGS AFTER L-4 AREA D-E |

H,
v

o VIDIGAGE READINGS AFTER L-9 AREA D-E

[ I l 1 | | l

LENGTH SECTION FROM AREA D-E, —
SOUTHEAST QUADRANT

| | /R

‘_ BMI MICROMETER READINGS
ON TREPANNED SPECIMENS

SHELL

,,,,

,,,,,,,

,,,,,,,
””””

DISTANCE DOWN FROM SLIP-ON FLANGE (in.)

GIRTH WE_D
/

|

SALT PHASE "

FLANGED AND

DISHED HEAD 1
i

10 15 20 25 30

35 40 45 50 55 60 65 70

WALL THICKNESS LOSS (mils)

Corrosion Profile of Mark II VPP L Nickel Fluorinator.

_gg_
- 54 -

Unclassified
BMI CO1h

Fig. 19. Microstructure of Sample frcm Interior Surface of Mark
VPP Fluorinator 3 in. Below Slip-on Flange (Vapor Phase, Area A)
Etchant: Nitric-acetic acid. 100X.
- 55 =

Unclassified
BMI C615

/ ; / ... "“
R - . -
- / > -
- - L > \x
) ' v -
~ o = - >
’ <
- % s
) - -
-
- QR Y
e i - e
. -
. - . dw
- e T ® ...
B % »
g - .
L : e
8 o . N i PR
| “» \ 4 % . .
w PR ‘@
. : a8
z ® *

Fig. 20. Microstructure of Sample from Interior Surface of Mark II
VPP Fluorinator 19 in. Below Slip-on Flange (Vapor Phase, Area A). Etchant:
Nitric-acetic acid. 100X.
- 56 -

Unclassified

Fig. 21. Microstructure of Sample from Interior Surface of Mark II
VPP Fluorinator 34 in. Below Slip-on Flange (Vapor-Salt Interface, Area A).
Etchant: Nitric-acetic acid. 100X.
- 57 =

Unclassified

Fig. 22. Microstructure of Sample from Interior Surface of Mark II
VPP Fluorinator 42 in. Below Slip-on Flange (Salt Phase, Area )
Etchant: Nitric-acetic acid. 100X.
- 58 -

Unclassified
BMI C617

8V

Fig. 23. Microstructure of Sample from Interior Surface of Mark II
VPP Fluorinator Approx 47 in. Below Slip-on Flange (Salt Phase, Area A)
Vicinity of Girth Shell to Bottom Head Weld. Etchant: Nitric-acetic acid.
100 X.
- 59 -

N sections in all regions but penetration was deepest in the samples from the
- interface and salt regions. Maximum grain-boundary attack was found to be
33 mils in the salt-phase sample A-3b which was in the vicinity of the shell

. to head girth weld.
Some slight variation in grain sizes was found in the metallo-
- graphic samples. The largest grains were in the salt region of the fluori-
nator. Table VIII summarizes the grain-size data in terms of average ASTM

grain-size number.

Table VIII. Summary of Grain Sizes in Samples Removed from the
Mark ITI Volatility Pilot Plant L Nickel Fluorinator®

Location
(inches down from
bottom of slip-on ASTM grain-
. flange) Region size number
3 Vapor 3-5
- 15 Vapor 34
19 Vapor 3L
29 Vapor-salt interface 3=5
34 Vapor-salt interface 3L
L2 Salt o)
L7 Salt o

aGrain sizes from interior wall and exterior wall samples
were approximately equal.

A pitting attack was found on the interior wall near Area C in
the vapor region, about 3 in. down from the bottom of the slip-on fiange.
Figure 24 shows the appearance of the surface at 3X, and Fig.25 shows a metal-
lographic section through this pitted area. The latter indicates a level of
intergranular penetration similar to that found in other upper-vapor regions.
The depth of intergranular attack did not appear to vary from top to bottom of
the pits.
- 60 -

Unclassified
BMI N59673

Fig. 24. Photograph of Sample from Interior Surface of Mark IT VPP
Fluorinator 3 in. Below Slip-on Flange (Vapor Phase, Area C) Showing
Pitting-Type Attack. 3X. '
= Bl =

Unclassified
BMI CA28

Fig. 25. Microstructure of Sample from Interior Surface of Mark II
VPP Fluorinator 3 in. Below Slip-on Flange (Vapor Phase, Area C) in Region
of Pitting-Type Attack. Etchant: Nitric-acetic acid. 100X.
- B -

In the salt region of Area C, a few small intergranular cracks
were found on the interior of the vessel wall. A typical crack is shown in
the unetched condition in Fig. 26 and in the etched condition in Fig. 22.
Very pronounced darkening of the grain boundaries to approximately the same
depth as the crack was evident after etching.

Cross sections from the fluorinator shell longitudinal weld are
shown in Fig. 27 (Samples B-T and A-3b) which depicts sections in the vapor
and salt phases, respectively. The weld joint from the salt region shows
what appears to be an increased attack at the weld root. This confirms the
visual examinations of the Mark II fluorinator. Examination of the weld from
the salt section at high magnification revealed pronounced grain-boundary
darkening after etching to a maximum depth of 10 mils. Similar, deeper grain-
boundary darkening occurred in the base metal adjacent to the weld metal,
extending to a maximum depth of 33 mils.

The exterior of the vessel displayed a corrosive attack which
appeared to be intergranular in nature. Penetration varied from 1-6 mils, the
maximum occurring at or below the vapor-salt interface (Fig. 28, Sample A-3b).
However, there was one exception to this pattern. A grain-boundary penetration
of approx 12 mils in depth was found on the exterior surface opposite the

pitted region in the vapor near Area C.

5. Summary of Corrosive Attack

Table IX summarizes the corrosion losses of all types found in
the three major regions of the Mark II VPP L nickel fluorinator. The maximum
attack was calculated to be 60 mils/month based on exposure to molten salts
during the VPP "E" runs (1-6) and "L" runs (1-9) or 1.1 mils/hr based on
fluorine sparge time during fluorination of molten salts. The maximum attack

occurred in the salt region.

D. Discussion of Results

The Mark II VPP L nickel fluorinator displayed a maximum corrosion

attack during the described VPP runs of 1.1 mils/hr, based on F, sparge time

2
during fluorination, or 60 mils/month, based on molten-salt residence time
- 63 =

Unclassified
BMI C668

Fig. 26. Photomicrograph of Sample from Interior Surface of Mark II VPP
Fluorinator 42 in. Below Slip-on Flange (Salt Phase, Area C) Showing Crack
Easily Visible Before Sectioning. As-polished. 100X.
- B =

Unclassified

BMI N596T74

Fig. 27. Photomacrographs of Longitudinal Weld Sections Through Wall of
Mark IT VPP Fluorinator (a) From Vapor Phase 3 in. Below Slip-on Flange, (b) From
Salt Phase 47 in. Below Slip-on Flange. Etchant: Nitric-acetic acid. 5X.
Unclassified
BMI €618

}
P

Fig. 28. Microstructure of Sample from Exterior Surface of Mark II
VPP Fluorinator Approx 47 in. Below Slip-on Flange (Salt Phase, Area A)
Showing Result of Air Oxidation. Etchant: Nitric-acetic acid. 100X.
Table IX. Summary of Corrosive Attack in Each Major Region of the Mark IT

Volatility Pilot Plant L Nickel Fluorinator

Total losses converted

Location Wall Intergranular : : .o b
Elevation o Penetration Total to mils/unit time
(inches below . Interior Exterior Corrosive mils/monthC mils/hrd
slip-on lass wall wall attack (molten salt (F2 sparge
flange) Quadrant Region (mils) (mils)  (mils) (mils) time) time)
3 S.W. Vapor L 3 1 8 5 0.1
15 S.W. Vapor 22 7 L 33 20 0.4
19 S.W. Vapor 25 8 3 36 22 0.4
29 S.W. Vapor-salt 20 8 2 30 18 0.3
interface
29 N.W. Vapor-salt 28 9 2 39 23 0.4
interface
34 S.W. Vapor-salt 53 16 6 75 L5 0.8
interface
) S.W. Salt 59 1k L T7 L6 0.8
Lo N.W Salt 52 16 1 69 b1 0.7%
L2 N.E. Salt &e 16 2 100 60 1:1
L7 S.W. Salt L5 33 6 8L 50 0.9
L7 N.W. Salt 52 16 1 69 b1 Q.75
L7 N.E. Salt 6L 23 3 90 54 1.0

a s
By micrometer measurement.

bIncludes exterior intergranular penetration.

“Based on molten salt residence time during VPP E(1-6) runs and L(1-9) runs.

dBased on fluorine sparge Time during fluorination of molten salts.

_99—
- 67 -

during VPP "E" runs (1-6) and "L" runs (1-9). Maximum attack occurred in the
salt-phase region of this vessel whereas the first VPP fluorinator experienced
maximum losses in the vapor-phase region. The rates of attack for both vessels
were of the same orders of magnitude. The corrosive attack in the Mark IT
fluorinator, as was the case for the' Mark I vessel, can be categorized into
bulk metal losses from the interior wall of the vessel and intergranular at-

tack on both the interior and exterior walls of the unit.

1. Interior Bulk Losses

Bulk metal losses are believed to be the result of continuous loss
and reformation of "protective" I\TiF2 films on the interior wall of the Mark II
fluorinator. As discussed in Section I, the nickel fluoride films formed on
the walls of the volatility fluorinators during conditioning and fluorination
operations could be removed by three and possibly four methods. These are:
removal (1) by rupturing or spalling, (2) by a fluxing action of the fluoride
salt baths, (3) by a washing action of the melts, and (4) by dissolution in
certain highly corrosive liquids condensable in the cooler regions of the
vessel.

Maximum losses occurred in the salt region of the Mark II vessel
which‘seems to indicate that the fluxing action of the fluoride baths was the
predominant method a} rémoving the protective films from the walls of the
fluorinator. The high vapor regions losses described for the first VPP
fluorinator were partially attributed to the presence of corrosive liquids
which condensed in the cooler regions of that vessel and some discussion of
whether similar liquids were preesent in the second vessel seems in order.

Evidence of the presence of these condensable liquids in the
middle vapor regions of both vessels were (1) the tenacious wall deposits,

(2) the segregation of chromium in surface and subsurface layers, and
(3) the bulk metal loss maxima. However, the bulk metal loss maximum in the
cecond vessel occurred a few inches below that of the initial fluorinator.

One major dissimilarity found was the lack of uranium segregation
in the vapor phase of the Mark IT vessei which was in contrast to the behavior

of the first fluorinator.- This latter fact may suggest the reason for the
o - 68 -

‘lesser middle vapor region‘éttack in the’seéond‘fluorinatof when‘compéred with
the first vessei. Perhaps the additionél operating experience of the VPP
personnel plus the modifications in the vessel's interiér filumbing and ex-
terior appendages permitted UFLF to remain in the salt baths until more or

" less complete oxidation to volatile‘UF6 had occurred. ‘

The wall-thickness-loss profile of the Mark II fluorinator shows
an additional, smaller, maximum &t a point 25 in. below the bottom of the
slip—ofi'flange. This additional peak in the upper intérface region is
beliévéd to have been induced by low operational tempefatures a few inches
below the 25-in. level. There was no direct method of heat at 28 in. below
the bottom of the slip-on flange. This region was between the furnace
%indings used to heat the sa;x region and the rod-type heating elements used
to heat the vapor region. In addition, the furnace seal was present at this
point, resulting in a built;in heat sink. Thus, thé-corrgsion profile curve
ifirobably would have continued in a smooth curve from the 25-ifi. elevafiion on
\down toward the salt region except for this low-temperature region cited. A

similar low-corrosion area was found in the wall of the Mark I fluorinator

3

Cét approkimately the same elevation.

) ‘ The corrosion losses found in the salt region of the Mark IT
if;uorinator were the maximum found in the'system and on'the order of twice
that noted for the first fluorinator in the same region. Two reaéons are
proposed for the higher saltaphaée losses in the second fluorinator. First,
"~ the fluoride salt baths in contact with thé_lower regions of the second
fluorinator contained uranium for a period approx 67% longer than for the
-firSt_fifiorinator. As will be described in a later section, the presence of
uraniumrin fused fiuoride salt systems during bench-scale volatility cor-
rosion studies has enhanced corrosive attack. Second, although the total
quanfiity of fluorine sparged during Mark II was only slightly'higher than
‘that used during 0pera£ions with the Mark I vessel, the total time of fluori-

nation for the Mark II vessel was approx 50% longer than for the first

fluorinator. Since in both cases an average of 3:1 excess fluorine over that
- 69 -

quantity necessary for theoretical reaction was used, fluorine probably had
opportunity to remain in the salt longer during Mark II operations. This
would allow the corrodent a longer period‘of time to attack the fluorinator
wall.

2. Interior Intergranular Penetration

Upon etching samples removed from the wall of the Mark IT fluori-
nator, the grain boundaries appear heavily darkened to various depths. In
the vapor region, a maximum depth of 8 mils penetration was noted. The vapor-
salt interface region showed a maximum of 16 mils while the salt region dis-
played penetration proceeding to a maximum depth of 33 mils.

Suspecting that theflintergranular attack might be the result of
sulfur contamination, personnel at BMI attempted two tests to make positive
identification of the grain-boundary deposits. The sulfur-print technique,
whereby acidified photographic paper is pressed against the metal surface,
was used. Sulfides could not be detected using this method. Next, the re-
action of a solution of lead nitrate and nitric acid upon constituents at
the grain boundaries was observed under the microscope. The precipitate
formed appeared to be similar to but weaker than that observed on a nickel
tube known to be grossly contaminated with sulfur. Battelle Memorial Institute
also indicated that the structures observed at the grain boundaries resemble
those found in L nickel rods tested and studied at ORNL.uQ |

A section from the Mark IT fluorinator wall which had been ex-
posed to the salt phase was analyzed for sulfur content at several depths
by BMI personnel, (Fig. 24). Twice the weighf percent of sulfur was found
in the interior wall at an O to 5-mil depth as that found at the same depth
on the exterior Qall, 240 vs 120 ppm. Sulfur analyses’condficted at ORNL on
wall material removed from various depths in the middle vapor region of the
vessel (Fig. 16) did not disclose any increase in sulfur content on subsurface
interior samples when compared to subsurface exterior samples, 20 vs 4O ppm.

Unfortunately, the poor correlation obtained on sulfur analyses
from BMI and ORNL on exterior wall samples plus the weak case for sulfur
hEL. R. Trotter and E. E. Hoffman, Progress Report on Volatiiity Pilot

Plant Corrosion Problems to April 21, 1957, ORNL-2495, pp. 22, 206, 29
(September 30, 1958). )

70 -
contamination presented by other test methods dilutes.any blanket statement
which could be made regarding the role that sulfur played in the interior '
intergranular attack of the Mark II fluorinator. However, the serious
embrittling and potential corrosive effects of sulfur in contact w1th nlckel
at high temperatures are -definite and known facts. Therefore, sulfur con-
tamination should be stringently avoided in any of the VPP's nickel process

equipment,

3., Exterior Intergranular Attack

The exterior of the Mark II fluorinator shell also underwent

- intergranular atteck. This was noted by BMI~personnel'in their examination

of samples removed from the vessel. The general depth of the attack, - 1-6 mils,
was of the same order as that reported for the first VPP fluorination vessel.
In the analyses of scale from the exterior wall of‘the'Mark'II, NiOg was

found indicating that the exterior intergranular attack on the second fluori-

nator was due to air oxidation.

L, Grain-Size Variations

Grain sizes found in samples removed from the wall of the Mark I
“fluorinator varied from an average ASTM number of 56 to > 1. The large sizes
occurred exclusively in the vapor regien of.the-vessel. The second VPP
fluorinator showed a different grain-size pattefn. Average ASTM grain-size
numbers in this vessel varied from 35 to 2—h the largest occurring in the
salt reglon. 7
Although the second fluorinator was fabricated from the same

eat of L nickel and in a similar manner as the first vessel, initial thermal
.conditions were quite different in the respective vapor regions. Initial
heatup for the Mark I[was done without the benefit of an external heat source
in the vapor region while the Mark IT had rod—type'heating elements with a
total rating of 9 kw attached to the upper nalf of the vessel. Heavy re-
fractory insulation covered the rod-type elements. |

| From a classical metallurgical viewpoint, it appears that many

more nucleation sites were present in the vapor region of “the second fluorinato:

;-
- 71 -

as the result of the initial rapid heatup and more or less incomplete-
recovery when compared with the first vessel. The many sites affected the
production of many grains in constant volume which could only result in com-
paratively smaller grain sizes,

The somewhat larger grain sizes found in the salt region of the
Mark II fluorinator when compared to the same region of the first vessel and
to the resulting grain sizes noted in the vapor region of the:Mark II seem
to be the result of some coalescence during operations. It will be recalled
that the first fluorinator was at 600-725°C for approx 1250 hr while the
second vessel was at about the same temperature range for over 1900 hr,

The specific effect of grain size on corrosion in the fluori-
nation system has been reported in Section I as being conjectural. However,

maximum bulk losees on the walls of both fluorinators occurred at locations

-where'the largest grain sizes predominated. As such, the production and

stabilization of small grain sizes in nickel fluorination vessels séem a

reasonable precaution. Open annealing cycles for L nickel to ensure small

grain sizes have been reported. 3

E. Corrosion of internal Components from the Mark II VPP Fluorinator

Several internal components of the Mark II fluorinator have been
sxamined and an evaluation of the corrosive attack on these parts made by per-

(ref )

sonnel ol the Corrosion Recearch Division, BMI. Figure 3 shows the
location of most of these components with respect to the top flange of the
fluorinator. Of particular interest are the draft tubes and fluorine inlet
tubes which, by virtue of their position in the system, were subjécted to

initial contact by fluorine, and could be expected to sustain the greatest

corrosive attack. Two draft tubes, Mark ITA and Mark ITB, were used in the

‘Mark IT fluorinator during Phase I and Phase II of the fluorinator's lifetime,

(Table VI). The draft tubes were fabricated at ORNL from A nickel. Results

_n3”Annealing of Nickel, Monel, and Inconel,” Tech. Bull, T-20, The
International Nickel Company, Inc., New York, April, 1953.

l‘LL'LLet‘t,e?c' Report from I'. W. Fink, Batfelle Memorial Tnstitute, to
R. P. Milford, ORNL Subcontract No. 988 (October 7, 1959).

’
- 72 -

of the examinations of the draft tubes ang fluorine -inlet lines ereoschemati—
cally displayed in Figs.. 29 and 30 and show the configuration details of the
assemblies. Supportdng photomicrographs are presented in Figs. 31 and 32, |

, Considerable metal loss was found on the‘Mark ITA draft tube, but
little visible intergranular attack was noted. Converselg the Mark IIB
draft tube showed little dimensional change but a rather deep intergranular
"attack. The etching characteristics and general microsc0pic apbearance were
also guite different. While the Mark IIB tube etched normally for nickel
using a nitric-acetic acid mixture, the other draft tube did not, -and it was
necessary for BMI personnel to use an HCerNO (3:1) etch, which they commonly
use for Inconel to develop gralnnboundary detall However, spectrographic
'analy51s at BMIT showed the Mark IIB draft tube construction material to be A
mc:kel.mL i :
- " The mekimum bu}k metal loss on the Mark ITA draftltube by micrometer
average measurement was 77'mils and occurred near the upper support plate of
the assembly."Comparison of the measured -outside diameter of the tube with
the origihal diameter indicated thet most of the metal loss occdrred on the
outer surfaces of the unit. Micrometer measurements on the Mark IIB draft
‘tube showed max1mum.bulk losses of only 6 mils but a max1mum total inter-
granular penetration on both the 1nter10r and exterior wall of,29 mils. The
fluorlne inlet lines associated with the Mark IIB draft tube assembly showed
51mllar corrosive behavior when compared to the draft tube bodies. However,
the Mark ITA fluorine inlet llne showed a much more severe intergranular at-
tack on the inside of the pipe when compared with the outside.

A high-probe line and a thermocouple well, both made froh 3/8—in.
sched-40 A nickel, were inspected also by BMi,personnel for corrosive attack.
Both components operated atftemperatures essentially the same as fhose of the
fluorinator wall. The high-probe line was a high-pressure liquid level and
density-probe combination which extended from the top flange of\the fluori-
nator beneath the molten salt level. The line was open at the bottom to
allow cohtaef with the fused fluoride salt baths,'but pressurized nitrogen,

inside the tube, for the most part prevented the salts from entering the tube.
-l

- 73 -

Unclassified
BMI A32545

F, inlet, /2in, Schedual 40 ‘A" Nickel

pipe (wall thickness =109 mils, OD=
F/ 840 mils) |

\Upper-suppori-plote fin
-y (125mil "A'Nickel sheet)

I

!

|

|

|

|
”l?lll

21A4 2l A

O-mils of sound metal remaining

(average by metallography) Draft tube, 4-in., Schedule

40 "A' Nickel pipe
(J-mils thick , {(wall thickness =237 mils)
-mils thickness

- (average by micrometer)

__~Underlined numbers are
maximum visible intergranular
penetration in mils

Lower-support-plate fin
25-mil,"A" Nickel sheet

Y

AR

AW

Scale: Approximately (/2 size

Fig. 29. Corrosion Losses on the Mark ITA Draft Tube and Fluorine
Inlet -from the Mark II VPP Fluorinator. Considerable metal loss has
occurred at all areas but most of the intergranularly attacked grains
appear to have sloughed off. ! ‘
- T4 - Unclassified
- : BMI A32546

F, Inlet, 1/2in,Schedule 40 '8 Nickel

pipe |
__— (woll thickness=109 mils, 0D=84Omils)

' - R : Upper-su ort-plate fin

\\\m : \\\\. \\\ ofF (I25-mil i Nickel sheet)
- o Oy \
e N —E |

//I(/
3

O - mils of sound metal remaining
(average by metallography)

- [J-mils thickness

- (average by micrometer) WA

{overage by micror Draft tube, 3in. Schedule 40 A
" _-Underlined numbers are / Nickel pipe (WO“ thicknes=216 m|IS)
.. maximum visible intergranular

penetration in mils

A

0,

LRSI TITHITTTITH T T

Gl
//
57
X
T

| I3

A L A a L L i i iR it t R A S i iSRSy
0 . .
7 ASLILTIILLELLALALLALLLLRARRRL R RUF-L AW n'.\\‘\\‘\\\\\\\\\\‘\\\\\‘\\\‘\\“‘\\\\‘\\\\\\\

YIRS 1111110111

Lower-support-plate fin _
— (I25-mil A" Nickel sheet)

ber

\

" Scale: ;jpbrpximately _
1/2 size & ¢

i , ’ . .

Fig. 3%. Corrosion Losses on theMark ITB Draft Tub® and Fluorine
Inlet from the Mark II VPP Fluorinator. Metal loss was not as severe as
in the case of the Mark IIA assembly and more intergranular penetration
was noted. ‘ L , . '
- 75 =

UnClaSSified Unclassified
BMT C675 BMI CATL4
' ' i
i \
? b |

{ ot
l; i “; .J)J

Unclassified Unclassified

BMI C673 BMI 0672
.\

oy

Fig. 31. Microstructures of Samples from A Nickel Draft Tubes Used
in Mark IT VPP Fluorinator. (a) and (b) exterior and interior surfaces
of Mark ITA draft tube at Section 21B; (c) and (d) exterior and interior
surfaces of Mark ITB draft tube at Section 11F. Etchants: (a) and (b)
Hydrochloric-nitric acid, (c¢) and (d) Nitric-acetic acid. 100X.
- 76 -

Unclassified Je iy

BMI C676 Yoo

\ e

3 / %
s & ' 3
— @ @
£ G &
35 ol E
M 0N b 45)

" (a)

~ Unclassified Unclassified
B _BMI C677 BMI C678

)

Fig. 32. Microstructures of Cross Sections Through A Nickel Fluorine
Tnlet Tubes from (a) Mark ITA Draft Tube (b) Mark IIB Draft Tube.
Etchants: Nitric-acetic acid. TOX.
- 77 -

The thermocouple well was positioned in a similar manner to that described
for the probe line but the end was sealed with an A nickel plug. Table X
cites the corrosive losses found on the high-probe line and thermowell as
well as giving more detail on the losses found in the fluorine inlet lines.

The internal components from the Mark II VPP‘fluorinator sustained
total corrosion losses comparable to the walls of the fluorination vessel.
No increased attack was noted on the fluorine inlet lines or on the draft
"~ tube bodies as the result of the proximity to elemental fluorine during
operations, |

Of interest is the indication of much more corrosive conditions
pres?nfi during the L-1 through L-4 runs when compared to E-3 through E-6 or
L-5'£h;6ugh L-9 runs. Extended times of service at high temperatures for
the early "L" runs may explain these differences. Corrosion control speci-
mens of L nickel in place during the run groups mentioned, and reported in
Section IV, corroborate the variable corrosive behavior present during these

different operation groups.:

ITTI. Bench-~Scale Fluorination Corrosion Studies

The Volatility Studies Group, Chemical Development Segtion A of the
Chemical Technology Division, has continued to study process chemistry in
connection with the volatility process since their eérlj work indicated the
latter's feasibility. These studies have included the gathering of corrosion
data from small-scale experiments,

As stated in Section I, nickel-base alloys have shown superior cor-
rosion resistance to fused fluoride salts under dynamic flow conditions,
and nickel and nickel-base alloys exhibit generally good resistance to ele-
mental fluorine and UF6. In this connection, commercial purity nickel and
Inconel have been the primary matgrials of construction for facilities using
the individual corrodents mentioned above., Extensive studies at ORNL on
Inconel in contact with fused fluoride salts have shown that appreciable

quantities of chromium were removed through reaction wilh UFu and other
Table X.  Measurements of the ngh—Probe Line, Thermocouple Well, and Fluorine- Inlet Tubes
: From the Mark IT Volatlllty Pllot Plant Fluorinator®
Distance from Nominal Intergranular, Wall ’
Bottom Outside Diamd Penetration® Thickness ‘Total"
| of slip-on (mils) - (mils) lossesf  Corrosion
Specimen Exposure Flange (in.) Maximum Minimum TInside Outside (mils) . (mils)
High-Probe Line
LE-2b | ‘
1F Vapor, 18 v 677 673 . L L 3 11
oF Interface 32 67Tk 667 3 10 7 20
3F Liquid - 48,5 - - 669 663 L 12 5 21"
3F(weld) - Liquid 49 - - 1 9 - -
3F(tip) Liquid 50.5 : - - 9 12 - -
Thermocouple Wellb ' | .
LF Vapor - 18 679 676 L S 1 9
5F Interface 32 679 677 L 13 3 20
6F - Liquid 49,5 675 666 L 13 0 17
6F (weld) Liquid 50.5 - - 1 6 - -
! Fluorine-Inlet. Tube
Mark IIBC _ , _ -
TF Vapor - 18 : 845 8h1 L 6 1 11
. - 8F Interface 32 838 . 826 . 9 12 - 8 29
oF Liquid ~38 829 829 10 15 3 28
10F Liquid ~u5 0 - 831 831 7 11 7 25
12F , Liquid ~l9 825 825 12 11 5 28
Fluorine-Inlet Tube \ : L '
Mark ITAC . | | |
SA Liquid ~38 © 756 756 16 3 g 63
5B Liquid ~42 © 751 751 20 3 48 71
5C Liquid b5 776 776 14 3 L6 63
-- " Liquid - ~L9 - 789 789 7 .5 30 4o

i

aletter Report from F.-W. Fink, BMI, to R. P. Mllford ORNL Subcontract No. 988 (October 7, 1959).
bConstructed from 3/8 1n.-sched-h0 nlckel plpe, nominal o.d., 675 mils, nominal wall thickness 91 mils.

CConstructed from 1/2 in.-sched-40 nickel pipe; nominal o.d., 840 mils, nominal wall thickness, 109 mils.

dBased on micrometer measurements. : : '
€Based on optical microscopic measurements. :
fBased on optical microscopic measurements and subtracted from nominal original wall thlcknesses.

-

- gL -
- 79 -

oxidizing im.puritiesa.u5 The chromium removal was accompanied by the forma-
tion of subsurface.voids in the metal. Subétitutions of molybdenum in an
approximate ratio of 2:1 (Mo:Cr) for about half of the chromium content in a
typical chromium—containing_alloy, plus other modifications, provide an alloy,
INOR 8, which has exhibited no measurable:attack when in contact with molten
fluoride salts at temperatures up to approx 700°C. Also, INOR 8 has been used
as the material of construction for the VPP Dissolver-Hydrofluorinator, the
major vessel for the head-end cycle of the Volatility Program.

| * For these reasons, a 2-in.-diam Inconel fluorinator, ten l-in.-diam
A nickel fluorinators, and four l-in.-diam INOR 8 fluorinators were tested

and subsequently examined for comparative corrosion behavior.

A, Inconel Fluorinator

1. Test Method

’ The Inconel fluorinator was used in hot-cell studies and was
fabricated from 0.065-in. stock at ORNL. Prior to the exposure detailed in

.Table XI, fluorine flowed into the fluorinator as it was raised from ambient

Table XI. Exposure Conditions for the Inconel Fluorinator Vessel

Temperature, °C 600-800 |

Time of exposure at temperature, 187

‘hr with salt molten

Thermal cycles 70 ‘

Fused salt, nominal mole % - NaF-ZrF) -UF)
| - (4B-148-1)E

fluorine input; standard litersa ' 294 in h9 hr

“Uranium irradiated, not enriched.

Mbw. D. Manly et al., "Metallurgical Problems in Molten Fluoride Systems,"
Progress in Nuclear Energy, Series IV, Vol 2 — Technology, Engineering, and

Safety, pp.lbL—179, Pergamon Press, London, 1960.

_ 80 -

room_temperature.to operating temperature. This was done for leak-testing .
purposes although, at the same time, the interior veséel walls were probably
"conditioned" by producing films of fluorides. The total time involved for
this testing-conditioning‘treatmgnt was about 2 fir. .

After exposure operations, areas from the vessel were selected for
micrometer measurements and metallographic examination. Figure 33 shows a
cross section of the vessel and the location.of the sample areas removed for
study. A summary of corrosive attack found is shown in Table XII where total
losses are re@orted ifi mils per hour, based on fluorine sparge time, and mils.
. per month, based on résidence time in molten salts. Representative photo-

‘micrographs'aré grouped in Fig. 3k.

2. Discussion of Results

. ‘Maximum corrosive éttack ifi the Inconel fluorinator was encoun-
tered at the vapor-salt interface and occurred at a.rate of 0.2k mils/hr, based
on fluorine sparge time, or 48 mils/moqth, based on molten-salt residence time.
Intergranular penetrgtion seeméd the predominant mode of attack in the éalt
and vapor regions but no evidence .of intergranular-cérrosion'was found at the
vapor-salt interface. _ |

In this connection, the‘intergranular'penetration seemed dis-
similar to that found on the“interiéf walls of- the VPfi.fluorinators.' That is,
ofl the Inconel vessel there was evidence of the sloughing of* whole grains -of
material, a condition not observed on the full-size L nickei vessels. The
depth of the infergranular penétration on the Inconel vessel wés only a single
grain deep. 'The VPP fluorinators demonstratediintergranu;ar modifications
‘many grains in depth on samples removed from corresponding service regions con-
taining similar grain-size material. ' |

| In the Inconel bench fluorinator, it appears that after inter-
gfanular'penetration had proceeded to abofit one grain. in depth, many of the
affgcted grains could not be retained in position by the remainder of the
graifi—boundéry‘matérial and sloughed off, leaving new material ready for cor-

rosive attack. This method of metal loss would seem devastating to the vessel
- 81 -

UNCLASSIFIED
ORNL-LR-DWG 49162

12F N

10F N

~ 0.065-in. WALL
THICKNESS

12 in.

4 .
*4F AT ¢
% 32
’ ? Pd
/ / -
/ / f
/ /
/ /
‘ ’
¢ /
/ / -
g ¢ =
2F MMM S
0T
SPECIMEN 4F §\\‘\\‘\\\\\\\\\\\2§ . €
: S
[l
17
i1 1 |

FLUORINATOR

Fig. 33. Cross Section- of the Chemical Development Section-A Inconel
Fluorinator. Metallographic specimens were removed as shown.
-Table XII. Summary of Maximum Corrosion Results ona' Chemical

) Inconel-Fluorinator_

Development's -

4

Losses'Converted to

: ‘ - , Intergrénular‘ 3 ‘
. " Estimated ..  Wall Penetration Mils/Unit Time
Speci- Wall Thickness Interior Exterior Total -mils/month mils/hr
‘men . Specimen  Temp Loss _ wall wall Corrosion (molten- (Fg.sparge
‘No.  Location - (°C) (mils) (mils). (mils) -~ (mils) salt time) time)
12  top vapor 100 3 - - 3 12 0.06 .
' o
10 upper vapor 250 3 3 5 11 ouP - 0.12 ,
. 4 lower vapor 500 8 P - 10 40  0.20
2 .vapor-salt 600-800 ‘12 - - 12 48 - 0.2k
interface | .
1 salt 600-800 9 2 - 11 4 . 0.22

a, . - .
Four micrometer measurements taken in each area..

Does not include the exterior attack found on-the vessel wall.
Y -26551 UNCLASSIFIED
- 83 - ORNL—LR—DWG 32069

Y-26548

INNER SURFACE,
VAPOR AREA

Y-26550

OUTER SURFACE,
VAPOR AREA

|——— SALT LEVEL

INNER SURFACE, APPROX.
SALT-VAPOR INTERFACE

s B B | .

E lg lg ]§ :mcinzs

INNER SURFACE ,
SALT AREA

Fig. 3L, Typical Microstructures of Samples Removed from Chemical
Development Inconel Fluorinator. ZEtchant: Glyceria regia. 250X.
Reduced 32%.
- B -

in service, but the anomaly of the situation is that maximum wall-thickness
losses occurred in the Inconel vessel at the vapor-salt interface and at that
point no evidence of intergranular attack could be found.

No explanation can be given for the interior wall interface
anomaly described nor can one be given for the heavy exterior intergranular
attack which has localized in a cooler region of this Inconel fluorinator

(see Fig. 34).

B. A Nickel Fluorinators

1. Test Method

In addition to the Inconel fluorinator reported upon in
Section IITA, ten A nickel miniature fluorination vessels, each l-in.-o0.d.
and 0.035-in.-wall thickness, have been operated by the Volatility Studies
Group, Chemical Development Section A, in order to compare fluorinator cor-
rosion using different process flowsheets. Figure 35 shows the units in test
position, while Table XIII is a summary of the imposed test conditions.

The vessels, in all cases, were charged with 75 g of fluoride
salt that had been ground and classified to =8 +20 mesh. The vessels were
then placed in split tube-type furnaces and brought to the specified tempera-
ture under a nitrogen purge of 0.05 liters/min. At temperature, the nitrogen
was bypassed and fluorine was allowed to bubble through the salt at the same
flow rate. In those cases where uranium was added to the melt, 0.50 g
increments of UF& were added at intervals of 1 hr or more. As shown, some
reactors were placed in series both for convenience and to minimize the
fluorine consumption. The total elemental fluorine exposure in each case was
50 hr at a rate of 0.05 liters/min. During these 50-hr exposures on test re-
actors Nos. 5, 6, and 10, twenty-five UFu additions and UF6 volatilizations
were made.

In all tests, the salts were kept molten for over 200 hr while
under a nitrogen purge. This was done to facilitate the test procedure, that
is, avoid remelting salts for the next fluorine test period. Figures 36 and 37

illustrate the corrosion results obtained by micrometer measurements and
<« i35 -

UFh additions
UFM additions

1]
)
—

]

0

(]
o)
o

bl

g
—~
=

Fluoride salts

——

Unclassified
Photo 45840

Fluorinators

*

ol

M e o e A

Split Tube
Furnaces

Fig. 35. Apparatus for Comparison of Fluorination Corrosion on A Nickel
Miniature Reactors.
- B

Table XIII. Process Conditions for A Nickel Bench-Scale
Test Fluorinators

Hr of Molten Hr of Molten

Vessel Fluoride Temperature Salt Exposure Salt Exposure Position in

No. Salt (%) With N, Sparging With F, Sparging Test Series
1 * 450 222 50 1
L * 525 238 50 1
7 * 525 240 50 1
8 * 525 240 50 2
2 * 600 222 50 2
5 2 +0.5%U 600 238 50 2
9 ¥ 525 2Lo 50 3
10 3 +0.540U 505 240 50 1
3 RHK 600 222 50 3
6 1 +0.56U 600 238 50 1

*¥26—-37—37 mole % LiF—NaF-ZrFu: Composition 31, plus LiF addition.

*¥31—24—L5 mole % LiF-NeF-ZrF): Composition 31, plus LiF and ZrF)
additions.

¥**¥50-50 mole % NaF-ZrF): Composition 31, as received.
= UNCLASSIFIED
ORNL-LR-DWG 49163

TEST SALT TEMP.
NO. NO.  (°C)

| 2 450 v//A | |

(SLIGHT PITTING NOTED)

4 2 525
7 2 525
8 2 525

2 2 600

9 3 525

SSEERERIE L LEGEND
10 3 525 |V}l P2
0;‘7 : | R V- VAPOR
5% < > ! - VAPOR SALT INTERFACE
U S M : . S - SALT s
0 , ]
3 1 600 (vl il - METAL LOSS/MICROMETER MEASUREMENT
| -
s b A - INTEGRANULAR ATTACK/METALLOGRAPHIC EXAM.
MAXIMUM LOSS SHOWN IN EACH PHASE ‘
6 { 600 V[
+ I
0.5%
P v S i
0 5 10 15 20 25 30 35 40
mils/mo

(BASED ON SALT RESIDENCE TIME)

Fig. 36. Summary of Corrosion on A Nickel Miniature Fluorinators Using Different
Process Flowsheets..
H \
A
! % P s, -
UNCLASSIFIED
ORNL-LR-DWG 49164
~ ! ~
TEST " SALT TEMP .
NO. NO. (°C) -
1 2 450 |VE :
’ % Z]{SLIGHT PITTING NOTED)
s ) e -
4 2 s25 ‘I’
Sl }
7 2 s25 |V .
| \
St s ) ’
. | | A
8 2 525 Y7772
A ]
2 2 600 —
]
5 2 600 2 _ .
+ L T I R T S T L e L ey
oxa LI X
U
9 3 528
o 3 s2s |VE V22 -
059 'L 22272, LEGEND
3% s |- s V - VAPOR
| - VAPOR SALT INTERFACE
3 i 600 |V S - SALT
4 3 .
3 ] - METAL LOSS/MICROMETER MEASUREMENT
6 i+ 600 |V - INTERGRANULAR ATTACK/METALLOGRAPHIC EXAM.
¥ ! : MAXIMUM LOSS SHOWN IN EACH PHASE
0.5% S 20
v » ° N L L
) 0.025 0.050 0.075 0.1 0.125 0.150 0.175 0.2 . - 0.225 . 0.250 0.275 0.3
. mils/hr ‘
i : . ) (BASED ON FLUORINE E XPOSURE TIME)
Fig. 37. Summary of Corrosion.on A Nickel Miniature Fluorinators Using Differen
Process Flowsheets. . .
1
/
‘ ' i i ' ' t
. ~ v i t .
- 89 -

metallographic examinations of specimens removed from the vessels. The losses
are reported both in mils per hour of fluorine sparge and mils per month based
on total residence time in molten salts for comparison with other studies.
Figures 38 through 42 show representative photomicrographs from this test

series.

2. Discussion of Results

The A nickel fluorinators showed widely varying corrosion rates
as anticipated in planning this test series.

Comparison of the results on test vessel No. 6 with those of
vessel No. 3 indicated that uranium additions to equimolar NaF-ZrFu resulted
in increased corrosion occurring as intergranular attack at the vapor-salt
interface. A rate of 0.1 mils/hr based on fluorine sparge time was noted at
the interface of No. 6.

Comparison of vessel No. 2 with vessel No. 3 indicated that the
addition of 26 mole % of LiF caused increased attack at the vapor-salt inter-
face, as evidenced by bulk metal losses. Upon metallographic examination, no
evidence of intergranular attack was found in either vessel.

The addition of uranium to the LiF-NaF-ZrF (26-37—37 mole %)
salt in vessel No. 5 resulted in significantly increased metal losses plus
intergranular attack at the vapor-salt interface. Additional intergranular
attack was noted in the vapor phase of this vessel.

