ORNL-TM-2490

 

Contract No. W-7405-eng-26

4 ‘T }':}

METALS AND CERAMICS DIVISION

COMPATIBILITY OF HASTELLOY N AND CROLOY 9M WITH
'NaBF,-NaF-KBF, (90-4-6 mole %) FLUOROBORATE SALT

J. W. Koger and A, P. Litman

 

 

I.EGAL NOTICE

“ , . ’ | This report was prepared as an account, of Government -ponsored work. Neither the United
L ’ . i i Btates, nor the Commission, nor any person acting on behalf of the Commission:
f - A, Makes any warranty or representation, expressed or implied, with respect to the accu-
rlcy. completeness, or usefulness of the information contained in this report, or that the use
S , - pf any information, apparatus, method, or process disclosed in this report may not infringe -
| privately owned rights; or -
" . . B, Assumes any liabilitiea with respect to the usze of, or fior damages resulting from the’
~ use of any information, apparatus, method, or process disclosed in this report.
. As used in the Rbove, ‘“‘person acting on behalf of the Commiasion’ includes any em-
i ployee 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,
i disseminates, or provides access to, any information pursuant to his emiployment or contract
3 with the Commisaion, or his employment with such contractor.

s

o

- CAPRIL 1969

OAK RIDGE NATIONAL LABORATORY
- Oak Ridge, Tennessee
. .. operated by -
' UNION CARBIDE CORPORATION

(- - | ~ for the

U. S. ATOMIC ENERGY COMMISSION

o 9w

 

GISTRIBUTION OF THIS DOCUMENT 15 URTH

Ty
 

(
i

 

 

.‘J

i

*)

S s

<)
 

)

v

Abstract { e e

- . & @ . -

Introduction . « « + « & o o o « &

Experimental Procedure . . . . .

- Materials Sectibfi and Fabrication

Salt Preparation . . .
Operations . . . .'; c e e e e e
Resulfs’ .. .'. e e e e e

NCL-10 (Hgstelloy N) .

NCL-12 (Croloy 9M) . .

Dis cus S ion . - » - - ¥ - * - - -

Thermodynamics of System Corrosion .

Kinetics of System Corrosion . . . .

~ Salt Purification . .

Recommendations . . . « « « + « o . .

Acknowledgments . . . . . . .

-

) COHCIuSionS . - - s e . s e . e . . .

llllll

 

i

O O 3 v W K

Wowowwwwn e
BT IR WH OO
 

”

* e

~ 'COMPATIBILITY OF HASTELLOY N AND CROLOY 9M WITH
- NaBF;-NaF-KBF; (90-4-6 mole %) FLUOROBORATE SALT

J. W. Koger and A. P. Litman
. .ABSTRACT

The compatibility of a relatively impuré (> 3000 _ppm

impurities) fluoroborate selt, NaBF,-NaF-KBF, (90-4-6 mole %) s

with Hastelloy N and Croloy 9M was evaluated in natural cir-
culation loops operating at a maximum temperature of 605°C

- with a temperature difference of 145°C. The Croloy 9M loop
- '(NCL’—J.E.’) was completely plugged -after 1440 hr of operation ..
~and the Hastelloy N lo (NCL-10) was three-quarters plugged

after 8760 hr (one"ygar' ‘of operation. All major alloying

 elements of the container materials mass transferred during

operation as the result of nonselective attack by virtue of .
the initial oxygen and water contamination of the salt. Satu-
ration.concentrations of 700 ppm Fe and 470 ppm Cr were deter-

“mined for the fluoroborate salt at 460°C.. The mechanism of -
corrosion of the system is as follows. Initially, metal fluo-

ride compounds that are soluble in the salt are formed in the
hot leg. The reverse reaction occurs in the cold leg and
causes the metal to deposit and to diffuse into the cold leg.
This contimes until (1) an equilibrium concentration of one
or more metasl fluorides is reached in the salt at the cold-
leg temperature and these compounds start depositing (e.g.,
NCL-10) or, (2) the equilibrium constant of the reaction

deposited (g.g., NCL-lZ)

- changes so much with temperature that the pure metal is

 

, i J?NTRODUCTION

Nuclear reactors that use molten fluorides as fuels are under

: * development for 'thr_e' ‘dual purpose of power prodfié_fiioh and thorium con-
” 've'zjs;idnl.- 1 -:H_ea.t: genérafed ;ifi,-the ‘core region of such ,a,i_ziioitéh—éa.lt
| “breeder reactor is tra.nsferi;é&--”f'r_omthe fuel-containing fluoride
. selt to a 'secondary,co'olé.nfbfféiré\iit -of fluoride sa.ltmthcut fuel, tl;;at

 

IMSR Program Semiann. Progr. Rept. Feb. 29, 1968, .ORNL-'-4254.
 

then dissipates the heat to steam. 2 In the Molten-Salt Reactor Experiment

(MSRE), the nickel-base alloy Hastelloy N has proved to be an effective
container material for the fluorlde salt that contains the fuel and for
the LiF-BeF, diluent.

Many potential fissile® (contalnlng UF, ), fertile® (containing ThF4),
or combination! (containing both UF; and ThF,;) salts proposed for the
MSER have been or are being tested for their compatibility with
Hastelloy N and other container materials. The choice of a salt for
the secondary coolant, however, remains_open. Test programs naveronly
considered coolant salts like NaF-ZrF, and, recently, LiF-BeF.. While
these salts have demonstrated excellent compatibility with Hastelloy N,
there is need for cheaper fluoride mixtures that melt at 1owor tempera-
tures. On the baéis of low cost (approximately 50é/lb) and a relatively
low melting point (about 400°C, necessary for transferring heat to super-
critical steam), fluoroborate salts -especially'NaBF4 — with small
additions of NaF and/or KBF; have recently become of interest as secon-
dary coolants. Little is known, however, about the corrosion behavior
of these salts. _

The space diagram of the ternary NaF-NaEF,-KBF, systém (Fig. 1)
shows that the salt mixture under consideration lies very near the NaBF,
corner and has a melting point close to 390°C. Note that oxygen impu-
rities that form BF30H-compounds significantly lower the melting point
of the mixture below that shown by the diagram. The vapor pressure of
fluoroborate salts, especially NaBF,, is higher than that of other
fluoride salts,’ because in the temperature range of interest the sodium
fluoroborate maintains such an equilibrium with its components that a

significant amount of boron trifluoride gas is present in the system:

NaBF, =NaF + BF;(g) . (1)

 

°P. R. Kasten, E. S. Bettis, and R. C. Rdbertson, Design Studies
of 1000-Mr(e) Molten Salt Breeder Reactors, ORNL-3996 (August 1966).

3H. S. Booth and D. R. Martin, Boron Trifluoride and Its Derivatives,
Wiley, New York, 1949. : _

Y

-
 

7

)

,NaBF4 - NoF - KBF4 (90-4-6 mole %)

386’ /
398¢°
NoBF, / :

 

" B4ge

 

 

 

ORNL-DWG €8-5560R

 

4 . KBFy
408° )
384£'°8 ) A (576°)
4000 S
450° — - 4 ,
328° 700 180" e 00e 0\ 355
o _ — = 4400
. g0/ — o
650, .. B50° - ' s50°
L ol
900° - . ,
f 650°
9502 - L1000
© NoF A o
' . 780°
800°
NaF &= A \/ e L LN L \ KF
Fig. 1.. Space Diagramuof the NaF-NaBF,-KBF,-KF System.

Little is known about the corrosion behav1or of the BF3 gas. A few corro-

~sion exper:unents have 'been run wz.th very dry BF; gas »"

and llttle appre-

c:La.ble attack was found up to 200°C on metals such as brass y sta:.nless

steel nz.ckel, Monel, mlld steel, and many others. o -
ThlS report describes the first comprehens1ve study of the compu

__patlbllity of a relatively 1mpure fluoroborate sa.lt with Hastelloy N

: a.nd Croloy OM alloys.

‘The e:@erlments were de51gned to yield :Lnfor-

mation on temperature grad:.ent mass transfer, the major form of corro-

_ sion in fluorlde salt systems Two loops were operated'W1th NaBF4-NaF-KBF4

(90-4-6 mole ‘75) salt at a maxmm temperature of 605° C ‘W‘l‘bh 8 tempera.ture

,",dlfference of 145° C to obta.ln the data presented. These temperatures .

- ;reactor. -

' EXPERIMENTAL PROCEDURE

- match those proposed for fluoroborate sa.lts in 8 molten saJ.t breeder

. The natural circulation loops for this progrem were of the type .

'isho&nfin Fig. 2.

 

| Flowresul'bsfi'omthe difference in'density:"of-" the

“F. Hudswell, J. 8. Nairn, and X. L. Wilkinson, J. Appl Chem

| 1, 333 (1961)
 

 

 

 

 

 

 

 

 

 

 

 

 

™~
=+
o
w
o
-
o
=
Q.

 

Natural Circulation Loops in Test Stands.

2.

Fig.

 
 

”

-

galt in the hot and cold pertions of the loop. We estimated the velocity
of salt flow in the test:loops to be 7 ft/min. |

Materials Section and Fabi'icatioh -
" The Croloy 9M materle.l was from Bebcock & W:Llcox Company heat 18760

a.nd was vapor blasted before fabrication to remove oxides. The loop,

. '_NCL-12, was fabricated from-O. 750-in. -OD tubing with'a wall thickness of

0.109 in. The_-material "was_TiG welded and inspected to meet ‘existing

“ internal standards. The welded ereas were torch annealed before and:
‘after welding. .. - | S

The Hastelloy N materla.l was from Union Carbide Corporation,

Materials Systems Division heat 5096. The loop, NCL-10, was fa.brica.ted

from 0. 672-in. -0D tubing with a wall thickness of O. 062 in. , TIG welded
and inspected to meet the seme stendards as those required for the
Croloy 9M material.

The conmos:.tions of both alloys are given in Table 1.

Table 1. Chemical Composition of Alloy Test Materials

 

_ Composition (wt %)

 

 

Alloy — , ,
: Ni Cr M  .Fe C M S P si
Hastelloy N 70.8 7.47 15. 59 2,. 01 0.07 0.5 0.005 0.6
(NCL-10 :
Croloy 9M - ,8 87 0 98 89.00 0.09 0.48 O 012 0.010 0.47
- (NeL-12) - | | A | )

 

Sa.lt Prepa.ratlon

- The sa.lt was prepa.red 'by the Fluor:.de Process:mg Group of the

B . Reactor Chemistry Divlslon at ORNL. ~ This was their first experience
_ _w:.th a fluoroborate salt, and they prepared it by techniques established
__'for other _fluor:.de__s_a_lts._ = The___zja._w materials —~ relatively mpure NaBF, 9

 

SMSR Program Semiann. Progr. Rept. Jan. 31, 1967 ORNL-3626, p. 146.
 

‘NaF, and KBF, (90-4-6 mole %) (Table 2) — were loaded into a container
lined with nickel, which was then evacuated and purged several times
with helium.' Then, the materials were melted and heated to 400° C under
" a helium atmosphere. Next, the liquid was sparged with hélium."sinée
this caused a large increase in pressure, the reactor vessel was vented.
Large quantities of BF3, which had caused the increase in pressure, were
" then released. Later steps included sparges with a mixture of hydrogen
fluoride and hydrogen for several days at 550°C to remove oxides and a
sparge with hydrogen to remove structural metal impurities. The salt
was then transferred to a fill tank made of Hastelloy N in preparation
for filling the loop. _

The chemical analysis (Table 3) of the prepared selt disclosed two
important compositional changes germane to this test series:. (1) a
significant loss of BF; during preparation, and (2) a high content of
oxygen and water. Since little was known asbout the corrosivé behavior
of this salt or the effect of impurities and because information on its
compatibility with the container alloys was needed quickly, we decided

to use it as prepared.

Table 2. Composition of Salt Mixture Before Purification

 

Mole % | Wt %

 

Quantities ILoaded in Container

Compound
KBF,, | - 5.88 6.85
NaBF, 90.20 | 91.61
NaF | 3.92 | . 1.53

‘Chemical Analysis

Element
K | | o | 2.10
Na | 20.00 -
B | 9.60
F | 68.30

 
 

ot

n

 

- Teble 3. Chemical .Ana.lysis“of ‘Salt Before Fill

 

 

Element ' B Content (wt %)
kK R 2.20
Na o o 25.80
B o | | 9.65
Fo L 60440
NG < 5%
cr S - A
Fe I | e
s e
0 - ' 3000%
Bo . 900, 2100%

 

a'Pa.rts- p_ei-_'. million.
' OPERATIONS

The hot portion of each loop was heated by sets of clamshell heaters,
with the input power controlled by silicon controlled rectifiers .(SCR

units) and the temperature éont_rolléd by a Leeds and Ndrthrup Speedomax H

Series 60 type C.A.T. (current proportioning) controller. The loop tem-
peratfires were mea'sured by Chromel-P vs Alumel thermocouples that were
spot welded to the. outside of the tubing, covered by a la.yer of qua.rtz
tape, and then covered W1th stalnless steel shim stock.

- Before each loop was filled with salt, it was degreased with a.cetone

‘and heated to l?O‘?,C ~under v_s_a._c_:gum }:o_ remove any moisture in the system.
 

We checked for leaks with a helium mass spectrometer leak detector while
the interior of the loop was evacuated to < 5 X 1077 torr. All lines
from the fill tank to the loop that were exposed to‘the fluoroborate _ ¥
salt were of the same material as the loop and were cleaned and tested “
in the same manner as the loop. All temporary line connections were
made with stainless steel compression fittings. | |
The loops were filled by heating the loop, the salt pot, and all
connecting lines to a minimum of 530°C and applying helium pressure to
the salt pot to force the salt into the loop. Air was continuously
blown on the freeze valves leading to the dump and flush tanks to pro-
vide a positive salt seal. Tubular electric heaters controlled by vari-
able autotransformers furnished the heat to the cold-leg portions. Once
the loop was filled the heaters were turned off and the insulation was
removed to obtain the proper temperature difference by exposing portlons
of the cold leg to ambient air.
The first charge of salt was dumped after 24 hr in the loops at the
maximum operating temperature with some circulation and little tempera- /
ture gradient. This flush removed surface oxides and other impurities
that could have been left in the loops. The loops were then‘refilled : | ~
with qéw salt and put into operation. A helium cover gas under slight |
positive pressure (about 5 psig) was maintained over the salt in the
loops during operation. Each loop was operated at a maximum temperature
of 605°C and with a temperature difference of 145°C. A temperature pro-
file around the loops is shown around the schematic of the loop in Fig. 3.
During circulation each loop contained about 920 g of salt that con-
- tacted 1200 cm® of surface and traveled 254 cm around the loop.
Temperature excursions indicated a flow stoppage in NCL-12
(Croloy 9M) after 1440 hr. Attempts to drain the loop were unsuccess-
ful, and the loop was allowed to cool with the test salt in place. In
- NCL-10 (Hastelloy N),Va significant increase of temperature in the hot

leg accompanied by a simultaneous temperature decrease in the cold leg .
occurred after 8335 hr of operation. A perturbation of this type indi- t
cates a disruption in salt flow and can indicate plugging. This tem- -

perature cycle ceased after 1 hr, and no further incidents occurred Q;J
 

"

L4}

o

ORNL-DWG 68-11780

    
    

SURGE TANK

   

 

 

 

 

 

sosec L=l |
T —
TS
OsS
| g
| v
| . | sa0%c
VT
L o \
o CLAMSHELL !
0 _}7 HEATERS l
! ' 1
| B . -
e |
el | —NsuLaTioN |
32.51in. { I i,/
ANt _
i L
P!
s |
| | '
| |
| 1 :
LUl
540°C \' =
kN

 

 

DUMP TANK -

 

‘Fig. 3. Schematic of MSRP Natural Circulation Loop.

during the life of the lbop-,:The_looP’was shut down after 8760 hr

(1 year), and the salt was dréified‘intoda @ump'tank-in.a_fiormal manner.

'RESUI.TS :

Analyses of loop components and salt were made by standard chemical

analy31s, metallographlc examlnatlon, and electron mlcroprdbe analy31s.

‘The results are dlscussed below

'NCL-lO'(Hastelloy N)

Visual. -A partlal plug 4n the lcwer part of the cold leg (Flg. 4)

_closed approximately 75¢ of the cross-sectional area of the plpe._ Analysis
 

 

 

10

 

TUBE WALL

  

 

NO; CrFs PLUG

 

 

Fig. 4. Plug Formed in NCL-10 (Hastelloy N) Containing NaBF,-NeF-KBF,
(90-4~6 mole %) After 8760 hr at a Maximum Temperature of 605°C and Tem-
perature Difference of 145°C. 8X. Reduced 40%.

showed this emerald-green plug to be single crystals of NasCrFg (ref. 6).

Smaller amounts of two complex iron fluorides, NajFeFg and NaFeF,, were

also identified in the cold leg. | '
Chemical. — The analysis of the salt ceke (Fig. 5) collected from

- the dump tank of NCL-10 is given in Table 4. The concentrations of

nickel, molybdenuin, iron, and chromium in this salt were higher than

those of the salt before test. The average chromium concentration was

470 ppm. | |
Due to cooling, nickel was segregated in the top and bottom portions
of the cake, reaching 11.15 and 4.47 wt % in these portions, respectively. ’

‘The analysis indicated that the water content of the salt — 400 to 900 ppm -

 

6MSR Progrem Semisnn. Progr. Rept., Aug. 31, 1967, ORNL-4191, p. 228. O
L2

o Bottom layer 4.47°

11

 

 

BPhoto 74747

Fig. 5. Drain Salt Cake From NCL-10 (Hastelloy N).

Table 4. Impurity Analysis: NaBF,-NaF-KBF, (90-4-6 mole %)

 

Anelysis (ppm)

 

Ni Cr .M

- PFe

0

H,O

 

Before test 87 83 7

After test
8

Top slag  11.15 &

* 1000 1.35
Center layer 90 ~ ~ 210 160

® 1500 7300

146

4200

270

1500

1400
3000
4850

1200
1750

3540
3120

3550
3660

400
900
1800
11300

2800

2,6

 

Seight percent.
 

 

12

before test — increased to 1300 to 2800 ppm. The oxygen analyses showed
mich scatter, but we believe that the oxygen content also increased.

. Presumsbly these increases are due not to air inleakage but to other
factors discussed later in this report. |

The crossover line to the cold leg, the cold leg, and the crossover
line to the hot leg all showed slight increases in'wal} thickness due to
deposition of eomplex surface layers (Fig. 6). Chemical analysis of the
layers disclosed that they were primarily'metailic nickel (60 to 90 wt %)
and molybdémum. A small quantity of iron was present 1n.prox1m1ty to the
‘base metal, but chromium was conspicuously absent. o

Metallurgical. — Micrometer measurements of the hot leg of the loop
disclosed 1 to 2 mils of metal loss and slight surface roughening.

Metallographic examination of an area from the hottest section
_(605°C), however, showed a smooth surface with an 6ccasienal penetra-
tion along a grain boundary (Fig. 7). Microprobe traces’ of this hottest
section for all the alloying elements showed no compositional gradients.
The chromium trace for this section is givenrin Fig. 8. These results
indicate a general dissolutive attack in the hot-leg section.

An area at the entry of the cold leg (520°C) showed a duplex surface
structure, and the areas numbered in Fig. 9 were analyzed with the micro-
probe for the various elements (Fig. 10). The outermost layer, aresa 1,
is quite high in nickel ~ up to 87 wt %. We believe this is most likely
a region where nickel deposited during the test. Area 2 is a region con-
taining essentially only'molybdenum and nickel. Close to the original
metal surface there is an increase in nickel, a smaller increase in
molybdenum, and little change in chromium and jron. Note that with an
increase in nickel and molybdenum metal the concentration of chromium
and iron would show a decrease even though there were no chenge in the
actual amounts. '

Figure 11, a photomicrograph of the coldest section NCL-lO (460°C)
shows a spongy deposit on the surface and a loosely adherent corr031on
product beyond this region. At higher magnifications, it was seen that
the tightly attached spongy deposit varied in thickness from O to 6 um

7Da.ta not corrected for absoxptlon, secondary fluorescence or atomic
number effects.

iy
o N

 

  

 

 

 

“‘; ‘! ‘ 1 éll | ;;32.5in.
- 0007 inches '

 

  
    

 

 

 

 

 

 

. NeLo

| SALT: NaBF,NaF-KBF,; (90-4-6 mole %) -
TEMPERATURE :605°C, AT= 145°C, TIME =8760hr . ‘

  

 

 

 

L) )

o "

ORNL-DWG 67-9342R

 
 

LOCATION OF

 

 

Naz Crfg PLUG\_

 

 

 

 

 

 

Fig. 6. 'Surfa,ce' Layers in NCL-10 (Hastelloy N) After 8760 hr Operation.
 

 

 

 

 

 

 

 

 

 

 

 

   

Y-81402 __

I-.

 

 

 

 

 

e

 

[

0,007 INCHES
& 500x |

 

Jer

 

 

 

I

 

 

 

o

 

 

0i00175 INCHES
"2000x !

 

 

 

 

Fig. 7. Hastelloy N Hot-Leg Section NCL-10 (605°C) Shown at
Different Magnifications After 8760 hr in Fluoroborate Salt. Etchant:
Glyceria Regisa.
L t:

 

I . " +

 

ORNL-DWG 68- 9658

 

 

 

 

 

 

 

 

 

 

 

 

L 10
L ‘
o
. & L Oy Oy - O A 0
i_D . ;—..‘ o W ~ \J O O U L) o
gV &08 ‘
= o
s 2
o | &
x =~ 9 6
Q- ke
B é L
1 =
o z TRACES OF Ni,Fe, AND Mo
N o SHOWED NO GRADIENT
L s ~
:—-———\A————-
| - Bp o — — _ —
o {‘O : 10   ; | - : 20 - 30 S ‘  » 40

Fig.‘S; Penetration Curve of Chrom
Fluoroborate Salt. | R

DISTANCE ()

ium ithbfi7Beg (605°C) of NCL-10 (Hastelloy N) After 8760 hr in'

2
 

 

 

 

 

 

 

 

 

 

 

 

16

Y-83857

N

 

[

 

0.007 INCHES
I 500X

1w

>

 

 

 

 

 

0.00175 INCHES
I

 

 

 

 

(b)

Fig. 9. Entry to Cold Leg (520°C) of NCL-10 (Hastelloy N) Operated
for 8760 hr in Fluoroborate Salt. Etchant: Glyceria Regia. (a) 500x.
(b) Numbered areass were analyzed by microprobe. 2000X.

