ORNL-3791.txt

E
’,_
Zf
f
|
E

v

En

ORNL-3791
UC-80 — Reactor Technology
TID-4500 (46th ed.)

PRELIMINARY DESIGN STUDY QF A CONTINUOUS
FLUORINATION—VACUUM%DISTILLATION
SYSTEM FOR REGENERATING FUEL AND

FERTILE STREAMS IN A MOLTEN
SALT BREEDER REACTOR

C. D. Scott
W. L. Carter

RELEASED FOR ANNOUNCEMENT

IN NMUCLEAR SCIENCE ABSTRACTS

Top S

OAK RIDGE NATIONAL LABORATORY
operated by

UNION CARBIDE CORPORATION
for the

U.S. ATOMIC ENERGY COMMISSION

DISCLAIMER

This report was prepared as an account of work sponsored by an
agency of the United States Government. Neither the United States
Government nor any agency thereof, nor any of their employees,
makes any warranty, express or implied, or assumes any legal liability
or responsibility for the accuracy, completeness, or usefulness of any
information, apparatus, product, or process disclosed, or represents
that its use would not infringe privately owned rights. Reference
herein to any specific commercial product, process, or service by
trade name, trademark, manufacturer, or otherwise does not
necessarily constitute or imply its endorsement, recommendation, or
favoring by the United States Government or any agency thereof. The
views and opinions of authors expressed herein do not necessarily
state or reflect those of the United States Government or any agency
thereof.

DISCLAIMER

Portions of this document may be illegible in electronic image
products. Images are produced from the best available
original document.
ORNL~-3791

Contract No. W-7L05-eng-26

CHEMICAL TECHNOLOGY DIVISION

PRELIMINARY DESIGN STUDY OF A CONTINUOUS FLUORINATION-
VACUUM-DISTILLATION SYSTEM FOR REGENERATING FUEL AND
FERTILE STREAMS IN A MOLTEN SALT BREEDER REACTOR

C. D. Scott
W. L. Carter

JANUARY 1966

OAK RIDGE NATIONAL ILABORATORY
Oak Ridge, Tennessee
operated by
UNION CARBIDE CORPORATION
for the
U.8. ATOMIC ENERGY COMMISSION
Bons

iii

CONTENTS
Abstract =--ecmam et 1
Introduction wmeccmmm e e e - 3
The Molten Salt Breeder Reactor System --cemmeccmmcmm e e e o 5
Design Criteria —-----cmmmmmm e e e e 7
Basic Consideralions =-e-mmcmmmmm e e e e e e It
Process for the Fertile Stream m--c-cemmcm e e mm e e e e 9
Process for the Fuel Stream =ccememrcme o m e e e e e e e e e 9
Waste STOrage --mmmmm e e e e e e e e - 10
Operating POliCY ==mermm o s e e e e e e e e e e e 10
Process Dats —ceececm e e e — 11
Description of ProCess memememmcmcm o e e e e e e 11
Summary of the Process Flowsheet —=-em-mmemcommmcme e 16
FLlUuorination =--—== == oo o o e e 16
Purification of Uranium Hexafluoride by Sorption and Desorption --- 19
Vacuum Distillation ==mecmm e e e 25
Reduction of Uranium Hexafluoride and Reconstitution of the Fuel -- 28
Off -Gas Processing ==c-mmme e e e e e e 29
Waste Storage =--=-cmm oo e e e e e 29
Flow Control of the Salt Streams -—-—--cccmmmmcm e e e e e 35
Removal of Decay Heat --m--emmcm s cm oo e e e e e 36
Sampling of the Salt and Off-Gas Streams -----m;-—cmcccmrcc s e 36
Shielding, Maintenance, and Repalr of Equipment ------ceeeweeeoeou- 36
Materials of Construction =-meemecmemem e e e e 37
General Operating POliCy =me=cmcmmmmm e e e e e e e e e e e e 37
Process DesSign =—r-mm e e e e e e e e e 37
Fuel Stream -——c-ememmm o e e e 38
Fertile STream -—----cmcemm oo e e L9
Plant Design and Layoub =—=-c-mme oo e e 52
Cost Estimate —---romcmc e e 53
Process Equipment ==cecommmm e e e e e e e e e 53
Structure and IMprovements == --mce e m e c e e e 56
Interim Waste StOrage ———---em oo oo e e 56
Other Plant Costs =--remmeme e e e e 57
Total Fixed Capital Cost —--mmmmmem o e e 58
Direct Operating Cost —mmmm o e oo c e e e 58
Processing COBT ==mm e s o oo e e el 60
Conclusions and Recommendations ==—=mee—mmm oo oo oo - 62
Acknowledgement ===-em-m—m e e e e - 66
References ==---mmmm o e s e e e e ————————— 67

Appendix A. Design Calculations for Fuel Salt Fluorinator and
Cooling Tank m--mmc o e e e e e 73
Appendix B.

Appendix C.
Appendix D.

Appendix E.

Appendix F.

iv

Fission Product Heat Generation Rates in the Movable -

Bed Sorbers and NaF Waste Tanks ---scemecmccceccccmcmcman-— 78
Movable -Bed SOrber mmeemcmmccccc e e e e ————— 79
Sodium Fluoride Waste Contalners -—-meeccemccccmacmeeaa 79
Short-Term Cooling Station for Waste Sodium Fluoride-- 80
Interim Storage of Waste —weeemmcmmm oo 81

Estimation of Distillation Rate in Vacuum Still ~-weeeee- 83

Fission Product Accumulation and Heat Generation Rate

in Lithium Fluoride Pool in Vacuum Still --cceemmeemaaaan 89
Analytical Expression for Heat Generation Rate -=--w-- 90
Evaluation of Vacuum-5till Desigh «=-=mmccmccmmcmm e 95

Design Calculations for Waste-Storage System ----ceemeew- 99
Fuel Stream Waste System —cecmmmmmmmc e e 100
Fertile-Stream Waste System ==e-cwcccccmcc e 109

Physical -Property Data and Drawings ==e=ceceecceeoaceaaoo 110

PRELIMINARY DESIGN STUDY OF A CONTINUOUS FLUORINATION—
VACUUM -DISTILIATION SYSTEM FOR REGENERATING FUEL AND
FERTILE STREAMS IN A MOLTEN SALT BREEDER REACTOR

C. D. Scott
W. L. Carter

ABSTRACT

The purpose of this study was to make a preliminary
design and an engineering evaluation of a conceptual plant
for treating the fuel and fertile streams of a molten-saltl
breeder reactor. The primary requirements of the process
are to recover the unburned fuel (233UFM) and fuel -stream
carrier salts (LiFéBng% from the fuel stream, and the
LiF -ThF), plus the bred 33U from the fertile stream. Both
streams must be sufficiently decontaminated for attractive
breeding performance of the reactor. The plant was designed
to operate continuously as an integral part of the reactor
system, Titting into two relatively small cells adjacent to
the reactor cell. In this study, the plant capacity is
based upon treating 15 ft3/day of fuel salt and 105 ft3/day
of fertile salt removed as side streams. These capacities
are adequate for a 1000-Mw (electrical) power reactor.

As to the fuel stream, bvasically it is purified by
fluorination and vacuum distillation. The first step
removes uranium as volatile UFg; the second recovers the
Li¥ -BeFo by simultaneously volatilizing these two compo -
nents from the less volatile fission products. Fortunately,
the fission products so separated are primarily the rare
earths, which are the most serious neutron poisons. The
UFg from the fluorinator is accompanied by some volatile
fission product fluorides, primarily those of Mo, Te, Ru,
Zr, and Nb, which are removed by sorption on granular NaF
and MgFo,. Finally, the UFg is reduced to UF) by hydrogen
and recombined with the decontaminated LiF -BeFp carrier in
a single operation. Fission products are removed from the
plant by discard of NaF and MgF, sorbents and the still
residue, which is a highly concentrated solution of the rare
earths in Li¥F. Wastes are permanently stored underground.

With respect to the fertile stream, the process consists
only of fluorination followed by decontamination of the UFg
on NaF and MgFo sorbents. It is only necessary to remove the
bred ©33U sufficiently fast to keep 2 low concentration in

the blanket, thereby ensuring a low fission rate and negligible
poisoning by fission products. Discard of the barren fertile
stream at a slow rate suffices to keep the fission product
concentration at a tolerable level.

The chief conclusions of this study are: (1) that the
Tluorination-distillation process for the fuel stream and the
fluorination process for the fertile stream comprise a compact
and relatively simple system that can be engineered with a
normal amount of developmental work, and (2) that integration
of the processing plant into the reactor facility is both
feasible and economical and the logical way to take advantage
of the processing possibilities of a fluid-fueled reactor.
The nominal cost of this plant is presented in the following
summary of major items:

Process equipment and building space $5, 302, 000

Fuel salt inventory 89, 500
Fertile salt inventory 69,200
NaK coolant inventory 40, 000
Direct operating cost 788, 000 /year

These costs contribute about 0.2 mill/kwhr to the fuel cycle
cost when the reactor operates at an 80% plant factor and

~ capital charges are amortized at 10%/year. This cost is

sufficiently low to add to the incentive for developing the

molten-salt breeder reactor.

Some of the steps of the evaluated process are based on
well -established technology, whereas others are based on
extrapolations of laboratory and small -scale engineering data.
- Fluoride volatility and assoclated UFg decontamination by
- sorption are well -known operations, having been demonstrated
~in a pilot plant. However, vacuum distillation and liquid-
phase reduction of UFg to UF) have been demonstrated only at
the bench. Certainly, more development of these steps is
required for a complete process demonstration. A singularly
serious problem is the corrosive nature of the fluorine-molten
salt mixture in the fluorinator. However, this study shows
that this and other inherent processing problems can be solved
by proper design and operation of equipment. |
-3 -
INTRODUCTION

A reactor concept that has a high potential for economic production of
nuclear power and simultaneous breeding of fissile material on the thorium-
uranium fuel cycle is the molten-galt breeder reactor (MSBR). This reactor
utilizes two fluid streams. For this study, the stream compositions are:
(1) & fuel stream consisting of an LiF -BeF, (69-31 mole %) carrier that
contains the fissile component, and (2) a fertile stream, which surrounds
the fuel stream, consisting of a 71-29 mole % LiF -Th¥) mixture. In each
stream the lithium is about 99.995 at. % 7Lia The above compositions are

39

not unique; other MSEBR designs might use different compositions. The con-
figuration of the system allows a relatively high neutron leakage rate

from the fuel stream into the surrounding fertile stream, where capture

by thorium breeds additional 233U fuel. The reactor is operated at an
average temperature of about 650OC; fission energy is recovered in external
heat exchangers through which the fuel and fertile streams are circulated.
Sustained breeding performance of the MSER depends on the removal of
fission and corrosion products from the two fluid streams so that parasitic
neutron capture is kept to a tolerable rate. Portions of the two circu-
lating streams are continuously removed, processed for removal of fisslon
products, fortified with makeup Tissile and fertile materials, and returned
to the reactor in a cyclic operation.

A primary consideration of any process for recycling reactor fuel is
that minimal losses of all valuable fuel componente be attained without
intolerable capital investment cor operating expense. In the MSBR system,
this requirement applies to both fuel and carrier components. As mentioned

above, the fuel is 233U,

I

and one carrier component is the highly enriched
Ii isctope. The other major constituents, beryllium and thorium, are less
expensive; yet large losses of these cannot he tolerated either because of
the adverse effect on fuel ~-cycle cost and fuel conservation. The process
evaluated here accomplishes the obJjectives of conservation while providing
fission-product removal sufficient for a successful breeding system.
The work reported here is unique in that it examines a processing

plant integrated directly into the reactor system, which, in effect,

accomplishes on-stream processing. This method obviates the cumbersome
Lo

and expensive transfer of highly radiocactive material by carrier shipment;
furthermore, common use can be made of services and equipment necessary
to the reactor, thus avoiding the duplication that results in a separate
processing building. Also, since the spent fuel flows directly into the
processing plant, there is minimal out-of -pile inventory of valuable fuel
components.

Another interesting feature of this study is the use of the relatively
recent cdfiéept of wvacuum distillation as a means of purifying the carrier
salt. A modest vacuum of only 1 to 2 mm Hg is required, but the tempera -
ture (about lOOOOC) is higher than any normally encounteréd in handling
molten fluoride salts. The operation was explored first by Kel]_ey3 in
laboratory exPefiments that indicated LiFéBeFe decontamination from
fission products by factors of 102 to 103. The attractiveness of the
process lies in the fact that it involves only a physical operation that
is easily controlled and that can be made continuous. Fission products
can be concentrated in the still residue (primarily 7LiF) by a factor of
about 250 by using the decay heat of the fission products to volatilize
LiF—BeFE. This high concentration factor ensures a low discard rate for
the valuable 7L:'L. Cyclic operation of the still was assumed, allowing
the fission product concentration to increase with time. The corresponding
increase in the rate of decay heat generation limited the cycle time to
about 68 days. Although limited experience with the distillation step
indicates that high-nickel alloys are satisfactory structural materials
for use at this relatively high operating temperature, a more extensive
investigation is needed to define the design limitations.

A novel idea has also been studied in the evaluation of liquid-phase
reduction Qf UF6 to UFA by hydrogen. Initial bench-scale experimentse6
have,given promising data. The reaction is carried out by absorbing UF6
in a molten mixture of LiF—BeFQ—UFA at about 600°C followed by contacting
with H2 to reduce UF6 in situ. This technique avoilds the troublesome
problem of remotely handling solids (small UFA particles) that would be
net if the customary gas-phase reduction of UF6 were used.

Aside from indicating the feasibility and economy of the process,
this study also uncovered important design and engineering problems

assoclated with the scaleup of laboratory and batchwise operations to

4

o

_5_

larger, continuous ones. In this regard, recommendations are presented
at the end of the text, along with important conclusions. The most note-
worthy recommendation is that the key operations, vacuum distillation and
continuous fluorination, be given high priority in development.

The material that follows is arranged in this sequence: TFirst, a
brief description of the MSBR system is given to put the study in per-
spective; second, design criteria and ground rules are stated for each
phase of the study; third, a process flowsheet and a description of each
unit operation is presented; fourth and fifth, a description and pertinent
design data for each major component and the processing cells are listed;
sixth, equipment and building-cost data are presented; seventh, an over-
all evaluvation of the process is given in a set of enumerated conclusions
and recommendations arranged according to plant characteristics and
process operations. Six Appendices, giving detailed data and calculations,

are attached.
THE MOLTEN-SALT BREEDER REACTOR SYSTEM

The processing system of this study is designed to meet the
requirements of the molten-salt breeder reactor shown in Fig. 1. This
is a conceptual designl of a power reactor capable of producing 1000 Mw
(electrical) with a thermal efficiency of 45%. Basically, it consists
of a graphite matrix enclosed in a cylindrical Hastelloy N vessel for
containment. GCraphite occupies about 79 vol % of the core, fuel salt
about 15 vol %, and fertile salt about 6. The flow passages are such
that the fuel and fertile streams do not mix. The core is surrounded
radially and axially by a 3.5-ft blanket of LiF-ThFM mixture, and the
blanket is in turn surrounded by a 6-in.-thick graphite reflector. The
core is about 8 ft in diameter and 17 ft high; overall, the reactor plus
breeder blanket is about 16 ft in diameter and 25 ft high.

Fission energy is recovered in a battery of external heat exchangers
through which the fuel and fertile streams are continuously circulated.
The coolant may be either a molten carbonate or fluoride salt mixture
which transports the heat to boilers for producing steam. ©OSmall side-

streams of fuel and fertile fluids are continucusly withdrawn from the
ORNL DWG 65-3017

_ COOLING
FUEL _ ' rju"’S-SA‘\LT
PUMP(4eq)

BLANKET FUELBANKE —gp

PUMP{(2ea) — 3 it

o ;
CONTROL

ROD{leq)
7

REFLECTOR: :

Pb SEAL-
EXPANSION
JOIN

s
1

4 /T-\‘\ :
TR Il
1 . INT / % ] (.‘_,
: MODERATOR— | : _ _ _ |
_ R uu N : !
FUEL HEAT § '
! L+ EXCRANGER ' ' §§ 1
, CHANGER N H
IR =2 v
' - A : .
| WA BE e NAATT ;QJV}[
\ific——‘/ o LS - /‘I}{’
N - — —
oLAN ST 5 Few — | BLANKET HEAT
EXCHANGER
(2eq)

SECTION "A-A"

Fig. 1. Conceptual Design of the Molten-Salt Breeder Reactor.

-
._7...

circulation loops and routed to a chemical processing cell adjacent to the
reactor cell. After being processed for fission product removal and
reconstituted with makeup materials, the two fluids are returned to the
reactor via the fuel-makeup system. The processing cycle is selected to
give the optimum combination of fuel-cycle cost and breeding gain.

The data presented in Table 1 are typical for the MSBR and were the
bases for this study.2 Since the reactor concept is undergoing engineer-
ing and physics evaluation, these data represent no fixed design and are

subject to change as the studies progress.
DESIGN CRITERTA

The following discussion delineates ground rules and arguments for
the particular choice of process and design used in this evaluation.
Choices were made on the basis of existing knowledge and data. The study
presented here is expected to verify basic assumptions or indicate

Jjudicious alternatives.
Basic Considerations

One basic consideration concerns the fuel yield (the fraction of
fissile inventory bred per year) which, for a breeder reactor, is inversely
proportional to the total inventory of the reactor and chemical plant
systems. This characteristic is essential to the design of an MSBR pro-
cessing plant and suggests close-coupling of the reactor and processing
plant to give minimal out-of -reactor inventory. A fluid-fueled system
is readily amenable to this type of operation, and for this evaluation
the processing plant is integrated with the reactor plant. This design
permits fast, continuous processing, restricted only by the rather
stringent design requirements for fission-product decay-heat removal and
corrosion resistance.

The integrated plant occupies cells adjacent to the reactor cell, and
all services available to the reactor are available to the chemical plant.
Thegse include mechanical equipment, compressed gases, heating and ventilat-

ing equipment, electricity, etc. The cost savings for an integrated
-8 -

Table 1. Typical Characteristics of & Molten Salt Breeder Reactor

General

Reactor power, Mw (electrical)
Thermodynamic efficiency, %

Reactor geometry

Core diameter, ft

Core height, ft

Blanket thickness, ft

Moderator

Volume fraction of moderator in core
Volume fraction of fuel in core
Volume fraction of fertile stream in core
Reactor containment vessel

Fraction of fissions in fuel strean
Plant factor

Breeding ratio

Fuel Stream

Composition, mole %
LiF (99.995 at. % TLi)
BGFE
UF), (fissile)
Inventory at_equilibrium
Volunme, £t
LiF, kg
BeFo, kg
233y (as UFY), kg
235U (as UF),), kg
Other U (as UF)y), kg
Cycle time, days
Power, Mw (thermal)
Liquidus temperature, C
Density (calculated)

o (g/em3) = 2.191 - 0.000k t (°C) for 525 < t < 1200°C

Fertile Stream

Composition, mole %
LiF (99.995 at. % [Li)
ThE),
Inventory at equilibrium
Volume, ft3
LiF, kg
Th (as ThF)), kg
233y (as Ugu), kg
233pa (as PaF)), kg
Cycle time, days
Power, Mw (thermal)
Liquidus temperature, C
Density (calculated)

o (g/em3) = 4.993 - 0.000775 £ (°C) for 565 < t < 1200°C

Steam Conditions

Pressure, psia
Temperature, I
Condenser pressure, in. Hg abs

®

1000

L5
Cylinder
8

Ly

3.5
Graphite
0.79
0.15
0.06
INOR -8
0.972
0.8
~1.08

68.5%
31.22
0.31

3515
1000
1.5

Fpasic composition of carrier salt is 69-31 mole % (LiF-Bng).

Equilibrium composition for this cycle.

-9 -

facility are immediately apparent when one considers the large amount of
equipment and facility duplication required for separate plants.

A further basic consideration is that there will be no large
extrapolation of technology in the process design. Accordingly, the
process is based on treating the molten salt by fluorination and distilla-
tion, with the supporting operations of UF6 sorption on and desorption
from beds of pelletized NaF, followed by reduction of the UF6 to UFM' A
large amount of dats is available for each step except for the distilla-
tion and reduction operations, both of which have been demonstrated in

the laboratory.3’26

However, this study does assume the necessary
engineering extrapolations to convert from the current batchwise operations

to continuous operations.
Process for the Fertile Strean

The two streams of the bfeeder reactor require different processing
rates and must be treated separately to prevent cross contamination. The
first step in process for the fertile stream consists only of continuous
fluorination, which removes the bred uranium as the volatile hexafluoride.
No other treatment is needed if this step is designed to maintain a low
uranium concentration. To accomplish this, the stream is required to go
through the processing plant on a relatively short cycle, for example,
once every 20 to 50 days. The cycle time for this study is 22 days. The
fission rate in the blanket is low, and the fission products are kept at
a tolerable level by periodic discard of barren LiF-ThFM salt. A 30-year
discard cycle suffices. 1In the second step, the volatilized UF6 is sorbed

on NaF beds, desorbed, and finally caught in cold traps.
Process for the Fuel Stream

The fuel stream of the reactor is processed by fluorination and
vacuum distillation to recover both uranium and carrier salt sufficiently
decontaminated of fission products. A cycle time of 40 to 7O days is
required to maintain the fission-product concentration at a low enough

level for attractive breeding performance. The calculations of this study
- 10 =~

are for a 58-day cycle. The UF6 is recovered by NaF sorption and cold
traps, Just as for the fertile stream. Decontaminated fuel and carrier
are recombined in a reduction step that converts UF6 to UFM and. further
purifies by reducing corrosion products (iron, nickel, chromium) to their
metallic states. Makeup fuel and carrier are added at this point, and
the stream is returned to the reactor.

The time that the fuel stream spends in the processing plant is kept

as short as practicable to minimize out-of -reactor inventory.
Waste Storage

The chemical plant provides its own storage system for process
wastes. Incidental wastes, such as slightly contaminated aqueous solutions
and flush salts, are assumed to be handled by the reactor waste system,
thus such facilities are not duplicated for the processing plant. OSeparate
storage is provided for fuel-and fertile-stream wastes, which are primarily
LiF plus fission products, and Li¥F -ThF,, respectively. The facilities are
designed for a 30-year capacity and afe located underground a short
distance from the chemical processing area.

The fuel -stream process also produces a less radiocactive waste than the
LiF~fission-product mixture. This waste is in the form of pellets of sodium
fluoride and magnesium fluoride pellets used for decontaminating uranium hexa -

fluoride. Interim storage of 5-year duration is provided for these solids.

Fission-product decay heat is removed either by forced air or natural

convection, as required by the heat load.
Operating Policy

Certain ground rules consistent with convenlent and safe operation
were adopted for this study. Maintenance operations are facilitated by
assigning unit-process steps to either a high- or a low-radiation level
cell. Operating and maintenance personnel, who are not required on a full -
time basis, are to be shared with reactor operation. No water or agueous
solutions are to be admitted to The process cells; fluids required for heat ‘I’

transport will be either air or sodium-potassium eutectic (NaK). A barren
- 11 -

fluoride salt, for example, an NaF -KF mixture, would be an acceptable

substitute.
Process Data

The primary concern of a processing cycle for short-cooled fluoride
mixtures is in dissipating fission product decay energy so that process
operations can be controlled. A maximum of about 6,5% of the total energy
of the system is associated with beta and gamma energy in the fission
products; this amounts to about 140 Mw (thermal) in the MSBR fuel stream.
Most of this energy i1s emitted quickly, decreasing about 82% in 1 hr and
95% in 1 day. The data for this study were calculated for the reactor
described in Table 1, using the PHOEBE Code, which computes gross fission -
product decay energy as a function of exposure and time after discharge
from the reactor. The data are presented graphically in Figs. 2a-b and
3a-b for fuel and fertile streams, respectively.

The graphs give an upper limit for heat generation because the
calculations do not account for possible intermittent reactor operation
attributable to the 80% plant factor. In addition, the graphs include
decay energy assoclated with gaseous products that are sparged in the
reactor circulating loop and with those fission products that might
depos;t on surfaces throughout the system. It is difficult to separate
this energy from gross energy until more is known about the behavior of
fission products in molten fluorides.

A process flowsheet showing material balances for the fuel and fertile

streams in the processing plant is included in Appendix F.
DESCRIPTION OF PROCESS

The processing facility must have the capability of removing the major
portion of fission products from the molten fuel salt and returning the
purified salt to the fuel system after necessary reconstitution with 233U
and carrier salts. As to blanket-salt processing, the facility must achieve
recovery of the major portion of the bred uranium for recycle to the fuel
‘stream or for sale. These goals can be met with present technology or with

processes that can be reasonably extrapolated from current technology.
ORNL DWG 65-2987 R1

A ER SRR IR

I 1

i
i

10 SEC

w0 W oW @ ™ <

(g, ™8) 31V¥ NOILV¥INIO 1¥3H

o oy

2

"

TIME AFTER DISCHARGE FROM REACTOR (days)

in'MSBR Fuel Sfiream.

i o
s
Md
@]
(TGN
O
g
op~
a0
@
=}
O
=3
O
—~
3
=
=
&
RE
Ay IRy
T =
=
>y
g O
O\G
O i
—Qd
© o
o
o O
& Oy
5
jien]
5
-lf
[43]
0w
o D
B \O
QW&,
&8
&0
o
(&

stream vol
ORNL DWG 65-2988R1

6 7 891

T

=

Tt
rim ]

Fr T

25

B

5

MANRRE]

30

-~
£

i
1

)
il

5 6 7 9]04

?
|
it

e
A
@
T
3 4

5

G0N © oW B o~

n,c—u._: nig) 3LV NOILV¥INIO LVIH

102
10!

TIME AFTER DISCHARGE FROM REACTOR (days)

Fuel -

MSBR Fuel Stream.

in

a

y Heat
; pover, 2160 Mw (thermal); cycle time, 58 days.

Fission ~Product Deca

671 £13

Fig. 2b.
stream volume,
7 891

ORNL DWG 65-2989 R1

T

6 7'-':_;.9 1 -

3

[
R e

Fertile-

| ile-sfiream;~_
‘cycle time, 22 days.

AR AR LR

ot

TIME AFTER DISCHARGE FROM REACTOR {(days)

5

6 7 8

5

i

1

il

v Heat in MSBR Fert

power, 62 Mw (thermal)

e
2

.

AR IRIRANARSL]

Fission-Product Deca

1783 £t3

]

3a.

stream volume,

oL
3

T

M

Fig

R

il

TIHL

b0 seC

A_m,a?_f nig) 3LVY NOILY¥INO LV3H
HEAT GENERATION RATE (Btu hr-]fl_s)

10

ORNL DWG 65-2986 R1
6 7891

9 EEEag i _: : ==asa8ss = = £ =2 == BEaT T
8 i L" :::;:— st 53t == =gt sieed & = = z .
7. == Eas o H S5 =S + ; = o eI SeSsR st Tk B S If‘ ‘
= f T & T T . 5L "u‘ E t = T T T - |
6 : ' ' : . = =
i wal — II T - #
i t ] S i i
1 ANNETY r -
I I | 1 Hil I
4 H L N
=== H o S :A:::{;}tf:;;_ - H ve @ = ses H :
= 1 1 : e ni
== S
= ”. R - i
it 1 UREAD LAY 1o e B
i i S : ;
2 -1 Tr I { i i f Il ()
| T 4 I L i i
Eh.Y 1l T 1
T 1 I 1M
| ir ' T
EaREN i i i AR R
e | i ! HHEREH
]0]1 . i 1 I | A
9. i £ ‘ ESEE; : = = HE
8 B : = 5
= = N =i e o B SEmaT Es EEHEE HH H
7 s = f SN R auaaa it T a1 ) o o o o 1 jEa s o T
8 ik t ] ! = i :
Eheus + g : T
FrifE ! — b i s :
5 & S | T
ANER BN P T Bais b HARHE FEL e e 1 1 T
It T AESR REN 1 s : Hr
4 T T T I I H F Tr 1
- 1 Yol == = T e H
e : = e amnoSesee : = &
= == : : ERs s 2 = Ho t 3 T
t T T i T T .7 T b 1
3 E 2 L =  FEERRLE R ERSREER Nt Lk ik ; :
=5s: = EaaEsn T S i i T :
e s " i e I Hi
i HHHHE T
” pheit o i
it it i
Hit i e
i r ¥ ! 1] I 1
! ‘;é‘ . i ] |
i [T TR i i k
] = I I R i
0 i CHI | i | L _1 i}
100} R T _ 1071 SRR E :
8 = = =i £ : : 2 ==
=== e SSses k A 22t -
T t T H asdsins b T 11
6 ' i , = Thi=es i i
5 - it e it T i £f
I l‘ll ‘I f I
4 T } Hh : I ‘ ] } 1 M
E e T i 8 I e S : . L
= = :
3ifis rf e ? t +
3 = i e : ot .
i ! i it
= : HHER !
FE R i - f
2 iy L; : B -
i 1 T Rd i | i M i
1 H HEsIHH 1
: ] | I | |||1 i ] ]|]| i
HH HH
fl i LI { T ) Y \q_'_rr . 30 5 !,] T L
101 E A 1 R A ARSI Hulnflfl 'rmmmmmmmfl*mrrm 1 mmmmmmmmnmmmmu VA0 840 RS RURRTARRD
3 4 € 8 9]02 5 6 7 89 5 2 4 7809

10

Fig. 3b.

stream volume, 1783 ft3;

103 6 78 Yt

TIME AFTER DISCHARGE FROM REACTOR (days)

Fission-Product Decay Heat in MSBER Fertile Stream.

Fertile-

power, 62 Mw (thermal); cycle time, 22 days.

<1 -
'Sfimmary of-the_Process Flowsheefi

In Fig. b4 it can be seen that fuel—stream proce851ng is achleved by

233

first removing the uranium (as UF ) and volatile fission products from
a side-stream of the molten salt by fluorlnatlon. Then the carrier salt
is vacuum-distilled frbm the residual fission products. Next, the UF6,
further purified by a sorption-desorption process based on the use of NaF
pellets, is dissolved in the carrier salt. TFinally, the UF6 is reduced
to UFA by hydrogen, thereby reconstituting the fuel salt.

Blanket salt is processed concurrently with the fuel salt by
fluorinating a side-stream. The UF6 gas 1is separated from the volatile
fission-product fluorides by sorption and desorption, using beds of NaF
pellets as mentioned above. The blanket salt needs no further purification,

233

and a portion of the UF6 is sent to the fuel stream by dissolving it in
the carrier salt and reducing it with hydrogen. The excess over that
needed to refuel the reactor is sold.

The two chemical reactions (fluorination and reduction) in the
process are simple, fast, and quantitative. The other interactions are
physical and require only heat and mass transport; however, in the case
of distillation, a rather high temperature (about lOOOoC) is involved.

The salt is extremely stable at any temperature anticipated in this
process, and other physical properties, primarily vapor pressure and

solubility for fission products, are in accord with process requirements.
Fluorination

As noted above, uranium and volatile fission products are removed
from both streams by fluorination. A batchwise, molten-salt fluoride -
volatility pfocess for recovering uranium has been under development for
several years. Currently the process is in the pilofi plant phase at
ORNL.lY’18 One of the steps in this process is batch fluorination of the
molten fluorides to remove uranium hexafluoride. As noted before, uranium
~in these streams is in the form of UFl¥ | This uranium and some»of the
flSSlon products are converted by fluorlne ‘to higher-valent, volatile

- fluorldes Whlch leave the salt and go to the uranium hexafluorlde N

ABSORPTION CoLD
TRAP

ORNL DWG &5-3016

NON CONDENSABLE
GAS TO STACK

SORBED FP
IN NaF WASTE

L BLANKET SALT REACTOR

FPUFg,Fy

|

|
]
L= ¢
/2]
._
ud
FLLUORINATION %
3
m
[}
u
&
o
-
0.
SALT
WASTE
STREAM

DISTILLATION

SORBED FP
iN NaF WASTE

FLUORINATION

BARREN SALT

STILL
BOTTOMS
TO WASTE

ot

H2 HF, FP

Fz FP

OFF~GAS
TREATMENT

CAUSTIC
SCRUBBER

PURIFIED FUEL SALT

CAUSTIC
WASTE

I}

CAUSTIC
MAKE-UP

< FUEL SALT

]

Fig. 4. Major Steps in MSBR Fuel and Blanket Processing.
separated from blanket and fuel salts by fluorination (blocks 1 and 2),

further purified by sorption, and collected in cold traps.

MAKE—UP

Uranium is

The barren

fuel ~carrier salt is distilled away from the bulk of the fission products
(block 3). The cold-trapped UFg and fuel-carrier salt are recombined in
reduction step (block 4) and returned to the core.

LT -
purification system. The reaction'for_producing UF6'in the present_batch
processing 1ls the same as that_for.the proposed continuous system and may

be represented as follows:
flFh + F2 w>GF6 .

Certaln fission- product fluorldes are also volatile. The principal
ones are: Ru, Nb, Cs, Mo, and_Te. Z;rconlum is volatilized to a lessér
degree. RIS o -

o Contlnuous proce881ng of the MSBR fuel salt can be best achleved 1f
the flu@rlnatlon process 15 made contlnuous, preferably W1th counter- :

. current contactlng of the molten salt and fluorlne. Such a_process_has

Vnot'béefi_deyéloped;.and the @resefit_volatility process at ORNL is entirely
a a batch proééss. Tberefore} dévelopment work will be necessary to provide
for continuous fluorination. - o o

_ Rate data for molten—salt fluorlnatlom are limited and conflicting,
. aithough the reported rates_have been sufficient for the batch_process,

19

-Mallen 8 data bn the flubrination'of falling'droplets of molten salt
support the view that the reaction itself is very fast; whereas, the

. slower rates that result from sparging a pool of molten salt with fluorine
.(Cathers_gz_§£:,20 Pltt,gl and Moncr1ef2 ) can be explained by assuming a
massatransfer—controlling rate mechanism due to inefficient mixing of the
_gas_and liquid phases. Contact times (fluorine and molten salt) of 1 to
2 hf:have been.shown to be adequate for uranium removal down to 10 ppm in
su¢h'batch tests. The countercurrent, continuous operation, énvisioned
here, wéul@ probably give better contact. It is difficult to predict.the
fluofiné'utilization in a continuous fluorinator; however, it should be
better'than that for batch fluorination, which suggests that a utilization
of . 33-1/3% will probably be achieved.

A relatlvely low mass Tlow rate of gas must be malntalned in the
fluorinator to prevent salt entralnment in the off-gas. The highest mass
flow rate that has been used successfully w1thout enfiralnment 1s 0. 28 Slpm
per in. 2 of fluorlnator cross sectlon 23 |
Experience in the Volatility Pilot Plant at ORNL showed that
- corrosion on bare metal walls (L-nickel) in the_fluorinator.1s_relatively

'fhigb. Kessie §§_§£:2M-developed_a technique for keeping a frozen
-19 -

protective layer of salt on the metal wall, and we adopted their approach
for this study. This frozen wall of salt, which is kept so by a high
cooling rate on the metal wall, prevents gross corrosion. Internal
heating keeps the rest of the salt molten. For the fuel stream, ficsion-
product heat is sufficient for this purpose, but for the blanket stream
it may be necessary to supplement decay heat by using suspended electrodes
for resistance heating. Since a maximum of 1.5 kw of heat per ftg of wall
area can be removed from such a system, the fuel stream will have to be
delayed in a cooling tank until the heat generation due to decay heat is
at a sufficiently low level. As to the blanket salt, decay heating will
be ingignificant.

The continuocus fluorinator is a tall column into which molten salt
is fed at the top and flows to the bottom; fluorine is introduced at the
bottom and passes to the top, accompanied by thé volatile fluorides (Fig.
An expanded deentrainment section is added to the top, and the body of the
column is Jacketed with a coolant to maintain the frozen salt wall. A
gravity leg is used in the molten salt outlet to hold a constant salt leve

in the column.
Purification of Uranium Hexafluoride by Sorption and Desorption

The UF6 from both streams is purified in the same way before being
returned to the fuel stream. OSince the UF6 that leaves the fluorinator
contains volatile fission products, it is purified by a series of sorption
steps. These are batch steps, but the process is made continuocus by
using parallel beds alternately.

The first separation occurs in a NaF absorption system where the gas
stream passes through fixed beds of NaF pellets. This system consists of
two distinct zones, one held at L00°C and one at about 100°C. In the
higher -temperature zone, corrosion products, entrained salt, and niobium
and zirconium are irreversibly removed from the fluorinator off-gas while
the UF, and other fission products pass through (Fig. 6). In the second
NaF zone (lOOOC), UF6 and some of the molybdenum are held up by sorption
while the remaining fission products pass through. At this point, the

UF6 has been separated from all the fission products except molybdenunm.

5)-

1
- 20 -

ORNL DWG 65-3037

Fig. 5. Continuous Fluorination. NaK coolant, flowing through the
jacket, freezes a layer of salt on the inner surface of the column, thus
protecting the Alloy 79-4 from corrosive attack by the molten-salt—fluorine ‘
mixture.
F2, Ru, Te, Cs

F2, UFg, Mo

100°C DURING
SORPTION, < NaF
150°C DURING ABSORBER

DESORPTION

\ 400°C

FLUORINATOR ey
OFF-GAS OR Fy )

v

NaF WASTE
(Nb,Zr)

Fig. 6.
Fission Products.
one for the fertile.

¢

ORNL DWG 65-3015

F2 < Jl NaF I¢
UF, TRAP
SECONDARY
COLD TRAP,
prk ~60°C
ABSORBER PRIMARY COLD
(150°C) TRAP, -40°C
UFg, Fo o
N
;_I
¥
MgF2 UFg TO
WASTE REDUCTION UNIT
(Mo)

Purification System for UF6 with Disposition of Volatile
Two such systems are used, one for the fuel stream and
- 20

Molybdenum fluoride is removed from the UF6 by isolating the 100°¢ ‘
NeF zone, desorbing the UFg and molybdenum fluoride (by raising the
temperature to >lBOOC and passing fluorine through the bed), and passing
desorbed gas through a fixed bed of MgF,, which is held at about 150°¢C.

The MgF, sorbs the molybdenum but allows the purified UF, to pass through
into the cold traps.

The NaF pellets used in the high-temperature zone must be replenished
periodically since they accumulate fission products and corrosion products.
This discafd constitutes one of the waste streams. The two NéF sorption
zones may be integrated into a single unit, one zone on the other, and, as
Nal' is discharged from the lower zone, it can be replenished by Nal' from
the low-temperature zone, which is in turn fed with fresh NaF pellets.

Such a system has been used effectively in pilot plant operations, and a

similar system would be desirable for the MSBR processing facility.l7’18
In our concept, the system has a movable bed of NaF pellets, and, after

each sorption cycle, some of the lower NaF is mechanically ejected to

waste. Annular design with alr cooling would probably be necessary to

allow removal of fission product heat (Fig. 7).

As the UF6 leaves the last sorption trap it must be collected and
ultimately used as feed for the fuel -reconstitution step. This is done
by collecting the UF6 in cold traps. Two cold traps are connected in
series. The first, or primary trap, is operated at about -hOOC, and the
second, or backup trap, is operated at about -6OOC. The principal design
consideration is the heat transfer rate. Conventional designs are avail-
able for such units in which there are internal cooling fins for collecting
UFg, and Calrod heaters for vaporizing the UFg for removal (Fig. 8).
Uranium hexafluoride from the heated cold trap is fed directly to the
reduction process. This calls for at least three primary cold traps for
continuous operation: one for collecting UF6, one for feeding UF6, and.
one in transition between these two functions.

A fixed bed of NaF at ambient temperature very effectively removes
trace amounts of UF6 from gas streams. Such beds are used as backup UF6
traps in the fluorine exhaust from the cold traps and in other process
streams that might; contain UF6. Uranium is recovered from such traps by .

using the NaF as charge material for the main absorption beds.
- 23 -

ORNL~LR~DWG B50451R-3
YNOF CHARGING CHUTE

1Y5-in. NPS, SCHED -40 ( AIR COOL-
ING AND THERMOGOUPLES )

TO UFg TRAPS  =—p

DESORPTION CYCLE

5-in.NPS, SCHED-40 INCONEL
100°0OR 400°C.

TRANSITION
ZONE

/ INCONEL-X PISTON

5-in. NPS, SCHED-80 INCONEL

HYDRAULIG CYLINDER

1! ~ 35°

WASTE RECEIVER

Fig. 7. Movable-Bed Temperature-Zoned Absorber. When the lower zone
of the bed becomes loaded with fission products, the hydraulic cylinder
operates the piston to discharge that portion of the bed into the waste
carrier. Fresh NalF is added at the top. This apparatus has already
been tested in the ORNL pilot plant.
ORNL-LR-DWG 1909! R-I

QUTLET

THERMOCOUPLE

OUTLET END
HEATER
FILTER
CARTRIDGES
— REFRIGERANT TUBES (4}
5-in-SPS COPPER PIPE
12 0 12 24

Ewm—m_mom)
‘ SCALE (in.)

This design has already been

Fig. 8. Cold Trap for UFg Collection.
successfully used in the ORNL pilot plant.

- 1—(8_
- 25 -

Since there is excess fluorine in the fluorinator off-gas (33-1/3%
utilization), fluorine is recycled. This recycle contains some fission
products, so it is necessary to remove a side-stream (10%) Lo prevent
their buildup. Fresh fluorine is used for the desorption step, for all

processing in the blanket section, and for fluorine makeup.
Vacuum Distillation

The vacuum distillation step applies to the fuel stream only and is
used to separate the carrier salt from the fission products after the
uranium is removed by fluorination (see above ). The LiF and BeF2
volatilize, leaving fission products in the still bottoms. This residue
consists largely of rare-earth fluorides. If the relative volatilities
of the fission products, compared with the volatilities of LiF and Bng,
are low, then a good separation can be achieved in a single-step
distillation without rectification.

At the average operating temperature (sbout 650°C) of the MSBR itself,
the solubility of rare-earth fluorides in fuel salt is only a few mole
percent; however, at lOOOOC, the solubility in LiF alone is about 50 mole %

for the more insoluble compounds, for example, L&F3, Pr¥F_, and CeF Other

rare-earth (EE) fluorides have even higher solubilities 2t this teiperature.
Physics calculations on the 58-day-cycle MSBR indicate that at equilibrium
the molar ratio of L:L'F:(RE)F3 in the fuel is about 1400, a number consid-
erably greater than the 1:1 ratio permitted by the solubility limit at
1000°C. It is therefore apparent that, based on solubility data alone,
distillation at about 1OOOOC can tolerate an extremely large rare-earth
concentration factor before precipitation occurs. The design of the
distillation unit is concerned primarily with determining the appropriate
configuration that will permit a large fission-product concentration
factor and, at the same time, provide adequate heat-removal capability for
the short-cooled Tuel.

The still design developed for this study is shown in Fig. 9. It is
charged initially with L ft3 of Lil that has tThe same isotopic composition
as that in the reactor fuel; this volume Tills the tubes to a depth of

about an inch above their tops. The pressure above the LiF pool is reduced

p)
- 26 -

ORNL DWG 65-1802R2

34" ID -
j «—-‘fw\ AIR COOLANT

|~

B il

fiy—

| AIR OUTLET
* NaK OUTLET
©
817 TUBES
1/2" x 16 GAUGE
INOR-8
1/2" WALL
1/2" INOR-8
TUBE SHEET W

FUEL SALT ,
INLET e==xo) =& NoK COOLANT

LiF WASTE DRAIN
- 30" 1D -

Fig. 9. Vacuum Still for MSER Fuel. Barren fuel-carrier flows
continuously into the still, which is held at about 1000°C and 1 mm Hg.
LiF -BeF, distillate is removed at the same rate that salt enters, thus
keeping the volume constant. Most of the fission products accumulate in
the still bottoms. The contents are drained to waste storage when the
heat generation rate reaches & prescribed limit. This concept of the
vacuum still has not been tested. '
- 27 -

to 1 o 2 mm of Hg by evacuating the product receiver (see Fig. D-3 in
Appendix D), and the temperature is adjusted and held at sbout 1000°C.
Fluorinated fuel salt 1s continuously admitted to the LiF pool in the
still, and distillation is allowed to proceed at the same rate as the
inlet fuel rate so that there is no net volume change.

The operating principle 1s to allow the rare-earth fission products,
which have a much lower vapor pressure than either LiF or BeFE, to con-
centrate in the still. The liguild pool in the still reaches an equilibrium

concentration in LiF and BeF and the Two components then distill at the

P
rate at which each is enteriig the still. The condenser is a conical
region located Jjust above the evaporating surface; it is kept at about
85000 by forced convectlion of heated air. Distillate 1s collected in a
circular trough and drained to a product receiver. The still is operated
in this way until the heat-generation rate due to fission-product decay
reaches the heat-removal capability of the NaK cooling system. This
point occurs after about 67 days of continuous operation, at which time
the fission product concentration in the still is about 1& mole ¢, a
value considerably less than the approximately 50 mole % solubility limit
at 1000°C. At this time, the contents of the still are drained to a
permanent waste receiver, and the cycle is repeated. These calculations
are conservative since they are based upon gross fission-product heat
release and do not subtract the effect of Tthose fission products removed
or deposited in the reactor before chemical processing.

The attractive feature of carrying out the distillation in this way
is that it minimizes the volume of expensive LiF relegated to waste.
Since distillation is carried out after fluorination, less than 0.1% of
the uranium removed from the reactor should enter the still, and at 1000°C
the vapor pressure of UFM is fTavorable to the recovery of a significant
portion of this fraction, reducing the overall uranium loss. Beryllium
fluoride losses should be insignificant because the vapor pressure of
Bell, ot 1000°C is about 100 times that of LiF. Distillation probably
will not effect decontamination from CsF and RbF; each of these com-
pounds has a greater vapor pressure than either BeFE or Li¥. Because
rare earths are concentrating in the still as a function of time, their

decontamination facltor in the product will decrease with time. It is not

possible with existing data Lo asgess the magnitude of this effect.
- 28 -

Detailed calculations for the still design are given in Appendices C
and D. Vapor pressure data for principal components of the fuel streanm
are included in Appendix F.

Distillate from the still is collected in an evacuated tank operating
at still pressure. When filled, the receiver is isolated from the still,

and the LiF -BeF, mixture is transferred by gravity flow or pressurization

to the reductioi and fuel -makeup operations.

An in-cell waste receiver is provided for the initial cooling of the
still residue before transfer tb the underground waste-storage facility.
The tank has a b-fto volume, allowing a one-cycle delay (about 67 days)
inside the cell where heat is conveniently removed by the circulating NaK
coolant. During this delay, the heat generation rate decreases from
3.2 % 10/ Btu/hr to 6.8 x 106, imposing less stringent design requirements
in the permanent waste recelver. The interim recelver is a shell-and-

tube type, similar Lo the still; hovever, no condensing surface or

provisions for air cooling are needed.
Reduction of Uranium Hexafluoride and Reconstitution of the Fuel

The combined UFé;streams,'that form the fuel sgalt and the Tertile
salt, are reduced to UFM’ and only a sufficient amoung3§o maintain
criticality is returned to the reactor. The excess U from the fertile
stream is sold. The usual method for this reduction has been by reaction
with the excess Hy in an HgéFg flanme:

(H2 + Fg) .
UF6 + H2 > UFu 4+ 2HF

This reduction is carried out in a btall column into which UF6 and H2 are

introduced into an HQAFE flame, and dry UFM powder is collected as the
product. However, according to our proposal, a more convenient method
for preparing UF) for the MSBR is by reducing the UFg to UF) with H,
ggzgz_the UF6 is dissoclved in the molten salt. There are some experimental
data indicating the feasibility of such a process; however, the kinetics

of the absorption and of the reduction must be further investigated.26 It

is possible that this two-step process could be carried out continuously
- 29 -

in a single column in which the molten salt flows upward, the UF6 is

introduced and dissolved in the bottom of the column, and the H2 is
introduced at an intermediate point to reduce the UF6 (Fig. 10). Some
of the reconstituted salt has to be recycled to the column to provide

enough dissolved uranium for proper UF6 gbsorption.
Off -Gas Processing

Most of the off-gas from the processing plant comes from the
continuous Tluorinators; smaller amounts are formed in various other
processing vessels. The gases are processed o prevent the release of
any contailned fission products to the atmosphere. Excess fluorine used
in the fluorinators is recycled through a surge chamber by a positive
displacement pump, and a small side-stream of the recycling fluorine is
sent through a caustic scrubber to prevent gross buildup of fission
products. Each of the processing vessels and holdup tanks have off -gas
lines which lead to the scrubber for removing HF, fluorine, and volatile
fission products.

The scrubber operates as a continuous, countercurrent, packed hed
with recirculating agqueocus KOH. A small side-stream of KOH solution is
sent to waste, and the scrubber off-gas is contacted with steam to
hydrolyze fission products such as tellurium. A filter removes the hydro-
lyzed products. The noncondensable Tission products are sent to the off-

gas facility for gases generated by the reactor.
Wagte Storage

Four waste streams requiring storage leave the processing facility:
(1) NaF and MgF, sorbent from the UFy purification system, (2) aqueous
waste from the KOH scrubber, (3) molten-salt residue from the distilla-
tion unit, and (4) molten salt from fertile stream discard. The aqueous
waste stream is small, and it is assumed that adequate capacity exists in
the system for storing reactor waste. The other three wastes are stored
in separate underground facilities adjacent to the procegssing cells.

Since the values in the waste from the fertile stream - TLi, thorium,
ORNL DWG. 65-3036

Fig. 10. Continuous Reduction Column. Barren salt and UFg enter
the bottom of the column, which contains circulating LiF—BeF24UFu. The UFg
dissolves in the salt, aided by the presence of UF), and moves up the column
where it is reduced by hydrogen. - Reconstituted fuel is taken off the top of
the column and sent to the reactor core. Preliminary data indicate that

this design is promising.
- 31 -

2
and 33U -~ will be worth recovering at some future time, some very
tentative ideas about how they may be recovered are presented at the

end of this section.

Nal and MgF . Pellets

Spent NalF and MgF2 pellets, which retain the volatile fission and
corrosion product fluorides from the UF6 gas stream, are stored in 1lO-in. -
diam by 8-ft-high stainless steel cylinders in a concrete vault adjacent
to the still-residue waste tank. (See Dwg. No. 58080D in Appendix F.)

The cylinders are loaded inside the processing cell and transferred to the
underground area at approximately 90-day intervals. The vault is designed
to contain a 5-year collection of cylinders (160 cylinders at 100% plant
factor); after 5 years, the older cylinders are removed and transferred
to a permanent underground storage site. The integrated heat generation
when the vault contains 160 cylinders is about 1.73 x 106 Btu/hr. Forced
circulation of about 11,300 scfm of air at a temperature rise of 12500 is
used to remove this energy. The containers are constructed with a hollow
core, allowing coolant to pass through the cylinder as well as over the

outside.

Aqueous Waste from Off -Gas Scrubber

This waste, fission products in a strong sclutlon of KOH, will be
stored along with other aqueous wastes from the reactor system and
reprecsents an Insignificant contribution to the total amount of aqueous

wasgte.

Fuel -Stream Waste

Residue from the vacuum still is stored in bulk in a facility similar
to one evaluated previously by Carter and Ruch.lh A single, large tank
equipped with adeguate cooling tubes and adequate for a 30-year accumula -
tion of waste 1s provided. The 30-year capacity was chosen since that is
the expected lifetime of the reactor. After filling, the salt might
remain in the tank for additional decay or be disposed of by whatever

method is currently acceptable.
- 32 -

Decay heat is removed by forced air convection. The heat load (Btu/hr)
continually increases over the filling period but decreases rather sharply
when no further additions are made (Fig. E-1, Appendix E). The time
behavior of the integrated specific heat generation rate (Btu ™t ffi"3)
for a b-year collection period is shown in Fig. 11. This 1is a smoothed
curve for L-ft° additions every 67.4 days, followed by an extended decay
period during which no waste is added to the tank. The upper portion of
the curve might be extrapolated with little error to accommodate longer
filling periods. Filgure 11 shows the spedific heat generation for the
case of no dilution with inert salt; however, during the initial stages
of filling, it is necessary to add an inert diluent, for exampie, NaX' -KF
eutectic mixture, to lower the specific rate to a tolerable value. It
was calculated that 26k ft3 of diluent is required for the 520 ft3 of
LiF ~-fission product residue to be collected over the 30-year period. The
fission products being collected exhibit the decay behavior shown in
Fig. 2, which is repregsentative of gross fission products and does not

account for those that have been removed by processing or other mechanisms.

Fertile -Stream Waste

The fertile-stream discard is also stored in a large underground tank,
adequate for 30 years of waste collection. The tank is 13.5 ft in diameter
and 13.5 ft high. Since uranium is removed from the blanket on a rather
fast cycle, the fission rate in the blanket is low, making the waste
activity several orders of magnitude less than that of the fuel waste.
Cooling is provided by natural air convection around the tank and through
cooling tubes. All metal surfaces are expected to be coated with a layer
of frozen salt that will furnish excellent corrosion protection.

The integrated heat producfion rate due to fission products for the
30 ~year period is 5.9l x 3_0lL Btu/hr. Since this epergy is associated
with 1783 ft3 of LiFmThFu mixture, the moderate specific rate of 33.1 Btu

hrwl ft"3 presents no design problems. When first removed from the

1

reactor, the heat production rate of the waste is about 1600 Btu hr’ ft-3,

but this value decreases by a factor of 10 in about 4 days.
-3

ft

ORNL DWG 65-1844 R2

-1

SPECIFIC HEAT GENERATION RATE (Btu hr

E = 5-YEAR  HEHEEEEL e
i COLLECTION PERIOD il : %” e
L *
e I ]
i |
e j T
il i i E e 1
' == 10 =
| 1 YEAR L 2[5 ST By 15 130 B0 i o0 |
, I mlgjfs il I . i T e
EMHIE] 1 i 5 L] RRAE NG R N 4] I (E) ] EWREAIR . i s slizilells
10 108 10 10 10

TIME AFTER DISCHARGE FROM REACTOR (days)

Fig. 11. Specific Heat Generation Rate of Fuel-Stream Fission Products
in Waste Tank. Waste is accumulated in 4 ft° batches every 67.4 days. It
is then held in the processing cell another 67.4 days for further cooling
before draining to the waste tank. The reactor operates on a 58-day cycle
at 2160 Mw (thermal).

¢e -
e

Cooling System for the Waste

As mentioned above, the waste-storage system_is designed for cooling
by forced air draft. Air is easy to handle, COmpatible with construction
- materials for extremely long tlmes, and presents a minimum hazard 1n case.

~of contamination by the waste. Blowers, capable of supplylng 76 000 scfm
at a pressure drop of about 30 1n. water, are located upstream from the
waste vaults The alr is forced through the vaults and coollng tubes in
the Waste tank and 1is exhausted to the atmosphere via a. tall stack, which
also dlSposes of gases exhausted from the reactor system The ex1t duct
contains the necessary radlatlonamonltorlng 1nstruments and absolute '

fllters for remov1ng partlcles

-Pés’sible'Ultimate Treetment of Waste from Ferfiile'smeam :

Slnce a 30 -year accumulatlcn of wagte from the fertlle stream w1ll

33U, and thorlum), not too highly

contain recoverable values (TLl,:
contaminated with fission products, it may be worthwhlle_to consider a

- recovery system before relegatlng this waste.té permahent burial. Afiy.
significant uranium value would probably be.recévefed.by fluorination,
but the recovery of thorium and lithium requires:further process develop-
ment. A potential method for Li-Th separation is the.incompletely

15

investligated HF -dissolution process, based on_the:principle of leaching
LiF from THFA and rare-earth fissiqn products.with_ahhydrous hydrofluoric
acid. This process, however, leaves the thorium contaminated with fission
products, making it necessary to resort to an aquebus system (solvent
extractlon by the Thorex process) or to develop a thorlum recovery process
that utlllzes fluorlde chemlstry From & purely economlc v1ewp01nt,
thorium would be retalned in the waste tank until its recovery became more
economlcal than mznlng new thorxfim

On a 30-year dlscard cycle, an entire fertlile -stream lnventory of
: ThFh and LiF will aecumulate 1n the waste tank. In addltlon; the waste

33 233,

will contaln "Pa, and flSSlon products in amounts that depend on
the breedlng ratlo, eff1c1ency of the fluorination step, and the ‘blanket
pOWer. The uranlum loss is based upon a. 90% efficiency in fluorlnatlon,

A value belleved to be conservatlve,_ The_largest loss of fissionable
- 35 -

material, however, is through protactinium, which we assume to be non-
volatile as the fluoride and which consequently is discarded in direct
ratioco to its concentration. The amount of fission producis is calculated
for a blanket power of 62 Mw (thermal) and an 80% plant factor; to account
for those fission products that have volatile fluorides or which are
plated out on parts of the system, a nominal figure of 80% is used for
the fraction that finally reaches the waste tank.  The 30-year inventory
of the waste tank is shown in Table 2.
Table 2. Inventory in Fertile-Stream Waste
Tank After 30 Years of Collection
Waste volume = 1783 ft3

Amount Unit Value Value

(kg) ($/ke) (10° $)

Th (as ThFu) 141,200 10 1.h41
14 (as LiF) 10, 400 120 1.25
233y 4 233pg (as UrF), + PaFM) 116 12,000 1.39
Fission products {as fluorides) 450 -
4.05

In view of the figures in Table 2, the design presented here for the

fertile-stream waste system is not optimal. Nearly 99.5% of the 233U + 233Pa
233

value 1in the waste tank is attributbtable fo Pa discard; this loss can be

reduced to negligible proportions by providing in-cell decay storage

233,

followed by refluorination to recover the daughter

1-year holdup (about 60 ft3) would allow more than 99.995% of the 33pg, o
233
U.

. For example, a
decay to It probably would not be necessary to provide additional
fluorination eguipment for this volume because the refluorination could be

scheduled in existing equipment during reactor shutdown.
Flow Control of the Salt Streams
Flow control of the molten salt streams is by freeze valves coupled

with a controlled pressure drop. This can be achieved by the simple

freeze valve currently used in the MSRE (Molten Salt Reactor Experiment),
- 36 -

coupled with a flow restriction such as an orifice or venturi. A dynamic
freeze valve in which a controlled layer of salt is built up in a cooled
section of the line is bveing investigated; if successful, it will allow a

grester eage of flOW'control.ET
Removal of Decay Heat

Heat removal is a major problem in all process vessels that contain
short-cooled, highly irradiated fuel. In many cases, the heat flux and
operating temperature will be high, making it difficult to use water or
| air as primary coolants. Water has an additional disadvantage in that 1t
is incompatible with the process fluids, creating a hazard should there
be a leak in the salt lines. Therefore, a sodium-potassium eutectic
mixture, 22.3-77.7 wt % Na-K, was chosen as the primary coolant for
process vessels at temperatures above SOOOC where large amounts of decay
heat must be removed. This.codlant also has the capability of initially
heating the system to 600°¢ for startup purposes. Air was chosen as the

coolant for low heat fluxes at temperatures less Than 5OOQC.
Sampling the Salt and Off-Gas Streams

A rather complicated mechanism is required to remove analytical
samples from a highly radioactive molten-salt system. A technique similar
to that tested for the MSRE will be used,EB and the off-gas streams will

be sampled conventionally.
Shielding, Maintenance, and Repalr of Equipment

All process equipment that handles material that approaches the
radiation level in the reactor core will be shielded by about the same
amount of shielding as for the reactor, and maintenance will e indirect.
Process vessels in this area needing repair will be removed and sent to &
decontamination facility before repair. These include the fuel-stream
delay tanks, fluorinator, sorption-descorption systems, and distillation

system. All other equipment for processing radioactive materials is
- 37 -

contained in an area of direct maintenance with less shielding. In the
direct-maintenance area, decontamination for maintenance 1s achieved by

a molten-salt flush and an aqueous wash.
Materials of Construction

All process vessels and lines in contact with molten salt are made
of INOR-8 except the fluorinatocrs, which are made of a special material,
lloy 79-&.28 Vessels and lines'that contain Tluorine or fluoride -bearing
gases are made of Inconel or Monel, and cold traps are made of copper.

Other equipment is made of appropriately compatible material.
General Operating Policy

The overall policy for operating the [luorinator, vacuum still, NaF
beds, and related equipment is based on the projected simplicity of opera-
tion and small slze of the equipment. The system is designed for a
campalgn -type operation of one month's duration, without shutdown except
for emergency maintenance. There will be no access to the operating areas
during this period. At the end of the operating period, the entire systenm
will be closed down, routine maintenance accomplished, feed hoppers
replenished, accumulated waste transferred to waste storage, etc. The

operating cycle is then repeated.
PROCESS DESIGN

Process equipment for this design study was patterned as much as
possible after previously designed and tested equipment. Fach major
component was studied for its application to the continuous processing
requirements of the MSBR to ensure a practical design. Detailed designs
were made to the extent that overall size, configuration, heat transfer
requirements, flow rates, etec., were defined %o allow a reasonable
estimate of equipment cost. The waste system was considered separately,
and liberal use was made of a previous, similar :sfi:uéLyll‘L Tor its design

and cost dats.
- 38 -
Fuel Stream

Degign Criteria

Design criteria for process equipment were hased on past experience.
Equipment capacity was based on handling & flow rate 30% larger than that
required for 100% continuous operation as specified by reactor physics
calculations. Pertinent basic data adopted for this design are as follows:

Salt flow rate: 15 ftB/day |

Pressure: 2 atn | o o

Density of pelleted sorbents, NeF and MgF,: 1 g/cc

Temperature range for NaK .coolant:

As coolant: 200°C .
As heat source: 800 C o
Coolant temperature rise in heat transfer operations: 300 C
Normal convective heat transfer coefficient: 10 Btu he T rp e Opt
Electrical heat needed: 1.5 kw per ft2 of longitudinal area for
temperatures higher than 5OOOC

Fission product heat: 50% of total is removed with volatile fluorides
in the fluorination step

Major Process Lguipment

There are 41 major pieces of fuel-stream process equipment (Fig. 12).
‘Most of the equipment design is.straightforward and based on conventional
techniques.29-32 Each component.shown in the processing flowsheet
(Fig. 12) is listed below, with its purpose, design basis, and description.
The ldentifying number accompanying the equipment name corresponds to the
éircled numbers in the-figure. Design calculations that involve unusual

technigues or complexity are shown in the agppendices.

1. Flow Control

Purpose: The flochontrol device meters the flow of a molten salt stream
to or from a process vessel.

Description: A piece of process pipe, l/2-in., sched -40, 1 £t long. Pipe
is Jacketed with 1-1/2-in. sched-bO pipe made of INOR-8 and has two
coolant connections. '

2. Coolaflt Tank

Purpose: To allow delay for fission-product decay of the molten-salt
stream from the reactor core before its introduction to the fluorinator.
39 -

ANMONIUM
FLUSH OXALATE
SaLT AND WATER = TO PROCESS HOT Fg.FP
COOLING Fp SURGE| |et \ 2
TANK J b
sauEops T ASWERS (15)COMPRESSOR
FLUSH F :
ALT FLUSH o Fa.FP o
MAKE-UP MAKE 1P (
ANK TARK Fo UG FP
4 7
FLI TO PROCESS Fp UFg FF
ConTROL VESSELS FLOW NeF ™
CONTROL
UFg TRAP
TO PROCESS
VESSELS
— NaF —ag—
Fadifg,FP g HEBTER Mg Fz*!' o, 70 SOREERS
= 'S M3 o FP TRAP
F2,UFg,FP b ié UF4 FP @ 4
NoF —={33) ‘;;" o & REFRIGERATION coro
i i c
& MgF FO TRAP 5 9 Ay £ I | TRAP
w x o
@ 5] HE@ER 5
© | s o
o Ple F2 i REFRIGERATION ® . 3 s 3y D
B g > -
| = / i
& e Lfiflfl <N
P w— @ g [
s FLOW a WASTE o p— 10 WASTE
CONTROL @ TANK ) STORAGE e 1%
F HEATER TO WASTE ‘ b
o STORAGE oLD -
5 T%’S
z ¥
g = 5.( >
c TO WASTE g
> “ STORAGE &
w = w ¥
&
LiF, ThPg PRODUCT O 1 l
CONNER ‘
g G 0000000 o
N : MAKE-UP MaF WASTE TANKS AND
=] \ TANK Al COOLING STATION
el A -
=4 FLOW {ne TER
CONTROL.
- Hp HF 0] ) ||
TO WASTE T X
STORAGE Ha,XE £
@ BLANKET Hp FP
wgnz FP HYDRATER
TANK REACTOR 2
3 J )
TOBLANKET . _
TO WASTE My HF,UFg
HEATER STORAGE
<5 4
1 scausaNme
UM M%KE-’;UP Efrve @
® - 2
H He : >
¢ y NaF 7O ) L‘
SORBERS .
oy VACOUM PUMP UFg ¥
, Hp Be = '
Supply Sopoly - Y TQ WASTE N
Tank Tark . OWALI I MS,UFz STORAGE el Fa.FP k
Process ‘ ‘ CONTROL > 7 et .
NaF €9 @ ] @ o P
Hot Coot 2 @ } . WAETE
Nk N COLLECTION Co | HERVER g HOLD-UP
Pry §2 £
‘‘‘‘‘ 2 <
. & 5 Eo | ur ®
C°¥GC'=0" 5 2 il il - REFERENCE DRAWINGS RO.
onk oy <
=® 1 dem 0OAK RIDGE NATIONAL LABORATORY
3 ) . ! g@ KOH,FP TO OPERATED BY
- ] ek WASTE STORAGE UnioN CARBIDE NUCLEAR COMPANY
FUEL )% DMSION OF UNION CARBIDE CORPORATION
E SgLT TO g ~ Pégp 0AK RIDGE,
- DUMP TANK
— LIMITS ON DIMENSIONS UNLESS |MSBR Integrated Processing Ptont BLOG. 4500
HF STORAGE £ O,
! pump JF STORAG €5 STORAGE @ : N GTHERWSE SPECEFIED:
FRACTIONS
10, REVISIONS pATE JaPeD| APPD MSR FUEL AND BLANKET
ARMER | 3-11-69 : i ANBLES % SUBIHTTED APFRGTED FPPROVED
£ FATE IO | RS | DA ’ -
- DATE | TTDATE | APPROVED | DATE SCALE: NO SCALE APPROVED I l 58025| D‘ 1
E) i

Fig. 12.

Major Components in Fuel- and Fertile -5tream Process.

- bho -

Design Basis:

Type of tank: Healt exchanger with well-mixed contents
Salt residence time: (See Appendix A)

Temperature: 600°C
Coclant: Nak L 1 -3
Heat load: 5.14 x 10 Btu hr — ft
Material of construction: INOR-O

3

Description; A tank having 22.5 £t~ of liduid volume, with bayonet coolers

and two pipe connections.

3. Fuel-Balt Fluorinator

Purpose: To remove all uranium and volatile fission products from the fuel
salt by continuous fluorination.

Design Basis:

Salt residence time: 2 hr

Fluorine utilization: 33-1/3% o
Maximm mass flow rate of F: 0.277 slpm/in.
Temperature: 600°C

Coolant: Nak i _ 3
Heat joad: 5.31 x 10 Btu hr ki
Material of construction: Alloy 79-4

1

Description: A column made of 6-in. sched-40 pipe, 11-1/2 £t long, which
is Jacketed with an 8-in. sched-4O pipe. Column is expanded above
jacketed section to 8-in. sched-k0 pipe, 1 ft long. Column is supplied
with' 5 kw of electrical heat and has 5 pipe connections.

4., Burge Tank

Purpoge: To allow molten-salt surge capacity between the fluorination and
distillation steps.

Design Basis:

Temperature: 600°C

Surge capacity: 1 day's continuous flow (15 ft3)
Coolant: NaK L 1 3
Heat load: L4.4 x 107 Btu hr — ft

Material of construction: INOR-3

Description: A tank having 15 ft3 of liquid volume, with bayonet coolers
and two pipe connections.

5. Movable-Bed Sorber

Purpose: To separate some of the volatile fission products and corrosion
products from UF6-

Design Basis:

Number of sorption zones: 3

Cooling load in high-temperature trap: 5.52 x 10” Btu/hr
UFg load based on l2-hr cycle: 13.5 kg

NeF loading: 0.5 g UF6/g NaF
- b1 .

Coolant: Air

Average temperature: 400°C in bottom zone, 100 to 150°C in two top
zones

NaF Usage: 20% of one zone volume per day

Material of construction: Inconel

Description: An annular column made of two concentric pipes. The outer
pipe is 10-in. sched-40, 8 £t long; the inner pipe is 6-in. sched-Lo,
8 ft long with the bottom 30 in. finned. Bottom mechanical solids
ejection; 3% kw of electrical heat.

6. NaF Supply Tank

Purpose: To maintain a supply of NaF pellets for the movable-bed absorbers.

Design Basis:

NaF supply period: 30 days

NaF usage: 10.75 kg/day

Temperature: Ambient

Material of construction: Stainless steel

Description: A tank, l~1/2 £t in diameter and 3 £t high, with a conical
bottom and four 2-in. star valves.

7. Na¥ Waste Tanks and Cooling Station

Purpose; To provide short-term waste storage for the solid sorbents which
contain sorbed fission products.

Design Basis:

Tank capacity: 107.5 kg NaF

Coolant: Air

Temperature: <L00°C on surface

Capacity of cooling station: & tanks

Heat rate of cooling station: 2.0 x 10  Bbu/hr (Appendix B)
Material of construction: Inconel

Description: Vessel made of two concentric pipes; outer pipe is 10-in.
sched-20 pipe, 8 £t long; the inner pipe is 3-in. sched-20. Vessels
supported in a metal frame.

8. Fission Product Trap

Purpose; To remove the fission product Mo from a UF6 fission product stream
by sorption on MgFE.

Desgign Basgis:

Temperature: lBOOC

Coolant: Air

Capacity: 20 kg of MgF2 pellets
Material of construction: Inconel

Degcription: Vessel made of two concentric pipes; outer pipe is 6-in.
sched -40 by 5«1/2 ft long; inner pipe is 3-in. sched-40. Vessel heated
by 5 kw of electrical heat.
~ ho .

9. Primary Cold Trap and UF, Vaporizer

Purpcse: To serve as the primary cold trap for collecting UF, from the
sorber effluent and to be the.UF6 feed station for the reduction column.

Design Basis:

UF6 capacity: _1365_kg o
Temperature: =40 C to 100°C
Material of constructilon: Copper

Description: An internally finned vessel, 5 in. in diameter and 14 Tt long,
with 15 kw of electrical heat

10. Refrigeration Unit

_Purposé: To provide coolant for the fuel-salt primary cold traps. -
Description: Freon-type, rated at 60,000 Btu/hr at -45°C.

11. Secondary Cold Trap

Purpose: To act as a backup to the primary cold trap.

Description: An internally finned copper vessel, 6 in. in diameter and
6 ft long, with 5 kw of electrical heat.

12. Refrigeration Unit

Purpose: To provide coclant for the fuel-salt secondary cold traps.
Description: Freon-type, rated at 8,000 Btu/hr at -75°C.

13. UFg Trap

Purpose: To remove trace amounts of UF6 that escape the secondary cold
trap.

Design Basgis:

Temperature: Ambient
NaF Capacity: 20 kg
Material of comstruction: Inconel

Description: A vessel made of L-in. sched-kO pipe, b £t long.

14, Fluorine Surge Tank

Purpose: To provide a surge capacity for the recycle F2 stream.

Design Basis:

Capacity: Ll-hr supply
“Temperature: Ambient
Material of construction: Inconel

Description: A tank, 1-1/2 £t in diameter and 4 ft long.

15. ZFluorine Compressor

Purpose: To recycle fluorine from the fluorinator off-gas back to the ‘l’
fluorinator inlet. -
- b3 .

Deslign Basis:

Type: Diaphragm with remote head
Capacity: 6 to 10 £t3/ar
Temperature: Ambient

Material of construction: Monel

Description: A diaphragm pump with remote head.

]

16. Vacuum Still

Purpose: To separate fuel-carrier salt from less volatile fission products.
Design Basis:

Pressure:; 1 mm Hg

Temperature: 1000°¢C

Liquid volume: L 13

Heat transfer: Appendix D

Distillation rate: 15 ft3 of salt per day
Coolant: NaK

Operation: Continuous addition of feed and continucus removal of distillate;
periodic removal of residue.

Description: An INOR-8 vessel of ghell-and-tube design, 2.5 £t in diameter
and 2.2 Tt high; 267 4= of heat transfer area in liquid section and
6.9 £t~ in condensing section; 45 kw of electrical heat.

17. BSalt Makeup Tank

Purpose: To prepare barren salt in the nonsghielded operating area for use
as salt makeup.

Design Basis:

Temperature: 9OOOC
Capacity: 1 ft3
Material of construction: INOR-3

Description: A tank, 1 £t in diameter and 1.5 £t high, with 8 kw of
electrical heat.

18. ILiF Makeup Tank

Purpose: LiF makeup for the distillation unit.
Design Basis:

Temperature: 900°C
Capacity: L £t3
Material of construction: INOR-8
Description: A tank, 1.5 ft in diameter and 3 ft high, with 20 kw of
electrical heat.

19. Coolant Tank for Fuel-Stream Waste

Purpose: To provide short-term storage and cooling of the waste stream
from the vacuum still.
ERUN

Design Basis:

Temperature: 1000°C

Capacity: U £t3

Heat removal: See Appendix D
Meterial of comstruction: TNOR -8

Degcription: A tank having h i3

of liguid volume, with bayonet coolers
and 40 kw of electrlcal heat R -

20. LwaBeF Makeup Tank

 Purpose°' To prov1de surge capa01ty and barren salt makeup for the molten_
salt fed to the reductlon column. :

Design Basis:

:Temperature: 600°¢
Capacity: 8 hr of retention | S |
- Material of construction: INOR-8 -

: Deseriptiéfi A tank, 1.5 ft in diameter and 3.5 £t high, with L0 kw of
_ electrlcal heat

2L. Vacuum Pump

Purpose:  To provide vacuum for the vacuum still.
: Design Basis;

 Displacement: 4O cfm
Pressure: <50 u Hg.
Material of construction: Steel

Description: Commercial, oll-pumped vacuum unit.

22. Reduction Column

Purpose: To reduce the purified UF6 in the still distillate to UFM This
' reconstltutes the fuel.

i De81gn Ba51s

.Temperature 600°C

Residence time for dissolving: 30 min
'_Resiéeaée'time for reduction: 1 hr B -
UFg rate: 20.9 kg/day : A

~UFg ma.ss flow rate: T g in. = hr-t

H, utilization: 50%

Material of construction: INOR -8

Description: A column, made of 12-in. sched~h0 pipe, 8 Tt hlgh, expanded
at top to 18-in. sched-40 pipe by 18 in, long Heated by 50 kw of
electr;cal_heat
23. Molten-Salt Pump | |
Purpose: 'To recy¢le molten salt in the feductiom system. e , .‘I'

- 45 -

Design Basis:

Type: Open bowl with helium purge
Temperature: 600°C

Capacity: 2 ft3/hr at 20-ft head
Material of construction: INOR-8

Description: Similar to pumps developed for Molten Salt Reactor Experiment.

2L, Collection Tank for Recycle Fuel

Purpose: To provide surge capacity and a means of transferring molten salt
to the reactor dump-bank.

Design Basis:

Temperature: 600°C
Capacity: 12-hr retention
Material of construction: INOR-8

Description: A tank, 2 £t in diameter and 2.G ft high, with 27 kw of
electrical heat.
25. Filters

Purpose: To remove particles from processed molten salt before its return
to the reactor.

Design Basis:
2

Filtration area: C.1 ft
Material of construction: Porous nickel
Description: Porous metal filter in a 2.5-in.~diam by 1-ft canister, with
0.1 ftg area and a 2-kw electrical heater.
26. Off-Gas Scrubbing Column

Purpose: To remove volatile fluorides from waste-gas streams.

Design Basis:

Temperature: Ambient

Type: Countercurrent packed column

Packing: 1/2-in. Raschig rings

Source of fluoride gases: 10% of F, (derived from fuel-salt
fluorinator), all the HF from rediuction column, all the HF from
blanket -salt sparging. The HF rates Trom these two sources are
equal.

Gas mass flow rate: Less than 50% of flooding

Agueous purge: 1 gal of agueocus KOH per hour

Cooling: Water

Material of construction: Monel

Description: A column, which is 3-in. sched-4O pipe, by 6 £t high, having
an expanded section of 6-in. sched-40 pipe by 1 £t high on top. Column
is jacketed with 5-in. sched-h4O pipe, 6 ft high, and packed with 1/2-in.
Raschig rings.
r27‘ FlSSlon Product Hydrator

Purpeee To hydrate all volatlle flSSlOflmprOduCt fluorldes that leave
the scrubbing column. :

Deslgn Ba51s;_
Temperature: lOO C ' o
Steam rate: Equlvalent to 1 gal of HEO per hr :
- Materlal of eonstructlon Monel
Descrlptlon A vessel vhich is hwln schedéhofpipeibyra Tt high with a
1-kw electrlcal heater g ' R o :

'28.' Condenser o

Purpese Remove the condensable materlal from gases leavzng the Tission-~
Product hydrator. ' . _

- Desxgn.Ba51s.

Temperature: Ambient
Coolant: Air
Materlal of’ construetlon Monel

-]_.Descrlptlon .A_hmftg heat exchanger.

29. Absolute Filter

Purpose: To remove particles from the noncondensable gases leaving the
condenser.

Design Basis;

Temperature:; Ambien

Filter area: 0.1 It

Filtering medium: Fiberglas

Material of construction: Stainless steel

Description: A commercial unit.

30. Makeup Tank for KOH

- Purpose: To provide surge capacity and makeup volume for the scrubbing
column. - ; |

Deslgn Basis:

Temperature: Ambient
Capacity: O hr of scrubber flow
Material of construction° Monel -

Degcription: A tank 1 5 iV 1n dlameter and 2 ft high.

31. Pump for Solutlon in Off«Gas Scrubber

Purpose' To 01rculate the aqueous KOH 1n the ecrubblng syetem

De51gn Ba51s

Temperature Ambient
Type _ Canned_rotor
- L7 -

Capacity: 7 gal/hr at 5-ft head
Material of construction: Monel

Description: Commercial unit.

32. Aqueous -Waste Tank

Purpose: To provide short-term holdup of agueous waste.
Design Basis:

Temperature: Ambient
Capacity: 10 days
Material of construction: Monel

Description: A tank, 3 ft in diameter and 5 ft high.

33. Fluorine Storage and Supply

Purpoge: To provide the primary fluorine supply for the systen.
Type: Tank and trailer.

Description: Tank trailler containing gaseous fluorine.

34, Hydrogen Storage and Supply

Purpose: To provide the primary H2 supply for the system.
Degign Basis:

Type: 200-scf cylinders with pressure-reducing station
Capacity: U8-hr supply
Material of constructilon: Steel

Description: Four high-pressure cylinders; commercial units.

35. Hydrogen Fluoride otorage and Supply

Purpose: To provide HF supply for the system.

Design Basis: Continuous supply from 200-1b cylinders in constent -
temperature steam bath.

Description: Two commercial cylinders.

36. NaK Cooler

Purpose: To remove heat from the NaK coolant.

Design Basis:

Type: Outside, air-cooled

NeK capacity: 5251 £t3/hr

NeK temperature: Enter 500°C, exit 200°C
Average air temperature: 27°C

Material of construction: Steel

Description: Air-cooled heat exchanger with 9200 ftg of heat transfer
area and a 60-hp fan.
S48 -

37 NaK Hea%er

| Purpose: To ‘heat the NaK so 1t can be used as a heating fluild for startup
cand checkout : .

| De51gn Ba51s

:Type. Heat exchanger using electric heat
NaK capacity: = 180 ftg/hr

- NaX temperature ‘Enter at 500°C and leave at 800 C
Temperature on hot side of exchanger: 900°C
Materlal of constructlon Stalnless steel i

j_Descrlptlon A heat exchangfir with 270 ft2

of heat transfer area and
355 kw of electrlcal heat _

.38. NaK Collectlon Tank

PurpoSe To collect N&K coolant and to serve as surge capacity.

De81gn Ba81s

Temperature | go e
Volume:. = 200. 4 .
Materlal of construcfilon Steel

"Descrlptlon A tank, 5 ft in dlameter and 10.5 £t long.

| _ 39. Nak Pump |

- Purpose: To circulate the NakK through process equipment and heat
-exchangers.

‘Design Basis:

Type: . Centrifugal

Capacity: 650 gal/min with 50-ft head
Temperature: 500°C

Material of construction: Stainless steel

Description: A commercial centrifugal pump, powered by & 15-hp motor.
40. NeK Supply Tank

Purpose: To provide surge capacity for NeK when it is used as a heating
fluid AR | : '

De81gn Ba81s

Temperature SOOQC
~ Volume: 10 ft°
Materlal of constructlon Stalnless Steel

Descrlptlon A tank 2 f@ in dlameter and M 1t long, with 36 kw of
electrlcal heat '

1. .N&K Supply Tank

Purpose: -TQ $QPply¢NaK coolant to process equipment
- 49 -

Design Basis:

Tenmperature: 200°¢
Volume: 100 £3
Material of construction: Steel

Description: A tank, 3.5 ft in diameter and 10 ft long.
Fertile Stream

Design Criteris

Process equipment for the fertile stream was also designed for 30%
greater capacity than that specified by reactor physics calculations for
100% continuous operation. The design rate is 105 ft3 of fertile salt
per day. Other basic design values are the same as listed above in the
design criteria for the fuel stream.

There are 14 major items of process equipment; the function of each
is shown in Fig. 12. The listing below gives a description of each item,

and the identifying numbers correspond to the circled numbers of Fig. 12.

L2, Blanket-Stream Fluorinator

Purpose: To remove uranium from the blanket salt by continuous
fluorination.

Design Basis:

Salt residence time: 2 hr

Fluorine utilization: 33-1/3%

UF, rate: 15.15 g-moles/day

Temperature: 600°C

Coolant: NakK 5
: Meximum mass flow rate of Fy: 0.277 slpm/in.
Coolant load: L2 kw
Resistance -heating load: kL2 kv
Material of construction: Alloy-79-U

Description: A column, 12-in. sched-LO pipe, 16.5 ft high, which is
Jacketed with ih-in. sched-4O pipe, 16.5 in. high. Column is enlarged
at top to 2k-in. sched-4O pipe, 18 in. high. Heated by 42 kw of low-
voltage alternating current spplied to 14 ft° of nickel electrodes;
supplemental heat supplied by 4 kw units.

43. Movable -Bed Sorber

Purpose: To separate some of the volatile fission products and corrosion
products from UF6.

. Design Basils:
Number of sorption zonesg: 3
UF; load, based on 2L hr of operation: 221 g
.;e5o )

NaF loading: 0.2 g UFé-per-gram of NaF

Coolant: Air SRATE R R o _

Average temperature: ~MOOOC in bottom zone, 100 to l50?C in two top
zones | T |

NaF usage: 20% of one zone volume per day

Material of construetionf ~Inconel

Description:_ Column made of two eoncehtrie pipes, outer one is 5-in.
sched -40 pipe, 3 £t high; inner one is l.5-in. sched-4O, 3 ft hlgh.
Mechenzcal egector for SPent NaF 5 kw of electrlcal hea%

IV FlSSlonéProduct Trap

"Purpoee To remove flSSlOn products from a UF6—f1351on—product efiream by
sorptlon on MgF2 ' : : : :

eDes1gn Ba81s

Temperature 150 °c _
Capaezty 5 kg of MgF2 pellets
Materlal of construetlon Inconel

l Descrlptlon A vessel which is L-in. sched-LO pipe, 2 ft high, warmed by
2 kw of electrlcal heat ' ' : '

'M5, Refrlgeratlon Unlt

'T_Purpose To prov1de eoolmng for the blanketustream primary cold trap.

_Descrlptlon Freon-type W1th 2L, 000 Btu/hr COOllfig capacity at _45 C.

L6, Refrigeration Unit

Purpose: -To.provide cooling for blanket-stream secondary cold trap.

Description: Freon-type with 4,000 Btu/hr cooling capacity at -75°C.

W7, Product Receiver

ePurpose To recelve the UF6 product from the blanket -gtream cold traps
'De51gn Ba51s

Temperature Ambient
- Capacity: 10 days' production
Materlal of constructlon Monel

- Descrlptlon A,vessel,_h-ln, schedfho pipe, 2 ft long.

' :%8; .§E6 Trap

.Purpoee To remove trace amounts of UF6 from the coldmtrap off-gas
.De51gn Baels ' ' '

. NaF capacity: eS.kg._
Temperature:  Ambient
Materlal of construetlon Inconel

- Descrlptlon A Vessel, L-in. sched-L0O pipe, 2 £t long
- 51 -

49, Makeup Tank

Purpose: To act as surge tank, molten-salt makeup tank, flush-salt makeup
tank, and means of molten-salt transfer for the purified blanket salt.

Design Basis:

Temperature: 600°C
Capacity: 12 hr retention
Material of construction: INOR-8

Description: A tank, 3.5 £t in diameter and 6.5 ft high, with 105 kw of
electrical heat.
50. Filter

Purpose: To remove particles from the processed molten salt prior to its
return to the reactor.

Design Basis:
Filtration area: 1 ft2
Material of construction: Porous nickel
Description: A porous metal filter in a 2.5-in.-diam by 1-ft-high
canister, heated by 2 kw of electrical heat.
51l. Waste Tank for Fertile Salt

Purpose: Short-term storage for fertile salt waste.

Design Basis:

Temperature: 600°C
Capacity: 10 days' retention
Material of construction: INOR-O

Description: A tank, 14 in. in diameter and 2.5 £ high, with 14 kw of
electrical heat.

52. Makeup Tank for LiF -ThF,

Purpose: For blanket-salt makeup in the nonshielded operating area 1o
supply the main makeup tanks.

Design Basis:

Temperature: _600°C
Volume: 1 £t
Material of construction: INOR-8

Description: A tank, 1 ft in diameter and 1.5 ft high, with 7.5 kw of
electrical heat.

53. Makeup Tank for Aqueous Flush

Purpose: For mskeup and supply of an aquecus flush for decontamination.
Design Basis:

Temperature: bient
Volume: 62 ft
Material of construction: BStainless steel
s -

Description: A tank, 3.5 ft in diameter and 6.5 £t high. .

54, Line Heater

Purpose: To heat sections of process lines for transferring molten salt.
Design Basis:

Type: Clam-shell electrical
Reting: 5 kw per 10 £t of 1/2-in. line

Description: Heaters_furnished in 10-ft sections.

55. Cdoling Air Blower
-_Purpose: To supply édoling air where needed in the process.
Design Basis:

Capacity: 10,000 £t5/min
Pregsure drop: 10 in. H,O -
Material of construction: OSteel

Description: Radial-flow blower with a capacity of 10,000 ft3/min at

12 in. HQO.

PIANT DESIGN AND TLAYOUT

The inherent advantage of a fluid-fueled reactor is in close-coupling
the reactor and processing plants, thereby reducing capital investment,
inventory, and operating costs. This conceptual MSBR processing plant is
located in two cells adjacent to the reactor shield; one contains the high-
radiation-level operations, and the other contains the lower-radiation-
level operations. Each cell is designed Tor top access through a remov-
able biological shield having a thickness equivalent to 6 ft of high-
aensity concrete. Both cells are served by a crane used in common with
thé reactor_plant. Process equipment is located in the cell for remote _
removal and replacement from above. No access into the cells will be
'required; however, it is possible with proper decontamination to allow
1imited access into the lower-radiation-level cell. A general plan of
the processing plant and a partial view of the reactor system is shown
in Dwg. 58050D in Appendix F.
The highly radioactive cell contains only fuel -stream processing
equipmenfii:-fhe fluorinator, still, waste receiver, Nal and MgF2 sorbers,
and associated vessels. The other cell houses the blanket-processing ‘I'
equipment, fuel - and.fertile—stream cold traps, UF6 reduction equipment;

and fuel - and fertileéstream.makeup vegsels.

e
- 53 -

The high-radiation-~level cell has a cross section of about 19 by
22 ft; the less radioactive cell measures about 27 by 30 ft. FEach cell
is about LO £t high, the same as that of the adjoining reactor cell. A
"cold" operating area, located along the face of the cells, has dimensions
12 by 62 ft and contains the cold makeup equipment, product receivers, and
process-gas supply. All process operations are carried out remotely from

the reactor control roomn.
COST ESTIMATE

One of the major goals in this design study was to estimate the cost
of an integrated facility for processing the fuel and blanket salts for
this conceptual 1000-Mw (electrical) MSBR. The estimate arrived at
includes the total fixed capital cost and the annual operating cost
which are, respectively, $5,301,510 and $787,790 per year. In general,
conventional estimating practices were used except where past experience
in the nuclear energy field indicated changesu29”32

It is difficult to separate the process bullding and its eguipment
from the total reactor facility for cost-estimating purposes, since the
building must be an integral part of the reactor installation. To
determine the relationship between the costs of integrated processing
plant and the reactor bullding, it was assumed that the Molten Salt
Converter Reactor (MSCR) facility, designed by Sargent and Lundy
33

Engineers, represented the nonprocessing part of an MSBR. Any required
addition to the structure, facility, or operating cost of the MSCR due to
the addition of the processing facllity was considered a cost of the

processing facility.
Process Equipment

The installation charge for a component to be installed in an
indirect-maintenance area was set at 50% of the selling price and at 25%
for equipment to be installed in the direct-maintenance and cold areas.
A list of all major installed components and their costs is given in

Table 3. The total cost of the installation of major process equipment
is $853,730.
- 54 -

- Table 3. Installed Cost of Major Process Eaguilpment

Eguipment Number | No. of Installed

Fission-product trap

- From Fig. 12 Description Units Cost
1 Flow control 15 & 3,160
2 Cooling tank 1 184, 000
3 Fluorinator 1 5,420

oo Surge tank 1 139,000
5 Movable -bed. sorber 2 21, Loo
6 NaF supply tenk - 1 2,750
T NaF wagte and cooling station 9 12, 750
8 Fission-product trap 1 820
9. Cold trap and UFg vaporizer 3 31,930
10 Refrigeration unit 1 6, 300
1L Cold trap L 12,280

12 Refrigeration unit. 1 5, 000 .
13 - UFg tbrap | 2 420
1h Fluorine surge tank 1 1,910
15 Fluorine compressor 1 1,850
16 Distillation unit 1 58, 700
17 Makeup tank 1 1,570
18 Makeup tank 1 L, 390
19 Cooling tank 1 58, 000
20 Makeup ‘tank 1 6, 590
21 Vacuun pump 1 2, 640
22 Reduction column 1 9, 050
23 Molten-salt pump 1 15,000
2l Collection tanks 2 11,600
25 Filters 2 900
26 Scrubbing column 1 1,240
27 Fission-product hydrator 1 400
28 Condenser 1 800
29 Absolute filter 1 200
30 Makeup tank 1 1,410
31 Pump 1 200
32 Waste tank 1 L, 540
33 Fluorine storage and supply 2 13,800
3L Hydrogen storage and supply L 780
35 HF storage and supply | 2 1,280
36 NeK cooler | 1 60, 500
37 NeK heater 1 38, 670

38 Collection tank 1 6, 380

39 NeXK pump = - 1 2, 000 .
40 Supply tank 1 6,670
L1 Supply tank "1 3, 430

Lo Fluorinator 1 16,600 .

-ti -~ Movable -bed sorber 1 2,120
- 1

- 55 -

Table 3 (continued)

Bquipment Number No. of Installed
From Fig. 12 Description Units Cost

45 Refrigeration unit 1 3,600
L6 Refrigeration unit 1 3, 300
W7 Product receiver 1 200
L8 UFg trap 1 250
L9 Makeup tank 3 63, 300
50 Molten-salt filter L 1, 060
51 Waste tank 1 2,940
5e Makeup tank 1 1,570
53 Agueous -flush makeup tank 1 7, 100
sk Line heaters 13 6, 500
55 Cooling-air blower 2 L, 760

Total installed equipment cost $853, 760

Several of the process vessels and auxiliary equipment are similar
to equipment previously purchased by ORNL or previously estimated in
another design reportBES These estimates were updated, and in some cases
conventional equipment sizes were extrapolated to meet the requirements
of this larger plant. The process vegsels in which high heat fluxes were
anticipated were considered to have costs comparable to that of the MSRE
drain-and-fill tank.

For Vessels and tanks of conventional design, the cost was computed
from the cost of material plus estimated charges for fabrication and
installation. A summary of the fabrication charges used is given in
Table 4.

Additional estimating criteria were:

Vessel -wall thickness: 1/2-in. or sched-40 pipe
Vessel -end thickness: <36-in. diam, 1/2 in.
>36-in. diam, 3/L in.
Equivalent vessel weight for
entrances and exits: <100 1b vessel, 10 1b
100-1000 1b vessel, 20 1b
. >1000 1b vessel, 40 1b
Cost of electrical heaters: $100/kw installed
5 56 B

29, 32

Teble L., Summary of Fabrication Cost for Process Vessels

Fabrication Cost

($/10)

o _ - for a Vessel Welght of

‘Meterisl o <1000 1 >1000 1b
Mi1d steel R S 2.0 1.00
Stalnless steel nlckel Monel, and Inconel L  3;50 ~2.50
 INOR-8 | , - o - ko 3.50
Mloy 794 - © 5.00 4.00

Structure and Improvements

. Structure and improvements costs for the processing facility w

ere

~determined by assuming that these costs are directly related to corre-

sponding costs developed for the MSCR by Sargent and Lundy Engineers.

'Adflition of the procesgsing facility proposed here called for an additional

16.56% of floor space in the reactor building, and this factor was

to determine the cost from the corresponding MSCR data. The cost o

33

used

T the

crane and hoist was also increased by this amount. This additional space

includes a 10% increase in analytical and decontamination facilitie

The processing facility addition was 7.3% of the total building are

SB

a;

therefore, cost of the grounds and stack was increased by 7.3%. The

total increase in cost of structure and improvements due to the processing

~ facility was $550,770, which was obtained as follows:

Structure 0.1656 x $2,932, 400 = $485,610
Crane and hoist | 0.1656 x 195,000 = = 32,290
Grounds = 0.073 x 501,500 = 36,610
‘Stack | 0.073 x 31,000 = = 2,260

Total structures and improvements 8 o $556, 770 - -

Interim Weste Storage

The 1nterlmwwaste«fac1llty cost was estimated separately. _'Thi

estlmate was based on previous cost estlmates for similar fac1lmt1e

g .

o 4,32
- 57 -

Total waste facility cost was $387,970, with the following cost breakdown:

Tank cost $213, 200
Excavation and backfill 32, 350
Concrete 18,070
Transfer lines 9, 000
Instrumentation 10, 000
Cooling system 100, 650
Structural steel 4, 700
Total $387, 970

Other Plant Costs

Slgnificant cost items in the indirect-maintenance area of the plant
are the process line "Jjumpers", which have remote connectors necegsary
for indirect connection of pipelines, instrumentation lineg, and electrical

lines. The following cost schedule, based on experience at ORNL, was used

2
for estimating the cost of these connection8:3
Major pipelines $1,500/1ine
Multiple pipe and instrumentation lines $l,700/set
Electrical heater connection, including the
heater $2,000/set
Other process-piping cost schedules were:
Motor -operated control valves $500 each
Coolant air ducts $10/f%
MaJjor process lines in direct maintenance
area (<20 ft) $400 each
Gas lines in direct maintenance area (<20 ft) $200 each
NeK coolant lines (<20 ft) $800 each

The above cost schedule results in a total process-piping cost of $155,800.
The electrical auxiliaries consisted of the electrical substation,

- switching gear, feeders, and indirect connectors and Jjumpers. Cost

schedules used for these auxillaries were:29
Electrical substation and switching $36.60/kw
Overhead feeders $6.1/ft
Underground feeders $12.2/f%

The total cost of electrical auxiliaries is estimated to be $84, 300.
Process instrumentation is estimated to be $272,100, radiation
monitoring to be $100,000, and sampling connections to be $20,000.
The costs of service lines and high-temperature insulation are based

‘I' on the installed cost of the process equipment in the main processing
. 5 -

areas-and in the waste storége facility. The cost of service lines, taken
at 15% of the total installed cost of this equipment, amounts to $128,060;
the cost of the insulation, taken at 6% of this installed cost, amounts to
$51,220.

Total Fixed Capital Cost

The total direct plant costs are estimated to be $2,609,980 (Table 5).
Past experience was used to determine percentage costs of indirect capital

' items. These percentages are higher than those for other chemical indus -

tries but_re@resent actual cost expéerience in other ORNL projects.32

 Constructiofi overhead is estimated at 30% of total direct plant cost to
give a total construction cost of $3,392,970. Engineering and inspection
at 25% of total construction cost is $8L83,2L40, which results in a sub-
total plant cost of $h,2hi,216. Contingency at 25% of the subtotal plant

cost is $1,060,300, and the total construction cost is $5,301,510.

 Inventory costs include the cost of the molten salt held up in the

system and the initial cost of the NaK coolant. Total fuel-salt holdup

for this system is 63 3. At $1,Aeo/ft3, the charge is $89,460. Blanket-

salt holdup is 120 ft3, and at $56o/ft3 the cost is $67,200. TFuel- and

-~ blanket -salt charges do not include the cost of fissile material. The

400 £t holdup of NaX at $lOO/ft3 costs $40,000. The total inventory

cost is then $196,660, giving a total fixed capital cost of $5,L98,170.

Direct Operating Cost

The direct operating cost includes the cost of operators, chemical
consumption, waste containers, utilities, and maintenance materials. The
number of operating and support employees 1s based on a work schedule of
four shifts of LO hr each per week. These include immediate supervisory,
operating, maintenance, laboratofy, and health physics personnei, plus
two people.for routine clerical and janitorial work (Table 6).

No attempt was made to prorate the cost of higher supervisory,

clerical, or plant protectlon personnel for the procesgsing facility since

. some of this cost is included in labor overhead costs.
_59...

Table 5. Total Fixed Capital Cost

Installed process equipment
Structure and improvements
Interim waste storage
Process piping
Process instrumentation
Electrical auxiliaries
Sampling connections
Utilities (15% of installed process equipment)
Insulation (6% of installed process equipment)
Radiation monitoring
Total direct plant cost

Construction overhead
(30% of total direct plant cost)

Total construction cost

Engineering and inspection
(25% of total construction cost)

Subtotal plant cost
Contingency (25% of subtotal plant cost)
Total construction cost
Inventory cost
Molten fuel salt (at $l,h70/ft3)
Molten blanket salt (at $560/ft3)
NeK (at $100/£%°)
Total inventory cost

Total fixed capital cost

$ 853,760
556,770
387,970
155, 800
272, 100

8, 300
20, 000
128, 060
51, 220

100, 000

782,990

848, 240

1, 060, 300

89, 460
69, 200

10, 000

$2, 609, 980

$3, 392,970

$h, 241, 210

$5, 301, 510

$ 196,660
$5, 498, 170

....60 - !

Table 6. Employment Costs

| |  Cost
‘No. ($/year)
Production | ; . :
Shift supervisor | b 32, 000
Operators o S 16 96, 000
.'Malntenance workers - 8 | L8, 000
Laboratpry analysts ok 24, 000
Health physics workers 2. 12,000
Others | 2 10, 000
Total 36 222, 000

The cost of utilities, waste contalners, and consumed chemicals is
'f:'based on a 300 day/year operation for both the reactor and processing
plant. |

| Total direct operating cost for one year is $610,190 (Table 7);

this includes fuel and fertile salt makeup.
Processing Cost

The processing cost per year is estimated at $1, 447,570 (Table 8).
This cost is obtained by combining the direct operating cost, the indirect
cost of labor overhead (80% of direct labor cost), the fixed cost due to
depreciation (10% of fixed capital per year), taxes (1% of fixed capital
per year), and insurance (1% of fixed capital per year). The percentage
used. for the 1nd1rect labor cost is arbltrary, however, it is w1th1n the
range of usual practlce

On the basis of 300 days of operation per year for the 1000-Mw
(electrical) MSBR, the fuel -processing cost is 0.201 mlll/kwhr. The fuel-
cycle cost is composed of this coéfi plfis the in-reactor inventory of fuel,
fertile, and carrier salts, plus makeu@ fertile and carrier salts, and

233

less the credit for excess UF6 préduced In—reactor inventory, makeup,

and credlt were not conSldered in thls study.

- 61 -

Table 7. Annual Direct Operating Cost

Labor $222, 000
Chemical consumption

Fluorine (at $2.00/1b) $ 4,080

KOH (at $0.10/1b) 980

Hydrogen (at $0.01/7t°) 720

HF (at $0.26/1b) 1,000

NaF pellets (at $1.00/1Db) 5, 780

MgF, pellets (at $1.00/1D) 420

Inert gases (guess) 830

Fuel-salt makeup (at $l,h20/ft3) 28, L0oo

Blanket -salt makeup (at $560/ft3) 27, 350

NaX mekeup (guess) 830 70, 390
Waste containers , 28, 270
Utilities

Electricity (at $0.01/kwhr) 73, 300

Others (guess) 7, 000 80, 300
Maintenance materials

Site (guess) 2, 500

Building (at 2% of building cost) 10, 810

Service and utilities (at 4% of service

and utilities cost) 35, 880
Process equipment (at 15% of process
equipment cost) 160, 0Lo 209, 230

Total annual direct operating cost $610, 190

Table 8. Annual Fuel~Processifig Cost

Direct operating cost = $ 610,190
Labor overhead (80% of labor cost) B | 177,600
Fixed costs - | L o
Depreclation (lO%/year of flxed capital) o 5&9,820
Tax_(L% of fixed capital) | R     54,98@
_insfirance.(l% of £ixed capital . : '5h;980
Total B B : g1, W7, 570

Within the applled ground rules, these costs are belleved to be a
_reasonably‘accurate representatlon of the cost for regeneratlng the fuel

 and blanket 1n an 1ntegrated MSBR processing plant. A more thorough

o stbudy would have_lncluded_detalled design of equipment and layocut of the

g._integrafied processing plant, the reactor and its auxiliaries. Such a

_'thorough analysis was beyond the scope of this study.
CONCLUSIONS AND RECOMMENDATIONS

The central issues in this preliminary study were to analyze the
feasibility and cost of a conceptual system for continuously regenerating
the fuel and fertile streams in the Molten Sall Breeder Reactor. Briefly,
the system consists in (1) fluorinating, distilling, and reconstituting
the molten fluorides used in the reactor core, and (2) recovering the
233U from the molten breeder blanket by flucrination and using the
uranium to reconstitute the core salt. The excess is to be sold. ~ The
power of the breeder reactor was set at l,OOO Mw (electrlcal) for thls
study . | | | |
A number of basmc conclusions and essential recommendatlons were
developed The conclu81ons relate to the projected fea81blllfiy and. |
ant1o1pated costs in terms of establlshed technology and cost-accountlng
pracfilces, and The recommendatlons refer to what must yet Dbe learned W1th_
reSpect to technology and. chemical data before a complete englneerlng .

analy81s can be made It is our op;nlon that it w1ll.be.very_useful_to
- 63 -

begin filling these gaps in the knowledge because of the promising

simplicity and low cost shown by this study.

In the conclusions and recommendations presented below, the

recommendations are underlined. The first Tour numbered paragraphs

relate to general characteristics of this fuel-processing plant, and

the others to specific unit operations.

1.

FEASIBILITTY. Fluorination followed by distillation is a
feasible process for regenerating MSBR fuel (LiF-BeF,-UF)).
Fluorination alone is sufficient processing for the fertile

salt (LiF -ThF),). Reactor physics calculations indicate that
attractive breeding ratios can be obtained for such a process.
Engineering problems in the processing plant appear to be
amenable to solution through a well -planned developmental
program at the unit-operations level. Fluorination-distillation
should be developed as the processing method for the MSBR. Con-
currently, other attractive processes should be investigated at
the laboratory and/or engineering stage as potential alternatives.

INTEGRATED PLANT. Integrating processing and reactor facilities
is of primary importance in lowering the processing cost. Com-
plete advantage is thereby taken of the ready adaptability of a
fluid -fueled reactor for continuous processing with corresponding
minimum inventory. The relatively small size of this side-
strean processing plant, about 12 13 salt per day for a 1000 -Mw.
(electrical) reactor, is amenable to integrated construction,
thereby separating the economic dependence of the processing
industry upcn a large amount of installed electrical capacity.
The same {inancing convention that applies to the power plant
applies tTo the processing plant; this type of financing is
normally available at a lower rate than is available for a
separated processing facility.

ECONOMY. 'The estimated capital cost, excluding inventory, of
the plant is $5, 301,000, and operating costs are about $788, 000
a year. It is significant that the capital investment in the
integrated processing plant proposed here is only about W% of
the total cost of the reactor system it serves.

CORROSION. There are at least three areas in the chemical
processing plant in which corrosion behavior of construction
materials should be studied. These are the vacuum still, the
reduction unit, and filters. The still temperature of about
1000%C is much higher than has been contemplated for any other
part of the MSBR system, and the resistance of INOR-8 and
nickel to corrosive attack at this temperature is not known.
A reducer, hydrogen, enters the reduction unit and probably
helps limit corrosion there, but this should be verified.
Filters are subject to attack because of the large surface
area exposed to the fluid being filtered.

- 64 -

 FLUORINATION DEVELOPMENT. Batchwise Tluorimation of molten
- fluoride salts for uranium recovery has been rather thoroughly

investigated at ORNL; however, 1t is recommended that engineering
development of a continuous fluorinator be given high priority.
The need for continuous Tluorination is evidenced by the require-

~ment of low fuel and carrier ealt.inventories'in_the;proeessing"
plant. The reactor fuel system contains about 650 ft3 of salt,
and, without continuous  fluorination, the out-effplle inventory

could possibly be as large as the in-pile inventory. This
study indicates that in a continuous fluorination-distillation
process the holdup represents about lO% of the reactor Tuel

: volume. -

Itlsrammmmmdtmmimefmmahwfllcmmqm:mra
contlnuous Tluorinator be developed and demonstrated. ThlS
concept calls for a 1/2- to 3/%-in.-thick layer of frozen salt

- on.the 1n51de wall of the fluorinator to prevent the serious

Corr081ve attack by the molten mixture during fluorination.

_Ba81c information needed includes fluorination rate data, process

control in contbinuous operation, and method of establlshlng and

_malntalnlng & frozen wall.

Fluorlnatlon of the fertlle salt 1ntroduces probleme
81mllar to those encountered for, the fuel stream. However,
fertile-stream processing rates are 8 to 10 times higher, and

 the flSSlOprrOduct activity is several orders of magnitude

. less..  On the other hand, a lower fluorlnatlon eff1c1eney can

be tolerated in blanket processing.

Another method for continuous fluorination is the gas-
phase continuous operation in which Tluoride microspheres are
generated and fluorinated as they fall through a tower. This
process should be recognized as a potential alternative to the

- continuous method of fluorination studied here, but its

development should be subordxnate to that of the frozen-wall
concept

DISTITIATION DEVELOPMENT. The vacuum-distillation concept for
separating the LiF -BeFo (fuel-carrier salt) from fission products

~is feasible_from an engineering viewpoint. The theoretical net

discard of 'Li in the still waste is low epnough that its cost is
insignificant compared with other fuel cycle costs. Thermal

”7problems require that sufficient volume be malntalned for wetting
ca 1arge heat ‘transfer surface, and the bulldup of flSSlon products

in this volume will almost surely have an adverse effect on the
deeontamlnatlon factor of the distillate. Relative volatility
data are needed for the multlcomponent mlxture, LiF -BeFo ~figssion
products, in which compositions are in the range 99 5~O 50 mol %
to 84.5-0.5-15 mol % LiF ~BeFs-rare earth fluorides. It is |
strongly recommended that & conbinuous vacuum Sbill be built and

. operated to demonstrate a workable de81gn and o dbtaln rate and

J-entralnment data

10.

i1,

- 65 -

REDUCTION OF UF, TO UF,. This reaction is quantitative when UF6
is fed into an Fo flame contalning excess Hy, producing a
powdery EFM product. The solid product cakesg and adheres to
vessel walls, which should be avoided if possible in a remotely
operated system. The liquid-phase reactlion proposed here ig
more suitable for remote operation and should be developed. The
operating conditions that need study are: temperature, U, con-
centration, reaction rate, nozzle design, circulation rate,
contactor design, and gas-liquid separation.

The reducing conditions that exist in this operation are
consistent with those required for purging nickel, chromium, and

iron corrosion products from the fuel. Therefore the potentiality

of using the reduction unit for simultanecously giving the fuel a
final cleanup should be investigated.

SOLID -LIQUID SEPARATION. The general area of high-temperature,
solid-liguid separation in remote operations needs development.
Filtration techniques should surely be investigated to determine
operability and reliability in molten-fluoride systems.

FISSICN PRODUCT BEHAVIOR. A bpetter understanding of fission
product behavior throughout the processing plant is needed. In
particular, date are needed on the ways in which the variocus
nuclides partition in the several processing steps and on the
efficiency of removal. A more basic study is concerned with the
behavior of the fission products in the reactor environment to
determine whether or not certain nuclides remain in the reactor
system.

URANTUM HEXAFLUORIDE PURIFICATION. The NaF and MgFo sorption
units provide adequate decontamination for UFg. The batchwise
units can be operated satisfactorily for both fuel and fertile
streams of the MSBR; however, a coatinuous, temperature-zoned
system would reduce the frequency of cell entry. Probably the
largest uncertainties in UFg purification are in the removal of
tellurium and ruthenium; nmeans of removing them should be
developed.

The portion of the fertile stream used as fuel makeup need
not be passed through a sorption system because of the very
small amount of fission-product contamination. However, the
UF6 product that would eventually be handled by contact would
need purification.

PROTACTINIUM REMOVAL. The most significant advancement in
fertile-stream processing can be made in the development of a
process that removes protactinium. To be effective, the process
must remove protactiniue five to ten times as fast as its decay
rate; that 1s, the blanket volume would have to e processed
every four to eight days. Simplicity and ease of operation are
obvious requirements. Thus, a process based on forming protac-
tinium oxide by ilon exchange appears promising and should be

- 66 -

Effective removal would obviate the need of sending any
portion of the fertile stream to waste and possibly reduce
makeup cost to that represented by thorium burnup alone. Such
a process 18 not required for economic operation of the MSBR
since fluorination and fractional discard can adequately control
the fission product concentration; however, the potential
advantage of reduced waste cost and improved breeding performance
argues for basic development of a process.

12.  FUEL-STREAM WASTE SYSTEM. The most economical method of waste
mapnagement consists in bulk storage in a large heat exchanger
tank. The heat generation problem is so severe that the plant
waste must be diluted with an inexpensive, inert material, which
complicates future processing for reeoverg of any contained
values. The calculated loss of Ll and. U’in the fluorination-
distillation process is small, being only 1.5 to 2% of the fuel-
cycle cost.

The possible use of fission-product decay heat should not
be overlooked in the evalustion of an MSBR. The accumulated
waste generates about 4.5 Mw (thermal) throughout most of the
filling period of the waste tank. Thus salt-storage tempersture

“can be maintained high enough to make the waste tank a source of
high -temperature energy.

13.  FERTILE -STREAM WASTE SYSTEM. The fertile-stream waste contains
a significant inventory of valuable materials whose recovery is
probably warranted. At the end of the 30-year filling period,
the waste tank contains about 141,000 kg of Th, 10,400 kg of
114, and 116 kg of 233U; the “33U is isotopically pure material,
having been formed almost entirely out of the Tission zone.
Although not considered in this study, an in-cell decay period
of about six months followed by refluorlnatlon appears to be
advisable for greatly reducing the amount of “33Pa that enters
the waste tank.

14. PROCESS CONTROL. This aspect of plant operation was given only
a cursory review in this study, and no areas of unusual control
difficulty were observed. A flow-control device for the molten-
salt stream is needed, and the dynamic freeze-valve conceph
should be developed. Analytical and sampling regquirements
require a more thorough study than was given here. :

ACKNOWLEDGEMENT

W,IG. Stockdale assisted the authors in the development of the cost
of the plant and equipment, and his help is gratefully acknowledged.
10.

11.

12.

13.

1h.

15.

- 67 -
REFERENCES

. S. Bettis, ORNL, personal communication.

H. F. Bauman and E. 5. Bettils, ORNL, personal communication on
unpublished work still in progress.

M. J. Kelly, ORNL, personal communication on unpublished work still
in progress.

E. D. Arnocld, ORNL, personal communication on unpublished work.

R. E. Thome, Rare Barth Halides, ORNL-3804 (Mey 1965), p. L2.

C. B. Jackson (ed.), Liquid Metals Handbook, Sodium (NaK) Supplement,
3d ed., U.S.A.E.C., Washington, D.C., 1955.

K. A. Sense, M. J. Snyder, and J. W. Clegg, "The Vapor Pressure of
Beryllium Fluoride,"” J. Phys. Chem. 58, 223-k (1954).

O. Ruff, G. Schmidt, and S. Mugdan, "Arbeiten aus dem Gebeit Hoher

Temperatur. XV. Die Dampfdrucke der Alkalifluoride,"” Z. Anorg. Allgem.

Chem. 123, 83-8 (1922).

H. von Wartenberg and H. Schulz, "Der Dampfdrucke Einiger Salze. II,
Z. fur Elektrochemie 27, 568-73 E1921).

K. A. Sense and R. W. Stone, '"Vapor Pressures and Molecular
Compogitions of Vapors of the RUF-ZrF) and LiF@Zth Systems, "

J. Phys. Chem. 62, 1k11-8 (1958).

K. A. Sense and R. W. Stone, "Vapor Pressures and Molecular
Compositions of Vapors of the Sodium Fluoride -Beryllium ¥Fluoride
System," J. Phys. Chem. 62, 453-7 (1958).

M. M. Popov, F. A. Kostylev, and N. V. Zubova, '"Vapor Pressures of
Uranium Tetrafluoride," Zhur. Neorg. Khim. 4, 1708-9 (1959).

W. R. Grimes, Reactor Chemisgtry Division Annual Progress Report for
Period Ending January 31, 1960, ORNL-2931L (Apr: 29, 1960), p. 50-1.
W. L. Carter and J. B. Ruch, A Cost Estimate for Bulk Disposal of

Radioactive Waste Salt from Processing Zirconium-Uranium Fuel by the
ORNL Fluoride Volatility Process, ORNL-TM-9L8 (Sept. 25, 196k)
(Confidential).

D. 0. Campbell and G. I. Cathers, ORNL, personal communication of

unpublished data.
16.

17.

18.

19.

20,
22,

s
2l

25,
26,
| '(9E6'G?S to UF) in s Molten Salt, ORNL-TM-1051 (1965).
o7.
28.
29.
30,

- 31.

2.
33

;.68_-.

G. Burrows, Molecular Dlstlllatlon, p. 18~31 Oxford University Press,

Great Britain, 1960

- W. H. Carr, S. Mann, and E. C. Moncrief, "Uranium -Zirconium Alloy Fuel

Processing in the OBNL Volatility Pilot Plant," paper presented at
the A.I.Ch.E. Symposium;_"Volatlllty Reproceselng of Nuclear Reactor
Fuels, New York; N. Y. (Dec 1961) | e l

R. P. Mllford S. Mann, J. B. Ruch, and W. H. Carr, Jr. ; ”Recovefing

eUranlum Submarlne Reactor Fuels,” Ind. Eng Chem. 53, 357 (1961)

J. C. Mallen, Volatlllzatlon of Uranlum as the Hexafluorlde from
Drops of Molten Fluorlde Salt " - paper presented at the A C. S Natlonal
Meeting in Chicago, I1l. (Sept 2, 196L4).

G. I. Cathers, M. R. Bennett and R. L. Jolly, The Fused Salt-Fluoride
Volatlllty Process for Reeoverlflg Uranium, ORNL—2661 (1959).

ZM E. Whatley et al ,-Unlt Operations Sectlon Monthly Progress Report,

--September 1963, ORNL—TM~785 (1964).

"E. C. Moncrlef Results of Volatlllty Pillot Plant Cold Uranium
:'Flowsheet Demonstratlon Runs TU-6 and TU-7, ORNL-CF -61-9 37 (1961).
CW. WL Pitt, unpublished data, ORNL (1965).

R. W. Kessie et al., Process Vessel Design for Frozen Wall Containment

of Fused Salt, ANL-6377 (1961).

M. J. Kelley, ORNL, personal communication (196L).

L. E. Mcleese and C. D. Scott, Reconstitution of MSR Fuel by Reducing

C. D. Scott, unpublished data (1965).
W. 'L"Carter, R. P. Milford, and W. G. Stockdale, Design Studies and

Cost Estlmates of Two Fluoride Volatility Plants, ORNL-TM-522 (1962).

_ John Happel, Chemlcal Process Eeonomlcs, Wiley, New York, 1958.

jR S. Arles and R. D. Newton, Chemlcal Englneerlng Cost Estlmatlon,
TMcGrawafllll New York, 1955.

eC H. Chilton (ed. ), Cost Englneerlng in the Process Industrles,
'-Mcerawefilll New York,_l960 | | |
jW G‘ Stockdale, ORNL, personal communleatlon (1965)

.Sargent and Lundy Engineers, Capital Investment for lOOOAMwe Molten
"Salt Converter Reference De51gn Power Reactor, SL. 199M (1962)

34,

35-

36.

37

38.

39-

- 69 -

Octave Levenspiel, "Chemical Reaction Engineering," Wiley, New York,
1962,

W. R. Gembill, "Fused Salt Thermal Conductivity," Chem. Eng. 66 (16),
129-30 (1959).

B. C. Blanke et al., Density and Viscosity of Fused Mixtures of

Lithium, Beryllium, and Urénium Fluorides, Mound Laboratory Report
MIM-1086 (Mar. 23, 1959).
S. I. Cohen, W. D. Powers, and N. D. Greene, A Physical Property

Summary for ANP Fluoride Mixtures, ORNL-2150 (Aug. 23, 1956).

W. R. Grimes, Reactor Chemistry Annual Progress Report for Periocd
Ending January 31, 1963, ORNL-3417, p. L7. ‘ "

M. E. Lackey, