ORNL-TM-3488.txt

 

»)

»)

»n

#)

.8

¥)

 

ORN L;TM-3488

Contract No. W-7405-eng~26

METALS AND CERAMICS DIVISION

MASS TRANSFER BETWEEN H.ASEIELIOY N AND HAYNES ALIOY No. 25
IN A MOLTEN SODIUM FLUOROBORATE MIXTURE

J. W. Koger and A, P, Litman

 

i | This report- was prepared as an - account of work | °
~. | sponsored by the United States Government, Neither | |
-the United States nor the United States Atomic Energy |
-} Commission, aor any of their employees, nor any of |
. - their contractors, subcontractors, or their employees,
»| makes any warranty, express or implied, or assumes any.
. |- legal- liability - or responsibility for the accuracy, com-
pleteness or usefulness of any information, apparatus, | .
product or process disclosed, or represents that its use |
would not infnnge private!y owned rights,

g et b s
Lo

s mT— e

 

 

 

e T e ouSwem T Beg adigl o o

. OCTOBER 1971 T T e

~OAK RIDGE NATTONAL LABORATORY
- Oak Ridge, Tennessee
. operated by
UNION CARBIDE CORPORATION
for the
U.S. ATOMIC ENERGY COMMISSION

 PISTRIBUTION OF THIS DOCUMENT 1S UKL |
 

 
 

————E

 

 
 
&

")

 

i

 

»

b

Fi)

AbStI‘a.C‘t e’ o & & e @ o ¢ e @

Introduction «. « « ¢ ¢ ¢ « &

' Experimental Details o . . .

Loop Febrication . . .
Salt Preparation o e o
Loop Operations_. o e e

_Test.Results e e e e a e

CONTENTS

Haynes Alloy No. 25'Samples

Hastelloy N Specimens . . .

-~ Salt Analysis'. . . .
Summary of Test Results . .

Discussion . + « v o ¢ & o &

Prior Studies « ¢ « o &

Corrosion Mechanisms ., e .

Significance of Haynes Alloy No. 25

*

in the Hastelloy N

TeS‘tSyStem..'.....'.............‘..

o L J e - * . *

Conclusions .

Acknowledgments

-

® o o ¢ o s 8 8 & e e e o & o ° & w s s =

+d
5

 

N0t NN DN

bR G R&EE

23
25
25
 

 

 

MASS TRANSFER EETWEEN HASTELIOY N AND HAYNES ALIOY No. 25
IN A MOLTEN SODIUM FLUOROBORATE MIXTURE

 

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

The compatibility of Haymes alloy No. 25 and Hastelloy N
with fused NaBF,-8 mole % NaF was determined in the range
605 to 460°C, The cobalt~base alloy was inadvertently
- incorporated in the Hastelloy N thermal convection loop and
was exposed to the fluoroborate salt mixture for 3660 hr.
' The Haynes alloy No. 25 suffered damage by selective
~ leaching of cobalt and chromium, which migrated to the
- Hastelloy N, The mechanism of corrosive attack was activity-
gradient and temperature-gradlent mass transfer. Haynes
alloy No. 25 is more suscepmlble to attack by the fluoroborate
~mixture than Hastelloy N. The presence of the small amount
of Haynes alloy No. 25 in the system did not compromise later
experiments on the monometalllc Hastelloy N system. Penetra-.
tion of deEOS1ted cobalt corresponded to a diffusivity of
5,6 % 1071° em /sec in Hastelloy'N at 465°C

 

s INTRODUCTION

Two thermal convectlon loops, NCL-13 and -14 began operation in
October 1967 to determlne the compatlbillty of standard and titanium-

"modlfled Hastelloy N alloys W1th NaBF4—8 mole % NaF salt, a candidate

secondary coolant for molten-salt reactors. The loops, which are

_plctured in Fig. 1, operated w1th maxjimum temperatures of 605 C and
”1nduced temperature dlfferences of 145°C | '

~Both' the heated and cooled sectlons of the loops contalned removable

:Hastelloy N SpeC1mens._ These speclmens were w1thdrawn perlodlcally

along w1th salt samples to follow corrOS1on processes as a functlon of

’5time.' After ‘some 4000 hr of operatlon, the Hastelloy N spe01mens in

the hottest and coldest reglons of the loops were removed and sub jected

“to detalled metallurglcal analy81s.- Portlons of the speC1mens were

 

1Now with the USAEC, Washington, D. C.

_ 2H E, MbCoy, Jr., and J R. Welr, Jr., Materlals Development of
Molten-Salt Breeder Reactors, ORNL-TM-1854 (June 1967).

 

 
 

 

 

 

 

 

 

gy 7=Photo 75125A

L HOT LEG

Fig. 1. Hastelloy N Natural Circulation Loops NCL-13 and =14,

Containing NaBF;—8 mole % NaF at a Maximum Temperature of 605°C
with a Temperature Difference of 145°C,
»h

P

ok

"'Hg 31, 1969, ORNL-4449, pp.- 200208

sent for microprobe analysis to determine possible composition gradients

~due to mass transfer. In1t1al results showed ‘that a large amount of

cobglt had depos1ted on both the hot and cold 1eg specimens, The source

-of the cobalt was traced to the 1/8-in.-d1am rods that held the removable

speC1mens. These rods were determlned to be Haynes alloy No. 25 rather
than the specified Hastelloy N Further investigation revealed that

the source of the Haynes alloy No. 25 was a misidentified storage carton.
The specimen hanger rods were replaced with HastelIOy'N'and the experi-
ments were continued.

We have taken advantage of the situation to obtain 1nformation on
the corrosion of cdbalte and n;ckelébase alloys simultaneously exposed
to & molten fluoroborate salt. Further details on the compatibility of
Hastelloy and otheriallbYvaith fluroborate salts have been reported.> -t

 

33, M. Koger and A, Pv Litman, Compatibility of Hastelloy N and
Croloy 9M with NaPBF,-NaF-KBF, (90-4-6 mole %) Fluoroborate Salt,
ORNL-TM=2490 (April 1969). |

4J. W. Koger and A. P. thman, Catastrophlc Corrosion of Type 304
Stainless Steel in a System Circulating Fused Sodium Fluoroborate,
ORNL-TM~2741 (January 19707' :

J. W Koger and A, P. thman, Compatiblllty of Fused Sodium
Fluoroborates and BF3 Gas w1th Hastelloy N Alloys, ORNL-TM-2978
(June 1970). |

6J. W. Koger and A. P Idtman, MSR Program Semlann. Progr. Rept.
Feb. 29, 1968, ORNL-4354, pp.,221—25

7J. W. Koger and A. P. thman, MSR Program Semiann. Progr. Rept.

Aug. 31, 1968, ORNL-4344, pp. 264—66 and 285-89,

83 W, Koger and A. P Ldtman, MSR Program Semlann. Progr. Rept.
Feb, 28, 1969 0RNL-4369, pp. 24653, | L -

.—_—;_-,

 %J. W. Koger and A. P: Litman, MSR Program Semlann. Progr. Rept.

105, w. Koger, ‘MSR Program Semiann, Progr. Rept. FEb 28 1970

”7dORNL-4548, pp. 242-52 and 265-72.

113, W, Koger, MSR Prqgram Semlann. Progr. Rept Aug. 31, 1970
ORNL-4622, pp. 16878, |
 

4

EXPERIMENTAL DETAILS

The test devices used in the experiments were thermal convection
loops in a harp configuration, with surge tanks atop each leg for sample
and speéimen access, The flow was generated by the difference in
density of the salt in the hot and cold legs of the loop, and the salt

flow velocity was approximately 7 £t /min,

Loop Fabrication

The loops were fabricated from 0.606-in.-ID Hastelloy N tubing with
a 0,072-in. wall thickness. The annealed material, heat 5097, was
TIG welded to Specifications PS-23:and PS-25 and inspected to MET-WR-200
‘specification. The finished loop was stress relieved at 880°C for 8 hr

in hydrogen.
Salt Preparation

The fluoroborate salt mixture used in the test program.was furnished
by the Fluoride Processing Group of the Reactor Chemistry Division, and
its compositioh béfore test is given in Teble 1. To mix and purify
the salt, the raw materials were first heated in a nickel;lined'vessel
to 150°C under vacuum and held for 15 hr. Then the salt was heated to
- 500°C, agitated with helium for a few hours, and transferred to the fill
vessel. At 600°C, the BF3; pressure is approximately 200 torr.

Table 1. Salt Analysis Before Test

 

 

Element C°?;§nt Element C‘(’g;;?t |
Na . 21.9 Cr 19
B 9.57 Ni 28
F 68,2 Fe 223
0 459
Mo < 10
Co < 10

 

O,
 

 

.4

s

"loop.Operations

The loops were heated by'pairs of clamshell heaters placed end to

end W1th the input power controlled by silicon controlled rectifier
'units and the temperature controlled by a current proportioning controller.
"The loop temperatures were measured by Chromel vs Alumel thermocouples
~ that had been spot welded to the outside of the tubing, covered by a
" layer of quartz tape, and then'covered with stainless steel shim stock.
Tubular electric heaters cbntrolled by variable autotransformers

furnished the heat to the cold leg portions of the loops.
Before filling with salt, the loops were degreased with ethyl

alcohol, dried, and then heated to 150°C under vacuum to remove any traces

- of moisture. A helium massiSpectrometer leak detector was used to check

for leaks in the system. . |

The procedure forlfillinglthe loops consisted of heating the loop,
the salt pot, and all comnecting lines to approximately 550°C and
applying helium pressureto;the salt supply vessel to force the salt into

the loop. Air was continuonsly‘blown on freeze valves leading to the dump

~and flush tanks to prov1de & pos1t1ve salt seal.

All fill lines exposed to the fluoroborate salt were Hastelloy N.
All temporary connections from flll_line to loop were made with stainless
steel compression fittings., S .

‘The first charge of salt was held for 24 hr in the loops at the

a,max1mum operation temperature and then dumped. This flush salt charge

- was intended to remove ‘surface. oxides or other impurities left in the
.~ loops. The loops were then refilled with fresh salt, and operat1on began.
i_,fOnce the 1oop was filled the heaters on the cold legs of the loops
_-Eiwere turned off. As much 1nsulation was removed as necessary to obtain
i;_the prOper temperature difference by'exp031ng the cold leg to ambient
. air. Helium‘cover gas of . 99 998% purity and under slight pressure

| (approx 5 ps1g) was maintained over the salt in the loops durlng operation.

Each loop contained lA Hastelloy'N specimens 0. 75 x 0. 38 X 0.030 in.,

'”f'each with a surface area of 0.55 in.? (3.5 em ) Seven specimens were

attached at different vertical posztions on l/B-in. rods (later found

to be Haynes alloy No. 25). This array could be placed into or removed
 

 

 

- from the loops during operation by means of a double ball valve
arrangement, One rod was inserted in the hot leg and another in the
cold leg of each loop. The surface area of the rod exposed to the salt
was one-ninth that of the loop. The compbsition of the Hastelloy N

loop tubing is compared with the nominal composition of Haynes alloy No. 25

in Table 2.
Table 2, Alloy Compositions

 

Content, wt %
Ni Mo. Cr Fe Co W Si  Mn

Alloy

 

Hastelloy N 70.8 16.5 6.9 4.5 0.1 0.1 0.4 0.5

Haynes Alloy 9.0 0.5 19,0 1.0 53,0 14.0 0.3 0.5
No. 25

 

The loops were operated at a maximum température of 605°C and a
temperature difference of 145°C, with the Hastelloy N specimens and
" Haynes alloy No. 25 rod exposed to the salt for 3660 hr.

TEST RESULTS

Preliminary results of analyses from rods and specimens of both
loops (NCL-13 and -14) were identical., Thus, we completed detailed
analyses only on the materiasls from NCL-13, ;

Haynes Alloy No. 25 Samples

After 3660 hr of salt exposure and discovery of the material mixup,
samples of the 1/8-in. Haynes alloy'No. 25 specimen holder rods were
taken from various positions and analyzed in detail. Figure 2 shows
the 1ocat10ns of the Hastelloy N sPeC1mens, the Haynes alloy'No. 25

rod, and the portions removed for analysis. |
| Figure 3 shows the as-polished and the etched microstructures of
the Haynes alloy No. 25 rod (sample 3) located at the top of the hot leg
(598°C) - Three characteristics are apparent from examination of all the

mlcrostructure° (l)..about 0.2 mil of thickness of the material was

O
 

" :; , = ' o ' T o " ORNL-DWG €8-3987R2

SAMPLE
NO.

  

SAMPLE
NO,

   

&
y
N
b3
N

  

L PIIPEI S OIIIE

4

   

CLAMSHELL -
" HEATERS

  

i)

3Qin.

 
 
  

INSULATION

  

CORROSION
SPECIMENS

   
   
 

13

 

 

 

 

 

 

 

SAMPLER 44

* BLACK AREAS ARE PORTIONS OF THE
- HAYNES ALLOY NO.25 RODS REMOVED -
FOR ANALYSIS ,

 

 

 

. FLUSH -
2 ' o - o e L TANK - OUMP

 

 

 

 

 

 

: o ‘Fig. 2. Thermal Convection Loop and Salt Sampler, Including
‘ :}- - Location of Metal Specimens and the Temperature Profile. -
 

 

g Y-89516

 

@)

Y-89517

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Fig. 3. Microstructure of Haynes Alloy No. 25 Exposed to
NaBF;-8 mole % NaF at 598°C in NCL-13 for 3660 hr. 500x.
() As-polished. (b) Etched with hydrochloric acid and hydrogen
peroxide. ‘ , g -

O
 

¥

+¥

&)

lost, (2) corrosion productS”had deposited and (3) there was some

attack along the graln boundarles (seen in the asepollshed sample).

~The largest pit (not shown) was about 2 mlls deep.f

Figure 4 shows the metallographlc appearance of Sample 1, which

 was exposed to BF3 gas at 604°C in the. hot leg- surge tank., The upper

portlon of the flgure shows the area of maxlmum attack where about 8 mils
of metal was removed and other materlal'was dep081ted The lower portion
of the figure is 1nd1cative of most of the sample w1th about 1 mil of
attack and some deposited materlal

Samples 1 and 8 (see Flg. 2), exposed only'to BF3 gas in the upper-
portlons of the surge tanks,_were noticeably darker than the other
samples, The d;fference,1n;surface.character of the materials exposed
to the gas and the'liquidfsalt]is‘seen_in Fig;'S.: The Haynes alloy No. 25
rod was analyzed.by'x-ray fluorescence to determine relative concentra-
tions of Co, Cr, W, Ni,-and'Fe. The fluorescence results were compared
against as-received Haynes alloy No. 25 which was aSS1gned the composi-
tion given in Table 2. The results, which represent a surface zone

3 to 5 mils deep, are glven in Table 3 along with the temperature of

“the salt at each pOS1t10n of the rod Note that the concentratlon of

tungsten in all samples 1s unchanged from the before-test level of 14%.
Thus, the tungsten concentration was used as a standard-an the analysis.!?
Samples 1 and 8 showed a significant loss of chromium,;from about 19 to
ebout 4 weight units, and cobait from 53 to 20 and 45 weight units,

respectively,  Sample 1, at the hlghest temperature, 604° C, experienced

~ the greatest loss of mater1a1 - Samples 2, 9, and 10, which were
s TexPosed to relatively stagnant salt in the surge tanks above the 1loop,
. all lost chromium and cobalt.:,r—

Samples 3 through 7 and A1 through 14 were exposed to c1rculat1ng

salt at various temperatures. Sample 3, at the. hottest position, 598°C

lost nlckel, cobalt, and chromlum. Samples, 4, 5, 6 and 7 of the hot

 

. 12A standard WElght un1t of 100 was used for the uneXposed sample.

 The exposed and unexposed samples all contained 14 units of tungsten, .

so this allowed calculation of the amounts of the other elements. We
were then able to determine if a sample showed a net gain or loss of a
certain element.
 

 

 

 

 

 

 

 

 

 

 

 

 

Y-90189

 

 

 

Fig. 4. Microstructure of Haynes Alloy No., 25 Exposed to BF3 Gas
at 604°C in NCL-13 for 3660 hr. 500x. (a) .Area of greatest attack.
(b) Remainder of sample. Etchant is hydrochloric acid and hydrogen
peroxide. ' : '

O
o

 

 

o . L Fig. 5. Haynes Alloy No., 25 Rod and Hastelloy N Specimens Exposed
to NaBF;—€ mole % NaF and BF3 Gas at 500 to 605°C in NCL-13 for 24 hr.

 

leg lost cobalt and"Chromium'andtgained nickel. Sample 11 in the cold
leg is much like Sample 7 from the hot leg; both were exposed at about
the same temperature. However, Sample 12 was exposed to a slightly

P

lower temperature, 487°C and the analyzed ares was almost all nickel
and tungsten. Samples 13 and 14 which were exposed at nearly identical
temperatures, showed an increase in iron and nickel and decrease in cobalt

and chromium concentrations compared to before test.

_:Hastelloy N Specimens

 

 

7 Figure 6 shows etched and as-polished microstructures of the |
flHastelloy N specimen from the hottest posxtion (604°C) - The edge exposed
during test was more heaVily'attacked by‘the metallographic etchant than

 

"was the underlying base metal, ThlS faster etching response is apparently
due to chromium depletion of the alloy, Figure 7 1s the as-polished |
meicrostructure of the’ Hastelloy'N specimen in the coldest p051tion -
.(465°C) A uniform deposrt of apprOX1mately 0.25 mil is apparent on
“the surface of the sPeczu.n:u.en.,t
Electron microprobe studles revealed that the thin layers on the

»

‘7 } - edge of the hottest and coldest Hastelloy N specimens contained:

 
 

‘Table 3. Temperature, Position, and Composition of Haynes Alloy No."25 Sampleé

 

 

 

a b | Absolute Amount of Each Element

ASample  Position Tem?fg?ture in the Fluoresced Area®

W Ni Fe Cr Co

Before Unexposed Haynes Alloy No, 25 | 14 9 1. 19 53
1 Hot leg surge tank vapor 604 14 11 2 3 20

phase (BF3; and He) | | ‘

2. Hot leg surge tank 604 4 11 1l 1 7
3 Hot leg 598 - 14 6 1 14 30
4 Hot leg - 579 14 10 1 10 25
5 Hot leg | 560 Y% 12 1 4 25
6 Hot leg 53 4 13 1 9 29

7 Hot leg 524, 1% 20 3 %7 30
8 Cold leg surge tank vapor 538 X9 3 4 45
phase (BF3; and He) o |
2 Cold leg surge tank 538 | 14 12 2 5. 23
10 ‘Tubing between surge tank 538 . 14 13 1 5 20
and cold leg | ‘ _ o
11 Cold leg : ' 516 14 20 1 5 24
12 Cold leg _— 487 % 21 1 1 4
13 Cold leg 476 Y% 26 4 5 41
14 4 8 41

Cold leg 485 1% 23

 

gSampleS‘z, 9, and 10 in nonflowing'salt.'
Pa11 samples exposed to molten salt, unless noted.

®Based on 100 weight unlts for unexposed sample and referred agalnst as-recelved
Haynes alloy No. 25 as standard.
 

 

 

 

 

 

»

*

 

 

 

(2) As polished. (b) Etched with glycerla regia.,

   

   

 

Fig. 6. Microstructure -of Standerd Hastelloy N Exposed to
‘NaBF;—8 mole % NaF at 604°C in NCL-13 for 3660 hr, 1000x. -
 

 

14

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Fig. 7. Microstructure of Standard Hastelloy N Exposed to
NaBF;—8 mole % NaF at 465°C in NCL-13 for 3660 hr. As-polished.
500x. ' . . : ,

appreciable cobalt. There was an average of 1.8 wt % Co in a band
6 pm thick on the hot leg specimen and 7.3 wt % Co in a band of the
same thickpess on the cold leg specimen. These results were substantia-
ted by qfialitative x-ray fluorescence mgasurements,‘which showed more
cobalt than iron (approx 5%) in a band near the surface. Laser spectro-
graphic analysis showed substantial cobalt at depths less than 20 um
into the material. The amount of cobalt in the Hastelloy N specimen
located at the bottom of the_cold leg (approx 465°C) was determined as
a function of position by microprobe and is given in Fig. 8. A cobalt
composition gradieht in hot leg specimens, obtained by the microprobe,
was not well defined and will be discussed later,

Using the penetration curve of Fig. 8, we determined the diffusion
coefficient of cobalt in-Hastelloy'N. A constant surface concentration

of cobalt was assumed to integrate Fick's second law,

O
 

L1

n

Co CONCENTRATION (wt %)

30

25

20

15

{0

oL
0

ORNL-DWG 6814335

 

 

" ORIGINAL SURFACE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

10 15 20 2

PENETRATION DISTANCE (microns)

 Fig. 8. Cobalt Gradient Produced in Standard Hastelloy N at 465°C.
 

16

ac_Dazc,
TR

which relates concentration to time and distance. The appropriate

solution in this case is

C —=Cy = (CS V— CO) [l -er‘f(x/éx/D—'E)] ’

where C = cobalt concentration at a distance x centimeters below the
surface after diffusion has occurred for t sec,
CS = surface concentration,
Co = initial cobalt concentration in the Hastelloy N, and
D = diffusion coefficient, cm?/sec.

The diffusivity of cobalt in Hastelloy N was calculated torbe
5.6 X 107%° cm?/sec at 465°C,

Salt Analysis

Affer test, less than 50 ppm Co was found in the salt, The
significance of this is discussed in the next section. Table 4 shows
the composition of the salt after circulation for 42OQ hr., Comparison
with the salt analysis before test (Table 1) shows increases in the
chromium concéntration ffom 19 to‘232 ppm and iron from 223 to;314 ppm.

Table 4. Salt Analysis After Test

 

Content - Content

 

Element (%) Element = (ppm)
Na, 21.0 Co < 50
B 9.29 Cr 232

-F : 68.6 "Fe 314
| Mo . <20

NN <25

0 497

Hy0 800

 

O,
 

o

17

SUMMARY OF TEST RESULTS

The chemical and metallurgical analyses show that chromium and
cobalt were leached from the Haynes alloy No. 25 at all test temperatures.
Although the percentage of chromium lost was greater than that of cobalt,

the total mass of chromium lost was less. The cobalt and chromium

migrated to_the_Hastelloy N specimens and loop piping. There was also

some -evidence of highly locelized nickel transfer from the hot section
to the cold section of the Hayhes alloy No. 25 rod. Iron deposited on
the Haynes alloy No. 25 in the cold section and was probably supplied |
by the leaching of iron from the Hastelloy N 1oop'piping by the salt.
Attack of Haynes alloy No. 25 by BF3 in the vapor phase as evidenced
by'loss of alloy constltuents was less severe than the salt corrosion,
However, more dlscoloratlon and surface roughening were noted on the
samples exposed to the gas. The Haynes alloy No., 25 suffered much more
damage than the Hastelloy'N in the vapor phase, Haynes alloy No. 25
appears to be more suSceptible_to attack by the fluoroborate mixture
than Hastelloy N. -

~  DISCUSSION

Prior Studies

Past work!® at ORNL measured the chemical corrosion of various
materials under the condltlons exPerlenced durlng the fluorinatlon of

r'molten-salt fuels in the Flnorlde Vblatllity Process. The salt used

was equimolar NaF—ZrF4 contalning O to 5 mole % UF,. Several'cobalt-

:_contalnlng alloys were tested at 600 + 100°C, and ‘the behavior of those
 with less than 20 wt % Co wa.s 31mllar to that of Hastelloy'N However,

as the cobalt content exceeded 20 wt %, the alloys showed a much greater

1tdegree of attack than Hastelloy'N

 

13A P. Latman and A ‘E Goldman, Corrosion Assoc1ated With:
Fluorlnatlon in the Oak Ridge National Laboratory'Fluorlde Volatlllty

Process, ORNL-2832 (June 5, 1961)
18

”~ .

In a recent testl* at ORNL, samples of various materials including Kfiij
Hastelloy N, Haynes alloy No. 25, and graphite were placed in the vapor
and 1izuid zones of a vacuum distillation experiment that used a
LiF~-BeF,-ZrF, salt. The temperature ranged from 500 to 1000°C over a .
period of 4300 hr in the molten salt and 900 to 1025°C for more than 300 hr
in the vépor. The Haynes'alioy'No. 25 was the most heavily corroded of
the metals tested and was also brittle at the end of the test. Fracture
of one Haynes alloy No. 25 specimen caused a loss‘of some of the other
specimens during the experiment. Figure 9 shows the Hayhes alloy No. 25

specimens before and after test.

 

143 R. Hightower, Jr., and L. E. McNeese, Low-Pressure Distillation
of Molten Fluoride Mixtures: Nonradiocactive Tests for the MSRE
Distillation Experiment, ORNL-4434, pp. 30-33 (January 1971).

 

 

 

Photo 93712

VAPOR SALT *

 

AFTER TEST '

Fig. 9. Haynes Alloy No. 25 Specimens Before and After Test in
the Vapor and Liquid Phases of a LiF-BeF,; Salt. In vapor 300 hr at
900 to 1025°C and in liquid 4300 hr at 500 and 1000°C,.

O,
 

o

19

Thus, in other studies where both Hastelloy N and high cobalt

alloys were exposed to molten fluorides under highly oxidizing conditions,

Hastelloy N was much morejcorrosion resistant,

Corrosion Mechanisms

In polythermal flowingwsaltrsystems, corrosion commonly involves
temperature-gradient mass transfer, Figure 10 shows a schematic of

this process. For monometailic systems, the constituents of the salt

or impurities may react w1th one or more constituents of the loop

mater1al to form salt-soluble compounds. For example, the following
reactions may occur in a salt containlng UF, and FeF, exposed to
Hastelloy N; e
UF, + Cr - CrF, + UF; , | (1)
‘FeFy + Cr - CrFp + Fe . . (2)

The equilibrium constant:offiCorrosion reaction {1) is temperature
dependent. Thus, when the salt is forced to circulate through a tempera-
ture'gradient products from the reverse reaction may deposit in the

cooler regions of the system. Since the equlllbrlum constant for the

" chemical reaction 1ncreases-w1th increasing temperature, the chemical

activity or concentration of the attacked element in the container mate-
rial will decrease at hlgh temperatures and 1ncrease at low temperatures;

that is, in the hotter regions the alloy surface becomes depleted and

metal from the 1nterlor of the wall dlffuses toward the surface, and in
- the colder regions. the alloy surface becomes enriched with the diffusing
' -fmetal There is, of course, an’ 1ntermediate temperature at which the’

- initial surface composztlon of the structural metal and the attacked ele-
 ment is in equllibrium w1th the salt If the temperature dependence of the

o mass transfer reaction 1s small the rate of metal removal from the salt
 

20

ORNL-DWG 67-6800R

 

 

 

 

 

 

HOT SECTION

DIFFUSION TO SURFACE

 

 

SOLUTE ESCAPE THROUGH
-~ NEAR-SURFACE LIQUID LAYER

 

 

d—

 

— DIFFUSION INTO BULK LIQUID

D

TRANSPORT
TO COLD PORTION OF SYSTEM

 

 

\X\\\\\\\\\\\\\\\\\\\\\\\‘l\\\

 

COLD SECTION — —
— — {x SUPERSATURATION — ———

 

      
  

 

e NUCLEATION
o GROWTH TO STABLE CRYSTAL SIZE

.- —0R

_«SUPERSATURATION AND DIFFUSION

THROUGH LIQUID

NUCLEATION AND GROWTH

ON METALLIC WALL — —— - —
OR DIFFUSION INTO WALL

 

 

 

— — e r—————————————— -— —c — e -—_

Fig. 10. Temperature-Gradient Mass Transfer.

stream by deposition in the cold region will be controlled by the rate
at which the metal diffuses into the cold region wall. Many examples -
of temperature-gradient mass transfer by fluoroborate salt systems
contained in Hastelloy N were cited in the introduction.

In the system under study — Haynes alloy No. 25-Hlastelloy N-
fluoroborate salt — an additional mass<transfer mechanism is present
which is termed dissimilar-alloy or activity gradient mass transfer. While
this has been studied in detail for several alkali metal systems, little

O

O,
 

 

#”»

w

21

work has been done in molten salts. Mass transfer of this type requires -

. the presence of different alloys'in the same fluid and is not contingent

on the presence of a temperature gradlent Figure 11 shows a schematic
of this process. The sequence of events in dlSSlmllar—alloy'mass
transfer involves removal of an element from one material and deposition
on a second material; that is, movement from a region of high activity
_to‘one of low activity. Examples of dissimilar metal interactions are
common in niobium and type.Blfinstainless steel systems exposed
simultaneously to SOdiumeetassium alloy’® and in the other alkali

16 As might be suspected when the opportunity for both

metal systems.
forms of mass transfer arlses, as in the case at hand corros1on inter-
actions become very compllcated

In the present studies, chromlum‘and cobalt were removed from the

'Haynes alloy No. 25 by'reactlon with salt and were deposited on the
Hastelloy N.  This occurred because of the smaller concentration of those

elements (lower activity) in the Hastelloy N, This activity gradient

mechanism would have occurred without artemperature gradient. However,

the effect of the temperature gradient was also evidenced by heavier
deposition of chromium and cobalt on the colder Hastelloy'N surfaces.
Asrmentioned earlier, ne large cobalt concentration gradient could
be obtained from the Hastelloy N hot leg specimens, and more cobalt was
found on the cold leg spe01mens than in the hot leg. Thus, it appears

‘that the net result of the cobalt depositing on the Hastelloy N due to
_the activity'gradient fiechanisfi.and the cobalt mass transfer due to the
, 'temperature gradlent mechanlsm.was a contlnued depletlon of cobalt from
"'the Hastelloy'N in the hot section and a deposition in the cold section.
~ This depletlon could occur 1f most of the cobalt that transferred from
 the Haynes alloy'No. 25 d1d 80 inltlally.r Chromlum'behaved in the
';same manner in the hot leg. In an all Hastelloy N system, a chromium
' ??gradlent is usually expected,ln the ‘hot leg because of the chromium

 

155, R, DiStefano, Mass Transfer Effects on Some'Refraétery'Metal-

. Alkali Metal Stainless Steel Systems, ORNL-4028 (November 1966).-

163, H. DeVan, Compatiblllty of Structural Materials with 30111ng
Potassium, ORNL-TM=-1361 (April 1966).
 