 

 
 

) 1)

 

 

 

 

0 @ ORNL-DWG 68-9659
/ ) TS
| \= I _ | / ™~ Ni _
- . I | | s | So--—o-—-.-—-o..___‘,__.__. & o
N/ ——
N

 

-
e

H L
o . :

e LAYER ————=1  oriGINAL -
b sureace
g

ESTIMATED

 

COMPOSITION (wt %)

 

20 }

 

 

0.

Fig. 10. Penetration Curve of Constituents of Hastelloy N at Cold-Leg

0

e ® e =3 ::....‘-—.L—'“‘ =
_ 0

 

 

e © o s e e g U s
’."-. ¢ .
' Fe

 

 

 

 

20

3

DISTANCE (1)

0 | 40

Entry (520°C) of NCL-10 Operated for 8760 hr in Fluorcborate Salt.

LT

 
 

 

 

 

 

 

 

18

Y-81411

 

1-5

I

 

 

 

 

 

 

|

0.007 INCHES
& 500X

 

 

 

 

 

o

 

 

 

1%

 

 

 

 

Y-824u0

 

o

 

F2060ox 1

0.00175 INCHES
5

 

 

Fig. 11. Coldest Section (460°C) of NCL-10 (Hastelloy N) Shown at
Different Magnifications After 8760 hr in Fluoroborate Salt. Etchant:
Glyceria Regia. ’

an

3

 
 

"

w)

 

19

and that it resembled the area seen in the entry to the cold leg, &lthough

~ much thinner. The spongy layer is high in nickel, with some molybdenum,

and sppears, like the;area 1l of Fig. 9, to have resulted from nickel
deposition (Fig. 12). There was only 63% metal in this spongy layer;
the balance was fluoride compounds. The original surface is high in
nickel and molybdenum and low in chromium and iron and is an area where

nickel and molfibdenum have diffused into the matrix. No analysis of the

loosely adherent material Was possible

The last section of the 1oop examined'was an area at the entrance

_ _to the hot 1eg Although the salt was being heated.here, the temperature

was st111 low enough (< 530° C) to make this an area where materlal'was

' belng depos1ted. Again, a duplex layer'was seen (Fig. 13) end found to

be predominantly nickel, 657tor90% metal and the balance fluoride com-

. pounds (Fig. 14). X-ray diffraction analysis of the fluoride compounds

disclosed an unidentifiable crystal structure, probably a mixture of
several fluoride compounds. The surface layer had somewhat more nickel
and molybdenum‘than originally,'and, thus, smaller concentrations of

chromium and iron.

+ NCL-12 (Croloy 9M)

Visual — When this”loop'Was sectioned for examination, a dark

gray'plug that completely fllled the cross-sectional area fOr a vertical

_: distance of 1 in. was found 1n the cold leg (Fig. 15) Also, small,
~ green crystals were seen in the drain line, and a metallic layer was
'found against the 1n31de of the tubing in the cold section and in the

The th1n metallic 1ayer-about 2. 5 mils thick was dep081ted on the

in31de of the tubing in the entire cold-leg section (over one-half the
loop) In some places, the materlal had become detached from the tubing
f:;and'was found in the frozen salt (Flg 16) The metallic layer repre-
| sented sbout 4% of the total mass of the mater:l.al (mostly salt) removed
o Vfby'mechanical mesns) from the loop. Chemical analysis revealed the layer
-tobe90wt%Feanlewt%Crmetal.
 

ORNL—DWG 68—9660

 

_ |
80 r63 7% METAL
| BALANCE
FLUORIDE
SALTS

 

 

 

|
|
0 | I

/\

 

ESTIMATED
ORIGINAL
SURFACE

 

 

40\!

 

COMPOSITION (wt %)

SPONGY
et e e———
LAYER

 

 

20

 

 

 

 

 

 

 

 

 

 

 

   

20 R - 20
DISTANCE ()

Fig. 12. Penetration Curve of Constituents of Hastelloy N in Coldest Section
(460°C) of NCL-10 Operated for 8760 hr in Fluoroborate Salt.

0c

 
 

 

 

 

 

 

 

 

[-

 

0.007 INCHES
5 500X I

[on

 

 

 

t2000x 1

0.00175 INCHES
I

 

 

 

 Fig. 13.  Entry to Hot Leg (530°C) of NCL-10 (Hastelloy N) Shown at
Different Magnifications After 8760 hr in Fluoroborate Salt. Etchant:
Glyceria Regia. ' ' ' ' o
 

 

22

ORNL—~DWG €8-9€61

 

80

  

 

 

 

 

 

 

 

 

e . ‘—.—Ii—.—fl——‘—.-;d'—'_.fi_._
¢ o “Ni
|
|
i
|
65-90% METAL |
60 |— BALANCE  +
FLUORIDE SALT | ESTIMATED
= ‘ ! ORIGINAL
5 -~ LAYER , SURFACE
= I
5 |
- {
o 40 1
o |
a ]
s I
O
o i
|
|
|
20 :
| .
P __,..—-:—1.-—0—o—o—-c—dr—ojo—._._.,
V4 ' .‘3—0'—.--0—0—0—-.—-0-—0—-1.—.—0—0 “s...
0 * s__-—l-‘
0 10 : 20 30 40
" DISTANCE (p)
- Fig. 14. Penetration Curve of Constituents of Hastelloy N in Hot~-

Leg Entry (530° c) of NCL-10 Opera.ted for 8760 hr in Fluoroborate Salt.

 

P
£

Fig. 15.

(croloy oM) Qperated in Fluorcoborate Salt for 1440 hr at a Maximum Tem-‘

 

     
 

 

FLUOROBORATEVSALT ’

 

Iron Dendrite Plug in Coldest Section (460° C) of NCL-12

perature of 605°C and a Terperature leference of 145°C.

%
 

 

 

 

 

 

DEPOSITED METAL LAYER

 

Fig. 16. Cross Section of Tubing of NCL-12 (Croloy 9M) Showing
Deposited Metal Layer. Loop operated for 1440 hr in fluoroborate salt
at g maximm temperature of 605°C and e temperature difference of 145°C.
Reduced 22%. (a) Cross section of NCL-12 tubing. (b) Enlarged view of
metal and salt interface. 30X, = S , - '
 

24

The dark-gray'plug located in the coldest part of the 1oop'was com~ -
posed of dendritic crystals. we found sxmllar crystals adhering to spec1-
mens in the hot leg (Fig. 17), but we assume that these crystals, growing
and circulating in the salt stream, attached themselves_tb-fhe-specimens
while the loop was cooling. -

Chemical. — Chemical analysis showed that the dark-gray plug and the
material on the specimens were essentially pure iron with less than 1%

of other elements (shown below).

Elements ~ Content (wt %)
Fe ) 99.00
B 0.03
Cr < 0.05
Mn < 0.01
Mg 0.05
Pb < 0.02
Si 0.02
Cu © 0.05
Mo 0.02

The results of chemical analysis of the green crystals in the drain leg

are given below.

Elements wt %
Na 7 7 1015
B | 24
F o 45.9
K | ~ 0.058
Fe - | 18
Cr | 12
Mn , | 1.5

The stoichiometry of this green depbsit, based on this analysis,
color, and other factors, is roughly 2NaF-FeF,-CrF;-BFs, which corresponds

to chromium and iron fluorides mixed with the salt.
 

 

 

 

73674

TR R

 

 

' Fig. 17. Pure Iron Crystals from NCL-12, Which Operated for 1440 hr
in Fluoroborate Salt at a Maximum Temperature of 605°C and a Temperature
Difference of 145°C. 10X. Reduced 12%. (a) Crystals adhering to speci-
men, and () crystals removed from specimen for photographing.

