ORNL-CF-71-8-10.txt

-

 

. : :DE L . 07 iia
v IWE'Eflnllnufiu;h--*““flfi*J (}Efil?iltr

; r o PN ';i'fldj ?f-\.. :‘?;.-:1_ mw_r
M&g? 3 L L |

OAK RIDGE NATIONAL LABORATORY
et TIhEerarr—taetreT

UNION CARBIDE CORPORATION

uumglsml 0 R N L%C }'\

rost omic sox CENTRAL FILES NUMBER

OAK RIDGE, TENNESSEE 37830

 

71-8-10

 

 

 

DATE: August 6, 1971

SUBJECT: A Review of Possible Choices for Secondary Coolants
for Molten Salt Reactors

TO: Distribution

FROM: J. P. Sanders

Abstract

A review has been made of various choices for secondary coolants
in a molten salt reactor system which has as its primary fuel salt a
7LiF-BeF,~-ThF, -UF, mixture. Special consideration was given to in-
hibiting the migration of tritium from the primary salt through the
secondary salt to the steam system. At the same time the performance
characteristics of these coolants, the associated inventory costs,
their compatibility with the fuel salt, steam, and the cell environ-
ment, and the materials required to contain the coolants were summa-
rized for future reference. Three groups of coolants were considered;
these were (a) molten salts and sodium hydroxide, (b) liquid metals,
and (c) gases at a high or moderate pressure.

Of the coolants which were considered, there appeared to be no
real incentive for a choice other than the sodium fluoroborate which
is currently being studied. It appears that it might be possible to
contain a small concentration of hydroxl ion in complex in this coolant
which could be continuously removed and replaced to function as a trap
for the tritium. A present disadvantage of the sodium fluoroborate
appears to be its high liquidus temperature of T725°F. A lower melting
point fluoride mix of LiF, NaF and BeF, with a liquidus temperature of
640°F or a mix of LiCl and KC1 with a liquidus temperature of 680°F
appear to be alternatives.

M @‘%m\aii\“

 

.'P
ii

Sodium with a melting point of 208°F is the best choice of the liquid
metals; several design problems are introduced through the choice of
sodium as a secondary coolant and no real incentive exists to select this
as a coolant unless the lower melting temperature is desired. Of the
gases, helium is the best choice; however, with this, or any other gas,
it is necessary to design a secondary loop which operates at relatively
high pressures. Accurate evaluation of helium as a secondary coolant
will require considerable investigation into possible designs for such
a secondary system.

Keywords: *review, ¥coolants, ¥MSBR, *secondary salts, physical
properties, performance, tritium, capital costs, materials, inventories,
fluoroborates, fluorides, sodium hydroxide, /chlorides, liquid metal,
gases.

 
i w o+

O

iii

CONTENTS

INTRODUCTION. c o e v e vt ettt st s svanecnnnnnnas
PHYSTICAL PROPERTY VALUES.....cvtcveeneannnn.

----------------

COMPARISON OF PHYSICAT, PERF'ORMANCE OF COOLANTS....cocoveeeen

INVENTORY COSTS . vt iiie it it nnanssneosnanens

THE SIGNIFICANCE OF THE MELTING OR LIQUIDUS TEMPERATURE QF THE

oooooooooooooooo

SECONDARY COOLANT IN MOLTEN-SALT REACTOR SYSTEM DESIGN.....

EFFECTIVENESS OF COOLANTS IN TRITTUM HOLDUP.

COMPATIBILITY OF SECONDARY COOLANTS WITH THE
STEAM, AND THE CELL ENVIRONMENT............

7.1 Compatibility With the Primary Fluid...

oooooooooooooooo

PRIMARY SALT,

T.1.1 With the molten salts as secondary coolants.....

7.1.1.1 Sodium fluoroborate....

7.1.1.2 Lithium fluoride—beryllium fluoride....
7.1.1.3 FLINAK and the "low melting point”

FlUuoride . i vee e cenens

7.1.1.4 The chloride salts.....

oooooooooooooooo

7.1.1.5 ©Sodium hydroxide, the nitrates, and

the carbonates........

oooooooooooooooo

T7.1.2 With liquid metals as secondary coolants........

7.1.2.1 Sodium and NaK.........

T.1.3.1 Helium....:.eoeeeoeneaes
7.1.3.2
T.1.3.3
7-1.3.4
7.1.3.5 Nitrogen...............
7.1.3.6 Carbon dioxide.........
7.1.3.7
7.2 Compatibility with Steam...............
7.3 Compatibility with the Cell Environment

ConclusionNS . veeeesoenses

----------------

----------------

oooooooooooooooo

oooooooooooooooo

----------------

----------------

16
DD

28
28
29
29
30

30
31

31
33
33
34
34
3k
35
35
36
36
36
36
36
37
37

 
iv
Page

8. MATERTALS OF CONTATNMENT ++veeveceronncncceasesassssnenasaes 30

8.1 (Containment Materials for the Various Secondary

Coolant Systems .e..eeeveen. Ceeeeeeneas Ceresee e 38
8.1.1 Materials of construction for molten salts
as secondary coolants .eceeieasnos ceeresessesas hO
8.1.1.1 Fluoride s81t5 .vevvrerrcrsasnss ceeenen 40
8.1.1.2 Chloride s8ltsS coseeeececeesss ceereeee. U2
8.1.1.3 Sodium hydroxide ..seeeececec.. eeeeess 43
8.1.1.4 Nitrates ...eeeeeeeecesss Ceiiereeseeees U5
8.1.1.5 Carbonates ....veeesnssens ceeearane vee. 45
8.1.2 Materials of construction for liquid metals as
secondary coolants ..ecvee. teeesscsennansasass U5
8.1.2.1 SodiUmM ceveeececscoceronons teesseranaeas 46
8.1.2.2 Sodium—potassium (NaK) ..... ceeeaneeee. U6
8.1.2.3 Bismuth ..coveeueenennn Ceeesasenaas ve.. 46
8.1.2.4 Bismuth-lead ............. R ¢
8.1.2.5 Lead seveverrenceneaanan Ceieeaseeareans L7
8.1.2.6 MErcCury «eoeeeesess e ceeoe. kb7
8.1.3 Materials of construction for gases as
secondary co0lants ...eeecesonnss Ceeceeneiennan 48
8.1.3.1 Helium vuoveveveerennnenn et rie e . L8
8.1.3.2 Hydrogen ..... Ceereteete e cee... U8
8.1.3.3 Water VADPOT +evessensocosstsasasoasssss 48
8.1.3.4 Air ........ C et eae e e 48
8.1.3.5 NitrogeD +vveeceecnserisssssncononas ... L8
8.1.3.6 Carbon dioxide ...cveveeeeeenn. R

8.2 (ontainment Materials for Various Secondary Coolants
in Conjunction With the Primary Salt ....eeeeecesesses U9

8.2.1 Materials with molten salts as secondary

coolants ........ et reesee et .. L9
B8.2.1.1 Fluoride SaltS +vvesesrevonosseoseasanns L9
8.2.1.2 Chloride 5281tS v.veeeseessvens ciereenes 51

8.2.1.3 Sodium hydroxide ...eeveeeervoeesssases 51l

8.2.1.4 Nitrates and carbonates ....ceveeseeee. 51

 
Page
8.2.2 Materials with liquid metals as secondary
coolants .iiiiieeens chiescobsiiceteaaronseasaas DL
8.2.2.1 Sodium and NaK .eeeeeeesesvenceosvsnses oL
8.2.2.2 Bismuth and bismuth—lead .......... cee. ol
8.2.2.3 METCUTY +vvennoevroeanens ceeesennneers. 21
8.2.2.4 Tead sveeevevivernenns e cesees coe 01
8.2.3 Materials of construction for gases as secondary
COOLANES tutreresronnnnannncensonssoossonannnss o2

8.3 Containment Materials for Various Secondary Coolants
in Conjunction With Steam ..viieveereeirenececreeneans 92

8.3.1 Materials with molten salts as secondary

coolants ...... Gt csee e s esocea st etoe e enenon cas 02
8.3.1.1 Fluoride SaltS .e.ieeeeeeeeneeens ceieea. 52
8.3.1.2 Chloride 58lts .cvveveernnennnnn. ceeses D3
8.3.1.3 Sodium hydroxXide +vveveeeeeeonenrerenns 53
8.3.1.4 Nitrates and carbonates .....ceveeeeeas 53
8.3.2 Materials with liquid metals as secondary

coolants ..eeeeenns Cererie e Ceeteerecteeae ver 93
8.3.2.1 Sodium and NaK ...... e caeeo ee. 93
8.3.2.2 Bismuth, lead, and bismuth—lead ....... 53
8.3.2.3 Mercury ....... C et ceeeaes 23
8.3.3 Materials with gases as secondary coolants ..... 23
8.4 Literature SOUTCES .eeveverorerennnronnreoonesns Ceeenes 5k

8.5 Summary and Conclusions Concerning Materials of
CONStrUCEION tivvertenoeeosarenacoanossannosseasananas 5k

OTHER POSSIBLE COCLANTS OMITTED FROM THE PREVIOUS

PRESENTATION =+ s e v e eesenncaseaeeeeannnnneeeeeeseaaneeeeeess 61
9.1 Alternate Fluoride Salts ..... e 61
9.2 The Cyanides ..oeeeecoesscasseasssssnessscsssonosssosss 63
9.3 Organic Secondary COOLaNtS +eeiesssonssnossosssssasssas 63
9.t Sulfur ..eeviiieeoon Ceeereeona C e teeetecees o stessenns 63
9.5 HAloAlUMINALES +severnnreeennuneeeennonneennnnnens vo.. Oh
SUMMARY AND CONCLUSIONS « v v oossonevanoasocssonocanaanaassans 65

ACKNOWLEDGMENTS «.ov0... S e e e ea e s ‘10

 
Teble

Table

Table

Table

Table

Table

Table

Table

Table

vi

LIST OF TABLES

Possible Choices for Molten Salt Reactor Secondary
Coolants Which Were Considered in this Study ...... .

Physical Properties of the Coolants at 1000°F
(538%C) iviienrereenennenenennannnns Ceeeiearereraen

Melting Temperatures for the Coolants ...............
Comparison Factors for Secondary Coolants at 1000°F..

Cost of Secondary Coolants, Estimated Total
Inventories, and Estimated Total Inventory

COSTS teeevoseoesrsessessosnsosscosossssssssasossscsssess

Alliowable Stresses and Design Pressures for Various
Materials .ieiieireneereeosssnnosssnrsscsnsssnnsasnnss

Performance of Materials in Contact with the
Secondary Coolant Only +evveeeavsevsescssnas cheaaea .

Performance of Materials in Contact with the
Secondary Coolant and the Fuel Salt .........c..000

Performance of Materials in Contact with the
Secondary Coolant and with Steam ....ccivveeeveerens

Page

12

15

56

58

29

60

 
1. INTRODUCTION

A review of possible choices for secondary coolants for molten-salt
reactors has been initiated. Several problems prompted this coolant
study. ©Sodium fluoroborate, which had been investigated most extensively
as a possible secondary coolant, seemed to be more corrosive than was
desirable; there was concern about the effects of the mixing of this
coolant with the fuel salt in the event of a primary heat exchanger tube
rupture; and the pressure of the required BF, cover gas would be greater than
one atmosphere at temperatures of 1300°F and above which might develop
under abnormal operating conditions. An additional problem which prompted
this study was the consideration of the secondary coolant as a tritium
trap or a tritium reservoir for that tritium which migrated from the pri-
mary salt through the heat transfer area of the primary heat exchanger.

These have not been the only characteristics of the secondary coolant
considered in this study; an effort has been made to. bring to- |
gether as much additional information from all possible sources which
would concern the performance of the secondary coolant in the molten-salt
reactor system. This information includes the physical properties of the
coolant; possible materials of contaimment for the coolant itself and for
the coolant in conjunction with both the primary salt and with steam; the
compatibility of the coolant with the primary salt, the steam, and the
reactor cell environment; the cost and availability of the coolant;
safety considerations in the use and handling of the coolant; and any
other general information.

The ultimate purpose of this study is to evaluate the various coolants
in a reactor system design which will be optimized for the characteristics
of that coolant. Since it will not be practical to develop a design for
each choice for the secondary coolant, the coolants have been grouped into
three categories, viz., molten salts, liquid metals, and gases, and it is
postulated that a design which will be developed for the category will be,
at least, partially optimum for each of the coolants in that cétegory.

A comparison of the characteristics of each coolant in the optimized de-
sign of the category should then provide a basis for the selection of the

ultimate optimum coolant.
As the first phase of this study, the characteristics of the more
obvious choices for coolants in each of the three categories have been
collected and tabulated. The choices include nine molten-salt mixtures,
six liquid metal coolants or mixtures, and six different gases each con-
sidered as a high-pressure gas (1000 psia) and a low-pressure gas (250
psia). A list of these choices is presented in Table 1 which gives the
composition of the mixtures following a special designation which will
be used in referencing the particular coolant in subsequent tabulations.
The designations have been selected to be compatible with the characters
printed by the standard teletype machine so that subsequent tabulated
data can be printed by this unit.

This report has received an initial review by several members of
the MSR Program and has been revised to incorporate many of their comments.
It is issued at this time as a Central Files (CF) report for comment by
the entire staff prior to final revision and publication as a Technical
Memorandum (TM). Comments or corrections that will make the report more
useful will be appreciated. All information in the report should be con-

sidered tentative until the revised report is issued.
 

 

 

 

Table 1. Possible Choices for Molten Salt Reactor Secondary Coolants
Which Were Considered in this Study
Designation Composition
NA BF4-NA F 92 m/o NaBF, , 8 m/o NaF
LI F--BE F2 66 m/o LiF, 34 m/o BeF, |
FLINAK 46.5 m/o LiF, 11.5 m/o NaF, 42.0 m/o KF
NA+K F--ZR F4t  10.0 m/o NaF, 48.2 m/o KF, 42.0 m/o ZrF,
NA ¢H 100 o/o NaOH
LI CL--K CL 59.0 m/o LiCl, 41.0 m/o KC1
NITRATES 6.9 m/o NaNO,, 48.5 m/o NaNO5, 4kh.6 m/o KNO,
IMP FLUORIDES 35.0 m/o LiF, 27.0 m/o NaF, 38.0C m/o BeF,
CARB@NATES 43.5 m/o LiysCO,, 31.5 m/o NayCO,, 25.0 m/o K,CO4
S@DIM 100 o/o Na
NAK 4 w/o K, 56 w/o Na
BISMUTH 100 o/o Bi
PB--BI 55 w/o Bi, 45 w/o Pb
MERCURY 100 o/o Hg
LEAD 100 o/o Pb
HELIUM(HP) 100 o/o He at 1000 psia
HELIUM(LP) 100 o/o He at 250 psia
HYDR@GEN (HP) 100 o/o H, at 1000 psia
HYDRJGEN(LP) 100 o/o H, at 250 psia
WATER (HP) 100 o/o H,0 at 1000 psia
WATER (LP) 100 o/o H,0 at 250 psia
ATR(HP) Air (79 m/o N,, 21 m/o 0,) at 1000 psia
ATR(LP) Air (79 m/o Ny, 21 m/o 0,) at 250 psia
NITRGGEN(HP) 100 o/o N, at 1000 psia
NITRGGEN(LP) 100 o/o N, at 250 psia
c@g2(HP) 100 o/o CO, at 1000 psia
cg2(LP) 100 o/o CO, at 250 psia
Note: m/o designates mole per cent; w/o designates weight per cent;

where the desighation is not significant as for pure compo-
nents and compounds, o/o indicates per cent.
2. PHYSICAL PROPERTY VALUES

Physical property values for each of the coolants are presented in
Table 2. With the exception of mercury, all physical properties are
given for the coolant at 1000°F; property values for mercury are given
for a temperature of 600°F because values were not readily available at
the higher temperature and no accurate method of extrapolation to this
temperature existed.

For the most part, physical property values for the molten-salt
mixtures were obtained from the following references:

1. S. Cantor, Editor, "Physical Properties of Molten-Salt Reactor
Fuel, Coolant, and Flush Salts," USAEC Report ORNL-TM-2316,
Oak Ridge National Laboratory, August 1968.