 

ORNL-DWG 65-13233R2

KR I\ AKX
0.:.0.0 >0

             
   

     
 

   
 
  

   
     

| LIQuiD == o
ALLOY A —_— e = - b ~ ALLOY B
CONTAINING T = K CONTAINING n
COMPONENTS . MOVEMENT BY 3 COMPONENTS
2

    
   
 

”
%

L/

    
  

\/

     

DIFFUSION OR CONVECTION Nt Ny 4o +N; 4 Ny

05

     
     
  

  

M+ Myt +M; 485 7]

. —_——= - _———* [
— [

e - s

] —_— = e a = . K

-2 L R

REACTION OF A% i

WITH LIQUID I X

  

- — -

DIFFUSION INTO
ALLOY B

  

DRIVING FORCE FOR THE TRANSFER OF ANY COMPONENT 4%
IS THE DIFFERENCE IN CHEMICAL POTENTIAL OF 4% IN ALLOY A

COMPARED WITH ALLOY B.

Fig. 1l. Dissimilar-Alloy Mass Transfer.

removal. However, this gradient may not be as large as expected and may
disappear if the chromium diffuses to the surface as fast as it is removed
and possible even at a slightly higher rate.l? However , the chromium
gradient was almost completely masked in this system for the above reasons
and by the chromium deposition from the Haynes alloy No. 25 back to the
hot leg. As mentioned ea.r]ier, the microstrucfiure (Fig. 8) does show
evidence of some depletion at the edge of the specimen, but the overall
chromium composition of the specimen showed little change. Nickel and
iron, if removed from the Hast'elloy.N, would be predicted to depoéit on
the Haynes alloy No. 25 by virtue of the activity gradient mechanism,

. This was not observed experimentally, although the effects may have been

swamped by the greater rate of ‘chromium and cobalt transfer.

 

17G. M. Adamson, R. S. Crouse, and W. D. Manly, Interim Report on
Corrosion by Zirconium-Base Fluorides, ORNL-2338 (Jan. 3, 1961).

O
 

"

23

Significance'of Haynes Alloy No. 25 in the
Hastelloy N Test System .

The solid line at Fig. 12 shows the time dependence of experimental
weight changes of Hastelloy N specimens exposed to salt in loop NCL-13.
Note that all specimens showed a net weight gain during the first 200 hr,
As mentioned earlier, in pemperature;gradient mass~=transfer systems
specimens in_hot‘portions ef the loop are expected to lose weight while
those in the cold secfiicnjshould gain weight. chever,.after this
initial period of weight gain the samples in the hot leg started losing
weight, while the cold leg specimens continued to gain weight. This
initial weight gain lends credence to the idea that most of the cobalt
that transferred from.the"Heyees alloy No. 25 did so initially. It is
noted that our experlmental.welght changes reflected this deposition
from the Haynes alloy No. 25. The actual weight loss in an all
Hastelloy N syetem.would be larger and the actual weight gain would be
smaller, Thus, a constant factor was calculated and subtracted from
all our welght changes, resultlng in the dotted line. This constant
factor was calcelated u31ng an iterative trial-and-error method to obtain

a mass balance on the syetemefieting the following equation:

AW = AW 4 AC

system loss system gain salt ,
- where - ' H
Aweyetem.loss-;_weighf lcss‘for'SPecimens and 1e0p components,
-Awsystem gain weigh‘b ’.fgam for specimens and components,
AC

selt = content change in salt

-’This exercise also allowed us  to conclude that our mass—transfer rate
- would not have been exce551ve 1f ‘the cobalt alloy had not been in the
1system., The reason that the welght changes due to Haynes alloy No. 25
.:'f were ‘80 small was that 1ts surface aresa exposed to the sale is one-ninth

that of the Hastelloy N. Recent,work‘has shown that these welght
 

24

ORNL-DWG 68-14334R

 

 

465°C

 

 

 

 

o
- —

 

 

 

 

 

 

 

a— "‘- —_‘
NE ”’.——
< =
o
E o
& \\ EXPERIMENTAL WEIGHT CHANGE ~
‘E \\ ~— —— ESTIMATED WEIGHT GHANGE
5 NN WITHOUT COBALT IN SYSTEM
= N
T - \ -
o , :
= ~
, ~ \\\;\
-2 ™~ o

‘\\\
\ .
\ .
_3 ~§\
\\\
604°C N~

 

 

 

 

 

 

 

 

 

 

0o 500 1000 1500 2000 2500 3000 - 3500
TIME (bhr)

Fig. 12. Weight Changes as a Function of Operating Time of
Hastelloy N Hot-Leg and Cold-leg Specimens Exposed to NaBF;—8 mole % NaF
at 604 and 465°C, Respectively, in NCL-13,

differences have not substantially affected any later reaction-rate
constants calculated for fluoride salt from corrosion studies in this

system. 18

 

187, W. Koger and A. P. Litman, MSR Program Semiann. Progr. Rept.
Feb. 29, 1968, ORNL-4354, pp. 221-25.

 
 

 

-

»t

"

25

CONCLUSIONS

1. Haynes alloy No..25 in the fluoroborate salt-Hastelloy N alloy
test system suffered damage by loss of significant amounts of cobalt
and chromium, which migrated to the Hastelloy N by virtue of activity-
gradient and tempereture-gradient mass transfer,

.2. Haynes alloy No. 25 is more susceptible than Hastelloy N to
attack by the fluoroboraste mixture.