~ -
 

 

26

Table 5 gives the composite of the salt before and after operation.
The significant changes in the salt chemistry due to test are the incresses
in the chromjum and iron content from 54 to 255 and 28 to 700 ppnb respec-
tively; ' |

Table 5. Cdmposition of Salt Before and After Operation in NCL-12

 

Composition (wt %) - Composition (ppm)

 

 

K N B F ©Ni Cr Fe M S 0

 

_ , - Before Test
' Theoretical 2.10 20.00 9.60 68.3

(calculated) | |

‘Before Fill 2.20 25.80 9.65 60.4 <5 54 28 3000

During Fill 1.98 18.83 9.38 66.2 87 83 146 1400
After Test

Hot leg  1.55 19.72 9.29 67.1 265 700 < 20 7 3000

Cold leg 1.89 20.71 9.27 67.3 255 700 < 20 < 2 3200

 

Metallurgical. — Metallographic examination of the hot-leg loop
tubing (Fig. 18) disclosed a fairly smooth surface, and micrometer mea-
surements showed an average 2.5 mil loss from a nominal;pipe diameter.

Electron microprobe analyses were made on tubing fromrthe hot- and
- cold-leg sections of NCL-12: the hot-leg analysis (Fig. 19); consistent
with the metallographic results, showed no iron or chromium concentration
gradients; the cold leg analysis (Fig._20) showed an increase of about
4 wt % in iron concentration and a decrease of about 4 wt % in chromium

concentration (i.e., an iron-rich surface layer)-

iDISCUSSION

i The tests described are the first study of the compatibility of a
fluoroborate salt with container materials of interest to molten=salt
reactors. Unfortunately, little was known at the time of the test about

the purification'of the salt or the characteristics of mass transfer in

A
 

 

 

 

[ I

 

 

 

I~ 100X

e

 

 

Fig. 18. Hot Leg (605°C) of NCL-12 (Croloy 9M), Which Operated for

1440 hr in Fluoroborate Salt.

a.nd ethyl alcohol.

-Etchant: Picric acid, hydrochloric acid,

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

DISTANCE (p2)

o4 'ORNL-DWG 68-9663

25 90 foo—o - oo l oo o o 2|
'&(fl —~ o0 o o o "0.©
= & ° ° °
— 0|0
n2 % g6 '
h.lg S

s 8¢ |

| & el 1l 8 e.e o Y

| E 3 et eyt — et _ L,

: - : 8 8 ] Cr
s
l Lo Lo 4 T s

l 0p - 0 b— — _ L

i 0 100 = 200 300 -

400

Fig. 19. Penetration ‘Curve for Iron and Chromium in the Hot Leg
(Croloy,?M) » Which Operated for 1440 hr in Fluoroborate

(605°C) of NCL-12

Salt.
 

28

ORNL-DWG 68-9662

 

 

 

 

 

 

 

 

 

 

 

 

 

94
o S o | o .
90 \ 5 : ' o - o
< O~ 0 0 o o 0
& &cc - Y g0 o " Y0o0o ¢ o 3 Fe
= o o c o
~ 86
C':D 12
&
1t 8 7 .
Q ®
= /
19 °
© {
®
4
0

 

0 100 200 - .300 | 400
| | | DEflANCE(yJ

Fig. 20. Penetration Curve for Iron and Chromium in the Cold Leg.
(460°C) of NCL-12 (Croloy 9M), Which Operated for 1440 hr in Fluoroborate
Salt.

systems containing such salts. The use of loops of an old design pre-
vented our having any removeble specimens, but permanent hot-leg specimens
were used in NCL-12.

Our expectation, based on previous experience w:ith fluoride salts,
was that temperature gradienf mass transfer of the least noble constit-
uent (i.e., chromium and iron) would be the limiting factor. But salt
analyses after the test indicated that all the major alloying elements
of the container materials had mass transferred. The nonselective attack
was also: confirmed by metallographic examination and microprobe analysis
of the loop specimens and piping. In view of the mofrement of nickel and
molybdenum, it is obvious that highly oxidizing conditions, due to water
and oxygen in the salt, were present during the operation of the loops.

In analyzing NCL-10 and -12, we reviewed earlier studies® (1late

1950's) to single out the cases of mass transfer of nickel and molybdenum.

 

83. H. DeVan, unpublished data, 1957-1959.

i
 

-y

i

 

29

‘We. found several instances where this had occurred in salts containing

KF. One such 1oop, constructed of Hastelloy N, operated for a year with

‘a NaF-LiF-KF salt (11. 5=46. 5-42 mole %) at & maximum temperature of 690°C

w:Lth a temperature dlfference of 90°C. No anslysis for oxygen or water

‘was made before test. Examination after test disclosed green crystals

embedded in the salt. Sa]:.t_analyses showed significent increases in
nickel, iron, molybdenum, and chromium content. X-ray analysis showed

that the green crystals were mixtures of sodium and potassium chromium

fluoride complex compounds and that most of the KF present in the salt

'was actually KF-2H>0.

, Another loop, constructed. of Inconel s (N1—18% Cr—lO% Fe) operated
about one-half Yyear with the seme sa:l.t and at the same temperatures as

above. After test, metallographlc examination showed heavy attack in

-the hot-leg portlon of the 1oop After test only the chromium and iron

contents of the salt were determined; the chromium content had increased
significantly from 60 to 900 ppm. X-ray analysis of the salt disclosed

7 -that the salt was about 15% KF 2H20 The relatlvely well~ deflned x-ray
-pattern of this hydrate suggested that it was not formed when the cold
melt was exposed to air but was carr::.ed as a part of the salt mxture

& The compound KFe 2H20 is known to be thermally stable.

The literature states that KF easily forms a series of crystal hydrates,

whereas LiF and NaF crystallize anhyd.rously.g It is also noted that it

is easy to form KBF30H and‘NaBF3OH and that they are quite stable.
Thus we have seen. that dur:mg the molten-sa.lt reactor corros:l.on

durlng ‘operation.

f-program, salts conta:ming KE‘ have on occaslon ‘been very sggressive toward
|  metals. ' We believe that the reason for this is the combined water asso-
clated with that alkali metal fluoride selt. We agaln stress that we

' fou.nd no leakage of air 1nto any of the loops. This suggests that

hydrated KF_ is not removed by the pur:l.flcatlon process. Paradoxically,

. it appears that combined water 1n the fluoroborate salt m:.xture can be
S released and w:.ll react to produce HF by |

Hgo +, BF4 —-HF + BF30H .]_ (@)

 

| 9I. G. Ryss, The Chem:Lstry of Fluor1ne and Its Inorgan:r.c Compounds,

pp. 521-28 and 815-16, AEC-tr-3927, Pt. 2, (February 1960).
 

 

30

The generation of HF in the system leads to the following reactions with

~the elements of the container material:

2HF + Ni <NiF, + Hp , - (3)
6HF + Mo = MoFg + 3Hz , (4
2HF + Fe =FeF; + Hp , \ (5)
2HF + Cr =CrF, + Hy . | (6)

~ Note that, in the temperature range of interest, changes in the
standard free energy of formation are not favorsble for all the reactions
as written. waevér, studies on the Fluoride Vblatility Processing
schemel® at ORNL showed that in a Hastelloy N hydrofluorinator containing
fused fluoride salts and HF, chromium and iron were rapidly leached from
the alloy at elevated temperatures (500 to 650°C) evenjwhen the HF activ-
ity was quite low. The main reaction is the oxidation of the metal by
the HF to form fluorides soluble in the melt. It was found that NiF» is
produced by driving the reaction through the continuous removal of hydro-
gen and reaction products, since the free energy of formation (above
490°C) of the nickel fluoride reaction by Eq. (3) is not favorsble. For
the same reason, molybdenum can also be forced to react with HF even
though a positive change in free energy is involved. Evidence of small
but finite dissolution rates of molybdenum metal during hydrofluorination
conditions have been reported.ll,1? i' '

In light of the discussion above, we can now consider why two dif-

ferent types of products (i.e., metal and fluoride compounds)’were mass

transferred in these (NCL-10 and 12) thermal convection loops.