2. J. R. McWherter, "MSBR Mark I Primary and Secondary Salts and
their Physical Properties," internal memorandum, MSR 68-135,
Rev. 1, Sept. 27, 1968; revised Feb. 12, 1969.

3. H. A. McLain, "Revised MSBR Secondary Salt Viscosity Corre-
lations," internal memorandum, MSR 69-103, Oct. 27, 1969.

L. J. W. Cooke, internal memorandum, MSR 68-123.

5. H. F. McDuffie, et al., "Assessment of Molten Salts as Inter-
mediate Coolants for IMFBR's,  USAEC Report ORNL-TM-2696,
Oak Ridge National Laboratory, Sept. 3, 1969.

6. S. Mclain and J. H. Martens, Editors, Reactor Handbook, 2nd
Ed., Interscience Publishers, New York (1964).

7. H. W. Hoffman and S. I. Cohen, "Fused Salt Heat Transfer--
Part III: Forced-Convection Heat Transfer in Circular Tubes
Containing the Salt Mixture NaNO,—NaNO,—KNO,," USAEC Report
ORNL-2433, Oak Ridge National Laboratory, March 1, 1960.

 

Additional information for the liquid metal properties was obtained
from:

8. John G. Yevick, Editor, Fast Reactor Technology: Plant Design,
The MIT Press (1966).

 

Gas properties were obtained from:

9. J. Hilsenrath and Y. S. Touloukian, "The Viscosity, Thermal
Conductivity, and Prandtl Number for Air, O0,, N,, NO, H,, CO,
CO,, H,0, He, and A," Trans. ASME, August 1954, pp. 967-985.

10. National Bureau of Standards, "Tables of Thermal Properties
of Gases,'" Circular 56.L.

11. International Critical Tables.

 
12. Fluid Prgpe}ties Handbook, Missile and Space System, Douglas
Aireraft Company, Inc.

 

13. C. A. Meyer, et al., Thermodynamic and Transport Properties
of Steam, ASME, 1967.

14. E.R.G. Eckert and R. M. Drake, Jr., Heat and Mass Transfer,
ond Ed., McGraw-Hill Book Company, Inc., New York (1959).

 

 

Most of the physical property data are well established with the
exception of the thermal conductivities. For some fluids, various refer-
ences indicated values for the thermal conductivity of a certain coolant
which differed by a factor of L4 or more. 1In these cases, the value of
more recent origin or of the moré careful investigator was selected.

All of the last group of proposed coolants are gases at normal ambi-
ent temperatures (or slightly above ambient for COQ). The melting point
for the remainder of the coolants is given in Table 3. With the exception
of mercury, all of the coolants listed in Table 3 have vapor pressure
less than 1 atmosphere at the normal operating temperatures of the secon-
dary loop of the current design for MSR. Mercury has a vapor pressure

of 1 atmosphere at 6TL°F and a vapor pressure of 12 atmospheres at 1000°F.

 
Table 2. Physical Properties of the Coolants at 1000F (538QC)

 

 

Thermal Heat

Secondary Density Viscosity Conductivity Capacity
Coolant (1b/cu ft) (1lb/ft-hr) (Btu/hr~-ft-F) (Btu/lb-F)

NA BF4-NA F 116,90 3.355 . 2660 .3600
LI F--BE F2 124.06 28.780 . 5780 . 5700
FLINAK 133,36 16.560 . 7500 .4410
NA+K F-ZR F4 159,82 22,730 .5780 .2500
NA OH 91.52 3.797 .6048 .5074
LI CL--K CL 100,29 4,707 .4870 . 3015
NITRATES 108.33 1.030 .1900 .3730
LMP FLUORIDE 124,80 28.180 .4000 .5900
CARBONATE 125,00 18,272 .2000 .4130
SODIUM 51,20 .558 37.80 . 3350
NAK 48.80 .479 16.40 .3070
BISMUTH 608,00 4,320 9.00 .0369
PB-BI 623.00 3.010 8.14 .0333
MERCURY (600F) 802,00 2.090 8.10 .0320
LEAD 650,00 4.100 8.90 .0346
HELIUM (HP) .25510 .0958 .1740 1.2424
HELIUM(LP) .06378 .0958 .1720 1.2421
HYDROGEN (HP) .12868 .0425 2380 3.4800
HYDROGEN (LP) 03217 . 0425 .2380 3.4775
WATER (HP) 1.20550 .0741 .0460 « 5650
WATER (LP) .29085 .0730 .0440 .5250
AIR(HP) 1.85100 .0877 .0330 « 2650
AIR(LP) 46300 .0800 .0320 .2640
NITROGEN (HP) 1.78800 .0930 .0330 2770
NITROGEN (LP) .44700 .0900 .0320 .2750
CO2 (HP) 2.80900 .0810 .0381 . 2860

CO2 (LP) . 70200 . 0800 .0380 .2810

 

 
Table 3. Melting Temperatures for the Coolants

 

 

Melting
Secondary Temperature
Coolant (Degrees F)
NA BF4-NA F 725
LI F~=-BE F2 856
FLINAK 851
NA+K F--ZR F4 725
NA OH 318
LI CL--K CL 680
NITRATES 288
IMP FLUORIDE 640
CARBONATE 747
SODIUM 208
NAK 66
BISMUTH 520
PB-BI 257
MERCURY -38
LEAD 621

 

For the molten salt mixtures, the melting
temperature that is indicated is the liquidus
temperature for the composition in Table 1,

 
3. COMPARISON OF PHYSICAL PERFORMANCE CF COOQOLANTS

In order to compare the performance of the various choices of cool-
ants within any one group on the basis of their physical properties alone,
certain factors have been formulated. It was assumed that the two aspects
of the coolant system which would affect the economics of the reactor
operation were the heat transfer area requirements and the circulator
power requirements.

If we assume that the same temperature regime is maintained for all
reactor designs employing all choices of coolants within a group, then the
heat transfer area requirements in both the primary and secondary heat
exchangers are related directly to the reciprocal of the film heat transfer
coefficient (neglecting the contribution to the resistance to heat transfer
by the tube wall and by the film of fluid on the other side of the wall).
According to current designs the secondary coolant will be in the shell of
the primary heat exchanger to minimize fuel salt inventory and in the shell
of the secondary heat exchanger so that the high pressure steam will be
inside the tubes.

The heat transfer coefficient for flow outside of tubes is given by

*
the equation,

 

h D /EB_ 1/3 DG .. 0.6
K, O (Sf (T;’) (1)

where h is the mean heat transfer coefficient for the tubes in the bundle,
DO is the outside diameter of the tubes in the bundle, k is the thermal
conductivity of the fluid, Cp is the heat capacity of the fluid, and p is
the viscosity of the fluid. The subscript f indicates that the fluid
properties are to be evaluated at a theoretical film temperature which

is defined as the average of the mean bulk fluid temperature and the sur-
face temperature of the tube wall. Gmax indicates the maximum mass
velocity (mass per unit time per unit minimum flow area) of the fluid in
the bundle. Gmax can also be written as meax where p is the fluid den-

sity and Vmax is the maximum linear velocity of the fluid in the bundle.

 

*W. H. McAdams, Heat Transmission, 3rd Ed., McGraw-Hill Book Company,
Inc., New York (1954), p. 272.

 

 
The density of the fluid should be evaluated at the mean bulk fluid tem-
perature since it is a measure of the total momentum in the fluid. It

has been suggested that all of the coolant properties can be evaluated at
the bulk temperature if the exponent of the Prandtl number (Cpp/k) is
changed to a value of 0.4 when heat is transferred into the fluid and a
value of 0.3 when heat is transferred out of the fluid. In the secondary
loop, heat is transferred into the coolant in the primary heat exchanger
and out of the coolant in the steam generator. Since the area requirements
in the primary heat exchanger will probably be more critical in the reactor
system design, an exponent of O.4 will be used. For any given reactor
power, the heat load for the various coolant choices will be approximately
the same (they will differ only by the differences in pumping power re-
quired in this loop), and within any of the three groups of coolants, the
temperature regimes will be about the same, the heat transfer area require-

ment can then be expressed as

0.2.0.4

A x b D (2)
O 06 037 203

P
If the assumption is made that for all system designs for coolants in the

 

same group the same outside tube diameter and the same work per unit vol-
ume (pVe) is maintained for all choices within that group, the total heat

transfer area requirement can be expressed as follows:
0
AR —ps (3)

The assumption that the same work per unit volume exists for the various
choices of coolants within a group can be interpreted as assuming that
pumping power requirements are similar.

The value of this area factor, based on the values and units indi-
cated in Table 2, is shown for each of the coolant choices in Table k.
These values should not be used for comparing the performance of coolants
among the three groups (i.e., between liquid metals and gases or between
molten salts and liquid metals); they should only be used as an indication

of the relative area requirement of a coolant within a group.

 
10

An expression for the work requirement of the circulator in the sec-
ondary loop would be the product of the volumetric flow rate and the pres-
sure loss around the loop. If the major contribution to the pressure loss
is fluid friction, an empirical expression for the friction factor of the

form
-002
" DVp)
&= L
( H ( )

can be used to generate a factor indicating the relative circulator power.

For this calculation, in addition to assuming the same equivalent hydraulic
diameter for each system, we shall assume the same effective flow area and

the same mass velocity (i.e., linear velocity multiplied by density). Im-

plicit in this set of assumptions is that the product of the heat capacity

and the temperature change around the loop is a constant. We then find

that

Work X~ ET (5)

P

If we call this the work factor, we find the relative values listed in
Table L. Again the factors among various groups should not be compared.
If for any of the groups of coolants, the area requirements and

circulator power requirements are of similar importance, we may combine
the equations used above with the assumptions of similar hydraulic diame-

ters and similar mass velocities to get a combined term which has the form

0.40
Combined * — 2°55 0% (6)
0 k CP

Again this factor is listed in Table 4 for the three groups of coolants.
It is felt that this factor is the more significant of the three. Note
that the last factor cannot be obtained by multiplying the first two

 

since the first was based on the assumption of similar work requirements
prer unit volume of coolant and the last two were based on similar mass
velocities.

If it is assumed that the system is to be designed with similar pres-
sure losses, then the comparison factor reduces to the density multiplied

by the heat capacity (this is actually the volumetric heat capacity). The

 
11

reciprocal of this product is listed as the transport factor. This indi-
cates, if the pressure gradient around the secondary loop is an important
design consideration, a coolant with a low value for the transport factor
will be the most favorable choice.

If systems with similar head losses are assumed, then the heat capac-
ity is the important physical property. If the secondary circulator is to
be limited by design considerations to some maximum developed head, the
fluid with the highest heat capacity will be the most favorable choice.

The Prandtl number also is &an indication of the relative ease of heat
transport to momentum transport (pressure drop) on a molecular basis.

The coolant in any group with the lowest value of the Prandtl number, will
obviously contribute the most on a molecular basis to the transfer of
heat. Fluids with higher values of the Prandtl number will require more
circulator work, adding momentum to the coolant, to achieve the same

heat transfer.

 
12

 

 

 

Table 4. Comparison Factors for Secondary Coolants at 1000F
Secondary Area Work Combined Transport Prandtl
Coolant Factor Factor Factor Factor Number
NA BF4-NA F 1.70 1.27 1.502 1.750 4,541
LI F--BE F2 1.34 1.74 1,588 1.041 28,382
FLINAK 1.11 1.35 1.000 1.252 9.737
NA+K F-ZR F4 1,65 1.00 1.040 1.843 9.831
NA OH 1.00 2.13 1.588 1.586 3.186
LI CL--K CL 1.42 1.85 1.914 2,435 2,914
NITRATES 1.66 1.17 1.378 1.822 2.022
LMP FLUORIDE 1,64 1.71 1.908 1,000 41,565
CARBONATE 2,63 1.56 2,793 1.426 37.732
SODIUM 1.00 188.41 116.946 1.496 .00495
NAK 1.68 201,16 213,023 1.713 .00897
BISMUTH 4,10 2.01 2.436 1.144 01771
PB~BI 4,19 1.78 2.190 1,237 .01231
MERCURY (600F) 3,68 1.00 1.000 1.000 .00826
LEAD 4,11 1.74 2,071 1.141 .01594
HELIUM (HP) 1.75 125.4 122.18 2.535 .6840
HELIUM(LP) 2.66 2005.9 4521,98 10.141 .6918
HYDROGEN (HP) 1.00 418.9 287.07 1,794 .6214
HYDROGEN(LP) 1,52 6701.2 10554.47 7.181 .6210
WATER (HP) 3.17 5.3 5.92 1.180 «9101
WATER (LP) 5.12 91.4 250,96 5.261 .8710
AIR(HP) 4,76 2.3 3.43 1.638 - 71043
AIR(LP) 7.23 36.7 123.86 6.573 .6600
NITROGEN (HP) 4.78 2.5 3.78 1,622 . 7806
NITROGEN(LP) 7.36 40.3 139.96 6.535 .7734
CO2 (HP) 3.68 1.0 1.00 1.000 .6080
CO2 (LP) 5.61 16.0 36.93 4,073 .5916
Minimum Values of the Factors in Each Group are:
Factor Molten Salts Liquid Metals Gases
Area 5.9748E=~01 4,7867E-02 1.4133E+00
Work 7.3124E-05 1.8017E-06 7.6664E-02
Combined 1.5127E-05 4.2663E-08 2.9245E-01
Transport 1.3581E-02 3.8965E-02 1.2448E+00

 
 

13

L. INVENTORY COSTS

An important consideration in making a choice among some of the pos-
sible secondary coolants is the inventory charge associated with the use
of that coolant. This total charge is a product of the total volume of
secondary system, the density of the secondary coolant, and the cost of
the coolant per unit weight. The product of the first two terms is, of
course, the total weight of coolant which is required for the system.

In Table 5, an estimate is shown for the relative volumes of secon-
dary systems for lOOQ—MW(e) reactors for the various choices of coolant.
These estimates are based on extrapolations from available designs or
design estimates plus a consideration of the "Area Factor'" and the "Com-
bined Factor" of Table 4. Multiplication of this estimated volume by
the densities in Table 2 produces the estimated weight of the coolant
inventory which is shown in Table 5.

The unit costs for the coolants were obtained with the assistance of
J. M. Campbell of the Union Carbide Purchasing Department; he obtained
preliminary bids from vendors where possible. All costs exclude shipping
costs; however, the incremental difference in cost which would be made by
including this factor is probably less than some of the present cost un-
certainties. Most of the vendors were not prepared to quote on the basis
of the shipment of quantities as large as those indicated in the inventory
lists. Bids indicated that for most items the unit cost would decrease
as the size of the order increased; where this factor appeared to be sig-
nificant, extrapolation of the available information was made to estimate
the cost in quantities as indicated in Table 5.

For some of the compounds, very small differences in the amount of
impurities made quite significant changes (by a factor of 10) in the unit
cost of the material. Since it was assumed that any reactor plant would
have some continuous processing facilities for the purification of the
coolant, choices were made to accept the compound with slight impurities
if it appeared that the impurities could be removed with relative ease
during chemical processing and if it appeared that the impurities would

not be detrimental to the secondary system during this cleanup procedure.
1k

Some of the unit costs listed in Table 5 may appear to be unusual due
to the presence of certain impurities. For example, the cost of "air" is
relatively high because atmospheric air could not be used as a secondary
coolant due to the activation of the small percentage of argon which it
contains. Air, for a secondary coolant, would have to be synthesized
from liquid oxygen and nitrogen. Carbon dioxide and hydrogen are both
relatively expensive due to the difficulty of removing small amounts of
water vapor which occur in the commercially available products.

Special consideration must be given to those choices for secondary
coolants which contain lithium. It appears at the present time that the
primary salt will be LiF-BeFo (the eutectic of lithium fluoride and beryl-
lium fluoride) with the addition of uranium and thorium fluorides. The
lithium in this mixture must be enriched in lithium-7 to decrease the
number of neutrons which are absorbed in this component. If lithium is
also a constituent of the secondary coolant, and if the secondary cool-
ant becomes mixed with the primary salt through leakage in the primary
heat exchanger, then the lithium in the primary salt will become diluted
with the lithium-6 of the secondary coolant unless the lithium in that
coolant is also enriched in lithium-7.

