ORNL-TM-5325.txt

   

L7 =/ S ORNL/TM-5325

Evaluation of Alternate Secondary
(and Tertiary) Coolants for the
Molten-Salt Breeder Reactor

 

A. D. Kelmers
C. F. Baes

E. S. Bettis

J. Brynestad
S. Cantor

J. R. Engel
W. R. Grimes
H. E. McCoy
A. S. Meyer

 

 

 

T
w

e - OAK RIDGE NATIONAL LABORATORY

- ) . OPERATED BY UNION CARBIDE CORPORATION FOR THE ENERGY RESEARCH AND DEVELOPMENT ADMINISTRATION

 
 

Printed in the United States of America. Available from
National Technical Information Service
U.S. Department of Commerce
5285 Port Roval Road, Springfield, Virginia 22161
Price: Printed Copy $5.00; Microfiche $2.25

 

 

 

 

This report was prepared as an account of work sponsored by the United States
Government, Neither the United States nor the Energy Research and Development
Administration/United States Nuclear Regulatory Commission, nor any of their
employees, nor any of their contractors, subcontractors, or their employees, makes
any warranty, express or implied, or assumes any legal liability or respons:bility for
the accuracy, completeness or usefuiness of any information, apparatus, product or
process disclosed, or represents that its use would not infringe privateiy owned rights,

 

 

ot
ORNL/TM-5325
uc-76

Contract No, W-7405-eng-26

Chemistry Division

EVALUATION OF ALTERNATE SECONDARY (AND TERTIARY)
COOLANTS FOR THE MOLTEN~SALT BREEDER REACTOR

A, D. Kelmers
C. F. Baes

E. S. Bettis
J. Brynestad
S, Cantor

J. R. Engel
W. R. Grimes
H, E, McCoy
A. S, Meyer

    
  
  
  
   
    
 

..__-———-'-"—"'NOTICE T work
This repott was prepared as an account of #

gojs;\so[repd by the United States (?ovcrnment. };:enher
the United States nor the United States Energy

Adminietration, not any of
Reseasch and Devclon e of their contractors,

Lied, or assumes any legat
the accuracy, completeness
tion, apparatus, product or
ould not

warranty, express of imp
Bability or nspnns\b\}fi;_; for
or usefuiness of any informa 8
PIOCESS disclosed, o1 represents that its use w
infringe privately owned rights.

APRIL 1976

OAK RIDGE NATIONAL LABORATORY
Oak Ridge, Tennessee 37830
operated by
UNION CARBIDE CORPORATION
for the
ENERGY RESEARCH AND DEVELOPMENT ADMINISTRATION

 
iii
PREFACE

Much of this report was written during the period September 19Tk
through March 1975. Since then, additional information has been developed
which bears upon a number of the alternate coolant considerations. Prog-
ress relating to several pertinent topics is summarized in the following
paragraphs.

The discussion of NaBFu-NaF (92-8 mole %) as a potential secondary
coolant (Section 5) indicates that, at the time this report was drafted,
no data were avalilable for estimating the tritium retention characteris-
tics of this salt, afid that the absence of tritium trapping could be a
disadvantage for a single coolant. (Tritium trapping in the coolant salt
is expected to be one of the best potential methods for limiting tritium
transport into the steam system and then into the environment.) Since
that time some preliminary experiments have been performed in engineering-
scale eguipment at ORNL which indicate that this salt mixture does have
substantial tritium trapping capability. These experiments, which were
started in July, 1975, in the Coolant-Salt Technology Facility, involve
the addition of tritiated hydrogen to high-purity NaBFh—NaF eutectic that
contains no deliberate additives to enhance the tritium retention. The
tritium-hydrogen mixture is added by diffusion through a metal tube to
simulate the diffusion through MSBR heat-exchanger tubes, but at flow

>

rates (per unit of tube area) that are th to 10 times those to be
expected in a reactor.

Results of a recent steady-state experiment, in which tritiated

hydrogen was added to the system for more than 4 weeks, with the salt at
iv

811°K (1000°F), indicated that up to 96% of the tritium was trapped in
the salt and subsequently released to the loop off-gas. The apparent
ratio of the concentrations of combined and elemental tritium in the
circulating salt was about 4000. This ratio is important in determining
the rate at which tritium can be removed in a stripping system as opposed
to the rate at which the elemental form can escape into the steam system.

While the information that is currently available is inadequate for
accurate extrapolation to the rate of tritium release to the steam system
of an MSBR, it appears that the sodium fluoroborate salt mixture would
have a substantial inhibiting effect on such release and that environ-
mentally acceptable rates (<10 Ci/d) could be achieved with reasonable
effort.

—

Also, with regard to the use of NaBFh-NaF (92—8 mole %) as a second-
ary coolant (Section 5), concern was expressed about potential reactions
between coolant and fuel salt LiF-BeF,-ThF)-UF) (71.7-16.0-12.0-0.3 mole
%) in the event of mixing. Laboratory experiments have shown: (i) The
rate of evalution of BF3 gas on mixing was low, about 30 minutes were

required to complete the reaction

NaBFh(d) - BFB(g) + NaF(d)

coolant fuel salt

Presumably the rate-limiting step was transfer of NaF across the salt-salt
interface, thus, in a reactor system with turbulent flow, the release of
BF3 might be more rapid. However, the results are encouraging relative

to MSBRs in that very rapid gas release resulting in significant pressure

surges was not experienced. (ii) No tendency was observed for the fissile

fuel salt constituents thorium or uranium to redistribute or to form more
concentrated solutions, or to precipitate following mixing of coolant
salt into fuel salt. (iii) Apparently an oxide species forms in the

coolant salt phase which is more stable than U02, since no UO, precipi-

2
tation was observed even when molten coolant-fuel salt mixtures were
agitated while exposed to air for several hours. Thus, large amounts

of oxygenated compounds could be added to the fluoroborate coolant salt
for the purpose of sequestering tritium, since leakage of such a coolant
gsalt into the fuel salt would not lead to precipitation of uranium or
thorium.

Investigation of NaBFu melts by x-ray powder diffraction, infrared
spectroscopy and Raman spectroscopy have identified the stable ring
compound Na3B3F6O3 as the probable oxygen-containing species in coolant
melts. Measurements of condensates trapped from the CSTF loop show a

> relative to the salt, suggesting that a

tritium concentration of 10
volative species may be selectively transporting tritium from the loop
through the vapor. Recent results indicate that BF3'2H20 may exist as a
molecular compound in the vapor and could be responsible for the tritium
trapping.

All the above results are favorable for the use of NaBFh-NaF (92-8
mole %) melts as an MSBR secondary coolant. Satisfactory tritium trapping,
an important coolant criterion, appears highly'probable, vigorous chemical
reactions or pressure surges were not encountered on mixing coolant and
fuel salt, and precipitafion or segregation of fissile components was not
encountered.

A ternary salt NaF-LiF-BeF, (45-22-33 mole %), has been considered

as an alternate coolant (Section 6). Two published references gave
vi

freezing points of 290°C and 340°C for a melt of this composition. Since
the freezing point is an important coolant criterion, differential thermal
analysis techniques were used to reinvestigate the portion of the phase
diagram near this composition. The results confirmed that the composi-
tion NaF—LiF—BeF2 (45-22-33 mole %) had the lowest liquidus temperature

in this region of the phase diagram. A freezing temperature of about

335°C would be a practical value for engineering considerations.
4,

vii

CONTENTS

SUMMARY AND RECOMMENDATIONS .« &+ « « + « o+ o
1.1 Comparison of Best MSBR Coolants

1.2 Recommended Future Work . . . . . . . .
INTRODUCTION &« « & « & o o o & s o s o« o« o
2.1 Purpose and Organization of the Report
2.2 Terminology « « ¢ « ¢ + + v a4 a4 e
COOLANT CRITERIA . .+ + ¢ ¢ o & o o s o o« @

3.1 safety Significant Events . . . . . . .
3.2 Anticipated Off-Design Transients . . .

3.3 Design Characteristics . . . . . . . .
REJECTED COOLANT CANDIDATES . « - - .« . «

4.1 Single Secondary Coolants . . . . . . .
4.2 Dual Coolant Configurations . . . . . .

STATUS OF FLUOROBORATE COOLANT . . . . . . .

5.1 Safety Significant Criteria . . . . . .
5 . 2 Off-DESign Tra]flSientS . . . . . . . .
5.3 Design Factors . + + « « « o o & o+ &

7_.
ALTERNATE SINGLE COOLANT: NaF- LiF-BeF, . .

6.1 Safety Significant Criteria . . . . . .
6. 2 Off—DeSign Tra]'lsients » . . . . . . * .
6.3 Design Factors « « « « « ¢« « « « « « .

ALTERNATE DUAL COOLANT CANDIDATES .+ + « « &

7.1 Secondary Coolant S e e e e s e e e
7.2 Compressed Helium Tertiary Coolant . .
7.3 Molten Salt Tertiary Coolant . . . . .
7.4 Liquid Metal Tertiary Coolants . . . .
7.5 Fluidized-Bed Tertiary System e e e
ACKNOWLEDGEMENTS . « & ¢« 4 ¢ & ¢« o o ¢ & o &
RE:FEMNCES . . . - . - - . - . - . » . . - .

13
13
14
15

17
18
23

27
27
36
38

39
40
45

48

48
49
51

56

57
59
65
69
73

74

75
viii

DEDICATION

We dedicate this report to our late friend and colleague,
A. S. (Al) Meyer, whose untiring efforts in developing molten-salt

reactor technology and unfailing good humor will always be remembered.
1. SUMMARY AND RECOMMENDATIONS

1.1 Comparison of Best MSBR Coolants

The three most promising coolant selections for an MSBR have been
identified and evaluated in detail from the many coolants considered in
this report for application either as a secondary coolant in 1000-MW(e)
MSBR configurations using only one coolant, or as secondary and tertiary
coolants in an MSBR dual coolant configuration employing two different

coolants, These are: as single secondary coolants,

 

(1) a ternary sodium-lithium-berylliumfluoride melt (Nal.SuLiO.BGBeFH)’
(2) the sodium fluoroborate-sodium fluoride eutectic melt, the present
reference design secondary cooclant,
and, in the case of the dual coolant configuration,

(1) molten lithium-beryllim fluoride (Li Bqu) as the secondary

2
coolant and helium gas as the tertiary coolant.
A straightforward comparison of advantages and disadvantages has been made
in the case of the two single secondary coolant candidates and this compari-
son favors the ternary fluoride melt by a slight margin, primarily because
of problems to be encountered if the fluofoborate secondary coolant and
fuel salt were to become mixed. The application of sodium fluoroborate
melt as the single secondary cooclant may offer some potential, yet unproven,
advantage in tritium trapping which could offset other less desirable
characteristics. The coolants selected for the dual coolant configuration
appear, on the basis of present knowledge, to resolve all technological

problems and, in addition, to offer operational advantages; however, the

addition of another heat-transfer loop to the MSBR would decrease thermal
efficiency and entail an increase in both capital and operating costs. A
direct comparison between the more attractive single secondary coolant and
the dual, secondary and tertiary, coolant concepts cannot now be made since
technological factors on the one hand must be evaluated relative to economic
considerations on the other., The future work necessary to reach such a

comparison has been identified and is detailed in Section 1.2,

1.1.1 Single coolants

 

In this MSBR configuration, heat is transferred from the fuel salt (the
primary coolant) to the coolant (the secondary coolant) in several primary
heat exchangers. The heated coolant is then circulated in independent loops
to the steam generators. A side stream could be withdrawn and processed to
maintain the desired coolant redox potential.

Many potential secondary coolant candidates were considered and re-
jected (Section 4), primarily due to (i) potentially safety significant
incompatibility with the fuel salt or (ii) unacceptable corrosivity. Liguid
metals or compressed gases appeared unacceptable for these reasons. Of the
numerous molten salts considered, two appeared substantially better and
were evaluated in detail, These are sodium fluoroborate [NaBFq—NaF (g2-8
mole %), the current reference design coolant] which is considered in
Section 5, and a ternary sodium-lithium-beryllium fluoride (Nal.auLiO.BB
BeFu) which is reviewed in Section 6,

A one-on-one comparison of these two candidates was carried out in
terms of selected coolant criteria (Section 3) and is presented in Table
l1.1. Consideration of potentially safety significant events, off-design
transients, and design factors are covered in three groups of items. In

the case of safety significant criteria, items la and 1b in Table 1.1,
Table 1.1. Comparison of NaBFh—NaF eutectic and Na1.3hTLio.66Bth

 

Criteria

Safety Significant

a. Change in nuclear
reactivity in case of
leak into primary
system.

b. Chemical reactions in

case of fuel salt
coolant interleakage.

Off-vDesign Transients

a. Leak. of coolant into
primary system.

b, Leak of fuel salt
into coolant.

¢. Leak of steam into
coolant.

d. Leak of coolant into
steam system.

e. Leaks to cell
atmosphere.

NaBFh-NaF

Presence of loB precludes any increase in reactivity
due to bubbles or wvoids in the core from BF3 gas.

Pressure surges caused by the release of BF
threaten primary system boundary.

5 might

Released BF, will dissc¢lve in the fuel salt, may
penetrate tge core graphite, and could harm the
charcoal bteds. BF, in the fuel salt can be mostly
removed by inert-g sparge; lesser amounts can be
burned out neutronically.

BeFe, UFh, ThF),, di-and trivalent fission product
fluorides, noble-metal fission products are all
insoluble.

Corrosion of metals, formation of soluble oxides and
low partial pressure {<1 atm) of HF. Heat exchange
surfaces could bte fouled by insoluble corrosion
product fluorides, NBBCrFG and perhaps NaNiF3.

Na, sl ge®eF)

None,

None

No evolution of gas to affect graphite

or charcoal. NaF in the fuel ssalt
will reduce breeding gain
by a minor amount.

Only ncoble-metal fission products
are insoluble, all fissile materials
are completely soluble.

Corrosion of metals, formation of
insoluble BeQ and low partial
pressure of HF. Heat exchange sur-
faces could be fouled by BeO, and
perhaps by NaBCrF6 and NaNiF3.

Very little is known about the effects (such as stress corrosion cracking) of filuorides of

steam systems.

Coolant reacts with moisture to yield acidic wvapors.

Coolant reacts with moisture to
produce Be0 and acidic vapors.

Comparison

IN&LEhLiO=66Bth nay be
the safer coolant.

Effects of mixing fuel
salt and coolant are
less troublesome with

Na, 4,1, ggBeF),-

—

Effects of mixing steam
>and coolant are roughly
the same for both
coolants.

 

No advantage or serious
problem with either
coolant.
Teble 1.1 (Continued)

 

Criterisa

Design Factors

a. Corrosivity

b. Freezing Point

c. Heat Transfer and
Hydrodynamic
Properties

d. Vapor Pressure

e. BRadiation Stability

f. Tritium Trapping

g. Cost and Availability

NaBFh—NaF

Boron in coolant can be reduced by metallic chromium
and by some minor alloy constituents.

38Lo¢(723°F)

All properties measured

Thermel conductivity: = 0.4 Wm -1 oK -1

Heat capacity: Cp 0.36 cal g~1 °x~1

Viscosity: nh5h°c = 1.91 cp; n621°C = 1.07 cp
. = -3

Density: Pss000 1.86 g cm

Sizeable BF

3 decomposition pressure.

No chemical decomposition due to garmas.

Neutrons transmute 10
in corrosivity.