 

105, P. Litman and R. P. Milford, "Corrosion Associated with the
Oak Ridge National Laboratory Fused Salt Fluoride Volatility Process,”
paper presented at the Symposium on Fused Salt Corrosion at the Fall
Meeting of the Electrochemicsl Society, Detroit, Michigan, Oct. 15, 1961.

114, E. Goldman and A. P. Litman, Corrosion Associated With Hydro-
fluorination in the Oak Ridge National Laboratory Fluoride Volatility
Process, ORNL-2833 (November 1961).

12p. P. Litman, Corrosion of Volatility Pilot Plant MARK I INOR-8
Hydrofluorinator and MARK III L Nickel Fluorinator After Fourteen
Dissolution Runs, ORNL-3253 (Feb. 9, 1962).

 
 

o -

. -

 

31

Thermodynamics of System Corrosion

The cycle'of mass transfer initiated by the corrosion of metals by
fluoride salts is kncwndte‘hegin'in;the hotter regions of a loop system
with the formation of'structural metal fluorides that are soluble in the
salt: ' '

M+F =M , (7))
where Mis the attacked metal of the container.

The equilibrium constant for Eq (7) increases with increasing tem-

perature; thus, in a multicomponent alloy the concentration of the
attacked constituent, M, Will_decrease at loop-surfaces at high tempera-
ture (weight loss) and increases at surfaces at lower temperatures (weight
gains). At some intermediate temperature, the initial surface composi~
tion of the structural allow'w1ll be in equllibriumJW1th the fused salt
(no weight change). | |

In the cooler reglons of & closed system, the rate at which the

- metal is depos1ted (steady state condition befbre plugglng) is ‘generally

equal to the rate at which 1t diffuses into the metal — g necessity to

maintain its equilibrium activity on the surface. - Thus the rate of

diffusion of the metal in the alloy in the cooler regions usually con-

ltrols the rate of corros1on in the hot zone. At this stage in the temr

perature-gradient mass-transfer process,'weight gains would be found on
specimens in the cold section.but little if any surface change (depos1ts)
would be seen. This generalization is valid only'when the equilibrium
concentration of a metal fluoride compound in the salt [produced‘by |

. (7)1 does not exceed the saturation concentration of a metal fluoride

rai;compound in the salt at the lcw temperature.

‘Early studies reported.by Grimes13 clearly illustrate these p01nts.'
Table 6 shows the chramxum concentrations for two salts as functions of

'the equilibriumflbetween the salts and Inconel or pure chromjum. Note

that when the alkali-metal fluoride salt (NaF-KF-LiF-UF4) is in contact

'JFW1th Inconel at 800°C it'will support a higher concentration of chromium

 

3y. R. ermes, ANP Quart. Progr. Rept June 10 1956 0RNL-2106
Pp. 96-99. _
 

 

 

32

Table 6. Equilibrium Concentrations of Chromium
Fluorides With Two Salts

 

Chromium Concentration (ppm)

 

 

NaF-KF-LiF-UF, NaF-ZrF, -UF,
Calculated salt equilib-
rium with Inconel
N. = 0.16 at 800°C (ref. a) 1660 1400

Cr
Experimental salt equilib-
rium with pure chromium

Ny, = 1.0 at 600°C (ref. a) 1100 | 2400

 

a‘Conccantra.t:ihon of chromium in mole fraction.

fluorides (1660 ppm Cr) than it will when it is in contact with pure

chromium at 600°C (1100 ppm Cr). Accordingly, circulation of such a

salt in an Inconel loop would result in the deposition of essentially

pure chromium metal in the cold zone. 1In that case, the total rate of

attack would be controlled simply by diffusion of chromium to the inter-

face betweén metal and salt in the hot zone. We believe that the deposi-

tion of pure iron in NCL-12 (Croloy 9M) occurred this way. This is not

& new phenomenon. Rapid plugging by deposition of metal dendrites occurred

frequently in the late 1950's especially in iron-base alloy loops.l%4,1>
Examination of the data for the other salt shown in Table 6

~ (NaF-ZrF,-UF;) leads to the conclusion that chromium metal would not

plate out in that system and that any product of mass transfer would be

_a fluoride compound. At 600°C this salt in equilibrium with pure chromium

will support a higher concentration of chromium fluorides than it will

 

4G, M. Adamson, R. S. Crouse, and W. D. Manly, Interim Report on
Corrosion by Alkali-Metal Fluorides: Work to May 1, 1953, ORNL-2337
(March 20, 1959).

15G. M. Adamson, R. S. Crouse, and W. D. Manly, Interim Report on
Corrosion by Zirconium-Base Fluorides, ORNL-2338 (Jan. 3, 1961).

 
 

.

i

33

~in contact with Inconel at 800°C. For gross deposition of compounds to

-occur, the concentration of the compound at the temperature of interest
(cold zone) must exceed the saturation concentration of the metal fluoride

 corrosion product present in the system.

We believe that the mechanism in NCL-10 (Hastelloy N) is similar to
the one discussed above for the NaF;Zer system and that in time the con-
centratlon of chromium.-or more accurately the concentration of mixed
metal chromium fluoride -1n the fluoroborate salt at 605°C exceeded the
saturation concentration at 460°C and allowed deposition of large quan-
tities of complex compounds. | |

Both NCL-10 and NCL-12 operated with the same salt at the same tem-

perature, yet they'were plugged by different mechanisms at different

rates. A chromium-rich plug was found in the Hastelloy N loop (containing |
7% Cr—5% Fe) and an iron plng'was found in the Croloy 9M loop (containing
9% Cr-bal Fe). Thus, it appears that when these fluorcborate salts

~are contained in alloys with between 7 and 9 wt % Cr, the iron content
of the alloy controls the comp031t10n of the temperature-gradaent

mass~transfer dep081t.

| Kinetics of System Corrosion

With the fOreg01ng data and evaluatlon, it is now p0381b1e to deter-
mine when plugging started ;n_NCL-lO and =12 and the dlsp031t10n_of each
element at various times. . R

 We calculated the interim concentrations for chromium and iron in

_ NCL-12 from knowledge of . the amount of these elements after test, the
"welght of the iron plug, and an assumed reaction-rate constant and mode
of chemical attack (from later experlments) 16 The results of the cal-

culatlons, presented in Table 7, show that the saturation value of iron

" in the cold section, 700 ppm or 1120 mg, was exceeded shortly after
:. 130 hr; at that tlme pure 1ron started dep081t1ng and eventually caused
e complete plug. o | R

Table 8 shows the same klnd of calculatlons for the operatlon of
NCL-lO. In thls_case, it,appears that the saturation value of chromlum

 

167, W. Koger and A. P. Litman, MSR Program Semiann. Progr. Rept.,

__Féb 29, 1968, ORNL 4254 PP 218—225.
Table 7. Calculated Concentration of Alloying Elements from NCL-12
Present in the Salt at Various Times

 

‘.Time - Maximum  Aversge Area Total Deposited in

Concentra.t ion in Salt

 

 

 

m b Deposited
(hr) Attack Attack Attacked™ Material  Ribbon Form Fe Cr As Plug
" (mg/em?)  (mg/em?) = (cm?) Iost (mg) (mg) (mg)
(mg) (ppm) (mg) (ppm)
1.4 2 1 650 650 250 360 225 40 25
57.6 4 2 650 1300 500 720 450 80 50
130 6 3 650 1950 750 1080 675 120 75
1440 20 10 650 - 6500 2450 1120  700° 400 250° 2530

 

B0aloulated from M = 24/% o ovkt
where K = rate constant, 10~? cm?®/sec at 607°C,

MW = weight loss,

¢ = concentration of chromium snd iron,
p = density of Croloy 9M, and

t = time.

bIncludes ‘surge tank and lines and one-half of total area.

By chemical analysis.

e

 
 

®a !7 | . - ‘. , 1

 Table 8.

Ca.lc_:ulated Concentratlon of Alloying Elements from NCL-lO
Present in the Salt at Va.rlous Times

 

Ares .

Concentration in Salt

 

Ma.:dnmm Average : '.

 

 

 

Time ' 'b . Total Deposited on
(hr) Attack®  Attack. Attacked Material - Cold ; Fe Cr = Ni
(mg/mnz) (mg/cm?) (cm ) 1 Dost (mg) ‘Surface — —— g ‘
. Lot i mg) (ppm) (mg) - (ppm) (mg)  (ppm)
4000 :33':@1 L1900 550 ‘,..':-10,450 2100 350 383 420 459 6350 6950 1150 1260
8760 .56 28 . 550 15,400 3300 509 555 427 470 9300 10,000° 1560 1700°
aée.'ltmiated fr‘oifi Mz = Z/J- o o‘/f{_
where K= ra.te consta,nt, 10-12 2/:sec: at 607°C,
W = Welght loss, |
e = concentration of chromium, iron, nickel, and mol:fbdenum,
P o= density of Hastelloy N, a.nd
| -t = time.
1"’In(:].v.n:'ie:se surge tank and lines and one-half total a.rea..

| By chemica.l ana.lysis. B

ge

 
 

36

in’ the cold leg, 470 ppm, was reached after about 4000 hr of operation;
at that time, the chromium, as Na3CrFg, started deposmtlng in large
amounts. 1’

It is interesting to note that in both cases the amounts of the
alloylng elements in the salt and in the plug were in dbout the same
ratio as they are in the base metal. Capsule tests and other loop tests |
have shown that when conStituents of an alloy react with the liquid medium
they will be found in the liquid and/or deposited in the ratio at which
they existed in the alloy.l® This behavior was found in both loops of
this experiment. '

Nickel and molybdenum mass transfer products, not normally found in
relatively pure fluoride salf systems, also fbllcw this pattern. As has
been stressed here, the presence of gross quantities of nlckel and molyb-

denum indicates strongly oxidizing conditions.

Salt Purification

The salt used in these experiments contained many impurities that
the old techniques, successful with other fluoride salts, did not remove.
Since these experiments, improvements have been made in the fluoroborate
salt preparation. Only very pure salt (99.9%) is now used as starting
material. In fact, the salt has so few impurities that no steps are

 

17 Several suggestions, besides the removal of the water or oxygen
from the system, are offered to improve the service of the container
materials using this fluoroborate salt. The most obvious change is to
lower the hot-leg temperature to 540°C thus lowering the reaction-rate
constant an order of magnitude. It would then take three times as long
to duplicate the previous plugging. Another possibility is to raise the
temperature in the cold-leg section. This would raise the saturation
value of the elements in the salt, and depositing would not occur until
later. But this probably would not improve the life by the same factor as
above. Probably the most important improvement is removing corrosion
Products from the system on a continuous or batch-wise basis since
deterioration (thlnnlng) of the container wall by dissolutive corrosion
is not a problen in these systems.

187, R. Weeks and D. H. Gurinsky, "Solid Metal-Liquid Metal Reactions

in Bismuth and Sodium,"” pp. 106161 in Liquid Metals and Solidification,
American Society for Metals, Metals Park, Novelty, Chio, 1958.
 

R

 

37

taken during processing,to remove the structural metals. Purification

procedures for removing water and oxygen are now under study. A new
procedure is alse used in the.melting and preparation of the fluorcborate
salt to prevent'loss‘of BF3 vapor and the ensuing change of composition.
In brief, the loaded salt is evacuated to about 380 torr and heated to
150°C in a vessel lined with nickel and is held for about 15 hr under

these conditions. -If the rise in vapor pressure is not excessive, (no

volatile impurities) the salt is heated to 500°C, while still under

vacuum, and agitated with helium for a few hours. The salt is then
ready for transfer to the fill vessel.

CONCLUSIONS

1. Natural clrculatlon loops, fabricated from. Hastelloy N and
Croloy 9M, that c1rculated impure (> 3000 ppm.1mpur1t1es) NaBF, -NaF-KBF,

(90-4-6 mole %) at a max1mum;temperature of 605°C with a temperature

difference of 145°C evidenced serious temperature-gradient mass transfer.
The mass transfer involved migration of all major constituents of the con-
stituents of the container materials and resulted in restricting flow in
the Hastelloy N-leop by'dePOSition of Na3CrFe¢ crystals andfcomplete plug-
ging of the Croloy 9M loop by iron dendrites. -

2. The nonselective corrosion observed was due to the presence
of water, chemically'bound'to the fluordborate salts, that reacted
durlng heating to form HF. |

3. The driving fbrce fbr mass transfer'was the temperature depen-_

‘dence of the equllibrrwm constant between the conta1ner material con-
?stltuents and the,most stable,fluorlde compounds that cen,form in the

‘system.

4. The saturatlon concentratlons for iron and chromlum in the test
salt at. 460 C'were found to be 700 and 470 ppm, respectlvely

- 'EECOMNDATIONS |
1. Other, morerhighlfi'purified'fluoroborates should be extensively

tested for corrosion before these coolants can be qualified for molten-

salt reactor service.
 

 

38

2. Based on knowledge to data, iron-base and iron-containing alloys
should.be avoided in molten-selt reactor coolant circuits that use -
fluordborate salts.

ACKNOWLEDGMENTS

_ It is a pleasure to acknowledge that E. J. Lawrence supervised con-
struction, operation, and disassembly of the test loops during'the course
of this program. We are also 1ndébted.to H. E. McCoy, Jr. and J. H. DeVan
for constructive review of the manuscrlpt

Special thanks are extended to the Metallography Group, especially
H. R. Gaddis, H. V. Mateer, and R. S. Crouse, and to the Analytical
Chemistry Division, Graphic Arts Department, and the Metals and Ceramics

Division Reports Office for invaluable assistance.
 

 

- 32.

1-3,

6-15.
16.
17.
18.
19.
20.
2L.
22.
23.
24,
25.
26.
27.
28.
29.
30.
31

33.

’ 3.
i 35.

36.
37.
38.
39.
40.
41.

42.

43.

45,

46,

4T,
48.

49,

. 50.

51.

- 52,
53,

54.
55.