Lowering the temperature of the reactor containing the same
LiF-bearing salt, LiF-NaF-ZrF) (26-37—37 mole %), to 525°C produced erratic
corrosion results. Two of the vessels, Nos. 4 and 7, showed uniform, compara-
tively small metal losses in all regions while vessel No. 8 showed an increased
attack, especially at the vapor-salt interface. Intergranular attack was also
exhibited by vessel No. 8. As shown in Table XIII,vessel No. 8 was in a down-
stream position from vessels Nos. 4 and 7, and the deviation in behavior may,
therefore, have been the result of carry-over and collection of certain con-

stituents conducive to greater corrosive attack.
UNCLASSIFIED
ORNL-LR-DWG 49826

Y-28366

AS RECEIVED MATERIAL

INNER SURFACE,
VAPOR AREA

INNER SURFACE,
APPROXIMATE SALT-VAPOR
INTERFACE

////////m/ ( )

L

é
?
7777777

wilgw| | Cle@K®@WwI[ [ |u= X
w s o 2 o [0

5 G [6mncHes[o I8 [8 |18 [8 |8 200x (8 [8 [
clolo|] | @O ool lo]| | |olo P

N
Q
\,\.‘ 5 /‘x

‘ =
INNER SURFACE, SALT AREA

Fig. 38. Typical Microstructures of Samples from A Nickel Miniature
Test Fluorinator No. 1. Etchant: Acetic-nitric-hydrochlorie acid. 200X.

Y-28366

AS RECEIVED MATERIAL

/

Y=28372

!

INNER SURFACE,
VAPOR AREA

Y-28370

\

\

N

————— N

\

\

NN

[+
§§2|LCH552§8§§§'200><
o lo lo | o |lo |© [0 |lo |o |

Fig. 39. Typical

Test ‘Fluorinator No. 5.

INNER SURFACE, SALT AREA

Etchant:

UNCLASSIFIED
ORNL-LR-DWG 49827

INNER SURFACE,
APPROXIMATE SALT-VAPOR
INTERFACE

Microstructures of Samples from A Nickel Miniature
Acetic-nitrie-hydrochloric acid.

200X.
Y=28366

AS RECEIVED MATERIAL

Y-28639
> . .

INNER SURFACE,
VAPOR AREA

0.045
0.014

L . )
L \ /
| ol y
\\ .,' % !
a‘ ¥ <
- v -
% o .
L 5
~ AT
l \ :' »
§
( )
g .
5 Ve SN
% 2.
) \ -
i 3
! ' -
! . ?
S Sl 3 78
5 b N -
i \ N 4

INNER SURFACE, SALT AREA

UNCLASSIFIED
ORNL-LR-DWG 49828

INNER SURFACE
APPROXIMATE SALT-VAPOR
INNERFACE

Fig. 40. Typical Microstructures of camples [rom A Nickel Miniature
Etchant: Acetic-nitric-hydrochloric acid. 200X.

Test Fluorinator No. 6.
UNCLASSIFIED
ORNL-LR-DWG 49829

Y-28366 Y-=28375

<
<

X /) < /
AS RECEIVED MATERIAL INNER SURFACE, 7\ b il
VAPOR AREA = \ s
» ‘ i "
\
A
(p
Y A
N Y o
i
[ T\
S
A

INNER SURFACE,
APPROXIMATE SALT-VAPOR
INT ERFACE

0.015

0.014
0.043
)—2—1

o

&

m

w
0.010
0.001

’
\ P
o
/
/ - .

INNER SURFACE, SALT AREA

Fig. 41. Typical Microstructures of Samples from A Nickel Miniature
Test Fluorinator No. 8. Etchant: Acetic-nitric-hydrochloric acid. 200X.

UNCLASSIFIED
ORNL-LR-DWG 49830

Y-28366

AS RECEIVED MATERIAL

INNER SURFACE,
VAPOR AREA

INNER SURFACE
APPROXIMATE SALT-VAPOR
INNERFACE

< N
o~ o Q
\
BRERERRRiSSSy

=

o
| | lo®@ @ & |©
2 15 [ mncHeslS S 8 1S IS
ocololo] | |loo o |o|o

INNER SURFACE, SALT AREA

Fig. 42. Typical Microstructures of Samples from A Nickel Miniature
Test Fluorinator No. 10. Etchant: Acetic-nitric-hydrochloric acid. 200X.
- 95 -

Lowering the process temperature to 450°C and utilizing the same
lithium-bearing salt (No. 2) reduced corrosive attack to the lowest levels |
found in this test series. These results were provided by vessel No. 1 where
the approximate losses wefe 0.02 mil/hr, based on fluorine sparge time. This
is especially significant since lithium-sodium-zirconium fluoride salt mix-
tures can be used in the volatility process at lower operating temperatures
than the NaF-ZrFu composite system because of the lower liquidus line of the
lithium~bearing system. _

Tests Nos. 10 and 9 at 525°C used higher LiF- and ZrFu-content
salts, 31—24-L45 mole % LiF-NaF-ZrF), with and without uranium, respectively.
Vessel No. 9 showed slightly increased attack over vessels Nos. 4 and 7 which
wefe operated at the same temperature. Metallographic examination of
vessel No. 10 revealed that intergranular penetration was present in all
phases of the interior wall, A

The A nickel miniature fluorinators demonstrated generally lower
rates of corrosive attack in simulated fluorination environments when compared
to the full-sized L nickel vessels or the latter's A nickel internal components
which were exposed to pilot plant fluorination conditions. The following three
reasons can probably account for most of this deviation: (1) the temperatures
of fluorination generally were somewhat lower in the bench-scale work than
“‘during pilot plant operations; (2)'somewhat better control over thermal cycling
and other process conditions was obtained during bench-scale Work; (3) the
feed salts used in the pilot plant work were contaminated by having been,used
in previous loop studies, the ARE, or having been contained for long periods
of time in the pilot plant type 347 stainless steel charge melt tank.

The A nickel miniatures also demonstrated greater resistance to
corrosive attack during fluorination than the Inconel vessel used in Chemical

Development hot-cell studies.

C. INOR-8 Fluorinators

1. Test Method

Four INOR-8 test fluorination vessels, each 1l-in.-o0.d. and

0.065-1in.-wall thickness, and of similar design to the A nickel reactors
- 96 -

‘reported in Section IIIB were also fabricated at ORNL for bench-scale
volatility corrosion studies. The bomposition_of INOR-8 used in test fluori-
‘nators was 7.5 wt % Ni—-15.3 wt % Mo—6.5 wt % Cr—3.7 wt % Fe-0.02 wt % C.

' A summary of the test conditions for the INOR-8 miniature fluori-

nators, as shown in Table XTIV, indicates that only vessel No. 3 contained

Table XIV. Process Conditions for INOR-8 Bench-Scale
‘ ‘ Test Fluorinators.

"Hr of Molten Hr of Molten

Vessel Fluoride Temperature Salt Exposure Salt Exposure Position in
No. Salt (°c) With N, Sparging With F, Sparging Test Series
1 * - 450 - 236 50 1
D *% 600 236 50 2
3 2 +0.5% U 450 290 | 50 1
4 *. 600 , 290 - 50 1
| *26-37-37 mole % LiF-NaF-ZrFuE From addition of LiF to Composition 31
salt. ' - |

*¥¥50_50 mole % NaF—Zth: Composition 31, .as received.:

| uranium. . Vessei-No..3 received 25 UFh additions which were subsgquently
’_fluorinated to UF6 in like manner as the A nickel miniatures previously re-
ported. Figures 43 and 44 illustrate the corrosion losses obtained by microm-
eter_meaSurementg and metallographic examinations on'specimens removed from
the walls of the vessei. The losses have been converted to milé lost'per
unit time for comparison purposes. Figufes 45" through 48 show typical etched

.microstructures found in the INOR-8 fluorinator specimens.

2. Discussion of Resfilts

The INOR-8 test fluorinators showed é variety of corrosion'méni-
festations depending on the test conditions. Considering the two vessels which
operated at 600°C, vessel No. Ejlwhich'contained equimolar NaF-ZrFu, showed
lgss attack than vessel No. 4, which contained LiF-NaF-ZrF) (26-37—37 mole %).
! UNCLASSIFIED
ORNL-LR-DWG 49165

TEST SALT TEMP
NO. NO. (°C) .
| 2 450 UV [

MV
I -~
s |
2 { 6CO
3 2 450
+
0.5% |
u Vo)
g
’ i
q 2 600

-7 r
- . LEGEND
UV-UPPER VAPOR
MV -MIDDLE VAPOR

| - VAPOR SALT INTERFACE
S-SALT

il -METAL LOSS /MICROMETER MEASUREMENT

ADDITIONAL ATTACK/METALLOGRAPHIC EXAM,
(CORROSION - PRODUCT LAYERS)
MAXIMUM LOSS SHOWN IN EACH PHASE
I | | | |
"0 5 10 15 20 25 30 35 40
miis/mo
(BASED ON SALT RESIDENCE TIME)

4
s

Fig. 43. Summary of Corrosion on INOR-8 Miniature Fluorinators Using Different
Process Flowsheets.

i
TEST SALT TEMP

0, 0,9,.0,.0.0,.0.0.0.0_

RERZD2Z] - ADDITIONAL ATTACK -SPONGY LAYERS / METALLOGRAPHIC EXAM
' ~© MAXIMUM LOSS SHOWN IN EACH PHASE ‘
T o 0.025 0.050 0.075 0.125 0.150 0.175 0.2 0.225 0.250- 0.275
: mils/hr
(BASED ON FLUORINE EXPOSURE TIME)

LEGEND

UV - UPPER VAPOR -
MV - MIDDLE VAPOR .
I - VAPOR SALT INTERFACE

S- SALT ' .

¢

- METAL LOSS/MICROMETER MEASUREMENT

Process Flowsheets,

Fig. 44. Summary of Corrosion on INOR-8 Miniature Fluorinators

4

0.3

T ZuUNCLASSIFIED
. ORNL-LR-DWG 49166
NO: NO. (°C) - T
1 2 4s0 [UVEi ‘
MV .
|
S
2 600 [UV

-86_

s
Y-28800

UPPER VAPOR AREA

6in. —— =

o ¢ ™ | Jo o [@ |~ |o 8 N |-

5 [ [o INcHESS (3 |8 |3 |§ |8 200X (3 I3

O lolo]| | ©]lc |o|o |0 |o o o
Fig. U45.

A
MIDDLE VAPOR AREA

VAPOR-SALT INTERFACE AREA

UNCLASSIFIED
ORNL-LR-DWG 498314

Y-29062

AS RECEIVED MATERIAL

Y-28803
e 7 i £y
" " 4‘.\; ‘:t< \ ‘ i 7‘ r e
- : ,
\\
\ Py | i
\
—" 1 (
\ g L
g SRR ;
! X4
| ¥ {
a3 \,‘
' | .
I ' o !’ !
L] '! N
4 {
’ % '!t e .
J ¥
e
o Mgk 1
S " : e
. ‘ oy
SALT AREA

Typical Microstructures of Samples from INOR-8 Miniature Test

Fluorinator No.

1.

Etchant:

Modified aqua regia.

Y-28805
; "fi“;"‘ 5
NG
o\
A ; A
B x
1
A
i $ *
1
\s \
“X
1
, L)
' ‘} £\
|
v
| \
{ BN} \
"
| .

gy
UPPER VAPOR AREA

6 in.

D IIIIIIIITIIS VISP IS IIIITI NI,

7

J

MR

0

Fluorinator No. 2.

Etchant:

S |5 [oINncHES[S 8 8 8 18 [8 200x (8 (8
clolo|] | olololcloele] | |olo
Fig. L46.

- 100 -
Y-28806
|
i
\ i
i > . ¥
- |
jai g
fr—k
H
LN,
% ;‘k.\v_
T
L
i : . ;
; . el Y ¢ » 1
= i -~
P .
\‘ )’ o
«" ] d {
4 e ' e
by b4 \< :
A
MIDDLE VAPOR AREA
Y-28807
'l'
'
1
TSt
-
&1
" .. 1
' G
3 1
"%
5
.
g ;T
*® :
> )_" :
'. 5
i y
&
."n
' . A
7S
i

VAPOR-SALT INTERFACE AREA

UNCLASSIFIEC
ORNL-LR-DWG 49832

Y-29062
b g
i
AR T
¢ ool <
. |
?\ «’ ‘)
BT A
o4
\ b ¥
e s
7 ‘ i 4\!
Ry { \ v
7 rl‘\‘ e
AS RECEIVED MATERIAL

Y-28808
» [
st} ‘
R
i |
- L
A 5% »
\:1 "
£ ]
l\ J ':; ’ i
o L )
3 ‘ g
N N\,"
}: /
. -~
e B } : N
1~;. (
z '
4“ {
"‘ ’ i ! &
-
I o -
SALT AREA

Typical Microstructures of Samples from INOR-8 Miniature Test

Modified aqua regia. 200X.
UPPER VAPOR AREA

- 101 -

. _
MIDDLE VAPOR AREA

Y- 28853
L
[ ~|‘ ‘
J o
(i o
3 ‘..!'/.'.
8 |
| |
-
. L
o He
y '
,,‘_
.'
s !
N ks
€ \ :
© N
\ g
\
3 {
R e

- N

g I \ il o

£ ¥R |

v |

‘\ '

‘ \ :

‘ VAPOR-SALT INTERFACE AREA
wilel]| | cRESELERI] | &
656|NCHESq8]é88_qzoox88
ocloloc] | cloloclolcle| | lols

Fig. 47.
Fluorinator No. 3,

Etchant:

UNCLASSIFIED
ORNL-LR-DWG 49833

Y-29062
S
!‘ )
L
S Va
: S
5 )\
) 7 .')
\ A/.",\,_ N
‘Lilu
\ LA ,‘ i
NS A AT /
G AL
7 ’ E

AS RECEIVED MATERIAL

Y-28854

i i g
\ e 3

"T)gi—\‘\

N

SALT AREA

Typical Microstructures of Samples from INOR-8 Miniature Test

Modified aqua regia. 200X.
|
UPPER VAPOR AREA

6in

=

|

=

|

PP I TIIIIIIIGSIIIIIIIIOITIINIIIII AN

|

— r‘JQm

0.015

0.014

0.013

0.010
0.009
0.008
0.007
0.006
0.005

n

o

O

5

£
INCHES
s

0.002

0.001

Fluorinator No. L.

Fig. L48.

Etchant:

Y-34026

a
MIDDLE VAPOR AREA

Y-28858
7 s
) - < }
§ue $ ¥
| -
| ul )
\
1
\ g
. 'A 1' 5
PRI
l‘l
-7\
- AT
5 )
] ! M
L Y
“ " 1]

VAPOR-SALT INTERFACE AREA

UNCLASSIFIED
ORNL-LR-DWG 49834

Y-29062

AS RECEIVED MATERIAL

Y-28859

SALT AREA

Typical Microstructures of Samples from INOR-8 Miniature Test
Modified aqua regia.
% 108 =

Maximum attack in the latter vessel occurred at the vapor-salt interface at
a rate of 30 mils/month, based on residence time in the molten salts. This
was the maximum attack rate for the entire INOR-8 series.

Reducing the operating temperature to 450°C and using the same
lithium-bearing salt significantly reduced attack as indicated for vessel No. 1
(Figs. 43 and 44). The meximum attack found in this vessel was also at the
vapor-salt interface at a rate of 5 mils/month. The addition of uranium to
the lithium-bearing salt described plus operation at L450°C more than doubled
the corrosive attack when compared to the nonuranium-containing salt. This
occurrcd in vessel No. 3.

Examination of the metallographic specimens in the as-polished
state, Fig. 49, disclosed spongy layers on the surfaces of the vapor region
specimens from vessels Nos. 2 and 4. These two vessels operated at the highest
temperatures of the test series. Upon etching with a modified aqua regia
solution, 5:1, HCl:HNOs,

Figures 46 and 49 display the microstructures for specimens from the two ves-

the spongy regions were for the most part destroyed.

sels and show that in only one region, the upper vapor area of vessel No. 2
(Fig. 46) did significant amounts of the spongy region remain.

Careful repolishing on the upper vapor area sample from vessel
No. 2 and examination of the surface disclosed a corrosion product which re-
solved itself into two distinet layers, as shown in Fig. 50. Just above the
INOR-8 basc metal was a spongy region which was composed of voids and solid
metal intermixed. Many of the voids appeared to be partially filled with
nonmetallic-appearing compounds. Above the spongy region, on the outermost
surface of the specimen, was an irregular layer that had the angular appear-
ance of metal crystals.

Upon etching the specimen with a mixture of 10% KCN and
10% (NHM)ESQOB’ 1:1 ratio in water, the outermost layer of the surface showed
different characteristics than the INOR-8 base metal. The comparatively mild

etchanl delineated grain boundaries in the ounter layer but left the base metal

unaffected, as displayed in Fig. 51.
Fig. L49.

As-polished.

w108 -~

Unclapgified

Unclassified

Y-28815

Typical Spongy Surface Layers from Vapor Region Samples of
(a) INOR-8 Test Fluorinator No. 2 (b) INOR-8 Test Fluorinator No. L.

500X.
- 105 -

‘Unclassified
Y-31202

Fig. 50.

Surface Layers on Sample from Vapor Region of INOR-8 Miniature

Test Fluorinator No. 2. As-polished. 500X.
- 106 -

Unclassified
Y-31657
, |

800
o
o}
o]
x

900

$00
+00
€00

200’
=
(2}

=
m
w

Fig. 51. Surface Layers on Sample from Vapor Region of INOR-8 Miniature
Test Fluorinator No. 2. Etchant: Potassium cyanide-ammonium persulfate.
500X.
- 107;.

- Samples of the outer corrosion product layer and of the spongy
subsurface region were obtained by meghanical milling and sérapping. Most
of the material comprising the outer layer was found to be strongly ferro-
magnetic in contrast to the base metal. These sampleé plus millings removed
from the exterior diameter of the No. 2 vessel wall, 5-10 mils below phe
surface, were submitted for spectrochemical analyses. All of the sampies
were quantitatively analyzed for Cr, Mo, Fe, and Ni by emission spectroécopy
using a porous cup electrode method of sample excitation. The results are
shown in Table XV.

- .Table XV. Analyses of Corrosion Products and Base Metal
Removed from INOR-8 Miniature Fluorinator No. 2

Sample Description and Chromium Iron Molybdenum Nickel
Location (wt %) (wt %) (wt %) (wt %)
Outer corrosion product layer | 3.9 3.3 8.4 84,3 )
Spongy subsurface layer 7.6 3.8 16.3 69.4
Base metal 6.6 3.75 15.45 75.45

Examination of Table XV shows that the outer corrosion product

layer was 13 wt % richer in nickel when compared to the base metal, INOR-8,

while the chromium, molybdenum, and iron contents in the corrosion product
decreased 40, 45, and 11 wt %, respgctiveiy. The spongy subsurface layer
showed generally small weight percent increases for the chromium, molybdenum,
and iron, respectively,-when compared to the base metal analysis. Nickel in
the subsurface layer was approx 8 wt % less than that found in the base metal,
The mode of, corrosive attack on INOR-8 in contact with the
simulated volatility process fluorination environment at 600°C appears to
involve selective losses of chromium, molybdenum, and iron from the nickel
solid solution. This may occur by formation of metal fluorides on the bulk
metal surface Which subsequently volatilize, or by the various methods of

fluoride film losses described in Section I.
- 108 -

The reason(s) for the slightly higher minor elemenfl concentra- .
- .tions in the spongy subsurface corrosion layer when compared with the base
metalrappeaf(s)ranomaloué to normal diffusion processes. Concentration .
gradlents produced by the initial selective losses -should become the driving
force for dlffu51on of the minor alloy, elements toward the exposed surface
of the reactor and lower than base metal concentrations of chromium,
molybdenum, and iron would be expected in the subsurface.region. It may be
that complex fluorlde compounds are formed in the subsurface region which do
not wvolatilize under the test conditions and thus simply tie up higher con-
centrations of chromium, molybdenum, and iron. However, since X-ray
- diffraction patterns bn.sampies from. the. sublayer disclosed only the presence
of a face-centered cubic material very similar to the patfern obtained on
the base metal, supporting evidence for this theory has not been obtained.

. Detailed treatment of the corrosive attack on INOR-8 deserves
: seperate study which may or may not be warranted considering the ultimate
.goal ef the entire test series covered in this section; that is, comparative
. behavior of A nickel, Inconel, and INOR-8 in contact with a volatility process
. fluorination environment. In regard to the iatter, A nickellhad the best
fifesistance to attack at pilot plant fluorination temperatures,ef 600°C; but
. at 450°C INOR-8.presented favorable competition to the nickel. There are
several advantages attendant to using INOR-8 as a fluorinator construction
material including its high strength and exidation resistence; Also, since
the construction material for the VPP hydrofluorinater is INOR;B,'it might be
possible to consolidate both hydrofluorination and fluorination operations in

a single vessel.