It is obvious that the use of the isotopically-enriched lithium in
the secondary coolant will increase significantly the cost of this compo-
nent. In Table 5, for those secondary coolants which contain lithium, the
costs are indicated for the coolant constituted both with natural lithium
and with lithium which is 99.995% lithium-7. The cost of the additional
inventory charges which are incurred when enriched lithium is used must
be balanced against the possible cost of replacing the primary coolant in
the event of a primary heat exchanger failure or, as an alternative, allow-
ing any lithium-6 which is introduced into the primary loop to be "burned
out" during subsequent reactor operation.

The relative total inventory costé are given in the last column of

Table 5.

 
Table 5.

15

Cost of Secondary Coolants, Estimated Total

Inventories, and Estimated Total Inventory Costs

 

 

Estimated Inventory Unit Total Inventory
Secondary Volume Weight Cost Cost
Coolant (cubic feet) (pounds) (S/pound) (million S)
NA BF4--NA F 8500 993650 «371 . 369
LI F--BE F2 10000 1240600 5.200 6.451
WITH LITHIUM-=7 11,920 14.788
FLINAK 8500 1133560 1.095 1.241
WITH LITHIUM-7 5,111 5.794
NA+K F--ZR F4 8500 1358470 2,074 2,817
NA OH 8000 732160 « 250 »183
LY CL--K CL 8500 852465 . 509 434
WITH LITHIUM-7 4,140 3.529
NITRATES 7000 758310 +150 114
LMP FLUORIDES 11000 1372800 4,342 5.961
WITH LITHIUM-7 7.944 10.906
CARBONATES 10000 1250000 0256 .320
WITH LITHIUM-7 3.370 4,212
SODIUM 4500 230400 .400 .092
NAK 4500 219600 . 800 176
BISMUTH 5000 3040000 6.000 18,240
PB~~BI 5000 3115000 3,380 10.529
MERCURY 5000 4010000 12,750 51.127
LEAD 5500 3575000 180 .643
HELIUM (HP) 20000 5102 9.500 .048469
HELIUM(LP) 20000 1276 9.500 .012118
HYDROGEN (HP) 20000 2574 14.157 .036434
HYDROGEN (LP) 20000 643 14.157 .009109
WATER (HP) 20000 24110 .010 .000241
WATER(LP) 20000 5817 .010 .000058
AIR(HP) 20000 37020 616 .022804
AIR(LP) 20000 9260 .616 .005704
NITROGEN (HP) 20000 35760 .688 .024603
NITROGEN (LP) 20000 35760 .688 .024603
C02 (HP) 20000 56180 1.438 .080787
CO2(LP) 20000 14040 1.438 ,020190

 

All prices are for components which are 99 per cent

pure.

are F,0,B, vendor,
Unless otherwise noted, prices for lithium compounds

are for the naturally occurring element,

No transportation costs have been included; prices

Enriched lithium

compounds are assumed to contain 99.995 per cent LITHIUM-7

priced at $120/kg.

lithium is included.

No chemical conversion cost for this

 
16

5. THE SIGNIFICANCE OF THE MELTING OR LIQUIDUS TEMPERATURE OF
THE SECONDARY COOLANT IN MOLTEN-SALT REACTOR SYSTEM DESIGN

The group of gases that are to be considered for secondary coolants
do not approach either a solid or liquid state (with the exception of
carbon dioxide) at reactor conditions from ambient to normal operation.
The liquid metals retain the liquid state (with the exception of lead and
bismuth) at temperatures near or below the normal boiling point of water
(212°F). The group of molten salts which are included in this study have
liquidus temperatures ranging from approximately 300°F to temperatures
greater than 850°F. It appears, therefore, that the values of the melting
or liquidus temperature will only have importance when comparing various
members within the groups of liquid metals or molten salts.

While it is not obvious that the local formation of a solid phase in
the circulating secondary coolant system will be detrimental to the effi-
cient performance of the system, the design of the system and the selec-
tion of a coolant which eliminates this possibility have certain advan-
tages. Such a system and coolant combination would not require feedwater
temperatures and flow control instrumentation with the complexity and
redundancy of systems that possess the potential for subcooling the
secondary coolant.

If a solid phase which does not adhere to the channel walls is
formed in the coolant, then this solid will be carried by the coolant to
other parts of the loop. Progressive formation of solid will lead to the
generation of a "slush" within the secondary coolant system. Again, this
effect may not be detrimental as long as the solid phase can be converted
back to the liquid phase through the addition of heat or an adjustment of
the relative coolant composition.

A system with the potential of freezing the secondary coolant can be
operated successfully without the production of a solid phase. Such a
system would have a feedwater temperature to the steam generator which was
below the melting or ligquidus temperature of the coolant. During normal,
full-power operation the thermal gradient that is required for transfer-
ring the total heat load in the steam generator is greater than the

difference in these temperatures. A control problem is involved in

 
17

satisfactory operation during rapid changes in power level that result
from elther a reactor scram or a turbine trip, or both. Under these con-
ditions the heat load is rapidly reduced and freezing will occur unless
both the feedwater flow rate and possibly also the secondary coolant flow
rate are coincidently reduced in accordance with a predetermined program.

As the difference by which the melting or liquidus temperature of the
coolant exceeds the feedwater temperature increases, the requirements for
the control of the flow rates and temperatures of both the secondary cool-
ant and the feedwater system during these transients become more stringent.
Greater sensitivity, reliability, and redundancy must be included in the
design of the two control systems, unless the consequences of the forma-
tion of the solid phase within the secondary coolant system are considered
to be of minor importance. The evaluation of these consequences can be
made only after a thorough analysis of the secondary coolant system under
all possible transient conditions has been made. Comparison of the per-
formance of different choices for coolants within a group can then be made
only on the basis of a fixed system design.

The dependence of the comparison of coolant performance on detailed
system design and the relatively high capital cost of feedwater and cool-
ant control systems associated with such a design can be eliminated if
the melting or liquidus temperature of the coolant 1is less than or
equal to the feedwater temperature. As indicated above, due to the resis-
tance to the transfer of heat between the feedwater and secondary coolant,
the feedwater temperature can be lower than the melting or liquidus tem-
perature without incurring the formation of a solid phase in the second-
ary system. As the feedwater temperature is lowered and the difference
gets larger, however, the control requirements become much more restric-
tive and the instrumentation required to accomplish this control becomes
more expensive. If the stipulation is made that the feedwater temperature
can be no more than some limiting difference, such as 25 or 50°F, below
the melting or liquidus temperature of the coolant, then the cost of the
instrumentation for the various systems will be comparative.

This assumption relates the melting or liquidus temperature of the
secondary coolant to a major system design parameter, the inlet feedwater

temperature to the steam generator. The inlet feedwater temperature

 
18

must, in turn, be related to the steam system pressure through the pres-
sure-temperature relationship for water and by the conventions of turbine
plant design. Using the restriction of 50°F for the maximum difference
between the melting or liquidus temperature of the coolant and the feed-
water temperature would require that secondary coolants with melting or
liquidus temperatures greater than 755°F be used with systems which had
feedwater temperatures greater than T05°F. Since TO5°F is the critical
temperature of water, feedwater pressures for these systems must be super-
critical; otherwise the feed is actually superheated steam and the entire
steam generator unit becomes a superheater. This results in generally
lower overall heat transfer coefficients and therefore larger heat trans-
fer areas for any particular thermal duty.

Conventional steam turbine plant designs employ throttle pressures
up to 2500 psi for subcritical systems and throttle pressure greater than
3500 psi for supercritical systems; the critical pressure of water is
3206 .2 psia. The pressure loss in once-through steam generators in con-
ventional turbine plants is in the range of 600 to 1000 psi. This pres-
sure loss is due to frictional loss of flow in the tubes, header losses,
and the loss due to orifices placed in the headers to suppress flow insta-
bilities. Feedwater pressures usually then range from 3000 psia or
greater for plants with 2400 psia throttle pressure and from 4300 psia for
supercritical systems.

The opinion has been expressed that first-generation molten-salt
reactor designs should not have the complicating factors in their design
and operation which are associated with supercritical turbine plant
cycles. If the feedwater temperature limitation that was proposed above
is coupled with the stipulation that the plant design incorporates a sub-
critical steam cycle, a plant design is still possible. Supercritical
pressure could be specified for the feedwater to improve the performance
of the steam generator, and the steam cycle could be operated at sub-
critical pressures by allowing a larger pressure drop at the throttle
valve of the turbine. Since this larger pressure drop at the throttle
valve would have to be supplied by the feedwater pump, the power require-
ments for this pump would be larger than the requirements for a sub-

critical cycle of similar capacity.

 
19

Present steam plant designs usually incorporate feedwater tempera-
tures ranging from 480°F for subcritical cycles to 550°F for supercritical
cycles. The specification of a feedwater temperature equal to or greater
than a value of 50°F below the melting or liquidus temperature of the
secondary coolant will result in departures from these standard design
values for most of the possible choices of molten-salt coolants and for
two possibilities for the liquid metals. These higher feedwater tempera-
tures can be obtained through modifications in conventional steam plant
design which may decrease the net power generated by the cycle and/or in-
crease the capital cost of the turbine plant.

One method of increasing the feedwater temperature would be to simply
add more feedwater heaters to the train. To limit the operating tempera-
ture of the feedwater pump, the last heaters would probably be placed in
the discharge line of the pump. The units would then be high-pressure,
high-temperature heat exchangers with significant heat transfer areas.

For this additional feedwater heating, more extraction steam would be re-
guired. To obtain the high feedwater temperatures with a reasonable heat
transfer area requirement for the last-stage feedwater heaters, either
prime steam or steam extracted from the high-pressure turbine would be
required.

Several methods of obtaining the high feedwater temperatures by modi-
fication of the steam generator exist.. All of these methods involve
heating the feedwater with steam from the unit; this steam is then re-
turned to the unit for further heating or resuperheating. The reentrant
type of steam generator performs this funection internally by supplying
surface inside the steam generator through which heat is transferred from
the hotter stream to the entering feedwater. The same function can be
performed external to the steam generator simply by extracting steam at a
point, passing this steam through a conventional heat exchanger to heat
the feedwater, and then returning the degraded steam to the unit. The
reentrant design has the disadvantages that the internal heat transfer
area is more expensive than a similar area in an external unit and that
there is no control of conditions and fliows for the four streams involved
as there is in an external system. The external system has the disadvan-
tage that the collection and redistribution headers for the extraction and

return of the hot stream are expensive.

 
 

20

Another possible design is one in which the prime steam is recycled
to a contact heat exchanger where it is mixed with the normal feedwater
to produce a fluid with the desired inlet temperature. The disadvantage
of this system is that a recirculation pump is required. This pump may
be placed either in the recirculated prime steam stream, or it may be
placed in the mixed stream from the contact unit. In the first configura-
tion, the capacity of the pump is lower, but its operating temperature 1is
that of the prime steam; in the second location, the operating temperature
is the same as the feedwater temperature, but the capacity of the pump is
equal to the total of the feedwater and the recirculation streams. The
pressure head of the pump in both configurations would be equal to the
pressure loss through the steam generator unit.

To illustrate the operation of this system, assume that it was
desired to raise the feedwater temperature from SO0°F to TOO°F for a
supercritical system with throttle conditions of 3500 psia and 1000°F.

The energy that must be added to the feedwater stream is 335.5 Btu per
pound of feedwater, and the amount of recirculated steam required would

be 0.558 pounds per pound of feedwater. Even allowing for a 1000 psi head
requirement for this recirculation pump, the theoretical work reguired is
L1118 ft-1b per pound of recirculated steam or 2300 ft-1b per pound of
normal feedwater flow. Since for a L0% efficient turbine plant with these
steam conditions, the net amount of energy extracted per pound of steam to
the turbine is about 543 Btu per pound, the power requirement for this
recirculated steam pump is less than 0.6% of the electrical power produced.

The recirculation pumps which would be required for such a process
are similar in capacity and operating conditions to recirculation pumps
used in current steam plant design; the head requirements for these pumps
.are greater than the units 1n present use. The contact heat exchanger
which is proposed for the design would be quite similar to attemporators
used to "kill" the steam in the event of a turbine-trip in presently con-
structed steam plants.

The conclusions which are drawn from this discussion are:

a. If future study of possible molten salt reactor designs indicates
that possible formation of a solid phase in the secondary coolant system

will constitute a hazard in the reactor operation, then either the design

 
21

feedwater temperature must be above the liquidus or melting temperature
of the coolant (or within a limited temperature range below the liquidus
or melting temperature), or else the flow and temperature control systems
for both the secondary loop and the feedwater system must be extremely
reliable.

b. Within the range of possibilities just stated, it is most prob-
able that secondary coolants which have melting or liquidus temperatures
greater than 750°F will find their most economical applications in con-
junction with supercritical steam cycles.

c. Selection of lower feedwater temperatures to obtain higher
turbine plant efficiencies must be balanced against the increased capital
costs of the instrumentation and the control systems for the feedwater
system and the secondary salt system.

d. Higher feedwater temperatures can be obtained by several methods
that are available within the scope of present technology. One method,
which is available, is the recirculation of prime steam to a contact heat
exchanger; this method of increasing the feedwater temperature from 500
to TO0°F would result in less than a 0.6% reduction in electrical power

production from a plant with an overall efficiency of 40%.

 
22

6. EFFECTIVENESS OF COOLANTS IN TRITIUM HOLDUP

To be effective in preventing the tritium which is formed in the pri-
mary coolant loop from migrating by diffusion through the heat exchanger
surfaces first to the secondary salt and then to the steam or the reactor
cell, the secondary coolant must exhibit one of the following effective
mechanisms. The coolant may contain atoms which are exchanged with the
tritium atom, and these "captured" tritium atoms may then either be re-
moved as a chemical component or they may be stored in the coolant for
its lifetime. The coolant can also retain the tritium if it contains a
constituent which will chemically combine with the tritium. Again, the
tritium can be removed from the coolant by a chemical process which re-
moves and replaces the constituent; or if the constituent is present in
large concentration in the coolant, the inventory of tritium may be al-
lowed to build up over the lifetime of coolant use.

The atom which obviously will be most easily changed with tritium is
the normal hydrogen atom which is chemically identical. The only coolants
which have been considered in this review that contain hydrogen atoms are
NaOH, Hp, and Hs0. If any of these three choices were employed as the
secondary coolant, the concentration of hydrogen atoms in the secondary
coolant would be so great that the law of mass action would cause the
normal hydrogen atoms to diffuse out of the system in preference to the
tritium atoms. The more rapid diffusion of the hydrogen which was re-
placed by the tritium would actually be enhanced by the comparative
welght ratio.

Collecting the tritium in such a '"reservoir" as the secondary coolant
would have one major disadvantage. The molecules of coolant which con-
tained the tritium (such as NaOT and HT) cannot be separated chemically
from the molecules containing normal hydrogen. The secondary coolant
would therefore be always contaminated with the tritium until it was re-
moved by natural radioactive decay (12.3 year half-life) or until it was
removed by physical processing of the coolant. Physical processing of
NaOH would probably not be practical. Since NaOH is the hydrous form of
Nag0, it might be possible to liberate all of the tritium and hydrogen

through the application of heat and appropriate reducing agents. Physical

 
23

separation of tritium and hydrogen has been considered practical. The
mixture of tritium and hydrogen could be separated by low temperature
distillation, or they could be combined with oxygen to form a mixture of
light and very heavy water, and the separation might be made with existing
equipment used for the production of heavy water.

If the aspects of physical separation of the tritium from reservoirs
are neglected temporarily, the consequences of the handling and storage
of a large quantity of secondary coolant can be considered. After con-
tamination with the tritium, the entire inventory of coolant must be
contained and stored for an indefinite period. While the secondary cool-
ant is in the reactor system, storage presents no problem. If a reactor
is shut down after a planned 30-year lifetime, the cost of storage of the
coolant would probably not be prohibitive. In the case of using hydrogen
as a secondary coolant, it could be converted to water for storage. The
principal problem arises, however, if the coolant must be replaced one or
more times during the design lifetime of the reactor. Replacement could
be required, if, for example, the secondary coolant became contaminated
with fission products from the primary loop. If the tritium could not be
effectively removed from these replaced volumes of secondary coolant, the
cost of storing an undetermined number of these volumes could be prohibi-
tive from an economic standpoint.