56.

v

-39

 

ORNL-TM-2490
INTERNAL DISTRIBUTION

Central Research Library 58. E. L. Compere
ORNL — Y-12 Technical Library 59. K. V. Cook

Document Reference Section 60. W. H. Cook
Laboratory Records o 6l. L. T. Corbin
Laeboratory Records, ORNL RC 62. B. Cox ,
ORNI: Patent Office 63. R. S. Crouse

R. K. Adams 64. J. L. Crowley
- G. M. Adamson 65. F. L. Culler

R. G. Affel 66. D. R. Cuneo

J. L. ‘Anderson 67. J. E. Cunningham
R. F. Apple 68. J. M. Dale

C. F. Baes - 69. D. G. Davis

J. M. Baker 70. R. J. DeBakker
S. J. Ball 71. J. H. DeVan

C. E. Bamberger 72. S. J. Ditto

C. J. Barton 73. A. S. Dworkin
H. F. Bauman 74. TI. T. Dudley

S. E. Beall 75. D. A. Dyslin

R. L. Beatty 76. W. P. Eatherly
M. J. Bell 77. J. R. Engel

M. Bender 78. E. P. Epler -

C. E. Bettis ~ 79. D. E. Ferguson
E. S. Bettis 80. L. M. Ferris

D. S. Billington 8l. B. Fleischer

R. E. Blanco 82. A. P. Fraas

F. F. Blankenship 83. H. A. Friedman
J. O. Blomeke 84. J. H Frye, Jr.
R. Blumberg _ 85.. W. K. Furlong
E. G. Bohlmann 86. C. H. Gabbard
C. J. Borkowski 87. R. R. Gaddis
G. E. Boyd 88. R. B. Gallaher
J. Braunstein - 89. R. E. Gehlbach
M. A. Bredig 90. J. H. Gibbons
R. B. Briggs .. - ' 91. L. O. Gilpatrick
H. R. Bronstein 92. P. A. Gnadt

G. D. Brunton 93. R. J. Gray -
D.. A. Csnonico 9. W. R. Grimes -
S. Cantor 95. A. G. Grindell
W. L. Carter 96. R. W. Gunkel -
G. I. Cathers- 97. R. H. Guymon
0. B. Cavin el - og., J. ‘P. Hammond -
J. M. Chandler - 99, B. A. Hannaford
F. H. Clerk 100. < P. H. Harley - -
W. R. Cobb 101. . D. G. Harman:
H. D. Cochran 102. W. O. Harms -

C. W. Collins

103. C. S. Harrill
 

104.
105.
106.
107.
108.
109-111.
112.
113.
114.
115.
116.
117.
118.
119.
120.
121.
122.
123.
124.
125.
126.
127.
128.
129.
130.
131-140.
141.
142.
143.
144.
145.
146.
147.
148.
149.
150.
151.
152-161.
162.
163.
164.
165.
166.
167.
168.
169.
170.
171.
172,
173.
174.

P.
R.
P,
D.
J.

.M.

H.
D.
P.
R.
A.
T.
W.
H.
W.
P.
R.
M.
M.
C.
T.
H.
J.
D.
S.
dJd.
R.
A.

w3

EPHAPPELEOSO

N.
E.
G.
N.
R.
R.
W.
K.
P.
W.

40

Haubenreich
Helms
Herndon
Hess
Hightower
Hill .
Hoffman
Holmes
Holz

Horton

Houtzeel

L.
R.

Hudson
Hantley

Inouye

H.
R.
J.
T.
J.
R.
W.
T.

J.

V.
S.
W.
B.
I.
S.

- W.

E.
A.
E.
J.
J.

o

HOHONEHN YR

2Py

Jordan
Kasten
Kedl
Kelley
Kelley
Kennedy
Kerlin-
Kerr
Keyes
Kiplinger
Kirslis
Koger
Korsmeyer
Krakoviak
Kress
Krewson
Lamb

Lane
Larson
Lawrence
Lawrence
Lin
Lindauer
Litman
Llewellyn

. Long, dJr.

Lotts
Iundin
Lyon
Macklin
MacPherson
MscPherson

Mailen

Manning |
Martin-

. Martin

Mateer

175.
176.
177.
178.
179.
180.
181.
182.
183.
184.
185.
186.
187.
188.
189.
190.
191.
192.
193.
194.
195.
196.
197.
198.
199,
200.
201.
202,
203.
204.
205.
206.
207.
208.
209.
210.
211.
212.
213.
214.
215.
216.
217.
218.
219.
220.
221.
222,
223.
22/,
225.

QEEEmEPa

T.
H.

PRUrQaYEp

HPEHEE RO P

J.
H.
C.
J.
W.
G.
A.
F.
G.
0.
P.
I.
R.

Dunlap

H. Mauney
McClain
W. McClung

. E. McCoy

. L. McElroy
. K.
J

MeGlothlan
McHargue
E. McNeese

R. McWherter
Metz

Meyer

. P. Milford

Moore
M. Moulton

W. Mieller

A. Nelms

H. Nichol

P. Nichols
L. Nicholson
D. Nogueirs

C. Oakes

Patriarca

M. Perry

W. Pickel

B. Piper

E. Prince

L. Ragan

I.. Redford
Richardson

Robbins

C. Robertson

Robinson

Romberger

Ross

Ssvage
Schaffer

Schilling
Scott

L. Scott

E. Seagren

E. Sessions

H. Shaffer

H. Sides

M. Slaughter

N. Smith

J. Smith

P. Smith

L. Smith

G. Smith
Spiewak

C. Steffy

P

AQEEQ

=
 

oy

226.
227.
228.
. 229,
230.
231.
232.
233.
234.
235.
236.
237.
238.
239.
240.

256.
257.
258,
259,
260.
261.
262.
263.
264
. 265.
266-267.

268.

269.
270.
271.
272.
273.

_' 2740
275,
276.

277.
_278.
- 279.
280.

281.

282.

283,

284—298

G.
dJd.
D.
C.
H.
A.
F.
c.

W.

W.
T.
A.
D.
C.
G.
D.

H.‘

M.
J.
W.
S.
E.
D.
R.
A.
R.

- C.

H.
A.
A.
R.
H.

 

41

Stoddart - - 241. C. F. Weaver
Stone 242, B. H. Webster
Strehlow 243, A. M. Weinberg
Sundberg 244, J. R. Weir
Talleackson 245, W. J. Werner
Taylor , 246. K. W. West

Terry . 247. M. E. Whatley

E.

- F .

M.

. B.

E.
M.

® S.

L.

G.
G.
F.
B.
M.

Thoma, . . 248, J. C. White
Thomason . 249. R. P. Wichner

Toth ~ 250. L. V. Wilson

Trauger ' 251. Gale Young
Unger ' -~ 252. H. C. Young
Watson : ' : 253. J. P. Young
Watson 254. E. L. Youngblood
Watts - 255. F. C. Zapp

- EXTERNAL, DISTRIBUTION

Allaria, Atomics International

Asquith, Atomics Internationsal

Cope, RDT, SSR, -AEC, Oak Ridge National Laboratory
Deering, AEC, OSR Oak Ridge National Laboratory

Dieckamp, Atomi¢s International -

Giembusso, AEC, Washington

D.
E.
L.
J.

W.

B.
G.
L.
W.
E.
M.

Haines, AEC, Washington

Johnson, AEC, Washington

Kitterman, AEC, Washington

Larkin, AEC, Oak Ridge Operations

McIntosh, AEC, Washington

Martin, Atomics International

Mason, Atomics International

Matthews, RDT, OSR, AEC, Oak Ridge National Laboratory
Meyers, Atomics International

Reardon, AEC, Canoge Park Area Office

Roth, AEC,- Oak Ridge Operations

Shaw, AEC, Washington

M.
L.
R.
E.
K.
F.

Simmons, . AEC, Washington

‘Smalley, AEC, Washington

Stemp, AEC, Canoge Park Area Office
Stansburg, the University of Temmnessee

Stevens, AEC, Washington

Sweek, AEC, Washington

Taboada, AEC, Washington -

F..

Wilson, Atomics International

- Leboratory and Unlversity'Div1sion, AEC, Oak Rldge Operations

Division of Tedhnlcal Infbrmatlon Extension _