-

IV. Volatility Pilot Plant Scouting Corrosion Tests

A.' Material Selection

In order to take advantage of the service conditions provided by the
VPP process runs and to achieve insight into the resistance of other materials

to the complex fluorination environment, & series of corrosion specimens was

located in the Mark I and Mafk IT fluorinators and examined,at'convenient
n

- 109 -

intervals. As mentioned in Section I, resistance to further attack by
fluorine on metals was felt to be imparted by passive fluoride films which
form on the metal. Protection has been shown to be dependent on the proper-
ties of the fluoride films, especially volatility, adherence to the substrate
metal, mechanical and thermal stability, and thickness.

For the scouting corrosion tests described in this section, selec-
tion of materials which contained constituents known to form low volatile
fluorides seemed especially appropriate. As a guide to volatility, the
melting and boiling points of the common ingredients in many commercial

J
materials of construction were reviewed. A partial listing taken from Brewer 6

d's shown in Table XVI. It can be seen that chromium and molybdenum, commonly

added for improved resistance tc air oxidation and/or improvement of high-
temperaturé properties, form highly volatile fluorides, especially at their
higher oxidation states. Nevertheless, because of the previous use of nickel-
base alloys containing chromium (Inconel) and/or molybdenum (INOR-8) in
the Aircraft Reactor Experiment (ARE) and the Molten-Salt Reactor (MSR) studies,
alloys containing these two constituents were included in the materials selected
for the scouting tests.

Table XVII lists the specimen materials, their trade names, where:
applicable, and thekgéneral alioy classifications to which they belong. Most
of the materials were nickel-rich alloys. The 90 wt % Ni—-10 wt % Co and |
80 wt % Ni—20 wt % Co alloys were nonproprietary and were fabricated from
melts made in Metallurgy Division facilities. The remaining materials were
obtained from commercial suppliers. The platinum specimen served as an elec-
trode probe in an attempt by VPP operating personnel to investigate the
electrochemical effects during the fluorination process. However, current
did not flow through the probe and corrosive losses were recorded in a regu-
lar manner for the platinum specimen. Table XVIII shows the nominal composi-

tions for all scouting test materials.

uéL. Brewer, "The Fusion and Vaporization Data of the Halides,'" The Chemistry

and Metallurgy of Miscellaneous Materials: Thermodynamics (ed. by Laurence

L.Quill) National Nuclear Energy Series, Div. IV 19B, McGraw—Hill, New York, 1950,
- 110 -~

Table XVI. Melting and Boiling Point Data of Various Metal Fluorides
Constltuents of Materials Used in the Volatlllty Pilot Plant
Scouting Corrosion Tests® :
(Converted from °K)

Meltlng Point "~ Boiling Point

Element Fluoride . (°c) . (°c)
Ni . NiF T 10277 - 1627b
Fe : FeFi ' 1102 - , 1827:

FeF3 1027 | 1327b |
Co CoF,, 1202b | - 1727b
CoF 5 - _ 1027 ’ v1327
- Cu "CuF , Di'sproportionates ‘
- CuF,, | oo 1377°
Al AlF -  geq® 1357
A1F3 > 1272 1272
Mg MgF, 1263 2227
Mn | MnF,, 856b 2027
MnF 1077b 1327b
Ti . TiF3 1227b~ 1427
TiF) Lot 28ub
Cr CrF, 1102 , 2127
CrF3 1100 14272
CrF), 277 297
CrF5 102 1172
Mo MqF5 77 | 227" .
MoF 17 , 36
Pt - PtFi R > 727 > 727"
| PtF), > 627 : a 727"
PtF, Decomposes a% o80°c(¢) )
Au AuF 727 . > 727

aL Brewer, "The Fusion .and Vaporization Data of the Halides," The Chemistry
and Metallurgy of Miscellaneous Materials: Thermodynamlcs (ed. by Laurencé L.
Quill) National Nuclear Energy Series, Div. IV 19B, McGraw-Hill, New York, 1950.

‘bEstimated or Obtained by Extrapolation of Experimental Data by L. Brewer.

CBernard Weinstock and J. G. Malm, "Some Recent Studies with Hexafluorides,"
‘Basic Chemistry in NucleaT Energyf728, Pp. 125129, 2nd United Nations Inter-
national Conference on the Peaceful Uses of Atomic Energy, Geneva, 1958. '

N

- 111 -

Table XVII. Corrosion Scouting Test Specimen Materials Used in the
Volatility Pilot Plant Mark I and Mark II Fluorinators

Platinum

~ Material Classification
L Nickel _ Ni:
INCO 61 Weld Wire Ni
Gold-plated L Nickel Ni
90 wt % Nickel-10 wt % Cobalt Ni-Co
80 wt % Nickel—20 wt % Cobalt Ni-Co
Cobanic Ni—Cq
‘Monel Ni-Cu
D Nickel Ni-Mn
Nimonic 80 Ni-Cr
Inconel Ni-Cr~Te
Waspalloy Ni-Co-Cr
(+ Ti, Al, and Fe)

INCO 700 Ni-Co-Cr
Hymu 80 ‘Ni-Fe-Mo
Hastelloy B Ni-Mo-Fe
Hastelloy W Ni-Mo-Fe
INOR-2 Ni-Mo-Cr
INOR-8 Ni-Mo-Cr-Fe
Hastelloy X Ni-Mo-Cr-Fe -
‘OFHC Copper Cu.

Pt

.4

-

" tTable XVIII. Nominal Composition of Corrosion Specimens from the Volatility Pilot Plant Fluorinator ’ P

Nominal Composition wt %

Material Ni Fe

Co Cu

AL Mn~- Pt Ti Mo Cr C W vV Zr S Si P B
L Nig - 99.47  0.11 . 0.17 . 0.023 0.0075 < 0.01
INCO 61 '93.00 1.0 0.25 - 1.5 1.C 2.0-3.5 0.15 0.01 0.75
weld wire (min) ’
Au-plated . ,
L NiP Gold plate = ~ 0.0015 in.
D Ni 95.0¢% © 0.05 0.02 L.75 0.1 0.005 0.05
OFHC Cu : 99.9* ' '
Hymu 80 79 . 16 0.50 4.0 0.05 0.15
. _ (bal) -
Cobanic 55 ks 0.1
INCO 700 L5 1.5 29" 2.5 2.3 3.0 16 0.08 -
80 Ni—20 Co? 79.59 20.13 - 0.026
90 Ni—10 Co2 90.02 9.63 0.020 ,
Monel?® . 67.66 1.40 . 0.47 30.3C 0.93 . 0.19 . 0.01 0.11 ‘
Inconel 7€ "7 0.1 0.20 15 0,06 - . 0.007 0.2
Hastelloy 48 18.76  1.01 0.64 8.82 21.81 0.11  0.20 0.008 0.78 0.011
) (bal) : ' ) . , -
Waspalloy 57 2 12-15 0.1 1-1.5 .5 1 2.75-3.25 - 3.5-5 18-21 0.1 0.1 0.03 0.75 0.008
(bal) “(max) max ) (max) (max) (max) | (méx) (max)
Hastelloy 65 5 - 28 0.4 0.1 '
B
Hastelloy 60 5.5 2.5 1 25 5 0.12 o.‘6 .
w_ .
.INOR-2 79 : 6 5 0.1
INOR-84 9.8 5.1 . 0.83 0.08 16.5 6.9 0.08 0.005 0.14 0.008
Nimonic 80¢ 74.28 0.6 o.0b  1.k2 0.5 2.k2 . 20.38 0.05 0.007 0.28
Platinum -99.9% ’

aORNL Chemical analysis: <bBase metal same as L Ni asbove. SVendor chemical analysis. dWestinghouse chemical analysis (Heat 8M3).

€Includes cobalt.

cll
- 113 -

B. Test Method

. The scouting specimens consisted of lengths of rod, sheet, or split
pipe stock, or combinations thereof, each approx 48-56 in. long. They were
held in place in the fluorination vessels by two methods. During the Mark I
fluorinator's operations, all specimens were welded to the top blind flange of
the vessei as shown in Fig. 3. When the Mark II vessel was in use several
specimens were welded to the inspection port flange in a similar manner while
the remaining test rods were held in place by metal connectors gripping a fit-
ting prewelded to the end of the test rod. Tfie end fittings were A nickel
tubing l/2-in.-o.d. X approx 3 in: long and seal welded to prevent the escape
of process gases. Figure 52 shows a cut-away view of the Mark II VPP fluorina-
fior illustrating the placement gf corrosion specimens and a typical test
specimen prepared for insertion.

A tcfial of 31 test specimens was exposed to the fluorination environ-

ments. In each test grouping at least one L nickel specimen was included as a
control. Figure 53 shows a top view of the fluorinators indicafiing the placement
of the specimens in terms of polar geometry, the testing ofder and identification
number, and corrosion specimen numbers,  The corrosion specimen number identifies
particular fluorinator runs when specimens were in place and also indicateskthe
total mumber of specimens tested simultaneously. The major test groupings were
fluorination runs M-21 through M-48, C-9 through C-15, E-3 through E-6,
L-1 through L-4, and L-5 through L-9. Figure 54 details the process cycling
for each run group. Tables I and VI summarize the other process details during

these scouting runs,

‘C. Reactions to Environments

\ ~After exposure, the test specimens were rémbved from the fluorination
vessels and subjected to dimefisional analyses and metallographic study.
Figures 55 and 56 summerize the maximum corrosive losses found in each major
exposure region as determined by micrometer measurements and optical microscopy.-
Fignre 535 details the loss data as a plot of mils per month, based on individ-

ual molten-salt residence times, while Fig. 56 shows the loss data in terms of
- 114 -

UNCLASSIFIED
ORNL—LR=DWG 49167

'SPECIMENS WELDED DIRECTLY TO TOP FLANGE

TUBE PRE-WELDED TO SPECIMEN

) TO FIT THREADED FITTING
INSPECTION
MECHANICAL 17~ PORT
' THREADED i ~
FITTING : —
' l"”]l(//f///
i
|
EI
!
I
\ i :
N 1
17 R
PIPE‘\\\fé // , _ ; .
' / % ~ TOP ,' { °
an % N FLANGE ! I
. = . . \..:' i Iy ) R
BN @ T P
\ ' | . .
. . ‘ ’ | I
| SEAL WELD L
‘ “ |
SPECIMEN "N S
. ’ i [ 1

~ Fig. 52. View of Mark IT VPP Fluo,rinatofl_Ves,se-l Showing Methods of
Corrosion Rod Rlacement. ' ' ' .
’

- 115

Pl

DIFFUSER \
CONE —

UNCLASSIFIED
ORNL-LR-DOWG 49168

SCHEDULE
Order of Test ond Identification Corrosion Séecimen No.* Material
No.
1 M21-M48:1 “L"" nickel rod
2 M21-M48:2 "L nicke! rod ki
) 3 M21-M48:3 ~*L" nickel rod
4 M21-M48:4 “'L"" nickel rod
5 C9-C15:1 INOR-2 rod
[ C9-C15:2 “L"" nickel rod
7 C9-C15:3 ““L'’ nickel rod
8 C9:C15:4 Hastelloy ‘B’ rod
9 E3-E6:1 Inco-61 weld wire
10 E3-£6:2 nconel tube
n E3-E6:3 Monel rod
12 E3-E6:4 80Ni ~10Co rod
, 13 E3-Eé6:5 L' nickel rod
14 E3-E6:6 90Ni-10Co rod
15 "E3-E6:7 OFHC Cu rod
16 Eé Platinum
17 L1-L4:) INOR-8 rod
18 L1-L4:2 Au-plated **L"’ Ni rod
19 L1-L4:3 Inco-61 weld wire
20 L1.L4:4 Ince-700 rod
21 L1-L4:5 Cobonic
22 L3-L4 Inco-61 weld wire
23 L1-L4:6 “L" nickel rod
24 .LS-L9:1 ‘D" nickel sheet
25 L5-L9:2 INOR-8 rod
~ 26 LS5-L9:3 “L"" nickel rod
27 L5-L9:4 Wospalloy rod
28 L5-L9:5 Hastetloy "* X" pipe
29 L3-iv:é Mostrelloy W' sheer
30 ) L5L9:7 Nimonie '80"" pipe
31 LS-L9:8 Hymu *'80" rod

*Example: M21-48 refers to runs when specimen was exposed; {:1) refers t
is, four specnmens rested simulroneously during the "'M*’ runs.

o local group number, that

Fig. 53. Top V1ew of VPP Fluorinators Showing Placement of Corrosion

Test Specimens, Testing Order, and Specimen Number,
TEMPERATURE (*C)

UNCLASSIFIED ’

ORNL-LR-DWG 49202
] '
700 * 700
\ T
¥
600 Nl A | _ I a0 B fa Lt tMal 1 a —L L
. ) . L 1
L_‘LT__L:’_LL__ - “lv d 1. \ | SALT MELTING "’O‘I—NJ_'_H—'_‘_" r" R
500 [t MELTING PT 500 B — | .1 1 Y ;
1
RUNS M21-M48 \ . ¢ I RUNS C9-CI5 - \ T
400 SALT 1 NoF -2¢F, u 400 SALT: NoF -2¢F4 - UF, [N 0
. N°"-'“°‘i‘7~‘5°‘55°sml o \ [ \ 2 NOM.mole %' 48 - 48 - 4 . Y 3
Fa INPUT 716,77 iters, 10 hr g Fo INPUT: 31,850 STD.liters, 37 hr !
300 : 0§00 2 i el 1> 2
" 3 = . 1 I
200 | H \ oy & \ II L o3
S . . (BT} W
w Rt &
N t o S
00 . { fo £ 00 (s °3
H E LI ' )
P
0 0100 - 0 - , U
o 100 200 . 300 400 500 4] 100 200 300 400 500 600
' TIME {hr) TIME (hr)
. |
AtA T
- 700 200 2 A
Tl T A a A : T M f\
N T T w—1 nat
600 {{/ A ;"J‘ e 600 N —
sart L,. Twervepovy. IV Wb (W VL .
i H =1 — .
TING PT . .
500 MELTING PTy 5 500 :
: [} .
g \ i ¥ \
400 RUNS €3-€6 L —— S @00 - - ‘
E . SALT: NoF-2rFa-UFs T RUNS L1-L4
a " NOM. mole % : 48-49.5-2.5 & SALT: NoF - ZrF - UF,
§ 300 . Fp INPUT 19,650 STO.lders§ 20 _ % 300 NOM.mole %: 84 - 4l - 5 30 _
= A ] = - Fg INPUT : 16,250 STO iters, 34 hr ]
= \ 2 2 )
200 : 4 03z 200 20 3z
|| II \ ' 7 ‘ Ii‘
b g | . g
w
100 \ —{ w0 2 100 t - 0 Z
| ' § . S
: Hl 4 '
. 1 I l w o -
o c o
o 100 200 300 ) 100 200 300 400 500 600
TIME (Pe} . TIMF thr)
700 |- =83 ' ’
thg . . .
N, T oA | AT ATal A | - LEGEND . -
600 A-FLUORINATION SALTS AODED TO FLUORINATOR
LT - - b
U ¥ l mJ - | MELTING U U L"-J § Ay-3609 CrFy 3.5 Hp0 ADDED
500 17 POINT _ Az-10mg PuFy ADDED
m | . A3-19 PuFq ADDED
w . A4-10¢ PuF4 ADDED .
B 900 T-FLUDRINATION SALTS TRANSFERRED OUT OF FLUORINATOR
& ) \ \ \ \ § - FLUORINE SPARGE ¢
¢ 300 . RUNS L5-L9 30 _ 0 - NITROGEN SPARGE
SALT 1 NOF - 2rF4-UF ¢ z X _
,‘:‘ NOM. mole 7ot 52 - 45.5-2.5 1 HEIGHT OF BAR » FLOW IN {SLM), STD liters/min.
. A 5-2.5 2 ¥
£, INPUT 116,000 STD. liters, 18 br 42 WIDOTH OF BAR » FLOW [N hours . ,
200 i 20 & NOTE:NITROGEN SPARGED PRIOR TO EACH .
. " “ \ @ FLUORINE SPARGE ,NOT SHOWN . :
- - .
z
100 N ) v—© 2 .
| . g
-
o o -
o 100 200 300 400 500
TIME ()
‘
Fig. 5k. Comparison of Process Details for Various VPP Run Groups When
i B :
Corrosion Scouting Specimens Were in Place. -

)

91T -
) . i ., 8
! [
T
. i -
~— - . ~ '
’ ' UNCL ASSIFIED
ORNL-LR-DWG 37704
- S
\ ) ’

Ay o Dsctnon ) . i -

(KX W NICKEL v e . ; ////// AL //."////////////////.’//////////,
We N, 100 s g w0 e RSSO \\\\\\\\\\\\ \ SN \\\\\\\\\\\ Y

S f/////// // // A 1

¢ 'I'/'-‘"[:":; SHCIMEN N, 21 - il 8 VPECIEN COTRODED TO & SO * 74, ABDYE THE STATIC SALT LEVEL

+ "L NICKEL B ¥ _
VN, 100 SHECIMIN NO. CF - CI5i 2 - _ . - %

] " NICKEL il | | l I I | l
VN, 100 SPECUMEN NO, €9 - C15,) INTERGIANULAR ARD TOTAL SPECIVEN COTRESION 407 RECONCED, CORPARABLE 10 SPECIAN ND. 89 - £131 1

” L MICKEL
4 N, £0D SPECIMEN NQ. B - U4 3 \\\\\>\\\\\ EEETS: \\ \\\\\\\\\\ SN

* 2 ° me kR I
~ V4 IN. 100 ! SPECIMIN NO. LI - 14 4

n "L HICKEL VA | 4 J

174 1N, 100 ¥ SHQIMEN NO. L3 - L1 ) §\\\\\\’\\\\\\ S
[ / I :
T} | I I I 1 I

* ;"_;ofiot;l:' :‘;:’ e SPECIMEN NO. E3 = Eb: | 0 ! - ! = 1 l

: e s o 1T T T T T T 1 T T |
V2 IN. DlA. k00 . . Y I

" INCO NO, 41 WILD WIRE -
VA IN. Dis, ROD SPLCIMIN NO. L1 - 14,3 ND SPECIUEN CORROOLE TO & BOINT AT THE INIERFACE

'

3 INCR2 i ~

W4 IN. i, 80D SHCLAIN NO, CF - C1%: | 1 i
1 L} RASTELLOY "% WPECIMEN CORRODEE 10 4 POINT ~ T2 1N, FROM un'?c LTLE/EL
V4 IN. DA 500 SMCIMEN NO. CF - C15i ¢ Na
- 1 INGONEL

'/u‘e‘:'u"l: ml SHCIMIN NO. 83 -0 2 . SPECInEN CNITODED T3 4 POWNT &7 THE INTERFACE

" MamiL SMCMEN NG, E3 - {613
141N, DA, 100 . " Ty l SPECINEN CORFODER T0 4 POINT AT THE NTENFACE

.
. 12 i - 20Ce

SPECIMEN NO, E3 - B 4

- .2, O1a, 00

ta 0 N - 10Co
rs 0.1 IN. DIA. 10O SPECLMEN NO. [ - Ed1 4

—
(1] OF HC COPFER PPICLEN CORTIDED 1O & FONT - 1300 SROVE THE SHYF BALTIEVER
. V4 4, O, 100 SHCIMEN ND. £+ £ 7 na
SPECMEN CORROOED TO & POWT ~T0 (N, ABCYE TnE STATIC

g MLATINUM SPECIMEN NO. & Na o

0,08 IN. DLA, WiRL Na
” " NICKEL ! .

[N N NO. U -1t ) — —

0.185 IN. SHEET IMCMINND. L -th ) 5 i * — m| +
T} Au MATED “'L** NICKEL I I | I I I l l

0.0013 1N, WATE ON SPLCIMEN NO. 1) - 14 2 ; : - o -

1/4 IN. 40D : .
n INCO 700 N } SPECIMEM EXPOIED ONLY IN SAL T PrasE I l I | | | | 1 I | | I
tafN NG, LY =LA 4
~ VN, DIA, 100 e ' L - — - -
n COBMNIC SATIMEN ND. LI - 141 3 I l I i I I
NUSt
8 STRAND IRAID OF NA Y hEClurn E1POLED oLt TO YAPOR PruasE . -~ .
. - . 0,05 (N, Dla, Wil LT} W OUR SPECIMENS TESTED IN THIS UM GEOUP,
I INO®-8 Data FO8 SPECIMEN MIT - .M‘.l 1 TYPRCAL OF SPECIMENY 1, 1, ARD D,
o SHCIMEN NO. U1 = 4 ) I A\ .