B with no significant increase

Isotope—exchange and/or oxidizing additives necessary.
High solubility of oxide ion maey improve tritium
trapping. BF3°H O vapor species may also be significant.

8L00 £t cost $0.3TM.
elements.

Coolant consists of common

Na) 5)Ldg ccBeF)

The coolant will not react
with alloy constitutents.

290 - 340°C(554 - E44OF)

Two references disagree.

All properties estimated.
= 0.85 W L oL

Cp 0.46 cal g=1 °x-1
Mysyoq = 22 °P3 n6§1°c T-b ep
0550°C =2.1 g cm

Insignificant vapor pressure,

go chemlical decomposition due

O gammas .

No effects on chemical stability
due to neutronically induced
transmutations.

Isotope~exchange and/or oxidizing
additives necessary.

3

8400 ft~ cost $6M.

Comparison

Corrosivity less for
6BeF although
the a%fference may not be
significant since correosion
will probably be governed
by additives necessary to
sequester tritium.

Less (and possibly no) pre-
heating of feedwater will

be necessary if Na Li
Bth is the coolan% 3477066

Overall roughly equivalent.
Film coefficient and volu-
metric heat-capacity better
for Na hng 66B6Fh Kine-
nmatic v1scosity more
favorable for NaBFh-NaF.

Slight advantage for
Nay quldg, geBeF),-

Both acceptable.

Possible advantage for
NaBFh-NaF.

Clear advantage for
NaBFh—NaF.
the ternary fluoride is preferred over sodium fluoroborate which would
release BF3 gas in the event of fuel-salt-coolant interleakage. The
potentially safety significant release of BF3 gas is the most serious
negative factor in considering fluorcborate as a secondary coolant (see
Section 5.,1.2). In the case of off-design transients, BF3 release as a
result of minor leaks, item 2a, is again a problem which is absent with
the ternary fluoride melt. TFissile materials might redistribute between
immiscible phases after leaks in the case of fluorobbrate while they would
be completely soluble in the ternary fluoride melt, item 2b, again favor-
ing this coolant. In considering design factors, differences in corrosi-
vity of the melts toward Hastelloy N, item 3a, slightly favor the ternary
fluoride melt but the differences are minor and in either case corrosion
will probably be governed by the conditions selected to sequester tritium
in the secondary ccolant. The freezing point criterion, item 3b, clearly
favors the ternary fluoride melt since its lower freezing point would re-
quire less preheating of the feedwater. Heat transfer and hydrodynamic
properties, item 3c, and radiation stability, item 3e, are equivalent for
the two candidates and in elther case quite adequate for MSBR application.

The necessity of maintaining a fixed BF,_, vapor pressure over the fluoro-

3
borate coolant, item 3d, gives a slight advantage to the ternary fluoride
melt, The probable necessity of trapping some portion of the tritium in

the secondary coolant, item 3f, may favor fluoroborate, although at this

time tritium trapping has not been demonstrated experimentally in fluoro-
borate melt, Finally, cost and availability, item 3g, clearly favors

fluoroborate since beryllim and 99.99+ % 7Li would be required for the

ternary fluoride melt. The cost for the Nal.SHLlO.SBBeFH coolant 1s less
than that for an equivalent volume of LiQBeFu since the lithium content
is lower,

The result of this comparison is that the ternary fluoride melt is
favored by a majority of the criteria, especially those associated with
potentially safety significant events and with the ability to cope with
off-design transients., The minority of criteria that do favor the fluoro-
borate coolant are in the area of design factors, where various aspects of
the ternary fluoride coolant that are less suitable could be accommodated
by suitable engineering design considerations. While the ternary flucride
melt appears to be the more suitable coolant for an MSBR design employing
a single secondary coolant, the sodium fluoroborate coolant would likely

be preferred if it can be shown to aid significantly in tritium management

in an MSBR.

1.1.2 Dual coclants

In this MSBR configuration, heat is transferred from the fuel salt
(the primary coolant) to a secondary coolant in several primary heat ex-
changers. The heated secondary coolant is circulated in several loops to
intermediate heat exchangers where the heat is transferred to a tertiary
coolant which is then circulated in several loops to the steam-raising
system. Side stream processing might be required on both the secondary
and tertiary coolant loops to remove tritium, remove corrosion products
and/or adjust the coolant redox potential,

Three combinations of dual coolants were evaluated in detall for this
MSBR configuration (Section 7). In each case the secondary coolant was

lithium-beryllim fluoride; 7Li BeF, , the coolant used previously in the

27
MSRE. It was selected because of its complete compatibility with the fuel
salt in the event of mixing due to leaks in the primary heat exchanger or
other causes. In addition, its relatively high melting point helps de-
crease the possibility of fuel salt freezing during thermal transients.
All design factors also favored this secondary coolant with the exception
of cost due to its 7Li content, Three different tertiary cooclants were
evaluated, a compressed gas (helium, Section 7.2), a different molten salt
(a ternary carbonate melt, Section 7.3) and a liquid metal (molten sodium,
Section 7.4)., Of these three, compressed helium was by far the most at-
tractive candidate and is the only one recommended for further consideration.
Liquid sodium was considered to be less attractive due to severe chemical
incompatibility with both the secondary coolant and the steam in case of
leaks, and problems associated with thermal shock to structural components,
Tritium trapping methods are being developed for the LMFBR but may not be
adequate for MSBRs. A molten carbonate tertiary coolant was more suitable
than liquid sodium since it could readily afford methods of trapping large
amounts of tritium; however, it is chemically reactive with the secondary
coolant, releasing CO2 gas on mixing, and little information is available
concerning materials of adequate corrosion resistance to construct the
third loop. Thus the carbonate tertiary coolant concept was not felt to
warrant additional attention at this time.

The use of compressed helium (700 psia) as the tertiary coolant in an
MSBR concept coupled with molten 7LiQBePLL as the secondary coolant appears
to meet all technological requirements for an MSBR, but only at some addi-
tional cost for construction and operation., The advantages are:

(1) tritium can be readily trapped by the addition of 0, and/or H,O

2

at low concentration to the helium lcop; the proportions will be
dependent upon the rate of back diffusion of normal hydrogen
from the steam-system.

(2) the 7Li2BeF4 secondary coolant is completely compatible with

the fuel salt on intermixing, thus leakage of secondary coolant
into the primary circuit is not a serious matter

(3) steam leaks into the helium loop from the steam-raising system

would not cause a major increase in corrosion of the tertiary
loop nor would helium leaking into the steam system lead to
damage

(4) operation and control of the MSBR is simplified by the "soft"

coupling introduced by the helium loop

(5) start-up of the MSBR is easier and a much smaller auxiliary steam

generator would be required than in the case of an all molten
salt MSBR

(6) the possibility of fuel freezing on thermal transients is greatly

reduced |

(7) steam generator technology already developed for the HTGR could

be adapted for this MSBR configuration

(8) plant availability and maintainability would be improved by the

added passive barrier introduced by the tertiary loop.

The only apparent disadvantages to this MSBR configurétion are the
added cost and decreased thermal efficiency. The cost increase comes
primarily from the hardware required for the third loop and from the
decreased thermal efficiency. The decrease in thermal efficiency results
from the pumping power necessary to circulate the helium, Very pre-

liminary estimates (Section 7.3.2) indicate that the added cost may be
relatively modest, but a more detailed analysis is needed.
1.2 Recommended Future Work

In order to reach a final choice of a coolant or dual coolants for an
MSBR, additional cost information is required so that a quantitative com-
parison can be made among the three coolant selections defined in Section
1l.1. The following work 1s recommended:

(1) preliminary engineering conceptual designs and cost estimates of

a 1000-MW(e) MSBR with a single coolant configuration employing
either the ternary fluoride melt or sodium fluoroborate as the
secondary céolant. Similar information should be developed simul-
taneously for the dual cooclant configuration with molten LiQBeFL+
as the secondary coolant and compressed helium as the tertiary
coolant,

(2) definition of the tritium trapping capability of the coolants,
either secondary or tertiary, and experimental demonstration of
such trapping in helium, sodium fluoroborate and the ternary
fluoride melt, and

(3) experimental evaluation of the fuel salt-sodium fluoroborate
compatibility~questions.

The recommended work items should be carried out concurrently., The experi-
mental work called for in recommendations 2 and 3 would help supply accu-
rate information for conceptual design work and cost data. Simultaneously,
as the conceptual designs and cost estimates advance they will help guide
the experimental work to the most critical areas. A direct comparison of

the single vs dual coolant MSBR configurations and a final selection of an
10

MSBR coolant(s) can be made only after these recommendations are carried
out., Reduction of the comparison of the competing concepts to a comparison
of construction and operating costs provides the only means for quantitative

evaluation.

1.2.1 Conceptual designs and cost estimates

 

Development of capital and operating costs adequate to make a mean-
ingful comparison among the coolant choices and configurations is the most
important recommendation. Dollars are the only common denominator among
the disparate factors that must be evaluated and a comparison cannot be
made until cost estimates are available, Work of this nature done during
the preparation of this report was, of necessity, quite limited and is
useful only in suggesting that costs associated with the helium tertiary
loop may not be unattractive.

Adequate information should be developed to aid in the selection of
fluoroborate or the ternary fluoride melt in single coolant configuraticns.
It is anticipated that problems associated with fuel salt-ccoolant inter-
mising (Section 1.2.3) may play a dominant role in the selection and may
favor the ternary fluoride. The extent to which sodium fluoroborate can
assist in tritium management is also quite important. In evaluating the
- dual coolant configuration, heat transfer calculations and estimates of
the salt volumes for the primary and secondary loops will be important.
Also, sizing of helium-loop components, ducts, circulators, steam gene-
rators and tritium removal systems should be done with greater accuracy.
The cost and availability of 7Li compounds should be better defined,
Details to be considered in the conceptual design include, for example,

pressure relief mechanisms in the coolant loop in case of major steam
11

inleakage, how to deal with or prevent cooling the fuel salt below its
liquidus temperature and sizing the auxiliary steam system, Also questions
such as establishing relative levels of plant availability and maintain-

ability in different configurations should be considered.,

1.2.2 Tritium trapping
It is recommended that adequate experimental data be developed to
define the capability of tritium trapping in the three coolants: helium,
the ternary fluoride melt and the fluoroborate melt, Approximately 2400 Ci
of tritium per day will be generated in the MSBR fuel salt, Only about
0.1% of-this material can be permittéd to diffuse through the steam gene-
rators to the steam system, from which it would be discharged to the
environment (Section 3.3.5), Many factors affect the distribution of
tritium in the MSBR, These include:
(1) the Uu+/U3f ratio in the fuel salt, which controls the ratio of
TH/(T,H)F
(2) ability of the core graphite to sorb tritium and/or (T,H)F
(3) tritium diffusion through the primary system pressure boundary
to the cell atmosphere
(4) tritium trapping in the secondary or tertiary coolant
(5) decreased permeability to tritium in the steam generator tube
walls due to oxide formation on the steam side,
Currently, none of these factors has been adequately quantified. Experi-
mental work is under way to investigate items (1), (4) and (5). Para-
metric studies indicate that perhaps half or more of the tritium must be
trapped in the coolant in order to limit the environmental release to no

more than 1 to 2 Ci per day.
12

Since tritium trapping in the coolant has been established as a
significant criterion and the ability, or lack of ability, to sequester
tritium has been considered a major factor in favoring some coolant
candidates, experimental work on each of the five factors defined in the
preceding paragraph will be needed, including laboratory experiments and
circulating loops with provisions for removal of trapped tritium from the
respective coolant. Potential environmental impacts due to reactor
operation are currently receiving increased attention and establishment
of an acceptable level of tritium release and experimental demonstration
that this could be achieved in an MSBR are important aspects in the ulti-
mate selection of a practical and acceptable coolant or coolants for the

MSBR.,

1.2,3 Fuel-salt fluorcoborate mixing problems

Problems associated with the intermixing of sodium fluoroborate and
fuel salt need to be more carefully defined. Leakage of fluoroborate into
the primary circuit will generate BF3 gas. Over a wide range of equili-
brium conditions the resulting pressure may not be large; however, under
dynamic conditions much greater pressure transients could be developed.
Such situations may be safety significant and should be assessed carefully.
If fuel salt leaks into the secondary circuit, the fissile materials would
be relatively insoluble in the fluoroborate and would precipitate. Addi-
tional information is needed to understand the complicated salt system
formed after mixing and to define the concentration of the wvarious com-
ponents in the phase or phases which result. Information of this type will
help establish the seriousness of fuel salt-coolant intermixing, an im-

portant evaluation criterion.
13
2, INTRODUCTION
2.1 Purpose and Organization of the Report

The purpose of this report is to evaluate alternate secondary (and
tertiary) coolants for the Molten-Salt Breeder Reactor (MSBR). While
extensive experience has been accumulated for many years with molten fuel
salts,:L including operation of the Molten-Salt Reactor Experiment, the
selection and evaluation of an MSBR secondary coolant has received less
attention. A sodium fluoroborate melt, actually the eutectic composition
NaBPu—NaF (92-8 mole %), was proposed2 in 1965 and is the current reference
design coolant.3 It has been recognized, however, that this sodium fluoro-
borate melt is less than ideal in some r*esPectsq’5 and, therefore, an
evaluation of fluoroborate and alternate coolants was carried out, This
report comprises the finding of that evaluation,

First, a set of coolant criteria was established that would be perti-
nent regardless of what the coolant choices might be. The criteria were
divided into three categories: (i) safety significant events, (ii) anti-
cipated off-design transients, and (iii) design characteristics. These
criteria, presented in Section 3, were then used to evaluate various al-
ternate coolant candidates as well as to reevaluate the sodium fluoroborate
eutectic mixture. As the criteria were developed and applied, the first
two categories dominated many considerations and resulted in the rejection
of a number of coolant candidates (Section 4), The status of fluoroborate
coolant, relative to the criteria, is presented in Section 5. In Sections
6 and 7, evaluations of several potential alternate coolant concepts for

the MSBR are presented.
14

Based on the available information and reasonable estimates, the
coolants (fluoroborate and alternates) are evaluated by degree of compli-
ance with the coolant criteria and probability of successful development
and application to an MSBR. Also, recommendations are made relative to
the work needed, either experimental or conceptual design, to resolve un-

known areas to permit a final selection of an MSBR coclant or coolants.
2.2 Terminology

Definitions for some of the terms used frequently in this report
follows,

Primary Loop - The first circulating loop is referred to as the pri-
mary or first coolant loop, since it accepts the heat generated by nuclear
fission in the core. This loop contains the fuel salt, or primary coolant,
which is circulated to the primary heat exchangers where the heat is trans-
ferred to the fluid in the next loop.

Secondary Loop - The second circulating loop contains the secondary
coolant, which is the first coolant other than the fuel salt. In the con-
ceptual design this coolant is a fluoroborate eutectic melt and is used to
transfer heat from the primary heat exchangers to the steam-raising system.

Tertiary Loop (optional) -~ In some conceptual MSBR configurations a
third coolant loop is employed which contains the tertiary coclant, or the
second coolant transfers its heat via intermediate heat exchangers to the
tertiary coolant which then circulates between the intermediate heat ex-
changers and the steam-raising system.

Coolant - When used without a describing adjective, the term "coolant"

refers to the fluid (molten salt, liquid metal or compressed gas), within
15

the secondary or tertiary loop under consideration.

Steam-Raising System - This refers to the heat exchangers which
transfer heat from either the secondary or tertiary coolant to the water
or supercritical steam,

Steam System - This term describes the entire system in contact with
steam or water and thus includes the steam-raising system, superheaters,
preheaters, turbines, condensers, etc.