3. Because of the relatively small emount of Haynes alloy No. 25
in the system (one-ninth the surface area of Hastelloy'N), the early
presence of this material did not compromise, beyond the normal 10%
variation in quantitative data, experiments on the preeent monometallic

Hastelloy N'system.

ACKNOWLEDGMENTS

It is & pleasure to acknowledge that E. J. Lawrence supervised
construction and operation of the test loops. We are also indebted to
H. E, McCoy, Jr., and J;VH.VDeVan for constructive review of the
manuscript. o

Special thanks are extended to the Mbtallography'Group, especlally

H R Gaddls, H. V. Mateer, T. J, Henson, and R. S. Crouse, and to the

Anglytical Chemlstry'D1v151on, espec1ally'Harrls Dunn and Cyrus Feldman,
Graphic Arts Department, and the MEtals and Ceramics Division Reports

' Offlce for 1nva1uab1e assxstance.
 

 

 
 
 
 
 

  
 
 
 

1-3,
45,

ok

6—15.
16,

17.
-18-19,
20,

21.

22.
23,
24,

25.

26,

27.

28.

29,

30.

. 31,
- ' 32.
‘ 33.
34,

35,

36,

37.

28.

39,

40,
4],
42,

43,

45,
46,
4.
. 48,

 95-96,

 or-ss.

99,

- 100-104.

 

27

ORNL-TM~3488
DITERNAL DISTRIBUTION

Central Research Library 49, W. O. Harms
ORNL — Y-12 Technical Library 50. P. N. Haubenreich
Document Reference- Section 21l. R. E. Helms
Laboratory Records Department 52. T..T. Henson
Laboratory Records, ORNL RC - 53-55., M. R. Hill

ORNL Patent Office : 56, W. R, Huntley
MSRP Director's Office (Y-12) 57. H. Inouye

G. M. Adamson, Jr, 58, J. J. Keyes

J. L. Anderson o . 59-68, J, W. Koger -

R. F. Apple 69, A, I. Krakoviak
C. F. Baes S 70. E. J. Lawrence
S. E. Beall ' 71. M. I. Lundin

E. S. Bettis | 72. R. E. MacPherson
F. F. Blankenship " 73. W. R. Martin
E. G. Bohlmann ' 74, H. V. Mateer

G. E., Boyd ' 75. H. E. McCoy, Jr.
R. B. Briggs 76. C. J. McHargue
S. Cantor 77. A. S. Meyer

0. B, Cavin 78. L. E. McNeese
Nancy C. Cole - ' 79. R. L. Moore

W. H. Cook | - 80. F. H. Neill

R. S. Crouse | | 81, E. L. Nicholson
J. L. Crowley B 82. P. Patriarca

F. L. Culler o | 83. A. M. Perry

J. H. DeVan - 84, Dunlap Scott

J. R. DiStefano : - 85, J. H. Shaffer
S. J. Ditto o 86, G. M, Slaughter
W. P. Eatherly : 87. R. E. Thoma

J. R. Engel o - 88, D. B. Trauger
D, E, Ferguson - | 89. G. M. Watson

L. M. Ferris o ' : 90. A. M. Weinberg
J. H Frye, Jr. . 91. J. R. Weir, Jr.
L. O, Gilpatrick . - 92. M. E. Whatley
W. R. Grimes = , 93, J. C. White

A, G, Grindeln =~ 94, Gale Ybung
R. H. Guymon -~ =~

 EXTERNAL DISTRIBUTION |

'E G. Case, Dlrector, DiV181on of Reactor Standards, AEC
Washington, DC. 20545
D. F. Cope, RDT, S8R, AEC, Oak Rldge Natlonal,Laboratory
A. R. DeGrazia, AEC, Division of Reactor Development and
Technology, Washington, DC 20545

.. Executive  Secretary, Adv1sory Committee on Reactor Safeguards,

p | , | | 105.7

AEC, Washington, DC 20545
J. E, Fox, AEC, Division of Reactor Development and Technology,
Washington, DC 20545

 
 

 

106,

107.

110-111.

112.
113-114,

115,
116.
117,

118.
119-120,

28

Norton Haberman, AEC, Division of Reactor Development and
Technology, Washlngton, DC 20545

Kermit Laughon, RDT, OSR, AEC, Oak Ridge National Laboratory

A, P, Litman, AEC, Division of Space Nuclear. Systems,
Washington, DC 20545

T. W. McIntosh, AEC, Division of Reactor Development and

- Technology, Washlngton, DC 20545

H. G. McPherson., University of Tennessee, KnoxV1lle, TN 37916

Peter A. Morris, Director, Division of Reactor Llscen31ng, AEC,
Washington, DC 20545

J. F., Neff, AEC, Division of Reactor Development and Technology,
Washlngton, DC 20545

Sidney Siegel, Atomics International, P.0O. Box 309 Canoga
Park, cA 91304

‘M. Shaw, AEC, Division of Reactor Development and Technology,

Washington, DC 20545
Laboratory and University Division, AEC, Oak Ridge Operations
Division of Technical Information Extension