1/4IN. W, ROD

I ' I l ‘ i UGIHD
- YAPOR r
J ~ VAPOR $ALT INTERFACE

» INOr4 T
RN, O, ROC

SHCUMEN NO. LS -1 2

sanf
AN NOT avaiADLE
n o7 ARSI mee—] ’
l : . INTERGRANULAR ATTACK ~

° 1oTAL 1088 - |
‘ ‘[ J l MAKIMUY LOSS SHOWN [N EACH PHASE

" - WASPALLOY
/2 IN. DLA. 10D SECTIOR &

E“‘-(

- <c|n-<je-cfn-2|n-cju-<c|lo-<|w-<|u-c|lu-<|vw-<|n-<|v-<cjlu-<|n-c|u-c|lu-<|u=-c|u-<c|lu-<c]u-<|o-<|lo-<c]u-<|u-<c|un-<cju-<]|n-<
- - ) Tz
> -
[ \
-

SHCIMEN NO, LS - Lt 4

-hn HASTELLOY X
0,12 1N, WALL PHPE SICTION

SMCIMEN NC. L3 «L%: 5

” HASTELLOY W
/3 N, SHEET SECTIONS

SMCIMEN NO. 15 17 4
-~

» NIMONIC "0
N NOQ. -2
Toan N, walL TN TECMIN MO L in
. WALL SECTIONS

SPECMEN COPROTED 10 & POINT « L) IN. ABOVE TWE STATIC B4LT LEYEL
1

n HY MUY "
4N, DA, ROD SECTIONS

$HCIMIN NO. LS - 17td

[} 5 10 5 20 25 30 35 20 43 33 55 60 (1]
mifs £ mi
{BASED ON MCLTEN SALT l’lME)

Fig. 55. Summary of Corrosion on Specimens Exposed in VPP Fluorinators. .
- —
. . A -
- ’ . -
. . 1 . - -
. . .
.o ‘.
. DA Lt 2 2 B . .
. N .
. \ v ? . : UNCLASSIFIED
) ORNL-LR -OWG 3938t
. .
. ; .
. . .
IDENTE ICATION
NuMaTes oLsarnon -
12, L NICKEL '
Vi N, 100 SPCIMEN NO. 41 - Med: 10
4 AT NICKEL SMCUAIN NO. M1 + Ml -‘ Y SPEQMIN CORRODLO TO 4 POINT * 77 ABOVE THE STATK SALTVLEVEL
8 - gy N& t :
1/4 IN. 0D S| Na - i . . -
¢ e ekt SPCIMIN NO. €V - CI%i 2 h ’
V41N, 100 M R ) .
7 .M NICKIL v -
Ve, 100 SPCUAIN NO, € + CI% 3 = PERANULAR AND TOTAL SPECINEN CORACHION NOT RECOR0OED COMPARARLE TO SPECIUEN MO €Y - CI3
- s [ .
13 ek vz
SPECIMEN NO. £ = E6: § N . .
\ /4 N, 200 H
n -t Mgk v .
. Va1, 20D SMCIMEN NO. L1 - L4 & ye
E - NICKEL vy .
V41N, %00 SPECIMEN NO. LS - L% > ‘
c - - S .
' ICO NO. 41 WELD wits v
SPECIMEN NO. ED - Eé: | [}
VI N, DA, kOO s 1
z INCONO. Y WELOWRE o e M- .
VR IN. DAL 100 . s - .
" INCO NO, 4} WILD wirg vi i - A
V321N, 0w, 100 SPRCMIN NO. L8 =143 H P LOMN CORRODLD 10 4 PO AT Tf NTLRS ACE
3 Nor2 v '
VAN, A, 100« smeminno, er-crnt [ 1 | ) .
y . HASTELOY *a™ MCIMEN NO. €3 + 1 “' - TO A POOIT ~ 271N FROM $TATIC MLTLEVEL
. MEN . +«Clhe N Ay N - =
V4 IN. DA 100 sl na , . . :
1 INCONTL . Yl AR R T N LA T TN » .
1/4 IN, DIA, T\BE, SPCIMIN NO. €1 -t 2 ' PUOWEN CORRDDED 10 A POST AT THE INTERFAC
0,002 IN. WALL S| Na . . .
: " mMONIL ) Vi . ) . o
; < /4N, DA, €00 SPECIMEN NO. 63 - Eh ) 's PP EGuLN CORROOLD 10 & POWT 4T TNE INTEF7ACE .
. . NA . )
” ®N - 10Ce v ‘ .
SPECIMEN NO, D = th 4 ' )
, 0.2 1N, OWA, £OD sk 1
" 0N -10Cy . wtewanno. 0 ws |1 i '
. €+ th .
0.2IN. DWA., KOO s 1
1 orc comn v - T TOX CORTODRD, 10 £ 700N T =13 &% LBOVE T STATIC BT LEVIC
Va N, DA, £00 SPECIMEN NO., € - b 7 1
~ o : , ™
: v WECQMLN COf| D104 T~ DN 4l THESTATIC SALT LEVE! ‘
" MATINUM SPECIMEN NO. B v R00€0 10 4 PO oV
0.08 IN. DA, wizt < . ¥
7 D" MICKEL ‘ v
SMCIMEN NO. 15 ~ 19 | 4 .
0,103 IN. SHETT s [
'
" ‘aw MATED 1 NICKEL RGN O U1 - 12 v
N NO. Ll = Léy e
0.0003 IN, RATE ON —~
1/4 IN. 200 s T ; ENCKI %7/,
.
» INCO 700 HECIEN N0, L) < Lo 4 "‘ :: LOMDN E1RCID LT D4 BLT Prsd
AURTH : .
Y4 IN. DA, r00 s e PN N T P 27 S T e
n COMNIC v
. ¥ STIAND BIAID OF SPECIMIN NO. LT - Lt § : s 10 vare st
0,025 IN. DA, wite “FOUR SPECIMENS TESTED IN THIS RUN GROW.
P NOu-e v \ \ DATA FOR SPECIMEN 21 - a3 1 I3 TYPICAL OF SPECIMENS 1, 2, AND 3.
Ve in. bua, €00 L LUURICR I \ .
” v LECEND
roo SPECUMEN NO, LS - 1% 2 B v . vapom
. VB IN, OWA, s | - VAPOR SALT INTERFACE
. s -t
” WASPALLOY v . N t AVATLABLE
. U2 o0, o €00 secrions  TICWIN KO- 132104 ' . -
. . DA, THON! sl . wetar boss - [T -
— " INTERGRANULAR st 14CK - 227227777777
n MASTELLOY "X " .
0.12 N, waLL PRE SECTiON SOOI NO- L3 A : TORAL LGIS «
s T . MAXIMUM LOSS SHOWN [N EACH PHASE
’ » sty SHCIMIN NO, LS = L 8 v PERY ‘
1/3 1N, SHEET SICTIONS " s —
© NIMONIC " 90~ v POt COTROOLO 10 & PO T =
Clwrs 3 - L?
0.1 IN, WALL THICKNESS  PTUCIMEN NO. 3 <1007 ' .
- WAL HCTIONS s [ e R N T e e T A ST e
g v
n YA 90" . oy MECDINNO S s ] . I
B /8 IN. DIA. €O SECTH s

[-4

0.8 06 08 - 1.0

‘e 16 (2] T 20 2.2 2.4 2.8 28 30 32 3.4
mils/hr
{BASED ON FLUORINE TIME) 14

Fig. 56. Summary of Corrosion on Specimens Exposed in VPP Fluorinators.

4

- 119 -
9\

mils per hour, based on operational fluorine sparge time. In both figures
dotted and cross hatched bars represent bulk metal losses and visible inter-
granular attack,'reSPectively, while solid bars indicate a total loss of the
specimen occurred at the point indicated, ‘
Several specimens registered lower rates of maximum corrosive attack
than the L nickel control rods.. These specimens included Hymu 80, INOR-2,
INOR-8, Hastelloy W, Hastelloy X, Waspalloy, and the 90 wt % Ni—10 wt % Co
alloys. In particular, Hymu 80 was superior to all test specimens in all runs,
althdugh the specimen registered a maximum bulk loss rate of 11 miis/month,

based on molten~salt residence time, at the vapor-salt intérface. Total loss

of-"corrosion specimens, indicating corrosion rates of > 105 mils/month based

on molten-salt time, was found for one L nickel specimen, one INCO-61 weld
wire specimen, and for the Hastelloy B, Nimonic 80, Inconel, Monel, copper, and.
platinum specimens.

Selected photomacrographs and photomicrographs of samples from many
of these corrosion specimens are presented in Appendix A. In most cases, the

areas of maximum attack are shown.

D. Discussion of Results

A widé variation in resistance to the fluorination environment was
found on corrosion specimens used in the compatibility testing described. The
variations in corrosion noted were anticipated because of the widely divergent
proééés conditions. Because the VPP has essentially been|operated for.feasi-
bility studies, demonstratidn runs, and runs designed to recover uranium from
available fluoride salt mixtures, the determination of optimum process condi-
tions rather than the gathering of corrosion data has been the prevailing
philosophy. | )

In many groups of runs a wide variety of scouting materials was -
used which were in proximity.  This was recognized as being far from ideal for

corrosion testing, but little choice was available under the circumstances.
{

.However, Run Group M-21-48 did contain only the reference material, L nickel,
- 120 -

ana the results were almost as divergent as the data obtained in run groups
where as many as seven different materials were bresent,in the same system.

The L nickel reference specimen's mean rate of corrosive attack was
‘similar to thaf found on the walls of the corresponding pilot plant fluori-
nators, although iarge deviations from the mean existed. While maximum
corrosive attack occurred in the vapor region on the Mark I fluorinator and
infthe salt region of the Mark II Vessel{ in almost all cases the L nickel
éfiecimens showed maximum attack at the point of vapor-~salt inferface contact.
A notable exception was specimen M-21—-L48:4, tested in the Mark I fluorinator,
which exhibited failure in the upper vapor region. The addition of chromium
fluoride during the M-21-M-48 Run Group may have produced the vapor_region
failure thréugh mechanisms proposed in Section I.

| . The increased attack on the L nickel spe¢imens at the vapor-salt

1nterface possibly resulted from salt agitation at. the center of the vessel
1nduced by the fluorlne and nitrogen sparging through the draft tube. ThlS
agitation could result in 1ncreased erosion and vibration of the spec1mens
with resultant Jloss of protective fllms In addition, the location may have
allowed primary.contact of the specimens with the fluorine sparge before the
xlattef reached the vessel walls. However, the increased corrosion to be ex-
pected from this effect was not noted on the A nickel fluorine inlet tubes,
the draft tube walls, or other fluorinator internal components discussed in
-Sectlon IT. '

Tests on INCO-61 weld wire were intended to provide data on a nickel-
rich alloy with additions not readily available in commercial alloys. Besides
nickel, this material contains small amounts of Al, Ti, Fe, Mn, Si, and Cu
(Table XVIII). Since this material was used in fabricating the fluorinators,
these tests provided an indication of weld metal behavior, although the weld
metal after deposition would have a cast structure compared to the wrought
structure of the weld wire. The INCO-61 specimens showed generally analogous
bulk metal loss behavior when compared to the L nickel specimens. However, no

intergranular attack was found in the INCO-61 specimens.
5

- 121 -

The behavior of the gold-plated L nickel corrosion specimens indicated
that no protection was provided by the plating. Comparison with L nickel speci-
men number L-1-L-4:6 in Fig. 55 shows comparable losses.

’ On the basis of single exposures of the 90% Ni—10% Co and 80% Ni—20% Co
specimens; cobalt additions seemed to improve the resistance of commercial

nickel to the fluorination environment with the one exception of the high vapor-

_salt interface attack found in the 80% Ni—20% Co specimen. The cobanic alloy,

containing 45% Co, was tested only in the fluorination vapor phase and showed
greater losses when compared with L nickel. Thus, further investigation of
nickel-base alloys with less than 20% Co seems warranted. Because the most
corrosion-resistant Ni-Co alloys are experimental, the investigation should
ultimately include the determination of mechanical and physical properties and
fabricability.

The nickel-rich copper specimen, Monel, provided poor resistance to
the fluorination environment during a single test run. The specimen failed
completely at the vapor-sait interface.

The D-nickel, containing about 5% Mn, and developed for improved
resistance to sulfur attack in oxidizing atmospheres at elevated temperatures,
presented somewhat higher loss rates than the L nickel_sPeéimen simultanecusly -
exposed. The D nickel specimen presented a different geometry to the corrodents,
sheet vs rod, but that difference is not believed to have been significant in
producing the higher rates of attack. |

As expected, the Nimonic 80 alloy, containing 20% Cr and approx
2.5% Ti, exhibited substantially greater bulk losses than the corresponding
L nickel specimen, Reference to Table XVI shows that chromium and titanium
can form highly volatile fluorides at their higher valence states.

A high-temperature, oxidi;ation—resistant, nickel-chromium-iron alloy,
Inconel, has been mentioned as being useful in contact with fused fluoride salts.
During the single test in the fluorination environment, the specimen completely
failed at fhe vapor-salt interface, and also exhibited a rather high vapor-phase

attack.
- 122 -~

The Ni-Co-Cr alloys, Waspalloy and INCO 700, differed significantly
-in attack, Samples of INCO 700 were available only for a salt-phase test
(L-1-L-4:4) and in this region had a bulk loss rate > 100 mils/month which
considerably exceeded the L nickel specimen exposed sifiultaneously; The
Waspalloy specimen showed a rate of attack slightly less than that of a cor-
responding L nickel specimen. Further investigation of Waspalioy seems war-
ranted, although it should be remembered that the alloy ages at 750°C, a
temperature only slightly higher than the fluorinator operating range. Thus,
temperature excursions, which may be encountered in extended operations, might

_drastically reduce the ductility of the material,.

I

~

The Hymu-80 specimen, Ni-Fe- Mo, presented the best resistance to
corrosion of any of the specimens tested including the L nickel reference
material. This alloy is cofimercially available and six miniature fluorinstors -
‘have been fabricated for compatibility testing with NaF-ZrFu énd LiF-NaF-ZrFu
salt mlxtures by the Volatlllty Studies Group of Chemical Development Section A.

The essentially Ni-Mo-Fe alloys, Hastelloy B and W again showed the
‘variant behavior characteristic ‘of this test .series. The Hastelloy-B specimen .
had extremely poor resistance and failed completely 22 in. above the static .salt |
ievel. Since the specimen was exposed during the Mark .I operation, the high
etapor;phase attack evidenced at that time may well ha&e influenced the behavior
of Hasfelloy B. The Hastelloy-W specimen, in place during a relatively low
corrosion run series, L—B—L—9, exhibited slightly lower losses when compared
with the\L nickel control specimen. This was because of the intergranular
attack experienced by the L nickel. |

One of the early fianifestations of the Ni-Mo alloy series developed
at ORNL for fused-fluoride salt use was INOR 2. The‘sing}e INOR-2 specimen
| tested had corrosion resistance superior to either of the L nickel specimens
exposed concfirrently. Lack of additional material has hindered further corrosion
testing. | | -

The Ni-Mo-Cr-Fe alloys tested, INOR 8 and Hastelloy X, hed corrosion
resistance comparable to that of the L nickel rods exposed simultaneeusly. As
deec?ibed in Section ITT, miniature fluorinators were fabricated from INCR 8

and tested. These tests did not indicate any superiority over commercial nickel.
- 123 -

A single copper specimen, L-3-L-6:7, was exposed to the Mark II
environment. The specimen failed completely in the middle vapor region. Cu--
pric fluoride was considered most likely to form under the excess fluorine

conditions present during testing. Although CuF, has a melting point of

approx 950°C and should have afforded protection?to the parent metal, copper's
lack of resistance to oxidizing environments at fluorihation temperatures may
explain its poor performance.

The platinum specimen was completely severed by corrosive attack in
the upper vapor region. This wire was intended to serve as an electrode probe,
but never carried current. The reason for the extreme rate of attack on
platinum seems to be the instability of PtF6 which has been reported to de-
compose at 280°C (Table XVI).

Whereas the L nickel specimens exhibited intergranular attack, very
few instances of this mode of attack were noted on the other specimens tested.

Sections I and II describe thé mechanisms presumably operating on L nickel.

E. Future Studies

The corrosion rates for all scouting materlals tested to date appear
to be excessive for long-time use. Therefore, additional fluorination scoutlng
corrosion tests of the type described‘\have been planned. Particular ¢mphdsis
has been placed on binary nickel-rich alloys containing various amounts of
Fe, Co, Mn, Mg, and Al. The latter two elements have not been considered in
previous experiments because of the fabrication, age hardening, or other
difficulties characteristic.when alloyed in appreciable quantities with nickel.
Only a few nickel-base commercial alloys containing magnesium and aluminum in
combination with other ingredients have been available and still fewer have
been commercial as Ni-Mg or Ni-Al binaries. Yet, aluminum and magnesium are
known to form fluorides at their highest valencg states, which have higher
melting points than NiFé. |
To date, Fe, Co, Mn, Mg, and Al in the quantities shown in Table XIX

have been added to induction melts of nickel and cast into l-in. rounds. After

suitable- - homogenization treatments of the castings, the rounds were cold swaged
- 124 -

Table XIX. Proposed NQhCOmmercial Binary A;loys-for Corrosion Testing
in the Volatility Process Fluorination Environment-

Nominal Composition.

| . (vt %) . |
Alloy Group - Ni .Fe Co Mn Mg Al
NiFe N ‘ 95 >
| 90 - 10
| 80 20
“Wico S 95 5 ’
L 90 10
NiMn | : 98 2
- NiMg | 199.95 o 0.05
99.9 ' - 0.1
99 , 1.0
MAL o o o

97

- 125 -

to l/h—in.-diam rods and hydrogen annealed. The rods have been cut to proper
lengths, end fittings attached by seal welding, and are presently awaiting
placement in a VPP fluorinator.

1

In addition, high-purity vacuum-melted nickel has been obtained and

fabricated into the proper test specimen shape for examining the resistance

of that material to the fluorination environment.

V. Argonne National Laboratory Fluorination'Corrosion Studies

The Chemical Engineering Division of Argonne National Laboratory has

b7

done extensive work on nonaqueous processing of irradiated fuels. A portion
of their effort has been on studies of fluoride volatility processes. A re-
view of one gf the Argonne National Laboratory studies on materials compati-
bility in a simulated fluorination environment is included here for comparison

with ORNL data.

A. Test Method

Fluorination corrosion tests were conducted on coupons of L nickel,
D nifikel, Hastelloy B, and INOR l.(ref h8)_ The coupons were contained in a
vessel which had an inner liner and internal fiiping fabricated from A nickel.
The coupons, each 8 in. long x 0.5 in. wide x 0.032 in. thick, were wired to-
gether to form a simulated draft tube of square cross section. The composition
of the wire was not indicated. The nominal composition of the first three
materials listed above has béen gfven in Table XIX. Another of the early ex-
perimental alloys, INOR 1, studied in the development of INOR 8, had a nominal
composition of 78 wt % Ni—20 wt % Mo—0.5 wt % Mn—0.5 wt % Si—0.25 wt % Fe—
0.01 wt % C. "

The simulated draft tube was partially immersed in & bath of equi-

molar NaF—ZrFu so that the vapor-salt interface was approx 3 in. up from the

. J+7R. C. Vogel and R. K. Steunenberg, "Fluoride Volatility Processes for
Low Alloy Fuels," Symposium on the Reprocessing of Irradiated Fuels Held at
Brussels, Belgium, May 20-25, 1957, Book 2, Session IV, pp.498-559, TID-T753k.

8L. Hays, R. Breyne, and W. Seefeldt, "Comparative Tests of L Nickel,

D Nickel, Hastelloy B, and INOR 1," Chemical Engineering Division Summary

Report, July, August, September, 1958, ANL-5924, pp.49-52.

- 126 -
bottom of thé coupons. The bath was held at 600°C while fluorine or helium
was introduced centrally beneath the melt surfaces at rates of-approx 0.1
‘standard liters/minl During this test series, the fluoride salts were kept
molten for a total of 216 hr and for 63 hr of that time fluorine was sparged
- into the‘bath. - The process gases were introduced on a cyclic basis, i.e., -
for 7 hr/day fluorine was sparged while helium was.sparged for the remaining
17 hr of a day. |
" F;gures 57 and 58 illustrate the corfosion losées obtqined during

this test series. Figure 57 is a bar graph where bulk metal iosses as
~determined by micrometer measuremernt and additional losses as determined by
‘metallographic examination have been converted to mils/month of molten-salt
residence time. Figure 58 is a similar graph plotted as mils/hour of fluorine
sparge time. Rates haveubegn filotted insfead of original data for comparison

with ORNL @sta. '

\

B. Discussion of Results

The corrosion losses in fhis Argonne NationalaLaBoratory test series
suggest that INOR 1 was the most resistant to the test enyironment, while the
L nickél reference material fared poorly. This is attributed'tolthe extensive
“intergranular attack on the latter material, particularly at the vapor-salt
ifiterféce. |

Comparison of the Argonne National Laboratory corrosion results with
those obtainéd\at ORNL ;s difficult because of differences in test. procedures
which ihclude fluqrine flow rates, specimen locétion and geometry, and salt '
.compositions. However, Argonne's results on L nickel fit the corrosion
limits determined for the ORNL scouting corrosion refereqce'specimens in the
salt and at the salt-vapor interfaces. Vapor-phase specimens‘are nét readily
comparable,.since Argonne specimens were only 3 to 4 in. above the salt -
surface, ‘ ‘ ‘

The Argonne National Laboratory Hastelloy B and the D nickel specimens
genérally had less attack than scouting specimens at ORNL. The differences may
be attributed to more seve;e operating conditiops in the ORNL pilot plant
“fluorinator. Sinde.no INOR-laspecimen had been tested at ORNL, no comparisons

i
could be derived.
“ VE "*&Ezzzzzzzzzzzzzza ] [ B ]
cweker T B . \
\ ] f ARt , =7
| T\ "" """""" I ”" -
o nickeL |1 [ Mmmmwmmmwwwmwmwmmmwmmm
L 2
TVE Ew :
HASTELLOY B & R
, ] S
INOR | ) ~
5 EGEND
V- VAPOR‘
| - VAPOR SALT INTERFACE
S - SALT

] - METAL LOSS/MICROMETER MEASUREMENT

7~~~ - INTERGRANULAR ATTACK/METALLCGRAPHIC EXAM

(I - suBSURFACE vOIDS/METALLOGRAPHIC EXAM
MAXIMUM LOSS SHOWN IN EACH PHASE

| S N

0 5 - o 15 20 25 30 65 70
mils/mo
( BASED ON SALT RESIDENCE TIME)
Fig. 57 Summary of Corrosion on Couporns Exposed to & Fluorlnation Environment

v by “the Argonne Natlonal Laboratory
- : UNCLASSIFIED
ORNL-LR-DWG 49170

TP | T 1 [
e T P )

. D NICKEL |

HASTELLOY B |1 [ ar

INOR

_88'[..

LEGEND
V-VAPOR

i - VAPOR SALT INTERFACE
S-SALT

sy - METAL LOSS/MICROMETER MEASUREMENT

777 - INTERGRANULAR ATTACK/METALLOGRAPHIC EXAM

(T - SUBSURFACE VOIDS/METALLOGRAPHIC EXAM

MAXIMUM LOSS SHOWN IN EACH PHASE
] 1 1 1 1

0 0.1 0.2 0.3
mils/hr
(BASED ON FLUORINE EXPOSURE TIME)

Fig. 58. Summary of Corrosion on Coupons Exposed to a Fluorination Environment
by the Argonne National Laboratory. |
- 129 -

Close cooperation has been maintained with the Argonne National
Laboratory on the selection and testing of candidate materials of construction
for use in the fluoride volatility process. In addition to materials research
as a means of containing the fluorination environment, Argonne National
Laboratory has suggested three other approaches. These have been the use of
(1) cold wall fluorination vessel, (2) a spray tower fluorination, and

L9

(3) lower melting salts. Further work on these approaches has been covered

in the Argonne National Laboratory Chemical Engineering Division Summary Reports.

VI. Supplementary Volatility Pilot Plant Equipment

The operation of the VPP necessarily included the use of a number of
auxiliary components such as traps for radioactive products, absorbers, valves,
fittings, fluorine di;posal systems, and piping. Each of these are.éubject to
various corrosive conditions including fluorine, uranium hexafluoride, fused
fluoride salts, etc. Thus, successful operation of the volatility process
requires that materials be selected approPriate for the particular conditions
of sérvice. Construction of the present system was predicated on the knowledge
available at the time and as operating experience has been gained these parts
have been examined to verify the ofiginal assumptions. In most cases the metal
selected appears to give satisfactory service and failure analyses have sug-
gested alternate mefials in the case of those which proved unsuitable. The

details of the findings of this investigation are presented in Appendix B.
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 corrosibn 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.

49

A. Goldman, Trip Report, ANL, to W. D. Manly, June 12, 1958,

o

S . | -~ 130 -

. Thanks are due to the _meny personnel of the Chemical Technology Division
working on the ORNL Fluorlde Volatlllty Process who aided in gathering process
data or in reV1eW1ng this report. .

Special thanks are due J. H. DeVan, E. E. Hoffman, and D. A. Douglas, Jr., h
for their constructlve sppralsal of portions of this report and to the
Metallurgy Division Reports Office for their patient typlng and careful

preparation-of ,the material for reproduction.

. BIBLIOGRAPHY - .

G. I. Cathers and R. E. Leuze, "A Volatilization Process for Uranium
- Recovery," Preprint 278, Paper presented at Nuclear Engineering and
Sc1ence Congress, Cleveland, Ohio, December 12-15,  1955; also printed

in "Selected Papers," Reactor Operatlonal Problems, 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,
Belglum, TID-7534, Book 2, ;gp.560—573, May 2025, 1957.

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

L

SNHWIDHIS LSHL NOISOHH0D ODNILNODS ddA A0 SHIVHODOHIIWOLOHJI

¥V XIANHIdV
THIS PAGE
WAS INTENTIONALLY
LEFT BLANK
- 133 -

UNCLASSIFIED
Y-27610

Fig. 59A.
Corrosion Specimen (Runs L-5-1-9:3).
circle indicates original size.

As-polished.

Interior circumference of white

Cross Section of Interface Region Sample from L Nickel
10X.

UNCLASSIFIED
Y=27A1R8

B R

T

| T
.mclnzs

BBE

EREI

:

-]
o

Fig. 59B.

k1

Portion of Fig. 59A.

kR

Note intergranular attack. Etchant: KCN_(NHA)QSEOS

Typical Surface and Microstructure.
200X.
Unclassified
Y-27289

Fig. 60A. Cross Section of Salt Region Sample from Gold Plated L Nickel
Corrosion Specimen (Runs L-1-I-4:2). Interior circumference of white circle
- indicates original size. As-polished. 15X.

Unclassified |

. : i A‘.-. . . é - ('T-
- ( o ~ 3 - . o T P o~
. p-

- o * J ¢ |

Fig. 60B. Portion of Fig. 60A. Typical Surface and Microstructure.
Intergranular attack. Etchant: HNoj—HéSOh. 200X,
- 135 -

Unclassified
Y-27291

Fig. 61A. Cross Section of Interface Region Sample from INCO-61 Weld
Wire Corrosion Specimen (Runs L-3-L-4). Interior circumference of white
circle indicates original size. As-polished. 15X.

Unclassified

.001

Y-27507 [
UuZ
003
S
w
hI—
(&)
z
007
008
.009
010
il
012
013
014
Ol5
018
017
018
- -
(o]
.
N

Fig. 61B. Portion of 61A. Typical Surface and Microstructure.
Etchant: Concentrated HNO3. 200X.
Unclassified
Y-3537¢€

Fig. 62A. Cross Section of TInterface Region Sample from INOR-2 Corrosion

Specimen (Runs C-9-C-15:1). TInterior circumfere {
original size. 85X. nce of white circle indicates

Unclassified |

1=-35293

Fig. 62B. Portion of Fig. 62A. Typical Surface and Microstructure.
Etchant: Chromium regia. 200X.
« 187 =

Unclassified
Y-35296

Fig. 63. Section of Vapor Region Sample from Hastelloy B Corrosion
Specimen (Runs C-9—C-15:4) Showing Typical Surface and Microstructure.
Etchant: Chromium regia. 200X.
- 138 &

Unclassified pg;
Y-35300 o]

Fig. 64. Section of Vapor Region Sample from Inconel Corrosion Specimen
(Runs E-3-E-6:2) Showing Typical Surface and Microstructure. Etchant: Modified

agua regia. 200X.
- 139 -

Unclassified |20z
Y-3 5298 003

R T T —

Fig. 65. Section of Vapor Region Sample from Monel Corrosion Specimen
(Runs E-3-E-6:3) Showing Typical Surface and Microstructure.

Etchant: HC.H OQ:HNO3:HCl. 200X.
- 140 -

Unclassified |oo
Y-35297
Fa e — | P

- - s < - :
“'i ® - g i ” ..;4:',:— } s
3 —'~'~—""‘“' —, sl o2 b . ot i C—— — -
- g — - - - ’
—mtm— — s
-~ ’ -

2 - - — .‘“—-&“ '__ el —
Bt 23 ey — 0 ~ . :
— ———— ' i Py . S b . —— - S5
IR o - :
- g i -
-—.. . - .
e —" = . c———— p—
_— . . -
— — pesime .
e s — - 0.
R R i i - ® —
—ay -~ g /q . ¥ _‘?
o = - P P
iz~ T e - e ; . -
— SR e i - s - -
e = - o
et - . . .
s e A = ® - — . =
o | — -t - = --
- -l - e
. o ‘ o
. e R
— - —— - -
i - —— e e c——
- ——— - _—'-’—'-T — -
e e -+ — —— — =5
- i3 ig - - — - -
- - & e . - - -
- peallag o~ o . . - ’ P o - > “

T

INCHES

l

Fig. 66. Section of Vapor Region Sample from OFHC Copper Corrosion

Specimen (Runs E-3-E-6:7) Showing Typical Surface and Microstructure.
Etchant: NHMOH: H‘202 . 200X.
- 18]

Unclassified

Y-26905

Fig. 67A. Cross Section of Interface Region Sample from 80 Ni—20 Co
Corrosion Specimen (Runs E-3-E-6:4). Interior circumference of white circle
indicates original size. Combination of grain size and etchant presents a
relief appearance because of light. Etchant: HNOB:HéSOh. 15X.

Unclassified

Y-26906

"y AR S e
/ = 4 /’ "y’ \ §
L ;Y U3 / K\\/“\—\ 018

Fig. 67B. Portion of Fig. 67A. Typical Surface and Microstructure.

Etchant: Concentrated HN03. 200X,
- 142 -

Fig. 68A. Cross Section from Interface Region Sample from 90 Ni—10 Co
Corrosion Specimen (Runs E-3-E-6:6). Interior circumference of white circle
indicates original size. Etchant: HNO3:HESOM.

Unclassified :;fl

Y-26908

Fig. 68B. Portion of Fig. 68A., Typical Surface and Microstructure.

Etchant: Dilute HNOS. 200X,
- 143 -

Longitudinal Cross Section of Vapor Region Sample from Platinum
200X.

Fig. 69.
Corrosion Specimen (Run E-6) Showing Typical Surface and Microstructure.

Cathodic vacuum etch.
- 1hh

UNCLASSIFIED
Y<27611

Fig. TOA. Cross Section of Interface Region Sample from D Nickel
Corrosion Specimen (Runs L-5-L-9:1). Interior of white lines indicates
original size. As-polished. 10X.

UNCLASSIFIED
Y-27778

A o <
ltuumm j
Sthiumy PR

!-- .

f\ Sl ds

Fig. 70B. Portion of Fig. TOA. Typical Surface and Microstructure.
Etchant: KCN;(NHu)gsgoS. 200X .
Unclassified
Y-27288

Fig. 7T1A. Cross Section of Salt Region Sample from INCO-T700 Corrosion

Specimen (Runs L-1-L-4:4)., Interior circumference of white circle indicates
original specimen size. As-polished. 5X.

P X
A~

4
'a -

&
SIS € N - . i A5 B ) »* X Ve A 2

: - g~ R (L 5 > 1::.’; v 7y ;\ f{’,/,»":i)i: S pr he ’ 7“/\" \\}"‘(’Q/;};;}( € |.ole
o L g » T YC A e \’E{({- ¥ RN - = e AT ..w\cu\@(-r e ~J'>>“.¢‘:’ ’
3 7, 4 ¥ g £ N g B 5 P a N a2, Rl Al S e 0N /ot
.o o "_"Ai'y; gyt % L P \ —y”‘sgg’/g‘" P, ..;*K__»'; ~ > %L& 5 “——k‘ "‘; wfim( qq\ {I' 017
{ »» 5 L nc = { — /| P > e = an3%s 4 4 \ A . b
BI=P <M g St WSSV TN WY S S S ARFLVAERS M\ IAN 20 ol oA A Y

Fig. 71B. Portion of Fig. 71A. Typical Surface and Microstructure.
Cathodic vacuum etch. 200X.
= 186 -

Unclassified
Y-27492

Fig. 72. Cross Section of Vapor Region Sample from Cobanic Corrosion
Specimen (Runs IL-1-L-4:5). Interior circumference of white circle indicates

original size. Etchant: Concentrated HNO3. 100X.
- 47 =

e,
L
o
G
o
0
0
o
—
g
)

Interior circumference of white circle indicates

o
polished

Cross Section of Salt Region Samples from INOR-8 Corrosion

Fig. T3A.
Specimen (Runs L-1-L-}4

15X

As-

original size.

.002

003

.007

008

.009

Ol2

o

AN

R

\ v

Q
Q
o
G
H
2
LR
S
R
B
S o
g5
£ w
2
e
=
,mm
g5
WC
Ih“'
. 3
P
I~ P
.E
o4
o
=
G
@)
o
O
o
¥
(&)
fo¥

Fig. T3B.
Showing intergranular attack.
- 148 -

UNCLASSIFIED
Y-27608

Fig. T4A. Cross Section of Interface Region Sample from Waspalloy
Corrosion Specimen (Runs L-5-L-9:4). Interior circumference of white circle
indicates original size., As-polished. 5X.

UNCLASSIFIED
Y+27640

EEE E

3

1]
INClHES

T
!

o
o
~

|

o 4

n

£

2
Iy

|

oo
o

I

o
(]

EF

Fig. T4B. Portion of rkig. (4A. ‘''ypical surrace and Microstructure.
Etchant: Aqua regia. 200X.
UNCLASSIFIED
Y-27612

Fig. 75A. Cross Section of Interface Region Sample from Hastelloy X
Corrosion Specimen (Runs L—S—L—9:5). Interior of white lines indicates original

thickness. As-polished. 10X.

UNCLASSIFIED
Y-27614

017

.018

Fig. 75B. Portion of Fig. 75A. Typical Surface and Microstructure.
Etchant: Chromium regia. 200X.
- 150 -

UNCLASSIFIED
Y-27613

Fig. 76A. Cross Section of Interface Region Sample from Hastelloy W
Corrosion Specimen (Runs L—5—L—9:6). Interior of white lines indicates
original thickness. As-polished. 10X.

UNCL ASSIFIED
Y-27615

Fig. T6B. Portion of Fig. 76A. Typical Surface and Microstructure.
Etchant: Chromium regia. 200X.
- 151 -

UNCLASSIFIED
Y-27636

Fig. T7A. Cross Section of Salt Region Sample from Nimonic 80 Corrosion
Specimen (Runs IL-5-1-9:7). Interior of white lines indicates original
thickness. As-polished. 10X.

UNCLASSIFIED
Y-27637

Fig. 77B. Portion of Fig. TTA. Typical Surface and Microstructure,
Ltchanl: Agua regla., 200X,
= 456 =

UNCLASSIFIED
Y<27607

Fig. T8A. Cross Section from Interface Region Sample from Hymu 80
Corrosion Specimen (Runs 1-5-1-9:8). Interior circumference of white circle
indicates original size. As-polished. 5X.

UNCLASSIFIED
Y<-27616

';k&!f$i\ L *\rTC;\

L 7 N
Ry Al 3 ! ™\ b
( i —  \ / p
N P g4fi\,~\rh l
\’2 5 « = % ¢ ¥
. g \,\ \‘\\““)\;“\T_
( AL T

Fig. 78B. Portion of Fig. 78A. Typical Surface and Microstructure.
Etchant: Aqua regia. 200X.

azmzfib.@m ddA KYVININTIIINS

d XIANHddV
 THIS PAGE
WAS INTENTIONALLY
LEFT BLANK
- 155 -

Supplementary Volatility Pilot Plant Equipment

Sections I and II have déscribed reactions to the process environment of
VPP L nickel fluorination vessels., Other importané components used in the
fluorination cycle of the fluoride volatillty process have been examined for
corrosion resistance to their local environments. The components include a
CRP trap, a waste-salt line, the absorber vessels, valves and fittings, com-
ponents in the fluorine disposal system, and selected sections from the
pirocess gas lines. Figure 79 illustrates the relative positions of these

components in the pilot plant.

Complexible Radioactive Products Trap

To date, two CRP traps have been utilized in the VPP. The trap examined
was the second used and was in service during the Mark II fluorinator's life- °
time, as has been described previouély in Seection TTA. This vessel was
fabricated from a 5-in. sched-40 Inconel pipe, :31 in., long and it contained
NaF pellets to scrub the fluorinator effluent gas. The trap was designed to

remove ZrF4 and CrF_. and provides a certain amount of radiocactive decontaml—

p)

nation. Table XX summarizes the exposure conditions for the trap. A complete
description of the performance of the CRP trap has been published.sojl’52
After the Mark II fluorinator was taken off stream, trepanned sections

were removed from the CRP trap and sent to BMI for corrosion analyses.

Figure 80 illustrates the sections removed. ' |
Macroscopic examination of the samples at 20X revealed outlining of the

Inconel grains and an etched surface with increased attack noted at Section B-T.

50C L. Whitmarsh, Reprocessing of ARE Fuel, Volatility Pilot Plant Runs,
E-1 and E-2, CF-59-5- 108 (May 11, 1959).

5lC L. Whitmarsh, Reprocessing of ARE Fuel, Volatility Pilot Plant Runs,
E-3 through E-6, CF- 59—u-73 (August 26, 1959).

52C. L. Whitmarsh, Uranium Recovery from Sodium Zirconium Fluoride Salt
Mixtures, Volatility Pilot Plant Rune, L-1 through L-9, CF-59-9-2
(September 30, 1959).

-
f 3

. | : : ‘ o S UNCLASSIFIED
, V . ORNL-LR—-DWG 30402A

N ‘
* , ‘ ' 7 ‘ - Fa .
. : DISPOSAL
| . . . : ' UNIT
CHARGE MELT TANK SPRAY NOZZLES~|. (MONEL)
(347 STAINLESS STEEL). - . MONEL : i "
: > | =0
OFF GAS
1 R |
. : : CHEMICAL : .
. PUMP -— CAUSTIC SURGE TANK . WASTE
! - (MONEL)
; Na Fa ' ~ .
u 1L Zrf4~CRP __MONEL
8 TRAP iR A _
z (INCONEL) . COLD TRAP o
» HEAT EXCHANGER SHELL (MONEL) 5 |
‘ l I MONEL " HEAT EXCHANGER BAFFLES (COPPER) —a. 1= -
N : . . : \J1
: o)
- - |
FREEZE ‘ :
VALVE FREEZE -
VALVE [
.
W
p=d
(@] \
Q
: =
FLUORINATOR { L NICKEL);
M " CHEMICAL TRAP
TO PRODUCT (MONEL)
WASTE CAN RECEIVER
(LOW-CARBON STEEL)

'ABSORBERS
“(INCONEL)

Fig. 79. ©Schemetic Drawing of the VPP Sho.'v}'ing' Relative Positions of Cpfirponents. -
Summary of Process Conditions for Volatility Pilot

Table XX. _
Plant Complexible Radioactive Products Trap
During VPP Runs E-1 through E-6 and L-1 through L-9

' .Time of Exposure During Operations
Room Temperature to At Operating At Operating
Operating Operating Temperature Temperature
Temperature - Temperature With N, Flow With F, Flow®
Region Exposure (°c) (hr) (hr) (hr)
. |
Top FQ,UF6, N2 305-395 ~ LOO * ~ 400 ~ 100 M
(outlet) ) | ,
Middie FQ,UF6,N2,NaF ~ 500b ~ 400 ~ 400 ~ 100
(solids) : :
‘Bottom F,, UFg, I, 385-500 ~ 400 ~ 400 ~ 100
(inlet) (solids)
> stream,

aFor'appirox 35. hr UF6 was added to F
Temperatures recorded only during 2 runs, E-1 and E-2
- 158 -

UNCLASSIFIED
ORNL-LR—DWG 49474

NaF PELLET LEVEL

3in.
26 in.

e

| : 5-in., SCHED 40"
| | / INCONEL PIPE"

Fig. 80. , Location of Specimens Trepanned from the Inconel CRP Trap.
- 159 -

Dimensional analyses were performed on the sections as well as a microscopic
examination. A summary of the corrosion losses found by BMI personnel is
presented in Table XXI. Included in the table is a column converting the
maximum total losses to a mils per hour rate based on fluorine exposure time
at operating fiemperature for the vessel. A maximum total corrosion attack
rate of O.h6 mils/hr occurred in the lower section:of the trap which sus-
tained the higher operating temperatures.

Optical microscopic examination of sections removed from the trap
revealed intergranular attack on both the interior and exterior surfaces of
the vessel in the lower regions. The upper region showed intergranular attack
only on the vessel's interior wall. A typical cross section of the interior
wall of the CRP trap is shown in Fig. 81.

Comparison of the corrosive attack on the Inconel CRP trap and the Inconel
bench-scale fluorinator. described in Section IIT indicated maximum rate losses
for the trap to be. approximately twice that for the bench-scale fluorination.
vessel. The most similar areas with respect to operating environments were
the lower vapor region of the fluorinator and the lower region of the CRP trap.
At those locations, similar temperatures (approx 500°C) and process gases
(FE’UF6) were present. -In addition, however, sodium fluoride pellets were
in contact with the wall of the CRP trap. The maximum bulk metal losses in
the two regions were in a ratio of 3:1 (Trap:Fluorinator) while the maximum
interior intergranular penetration ratio was > 5:1. These differences are
difficult to reconcile in view of the 2:1 ratio for fluorine contact time at
elevated temperatures for the two vessels.

No appropriate reason can be advanced to- explain the variations cited in
corrosive behavior for the CRP trap and the Inconel bench-scale fluorinator.
However, the results present additional evidence that Inconel should be used
with caution as the construction material for certain fluoride volatility

Process component S.

Waste-Salt Line

The removal of waste fluoride salt from the VPP fluorination vessels was

accomplished by pressure transfer through a freeze valve and wasfie-salt line
Table XXI. Summary of Corrosion Losses on Specimens
Trepanned from the Inconel CRP Trap :

Wall .Intergranular ‘Total
‘ Thickness Penetration® Maximum Maximum
Distance frou Losses® (mils) Corrosion  Corrosion
Speci- Top Flange Weld (mils) -Interior Exterior Losses "RateC€
men Locatlon (in.) Maximum Minimum  Surface Surface (mils) (mils/hr)
T-T Top 3 . 7. O 3 None 10 O;l
M-T Middle 8.5 - - 11 5 - -
B-T  Bottom - 26 o2k 9 11 11 46 0.46
Based on 12 mlcrometer measurements taken on each spe01men and subtracted from nomlnal R

original wall thickness of 958 nils. _ |

bAs determlned by optical microscopy.
Based on fluorine exposure time at operating temperature.

v-o'9-[_
- 161 -

Unclassified
BMI C626

Fig. 81. Typical Microstructure of Sample B-T from the Interior
Wall of the Inconel CRP Trap Showing Intergranular Attack. Etchant:
Hydrochloric-nitric acid. 100X.
- 162 -

into a low-carbon steel waste container. Pressures of approx 5 psig in the
fluorinator were normally used to start the waste-salt flow. The salt then
flowed through the waste line by siphon effect (See Fig. 79). The salt line
reported here was used during the VPP L Runs 1-9 and for a single transfer at
the end of Run M-6k.

The subject waste-salt line was 3/8—in. sched-40 Inconel pipe. The line
was held at temperatures of 550-835°C for approx 22.5 hr and exposed to
52

flowing molten salts for about 2.5 hr during the pressure transfers. It was
probable that residual salts remained in static contact with the interior wall
of the piping for all operations.
After Run M-64, the piping was removed from the pilot plant, sectioned
in nine places and specimens sent to BMI for corrosion analyses. Figure 82_
shows a schematic drawing of the waste line and the location of the specimens.
Table XXII lists the corrosion data established at BMI by micrometer measure-
ments and metallographic examinations. Maximum wall-thickness losses of
21 mils were found in specimen IX which was at the exit end of the waste-salt
line.
Metallographic examination of the ecimens showed slight intergranular
attack of l/g-mil penetration depth had occurred on the interior wall of all
the samples with the exception of specimen IX. In the latter, a 3.5-mil thick
corrosion product layer was found on the interior surface of the specimen.
Figure 83 shows the corrosion product layer and the subsurface structure.
The product layer was spongy in character and similar in appearance to surface
layers found on sections removed from early test Inconel freeze valves which
had been in contact with NaF, ZrF), UF) (50-46-U4 mole %) at 650°C.(ref 53)
Personnel at BMI suggested that the product layer might be the result
of selective leaching of chromium by the fused salt. They indicated that the -
appearance was quite similar to layers found on Inconel in contact with hydrogen
fluoride and fused fluoride salts where chromium leaching had been proven. -
Analyses of the metallic spongy deposit found on sections removed from the

Inconel freeze valve referenced above indicated considerable losses of chromium

53L. R. Trotter and E. E. Hoffman, Progress Report on Volatility Pilot
Plant Corrosion Problems to April 21, 1957, ORNL-2495, pp.li4—16 (Sept. 30, 1958).

TO FREEZE VALVE
FROM FLUORINATOR

<—|

24 in. T I

2.5 /

3, -in. SCHED 40 PIPE
8

NOTE: DIMENSIONS ARE
IN INCHES

ROMAN NUMERALS INDICATE LOCATION OF
SPECIMENS REMOVED FOR CORROSICN ANALYSES

Pig. 82.

UNCLASSIFIED
ORNL—-LR—DWG 56057

ROTATE 90 deg
COUNTER CLOCKWISE

'€9T-

red  bag
! I
| '

| i
| W —

WASTE -SALT
CONTAINER

Schematic Drawing of Inconel Waste Salt Line from VPP Fluorinator.

Roman numerals indicate location of specimens removed from line for corrosion analyses.
- 164 -

Table XXITI. Corrosion Loss Data for Waste-Salt Line
Material: 3/8-in. sched-40 Inconel pipe

Distance from Top Wall Intergranular
of Waste-Salt Thickness Penetration Total
Container Losses® (Interior Surface)P  CorrosionC

Specimen (im.) (mils) (mils) (mils)
I 191.5 L 2 6
IE - 170 6 2 8
IIL 159.5 L 2 6
IV 146 L i | 5
v 110,5 L i 5
VI Tsd 3 i i L
Vit LO.5 3 1 L
VIIT 39.5 6 1 T
IX 0] 21 o) 24

aBased on metallographic measurements and subtracted from nominal
original wall thickness of 91 mils.

b
No intergranular attack was observed on the outer surface of the pipe.

CTotal corrosion equals the assumed original thickness of 91 mils minus
the measured sound metal remaining.
- 165 -

Unclassified
BMI C623

Fig. 83. Microstructure of Sample IX from the Interior Wall of Inconel
Waste Salt Showing Corrosion Product Layer.
- 166 =

and iron in the corrosion product layer when compared with the nominal
analysis of Inconel, i.e., 0.41% Cr and 0.46% Fe in the product layer.5LF
Since air, probably containing water vapor, could enter the exit line
at the point of preferential losses, the mechanism involved appears to be an
effect of a strong oxidant in the fluoride salt contacting the Inconel. In
such cases, it has been reported that the nickel ion rapidly goes into
solution and then is reduced by chromium resulting in leaching of chromium
from the alloy.55
The attack described emphasizes the importance of preventing air and/or

moisture from entering the fluorination system.

Absorbers

The absorbers in the VPP are used to decontaminate the VF6 product by
56

sorption and desorption on NaF beds. Inconel was used for the construction

of these vessels.

Absorber Gas-Distribution Ring Failure. The Inconel gas-distribution

ring in the first absorber failed during VPP operations. The failure was
discovered after Run M-52 but is believed to have occurred during Run M-L47.

A schematic view of the absorber and the gas-distribution ring is shown
in Fig. 84. As shown on the photographs, Fig. 85-a and -b, approx 3 in.3 of
Inconel was melted in an area roughly 180 deg from the Jjunction of the dis-
tribution ring and the inlet pipe. The melted area was about 5 in. long and
the distribution ring was nearly severed by this action.

The fused salt visible in Fig. 85-b around the metal nodules which were
formed as a result of melting was found by analyses to contain 4 wt % Cr as

Cr¥ o, 10 wt % Ni as NiF,, and 3 wt % U as UF),. Process chemicals normally

54
55F. F. Blankenship, The Effect of Strong Oxidants on Corrosion of Nickel
Alloys by Fluoride Melts, CF-60-3-125 (March 30, 1960).

56&. P, Milford, Engineering Design Features of the ORNL Fluoride
Volatility Pilot Plant, CF-57-4-18, pp, 3-L.

Trotter and Hoffman, Op. Cit., p, 16.

- 167 -

UNCLASSIFIED
ORNL—LR—DWG 49172

GAS EXIT GAS INLET
i 4

WITH NaF PELLETS

Il — VESSEL FILLED

N p— o — - - S
O VN T T T I ITTIIIIII N

|
)
<
x
-
O
=
=
L
[0
=
0
(@]

THERMOWELLS \
il

©
£2
Z
o=z
32
xS
om
o
=40
uLas

Schematic View of VPP Inconel Absorber and Gas-Distribution

Fig. 8.4.
Ring Illustrating Area of Failure.
Unclassified
Y-23294

(a)

Unclassified
Y-23295

(v)

Fig. 85. Photographs of Failed Region from the Inconel Gas-Distribution
Ring of Absorber, (a) Top View (b) Bottom View, Showing Molten and Recast
Nodules and Adhering Salt.
oo .
-',4’, Ve o~ v
¥ »

8]

- 169 -

present in this area pf the absorber were NaF pellets, F2 and UF6. Table XXTIII
shows the prior history of the ring up to the time of failure. Three unusual
plant operating procedures were a part of this prior history:57 (1) substitu-
tion of hydrogen fluoride for low temperature fluorine conditioning, (2) allowing
fluorine to enter the absorber at approx 600°C rather than 375°C as specified
in the Standard operating procedures (the higher temperature was used to
minimize hydrogen fluoride from the NaF and to reduce the sodium fluosilicate
content of the NaF), (3) use of an accelerated cooling rate after desorption.
Only (2) is felt to be pertinent toward causing the failure described.
Metallographic examination and chemical analysis of the samples cut‘from
the failed pipe, revealed only that the Inconel was normal in chemiétry and.
microstructure. The microstructure was equivalent to aé—received stock. The
molten and recast nodules exhibited a typical cast structure.
The Inconel gas-distribution ring appears to have failed as the result
of ignition Witfi fluorine. Ignitidh is believed to have been initiated by an
unknown foreign substance, such as dirt or grease, introduced while the absorber
was open. The failure points up the necessity for extreme cleanliness when

handling fluorine.

Inspection of Absorbers. Following fluorination Run L-9, the absorbers

were visually inspected and thickness measurements taken with a"Vidigage."
The ultrasonic device was operated by members of the Nondestructive Testing
‘Group of the Metallurgy Division. Results of those examinations are cited in
Figé 86 and 87. The maximum detectible losses found were approx 30 mils in
the upper regions of the vessel, a rate of approx 0.06 mlls/hr based upon
fluorine residence tlme

Based on these results, it is believed that the vessels can continue to

be used in pilot plant operation.

57F. W. Miles and W.-H. Carr, Engineering Evaluation of Volatility Pilot
Plant Equipment, CF-60-7-65 (September 30, 1960) pp.9596.

- 170 -

Table XXITI. Exposure History for Absorber Containing Failed Inconel
Gas-Distribution Ring from Absorber FV-120 ‘

1. .nglve (12) abgorption-desorption runs using NaF which typicali&

included: .. - ‘ . ,

a. Absorption at 65-150°C; 2 hr

' (FE at ~ 15 standard lifers/min

UF6 + 3 NaF-————————> UF6' 3 NaF
b. Purge-sparge with Né at 20 standard llters/mln for 1. 5 hr '
c. Desorption at 100-L425°C; 12-14 hr

' F, at ~ 18 standard liters/min
A 3

UF6- 3 NaF —mm—> UF6 + 3 NaF

2. Vessel was removed from service and left exposéd while the remainder of
" the system was water washed, dried, and treated with Fo. Vessel was

then placed back in service. Co

3. One (1). special run whose purpose was preparation of NaF from NaHF,,.

a. Purge sparge with Né at ~ 3 standard llters/mln for 12 hr

25-125°C in 7 hr
at 125°C for 5 hr

\

b. ~“Preparation of NaF under Né at 3 standard liters/min by

NafF_ + A ——> NaF + BF |
2 - ‘ C ,
125-600°C in 9 hr
at 600°C for L4 hr
c. Condltlonlng of absorber at 635°C with F2

at ~ 15 standard llters/mln for 0.5 hr

m : T
Failure is-thought to have occurred at this time.
- 171 -

FV-120
FIRST ABSORBER UNCLASSIFIED
ORNL-LR-DWG 49473
Material:
Shells - %’ in. Inconel, rolled plate |t\(|;|f\EsT EC,)?I% T\:EEEASO
Bottom —~ :}8 in. Inconel, flat plate ;

Service Conditions:
NaF contact  ~ 5000 hr : :

UF6 contact ~400 hr

|
j

~310 k Naf PELLET
9 8ED LEVEL .
Pressure <0.1 in. H20 - 5psig jfi__l_ . _L*_
. Absorption 1 1 7 _l_(
Wall temperature 65-150°C l c
F, contact ~ 100 hr | ~
~ 100,000 std. liters FRONT VIDIGAGE | -
Desorption READINGS TAKEN :—[
Wall temperature 100~-425°C ALONG THIS LINE l ]
occasional excursion to 600°C £
Te)
N

F, contact ~ 400 hr ’
~ 170,000 std. liters

Visual Inspection

DISTRIBUTION |
Interior RING 1T

|
|
|
|
|

Shell — Covered with aon adherent yellow-brown deposit \\
except for the surfaces within 3 in. of the bottom. ] O]
No defects visible. - :
Bottom ~ No coating or defects visible. : |
Welds — All beads visible appeared to-be in good condition. I-—‘lOin.———|
Exterior
Shell - Covered with u thin Llack adherent film, presumably oxide.

Bottom —_ Covered with a thin black adherent film, presumably oxide.
Welds - All beads appeared to be in good condition.

Vidigage Inspection: - (Nondestructive Test Section operators expressed ‘‘no confidence’’ in front ond
back readings due to the wall deposit and operational difficulties.)

Wall Thickness in Inches

Distance' Location Remarks Front Series Back Series“: Bottom Series3
2 Neck ) 0.226 0.235 0.370
5 Reducer 0.222 0.220 0.374
8 Shell 0.236 . 0.224 .0.370
10 Shell’ 0.223 NA '
12 Shell, 0.230 NA
15 Shell 0.236 NA
: 18 Shell 0.240 NA

=0 Shell 023 0.235
23 Shell 0.241 0.238
25 Shell 0.242 0.240
28 Shell 0.244 0.234

1. Ininches, from bottom of flange.
2. These readings taken 180° from front series.
3. Random readings, £0.002 in.

NA: Not Available .

Fig. 86. Results of Vicual and Vidigage Inspection on VPP First
Absorber. . : '

.
- 172 -

Fv-121
. SECOND ABwRBER N UNCLASSIFIED
. ‘ A ORNL-LR-DWG 49474
Material: : GAS GAS THERMO-
Shells - ]/4 in. Inconel, rolled plate ] INLET EXIT WELLS
Bottom —~ % in. Inconel, flat plate
Service Conditions: ’
NoF contact ~ 3000 hr
UF6 contact ~250 hr : NoF PELLET
~ 200 kg BED LEVEL
.Presswe <0.1in. H,0 - 5psig
Absorption / ‘
Wall ;emperature ~ 65-150°C - | <
F t ~ . '
2 contact 40 hr FRONT VIDIGAGE | -

~ 36,000 std. liters READINGS TAKEN
Desorption ALONG-THIS LINE — -

Wall temperature  100-425°C . |
occasional excursion to 600°C . i

F, contact . ~250 hr
~ 90,000" std. liters

DISTRIBUTION

Visual Inspection

RING

o

Interior : \\ l

Shell - Covered with an odherent .yellow-brown .deposit 1 15 m
* except for the vessel neck and surfaces within 6 in. H

of the bottom. The deposit appeared thinner than f

that noted for FV-120. No defects visible. .o | . ! I
. {0in.

Bottom — No coating or defects visible.
Welds - All beads visible appeared to be in good condition.

Exterior

Shell - Covered with a thin black ‘adherent film, presumably oxide.
Bottom — Covered with a thin black adherent film, presumably oxide.
Welds — All beads appeared to be in good condition. -

Vidigage Inspection: (Nondestructive Test Section operators expressed ‘‘no confidence'’ in front and
‘ back readings due to wall deposit and operational difficulties.)

-

L Wull. Thicknes§ in Inches

Distance'! Location - ~ Front Scrics . Back Series? Bottom Series?
2 Neck 0.228 0.232 0.370
5 . Reducer 0.222 . 0.230 0.374
8- Shell 0.220 10.230 0.372
10 Shell ' 0.222 0.230
12 Shell TONA L 0.228
15 Shell NA - 0.228
_ .18 : Shell . NA 0.226
=2 Shell NA . 0.230
23 Shell 0.228 0.237
25 : Shell . 0.232 . 0.240
28 Shell 0.238 " - 0.238

1. In inches, from bottom of flange.
2. These readings taken 180° from front series.
3. Random readings,  £0.002.

NA: Not Available

Fig. "87. Results of Visual and Vidigage Inspection on.VPP Second
Absorber. : !
o

_]_73_.

Valves and Fittings

Aside from sporadic leakage and some plugging by unidentified deposits

-of valves carrying UF6, valves and fittings in the VPP performed satisfac-

o7

torily. Many screwed fittings carrying nitrogen were back-welded after
various amounts of plant operational time to eliminate such leakage.

Fluorine carrying valves of various body materials, i.e., Monel, nickel,
Inconel, steel, types 316 and 347 stainless steel, brass, and copper, were
welded or brazed into the associated piping. At temperatures greater thafi
150°C, only nickel and Inconel valves were in contact with fluorine under
normal 6perating conditions.

One failure, originally thought to be in a valve but later discovered
to be in the adjacént fittings, has been selected for treatment here because
of 'the materials involved and their reaction to service conditions.

Following Run L-4 in the VPP, leaded, yellow brass valve fittings of the
composition 60-63% Cu—2.5-3.7% Pb—0.35% Max Fe~0.5% other—bal Zn (vendor's
analysis) were removed from & nitrogen purge line, PE-3k4, because of defective
valve operation.

The inlet side of the valve, a packless diaphragm-seal valve with an
Inconel diaphragm, Duranickel stem point, and Monel body and trim, was placed

in conjunction with the NaF absorber outlet piping which carried F, and UF6..

Under normal operating conditions, the valve was not in contact wiih process
gases, but there was continuous nitrogen contact. The valve was shut only
when nitrogen pressure was low and there was the possibility of process gases
diffusing past the valve to the instrumentation units. '

During the early L-runs, persistent system plugs and low pressure in the

nitrogen line probably permitted F. and UF6 to contact the valve and its

fittings for an unknown period of iime. The operating service time on the
valve and its fittings was approx 6 weeks.

Visual examination of the valve disclosed no obvious defects. Therefore,
it was subjected to several times operating pressure (15 psig) by VPP person-
nel and found to function satisfactorily without leakage. The valve was

subsequenkly placed in stock to be used as an emergency replacement.
- 174 -

Visual observation of the fittings revealed that several components had
ruptures running the complete length of the walls. The fissures were normel
to stresses induced during installation: Figure 88.and Table XXIV show and
describe the defects noted. Metallographic examination of the fittings re-
vealed the cracks to be intergranular in nature. Dezincification apparently
occurred along.the inside ‘edges of the fissures, as evidenced by;copper-rich

" deposits which lined the fissures. (See Fig.'89.) Copper deposits were also
evidegt in several érain boundaries radiating outward from the flaw. GCeneral
corrosive attack was nored on the outside_of the inlet union, Fig. 902'

Although all evidence indicates that faiiure was by selective inter-
granuiar dezincification, the stresses induced during installation were believed
fo have provided a strong directional driving force. The use of stock Monel
fittings, because‘of-its increased corrosion resistance, was recommended for

. use in the environment described. ' 1 ‘
~ Theé asspciated Monel valve was found to be in good operating condltlon,

appeared satisfactory, and can continue to ‘be used for the mentioned service

- conditions.
{F " This failure serves to point out the necessity for constant vigilance in
-selectlng materials and components for a highly corrosive process scheme, b

.+$hls includes the selection of plplng and .joining couples which should, if at

all possible, match thé components in corrosion-resistant properties.