MSBR - A complete molten-salt breeder reactor facility of a nominal

233

1000-MW(e) capacity. The fuel is considered to be U in the nominal

fuel-carrier salt mixture LiF—BeFQ—ThFu (72-16-12 mole %).

3., COOLANT CRITERTIA

Any secondary coolant (or combination of coolants in secondary and
tertiary loops) used to transfer the heat generated in the fuel salt in
the primary loop to the steam-raising system must, obviously, satisfy a
number of requirements which will lead to a practical, safe and economical
design for an MSBR., Certain restrictions on the choice of coolant are
also imposed; these stem principally from safety requirements and antici-
pated events related to off-design conditions. Molten salts, liquid
metals, or gases could be selected as coolants and, to some extent, various
criteria would be more or less relevant to one or the other,

In this section, the coolant criteria - requirements and restrictions
- are detailed under three categories: (1) safety significant events,

(2) anticipated off-design transients and, (3) design characteristics,
The safety criteria are the most restrictive and absolute in nature and

are related to events which could lead to unacceptable consequences, The
16

criteria detailed under anticipated off-design transients are also some-
what restrictive but, in addition, involve "trade-off" features or aspects
of a coolant which can be accommodated by appropriate design. In general,
these criteria are formulated to preserve the integrity of reactor systems
during transients which can be anticipated. The third category, design
characteristics, includes c¢riteria related to optimizing the MSBR design
with regard to cost, thermal efficiency, etc., and are the least restric-
tive since desirable or less-desirable features of coolants and the asso-
ciated plant design must be evaluated and compared to reach a workable

and economically attractive design of an MSBR,

These criteria were developed, as much as possible, without consid-
ering any specific coolant or combination of coolants for an MSBR, but
rather by establishing a screening mechanism for evaluating coolant candi-
dates. Some of the criteria are absolute in nature; a coolant must meet
these criteria or it cannot be considered applicable irrespective of how
attractive it might appear in other aspects. Most of the criteria, how-
ever, are relative in nature and more or less favorable aspects of various
coolants can be adapted for use in an MSBR through suitable design., The
process of evaluating potential coolants is complicated somewhat by the
possibility of using two coclant circuits (secondary and tertiary) con-
taining different coolants to transfer heat from the fuel salt to the
steam-raising system. Thus, a palr of coolants could be selected, neither
of which meet all the criteria. A final compariscon and selection can be
achieved only when the various criteria can be reduced to capital and

operating costs; dollars are the common denominator in such a comparison,
17

3.1 Safety Significant Events

The criteria detailed in this subsection are limited to those related
to safety significant ewvents which could lead to temperature and/or
pressure excursions of sufficient magnitude to breach, or threaten to
breach, the primary system boundary. The characteristics of an MSBR are
inherently quite stable and safe; however, selection of an unsuitable
secondary coolant could compromise these characteristics in the event of
secondary coolant entering the fuel circuit. Thus, these criteria are
most useful in rejecting entire classes of potential candidate coolants

from further consideration.

3.1.1 Increase in reactivity

3.1.1.1 Precipitation of fissile material. The coolant must not be
capable of causing significant precipitation and/or segregation of fissile
material by the formation of compounds insoluble in fuel salt under any
event; e.g., leaks in the primary heat exchangers. For example, the tem-
perature and pressure surges in the core caused by the return of a few
(probably <10} kg of uranium which had precipitated outside the core,
might cause damage before being checked by inherent shutdown mechanisms.,
Use of a céolant potentially capable of causing such events would, there-
fore, likely represent an unsafe design. Precipitation of fissile material
in the coolant circuit probably would not represent a safety problem, al-
though shut down and clean-up of the secondary circuit would be required
to remove fission product contamination. Hidden precipitation of fissile
material would probably violate safeguard accountability requirements,

3.1.1.2 Gas injection into core, Another restriction results from

 

the fact that the MSBR core has a small positive reactivity response to
18

the introduction of wvoids or gas bubbles of materials with low neutron-

. 3 . .
capture cross-section, Thus inleakage of a gaseous coolant or production
of gases resulting in a void volume greater than about 10% over the entire

core could lead to unacceptable surges in fission rate and must be avoided.

3.1.2 Gas Egperation or in-leakqgg

 

The generation of large volumes of gases or release of a high-pressure
gaseous coolant as a result of a primary heat exchanger leak cannot be
permitted to over-pressurize and damage or rupture the fuel salt system
or inactivate or poison the fissién-product gas absorber beds of the off-
gas system. Such events would lead to release of radicactivity within the
MSBR containment and could treaten the primary system boundary. Minor in-
leakage or generation of gas could probably be accommodated by suitable

design,

3.1.3 Chemical reactions

 

Vigorous chemical reactions should not occur between fuel and coolant
as a result of a leak or other mixing event., In addition, coolant leaking
into the primary system should not react vigorously with the graphite
moderator. Chemical reactions which could lead to serious damage or des-
truction of reactor structural components could be considered safety signi-
ficant events due to the potential for release of radicactivity. Minor
chemical reactivity can probably be accommodated by suitable design if the

reaction products are innocuous or readily removable.

3.2 Anticipated Off-Design Transients
A variety of off-design conditions and transient events can be anti-

cipated to occur during the projected 30-year life of an MSBR, Examples
19

of such events are leaks in heat exchangers or thermal excursions on
reactor scram or turbine trip., The design of an MSBR and the selection
of a coolant or coolants must be able to accommodate such events with a
minimum of resulting down time or damage. These criteria are less re-
strictive than the safety-related criteria and involve, in many cases,

various features which can be accommodated by appropriate design.

3.2.1 Primary heat exchanger leaks

The conceptual design of a 1000-MW(e) MSBR power statiofi3 specifies
four shell-and-tube type primary heat exchangers each containing 5896
tubes. It is assumed that some tube failures will occur dyring the life
of the power stationj; therefore, the heat exchanger design incorporates
provisions for tube~bundle replacement or tube plugging by remote opera-
tion. Tube failures will, of course, lead to mixing of fuel and coolant,
Both massive and minor leaks must be considered. Minor leaks and their
results must be tolerable and repairable, Massive lgaks, such as the
collapse of an entire heat exchanger assembly, would not be expected to
occur; nonetheless, the possibility of such events must be recognized and
accommodated in the design although repair might be difficult. It is
felt that adequate assurance of unidirectional leakage cannot be achieved
by suitable adjustment of the fuel and coolant pressures; thus, leakage
in both directions through the primary heafi exchanger must be considered,
Similarly, leakage of the coolant into the reactor containment cell could
mix coolant and fuel in the drain tank, and the same criteria apply to
such an event.

The three most serious possible events as a result of primary heat

exchanger leaks - increases in reactivity, gas generation or in-leakage,
20

and vigorous chemical reactions have been considered above, Sections 3.l1.1,
3.1.2 and 3.1.3.

3.2.1.,1 Reduction in breeding gain. It is desirable that the breeding

 

gain not be substantially reduced as a result of the introduction of
nuclear poisons in the fuel which cannot readily be removed., This is a
less serious situation since power generation and operation as a con-
verter could be maintained by the addition of more uranium. In any case,
the valuable nuclear fuel should not be rendered useless or require re-
moval from the MSBR and expensive processing,

3.2.1.2 Clean-up of coolant. Leakage of fuel salt into the coolant

 

would introduce fission products in the coolant salt, If the coolant is
relatively inexpensive it could be discarded after recovery of fissile
material., If the coolant is more expensive and thus cannot be readily
discarded, methods for purifying the coolant should be available,

3,2,1.3 Fuel processing chemical plant., A side-stream of fuel is

 

continucusly withdrawn from the reactor and processed by fluorination for
the recovery of uranium, by reductive extraction for the removal and re-

tention of protactinium and by metal transfer for the removal of fission

products. Leakage of coolant into fuel could result in the transport of

dissolved or suspended coolant into these systems. It would be desirable
if significant chemical reactions or disruption of the function of the

chemical processing plant did not occur.

3.2,2 Steam-raising system leaks

 

The steam-raising system of a 1000-MW(e) MSBR contains a large heat-
exchange area having coolant on one side and supercritical steam on the

other, Leaks must be assumed to occur periodically during the 30-year
21

life of the MSBR. Although the predominant leakage would occur from the
high pressure steam intc the lower pressure coolant, it must be assumed
that some coélant could leak into the steam-raising system and be cir-
culated through the power plant and turbines. Leakage in either direction
must not result in vigorous chemical reactions which yield large quanti-
ties of heat or gaseous products. Such events could lead to wastage of
the structural material and/or disruption of the integrity of the system.
Mild chemical reaction could be acceptable if the reaction products are
tolerable or easily removable. Also, leakage should not lead to unac-
ceptably high corrosion rates in either the coolant loop or the steam
system. It is necessary that the coolant not be capable of causing
stress-corrosion cracking of the steam system following a mixing event,
Even ppb quantitities of some ions can lead to stress-corrosion cracking
of certain steam system alloys and, if such alloys are used, the coolant

cannot contain such ions.

3.2,3 Fuel-salt freezing

Events can be expected to occur which lead to rapid thermal excursions
and impose off-design conditions. The coolant or coolants and accompanying
system design must be able to accommodate such events, In the case of a
reactor scram, heat generation in the fuel salt via fission will be
abruptly decreased but heat removal will continue because of the finite
time involved in stopping the fuel and coclant circulation pumps and the
steam turbine, Damage to the primary heat exchangers caused by stresses
associated with freezing and thawing of the fuel salt would be difficult
and costly to repair and could lead to a radiological hazard. Thus, an

MSBR design should include features that preclude the possibility of fuel
22

salt freezing. Two conceivable methods of preventing freezing would be

to use (1) a molten salt coolant with a freezing point higher than that of
the fuel salt 502°C (935°F) or (2) a coolant with suitable heat-transport
properties.

For a single coolant system the first method is untenable since con-
ventional steam cycles demand a coolant with a considerably lower freezing
point (Section 3.3.2). The second method is conceivable if the reactor
design can accommodate a coolant with poor heat-transport properties., If
not, engineering precautions, possibly in the form of a dependable system
to by-pass coolant around the steam generators, would be required.

For a dual coolant design, this criterion could be satisfied by the
selection of a secondary coolant with a freezing point higher than 500°C
(932°F) while the tertiary coolant could have a lower freezing point com-
patible with the steam-raising system, In this configuration, freeze-up
would occur first in the intermediate heat exchanger. Volume change on
freezing and thawing must be accommodated by suitable design to prevent
rupturing portions of the intermediate heat exchangers or again a fast
acting by-pass system may be needed. A gaseous coolant in a tertiary
loop could possibly meet this requirement and might help circumvent the

problem of freeze-up in the intermediate heat exchange.

'3.2.4 Leaks to or from cell atmosphere

 

Leakage of coolant into the cell atmosphere - nitrogen plus oxygen
and traces of water vapor - should not result in violent chemical re~
actions or large volumes of gaseous products, Similarly, leakage of the
cell atmosphere into the coolant system again should not lead to vigorous

chemical reactions, and the reaction products should be readily removable.
23

An increased oxygen concentration in the coolant could lead to increased
corrosivity and could conceivably be unacceptable in the event of sub-
sequent mixing of contaminated coolant and fuel due to the very low
solubility of uranium, thorium and protactinium oxides in the fuel salt.

The products of minor inleakage should be innocuous or readily removable,

3,3 Design Characteristics

The criteria related to normal operating conditions are presented in
this subsection., For successful adaptation to an MSBR design, a coolant,
or combination of coolants in two loops, must meet all of these criteria
to some degree, Since it is very unlikely that any one candidate coolant
will be optimum in all categories, selection of a coolant to meet these
criteria involves an averaging of advantages of desirable features and
accommodation of the disadvantages where the coolant is less than optimum,
Particularly for the criteria in this section, consideration of the ulti-
mate capital and operating costs are necessary in comparing alternate

coolants,

3.3.1 Corrosion

The rate of corrosion of the coolant system boundary must be consis-
tent with the 30-year design life of the plant. Coolants which, because
of predictably high corrosion rates, would lead to rapid failure of plant
components are unacceptable. With some cooclant candidates, corrosion
considerations amount largely to a comparison of coolant cost and heat-
transfer and fluid properties with the increased cost of providing
greater metal thickness to withstand increased corrosion. Soluble cor-

rosion products may accumulate in molten salt coolants and must either
24

be removed or not interfere with operation of the coolant loop via

fouling of heat exchange surfaces,

3.3.2 Freezing point
Any freezing point below 330°C (626°F) would probably permit a feed-

water temperature as low as 304°C (580°F) which would be compatible with
conventional supercritical steam cycles.3 Higher coolant freezing points
can be tolerated, but with increasing cost and efficiency penalties as
the freezing temperature increases., If the freezing point is such that
the feedwater must be heated to 427°C (800°F) or higher, the increased
costs are probably prohibitive.sr Considering the thermal efficiency of
an MSBR and the ease of normal operation, there is no lower limit to the
freezing point. Gaseous cooclants, of course, offer another means of at-
taining the equivalent of a very low freezing point.' The criterion for
prevention of fuel freezing (Section 3.2.3) places additional restrictions
on the coolant freezing point in some design configurations., If a ter-
tiary coolant loop is employed, then only the tertiary coolant need have

a low freezing point.

3.3.3 Heat transfer and fluid transport properties

The coolant must have heat transfer and fluid transport properties
that are compatible with a practical and economical MSBR deisgn., These
properties include high thermal conductivity and heat capacity, and low
viscosity; such a combination..of properties implies high heat-transfer
coefficients and low pumping power requirements, Thus, for a gaseous
coolant to meet these criteria it would probably have to operate at

pressures of several hundred psi, Different coolants, molten salts,
25

liquid metals, or gases, will offer various combinations of these prop-
erties, For those candidate coolants which are not so extreme in some
property as to be summarily rejected, an analysis relating the cocolant
properties to capital ccsts (e.g., heat exchanger area) and operating

costs (e.g., pumping power requirements) of MSBRs would be required to

make a qualified selection between competitive coolants.

3.3.4 Vapor pressure and composition

The vapor pressure of a molten-salt or liquid-metal coolant should
be low at the highest temperature to be encountered during normal operation
or off-design excursions. Vapor pressures greater than one atmosphere
would complicate engineering design, but probably could be accommodated.
Even low vapor pressures, 0.0l to 1 atmosphere, will require recovering
the coolant vapor from the cover gas sweep stream and returning it to the
coolant to prevent depletion of volatile constituents. It would be ad-
vantageous if the coolant did not sublime which cculd result in formation
of a solid from the vapor in cooler regions with assoclated restrictions
in vent lines, erosion of pump shafts and seals, etc.

Consideration of a gaseous coolant implies a substantially different

design to accommodate the moderate-to-high pressures involved.

3.3.5 Tritium control

 

Because the fuel salt contains a high atomic density of lithium, a
significant quantity of tritium (2420 curies or about 0.25 g per day in
a 1000-MW(e) MSBRB) is generated in the reactor core. Since metals at
high temperatures are permeable to isotopes of hydrogen, a portion, cal-

culated to be in the range of 790-1500 Ci/day for the reference concept,
26

could reach the steam system.6’7 Virtually all of this tritium would be
released to the environment by normal system blowdown and/or leakage. The
associated release rate would be 50 to 100 times that for light-water-cooled
nuclear power stations and probably would be environmentally unacceptable.
If the tritium were allowed to accumulate in the steam system by requiring
total recycle of all steam discharges, the steam system probably would
become sufficiently radicactive (2 Ci/gal) to require personnel protection
during maintenance and would be expensive to maintain., Thus, it is prob-
able that the coolant system may be required to interdict in some way the
flow of tritium,

Various tritium trapping schemes have been proposed, including iso-
topic exchange or oxidation within the coolant and subsequent side-stream
removal of tritium as THO or TO- compounds. Further, the formation of an
oxide film on the steam side of the steam-raising system is expected to
decrease the tritium permeability in this portion of the system and thus
raise the partial pressure of tritium within the coolant, possibly to a
level where gas sparging or sorption could be used for tritium removal.