The better alternative for tritium holdup in the secondary coolant
appears to be the addition of a limited amount of a constituent which
either contains hydrogen atoms which will be exchanged with the tritium
or a constituent which will chemically combine with the tritium and which
would be present in relatively minute quantities. This constituent should
be one which could be removed with relative ease by chemical processing.
It should also be chemically inert with respect to the major component or
components of the secondary coolant and to the materials of construction
of the system. It should be stable in the thermal and irradiation envi-
ronment of the secondary salt system.

Several constituents are available in the category of hydrogen-con-
taining compounds which may be added in small concentrations to the
choices of secondary coolants which are being considered. The most obvi-

ous choice is hydrogen gas. Neglecting the consideration of hydrogen gas

 
ol

itself as a secondary coclant, hydrogen gas may be added in small concen-
tration to any of the other choices of gaseous secondary coolants without
adverse effect; a possible exception is carbon dioxide. With mixtures of
hydrogen and carbon dioxide under high temperature and pressure, there
would be the formation of some carbonic acid gas and carbon monoxide that
might induce corrosion of the system piping.

A1l of the choices for molten salt or liguid metal coolants would
require an overpressure of a gaseous phase at some point in the secondary
system piping. This gas-liquid interface is usually maintained over the
liquid phase in the pump bowl of molten-salt systems; otherwise the inter-
face would be at the point at which the secondary system pressure was regu-
lated. Normally this gas phase will consist of an inert gas such as
helium or nitrogen plus a partial pressure of the secondary coolant or one
of its degradation products. For example, in the case of sodium fluoro-
borate, a substantial partial overpressure of boron trifluoride gas is
required to maintain the proper coolant composition.

For most of the choices of liguid metal and molten salt secondary
coolants, it would be possible to add hydrogen gas to the gaseous phase.
If equipment was incorporated in the secondary coolant loop so that
hydrogen in an inert carrier gas could be added and stripped from the
liquid phase, tritium in the coolant would then be carried with the hydro-
gen to the gas phase. The hydrogen and tritium mixture would be continu-
ally removed and replaced; this hydrogen could contain a high percentage
of the tritium which was transported to the secondary coolant if the
proper concentrations and contact areas were provided. Separation of the
hydrogen gas from an inert gas such as helium or nitrogen could be accom-
plished by the catalytic oxidation of hydrogen to water followed by its
removal in a cold trap or drying medium. Removal of hydrogen from some
of the other gaseous mixtures might prove more of a processing problem.

One of the major disadvantages in the use of overpressures of Hs is
the possibility of its forming an explosive mixture with atmospheric
oxygen in the event of inleakage to the reactor cell. The concentration
of hydrogen can be kept so low, however, that explosive limits can never
be reached, and inert gas blanketing of the reactor cell can exclude

atmospheric oxygen in all but the most catastrophic accident.

 
25

Another constituent which would be compatible with most of the
choices for molten salt, if not the liguid metal coolants, is the addition
of hydrogen fluoride or one of the other hydrogen halides. Use of any of
these additives would depend upon their compatibility with the materials
of construction for the secondary loop. If the halide is already a con-
stituent of the molten salt coolant, the addition and removal of the hy-
drogen halide should be relatively easy, and the addition of the hydrogen
halide should not affect the corrosion rate of the coolant.

A hydrogen-containing constituent which might be a natural contami-
nant in many coolants is water. In very small concentrations, the hydroxyl
ion forms a complex with some of the components of molten-salt mixtures.
FPor example, in the lithium fluoride-beryllium fluoride system, a complex
Be(OH) 1is formed. A hydrogenous form of sodium fluorcborate can be ob-
tained through a complex with the hydroxl ion in the form of NaBFSOH_°
Normal "drying" procedures can be used to remove the moisture during
chemical processing, and a slight amount of moisture could be allowed to
recontaminate the salt before it was returned to the system.

Several of the choices for molten salt coolants present the possi-
bility of maintaining a small amount of the acid represented in the salt.
For example, small amounts of nitric acid could be maintained in the
nitrate salts without affecting their vapor pressure substantially or
their corrosion performance. Carbonic acid is not sufficiently stable
to be maintained in the carbonate mixtures, but it is possible that a
sufficient concentration of the bicarbonate ion HCOS‘ could be maintained
to act as a tritium trap.

In the category of liquid metal coolants, particularly the alkali
metals, a small concentration of the hydroxide would supply the hydrogen
atoms which could be interchanged with the tritium. The metal hydrides
constitute a group of hydrogen-containing compounds which might feasibly
be added to either the liquid metals or the molten salts. Also the
ammonium salt is a feasible constituent for use with some of the choices
for molten-salt coolants.

Tritium can be retained in the secondary coolant through the addition
of a constituent (probably in small concentration) with the oxidation

potential to convert the tritiumto its ionic state and maintain it in that

 
26

state. Oxidation in this concept refers to the process of stripping an
electron from the tritium atom and producing in the coolant the posi-
tively charged tritium ion which may then associate with negative ionic
species in the coolant. The added constituent must, therefore, have a
reduction potential--that is, an affinity for electrons--which is greater
than tritium (hydrogen). In the normal chemical processing of the second-
ary coolant, the oxidation-reduction potential of the coolant can be ad-
justed, the tritium will be released in the diatomic state and can be
stripped from the coolant with an inert gas stream.

Tritium which diffuses through the heat transfer surface of the pri-
mary heat exchanger will exist either as the dimer or a pseudo-dimer
through association with the electrons in the outer shells of the atoms
of the wall structure. Oxidation of the dimer structure in solution will
result in the association of the ionic component with anionic constituents
in the coolant to prevent further diffusion of the tritium. Tritium which
is formed by nucleonic reactions within the secondary coolant will prob-
ably first exist in the ionic state; however, it will be reduced by the
constituents of the coolant or containment material which are above
tritium (hydrogen) in the electromotive series. These components would
include lithium, potassium, sodium, magnesium, beryllium, aluminum, man-
ganese, chromium, iron, cobalt, nickel, tin, and lead, as examples.

The added oxidizing constituent, which is reduced in the process,
must have sufficient potential to convert the tritium to its ionic state,
and, at the same time, it must not convert the materials of construction
of the secondary system to ionic species. A limited number of "getters"
are available for the various choices of secondary coolants and the accom-
panying choices of materials of construction.

For example, C. F. Weaver indicated in a letter dated March 1, 1971,
that in a fluoride-containing molten-salt coolant, if it was desired to
convert To to TF in a system with nickel containment, it would be neces-
sary to have an oxidizing agent with a free energy of formation between
—65 kcal/gr-atom of fluorine and —62 kcal/gr-atom of fluorine. The first
value corresponds to the free energy of formation of TF at 1000°K, and
the second value corresponds to the free energy of formation of NiFs.

The only agent which meets this criterion is germanium which is, at best,

 
27

a borderline possibility. If the containment material is molybdenum, then
germanium is an acceptable material. If the system is lined with copper,
then possible agents are germanium, molybdenum, arsenic, or antimony; the
last 1s again a borderline choice. All of these oxidizing agents except
arsenic and antimony would exist in the molten-salt mixture in the tetra-
valent state; arsenic and antimony would be pentavalent. All would be
oxidized to the trivalent state in the reduction of tritium.

In the primary salt, it is possible to retain tritium through the
introduction of either uranium or thorium at a reduced valence to serve

as an oxidizing agent.
28

7. COMPATIBILITY OF SECONDARY COOLANTS WITH THE
PRIMARY SALT, STEAM, AND THE CELL ENVIRONMENT

The compatibility of the various choices for secondary coolants with
the primary system fluid, with the water in the steam generator, and with
the cell environment must be considered. It will be assumed that the pri-
mary system fluid is a eutectic mixture of lithium and beryllium fluorides
Li-BeF, and relatively small percentages (~10%) of thorium and uranium
fluorides. The cell will be assumed to be filled with an inert gas such
as helium or nitrogen or a mixture of carbon dioxide and nitrogen formed
by the combustion of hydrocarbons with air followed by scrubbing and

drying.

7.1 Compatibility With the Primary Fluid

 

There are two major points of concern if leakage in the primary heat
exchanger allows the secondary coolant to mix with the fluid in the pri-
mary loop. First, there is the concern that an exothermic reaction be-
tween the two fluids would propagate the leakage and mixing rate so that
ultimately either or both of the two systems would be ruptured. Second,
there is the concern that the addition of limited amounts of secondary
coolant to the primary system would result in a concentration of the
fissile material and produce a critical mass outside of the core region
or excessive reactivity within the core region. Fissile material could
be concentrated either by the precipitation of the material from the pri-
mary fluid (such as the precipitation of uranium oxide) or preferential
absorption of the fissile material in the small volume secondary coolant.

Of secondary importance is a consideration of the possibility of an
insoluble precipitate formed within the primary system which would prevent
further attempts to circulate the primary fluid through that portion of
the system, or the possibility that contamination of the primary fluid by
the secondary coolant would adversely affect the nuclear characteristics
of the fluid. In the first instance, a portion of the primary system
piping and associated components would have to be replaced, and in the

second instance, either the expensive primary fluid would have to be
29

chemically processed or replaced, or the adverse characteristics would

have to be tolerated in the future operation of the reactor.

7.1.1 With the molten salts as secondary coolants

 

7.1.1.1 Sodium fluoroborate. The sodium fluorcoborate (NaBFs,-NaF

 

appears to be the choice at present for a secondary coolant for molten-
salt reactors with LioBeF, plus fluorides of uranium and thorium as the
primary salt. Some experimental study has been given to the possible mix-
ing of these two salts; however, more detailed qualitative analysis of

the dynamic characteristics of the interaction are required. No violent
exothermic reactions occur when fluoroborates are mixed with steam or with
fluoride fuel salts. 1In fact, it has been discovered that fluoroborates
are immiscible with molten mixtures of lithium and beryllium fluorides
over a significant range of conditions. Uranium and other tri- and tetra-
valent elements are not extracted into fluoroborates, and the only high-
melting compound that might be formed is sodium fluoride. There is some
migration of LiF to the fluorcborate phase, and replacement of NaF by

LiF in the NaBFs complex results in an almost immediate release of BFs
gas. The detailed consequences of mixing of the fuel salt with the sodium
fluoroborate depend on (a) the rate of mixing of the two streams, (b) the
solubility of the BFs gas in the resultant phase, (c) the relative tem-
peratures of the two fluids, (d) the kinetics of reactions between the
distributed components, and (e) which fluid constitutes the continuous
phase. FPresent evaluation of all of these factors indicates that there

is no mechanism for the concentration of uranium to produce a critical
configuration and no compounds will be formed which cannot be redissolved
through the addition of appropriate chemical agents.

Operation of a test loop (containing residues of this fluoride salt )
with a flushing charge of NaF-NaBF, did, however, reveal the deposition of
a green salt in the upper region of the pump bowl. The composition of the
salt was essentially 7NaF-6(Th,U)Fs, suggesting that either entrainment of
the residue or solution-deposition of it had occurred, along with some
replacement of Li by Na; although more study of this phenomenon is indi-
cated, there is no information available to cause alarm over the possi-

bility of accidental mixing of fluoroborates with fluoride salts. For

 
30

MSBR use, moreover, the accidental introduction of fluoroborates into the
circulating fuel would cause a large reactivity decrease because of the
boron, and thus even a small leak would be quickly detected. The boron

could be easily removed from the fuel salt by treatment with HF.

7.1.1.2 Lithium fluoride—beryllium fluoride. Lithium fluoride—

 

beryllium fluoride mixtures (LoB) would be compatible with the primary
salt since it is the major component of the presently proposed fuel salt.
Introduction of the secondary salt into the primary system would have the
effect of reducing the fissile concentration in the primary salt. The
major adverse consequence would be the introduction of lithium—6 into the
primary salt in the event that the secondary salt did not contain lithium
which had been isotopically enriched in lithium—7. The result of such a
mixing would either require that the primary fuel salt be replaced or
that uranium concentration of this salt would be increased and the
lithium—6 which was introduced be allowed to "burn" out. The latter
course would probably be more favorable if the amount of lithium—6 intro-
duced were small; if the amount was large, then a detailed economic con-
sideration would have to be made to determine the course of action.

The use of I,B which was enriched in lithium—7 as the secondary cool-
ant would eliminate any mixing probléms. This would increase the inven-
tory charge for the secondary salt as indicated in Table 5. I-B as a
secondary coolant has the disadvantage of having a relatively high
liquidus temperature as indicated in Table 3. This liquidus temperature
can be lowered by adding higher concentrations of beryllium; however, this

addition results in a much higher viscosity for the mixture.

7.1.1.3 FLINAK and the "low melting point" fluoride. The comments

 

which might be made concerning FLINAK are essentially the same as those
for LoB. It has the compound LiF in common with the primary coolant, and
NaF and KF are analogous compounds. Mixing with the primary coolant

would probably produce no reaction, but the secondary coolant would have
to contain lithium enriched in “Li to avoid problems with the primary salt
nuclear characteristics following a mixing accident. These same state-

ments apply to the "low melting point" fluoride mixture.

 
31

7.1.1.4 The chloride salts. The problems resulting from the use

 

of a mixture of lithium and potassium chlorides as a secondary coolant
following a mixing accident would differ from those encountered with LB
and FLINAK only in that the chloride ion would be present. Following a
mixing accident, the chloride ion could be removed from the primary salt
through the use of HF gas. The presence of the chloride ion might intro-
duce some corrosion problems if trace quantities were left in the primary
loop; however, most of the proposed materials of construction for this
loop are resistant to this attack, and, if the chloride salt were used

as the secondary coolant, the materials of construction in the secondary

loop should also be resistant.

7.1.1.5 Sodium hydroxide, the nitrates, and the carbonates. The

 

compatibility of the primary salt with NaOH, the nitrates, and carbonates
may be considered as a group principally because they are all oxygen-
containing compounds. They differ to some degree in the availability of
the oxygen for chemical reaction--or, more exactly, their free energy of
formation with respect to the compounds in the primary salt. In con-
sidering the consequences of these compounds mixing with the primary
salt, it is necessary not only to consider the reaction products which
result from the combination of the chemical components, but it is also
necessary to consider both the mixing rates and reaction rates.

For example, the mixture of any of these secondary coolants will
result in the formation of both uranium and thorium oxides which are
essentially insoluble in either of the molten mixtures. If this mixing
occurs as the result of a small crack in the tubing of the primary heat
exchanger, then the formation of the oxides at the point of the mixing
may form a "scab" which results in a type of "self-healing" reaction to
small leaks between the systems.

On the other hand, if massive mixing occurs as a result of the double-
ended rupture of a tube in the primary heat exchanger, a large amount of
oxide precipitate will be formed. Depending upon the relative fluid
velocities, this precipitate may either be carried around the coolant loop,
or it may precipitate in the place where it is formed, or it may precipi-

tate in some section of the system where the fluid velocities are low.

 
32

If the precipitate remains in suspension in the loop, it is possible that
it can be collected and chemically reconverted to a soluble fluoride. If
it precipitated at the point of formation, say in the primary heat ex-
changer, to the extent that flow through that part of the system is
blocked, then the resolution of the uranium and thorium would be practi-
cally impossible. The melting point of the uranium and thorium oxides are
well above (~4500°F) the temperatures that the materials of construction
can contain, and adequate agitation for effective chemical resolution
would be difficult. This same problem would exist if precipitation of the
oxides occurred at any other inaccessible point in the primary system. It
might be feasible, if this precipitation and flow blockage did occur, to
remove the system component which was involved (such as a heat exchanger)
and replace it with another unit.

A fortunate circumstance is that the conditions which cause the pre-
cipitation of the oxides of uranium will also cause the oxides of thorium
to precipitate. Therefore, the formation of a critical configuration out-
side of the core is not probable.

Mixing of the secondary coolants containing carbonates, nitrates, or
hydroxides with the primary salt will result in the formation of some
gaseous components. The relative amount of gas evolved and the relative
evolution rate will decrease in the order just listed. In the first case,
the gaseous product will be carbon dioxide, in the second it will be
oxides of nitrogen, and in the last, it will be hydrogen fluoride.

Another potential problem for a violent chemical reaction exists if
the nitrates penetrate the primary circulating system to the point of
entering the graphite moderator lattice. This is not a probable situation
since this coolant mixture is less dense than the fuel salt; it will be
buoyed out of the core region as long as the fuel salt is present. If the
fuel salt has been dumped into the drain system, any inleakage of this
secondary salt should be carried with it. Finally, a rapid reaction be-
tween the graphite and the molten nitrate salts occurs only over a par-
ticular range of relative compositions and within a range of interactive
surface to volume ratios. If the nitrates are to be considered in detail,
an experimental investigation will have to be made to determine the limit

of these ranges.