-Flfiorine Disposal System

Excess fluorine and nitrogen from the VPP have been disposed.of b& co-
current contact with an aqueous KOH solution in a gas-dispoeal unit. (See
Figs. 79 and 91;)— The KOH solution was prepared in a caus;ic surge tank,
shown . in Figs. 79‘and 92, to a copeentration of 2-10% KOH and pumped to the
disposal unit. The caustic'solution was then sprayed into the dispesal unit
as the excess process gases entered the vessel. Fluorine in the waste gases

formed potassium fluoride in water solution according to

2?2 + 2KOH > F20 r 2KF + HEO

FQO + 2K0OH > 2KF + O2 + HEO

uoTufl 39Ul |

Unclassified

Y-26021
Leaded Yellow Brass——— [—————Leaded Yellow Brass—

H — — e
% 31% gi% @15 w:g g E'g'g
pars R o H o o K = g
N O H ™ B o o
= g‘E E;E H oo E‘fi ot B o
& @ e o b8 D = gBE
e = 5 o Sl
&+ n B

=

o

3

Fig. 88. Photograph of Gas Purge Line (P3-34) Valve and Atteched Fittings. (Note

fractures in inlet union, inlet nut, and inlet front ferrule.)

= GLT =
- 176 -

Table XXIV. Defects Found in Leaded Yellow Brass Fittings

Sub jected to F2 and UF6
Part Name Location Defect
*

Union Inlet side of valve Wall fracture through threaded end
covered by hex nmut. Fracture extended
through hex body and part of opposite
threaded end.

*
Hex Nut Inlet side of valve Wall fracture running entire length of

Front Ferrule

Back Ferrule

Union

ferrule.

*
Inlet side of valve Wall fracture running entire length of
feyrrule.,

*
Inlet side of valve Crack partially through wall.

Wall fraclure partially through threaded
end covered by hex nut. (Not shown in
Fig. 88 because of photo orientation.)

*
All fractures and cracks were essentially in line with one another (see

Fig. 88).
- 177 -

Unclassified
Y-27193

S
b4

.

Fig. 89. VPP leaded, Yellow Brass Inlet Back Ferrule Showing Intergranular

Fracture Lined with Copper-Rich Deposits Indicative of Dezincification.
Etchant: NHHOH-Hé-HéO. 250X%.

00!

- 178 -

Unclassified
Y-27189

Fig. 90. VPP Leaded, Yellow Brass Inlet Union Showing General Corrosive
Attack. Etchant: NHMOH—HéOE-HéO. 100X.
..]_79_

FY-150
GAS DISPOSAL UNIT

Material:

Shells — ‘/8 in. Monel, rolled sheet

Heads — 3/16 in. Monel, flanged and dished
Service Conditions:

Excess F, with N, from processing passed in co-current con-
tact with 2-10% KOH in tower at 0.5 cfm. Discharged at
a concentration of 2% KOH, 5-15% KF.

Circulating ~ 5000 hr

Not circulating ~ ~8000 hr

Temperature 35-55°C occasional excursions to 100°C
Pressure 0.1in. H,0

Visual Inspection
Interior
Could not be inspected at this time.
Exterior

Shell — Covered with an adherent green deposit. Nu defects
visible.

Head — Covered with an adherent green deposit. No de-
fects visible.

Welds — All appeared to be in good condition.

UNCLASSIFIED
ORNL-LR—DWG 49175

FZ_ N2 INLET
(ROTATE 90°)

KOH INLET
TOTAL OF SIX

NOZZLES, 15-in. OC

, O KOH DRAIN TO

Vidigage Inspection SURGE TANK
Wall Thickness in Inches*
SHELL SHELL SHELL HEAD
Distance' Circumferential Inter-KOH Nozzle F, Inlet-lmpingement Sories
Series? Series Series?
1 0.119 0.118 No. 1-2 0.126 Dished Top
0.118 0.118 0.125 0.126 0.182
0.118 0.119 0.126 0.127 0.184
0.127 0.126 0.184
9 0.124 0.125 0.127 0.127 0.186
0.126 0.126 0.127 0.183
0.125 0.125 No. 2-3 0.128 0.182
0.128 0.128
24 0.130 0.125 0.128 0.128 Flanged Side
0.129 0.127 0.126 0.128 0.185
0.128 0.130 0.128 0.184
0.129 0.189
39 0.125 0.126 No. 3-4 0.190
0.127 0.124 0.189
0.128 0.186
54 0.130 0.126
0.129 0.128 No. 4-5
0.127 0.129 0.129
0.128
69 0.127 0.129
0.129 No. 5-6
0.124
0.125
84 0.128 0.128
0.129 0.128
0.128 0.129
99 0.130 0.128
0.129
1. Distance, vertically down, in inches starting at top head to shell weld — reference only to shell
circumferential series.
2. Reudinys taken at 60%or 120°% ratated ~ 20° for each successive group.
3. Readings at 1 in. intervals starting 13 in. below the head to shell weld 180° from F, inlet.
4. All readings, £0.002 in.
Fig. 91. Results of Visual and Vidigage Inspection on VPP Gas
Disposal Unit.
Material:
Shells — ;l‘ in. Monel, rolled sheet
Heads — :?u in. Monel, flanged and dished

Service Conditions:

Continually %~ full of 2-10% KOH with F~
present. (One area subject to agitation)

Circulating ~50C0 hr
Not circulating  ~80C0 hr
Temperature 25-55°C
Pressure Atmospheric

Visuval Inspection

Interior
Could not be inspected at this time.
Exterior

Shell — One end covered with an adherent green deposit. No defects noted.
Heads — Both heads covered with an adherent green deposit. No defects noted.
Welds — All beads appeared to be in good condition.

UNCLASSIFIED
ORNL-LR—DWG 49176

VIDIGAGE READINGS TAKEN
ALONG THESE LINES

FRONT BACK

Leegeeeieeeieeieeieeeeeete

le—— 30 in.

v N £ D i

I
L—— 24 in, —=

94 in.

Vidigage Insection WELDS & s |
Wall Thickness in Inches? (’_IJ)
Distance' Front Series Remarks Back Series Remarks O
I
Top 0.189 0.194
5 0.190 0.195
10 0.188 0.196
14 0.187 Longitudinal 0.196 Longitudinal
18 0.187=—" weld here 0.197 _—" weld here
22 0.190 0.196
26 0.191 Agitsiion 0.195
30 0.192 e 0.195
34 0.189 0.194
Bottom 0.189 0.194

1. Distance, in inches, circumferentially from top of shell.

2. All readings, +0.002 in.

FV-152
SURGE TANK

Fig. 92. Results of Visual and Vidigage Inspection on VPP Surge

Tank.
- 181 -

which subsequently was discharged as chemical waste while the nitrogen re-
maining in the excess gases was exhausted to the atmosphere. The entire
fluorine disposal system was fabricated from Monel.

After VPP Run L-9, visual and ultrasonic thickness measurements were
taken on the gas-disposal unit and the surge tank to ascertain thelr suita-
bility for additional VPP service. The respective service conditions for
the two vessels and the results of those inspections are presented in
Figs. 91 and 92. A maximum wall-thickness loss of 5 mils was found near
the top head of the gas-disposal unit while no detectable metal losses were
found in the surge tank.

Previous studies have indicated that heavy corrosive attack can occur
on the process gas inlet pipe inside the gas-disposal unit.58 Therefore, &
inlet pipe fabricated from 1l-in. sched-4O Monel pipe in service for approx
4900 hr was removed from the disposal unit and subjected to visual and metallo-
graphic examination. Figure 91 shows the location of the gas inlet pipe.

Visual examination of the exterior of the pipe inlet disclosed the
presence of two holes, approx 1/32 in. and 3/32 in. in diasmeter, respectively,
on the upper half of the pipe. The holes completely penetrated the pipe wall
and were 1-1/2 in. and 2-1/2 in. from the exit end of the pipe. Inspection
of the interior of the nozzle showed an extremely rough surface and a severe
pitting attack extending from the end of the nozzle inwards for approx 3 in.
A white crystalline deposit encircled the interior at a point 4 in. from the
end of the pipe. ©Subsequent x-ray diffraction analysis indicated this deposit
to be wholly KHFQ.

Metallographic examination of sections removed from the pipe showed a
roughened surface and metal losses concentrating in circular cavities but no
evidence of intergranular attack. The holes described above were simple ex-
tensions of the localized attack. Figure 93 shows a magnified view of the
pipe sectioned through the center of the largest hole found in the inlet pipe.
A photomicrograph of the surface at the base of one of the pits is shown also

in this figure,

58L. R. Trotter and E. E. Hoffman, Progress Report on Volatility Pilot
Plant Corrosion Problems to April 21, 1957, ORNL-2495 (September 30, 1958).

- 182 -

Unclassified
Y-32237

SesampTE
ORIGINAL THICKNESS
(a)
" \ \ g i . NI PLATEF ; "' “"m Unclassified
\ W T ¢ ' p . “ ALY N Y-32029
‘-‘ o WK ’ ‘\ \ | L Ly / : ; o 8
s . 5 A “‘ . % ' ‘ { ! "' s - T —
Ll 'H' : \ AL i “\s‘ s g 2
: l‘ c“, % o o T » .‘j A . 8 40 g
\ ¢ ¥ ) o~ 1 . ; ,:( b g | o .002
+ - ~ 4 ._‘
L 7 e
r'\-s)\‘ Y d S .003
- . ¥ /
> =
N J . ’)
\ ‘Z’ .004
~ \.\ .008
I \ N
&
\ .006
P / : ///// x
B-dl
(b) °

Fig. 93. Cross Sections of Gas Disposal Unit Inlet Pipe (a) Photomacrograph
Showing Severe Pitting Attack. 15X. (b) Portion of Fig. 93A showing typical
surface at base of pit. 500X. Etchant: Acetic-nitric-hydrochloric acid.
.
7

i

-~ 183 -

Corrosion seemed to have been directional from the interior toward the
exterior of the pipe. Examination of the fluorine-nitrogen entrance end of
the pipe disclosed no noticeable or measurable attack.

Intermittently during operations, the gas-inlet nozzle plugged (presum-
ably by KF and/or KHF,,

frequency during the "E" and early "L" runs that a secondary nozzle was in-

deposits) and this condition occurred with such

serted in place of a lower KOH inlet. Examination of this secondary nozzle,

which had approx 100 service hours, disclosed a very slight attack similar

~to that found on the primary pipe.

The presence of KHF, inside the main gas-inlet pipe is believed to

2
occasion the high localized corrosion found in that component. It would

appear that an easy way for KHF. to form in the pipe would be by the alkali

2
deficient reaction of KOH with HF as

2HF + KOH ————> KIF, + H,c,o.(ref 59)
The formation of hydrogen fluoride seems possible from the reactions of
fluorine monoxide or fluorine with the available water of solution present.

Investigators have reported the hydrolysis reaction of FQO

FQO + HQO > O2 + 2HF

to be very slow at ordinary temperatures but a steam-F_ O reaction occurs so
-~ Fa

rapidly that it can be considered explosive.bo For reasons unknown, the re-
action of fluorine with water has been reported as not always occurring.
One source has indicated that fluorine reacts with water easily, giving oxygen
mixed with ozone, hydrogen peroxide, and other oxidizing substances.

If hydrogen fluoride is present in the gas-inlet pipe, then the high

degree of localized attack on the nozzle may be explained as the result of

59

6OA. B. Burg, "Volatile Inorganic Fluorides," Fluorine Chemistry- (ed. by

J. H. Siwmons) p.83, Academic Press, New York, 1950.

G. I. Calthers, Chem. Tech. Div., ORNL, Private communication.

61R. Landau and R. Rosen, "Industrial Handling of Fluorine," Preparation
Properties and Technology of Fluorine and Organic Fluoro Compounds (ed. by

C. S. Lesser and S. R. Schram) NNES VII-I,p.154, McGraw-Hill, New York, 1951.

62N. W. Sidgwick, The Chemical Elements and Their Compounds, Vol. II,
p. 1101, Oxford University Press, London, 1950.

- 184 -

hydrpgen fluoride corrosion highly accelerated bfi‘thg presefice of oiygen.

Also, at particular concefltrétions of hydrogen flubride in water, close to

the- azeotroplc composition (38 wt %),one investigator has reported corrosion

‘rates on Monel of 145 mils/yr at 120°C. (ref 63) It may be that. either or - &

both of these reactions may be responsible for the localized attack described. G
Considering the easy replacement of the gas-inletvnozzie and the major .

cérrosidn problems‘remaining to be solved in the VPP, no development work

haé beén planned toward'findifig an'imfiroved material of construction.

Process Gas Lines

. Sections of Monel pipe and copper tubing near the VPP absorbers, chemi;
cal traps, and cold .traps were removed from the process gas piping system .
after Run M-6L4 and sent to BMI for corrosion analyses. A schematic drawing
Showing,thé.locafion of these sections is presented in Fig. 94, ‘and a descrip-
tion of the piping and the individual process environments can be found in jfi
Table’ XXV. | -

The wall thicknesses of the sections were measured by BMI and no indica-
‘tion of.significanfi metal losses were found. The méfallographic examinations
conducted on the speéiméns discloééd‘no serious corrosive attack. Almost all
of".the SPecimens showed isolated pitting but the pits vere less than one grain
' deepL Slight indication of selective attack at the grain boundaries was |
* observed on specimens 1-1, l-é3 1-3, 2, and 3 which were located close to the
absorbersfland the first cold trép. However, the intergranular attack was

- never deeper than one grain, and for the most part, was less than 1 mil deep.

63W Z. Frlend "Nickel-Copper Alloys," Corros1on Handbook (ed. by
H. H. Uhllg);) 273, John Wiley and Sons, New York, 1948.

o

~ UNCLASSIFIED
ORNL-LR-DWG 49177

| CYLINDER |
ABSORBERS | | | | CHEMICAL
- | \- - TRAP

\\\\§ \§ - PrRODUCT

Fig. 9L. Diagrammatic Flowsheet Showing Lccation of Process Gas Line Specimens
Removed from the VPP Systenm. ‘

- SOT -
’

TABLE XXV. LOCATION, DESCRIPTION, AND EXPOSURE CONDITIONS FOR SPECIMENS FROM PROCESS GAS LINES

Specimen Description

Exposure

Fused-Salt

: Cooling After N
- D t 9
Number Location Material Fluori nation-UF ssorpron Desportion Product Removal
Sorption
1-1 Ahead of first absorber ]]/4 and 1 ]/2 in. sched-40 200 hr, 110-140°C, - 350hr, 110-140°C, 350 hr, 110-140°C, None
1-2 Monel pipe F,and UF (30 hr F, N,
1-3 with 90% UF ()
2 Outlet from second 1 ]/4 in. sched-40 Monel pipe 200 hr, 75-150°C, 350 hr, 75-350°C, 350 hr, 75-350°C,  None
absorber F, F and UF, N,
3 Inlet to first cold trap ]"‘/4 ond 1 in. sched-40 Monel 200 hr, 140-165°C, ., 350 hr, 140-165°C, 350 hr, 140-165°C, 190 hr, 140-1%5°C,
‘ pipe; 90 deg ‘ell’ and reducer F2 E2 and UF6 Ny, 0-55 psia, 'UF6
4 Qutlet from first cold ]/2in. sched-4C Monel pipe 200 hr, 110-135°C, 350 hr, 110-135°C[, 350 hr, 110-135°C, 190 hr, 110-135°C,
trap . F, F, = - N, ‘ 0-55 psia, UF -
5 Bypass valve of second 1/2 in. sched-40 Monel pipe 200 hr, 110-135°C, 350 hr, 110-135°C, 350 hr, 110-135°C, 190 hr, 110-135°C,
‘cold trap - F2 F2 N2 ‘ 0-55psia, UF6
6 Ahead of chemical trap 1/20nd 1 in. sched-40 Monel 200 hr, 100-150°C, 350 hr, 100-150°C,- 350 hr, ]00-]50°C, 190 hr, 100-150°C,
' pipe; straight “T' reducer and F, -~ F, N, ' 0-55 psiq, UF,
90-deg ‘ell’ A
7 Outlet from chemical 1 in. sched-40 Mone|~pipe 200 hr, 35-60°C, F, 350 hr, 35-60°C, 350 hr, 35-60°C, None
trap ' F, N,
3, . . Y o o .
8-1 Inlet side of product -% in. copper tubing None None None ]9L?Fhr, 66-100°C,
8-2 receiver 6
8-3
9-1 Outlet side of product 3/8-in. copper tubing None None None 190 hr, 66-100°C, -
9-2 receiver UF(,
9-3
y
]
~ l>:=é: r‘.'- ! + __' _i - :—‘ >

- 98“[ -
OS'D:e‘.‘.PP_MF;UF"‘—I'#FI!M';USWOQSUSZWWQWH@’J@NEtl:l@

DISTRIBUTION

Biology Library 6L—68.
Health Physics Library 69.
Metallurgy Library 70.
Central Research Library T1.
ORNL Y~12 Technical Library T2.
Document Reference Section 73.
Laboratory Records Department Th.
Laboratory Records, ORNL R. C. 75.
M. Adamson 76.

J. Barber (K-25) 7.

R. Bennett 78.

A. Bernhardt (K-ES) 79.

S. Billington 80.

E. Blanco 81.

F. Blankenship 82.

G. Bohlmann 83-85.

S. Borie 86.

C. Bresee 87.

N. Browder 88.

B. Brown 89.

A. Brown 90.

A. Bush al.

0. Campbell g2.

H. Carr 93.

I. Cathers 9k,

E. Center 95.

A. Charpie 6.

E. Clark a7.

S. Cockreham o8.
Cohn 99.

J. Culbert (K-25) 100.

L. Culler 101.

E. Cunningham 102
-A. Douglas, Jr. 103.

S E. Ferguson 10k.
H Frye, Jr. 105.

E. Goeller. 106.

E. Goldman 107.

T. Gresky ) 108.

R. Grimes 109.

E. Guthrie 110.

F. Hale (K-25) 111.

- 187 -

ORNL-2832

Metals, Ceramics, and Materials
TID-4500 (16th ed.)

CARHSEE GO IR UAE NP TGN E ORGP IS EECES DT YR

. R. Hill

Hollaender
Horton
Householder
Jolley
Jordan (Y-12)
Jordan
Keesee

Keim
Kelley

mb

Lampton
Lane

Lewis

. Lindauer

. Litman

Long

. MacPherson

CURBEEEEpR YR ON G S

Matherne
Miles
Milford
Miller
Miller
Moncrief
. Morgan
Murray (K-25)
Nelson

. Nicol
Patriarca
Phillips

M. Reyling
B. Ruch

E. Seagren
M. Shank
Shapiro (K-25)
Shipley
Skinner
Smiley
Smith

QEHNQQG Y

omlE4aY
112.
113.
114,
115.
116.
117.
118.
119.

128.
, 129.
© 130-131.
132,
133.

- 134,

- 135.
136.
137.
138.
139.
140.
141,
142,
143-719.

caRrEgn>

=

C)DUU"JUZLIE"UU)'T-:I

. Snell
Stainker
. Swartout
Taylor
Weinberg
Whatley
Whitmarsh
Winters

R

- 188 -

120.
121.
122.
123.
124 .
125.
126.
127.

EXTERNAL DISTRIBUTION

. E. Baker, GE Hanford

E. Bigelow, AEC, OAD

. Cope, AEC, ORO N
O. E. Dwyer, BNL

rsel Evans, GE Hanford

. W. Fink, BMI
Lawroski, ANL

. D. Miller, BMI
Seefeldt, ANL

Simmons, AEC, Washington
J. Steindler, ANL

K. Steunenberg, ANL

(B B w)

. C. Vogel, ANL

K. Stevens, AEC, Washington

MmO g > H
Unu:riblfi '

. A. Wilhelm (consultant) . =

. Youngblood

Burr (consultant) _
Gregg (consultant) A
Koenig (consultant)

Smith (consultant)

.'Smoluchowskl (consultant)

E. Stansbury (consultant)

iven distribution as shown in TID- MSOO (16th ed.) under Metals, Ceramlcs, - s
and Materials Category (75 copies - OTS)