However, none of these potential methods for tritium control has
been evaluated or demonstrated experimentally nor has the environmental
impact been carefully assessed to set a limit for tritium releasej; thus,
it is difficult to establish quantitative criteria for this aspect of the
coolant or coolants. It is apparent that it would be beneficial if tri-
tium diffusing into the coolant were sequestered and not allowed to enter
the steam system. Depending on future experimental results and environ-
mental considerations, this could become a mandatory requirement. This

requirement could possibly be met by different means in secondary or
27

tertiary loops depending on the selectlon of a molten salt or a gaseous

coolant for that loop.

3.3.6 Radiation and chemical stability

The coolant should be highly resistant to radiation damage under all
operating conditions. It should not evolve gases or form sludges or solids
that foul heat-exchange surfaces. Neither should it become increasingly
corrosive with use. Minor chemical changes during operation could be ac-

commodated by side-stream processing to remove decomposition products.

3.3.7 Cost and avallability

The material chosen for the coolant should be composed of compounds
or elements which are readily available in adequate quantity and which
preferably are available in high purity. While low cost is desirable,
the capital and operating costs associated with a given coolant will be
more important than the initial coolant cost. Initial cost is a relative

item and must be compared to other attributes.
4, REJECTED COOLANT CANDIDATES

4,1 Single Secondary Coolants
In this section many possible coolants having a wide variety of
chemical and physical properties are considered, Most are rejected be-
cause leakage of coolant into fuel salt could cause, or could threaten
to cause, safety-significant events, or else because the coolant would be
too corrosive, Three coolants used in other reactors, sodium, helium,
and H.,0, were unacceptable because of safety-related problems that could

2

arise if leaks occurred in the primary heat exchangers. TFor the same
28

reason, organic coolants and low melting oxides (nitrates, nitrites, and
carbonates) were rejected. The molten mixture of LiF and BeF2 used as the
MSRE coolant was rejected because of the penalty in thermodynamic effi-
ciency associated with its high freezing point. The low melting metals,
lead, tin, and bismuth, were rejected because of their corrosivity toward
structural alloys; mercury was dismissed on account of its scarcity in

the earth's crust. Hydroxides and salts which contain easily reducible
cations (Ni2+, Cu2+, Sn2+, Pb2+, Fe3+ and Bin+) were rejected because they
would not be stable in nickel- or iron-base alloys of construction. Al-
though a single factor was sufficient for rejection, many of these fluias

could have been dismissed from further consideration because of other

serious shortcomings.

4.1.1 Coolants used in other nuclear reactors

 

4.1.1.1 Sodium. Sodium and other alkali metals are capable of
chemically reducing all cations in the fuel salt except lithium in case

and UF., are

of their leakage into fuel salt. The uranium fluorides, UF4 3

the most readily reduced fuel salt components and the reduced forms of
uranium are either sparingly soluble (UF3) or insoluble (uranium metal)
in fuel salt. Accordingly, inleakage of sodium through a primary heat-

exchanger would precipitate UF, or uranium metal, either of which could

3

subsequently cause an unacceptable increase in reactivity. Beryllium

metal, from BeF could alloy with structural metals. Thus, sodium is

2’
not an acceptable coolant choice.
Sodium has additional serious drawbacks. The very high thermal

conductivity of sodium, while advantageous for compact heat exchanger

and steam generator design, is a distinct disadvantage in dealing with
29

thermal transients and avoiding possible fuel freeze-up., Another dis-
advantage of sodium is its vigorous reaction with water or steam to pro-
duct hydrogen and sodium hydroxide or sodium oxide, depending on whether
water or sodium is in excess., Very rapid corrosion may accompany leakage
of steam into sodium,

Sodium may not offer an effective means of sequestering tritium. Al-
though in the LMFBR the tritium problem may be solved via cold trapping of
sodium tritide, this approach would probably not trap tritium in the MSBR
adequately since it is present in quantitites that are 50 to 100 times
greater than in the LMFBR.

4.1,1,2 Helium (and other high-pressure gases). Helium cannot be
used as the MSBR secondary coolant because its leakage into the fuel-salt
could lead to unacceptable power surges associated with the positive re-
activity coefficient for voids or bubbles (criterion e,1,1,2), Similarly,
other high-pressure gases having low neutron capture cross-sectilons (C02,

H,, Ne, Ar) can be rejected,

99
Even if "fail-safe'", leak-free primary heat exchangers were developed,
there are strong incentives for favoring a low-pressure liquid secondary
coolant over a high-pressure gaseous secondary coclant. The lower volu-
metric heat capacity of gases would require substantial increases in the
fuel salt inventory and in heat-exchange surface area, and result in greater
power required to circulate the coolant. Although these disadvantages can
be dealt with, the sum of the costs to do so could be substantial,
4,1.1.3 7LiF—BeF2. These two fluorides are major components of the

fuel salt and the consequences would be minimal if a 7LiF—-BeF2 coolant

were Mixed with the fuel. This optimum coolant-fuel compatibility was the
30

chief reason for using a mixture of 7LiF and BeF2 (66-34 mole %) in the
Molten-Salt Reactor Experiment (MSRE)., In this application, heat was
transferred from the coolant at 546°C (1015°F) to a low-efficiency, air-
~cooled radiator. The coolant performed in a completely satisfactory
manner in the MSRE. However, in an MSBR the coolant should be able to
deliver heat to portions of the steam system at much lower temperatures
than 546°C, Molten mixtures of 7LiF—BeF2 could be used, but only at sub-
stantial cost for auxiliaries. The exact composition that could be used
in the MSBR coolant circuit would be a compromise between high freezing
point and high viscosity; compositions of interest would have LiF contents
of 60 to 67 mole % resulting in freezing points between 4U4O and 460°C
(825-860°F), The need to prevent salt from freezing in the steam-raising
equipment would require an abnormally high feedwater temperature, and
result in a decrease in the thermal efficiency of the reactor. Assuming
a supercritical steam cycle in which the feedwater would be preheated to
426°C (800°F), approximately the lowest temperature if LiF-BeF, were the
coolant, Robertson3 estimated that the net plant efficiency would be 41,3%
as compared to 44.5% in a system having a feedwater temperature of 371°C
(700°F). The unconventional size of the preheat equipment (especially
pressure booster pumps) would impose additional costs.

A second unfavorable factor associated with use of 7LiF—BeF2 is the
cost of the coolant inventory. Assuming an inventory of 83500 ft3 of
7LiF-BeF2 (66-34 mole %) and estimated prices of $120/kg "Li and $86/kg
Be, the coolant would cos‘t5 approximately $13 million.

Although a single coolant MSBR could not use a coolant composed

solely of 7LiF and BeF, because of its high freezing point and cost, this

2
31

coolant appears very attractive in the dual coolant configurations des-
cribed in Section 7.
4.1.1.4 EQO (water or supercritical steam). Leakage of water in any

form into the fuel salt will cause precipitation of the fissile material

 

in the fuel, and, perhaps, a substantial increase in the bubble volume
reaching the core due to steam formation. A second serious shortcoming of
water stems from the high freezing point, 446°C (835°F), of the fuel salt.
Both of these events are potentially safety significant. To prevent fuel
salt from freezing in primary heat exchangers requires that the coolant
water be a gas (steam) with all the economic disadvantages that are thereby
incurred (see Section 4.1.1,2). For these reasons, as well as the need

to ensure against any possible fission-product contamination of the steam-
power system, it has never appeared feasible to raise steam directly in the
MSBR primary heat exchanger,

4,1.1.5 Organic coolants., This class of coolants is occasicnally

 

proposed as a means of coping with the MSBR tritium problem.5 The rela-
tively low temperatures (400 to 450°C, 752 to B42°F) at which pyrolysis
occurs is a sufficient basis for rejecting these materials., In addition,
an organic coolant, if mixed with molten fuel salt, is likely to chemically

reduce and precipitate the uranium,

4,1.2 Low-melting metals and salts

 

 

4,1.2.1 Metals (Pb, Sn, Bi, Hg). These metals, with the possible
exception of mercury, would rapidly corrode the nickel or iron-base alloy
structural materials likely to be used in the coolant circuit at the design
temperature of the MSBR., Even if compatible constructional materials were

available, the costs of circulating these dense liquids would be very high,
32

As with sodium, the high thermal conductivity of these molten metals is a
disadvantage when thermal transients occur. Aside from its well known
toxicity, the scarcity of mercury in the earth's crust is sufficient for
rejecting this metal as an MSBR coolant, A similar argument perhaps ap-
plies to bismuthj; in addition, this metal expands upon solidification, a
property which would complicate the design of coolant circuits., For these
four heavy metals the shortcomings are serious enough to discount their
use as an MSBR coolant,

4,1.2.2 Oxide-containing salts (nitrates, nitrites, hydroxides, and
carbonates). When a coolant containing substantial oxide, or oxide within
a complex ion, mixes with the fuel salt, precipitation of a solid phase
containing uranium oxide is likely, and this could be a safety-significant
event,

These oxide-containing coolant candidates which freeze below 400°C
(7529F) have other serious disadvantages. Nitrates, either alone or mixed
with nitrites, could react violently with the moderator graphite if a leak
occurred in a primary heat-exchanger. Carbonates are also unacceptable
because of fuel salt-coolant mixing considerations. The reaction of
fluoride with carbonate could release 002 in quantities sufficient to in-
crease the bubble fraction in the core to unacceptable levels, Molten
hydroxide coolants, although ideal for managing the MSBR tritium problem,
must be rejected because of their corrosiveness to nickel- and iron-based
alloys of construction.

4,1.2.3 Salts containing reducible cations. There are a number of

 

low melting halides that are unacceptable because they would cause in-

tolerable corrosion of the structural alloy Hastelloy N. One example is
33

SnF2 whose melting point is 212°C (414°F). This salt can oxidize any of
the major components of Hastelloy N, In general, metallic halides con-
taining ions more oxidizing than Ni2+ will not be stable in nickel-base

alloys; the cations in this category include Sn2+, Pb2+, Fe3+, Bi™" and

2 . . .2 . .
Cu +. Salt melts containing substantial Ni ¥ concentrations are rejected
because they are likely to lead to mass transfer of nickel metal. The

same considerations apply to iron-based alloys; in addition, ferrous

halides would be unacceptable because of mass-transfer problems.

4,1.3 Other molten halide mixtures

There are several other molten halides mixtures that have freezing
points below 400°C (752°F) which cannot be rejected because of a failure
to meet an important criterion; nonetheless these mixtures have short-
comings which render them less attractive MSBR secondary coolants. These
molten halides are discussed very briefly below.

4,1.3.1 Mixtures of alkali chlorides., These are typically mixtures
of LiCl, NaCl, and KCl1l, The LiCl is required to obtain a sufficiently low
melting point and, because of the necessity for using lithium-7, these
coolants would bé fairly expensive. Leakage of coolant intoc the fuel salt
would cause a reactivity loss which would be primarily remedied by removal
of chloride from the fuel salt; unfortunately, potassium and sodium cannot
be removed from the fuel-salt by the fuel processing circuit and these
ions would reduce the breeding gain., A more serious situation could arise
if coolant leaked into the steam system. Ferritic alloys may be acceptable
although chlorides cause stress-corrosion cracking in many steam-system

materials at very low concentrations in water or in steam.
34

4.,1.3.2 Mixtures of NaCl and AlCLS. The main potential advantages

of these mixtures are low freezing point, 151°C (303°F) for an equimolar

 

mixture, low viscosity, and low costs. If a substantial volume of this
melt leaked into the fuel-salt, a solid of the cryolite structure (NaSAlFs)
would likely precipitate; if carried into the core, these solids could
restrict flow in the channels through the moderator., Although the vapor
pressure of these mixtures is moderately high, the vapor can be easily
prevented from condensing by heating cold surfaces (e.g., vent lines) to
about 180°C (356°F). The general considerations of stress-corrosion
cracking resulting from chlorides in the steam system apply to this coolant
as well., It is not clear how this coolant would trap tritium,

4.1.3.83 ZrF,-KF-NaF (42-48-10 mole%). This salt mixture is sometimes

 

considered as a possible inexpensive alternate coolant because of its
chemical compatibility with fuel-salt in case of a leak in the primary

heat exchanger. However, the breeding gain would be permanently decreased
by the presence of potassium in the fuel salt. A disadvantage of this melt
is associated with its condensible vapor, preponderantly ZrFu. The "snow"
that would form could block vent lines and cause problems in pumps that

circulate the coolant.

4,1.4 Fluidized-bed coolants

 

A fluidized-bed concept could, in principle, be used to transport heat
directly from the fuel salt to the steam system. This alternative was also
suggested by an independent design study.8 However, no definitive conclu-
sions were reached. To avoid the potential hazards associated with the
presence of high pressure steam-system piping inside the reactor primary

containment and adjacent to piping containing fuel salt, such a system
35

would probably necessitate two fluidized beds to perform the functions

of the primary heat exchanger and the steam generator. Heat would be
transferred to the bed material inside the primary containment and the bed
material would be transported to steam generators located outside the
containment. ‘

This approach appears to offer a number of advantages over other
single-coolant concepts. The favorable heat-transfer and heat-transport
characteristics of the fluidized beds would reduce the heat-exchange
surface-area and fluid-pumping requirements to values near those normally
associated with liquid coolants. Absolute gas pressures potentially could
be as low or lower than fuel-system pressures to minimize the possibility
of introducing gas into the primary loop and reactor core in the event of
a fuel-tube failure, The use of an inert fluidizing gas, prcbably helium,
and an appropriately inert bed material could significantly reduce struc-
tural-material constraints. For example, ordinary steam-system materials
could probably be used for the steam generators since resistance to cor-
rosion by molten salt would not be a requirement, With an inert bed

material, additions of HQO and/or 0, might be made to the gas phase to

2
interdict the flow of tritium from the fuel system to the steam system.
Alternatively, the bed material itself might perform the tritium trapping
function.

A major requirement of this concept would be compatibility of the
bed material with the materials of both the fuel system and the steam
system. As a minimum, there should be no vigorous chemical interaction

between the fuel salt and the bed material. In addition it should be

possible to separate the bed material from fuel salt without excessive
36

effort, and small amounts of bed materials should not strongly affect the
chemical, physical, and neutronic characteristics of the fuel salt. In
particular, the bed material should have minimal neutron-moderating capa-
bility to avoid either positive nuclear reactivity excursions 1f bed
material were to enter the core, or nuclear criticality in the fluidized
bed if large amounts of fuel salt were to leak into the bed, Other po-
tential effects of salt leaks, such as loss of fluidization would also
have to be shown to be acceptable. Bed materials that could react chemi-
cally with high-temperature steam (e.g. graphite) would also be unaccept-
able., Since the bed material would likely become activated and/or con-
taminated in use, it should have sufficient mechanical stability to pre-
clude the replacement and disposal of large amounts of solids.