 
 

33

7.1.2 With liquid metals as secondary coolants

 

7.1.2.1 Sodium and Nak. The introduction of either sodium or NaK,
when used as secondary coolants, into the primary loop will produce similar
results. The only difference is that the higher reactivity of the potas-
sium in the NaK may produce slightly higher reaction rates. 1In either
case, all of the metallic species in the primary salt, with the exception
of lithium, will tend to be reduced to their elemental forms by the intro-
duction of the alkali metals.

The uranium metal will be formed most easily and present indications
are that it will tend to agglomerate or sinter into particulate size.

Some thorium metal may also be formed with the uranium, but this will de-
pend upon the relative concentrations of these two species at the reaction
site. Beryllium will not begin to be produced in metallic form until all
of the uranium and thorium have been depleted from the primary salt at the
point of the reaction.

The consequence of these events is that a critical mass of uranium
can be collected outside of the core since the total inventory of fissile
material in a lOOO-MN(e) reactor approaches 1500 kg, and a critical mass
under partially reflected conditions may be on the order of 2 kg. If the
uranium does not collect in the system, then the possibility still exists
of introducing a surge of uranium into the core which may produce a sharp
reactivity increase. On the positive side of the picture, a small leak
would begin to remove uranium from the primary salt and cause a detectable
change in the reactor operaticnal mode.

Of the reactions which have been discussed thus far between the pri-
mary salt and possible secondary coolants, the reaction of elemental
sodium or potassium to produce metallic thorium, uranium, and/or beryllium
are the only ones which will liberate significant amounts of energy. The
maximum energy release will be less than 1 kcal per gram of sodium or
potassium introduced into the system. Because of the relatively low con-
centration of uranium and thorium, which must be reduced first, there will
not be a concentrated site for the energy release unless there is violent
agitation at the point of mixing. Also, because of the relatively high

thermal conductivity of the liquid metal, concentrated energy release
 

3k

sites will have their effect reduced if the liquid metal phase is con-
tinuous. The conclusion is that the potential of depositing large amounts

of energy at a local site in the mixed coolants is relatively low.

f.1.2.2 Bismuth and lead. Bismuth and lead, or a mixture of the

 

two, can be allowed to contact the primary salt with relatively little
effect as long as they do not contain any active metallic constituents
which might result from erosion of the containment material of the secon-
dary loop. Any metallic impurity which is in the elemental state will

have the tendency of reducing the thorium and uranium of the primary salt
to their elemental states. The thorium and uranium would then be extracted
in the bismuth or lead phase which would be essentially immiscible with

the primary salt. This effect can be eliminated by continuously pro-
cessing the coolant to restrict the concentration of metallic corrosion

products.

7.1.2.3 Mercury. The same statements which were made for lead and
bismuth also apply to mercury. The major difference in the response of
the primary system following the introduction of mercury would be due to
the higher vapor pressure of the mercury at the system operation tempera-
ture. The normal boiling temperature of mercury, where the vapor pressure
is equal to one atmosphere, is 6T4°F (357°C). At 600°C (1112°F) which is
about the mean temperature of the primary system, the vapor pressure of

mercury is almost 23 atmospheres (336 psia).

7T.1.3 With gases as secondary coolants

 

The problem encountered with the use of any of the gases as secondary
coolants for molten-salt reactors is the same as that which was just cited
for the use of mercury, the introduction of &a pressure surge to the pri-
mary system. In Table 2, the physical properties of the gases are listed
at a high pressure (1000 psia) and a lower pressure (250 psia). It is
obvious that the high-pressure gas is a more efficient coolant; however,
the consequences of a system at 1000 psia leaking to the primary system
are obviously worse than the consequences which might te encountered with

a system at 250 psia. It 1s obvious that the primary loop must be designed
35

so that the admission of the higher pressure of the secondary loop will
not result in major damage to the primary loop.

Several design alternatives have been discussed which would allow
the use of a high-pressure system in the secondary loop. First, the
higher pressure could be used as the design pressure for the primary loop.
This would result in a system which would be considerably more expensive
to fabricate; in addition to thicker wall requirements for a system with
relatively expensive piping material, the seals for pumps and valves
would be complicated. Second, the primary system could be equipped with
rupture disks which would allow the fuel salt to "blow off" into the
drain tank in the event of an over pressure. The problem with this design
alternative is that the inertial effects of the fuel salt might cause the
primary system to fail before the pressure wave reached the rupture disk.
The third alternative would be to pressurize the entire cell which con-
tained the primary and secondary loops ahd then design the piping and com-
ponents to withstand the pressure differential. This would mean that the
reactor would have to be scrammed and the system depressurized in the
event of a drop in pressure in the containment cell. A further complica-
tion 1s that there would be the tendency for leakage outward of the con-
talnment cell rather than inward, a more desirable containment situation.
A pressurized containment cell, however, appears to be the best solution

of the three which are cited.

7.1.3.1 Helium. If the problems associated with the higher second-
ary loop pressure are bypassed, the following comments can be made con-
cerning the use of the various gases as secondary coolants. Helium
appears to be the best choice of those gases listed in Table 1. It is
chemically inert, and, therefore, there will be no chemical reaction with
the primary salt. With the exception of hydrogen, it has superior heat
transfer performance; it transports more heat per unit weight of gas

circulated.

7.1.3.2 Hydrogen. Hydrogen does have the best heat transport per-
formance characteristics, and it is readily available. It does present

an explosion hazard when mixed with atmospheric oxygen over a large range

 
36

of concentrations. This would probably eliminate it for consideration

as a component in a nuclear plant.

7.1.3.3 Water vapor. Water vapor which is the least expensive
choice for a secondary coolant presents about the same problems when
mixed with the primary salt as the oxygen-containing molten salts which
were considered previously (hydroxides, nitrates, and carbonates ). Addi-
tion of water vapor to the primary salt would result in the precipitation
of the oxides of uranium and thorium.

7.1.3.% Air. The oxygen in "synthetic air" used as a secondary
coolant would not be as available for formation of the oxides of uranium
and thorium; however, this process would still exist. Because of the
similarity in heat transport performance, there would be no advantage in

using mixtures of oxygen and nitrogen in preference to pure nitrogen.

7.1.3.5 Nitrogen. Nitrogen is another inert gas with respect to
chemical reactions with the primary salt. It is less expensive than

helium, but it is much less efficient as a heat transport fluid.

7.1.3.6 Carbon dioxide. The use of carbon dioxide as a secondary
coolant would result in the same chemical reactions with the primary salt

as the carbonate molten salts.

7.1.3.7 Conclusions. The conclusion that might be drawn from these
statements about the gaseous choices for secondary coolants is that helium
appears to be the best choice as long as it is available and not prohibi-
tively expensive. From avallability studies which are made within the
gas-cooled reactor program, it appears that a supply of helium from natural
gas deposits will be available in sufficient quantity to support a rela-
tively large reactor power industry. At the time that this supply becomes
exhausted, it will be possible to extract helium from the atmosphere at a

cost which is not prohibitive.

 
37

7.2 Compatibility With Steam

 

The major points of concern in considering the compatibility of the
various choices for secondary coolants with water and steam in the event
of a leak or rupture in the steam generator turbine are (a) violent reac-
tions, (b) the creation of extremely corrosive reaction products, or (c)
the production of reaction products which may plug either the coolant
system or the steam system and make subsequent repair difficult. Previous
work* has indicated that the sodium fluoroborate, in particular, and the
fluorides, in general, do not undergo violent reactions on mixing with
water. DNone of the other choices of molten salts for secondary coclants
undergo significant exothermic chemical reactions when mixed with water.
If these salts are cooled below their liquidus temperature by the feed-
water stream, a solid phase will be formed; however, they can all be
melted under normal system operating temperatures and most are soluble
in water so that they can be washed from the steam system.

In the category of liquid metals, both sodium and potassium will
undergo a rapid reaction with water with the significant evolution of
heat and also hydrogen gas. The hydrogen which is evolved results in an
explosive hazard if an inert atmosphere is not maintained over the system.

No problem is encountered if any of the possible gaseous secondary

coolants are allowed to mix with steam or water.

7.3 Compatibility With the Cell Environment

 

The cell environment will most likely be a blanket of inert gas.
This could be nitrogen, which is obtained from an air liquification pro-
cess, or it could be the combustion products from the careful combustion
of natural gas in air followed by scrubbing and drying. In either case,
there should be no possibilities of chemical reaction or unusual corrosion
as long as moisture is excluded from the cell. A simultaneous rupture of

the steam system and the secondary coolant system could produce reactions

 

¥, F. McDuffie et al., "Assessment of Molten Salts as Intermediate
Coolants for IMFBR's,'" USAEC Report ORNL-TM-2696, Oak Ridge National
Laboratory, Sept. 3, 1969.

 
38

outside of the system piping; the events in this case would be similar to
those discussed with reference to the compatibility of the secondary cool-
ants with water and steam.

The exception to the preceding statements is hydrogen. If it is used
as a secondary coolant and a situation develops in which it is introduced
into the cell, then it must be considered that the same situation might
either introduce ambient oxygen to the cell or allow further leakage of
the hydrogen to the ambient. In either case, the development of an

explosion hazard is possible.

 
 

39

8. MATERIALS OF CONTAINMENT

A primary consideration in the selection of a secondary coolant for
a molten-salt reactor system is the choice of materials from which the
piping and components of the secondary system may be fabricated. Three
areas must be considered. One area includes the piping and components
which are outside of both the primary and secondary heat exchangers; in
this area any material which is compatible with the secondary coolant at
its operating temperature and pressure will be acceptable. The second
area 1s the tubing and headers of the primary heat exchanger where the
materials of construction are in contact with both the primary fuel salt
and the secondary coolant, and the third area of consideration is the
tubing and headers of the secondary heat exchanger or steam generator
where the material is in contact with both the secondary salt and feed-
water and steam.

If a different material of construction is selected for any of these
three areas, then consideration must be given to any problems which could
result at the point where the two different materials are joined in the
secondary system.

In the following paragraphs, it will be assumed that the materials
of construction are operating nominally in the range from 1000 to 1300°F
when the strength of the material is considered. In the steam generator,
additional considerations must be given to the performance of the mate-
rial in contact with feedwater in the range of LOO°F unless the steam
system is designed to preheat the inlet stream to temperatures in the
range of TOO°F as discussed in the previous section on the significance

of the liquidus temperature of the coolant.

8.1 Containment Materials for the Various
Secondary Coolant Systems

 

When those sections of the secondary coolant systems outside of the
primary and secondary heat exchangers are considered, only the resistance
of the material of construction by the secondary coolant must be con-

sidered; it can be assumed that the environment outside of the piping

 
 

Lo

and components can be kept sufficiently non-corrosive so that it need not
be considered. 1In addition to the resistance to corrosion by the particu-
lar choice for a secondary coolant, the strength of the material at the
operating temperature and pressure, the applicability of the material to
the fabrication of the piping and components, and the cost of the material

must be considered.

8.1.1 Materials of construction for molten salts as
secondary coolants

 

 

8.1.1.1 Fluoride salts. Molten fluorides are excellent fluxing
agents for the removal of oxide films. In the fluoride salt mixtures, the
corrosion resistance of a metal or alloy cannot be maintained by a protec-
tive oxide film. Corrosion is controlled by the thermodynamic driving
forces of the reactions involved. If the bulk of the metal fluorides in
the coolant are more stable than the metal fluorides of the structural
metal, there is little tendency for corrosion to occur. The relative ease
of fluoride formation by alloying constituents has been related to the
free energy of formation of the respective fluorides and increases in the
following order: tungsten, molybdenum, niobium, nickel, iron, chromium,
vanadium, zirconium, titanium, and aluminum. Results of corrosion studies
have indicated that corrosion in molten fluoride systems involves the at-
tainment of thermodynamic equilibrium between the fluoride melt and the
container metal and that the chemical activities of the alloying con-
stituents can be approximated by their atom fractions.*

In the more common high-temperature structural alloys, chromium has
the greatest tendency to form a fluoride, because of the higher free energy
of formation of CrFp, which will go into solution in the fluoride salts.
The presence of any oxidizing constituents in the molten salt will tend to
promote this reaction until the buildup of the reduction products tends to
stifle the formation of more '"noble" metal fluorides such as FeF- because
of the mass action effect. Thus, in chromium-containing base-metal alloys,
corrosion, for the most part, is selective to chromium. TImpurities in the
melt, such as iron or nickel, increase the corrosion of the alloy due to

the reduction of these fluorides by chromium in the alloy.

 

¥Chapter VII, Corrosion in Fluid Fuel Reactors (page 46).
 

L1

FeFg + Cr 2 CrFo + Fe

The dissolution of oxide films from the metal surface can also cause selec-

tive removal of chromium by the following mechanism,

+++ ++
2TFe + Cr 22Fe + 3Cr

where the original iron existed as the oxide in the film.

Nickel-base alloys generally have excellent compatibility with molten
fluoride salts, since NiF; has a relatively low free energy of formation.
As a consequence, the Hastelloy series of alloys has shown very good resis-
tance to corrosion by molten fluoride salts. Particularly good perfor-
mance was obtained from Hastelloy-B and C, and in particular, Hastelloy-N
(INOR-8) which was used as the material of construction in the Molten Salt
Reactor Experiment. Some attack has been indicated at temperatures above
700°C (1292°F).

Steel is not generally resistant to molten fluoride salts at high
temperatures. Static studies at 700°C revealed very rapid attack within
600 hours on most of the common types of stainless steels by a mixture of
lithium and uranium fluoride. Long-term tests with type 304L stainless
steel, however, revealesd a corrosion rate of about 2 mpy after exposures
of LU 000 hours at a maximum temperature of 675°C by a mixture of lithium
and beryllium fuel salts.

Molybdenum and niobium appear to be resistant to molten fluoride
salts at temperatures to 816°C (1500°F). Vanadium and tantalum were
attached to a depth of 1 mil in 100 hours at these temperatures. At T700°C
gold and rhodium resisted attack by a mixture of sodium and uranium fluo-
rides; platinum was slightly attacked, and copper was rapidly attacked.
Other studies indicated that copper possesses excellent compatibility with
fluoride salts.

For the containment of the sodium fluoroborate as a secondary coolant,
Hastelloy-N is the most favored material of construction, although Inconel
is a possibility. The present cost of Hastelloy-N, however, is about
$20 per pound. Inconel 600 has performed reasonably well as a pump impel-
ler even under cavitation conditions and exposure to mixtures of sodium,

zirconium, and uranium fluorides (50-46-4) at temperatures in the range of

 
Lo

600 to 760°C. Its long-term use as reactor system piping, however, is
still questionable. Alloys of nickel and molybdenum would probvably pro-
vide satisfactory service, and nickel would be satisfactory for tempera-
tures less than 1000°F. A duplex pipe of nickel inside of a stronger ma-
terial would probably be satisfactory and more economical.

For L-B in the absence of the fluorides of uranium and thorium which
are found in the fuel salts, experiments have indicated that satisfactory
materials of construction include Inconel, Ni-Mo alloys, Hastelloy-N, and
304 SS.

Based on the previous considerations, FLINAK could probably be con-
tained in Ni—Mo alloys, Ni for temperatures less than 1000°F, and
Hastelloy-N. Nickel-lined duplex piping is also a possibility.

The "low melting point" fluoride mixture which differs from the I-B
only in that NaF is substituted for some of the LiF could be contained in
the same materials that have been specified for the I-B. With this salt,
and for some of the other fluoride salts, there is the possibility that
some of the stainless steel can be used in the system piping and components
if the coolant can be kept free of oxidizing agents. Leakage of the pri-
mary salt into a secondary system which contained fluoride salts as second-
ary coolants would introduce (along with the radioactive fission products
which are undesirable in the secondary loop) the fluorides of uranium and
thorium which will result in the oxidation of both the chromium and iron
in the stainless steels. These mechanisms will be discussed in the follow-
ing section on compatibility of the materials with the secondary salt in

conjunction with the primary salt.

8.1.1.2 Chloride salts. As with fluorides, in reactions between
chlorides and free elements of other elements at high temperatures, there
is a tendency for the chloride which has the higher heat of formation per
equivalent of chloride to be formed. Thus, zirconium, titanium, manganese,
zinc, and thallium react with the chlorides of a large number of other
elements whose chlorides have a lesser heat of formation. Conversely,
those chlorides which have a low heat of formation per equivalent of

chloride tend to be reduced to the free metallic element.

 
43

Also, as with fluorides, the presence of air over a chloride melt
causes rapid corrosion of metals such as titanium, zirconium, and stain-
less steel. The presence of small amounts of water in a chloride mix can
cause failure of stainless steels due to intergranular stress corrosion
cracking.

As with the fluorides, Hastelloy-N appears to be a satisfactory mate-
rial of construction for loop components and piping which is in contact
with a molten chloride mix; this, however, is an expensive material of
construction. Again, it is possible that nickel with a small percentage
of alloying constituents might be used where temperatures are less than
1000°F, and a duplex tube with a nickel liner can be used at higher tem-
peratures. Type 347 stainless steel might be used if the salt is main-
tained free of oxidizing agents and moisture; use of this material, how-

ever, is questionable.

8.1.1.3 Sodium hydroxide. The reaction between NaOH and various

 

metals may be characterized by the following equation:
2 NaOH + Me -» NasQ + MeO (or NaMeO) + Ho

where "Me" indicates the metal, '"MeO" is the metal oxide, and "NaMeO"
represents the sodium metallite. IFach metal tends to form its most stable
or common oxide; the various metals differ only in the rate of reaction
and their valence state. In general, but not exclusively, those metals
which form stabie oxides are least corrosion-resistant while those which
form unstable oxides are most resistant. Thus, iron, chromium, beryllium,
titanium, manganese, and iron-base alloys that form stable oxides are not
corrosion-resistant and corrode according to the previous equation.

On the other hand, nickel, copper, gold, Monel 400, and probably
silver and cobalt that form unstable oxides also exhibit one or more of

the following secondary reactions:
MeO + Hy; — Me + H:0
Nas0 + 1/2 Hy — NaCH + Na

Hz0 + Na — NaOH + 1/2 Hy

 
L

The first two reactions are reversible for nickel, copper, and probably
cobalt and only these metals or some of their alloys attain a definite
equilibria (and low corrosion rates) because the sum of these three reac-
tions is the reverse of the reaction given at the first of this section.
Only these metals are protected from corrosion by a hydrogen partial pres-
sure above the molten NaOH.

Active metals such as zirconium may generate sodium more rapidly {(by
the middle equation asbove) than it can be oxidized (by the last of the
reactions shown above). Sodium hydroxide is also reported to be reduced
to sodium or sodium hydride by magnesium, chromium, tungsten, iron, cobalt,
or nickel. With gold and silver, the reaction forming NaOH and Na (middle,
above ) is not reversible because an alloy is formed with the metallic
sodium and the reaction goes to completion.

All of the above metals can react further if the hydrogen which is

produced is removed from the system as follows:
Nag0 + Me - MeO + 2 Na

This reaction will go to completion if the sodium is continuously removed
or is alloyed.

These reactions indicate that hydrogen is beneficial as a cover gas
with nickel, copper, and cobalt alloys. Conversely, an oXygen cover gas
increases the oxidation rate of the metals in the molten hydroxide.

The position of a metal in an electrochemical order in NaOH does not
appear to give an indication of its relative corrosion behavior. The
reported electrochemical order for metals in NaOH at 340°C (6LL4°F) is W,
Mo, Cd, Pb, Cu, Ag, Au, Pt, Fe, Ni, Ta, Nb, and Zr.

It appears that there are really no good materials for the contain-
ment of NaOH under the conditions which will be encountered in the second-
ary loop of a molten-salt reactor system. Nickel, which is used as a ma-
terial of construction for evaporators used to concentrate commercial NaOH,
1s not satisfactory from structural considerations at temperature approach-
ing or above 1000°F. It is possible that a nickel-base alloy with an
agent such as 15% molybdenum could be used; however, study supporting this
statement is not complete. Monel might be an acceptable metal at lower

temperatures also. For any of the choices, the maintenance of a hydrogen

 
L5

pressure over the NaOH will be helpful in limiting corrosion. ILoop tests
have also shown that corrosion rates increase in NaQOH systems both with

increasing temperature and with increasing fluid velocities.

8.1.1.4 Nitrates. The nitrates in the form of a mixture of 6.9%
NaNOs, 48.5% NaNOs, and L44.6% KNOs have been used for over 20 years as a
heat -transfer salt. It is marketed by the DuPont Chemical Corporation
by the trade name of Hitec, and it 1s sometimes designated as HIS for
High-Temperature Salt. Early literature indicated that it decomposed at
temperatures above 538°C (1000°F); however, more recent studies have indi-
cated that, with proper overpressures and careful exclusion of contami-
nants, particularly water, stability could be maintained at temperatures
significantly above 1100°F.

Operation of these systems has indicated that, at temperatures below
1000°F, most common materials of reactor construction can satisfactorily
contain the nitrate mixture; these materials include the stainless steels,
Inconel, and the high-nickel base alloys. The latter two should also be
satisfactory up to temperatures of 1200°F. No studies have been reported
for systems with velocities as high as those proposed in present secondary

reactor systems.

8.1.1.5 Carbonates. Low-alloy steels and stainless steels, as well
as the nickel-base alloys are compatible with carbonate systems up to tem-
peratures of 1200°F. At temperatures significantly above this, 800°C
(lh72°F), most of the materials of construction are attacked quite rapidly
by the carbonates. In the range of temperatures from 800 to 920°C, only
gold appears to have significant resistance to corrosion by the carbonate

mixture.

8.1.2 Materials of construction for liguid metals as secondary coolants

 

One particular disadvantage in the design of heat transfer systems
which employ liquid metal, particularly sodium and potassium, is the de-
velopment of excessive thermal stresses in the materials of construction.
This aspect of system design is a result of the very high film heat trans-
fer coefficient which is associated with the liguid metal; this same fea-

ture for liquid metals also contributes to their favorable performance in

 
L6

some parts of the system. The problem develops when it is necessary to
transfer heat to a fluid whose temperature differs considerably from that
of the liquid metal. Specifically, in the secondary loop, this problem
would develop in the steam generator. Because of the high heat transfer
coefficient to the liquid metal, the temperature of the tube walls of the
steam generator will remain very close to the temperature of the liguid
metal. If these same tube walls are contacted by water at the usual feed-
water temperatures (about L4O0°F), then excessive thermal stresses are cre-
ated over the pipe walls and in the adjacent tube sheets.

Several design alternatives are available to avoid the development of
excessive stresses in the steam generator. One of the more obvious alter-
natives is the same method which is used to eliminate the possibility of
freezing molten-salt secondary coolants on these same tube walls; this
method is to bring the feedwater to a temperature before it enters the
steam generator that is high enough to prevent excessive thermal stresses

in the tubing.

8.1.2.1 Sodium. Considerable experience has been obtained in circu-
lating sodium in heat transfer loops. For those sections of the loop out-
side of the primary and secondary heat exchangers, the materials of con-
struction may be the chrom-moly steels up to temperatures of 1200°F, the

18-8 stainless steels, Incoloy-800, or Hastelloy-N.

8.1.2.2 Sodium—potassium (N2K). Due to the similarity of the cool-

 

ants, the same materlals of construction may be used when the secondary

coolant is NakK.

8.1.2.3 Bismuth. Bismuth is quite corrosive to most metals and
alloys, but its éorrosivity can be reduced (particularly with respect to
steel) by the addition to the bismuth of zirconium or titanium in conjunc-
tion with magnesium. The zirconium (or titanium) is believed to react
with the nitrogen and/or carbon in the steel to form a protective layer
of ZrN or ZrC which provides a barrier between the bismuth and the ferrous
alloy substrate. The role of magnesium in conjunctior with zirconium or
titanium is that of a getter for the oxygen from the system; thus it pre-
vents any oxidation of the zirconium or titanium thereby destroying their

effectiveness.

 
b

8.1.2.4 Bismuth—lead. Essentially the same materials of construc-
tion that were cited for bismuth are permissible materials of construction

with bismuth—-lead mixtures.

8.1.2.5 Lead. Most of the materials of construction are applicable
for the containment of lead that were listed as materials for use with
molten bismuth. Of the ferrous alloys, the stainless steels are the least
resistant to attack due to the high solubility of nickel in lead and the
higher solubility of chromium (than iron) in lead. Inhibition of lead
with zirconium and magnesium additions is very effective in reducing
corrosion attack. A Croloy 2 l/h loop showed no attack after 27,765 hr
(>3 years) operation in lead containing 250 ppm zirconium plus magnesium
at a maximum temperature of 550°C (1020°F).

The refractory metal and metal alloys are resistant to attack by lead
up to temperatures of 1400°F. Unalloyed niobium and niobium with 1 w/o
zirconium showed no attack in one year of operation. Molybdenum is re-
sistant at the operating temperature. A small amount of dissolved oxygen
(6 ppm) inhibits corrosion through the formation of a protective layer of
PbMoO4. High oxygen levels or air contamination can result in rapid
oxidization of the molybdenum. Other resistant materials are tantalum,
tungsten, iridium, and ruthenium. Short-term tests indicate that both
zirconium and titanium are catastrophically attacked by lead; the first

by alloying and the second by pitting.

8.1.2.6 Mercury. Mercury and mercury vapor in topping cycles has
been contained successfully in commercial practice using 2-l/h Croloy when
certain inhibitors had been added to the mercury. In experiments for the
mercury loop of the SNAP-8 experiment, Croloy 9M was used and there was
some evidence of corrosive attack. Tantalum was suggested as an appro-
priate replacement material for the Croloy; however, this was not evaluated
experimentally. It appears, however, that tantalum would be a satisfac-
tory containment material for secondary loops of molten-salt reactors if
it was shown that the Croloy would not perform satisfactorily. Tantalum,
of course, presents the difficulties associated with most of the refrac-

tory metals in fabrication of piping and components.

 
L8

8.1.3 Materials of construction for gases as secondary coolants

 

In general, none of the pure gases which have been listed as choices
for secondary coolants are corrosive to the normal materials used in reac-
tor construction. Since the gases will be maintained at relatively high
pressures and temperatures, the major requirement that the material of
construction must meet is that it have the required strength at the op-

erating temperature.

8.1.3.1 Helium. For the containment of helium any of the steels
which possess sufficient strength at the operating temperature are

satisfactory.

8.1.3.2 Hydrogen. At elevated temperatures, hydrogen embrittlement
of the ferritic materials occurs. The Hastelloys will perform satisfac-
torily; however, hydrogen will diffuse through all of these materials at

the operating temperatures.

8.1.3.3 WVWater vapor. The materials adaptable to any of the current
commercial steam systems can be used with the water vapor systems. Use
of the stainless steel will be required at the proposed operating tem-
peratures.

8.1.3.4 Air. The oxidation of iron-base alloys might make the use

of the nickel-base alloys desirable for use with air.

8.1.3.5 Nitrogen. Since nitrogen is essentially inert with respect
to most structural materials, almost any material which is suitable from
a structural standpoint will be acceptable. If the possibility of intro-
ducing either oxygen or moisture into the secondary system exists, then

the stainless steels should be used.

8.1.3.6 Carbon dioxide. Since carbon dioxide is particularly d4iffi-
cult to "dry" completely, and since the possibility of introducing moisture
into the system exists if the gas secondary coolant pressure is lower than
the steam system pressure, the use of the stainless steels is probably

required for carbon dioxide.

 
L9

8.2 Containment Materials for Various Secondary
Coolants in ConJjunction With the Primary Salt

 

 

In the primary heat exchanger there will be some material of con-
struction, probably in the form of tubes, which will separate the primary
fuel salt from the secondary coolant. A direct-contact heat exchanger is,
of course, possible in the case where inert gases are selected as the sec-
ondary coolant; however, this would negate the desire to contain the vola-
tile fission products in the primary loop, and it would add to the problems
of tritium transport to the secondary coolant. And, unless some design
which incorporates bundles of double-tube heat exchanger is utilized, a
material of construction will be required for the fabrication of a tube
sheet to secure the tubes of the tube bundle. The tube sheet will also be
in contact with both the primary fuel salt and the secondary coolant, un-
less some type of inert gas blanketing is designed for one or both sides.

The importance of the fact that two types of components, that is,
tubes and a tube sheet, will be in contact with both the primary salt and
the secondary coolant is related to the fabrication and strength require-
ments. The tubes may be relatively thin-walled while the tube sheet, if
the pressures of the primary and secondary system differ significantly,
may be quite thick. The fabrication techniques for the tubes and for the
tube sheet also differ, so that, unless one material of construction is
adaptable to both processes, different metals may be required. An addi-
tional requirement is that it should be possible to make a satisfactory
Junction between the tubes and the tube sheet and alsc between the tube

sheet and the shell or header of the primary heat exchanger.

8.2.1 Materials with molten salts as secondary coolants

 

8.2.1.1 Fluoride salts. Since the primary fluid 1s also a fluoride
salt, it might be concluded that those materials which were listed as
acceptable for the remainder of the secondary coolant system with fluoride
salts would also be acceptable for this part of the system. The difference,
however, exists due to the fact that the fluorides of uranium and thorium

are in the primary salt and they should not be present in the secondary

 
50

salt. The UF4 in the fuel salt can be reduced by chromium, iron, and

other elements by the following methods:
Cr + 2 UF4 = CrFs + UFg
Fe + 2 UFq4 = FeFs + UFa

The iron fluoride in the melt can increase the corrosion of material

by the further reduction of this fluoride by
FeFs + Cr 2 CrFs + Fe

The dissolution of oxide films from the metal surface can also cause selec-

tive removal of chromium by
2

2 Fe+3(in oxide film) + 3 Cr &2 Fe + 3 cr’

Consideration of this mechanism would eliminate 304 stainless steel,
which was cited as & possible material for use with Li-BeF4 as a secondary
coolant, as a material of construction for the primary heat exchanger even
though the primary salt is essentially Li-BeF4 with the fluorides of ura-
nium and thorium added. Tpe remaining materials which were listed as com-
patible with LizBeF4 as a secondary coolant will also be acceptable in the
primary heat exchanger. These are Inconel, nickel-molybdenum alloys, and
Hastelloy-N.

For the sodium fluoroborate, two of the materials which were previ-
ously listed for the system piping should be acceptable in the primary
heat exchanger. These were nickel-molybdenum alloys and Hastelloy-N.
Inconel which was questionable for use in the loop piping is still ques-
tionable for this application. Nickel cannot be used because temperatures
will be greater than 1000°F during normal operation, and duplex tubing
employing nickel could not be used since both sides of the tubing are
exposed to fluoride salts.

The same comments apply to FLINAK as were made for the fluoroborate;
however, Inconel should probably be eliminated from consideration for
this coolant,

The reactions just cited would also eliminate the use of the stain-

less steels in conjunction with the "low-melting point" fluorides and the

 
51
fuel salt. The same materials are applicable here as were listed for IoB.
These are Inconel, nickel-molybdenum alloys, and Hastelloy-N.

8.2.1.2 Chloride salts. Hastelloy-N appears to be about the only
material which could be used in the primary heat exchanger of a fluoride-

fueled molten-salt system with chloride salts as secondary coolants.

8.2.1.3 Sodium hydroxide. Since there appears to be no really

 

"good" material for the containmment of sodium hydroxide at the operating
conditions in the secondary loop, there is no good material for use in
the primary heat exchanger. The most likely choice appears to be a

nickel-molybdenum alloy, if & satisfactory material does exist.

8.2.1.4 DNitrates and carbonates. TFor the nitrate and carbonate

 

salts, the material requirements for compatibility with the primary fuel
salt are much more severe than those of the secondary salt. The possible

materials are Inconel, nickel-molybdenum alloys, and Hastelloy-N.

8.2.2 Materials with liquid metals as secondary coolants

 

8.2.2.1 Sodium and NaK. Because of the reactions of the primary
fuel salt with material contalning iron and chromium, the materials of
construction which are compatible with both sodium and NaK and the pri-
mary salt are Incoloy-800 and Hastelloy-N. BEven though the Hastelloy is
more expensive, it would probably be the more favorable choice for this

application.

8.2.2.2 Bismuth and bismuth—lead. The only materials which would

 

be compatible with molten bismuth and the fuel salt are the refractory
materials such as molybdenum and graphite. Neither of these lend them-
selves to the easy fabrication of tubes or tube sheets or to junctions

with other materials.

8.2.2.3 Mercury. In contact with mercury and the fuel salt, again,
the refractory materials would have to be employed for tubing and tube

sheets. Materials which could be used are molybdenum and tantalum.

8.2.2.4 Tead. In conjunction with the primary salt, the use of
lead as a secondary coolant would require the use of niobium, molybdenum.

tantalum, tungsten, or some of the other refractory metals.

 
52

8.2.3 Materials of construction for gases as secondary coolants

 

For most of the choices of gases for secondary coolants, the material
of construction in the primary heat exchanger would be dictated by the re-
quirements to contain the primary fuel salt. A possible exception would
be for the use of air (nitrogen and oxygen) as a secondary coolant. Irra-
diation of the secondary coolant by the delayed neutrons of the fuel salt
would result in the formation of oxides of nitrogen. Pipes containing air
in an irradiation environment have been found to corrode due to these
compounds .

Hydriding with hydrogen as a coolant and nitriding of the material of
construction with nitrogen as a coolant have not been found in the case of
the use of Hastelloy-N. Both of these reactions are possible with some of

the other materials which are compatible with the primary salt.

8.3 Containment Materials for Various Secondary
Coolants in Conjunction With Steam

 

 

In the secondary heat exchanger or steam generator of the secondary
loop, heat must be transferred from the secondary coolant to water and
steam. This unit will unquestionably have tubes to provide the heat trans-
fer surface area. It is possible that the tubes may be joined in banks of
headers so that the fabrication of a tube sheet may not be necessary. It
1s more probable, however, at least for the molten salt and liquid metal
secondary coolants, that some type of tube sheet will be required. The
materials for the construction of these components must be compatible with
the secondary coolant on one side and water—steam on the other. With the
possible exception of the high-pressure gases, there will be a large

pressure differential across these components.

8.3.1 Materials with molten salts as secondary coolants

 

8.3.1.1 Fluoride salts. For the sodium fluoroborate, the use of
Hastelloy-N is possible for the steam generator tubing; however, the cost
of this material is approximately $20 per pound. It appears possible to
use a duplex tubing with Nickel-280 bonded to Incoloy-800 on the steam
side; this duplex tubing might cost about $4 per pound.

 
 

53
The same choices exist when LoB is the secondary coolant or for
FLINAK and the "low melting point" fluorides.

8.3.1.2 Chloride salts. For the steam generator materials, the
same choices appear to exist when chloride salts are used for a secondary

coolant as those for the fluoride salts.

8.3.1.3 Sodium hydroxide. If the metal temperature can be re-

 

stricted in the steam generator, an alloy of nickel with 15% molybdenum
might be acceptable as a material for the tubing. Monel might be used

near the feedwater temperatures.

8.3.1.k Nitrates and carbonates. The stainless steels, Inconel,

 

and simllar materials usually used for high-temperature steam-plant
tubing can be used in conjunction with the nitrates and carbonates as

secondary coolants.

8.3.2 Materials with liquid metals as secondary coolants

 

8.3.2.1 Sodium and NaK. With either sodium or NaK as secondary
coolants, acceptable materials for the steam generator tubing are 2-l/h

Croloy, Incoloy-800, and Hastelloy-N.

8.3.2.2 Bismuth, lead, and bismuth—lead. In the presence of bis-

 

muth or lead, it appears that 2-1/4 Croloy could be used as the steam

generator tubing if the proper inhibiting agents are added.

8.3.2.3 Mercury. With either lead or mercury as secondary coolants,

2-1/& Croloy could be used as the material for the steam generator tubing.

8.3.3 Materials with gases as secondary coolants

 

For most of the gases, the resistance to the corrosion by steam con-
trols the choice of materials for the steam generator tubing. In these
cases, the Croloy or stainless steels may be used. For the use of air as
a secondary coolant, oxidation of the tubing in the high-temperature
section may require the use of a nickel-base alloy. In the economizer
sections of most of these units, mild steel may be used as the material

of construction.
Sk

8.4 Literature Sources

 

Two major sources of information with respect to the performance of
various materials of construction in the presence of the various choilces
for secondary coolants and the fuel salt and steam were drafts of surveys
of the literature that are, as yet, unpublished. These surveys have ex-
tensive citations of the literature; however, since these original sources
of information were not thoroughly investigated, these citations have not
been transferred to this presentation. Copies of these drafts are in the
hands of various persons at the Laboratory, and it is anticipated that
they will be available in published form sometime in the future.

One draft was prepared by H. Shimotake and J. C. Hesson at Argonne
National Laboratory; it was titled: "Corrosion by Fused Salts and Heavy
Tiquid Metals (A Survey)," and it was dated September 1965. It was pre-
pared for publication as an ACS Monograph; however, it was not included.
Hopefully, it will be available as an ANL report.

The second draft is entitled: '"Chapter VII: Corrosion in Fluid Fuel
Reactors." It has just become available in published form. It is con-

tained in a book entitled Nuclear Applications by Werren E. Berry, pub-

 

lished in 1971 by Wiley—Interscience (New York).

8.5 Summary and Conclusions Concerning
Materials of Construction

 

 

For the purpose of summarizing the information that is presented in
this section, three tables have been constructed. These tables indicate
the performance of various materials of construction (a) in contact with
the proposed secondary coolants, (b) in contact with the secondary cool-
ants and the fuel salt, and (c) in contact with the secondary coolants and
steam. In many cases, the performance of various materials of construction
can be substantially affected by the presence of certain contaminants.
These points are discussed in detail in the text.

In the construction of the performance tables, a series of perfor-
mance criteria are listed on the next page. Where more than one criterion
was applicable, multiple designations are indicated. For example, the

designation 467 indicates that the performance of the material would be

 
 

25

satisfactory only if certain inhibitors were present in the secondary
coolant, that fabrication of the piping and components from the material
of construction would be questionable or difficult, and that the material
would be expensive compared to other choices. A "blank" in the table is
not intended to indicate that the material is unsuitable with a particular
choice of coolant; it means, rather, that a more obvious selection for a
material existed and that consideration of further materials of construc-
tion for use with that coolant was not necessary. Note that generally

the higher numbers indicate more favorable performance for the material.

Symbol Performance of the Material for Designations in Tables 7, 8, 9,

Blank Material was not considered for this use.

Material is not suitable for this use.

Material can be used at temperatures less than 1100°F.

Material can be used at temperatures less than 1200°F.

Material can be used only at pressures less than 300 psia.
Material can be used only if certain inhibitors are present.
Material can be used only if certain contaminants are excluded.
Material can be used but fabrication is questionable or difficult.
Material can be used but it is relatively expensive.

Material can be used but a better alternative exists.

Best material for the application.

W oOoO-10W1 Fw ko

In the introduction of these sections, it was stated that the mate-
rials should be considered over the range of operating temperatures from
1000°F to 1300°F. The latter temperature is above the nominal operating
temperature of many parts of the secondary loop; however, this temperature
may be approached during transient or adverse operating conditions in
almost any part of the secondary loop. The tubes of the steam generator
will, for example, approach this temperature only if the turbine-plant
feedwater system fails and the steam generator blows down and "dries out"
in the time period immediately following a reactor scram. It would be
desirable to establish the design temperature for a component or piping
at the maximum temperature which the material can reach. In some in-
stances, it may be economically desirable to establish a lower design
temperature and to make system instrumentation and control so reliable
that it is highly improbable that this temperature will be exceeded.

The relationship between design temperatures and pressures for the

various materials is emphasized in Table 6. For the various materials
 

56

Table 6. Allowable Stresses and Design Pressures
for Various Materials

 

 

 

 

. Design Maximum Maximum Design Pressures
(Agfigeééiig- Tempera- Allowable
nation) ture Stress Case A Case B Casg C
(°F) (psi) (psig) (psi)  (psig)
Carbon Steel 900 6,500 2,500 1,200 985
(A-106) 1000 2,500 960 450 380
1100 --
2-1/4 Croloy 900 13,100 5,050 2,400 1,990
(A-213) 1000 7,800 3,000 1,500 1,185
1100 4,200 1,617 800 638
1200 2,000 700 400 304
1300
18-8 stainless 900 15,500 5,970 2,800 2,360
Steel (316H) 1000 15,300 5,890 2,700 2,325
(A-312) 1100 12,400 4,770 2,300 1,885
1200 7,400 2,850 1,400 1,125
1300 4,100 1,580 760 625
Inconel 600 900 16,000 6,160 3,000 2,435
(B-167) 1000 7,000 2,700 1,300 1,065
1100 3,000 1,150 600 450
1200 2,000 770 400 304
1300
Incoloy 800 900 13,100 5,045 2,400 1,990
(B-407, Gr. 2) 1000 12,800 4,930 2,300 1,950
1100 12,700 4,890 2,275 1,930
1200 7,900 3,040 1,500 1,200
1300 4,600 1,770 300 700
Hastelloy-N 900 18,000 6,930 3,200 2,750
1000 17,000 6,550 3,000 2,585
1100 13,000 5,000 2,400 1,975
1200 6,000 2,310 1,200 915
1300 3,500 1,350 680 535

 

indicated, the allowable stresses at the indicated maximum temperatures
have been extracted from Section 8, Division 1, of the ASME Boiler and
Pressure Vessel Code (July 1, 1971). To facilitate the visualization of
these allowable stresses, the maximum allowable pressures have been

calculated for the following cases:
Case

57

Designates

Maximum allowable internal pressure in a pipe which
is 1.0 inch in inside diameter with a 0.25-inch wall
thickness. (This is approximately Schedule 160 pipe.)

Maximum allowable external pressure with 1 psia internal
pressure in the pipe indicated in Case A.

Maximum allowable internal pressure in a pipe with an
internal diameter of 12.0 inches and wall thickness of
1.0 inch (approximately Schedule 160 pipe).

For the other materials listed in the following table, values of

maximum allowable stress have not been established. The nickel would have

to be a low percentage carbon alloy with a small percentage of additional

material to increase its high-temperature strength. In the nickel-duplex

tubing, the type of material used in the duplex would determine the

strength of the resultant tubing. One of the materials cited in Table 6

would probably be the duplexing material.

 
58

 

Table 7. Performance of Materials in Contact with
the Secondary Coolant Only

Tui g} P wf O Q =

SlelEEC S bk | 8 |24

s | bl &l A8 bl 3131858«
secondary | 8 1L kool 5|3 | L EBL LS8 8B2]8
Coolant = o & o 218 .34326*38‘%* % § 2 ES

O o e ] A = 0o =S == > E- E4 = O
NA BFL-NA F ol 5 781 11781 9|67]67T|67T{67T] 6
LI F--BE F2 81 9 7 7 67 671 6
FLINAK T 1| 81 9|67 67| 6
NA+K F-ZR FL 5 T1 11 8| 9167 671 6
NA ¢H 8118{ 8| 9 68
LI CL--K CL 5 T| 1 9|67
NITRATES 8] 9| 8
IMP FLU@RIDE 5 7 8| 9] 67 67| 6
CARBBNATE 1| 2| 8| 9| 8| 7
SEDIUM 9! 8| 8| 7T
NAK 9| 81 8} T
BISMUTH 24| 41 o] o © 671 67 6
PB-BI 2kt 4} o} o] o 67| 67 6
MERCURY 2k 67
LEAD 2L Ll 67] 671 67
HELIUM(HP) 9 81 7
HELIUM(LP) 13} 231 9 T
HYDROGEN (HP) 79
HYDROGEN(LP) | 13| 13§ 13 T
WATER (HP) 7
WATER(LP) 13 7
AIR(HP) ol 7
ATR(LP) 13| 23] 9 8f 7
NITRPGEN (HP) oh| 9 781 7
NITROGEN(LP) | 13| 23{ 9 T
Cg2(HP) 21 9 7
Cg2(LP) 13| 23| 9 7

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 
 

Table 8.
the Secondary Coolant and the Fuel Salt

29

Performance of Materials in Contact with

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

E1dlagl8lTl | 8 |
218188 o= 808 Gl 6188
a nwm o 0 — o r—l:!r—ix g — 42 g =
Coolant o v Rel s S8 Hdieele 3l e 51813 8&
Ol dNjlddAl H I H I Higpagis=l=EAl = | B | BHIE O
NA BF4-NA F ol © 791 ol 2| o|67|67|67]| 67
LI F--BE F2 ol 9 7 7
FLINAK 0f O 791 O] 2| 0| 6767|6767
NA+K F-ZR Fh o © 791 Ol 2| o|67|67T]67]67
NA ¢H 0 O
LI CL--K CL 9
NITRATES 8 9 8
IMP FLUPRIDE ol 9 T 7
CARB@NATE 8 9 8
SPDTUM 81179
NAK 8179
BISMUTH 67 6
PB-BI 67 6
MERCURY 671 67
LEAD 6716767167 6
HELIUM(EP) 79
HELTUM(LP) 19
HYDRPGEN (HP) 79
HYDRPGEN(LP) 79
WATER(HP) 79
WATER(LP) 79
- ATIR(HP) 79
ATIR(LP) 79
NITRPGEN(HP) 79
NITRPGEN(LP) 79
c@2(HP) 179
cg2(LP) 79

 

 
 

60

Table 9. Performance of Materiasls in Contact with
the Secondary Coolant and with Steam

 

'E)i :c? ol © O =
PlelEC |25k | B (2]
s |2 BA D[]S |wdxlalalBd] 8]«
Secondary ,8 ioomg o 3'S:c>>’§:>>§.3'§= S & 3 .8
Coolant % 'T'oc';'cfi & & 8 -.34'_‘-3'5’-3%'5' g § 2 %
&) Al Al H H 28] MQZEZQ = B4 E = O
NA BF4-NA F T 9
LI F--BE F2 7 9
FLINAK T 9
NA+K F-ZRb4 7 9
NA ¢H 29
LI CL--K CL T 9
NITRATES 2] 9| 8
IMP FLUJRIDE 7 9
CARB@NATE 81 9
S@DIUM 28| 9 T
NAK 28 9| T
BISMUTH 2hg
PB-BI 2h9
MERCURY 29
LEAD 2kg
HELIUM(HP) 29
HELIUM(LP) 13} 29
HYDRZGEN(HP) 8{ 9
HYDRPGEN(LP) | 13 8] 9
WATER (HP) 28| 9
WATER(LP) 13| 28] 9
ATR(HP) 28! 81 8] 9
AIR(LP) 13] 28] 8} 8| o9
NITRPGEN (HP) 281 9
NITRPGEN(LP) | 13} 28] 9
Cg2 (HP) 2581 8] 8| 9
Cp2(Lp) 1351258 9

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 
 

61

9. OTHER POSSIBLE COOLANTS OMITTED FROM
THE PREVIQUS PRESENTATION

9.1 Alternate Fluoride Salts

 

A fluoride salt mixture somewhat similar to the "low melting point"
(IMP) fluoride salt which has been included in the previous presentation
has been suggested. In this alternate salt, the BeFs of the IMP fluoride
would be essentially replaced with XKF. The composition of the alternate
salt in mole percent would be 46.5% LiF, 11.5% NaF, and 42.0% KF. The
substitution of KF for BeFy essentially makes this a "high melting point"
salt since its liquidus temperature is 455°C (851°F). The viscosity of
this salt is considerably lower than the LMP fluoride, and its stability
when exposed to steam is better since there is no precipitation of the
beryllium oxide. Because of its high liquidus temperature (the signifi-
cance of this liquidus temperature has been discussed previously), this
alternate salt was not included.

Another low melting point fluoride can be obtained by replacing the
LiF in the IMP fluoride with BeFs; this would leave a salt mixture con-
taining only NaF and BeF,. The liquidus temperatures of these mixtures
would be below 400°C (752°F); however, the viscosity would be much higher
than desirable and excessive pumping power in the secondary loop would be
required. The advantage of using this salt is that it contains no lithium
to dilute the lithium of the primary salt in the event of a leak between
the two systems. If the decision was made to make all lithium fluorides
in secondary coolants from lithium enriched in lithium-7, then this salt
would be considerably less expensive than the IMP fluoride. The conse-
quences and alternatives for lithium-containing secondary salts have al-
ready been discussed. Compatibility and containment problems for this
salt would be similar to those for the IMP fluoride. Because of its very
high viscosity, however, this salt composition has been omitted.

It has been suggested that the composition of this NaF-BeFs mixture
could be altered by substituting ZrF4 or AlFs for some of the BeF-. This
would result in a mix which had significantly lower viscosities than the

NaF-BeFp mix with no drastic increase in the liquidus temperature. One
62

alternative mixture containing NaF, KF, and ZrF4 has been included; it
has a melting temperature about the same as the sodium fluoroborate. One
of the advantages of this choice is that the BFs cover gas that is re-
quired with the fluoroborate is eliminated. There has been some concern
in the use of both ZrF4 and AlFs that a solid may be condensed from the
cover gas over mixtures containing these compounds. This "snow'" problem,
at least for the use of ZrF4 in the concentrations proposed has been re-
viewed. From consideration of the vapor pressure—temperature, it appears
that significant precipitation of the solid phase in the vapor space will
not occur if the maximum temperature difference in any of the components
exposed to this vapor is less than 80°C (150°F).

The other alternatives were omitted from this presentation due to the
lack of sufficient physical property data. Conclusions concerning com-
patibility with the primary salt and steam and suitable materials of con-
struction for these alternatives would be quite similar to those for the
similar fluoride salts. The limit in lowering the liquidus temperature
by selecting various compositions among these fluoride salts is about
315°C (599°F) which can be cbtained with melts of acceptable viscosity.

For a system which incorporated a primary salt with the composition
of 58% XF, 31% ZrF4, and 11% (Th,U)F4, on a mole basis, a secondary salt
of 58% KF and 42% ZrF, was proposed. The liquidus temperature of this
secondary salt was about 400°C (752°F). It did not appear to have any
particular advantage for systems with Li_BF, as the primary carrier salt
in comparison with the other fluoride salts in the study.

Stannous fluoride has been suggested as a possible secondary coolant
because of its low melting point, 215°C (L419°F), and its availability in
very pure form due to the current large-scale use in toothpaste. Unfor-
tunately, this material is extremely corrosive to most materials of con-
struction. No nickel or iron-based alloys can be used in a system with
stannous fluoride; the only acceptable materials of construction are the
refractory materials such as molybdenum or graphite. Neither of these
materials is easily adaptable to the fabrication of the secondary system
components. It might be noted that PbF, and BiFs are similar to SnFp in

oxidizing and corrosive effect.

 
 

63
9.2 The Cyanides

The cyanides have been suggested as possible secondary coolants prin-
cipally because of the possibility of introducing hydrogen atoms to be
used as "tritium traps" in the form of either ammonium cyanide or hydrogen
cyanide. Sodium cyanide, the most readily available of the group, has a
melting point of 563.7°C (1046.6°F) which is too high for consideration.
The CuCN—NaCN system has a eutectic melting about 350°C (660°F), and the
AgCN-NaCN system has a eutectic melting about 300°C (572°F). Potassium
cyanide has eutectics with AgCN and CuCN melting below 300°C (572°F).
Since hydrogen atoms can be added to the other secondary coolants by
methods which have been previously discussed, the added biological hazard
of handling large volumes of cyanides plus some questions of stability in
the radiation environment appeared to disqualify the cyanides from further

consideration.

9.3 Organic Secondary Coolants

 

Organic materials such as Santowax OM or Santowax WR, which have low
vapor pressures at the proposed operating conditions and appropriate
melting temperatures, have been suggested. While the oversall stability of
these materials in an irradiation field is considered good, there is
enough radiolitically-~induced decomposition to form a significant fouling
film on the heat transfer surfaces of the primary heat exchanger. If
these surfaces could be cleaned at reasonable intervals, then these cool-
ants might be considered. However, the difficulties of obtaining access
to this part of the reactor system appear to be sufficient to eliminate

the consideration of these compounds.

9.4 sulfur

Molten sulfur was also proposed as a secondary coolant. The melting
point of this element is extremely favorable, 119°C (246°F). Again, how-
ever, the substance is extremely corrosive, and it appears that the only

suitable material of construction would be carbon.

 
 

6L
9.5 Haloaluminates

A mixture composed of 69.5% NaAlCls, 19.5% KA1Cl,, and 1.0% NH4A1Cl,
(mole percentages) has been proposed. This mixture has the advantage of
providing hydrogen atoms in the NH4A1Clg which can be removed by distilla-
tion and thus function as a tritium trap; it also has a liquidus tempera-
ture of 125°C (257°F). No information was available on materials of con-
taimment, but those materials which were applicable to the fluoride and
chloride salts are certainly applicable to this salt. The only compati-
bility problem would be the reaction with steam to release HCl. More
information is required on the vapor pressure of this mix, its thermal
conductivity and heat capacity, and an evaluation of its performance from

the corrosion standpoint.

 
65

10. SUMMARY AND CONCLUSIONS

This presentation is a review of various choices for secondary cool-
ants+~in a molten-salt reactor system. It assumes that the current best
choices for the secondary coolant is the sodium fluoroborate in conjunc-
tion with a primary fuel salt which is a mixture of lithium and beryllium
fluorides in the mole ratio of 2:1 plus the fluorides of uranium and tho-
rium. The study was prompted by a desire to determine a secondary coolant
which would perform as well or better than the sodium fluoroborate as a
heat transport medium and, at the same time, be effective in preventing
the migration of tritium that was formed in the primary loop to the steam
in the power cycle. It was desired that the alternative choice for a
secondary coolant should also avoid some or all of the additional problems
assoclated with the use of the fluoroborate. These problems include pos-
sible interactions with the primary salt, excessive corrosion rates, and
the need for a relatively high overpressure of BFs gas. Secondary objec-
tives of the study were to reveal, if it existed, a secondary coolant that
would have performance characteristics superior to sodium fluoroborate,
and to gather in one document information on the physical characteristics
and engineering performance of other choices for secondary coolant to
function as a basis for further study in this area.

Possible choices for secondary coolants were divided into three
groups; these were (a) molten salts, including sodium hydroxide, (b)
liquid metals, and (c) gases. It was realized that before a final deci-
sion could be made concerning the relative merits of any acceptable secon-
dary coolant, an engineering design of a molten-salt reactor plant that
wvas, at least to some degree, optimized for that coolant, would be re-
quired. If such a design were made for the best choice in each of the
three groups indicated above, then the same design should be useful in
evaluating any choice in that group.

Superior performance as a heat transfer medium for any coolant
should be related to either minimizing the heat transfer area required in
the system or the pumping power or both. These requirements can be re-
lated, by conventional engineering correlations, to the physical property

values for the various coolants at the average loop temperature. In the

 
 

66

category of molten salts, NaOH and FLINAK appeared to produce the lowest
area requirements and FLINAK and the mixture of NaF, KF, and ZrF, produced
the lowest requirements based both on area and power requirements. The
mixture of NaF, KF, and ZrF, had the lowest power requirements. The rela-
tive differences between the values were not large, however, and an appro-
priate system design could eliminate some of these differnces in perfor-
mance characteristics.

The liquid metals, on an overall basis, indicated much smaller heat
transfer area requirements; this would be expected due to the large heat
transfer coefficients that are obtained with these coolants. Sodium and
NaK also indicated the lowest area requirement in this group of coolants.
When power requirements are considered, less favorable factors are ob-
tained since these factors assume constant mass velocity in the piping.
The group of gases produced relatively high factors indicating the re-
quirements for heat transfer area and pumping power as would be expected.

When total inventory costs for the various choices of secondary cool-
ants were compared, significant differences (an order of magnitude) were
found only for a few of the choices; these were bismuth, bismuth—lead,
and mercury in the liquid metal group. In the molten salt group, if en-
riched lithium-7 was used in the formulation of those choices that con-
tained lithium, the inventory costs were increased by factors between 2
and 7 depending upon the lithium content of the coolant. For coolants
which were relatively expensive with the incorporation of natural lithium,
these added inventory charges corresponding to these factors were greater
than $5,000,000. As would be expected the inventory charges for the
gaseous coolants were significantly lower than charges for the other two
groups .

Some attempt was made to relate the melting or liquidus temperature
of the molten salts and the liquid metals to characteristics of the engi-
neering design of the system and the operating characteristics of the
system. On the tentative bases that it was undesirable to develop the
so0lid phase in the secondary loop and that the feedwater temperature must
approach the liquidus or melting temperature by at least 50°F, certain

regstrictions can be listed.

 
67

As a specific example, if the liquidus or melting temperature is in
the range of 750°F, this assumption implies that the feedwater temperature
should approach the critical temperature. To avoid poor performance in
the steam generator, the feedwater pressures must also be supercritical.

To function effectively in the prevention of tritium migration to the
steam system, the secondary coolant must either contain hydrogen atoms
which can be exchanged with the tritium atoms, or it must contain chemi-
cally active groups which will react to form stable compounds with the
tritium. To prevent the coolant from becoming saturated with tritium
during use, it must contain a large concentration of hydrogen atoms or
active groups so that it becomes a reservoir for tritium, or the hydrogen
atoms or active groups must be contained in a minor component of the sec-
ondary coolant which can be removed and replenished by simple chemical
processing.

Of the coolants that were considered, only sodium hydroxide, hydrogen
gas, and water vapor contain sufficient hydrogen atoms to act as reser-
voirs. A large number of possible sources of hydrogen atoms and a limited
number of compounds with active groups exist that can be added to the
various coolants and removed on a continuous basis. The most significant
point 1s that such a compound exists for the sodium fluorcborate in the
form of a complex hydroxl ion which forms NaBFsOH in the melt. Normal
"drying" procedures used to process the salt will break down this complex
and remove it; then a slight amount of moisture can be allowed to reform
the complex. If this low concentration of hydroxl ion does not increase
the corrosive characteristics of the salt, then these mechanisms appear
to be acceptable. Hydrogen fluoride gas for all of the fluoride salts or
hydrogen gas or a mixture of both for the liquid metals can be a source of
hydrogen atoms in the coolant which can be easily stripped and replaced
by equipment similar to that used to remove the fission product gases from
the primary loop.

Several consequences of mixing between the secondary coolants and the
primary fuel salt or the steam need to be considered. These include exo-
thermic reactions, precipitation of fuel material, and introduction of
poisons into the primary salt. Most of the fluoride salts are chemically

compatible with the fuel salt although some can reduce the uranium to a

 
68

metal. For those salts which contain lithium, much study will have to be
given to the question of whether the lithium of the secondary coolant
should be enriched in lithium-7. The oxygen-containing secondary salts
will precipitate the oxides of uranium and thorium; consideration will
have to be given both to mixing rates and reaction rates to determine the
exact consequences. Most undergo rather mild reactions with steam.

The active liquid metals will also produce metallic uranium and
thorium when mixed with the primary salt. Their reaction with steam is
significantly exothermic. The choices for gaseous coolants appear to be
compatible with both the primary salt and steam except that, again, those
which contain oxygen can cause the precipitation of the oxides of uranium.

When suitable materials for contaimnment of the various coolants are
considered, satisfactory alloys appear to exist for most of the choices.
The only coolant which must be eliminated on this basis is sodium hydrox-
ide; there appears to be no satisfactory containment material at the
operating temperatures. Additional consideration must be given to mate-
rials of construction for the tubes of the primary heat exchanger and the
steam generators where the material is in contact with the secondary cool-
ant and the primary salt and steam, respectively. This consideration
eliminates some materials which were acceptable for use in conjunction with
the coolant. Hastelloy-N (INOR-8) appears to be satisfactory in many
cases; however, it is relatively expensive. Nickel-molybdenum alloys also
appear to be acceptable with many of the fluoride salts. Hastelloy-N and
Incoloy-800 are acceptable for sodium and potassium in conjunction with
the fuel salt.

When a consideration is made of the various choices for secondary
coolants by groups, there appears to be at present no characteristic of
sodium fluoroborate which would exclude its consideration among the other
molten salts. Most of the other possible choices in this group have po-
tential problems which can be as difficult to relieve as those of sodium
fluoroborate. One of the disadvantages of sodium fluoroborate is its
relatively high liquidus temperature of 725°F; however, this is signifi-
cantly lower than the temperatures of 856°F for LoB and 851°F for FLINAK.
The mixture of lithium and potassium chloride has a liquidus temperature

of 680°F, and it appears to have good performance characteristics; its

 
69

only disadvantage is that it introduces another halogen in the system
which, for some materials of construction, may produce additional prob-
lems. All of the lithium-containing molten-salt coolants will result in
significantly higher inventory charges if it is determined that enriched
lithium is required to prevent contamination of the enriched lithium of
the primary salt. Both the carbonates and nitrates present compatibility
problems when mixed with the primary salt.

Of the possible molten-salt coolants, it appears that sodium fluoro-
borate is a good choice with the chlorides and possibly the low melting
point fluorides being the second and third alternative choices.

In the group of liquid metals, sodium is the obvious first choice.
The bismuth, bismuth—lead, or lead coolants are too corrosive as is the
mercury; in addition, mercury is too expensive and probably not available
in sufficient supply. Introduction of sodium as the secondary coolant
would present compatibility problems with both the primary fuel salt and:
the steam. In addition, incorporation of sodium as the secondary coolant
would lead to design problems in the elimination of excessive thermal 7
stresses in the tubes and tube sheets of the steam generator. The use of
sodium as a secondary coolant also introduces the problem of incompati-
bility with the steam and possibly also problems of compatibility with
the primary salt. At present, there appears to be no incentive to intro-
duce these problems into the molten-salt reactor concept by choosing
sodium in preference to a molten-salt unless the lower melting temperature
of the sodium proved to be a decided advantage.

While most of the proposed gaseous coolants appear to be acceptable,
helium and hydrogen have obvious advantages. Because helium has obvious
advantages in safety of handling and because the use of hydrogen would
provide a reservoir for tritium rather than a trap, helium appears to be
the better choice. The major problem related to the use of helium as a
secondary coolant is that of designing a high-pressure secondary system.
Design methods do exist for such a system, but considerable evaluation
will have to be given to the choice of a design which is optimized for

this coolant.

 
70O

11. ACKNOWLEDGMENTS

Considerable assistance was received in the preparation of this
report from Roy E.-Thoma, who provided physical property data, informa-
tion of phase relationships and reactions between various components.

S. Cantor also contributed to this body of information for some of the
components. Information on the acceptable materials of construction for
most of the systems was obtained from John W. Koger and Jack H. Devan.
J. W. Cooke provided estimates for some thermal conductivity data for
molten-salt melts that otherwise were not available. J. R. McWherter
and H. A. MclLain providéd information on present and previous consider-
ations of design alternatives for molten-salt systems and considerations
of coolant performance criteria. Many others of the molten-salt reactor
project contributed from their knowledge and experience to this informa-
tion. Dunlap Scott provided a point of departure and the overall aim

of the study.

 
O -1 OW1 W o

g EgoddHd P qgrossngHgHOOEHOEHAREZNDOQOY

 

. L. Anderson
. F'. Baes

E. Beall

. Bender

. 5. Bettis

. F. Blankenship
. G. Bohlmann

. E. Boyd

. B. Briggs

. Cantor

. Compere

. Crowley

. Culler

. DiStefano
. Ditto

. Batherly
. Ferguson
. Ferris

. Frye

. Furlong
Grimes

. Grindell
. Guymon

. Haubenreich
. Helms

. Holz

. Huntley

. Keyes

. Kirslis

. Koger

. Korsmeyer

=g HdEED "N O R2HEHYS W e

71

32.

Distribution

2PrEHTTgHEOODEN e Em R e

~

un

Ty Eqrgogxddgyy
tOoOHTD W HEH =T

Gale

TR A YO Q HH

ORNL~CF-T71-8-10

. Krakoviak

. Imndin

. MacPherson
. MacPherson
. Malinauskas
. McCoy

. McDuffie

McLain
McNeese

. McWherter
. Moore

Nicholson

. Perry — J. R. Engel
. Rosenthal

Sanders

ap Scott

. Shaffer

. Strehlow

. Thoma

. Trauger

. Weinberg

. Welr

. Whatley

. White — A, S. Meyer
. Wichner

Young

Central Research Library (CRL)
Document Reference Section (DRS)
Laboratory Records (IRD)
Laboratory Records — Record

copy (IRD-RC)