While an extensive survey has not been made, it appears likely that a
combination of fluidizing gas and bed material could be identified that
meets the fundamental requirements for an MSR application. However, since
fluidized-bed heat exchangers have not been used in power reactor appli-
cations, a substantial effort would be required to develop and demonstrate
the operability, reliability, maintainability, and safety of this concept
at levels commensurate with the requirements for nuclear systems. In view
of the fact that other, more conventional systems appear to be adequate
for MSBR applications, there appears to be no incentive for further consid-

eration of a fluidized-bed single coolant at this time,

4,2 Dual Coolant Configurations
The optimum, i,e, safest, and most economic, reactor configuration
may involve two coclants in series; a secondary coolant which transfers

heat from the fuel salt to a tertiary coolant which, in turn, produces the
37

steam to drive the turbogenerators. Thus two coolants could be employed,
neither of which meet all the coclant criteria. Dual cooclant configura-
tions which appear to be the most attractive will be discussed in detail
in Section 7. The purpose of this subsection is to identify the fluids

that are acceptable in secondary and tertiary loops.

4,2,1 Secondary coolants

Of all the fluids rejected in this section, only molten LiF—BeF2
mixtures are clearly acceptable secondary coolants. The criteria for re-
jecting all the others as single secondary cooclants also apply for re-
jecting them as secondary coolants in dual coolant configurations. They
either would pose safety problems if leaks occurred in the primary heat

exchangers or else they are toc corrosive toward structural alloys.

4.2,2 Tertiary (steam-raising) coolants

It was assumed that the prime purpose of an additional coolant loop
is to provide the means for sequestering tritium. Hence, any fluid that
cannot do this at reascnable cost while maintaining a tolerable rate of
corrosion was not considered acceptable, Nitrates and nitrites decompose
at too low a temperature to warrant serious consideration. Heavy metals
as well as the salts with reducible ions mentioned in Section 4.1.2 are
eliminated from further consideration on the same basis (corrosivity) by
which they were discounted as secondary coolants., Sodium is not likely
to sequester enough of the tritium to be an acceptable tertiary coolant.
Helium or the ternary eutectic (Na, Li, K)2003 are the two most attractive
tertiary coolants identified in this study.

The use of nitrate-nitrite mixtures as a steam-raising coolant has
38

been considered9 due to the ease of trapping tritium by oxidation to THO.
It appears, however, that the thermal stability of such mixtures is not
adequate to accommodate the design temperature of 621°C (1150°F). The
maximum temperature would probably have to be lowered to 482-538°C (900-
100°F) to prevent substantial degradation of the nitrate-nitrite mixture
with a concomitant decrease in overall plant thermal efficiency. Also,
very little information exists concerning potential materials of contain-
ment at these temperatures. For these reasons nitrate-nitrite mixtures
are not felt to be attractive tertiary coolants and were not considered

further.
5. STATUS OF FLUOROBORATE COOLANT

A salt mixture composed of sodium fluorocborate and sodium fluoride
was first proposed2 as the MSBR secondary coolant in 1965 after it was
recognized that the MSRE coolant, 7LiF-BeF2 (66-34 mole %), would be un-
desirable in a breeder reactor due to its high cost and high freezing
point. On the basis of a reportedlo eutectic temperature of 304°C (579°F),
low cost, and estimates of physical and chemical propertiesll that seemed

acceptable, molten NaBF, -NaF (92-8 mole %) appeared to be an attractive

4
secondary coolant fluid for the MSER,.

The principal advantages of the fluoroborate coclant are low cost and
low viscosity. Its actual freezing point 384°C (723°P)12 is somewhat
higher than desired, but the penalties associated with this freezing point
are not great. The chief disadvantages of this salt arise from events

associated with fuel-salt-coolant mixing. These include generation of

BF, gas and probable redistribution of fissile material between immiscible
39

phases., Most other properties of fluoroborate are acceptable with ade-
quate design considerations. A possible advantage is its potential for
trapping tritium. In the following section the NaBFu—NaF secondary coolant

is evaluated in terms of criteria detailed in Section 3.
5.1 Safety Significant Criteria

5.1.1 Increase of reactivity

 

Any mixing between coolant and fuel salt is very unlikely to lead to
an increase in nuclear reactivity. This coolant is inherently safe due to
the very high cross section of loB which virtually guarantees that fuel-
salt-coolant mixing would cause an immediate decrease in reactivity as soon

as coolant is swept into the core.

5.1.2 Chemical reactions

 

In the event that fluorocborate mixed with fuel salt due to leaks in
the primary heat exchangers or secondary coolant system, the reaction of
safety significance is the decomposition of NaBFu,

NaBFu(Z) + NaF(d) + BFa(g).

The reaction would be displaced to the right by dissolution of NaF in the
fuel salt mixture. From measurements of the solubility of BF3 in various
molten fluorides13 and from the preliminary studylu of phase equilibria
of coolant and fuel salts, the following will occur when coclant and fuel
salt are mixed in various proportions:

(a) an increase in BF_ pressure (except for quite low concentrations

3
of coolant in fuel salt)
(b) absorption of heat

(c) partial immiscibility of the resulting phases.
40

To better define the pressures generated by mixing fuel salt and coolant15
more measurements would be required, particularly dynamic mixing experi-
ments to investigate non-equilibrium conditions.

A BF3 pressure surge caused by a tubing fallure in a primary heat
exchanger might lead to further damage of the heat exchanger. It is
likely that a pressure relief device would be activated which would re~
lease gaseous fission products into the cell, If a break occurred in a
coolant circulation line such that a large volume (V500 fts) spilled out
onto the catch pan on the floor of the cell and drained into the fuel
drain tank through the thermally activated rupture disc at the top of the
tank, BF3 generated in the tank could carry gaseous fission products out
into the cell through the open drain,

The probability of such accidents are admittedly very small, The
chief value in considering them lies in devising the necessary engineering
safeguards to counter such possible dangers and in comparing fluoroborate

with other potential secondary coolant candidates.
5.2 0Off-Design Transients

5.2.1 Leaks in the primary heat exchanger

 

The effects and consequences of leaks outlined in this subsection
are caused by transients which can be anticipated, and are differentiated

from safety significant events. In this discussion, BF_, pressure surges

3
are not considered serious enough to cause breaks in the primary contain-
ment boundary. In the case of a small hole in a tube wall, coolant will
leak into the fuel salt and most of the BF3 and all of the NaF will dis-

solve in the fuel salt mixture. In case of a larger leak such as a tube
41

break, coolant will flow into fuel salt and fuel salt may flow into the
coolant, For still larger leaks two immiscible liquid phases may form.
Whenever BF3 is released in the primary circuit, some will be swept into
the off-gas system where it may interact with the charcoal beds and a
smaller amount may diffuse into the core graphite, but most of the BF3

is likely to dissolve in the fuel salt.

5.2.1.1 Gas generation or inleakage. (a) Dissolution in fuel salt.

 

 

The solubility of BF, is relatively high in fuel salt and even higher in

3
fuel salt containing a substantial concentration of NaF.13 For example,
at 704°C (1300°F) 4 moles of BF, confined to a volume half-filled with

fuel salt will be partitioned at equilibrium, 3 moles BF3 in the salt and
one mole in the vapor space., If 1 ft3 of coolant equilibrated with the
inventory of fuel salt (1720 fta) and if the only wvapor space in the fuel

circuit were a 1% bubble fraction in the salt, the BF, saturation pressure

3
would be only 0.23 atm (3 psi)., A far smaller amount of coolant (0,127
fts), if carried into the core, contains enough boron (0.6 kg) to render
the reactor subcritical. The 10.9 kg of sodium in 1 ft3 of coolant would
have a minor effect on reactivityj; about 2,000 kg of sodium in the core
would be required to cause the same negative reactivity effect as 0.6 kg of
boron. The consequences of a much larger leak are discussed in subsection

5.2.1.2.

(b) Interaction of BF, with charcoal. Some of the BF, released or

 

generated in the primary circuit and fuel drain tank will be swept into
the off-gas system, Adverse consequences might be inactivation of the

charcoal beds, or worse, selective sorption of BF, and desorption of the

3

xenon and krypton. Data for predicting the effect of BF3 on charcoal beds

is not available.
42

(c) Interaction of BF, with core graphite. No chemical reactions,

 

including intercalation, are known between BF, and graphite and chemical
effects on graphite should be absent.

5.2.1.2 Reduction in breeding gain. The breeding ratio of an MSER,

 

nominally 1,07, is reduced when fluoroborate coolant or any other source
of boron is present in the core. The amount of boron in the core that
changes the reactor into a converter, i.e., reduces the breeding ratio to
1.00, is quite small; 3 kg in the salt or 2 kg dispersed in the graphite.
The situation with sodium is quite different. It takes approximately

9000 kg of sodium in the core to reduce the breeding ratio to 1.00 and
thus sodium presents much less of a problem. The sodium in sodium fluoro-
borate, that enters the primary circuit would not penetrate the graphite.

(a) Boron and sodium in the fuel salt. In a "small" leak case,

 

1 ft3 of coolant is assumed to leak into the fuel salt. This amount of
coolant contains 4.7 kg of boron. If the boron is dispersed homogeneously
within the 1074 f't3 of fuel salt in the core, it would now contain 2,95 kg
of boron. More than enough will enter the core to bring the reactor sub-
critical. The boron content of the fuel salt can be greatly reduced by
allowing the fuel salt temperature to rise or by sparging with an inert

gas to effect BF, removal. These last two operations might be done in the

3
fuel drain tank or in the fuel storage tank. The BF3 volatilized from the
salt could be trapped in disposable NalP absorbers, Any lOB that cannot be
removed from the fuel salt by chemical or physical methods could be neu-
tronically burned out,

Larger inleakage of fluoroborate presents greater difficulties, For

example, if a double-ended tubing rupture occurred in the heat exchanger
43

and a volume of fluorcborate equal to the heat-exchanger shell (416 fts)
were to mix with the fuel salt, the mixture would contain approximately
2500 kg of boron and 5750 kg of sodium. Most of the boron and all of the
sodium would be in the fuel drain tank. As with a small leak, the boron
in the fuel salt could ke reduced to an acceptably low level by some com-
bination of sparging and heating. Even if all the boron were removed
from the fuel salt, the amount of sodium dissolved in the fuel salt is
such that the breeding gain of the reactor would be halved. Since sodium
cannot be removed by on-site reprocessing, continued use of the fuel salt
would depend on economic considerations.

(b) Boron in the graphite. No quantitative data exist to define what

 

BF, will do to the graphite that will be used in the MSBR. A 1000-MW(e)
reactor contains about 0.3 x lO6 kg of graphite and if about 2 kg of
boron is present on or in this graphite, the breeding ratio will decrease
to 1.00. If boron cannot be desorbed from the graphite, the reactivity

losses could be overridden by adding more uranium to the fuel salt,

5.2,1.3 Clean-up of coolant and coolant circuit. Any sizeable break

 

in a heat-exchanger tube is likely to leak fuel salt into the coolant.
Since all components of the fuel salt except LiF are insoluble in NaBFq,
the uranium in the fuel salt will most probably be entrained in the
coolant as a crystalline complex of UF#’ €eley Na2UF6. But even if the
LiF were not leached out by the coolant, the fuel would freeze when the
temperature in the coolant dropped below 502°C (935°F)., The contaminated
coolant could be discarded. Discard of the coolant in one of four coolant

circuits involves the loss of only about $100,000,

5.2.1.,4 Contamination of chemical plant. Entry of boron into the

 
By

on-site fuel-processing facilities will not create a serious problem.
The BF3 may separate from the UF6 in the fluorination and reconstitution
steps. In general, diminution of boron in the fuel salt to some ac-
ceptable level by prior sparging is the obvious provision for protecting

the chemical plant from boron and this should be readily achievable.

5.2.2 Other leaks

5.2.2.1 Inleakage of steam. Small leaks in the steam-raising system,

 

if detected and repaired within a short time, would not be very harmful.
Indeed, small amounts of water may aid tritium control in the coolant.
Large leaks, however, will greatly increaée metal corrosion and mass trans-
fer.l6 An associated complication is the formation of the relatively in-
soluble corrosion products NaSCrF6 and NaNiPS, which could foul steam-
generator surfaces and/or restrict narrow coolant flow channels. A method

for removing these solids may have to be developed.

5.2.2.2 Leakage of coolant into steam. A large break in a steam-

 

generator tube could introduce small amounts of coolant salt into the
steam-feedwater system counter to the pressure differential. The pH of
the water will be changed due to the formation of HF with fluoride-con-
taining salts and the water will also contain ppm concentrations of flu-
oride, There is no evidence that this could lead to stress-corrosion
cracking but this has not been studied sufficiently and experimental in-
vestigations would be needed,

5.2,2.3 Leaks to the cell atmosphere. Small amounts of coolant

 

leaking into the cell will react with moisture in the atmosphere and
generate acidic vapors that could cause minor corrosion of the metal

lining the concrete cell walls.
45
5.3 Design Factors

5.3.1 Corrosion

The inherent corrosivity of eutectic NaBFq-NaF melt toward nickel-
base alloys is small. In Hastelloy N, the minor alloy constituents Cr,
Ti, Mn, Si, and Hf will react with the salt. For instance, the following
reaction, at reactor temperatures has a negative standard free energy:

(1 + x)Cr(c) + 2NaF(d) + NaBFu(d) + Na CrFe(c) + CPXB(C)

3
If the concentration of chromium in the alloy is low (<0.1 atomic fraction),
this reaction is likely to be diffusion limited and of little significance.
Increased concentrations of chromium in the containment alloy could lead

to increased corrosion rates. Na CrF6 is sparingly soluble in flucro-

3
bor*a’cel'7 and plugging of flow channels might be a problem.

5.3.2 Freezing point

The NaBFu—NaF eutectic freezes at 384°C (723°F)12 which is too high
to be compatible with the highest feedwater temperature (v290°C, 580°F)
that could be used in an unmodified supercritical steam cycle, The design
modifications and costs necessary to preheat the feedwater to 371°C
(700°F) and the low pressure steam to 343°C (650°F) have been estimatedl8
and are a very small fraction of the total plant cdst. Although further
design modifications may be necessary to prevent possible damage caused by
salt freezing on cold spots in the salt side of steam generators, the
associated costs are prcbably modest,
5.3.3 Heat-transfer and hydrodynamic properties

The heat-transfer properties, thermal conductivity and thermal capa-

city of fluoroborate are'acceptable; the kinematic viscosity (viscosity
46

divided by density) is very favorable., Measurement of heat transferl9

in a tubular test-section of a pump loop showed that the performance of
fluoroborate was consistent with the Seider-Tate correlation which success-
fully correlates data for a wide variety of non-metallic fluids flowing

turbulently in pipes.

5.3.4 Vapor pressure and composition

 

The equilibrium decomposition pressures of BF3 over this coolant are
moderate; at the hot-leg temperature of the coolant loop, 621°C (1150°F)

p is 0.33 atm (5 psi).ll The BF
BF3

in the pumps since its critical temperature is -12°C (10°F), Present

3 cannot condense in the annular spaces
designs of coolant pumps call for helium or some other inert gas to sweep
downshaft toward the pump bowl. In the process, the sweep gas carries

BF3 out of the pump bowl and if makeup BF

slowly change composition and increase in freezing point,., The He—BF3 gas

3 is not added, the coolant will
mixture cannot be continuously discharged from the plant because of the

toxicity and chemical reactivity of BF Also, the amount of helium in-

3.
volved is probably too expensive to waste. Thus, it would be necessary

to develop a BF, recirculation system.

5.3.5 Radiation and chemical stability

 

The effect of the intense gamma radiation to which the coolant salt
would be exposed in the primary heat-exchangers was investigated by
E. L. Compere et al.zo Salt was exposed for 1460 hr in a Hastelloy N
capsule experiment at 600°C (1112°F) and no evidence of chemical decompo-
sition was detected from vapor pressure measurements or from metal weight

losses.,
47

Fluoroborate coolant will also be irradiated by delayed neutrons in
the primary heat exchangers. Their effect on fluoroborate chemistry has
not been studied experimentally. However, investigation of BF3 gas
suggests that the only significant chemical change in the salt would be
the release of fluorine by the reaction:

Br 4 n > "LiF + a4 T,
The number of delayed neutrons in primary heat exchangers will be about
lOl7 n/sec, thus the maximum yield of F2 from the coolant will be only ap-~
proximately 5 moles per year. This rather small amount of oxidant can

probably be handled by the processing methods that would be used to control

the redox conditions of the coolant.

9.3.6 Tritium

The sequestering of tritium in the fluorcborate coolant is as yet
undemonstrated. Ionic species containing oxidized hydrogen which could
exchange with and trap tritium are identified (BF30H—), expected (HFQ-),
or suspected (H+). There also may be polymeric oxyfluoroborate ions which
could bind gt or T+. It is not yet known, however, if one or more of
these species can be stabilized at a sufficiently high concentrations to
be effective for tritium trapping without exceeding the limits of the
oxidation potential of the coolant beyond which corrosive attack of nickel-
base alloys would become excessive, There is some experimental evidence2
that the proton of the BFSOH— species did not undergo the expected rapia
exchange with D+. Thus, there are a number of unresolved questions con-
cerning the potential of this coolant for tritium trapping, but this is

also true for other possible molten-salt coolants.
48

5.3.7 Cost and availability

 

The coolant components, NaBFq and NaF, are inexpensive and readily
available commercially, although the purity necessary for reactor use may
require some development efforts. In 1971, the estimated cost18 of the

8500 ft3 inventory was $0.37 M,

6, 'ALTERNATE SINGLE COOLANT: NaF-7LiF-BeF2

 

The composition of this coolant is NaF-7LiF—BeF2 (45-22-33 mole %) or

7.
3.31 Mg 66

except that sodium is substituted for approximately two-thirds of the

N BeF Thus, this coolant is very similar to the MSRE coolant,

W
lithium ions. A secondary coolant composed of these fluorides has a major
advantage of compatibility with fuel salt; i.e., the consequences of inter-
leakage or mixing with the fuel salt are minor or readily reversible,
Another advantage of this coolant is a very low vapor pressure. Disadvan-
tages are relatively high inventory cost and increased corrosivity when
mixed with steam. This coolant is reviewed in terms of the criteria de-

tailed in Section 3, as was the fluoroborate coolant in Section 5.

6.1 Safety Significant Criteria

7

Naj g L

io 66BeF4 is a very safe coolant that is compatible with fuel
salt. Mixing of fuel salt and coolant will not precipitate uranium, will
not release or absorb much heat, and will not generate gases by chemical

reactions. In case coolant is carried into the core, reactivity will de-

crease slightly because of the sodium,
49

6.2 Off-Design Transients

6.2.1 Leaks in primary heat exchangers

 

6.2.1.1 Secondary coolant into fuel salt. The effect of greatest

 

significance will be a reduction in the breeding gain arising from the
dissolution of NaF in the fuel salt; however, a considerable volume of
coolant is necessary. For instance, approximately 2000 kg of sodium in
the core could be necessary to render the reactor subcritical. Intro-
duction of this much sodium into the core would require the inleakage and
homogeneous distribution of about 6570 liters (230 fts) of secondary
coolant,

The fuel reprocessing system would not be significantly affected,
however, it would not remove the sodium from the fuel salt. A decision
would have to be made between continued operation with a lower breeding
performance by overriding the sodium poisconing via addition of more
uranium, reclaiming the 7Li and Be content by suitable processing methods,
or replacement of the fuel carrier salt.

Other effects of inleakage would be minimal. This coolant is miscible
with fuel salt and would not penetrate the moderator graphite. Also, fuel
salt contaminated with coolant would have virtually the same solubilities
of inert gases and hydrogen.

6.2.1.2 TFuel salt into the secondary coolant circuit. The damage to
the coolant and the coolant circuit would not be very great in case of an
inleakage of fuel salt. All constituents and fission products contained
in the fuel salt are soluble in the coolant salt except for '"noble'
metallic fission products Nb, Mo, Tc, Ru, Ag, and Sb. In such an event

it would probably be necessary to dispose of the contaminated coolant and
S0

replace it, After the leak was repaired, clean coolant salt could be
used to flush the loop. Any radioactivity remaining would be due to the

noble metal fission products adhering to metal surfaces.

6.2.2 Other leaks

6.2.2,1 Inleakage of steam. The effects of steam upon the coolant

 

and coolant loop would be corrosion of metal to form complex corrosion-
product fluorides and, if the solubility of oxide in the coolant were ex-
ceeded, precipitation of beryllium oxide. The reactions can be written as
follows: H,0(g) + M(alloy) + NaF(d) + BeF,(d) > BeO(s) + NaMF,(c) + H,(g),
where M represents Fe or Ni, and

3/2H,0(g) + Cr(alloy) + 3NaF(d) + 3/2BeF,(d) + BeO(c) + NasCrf.(c) + 3/2H,(g).

3
These reactions would probably proceed to completion because the corrosion-
product fluorides are very stable,23 and the hydrogen would readily escape
from the system. Thus, for every water molecule that leaks into the cool-
ant, slightly less than one atom of metal will be corroded. Cleanup of the

coolant system after such a leak could be a significant task.

6.2,2,2 Leakage of coolant into steam. The situation here is similar

 

to that of the fluoroborate coolant., Very little is known about the influ-
ence of these fluorides in relation to stress-corrosion cracking and such
information would have to be obtained prior to specifying the materials for
an MSBR steam system.

6.2.2,3 Leaks to the cell atmosphere. Small amounts of coolant

 

flowing into the cell will react with moisture to produce BeO and HF., The

HF may cause minor corrosion of the metal that covers the cell wall.
51

6.3 Design Factors

6.3.1 COrrosivitz

Pure Na BeF, is not corrosive toward all major and virtually

1.3ul%0, 665%Fy

all minor constituents of structural metal alloys. Inadvertent intro-
duction of oxidants (steam leaks) or deliberate addition of oxidants (for
tritium control) will substantially increase the corrosivity of the cocl-
ant. The oxidizing effects of steafi leaks have already been discussed in
Section 6.2.2.1, As with fluoroborate, oxidizing agents may have to be
added to Nal.BHLiO.66

secondary coolant circuit., It will have to be determined if effective

BeFu to sequester tritium which diffuses into the

concentrations of these oxidizing additives can be attained in this salt

and if they are tolerable to materials of construction.

6.3.2 TFreezing point

N L BeF, has a liquidus temperature that may be 290°C (554°F)

#1.3470,66° 4
but could be as high as 340°C (644°F). The lower temperature is taken

from a published S'tudym+ of the Na,BeF,-Li

5 4 2BeF4 phase diagram, shown in

Fig. 6,1. The higher liquidus temperature is derived from a preliminary

diagram of NaF-LiF-BeF,, Fig. 6.2, in a compilation25 of phase diagrams.

2’
The composition chosen for this (NaF—LiF-BeFQ) coolant is based on the
liquidus temperature valley presented in Fig, 6.1, If the liquidus

. . ° o .
temperature of Nal.BHLlO.GBBeFu is actually 340°C (644°F) as shown in
Fig., 6.2, then a shift to a slightly different composition NaF—LiF—BeF2
(41-23-36 mole %) should provide a liquidus temperature of 328°C (622°F),

perhaps still low enough to operate a supercritical steam cycle without

. 7. . .
freezing the coolant salt. If a Naf- LlP—BeF2 mixture is to be chosen as
59

ORNL-DOWG. 75-1590
600

 

      
   

 

 
 

 

 

 

 

 

 

 

500
400 B-NosLi(Bel;)f Liq.
a'ta-
NoLiBeli + Liq
300 -
N03Li(BeF4)2
200 |_a-NayBef, + -
y 7-N03Li(Bel-;)2 . .
NoLi Bel; + anaeF4
100 y-Na Li(BeF ),
y-NaBeF+ +NalLiBeF,
7’-N<:3Li(BeF4)2
o ] ] ] 1
0o 20 40 60 80 100
Na,BeF, Mol. % LioBefy

Fig. 6.1. Phase diagram Na,BeF) -Li,BeF) .
ORNL-LR-DWG 16424R

BeF,
542

DOTTED LINES REPRESENT
INCOMPLETELY DEFINED
PHASE BOUNDARIES AND
ALKEMADE LINES TEMPERATURE {°C)

COMPOSITION (mole )

 

NaF-LiF 3BeF,

A~

280 Q\\ 370

e - 275 ¥Ry, 290 T
i > NaF - BeF,

"2 NoF - LiF -2 BeF, -~ 380

: 355 a4 A
- W T\ 345

=S, 400 350
o g *A/%" 0

   
   

356
(LiF - BaFy)
274

400

 

 

   

 

345
450 T W " 500
2 ‘ 550 (5 NaF-LiF -3 BeF,)
459 0 2
2 LIF-BeF, —_—— e (NoF , % 2 NoF-BeF, 320
500 asg 2 '2';';‘9"‘:? \ —‘ 59 570
600
50 485 ’ /2 650
600 700
650 750
800
700 D\ 850
750 900
950
800
LIF 80O 750 700 €649 700 750 800 850 200 950 NoF
844 980

Fig. 6.2. Phase diagram of LiF-NaF-BeF,, system.
54

the secondary MSBR coolant, there is an obvious need to refine the

phase diagram data in the composition region arcund 33 mole % BeFQ.

6.3.3 Heat-transfer and hydrodynamic properties

Heat-transfer properties are good by comparison with other molten-
salt cooclants. The thermal conductivity is estimated to be twice that
of eutectic NaBFu-NaF and the heat capacity per unit mass is about 20%
greater. The viscosity of Nal.auLiO.GeBeFu is about 10 times that of
eutectic NaBFq-NaF, but certainly still acceptable. This property has
not been measured for this specific mixture, but the data for the NaF-
and LiF-BeF2 binary sys‘cemsg6 suggest that the viscosity should be almost

the same as that of the MSRE coolant and thus should be a satisfactory

heat transport medium,

6.3.4 Vapor pressure and composition

 

The equilibrium vapor pressure should be very low. By combining

data in the NaF—BeF227 and LiF-BeF228

of 0.002 torr is obtained at 621°C (1150°T), the outlet temperature of the

systems, an estimated vapor pressure

primary heat exchangers. The vapor composition has not been determined,

but BeFQ(g) should be the dominant vapor species.

6.3.5 Radiation and chemical stability

Experience gained in the MSRE29 and in the extensive in-pile radiation
testing programso of similar molten salt strongly suggest that as long as
the temperature is kept greater than 150°C, radiolytic decomposition of

this salt will be of no importance to the corrosion of structural metals.
55

6.3.6 Tritium

The trapping of tritium in this coolant will likely require concen-
trations of oxidizing ions such as OH or Ni2+ which could also oxidize
metals that will be in contact with the coolant. It is knownBl that sub-
stantial concentrations of OH can be maintained in the LJ'.F—-BeF2 system
at 500-700°C (932-1292°F) and similar concentrations can be expected in

molten NaF-LiF-BeF,. The solubility of metal ion oxidants Ni2+ or Fe2+

2
may be limited because these ions are stabilized in complex, relatively
insoluble solids, as NaNiF3 or NaFeF3.23 The concentrations of hydroxide

and oxide in this coolant may also be constrained by the need to prevent
precipitation of UO2 in case coolant leaks into the fuel salt. As with
the fluoroborate secondary coolant, the basic question is: Can the ionic
species which traps tritium be stabilized at sufficiently high concentra-
tion without causing intolerable corrosive attack? Experimental investi-

gation will be required to answer this question.

6.3.7 Cost and availability

 

The inventory of secondary coolant given in the conceptual designal
would cost approximately $6 million, assuming unit prices18 of $120/kg
for 7Li and $16.50/kg for BeF2. The relatively high cost of the coolant
inventory is an incentive for minimizing the volume of salt required in
the coolant circuits through appropriate design. Cost as well as vis-
cosity could perhaps be reduced by substituting MgFQ, AlF3, or ZrF4 for
part of the BeF2.

Beryllium is not an abundant element in the earth's crust and world
demand, if maintained at the rate of 1968, would deplete world reserves

. 32 . .
in about 25 years. However, beryllium mineral resources are at least
56

1000 times greater than the reserves. If these resources are developed,
the supply of beryllium would be adequate for all molten-salt reactor

needs. Fortunately, the cost of the ore is a small part of the cost of
the metal and, thus, the price of beryllium may not rise appreciably as

these resources are developed.
7. ALTERNATE DUAL COOLANT CANDIDATES

In the dual coolant MSBR configuration the reactor and primary loop
containing the fuel salt are considered to be similar to those detailed in
the conceptual design.3 The fuel salt is circulated through the primary
heat exchangers where heat is transferred to the secondary coolant. The
secondary coolant is circulated to intermediate heat exchangers where heat
is transferred to a tertiary coolant, and the tertiary coolant is circu-
lated to the steam-raising system, The incentives for considering the
addition of a third loop to an MSBR include a better opportunity for tri-
tium trapping, addition of another passive barrier between the fission
product inventory in the reactor and the steam system, improved chemical
compatibility between adjacent loop fluids in the event of leaks, and, in
some cases, improved plant operability and maintainability. Financial
penalties are, of course, associated with the addition of a third loop in
both capital and operating costs.

Three dual coolant MSBR configurations are considered in this section
as alternate reactor concepts that could satisfy, to varying degrees, the
coolant criteria. Each employs a molten salt secondary coolant and in each
case Li BeF, has been selected as the molten salt (Section 7.1). This was
due primarily to the excellent compatibility between this material and the

fuel salt if they should become mixed. Also, LizBeF has an adequately

4
57

high freezing point so that the problems of fuel salt freeze-up are sub-
stantially diminished.

For the tertiary coolant, a gas, a different molten salt, a liquid
metal and also a novel fluidized bed heat exchanger are considered. The
reactor configuration utilizing helium gas (Section 7.2) appears to resolve
all technological problems and to satisfy all coolant criteria; however,
only at an expense due to the additional hardware required for the tertiary
loop, reduced thermal efficiency and increased pumping power requirements.
Molten carbonates are considered (Section 7.3) as the tertiary coolant, and
liquid sodium also has been evaluated (Section 7.4), but both seem less
attractive and are unlikely to be suitable tertiary coolants. The fluidized
bed heat exchanger concept (Section 7.5) would require substantial develop-
ment before it could be applied to a reactor system; for this reason, no

work on this concept is recommended at this time.

7.1 Secondary Coolant
Lithium beryllium fluoride, Li2BeFu, appears to be near ideal as a

secondary coolant in this configuration. No other fluid appeared to be a

serious candidate (Section 4.2.1).

7.1.1 Advantages

7.1.1.1 Compatibility with fuel salt. The secondary coolant,
LiQBeFu, is completely compatible with fuel salt under any condition or
event, such as primary heat exchanger leaks. No precipitation of fissile
material can occur nor are reaction product gases evolved. Indeed, no

reactions occur other than mixing of two miscible liquids. Leakage of

coolant into fuel salt only dilutes the uranium and thorium, which could
58

be easily corrected by removing fuel carrier salt in the fuel processing
facility or by the addition of more fissile material. No nuclear poisons
would be added to the fuel salt and no purification or special processing
would be required. Leakage of fuel salt into the secondary coolant cir-
cuit would require shut down and clean-up due to radioclogical contamina-
tion, but no precipitation or other reactions would occur. It might be
advantageous to maintain the pressure differential so that leakage in most
cases would be predominantly from the secondary coclant into the fuel,
Then, in the case of minor leaks it might not be necessary to shut down
and repair immediately.

7.1.1.2 Avoidance of fuel salt freezing. The LiQBeFl+ secondary

 

coolant freezes congruently at #459°C (858°F) while the fuel salt freezes
incongruently, the first solids, a thorium-rich phase, appear at the
liquidus temperature of 502°C (935°F) while the fuel salt is not com-
pletely frozen until the solidus temperature of ~u445°C (v833°F) is reached.
Thus the selection of LiQBqu as the secondary coolant substantially
diminishes the problem of fuel freezing (criterion 3.3.2) since it would
freeze above the fuel salt solidus temperature. If it were necessary, the
liquidus temperature of the secondary coolant could be increased to 510°C
(950°F) by using LiF-Bef,, (70-30 mole %). With this composition, the heat
transfer properties of the secondary coolant should diminish markedly as

the temperature falls to near the fuel salt freezing point and thus fuel

salt freeze up could probably be further delayed or completely prevented,

7.1.2 Disadvantage
The lithium in the secondary circuit would have to be compose of

_ 6. .
99,99+ % 7Li to prevent ruining the fuel salt by the addition of Li in
59

case coolant leaked into the primary circuit. The added costs due to the
7 - - -
Li would depend on factors such as the composition and volume of the
; . 7. . .
secondary coolant as well as the price of Li compounds, but are likely

to be several million dollars for an MSBR.

7.1.3 Future work

The future work needed to more accurately evaluate LizBeFu, or other
LiF-BeF, mixtures, as a secondary coolant involves (1) a better under-
standing of the fuel freezing problem, (2) accurate cost estimates for
7Li compounds in the quantities that would be needed in an MSBR, and (3)

consideration of design factors that would minimize the volume of salt in

the secondary circuit.

7.2 Compressed Helium Tertiary Coolant
Of the compressed gases that could be considered as tertiary coolants,
helium at about 700 psia appeared by far the most attractive. Considerable
experience has been accrued over the years with helium in various gas-
cooled reactors and much development work has been done in the HTGR pro-

gram. The primary advantage of helium is its chemical inertness.

7.2.1 Advantages

7.2.1.1 Tritium trapping. Generation of tritium in relatively large

 

amounts (approximately 2400 Ci per day) is unavoidable in a molten-salt
reactor.6’7 Virtually all of the tritium must be sequestered and removed
to prevent it from entering the steam system from which it is released to
the environment., The use of the helium loop may be the easiest and possi-
bly the least expensive way of removing tritium from an MSBR plant since

the tritium can be captured in the helium and removed from this helium
60

loop. Some preliminary calculations have been made on the quantities of
water and oxygen that must be added to the helium to trap the tritium.
Assuming that no protective film forms on the steam side of the steam
generators to limit hydrogen diffusion from the steam, approximately 100
standard liters of oxygen per hour would have to be added to the helium
tertiary ldbps to trap tritium as tritiated water and the entire volume
of helium would have to be processed to remove the tritiated water once
every four hours. If hydrogen diffusion from the steam system is reduced
due to formation of an oxide film, less oxygen addition would be required
but a little water would need to be added. This situation, which repre-
sents the most likely operating conditions, would result in a water con-
centration in the helium tertiary loop of the order of 10 ppm and an
oxygen concentration of about 0.1 ppm in order to lower the tritium loss
to the steam system and environment to approximately 2 Ci per day or less.

7.2.1,2 Reactor control. A nuclear reactor has a very fast response

 

to control parameters because of the nature of the fission process. When
there is a tight coupling of the reactor to the steam turbine load, for
example, by means of circulating salt circuits, careful design is required
to prevent control "hunting" when sudden load transients are imposed on

the system. Control is much easier on coupled systems when one of the sub-
systems has a much larger time constant than the other. The time constant
for the helium loop is very large compared to that of the reactor or the
secondary coolant loop because of the nature of the heat transfer in a

gas loop and this eases the control problem and facilitates stable
operation,

7.2,1.3 Compatibility with the steam system. Since it 1s certain

 
61

that there will be leaks in the steam generators during the 30-year life

of an MSBR plant, it is desirable that such leaks not result in harmful
corrosion of either the salt loop or the steam system. If the steam
generator is heated by helium, there is no incompatibility problem and no
corrosion difficulty aggravated by a steam leak. The addition of steam

to the moist helium would not result in significant attack of the materials
of construction of that loop and the tritiated water could be removed in

é processing side stream.

7.2,1.4 Steam system design. A helium loop can be combined with any

 

kind of steam system desired, By using a helium tertiary loop, steam con-
ditions can be chosen to optimize the steam system, although it appears
that a supercritical steam system is very desirable if not necessary for
a salt-steam generator because of freezing point considerations of the
salt. Steam system components would likely have a longer life expectancy
in the case of a helium tertiary loop than with salt-heated steam generators
since no additional corrosive materials are introduced to the steam system
by a leak.

7.2.,1.5 Utilization of existing steam generator technology. If a
helium tertiary loop is used in the MSBR plant, the same type of steam
generators could be used as have already been developed for gas-cooled
reactors. The exact units used for HTGR plants may not be directly appli-
cable to an MSBR plant, but certainly much of the development work
necessary for such a unit could be adapted. The same is true for the
helium circulators and other components of the helium loop. There remains
a considerable amount of development work required to design and prove a

steam generator to generate steam directly from molten salt; for example,
62

problems regarding thermal stress, especially in tube sheets, will require
resolution,

7.2,1,6 Simplified startup procedures, Solid-fueled, water-cooled

 

reactors can start at room temperature and the entire plant can be brought
to operating temperature on nuclear power; however, molten-salt systems,
particularly systems with salt-steam generators, require more extensive
considerations. Because of the high freezing points of the molten salts,
the salt as well as the salt loops must be brought up to témperatures of
approximately 540°C (filOOOOF) before the loops are filled with salt. The
steam system must also be brought up to operating temperature by an auxil-
lary steam source before the secondary salt can be introduced intoc the
steam generators. The fossil fired auxillary boiler must be of the order
of 10% of the thermal rating of the plant, or 225 MW(t), for molten-salt-
heated steam generator designs.

Startup of the plant would be simplified if there were a helium ter-
tiary loop since an indirect heating source would not be required. The
helium tertiary loop could be started up cold using a much smaller auxil-
lary boiler to bring up the steam system and heat could be fed back "up-
stream" into the salt secondary and primary loops of the MSBR plant, The
boiler would only be of sufficient size to overcome the heat losses in the
various systems and thus represent a significant capital savings. Auxiliary
electrical heat would still be required for the salt loops but the entire
startup procedure would become much less complicated.

T.2,1.7 Plant availability and maintainability. The addition of a

 

helium tertiary loop would increase the plant availability and simplify
maintenance, primarily due to elimination of most of the corrosion problems

associated with other MSBR configurations., Steam leaks in the steam raising
63

system, a common generating plant problem, would not lead to the formation
of corrosive materials in the secondary loop nor can toxic materials be
introduced into the steam system that would require extensive clean-up
before repairs could be started following shut down after a steam system

leak.

7.2.2 Disadvantages

While the advantages of adding a helium tertiary heat transfer loop
are real, there are also some disadvantages., These are mainly in the area
of capital costs and can be accurately assessed from an economic standpoint,
It is not so easy to quantify the offsetting economic advantages, primarily
in safety and ease of operation, which result from the addition of the
helium loop.

It is obvious that additional components would be required in the
helium tertiary loop. Helium ducting, helium circulators, and intermediate
heat exchangers would be required. The coolant salt inventory would be
somewhat greater and the steam generators would have more surface area to
transfer heat from the helium to steam.

Preliminary conceptual design work gave estimates that seem to indi-
cate the cost of adding a helium tertiary loop to an MSBR is not prohibi-
tive. One intermediate exchanger will be needed to transfer heat from the
secondary coolant salt to the helium for each secondary salt loop. A heat
transfer calculation showed that about 106,800 ft2 of surface area would
be required for heat transfer from the secondary salt to helium. In the
conceptual design3 the primary heat exchangers had a total surface area
of 54,252 ftz. These exchangers were to be made of Hastelloy N. The salt

to helium exchangers likely could be made of stainless steel. This
o4

cheaper material would make the cost per square foot somewhat less for
the stainless units, but since fabrication costs are a large fraction of
total cost, the saving would be less than the ratio of the cost of the
materials. Thus, this very preliminary estimate suggests that the inter-
mediate heat exchangers for use with the helium loop would likely add
about 4% to the total direct cost of the MSBR.

An additional item of capital cost are the helium circulators. Based
on HTGR cost figures33 for these items, approximately 3% would be added
to the capital cost. Because of the lower temperature of helium approxi-
mately 593°C (1100°F) instead of 815°C (1500°T) as in the HTGR, a larger
mass flow of helium would be required so that six circulators of HTGR size
would be required.

For the same reason (lower helium temperature), more steam generators
would be required for comparable power in the MSBR than are needed for the
HTGR.S3 By increasing the steam generator cost proportional to the in-
creased surface area required, the additional direct capital cost of the
helium-steam generators would be about 8%.

The 7Li2Be]F1\t in the secondary loop is more expensive than the NaF-NaBF,
melt considered in the conceptual design.3 It is roughly estimated that
the LJ'.2BeFL1L used would add about $13 million to the cost over that of
fluoroborate.

To summarize the capital cost estimates, it appears that the addition
of the helium tertiary loop would add approximately 15 to 20% to the total
plant capital cost. From this would be subtracted a credit for the reduced
size of the auxiliary steam system. Differences in plant contrel and plant

protective factors have not been defined and may add debits or credits to

the costs, In addition to the capital cost items, there is an operating
65

cost penalty in the power required to operate the steam turbine driven
helium circulators which require 73 megawatts; this would reduce the over-

all thermal efficiency from 44.,4% to about 42,7%.

7.2.3 TFuture work

In this preliminary evaluation of alternate coolants, it has not been
possible to do a thorough optimized plant design study. A more detailed
conceptual design study would be required to accurately asSess the operating
advantages and economic penalties associated with this MSBR configuration.
It is essential that this be done, because only dollars offer a direct
method of comparison of these coolant schemes.,

Although tritium trapping in moist helium appears certain, an experi-
mental demonstration of the sequestering of tritium in the helium stream
along with a positive method for its removal from the helium should be
carried out. The removal of tritium must be done at such a rate as to
limit the diffusion of tritium into the steam to a value of 2 Ci/day or
less. The 240,000 square feet of steam generator surface presents a very
large area through which tritium could diffuse and the diffusion process
will compete with any remdval system employed to remove tritium from the

helium,
7.3 Molten Salt Tertiary Coolant

In selecting an MSBR configuration utilizing two different molten
salts in the secondary and tertiary loops, it 1s desirable to choose a
secondary coolant that is compatible with the fuel salt and to choose a
tertiary coolant that can sequester tritium and is chemically compatible

with the steam system, If the secondary coolant is chosen to be
66

compatible with the fuel salt and the tertiary coolant is chosen to be
compatible with steam, an incompatibility problem between the secondary

and tertiary coolant frequently exists at the intermediate heat exchanger
in case of leaks. This problem eliminated many coolant combinations from
further consideration and remains a major disadvantage with the combination
of two molten salts described in this section.

The two molten salt coolants selected are Li2BeF for the secondary

I
coolant and a ternary carbonate eutectic melt, LiQCOS—Nazcoa—K2CO3
(42.5-30.6-26.8 mole %), For the tertiary coolant. The chemistry of car-

bonate melts has been recently reviex»redau'_'36 and suggests several potential

advantages over fluoride melts for application as a tertiary MSBR coolant,

7.3.1 Advantages

7.3.1,1 Tritium trapping. An appreciable concentration of hydroxide

 

ion, OH , can be dissolved in the ternary carbonate eutectic melt, thus
providing a large reservoir of hydrogen in the tertiary loop for isotopic
exchange with tritium. For example, at a COQ/HQO partial pressure ratio
of 10 in the vapor over the melt, the OH concentration in the melt would
be 2.8 x 107> mole fraction or 480 ppm. The OH concentration was cal-
culated for the equilibrium

20H (d) + o, (g) = 0032“(d) + H0(g)

where the equilibrium constant has been determined36 to be 31.6 at 600°C,
Hydroxide, or trapped OT , could be removed from the melt in a processing
side stream. For example, sparging with dry CO2 at 1 atm pressure would
lower the 0T concentration to 4.8 ppm, assuming an equilibrium partial

pressure of 10-5 atm of HQO.
67

In considering tritium trapping, it has been assumed that the ex-

change reactions

T,(g) + OH (d) = TH(g) + OT (d)

and

T,(g) + H,0(g) TH(g) + THO(g)
are rapid in this carbonate medium. This would have to be demonstrated
experimentally. Direct oxidation of tritium to TQO or THO seems unlikely.
7.3.1.2 Compatibility with the steam system. The freezing point of
the tertiary coolant (392°C, 738°F), is adequate for supercritical steam
cyéles utilizing steam reheat of the feedwater to prevent coolant freezing
in the feedwater heaters. Leaks of steam into the carbonate coolant would
lead to initial CO2 evolution and transient corrosivity due to the presence
of OH . However, this could probably be corrected by COQ. Thus steam
inleakage could be considered as helping to reduce trtium diffusion to the
steam system by isotopic dilution. Leakage of the carbonate tertiary

coolant into the steam system, however, could have serious consequences,

as discussed in Section 7.3.2.2.

7.3.2 Disadvantages

7.3.2.1 Incompatibility of secondary and tertiary coolants., Leaks
in the intermediate heat exchangers will lead to mixing of secondary
coolant (LiQBqu) and tertiary coolant (a ternary carbonate eutectic
melt), resulting in the evolution of gaseous CO2 and precipitation of BeO,

Li BeF, (£) + M,CO (£) = co,(g) + BeO(c) + 2 LiF(d) + 2 MF(d)

2 2 73
The equilibrium constant for the reaction
2+ 2-
BeO(c) + COz(g) - Be” (d) + Co, (a),
68

estimated to be about 8 x 10—3, indicates that dissolution of beryllium
oxide from the system after a leak would not be practical since it would
require excessive CO, overpressures, in the order of 100 atm (1470 psi).
Thus, mechanical removal of precipitated BeO from the heat exchangers
and loop could be necessary in some situations.

In order to protect the primary heat exchanger from damage due to
002 pressure and thus avoid the possibility of releasing radiocactivity,
the secondary coolant circuit and probably also the tertiary circuit
would have to include some pressure relief mechanism to relieve the Co,
evolved on leakage. Leakage of secondary coolant into the tertiary loop
would also lead to increased corrosion in the tertiary loop. The problems
associated with coping with the incompatibility of these two coolants
represent significant design consideration; however, they are, in prin-
ciple, simpler than those associated with primary heat exchanger problems
because no radiocactivity is involved.

7.3.2,2 Materials of construction. It would be necessary to develop

 

a material of construction for the tertiary loop. It is possible that
alloys high in nickel and chromium, such as Incoloy 800, could be utilized.
Stainless steel has been shown35 to resist corrosion at 600°C (1112°F)
through the mechanism of passivation with chromium oxide. High chromium-
high nickel alloys may also be potential candidates for development.37

A serious problem may also exist in the choice of materials for the steam
system since leakage of carbonate into the steam system could lead to
hydroxide stress-corrosion cracking of many conventional steam system
alloys.

7.3.2,3 Vapor pressure., The vapor pressure above the eutectic due
69

to decomposition of CO 2-(4) to COz(g) is quite low and increases only

3
slightly with added hydroxide. At an OH concentration of 480 ppm, which
would be typical for systems containing OH for isotopic exchange with
tritium, the 002 partial pressure would be about 0,4 mm Hg (0.008 psi)
assuming that the only source of water in the vapor phase is due to

2

20H" ZHQO + 0°7.

7.3.3 Future work

No future work seems warranted at this time to further evaluate the
ternary carbonate melt as a tertiary coolant. The most serious negative
factors are the chemical incompatibility between the secondary and ter-
tiary coolants and the present lack of a material capable of withstanding
the corrosive carbonate melt for 30 years., Also, physical properties of
this melt are not well known, particularly heat transport and transfer
properties such as thermal conductivity, heat capacity and viscosity. It
is possible that the ternary carbonate eutectic could be utilized as an
MSBR coolant, but considerable work would be required and it now appears

less attractive than other potential ceoolants.

7.4 Liquid Metal Tertiéry Coolants

The use of liquid metals as a heat transport medium offer several
significant advantages when compared to molten salts or gases. These
include excellent heat transfer properties which permit the design of
small, very efficient heat exchangers, low viscosity and, in the case of
sodium, a low freezing point which simplifies operation and start-up.
As discussed in Section 4, no liquid metal appears applicable as a sec-
ondary coolant for MSBRs due to chemical reactivity with the fuel in the

case of leaks 1in the primary heat exchanger. Advantage could be taken
70

of these desirable properties if liquid metals were utilized as a

tertiary coolant, however there are also serious disadvantages. These
include chemical reactivity with the steam or with the secondary coolant
in the event of leaks in steam generators or the intermediate heat ex-
changer, and lack of a solution to the tritium management problem. Sodium
was evaluated since considerable experience is being accumulated in the
LMFBR program with sodium as a heat transport medium. Considerable effort
is currently involved in the design of steam generators. Two types of
solutions to the chemical incompatibility problem are being considered.
The first involves the design of carefully welded '"leak free" heat ex-
changers coupled with the capability for rapid detection of leaks and shut
down followed by easy repair. Such steam generators should be readily
applicable to this MSBR configuration. The second approach involves the
design of a more complex doubled-wall steam generator with inert gas
purging of annular channels to diminish the possibility of sodium-water

intermixing in the event of a leak.

7.4.1 Advantages

7.4.1.1 Steam generators. The use of sodium as the tertiary coolant

 

might allow the adoption of LMFBR-type steam generators. Due to the
favorable heat transfer properties of liquid metals these steam generators
would be considerably smaller than those required for steam generation
from molten salts or gases.

7.4.1.2 Smaller intermediate heat exchangers. Due to the favorable

 

heat transfer properties of socdium, the intermediate heat exchanger could
be of significantly smaller size, and thus lower cost, than in the cases

with either gaseous or molten-salt tertiary coolant concepts,
71

7.4,1,3 Operational advantages. The low melting point and the com-

 

pact design of the steam generators allow considerable freedom for opti-

mization of steam conditions and flexibility in operating characteristics.

7.4.2 Disadvantages

7.4.2.1 Incompatibility of sodium and water. Sodium and water react
rapidly with the evolution of heat and potentially combustible hydrogen
gas. In considering sodium-steam generators for the MSBR, it is assumed
that adequate safeguards to prevent significant chemical events will have
been developed by the LMFBR program. These could take the form of leak-
free steam generators cr more complex double-walled steam generators with
provision for annular gas flow. Such chemical reactions represent an
area of concern which is absent in all the other MSBR configurations con-
sidered in the report and are clearly a negative factor in considering
the use of sodium.

7.4.2,2 Incompatibility of sodium and Li_BeF,. In the event of

2
leakage in the intermediate heat exchanger sodium will react rapidly

with the LiQBqu to form insocluble DLe metal and NaF., Clean up and repair
of the secondary coolant system will probably prove expensive in any case
and, if the Be metal alloys significantly with the structural metal of
the secondary coolant loop, replacement of that loop might be required.
Since the secondary coolant loop includes the primary heat exchanger

with the associated radicactivity, repair or replacement could be very

difficult and expensive.

7.4.2.3 No solution to the tritium management problem., The use of

 

sodium as a tertiary coolant may not offer a useful solution to the tri-

tium problem in the MSBR. In the LMFBR, which has a tritium problem
72

50- to 100-fold less than the MSBR, it is planned to remove tritium by
cold~-trapping as NaT. However, most metal tritides are relatively un-
stable at elevated temperatures and thus may not afford a means of se-
questering a significant amount of tritium for the MSBR.

7.4.2.4 Thermal shock. The excellent heat transport properties of
liquid metals introduce a concomitant problem: thermal shock to heat
exchange surfaces. This complicates the design, limits the operating
characteristics and makes off-design transient events a more serious
problem than with molten salt or gas tertiary coolants.

7.4.,2,5 Incompatibility with the cell atmosphere. Molten sodium if
released from a leak in the tertiary system into the containment cell,
would react with the oxygen and moisture in the cell atmosphere. To
avoid this, the tertiary loop cell atmosphere would have to be an inert
gas. Alternately, ancillary sYstems could be required to prevent or
contain fires due to sodium leakage at any point in the tertiary loop,
its drain tanks and other auxiliary equipment. Again, equipment developed
for the LMFBR may be adaptable to the MSBR,

7.4.2,6 Sodium containment. A transition from nickel-base alloys

 

for containment of the Li_BeF, secondary coolant to iron-base alloys for

274
containment of the sodium tertiary coolant is essential at the sodium-

salt interface. This is not a major disadvantage and it might also be

desirable to make such a transition for economic reasons,

7.4.3 TFuture work

The disadvantages associated with use of sodium in an MSBR in the

manner discussed appear sufficient to discourage additional consideration.

If the LMIBR project finds attractive solutions to some of these problems

4
73

and if alternate MSBR concepts appear less attractive in the future,

additional consideration could be given to sodium as a tertiary coolant.

7.5 Fluidized-Bed Tertiary System

One possible variation of the helium tertiary loop might make use
of fluidized-bed heat exchangers in the steam-raising system., Since the
fluidized beds would be isolated from the fuel salt by the LiQBeF1+
secondary coolant salt, design limitations and materials compatibility
requirements would be substantially reduced from those discussed earlier
(Sec. 4.1.4) for a single-coolant fluidized-bed system. In this appli-
cation single-vessel fluidized beds could be considered, with the sec-
ondary-coolant tube bundles and steam tubes in the same vessel. Poten-
tially adverse nuclear effects associated with intermixing of the fluid-
ized bed materials and fuel salt would require essentially no considera-
tion. The absence of large quantities of long-lived radiocactive materials
(except possibly tritium) from the secondary ccolant would permit the
use of more conventional maintenance methods for the fluidized beds.

However, tertiary fluidized beds would retain some of the potential
disadvantages of single-coolant fluidized beds. Tritium trapping and
removal probably would have to be accomplished in the fluidized-bed loops,
so appropriate consideration of tritium contamination would be required,
In addition, any adverse effects on bed performance due to salt leaks
would have to be shown to be managable.

The use of fluidized beds would greatly reduce the cperating economic
penalty of a gaseous third-loop coolant associated with coolant pumping
power, but it is not clear that major capital equipment savings would be

realized. While fluidized beds have certain attractive features, an
74

extensive development program would be required to bring them to the
point required for use in fission reactor systems. Possible future de-
velopments of fluidized beds should be considered for application in

MSBRs but no work appears to be warranted at this time.
8, ACKNOWLEDGEMENTS

The committee wishes to acknowledge contributions by W. K, Furlong
and L. Maya to many of the calculations and L. M. Ferris for editorial
assistance. Also, the secretarial assistance of Nancy Pope was greatly

appreciated.
10,

11,

12,

13.

1k,

75
9, REFERENCES

R. E. Thoma, Chemical Aspects §f MSRE Operations, ORNL-4658
(December 1971).

MSR Program Semiannu, Progr. Rep., Aug. 31, 1965, ORNL-3872, p. 135,
R. C. Robertson (ed.), Conceptual Design Study of a Single-Fluid

Molten-Salt Breeder Reactor, ORNL-4541 (June 1971),

 

L. E. McNeese and Staff of the Molten-~Salt Reactor Program, Program
Plan for Development of Molten-Salt Breeder Reactors, ORNL-5018
(December 1974), p. 5-47,

J. P, Sanders, A Review of Possible Choices for Secondary Coolants for

Molten Salt Reactors, ORNL-CF-71-8-10 (August 6, 1971).

 

R, B. Briggs, Reactor Technology 14, 335 (1971).
R. B. Briggs, Reactor Technology 15, 79 (1972),
Ebasco Services, Inc., 1000 MW(e) Molten Salt Breeder Reactor Conceptual

Design Study, Final Report - Task I, (Feb. 1872), p. 4-2,

 

E. G. Bohlmann, Heat Transfer Salt for High Temperature Steam Generation,
ORNL-TM-3777 (December 1972).

V. G. Selivanvov and V. V. Stender, Zhur. Neorg. Kim. 3(2), 448 (1958),
S. Cantor et al., Physical Properties of Molten-Salt Reactor Fuel,

Coolant and Flush Salts, ORNL-TM-2316 (Aug. 1968).

 

C. J. Barton et al., MSR Program Semiannu., Progr. Rep. Aug. 31, 1967,
ORNL-4181, p. 158.

S. Cantor, J. Nucl., Mater. 47, 177 (1973).

C. E. Bamberger, C. F, Baes, Jr., J. P. Young, and C. S, Sherer, MSR

Program Semiannu. Progr. Rep., Feb., 20, 1968, ORNL-4254, pp. 171-73,

 
15,

16,

17.

18.

19,

20,

21,

22,

23,

24,

25,

26,

27,

28,

29,

30,

76

D. M. Richardson and J. H. Shaffer, MSR Program Monthly Rep. for

Oct. 1969, MSR-69-106, pp. 18-20.

J. W. Koger, Corrosion and Mass Transfer Characteristics of NaBS”eNaF
(92-8 mole %) in Hastelloy-N, ORNL-TM-3866 (Oct. 1972).

¢. J. Barton, J. Inorg., Nucl, Chem. 33, 1946 (1973).

R. C. Robertson (ed.), Conceptual Design Study of a Single Fluid Molten
Salt Breeder Reactor, ORNL-4541, p. 80 (June 1971),

W. R. Huntley et al., MSR Program Semiannu. Progr. Rep. Aug. 31, 1971,
ORNL-4728, p. 152,

E. L. Compere et al., J. Nucl. Mater. 34, 97 (1970).

W. K. McCarty, USAEC Report NAA-SR-1999 (1957).

J. P. Young, J. B. Bates, and G. E. Boyd, MSR Program Semiannu. Progr.

Rep. Aug. 31, 1972, ORNL-4832, pp. 52-53,

 

C. E. Bamberger, B. F. Hitch, and C. F, Baes, J. Inorg. Nucl., Chem. 36,
543 (1974).

N. A, Toropov and I. L. Shchetnikova, Russ. J., Inorg. Chem. 2(8), 238
(1957). |

R, E, Thoma (ed.), Phase Diagrams of Nuclear Reactor Materials,
ORNL-2548, p. 42 (1959).

S. Cantor, W, T. Ward, and C. T. Moynihan, J. Chem. Phys. 50, 2874
(1969).

K. A. Sense and R. W, Stone, J. Phys. Chem, 62, 453 (1858).

S. Cantor, D. S. Hsu, and W. T. Ward, Reactor Chem, Div. Annu. Progr.

Rep. Dec. 31, 1965, ORNL-3913, pp. 2u4-26,

 

P. N, Haubenreich and J. R. Engel, Nucl. Appl. Tech. 8(2), 118 (1970).

W. R. Grimes, Chemical Research and Development for Molten Salt Breeder
31.

32,

33.

34,

35,

36.

37.

77

Reactors, ORNL-CF-66-7-41, pp. 45-55 (1966).

A, L. Mathews and C. F. Baes, Inorg. Chem. 7, 373 (1968),

M. J. Bell, Availability of Natural Resources for Molten-Salt Breeder

Reactors, ORNL-TM-3563 (Nov. 1971).

High Temperature Gas-Cooled Reactor Plant, vol. V, WASH-1230, United
Engineers and Construction, Inc., 1973,
K. H. Stern, High Temperature Properties and Decomposition of Inorganic

Salts Part 2: Carbonates, NSRDS-NBS 30, National Bureau of Standards,

 

Washington, D.C. (1969),

G. J. Janz, J. Chem. Educ, 44, 581 (1967).

N. Busson, S. Palons, J. Millet and R. Buret, Electrochim. Acta 12,
1609 (1967).

William P. Banks, Materials Protection 6, 37 (1967).
 
13.

15.
16.
17.
18.
19.
20.
21.
22.
23.
2k,
25.
26.
27.
28.

NS
\O

67~

-

mHNEREuE OO

?!E:m F‘H'Utflk>§ib‘?‘@ QS GGy

65.

66.

68.

69-172.

79

ORNL/TM-5325
UC-76

INTERNAL DISTRIBUTION

F. Apple 30.
. F. Baes 31.
E. Bamberger 32.
S. Bettis 33.
Braunstein 34,
A. Bredig 35.
B. Briggs 36.
R. Bronstein 37.
Brynestad 38.
Cantor 39.
L. Crowley Lo.
L. Culler hl.
M. Dale Lo,
H. DeVan L3-46.
P. Eatherly Lt,
R. Engel 48,
G. Fee 49,
E. Ferguson 50.
M. Ferris 51.
0. Gilpatrick 52.
R. Grimes 53.
G. Grindell 5h,
H. Guymon 55,
N. Haubenreich 56.
R. Hightower, Jr. 57-58.
F. Hitch 59.
W. Hoffman 60-62.
R. Huntley 63.
R. Keiser 6l,

. - -

MI? ; ? ;3fl(fi'filfi oo @ a Zo@e@On=E o

D.
A.

. RC
. L. Matthews, ERDA-ORO

L. Keller

D. Kelmers

I. Lundin

G. MacPherson, UT
E. MacPherson
Mamantov

L. Manning
Martin

Maya

T. Mays

. McCoy

. McDuffie
McNeese

. Neill

. Pollock
Postma

W. Rosenthal
D. Shults

J. Skinner
N. Smith
Spiewak

B. Trauger
M. Weinberg, ORAU

00 E A

Central Research Library
Document Reference Section
Laboratory Records

Laboratory Records

(LRD-RC)

ORNL Patent Office

EXTERNAL DISTRIBUTION

Research and Technical Support Division, Energy Research and
Development Administration, Oak Ridge Operations Office,
Post Office Box E, Oak Ridge, TN 37830.

Director, Reactor Division, Energy Research and Development
Administration, Oak Ridge Operations Office, Post Office Box E,

Osk Ridge, TN 37830.

Director, Division of Nuclear Research and Applications, Energy
Research and Development Administration, Washington, DC 20545

For distribution as shown in TID-4500 under UC-76, Molten-Salt

Reactor Technology.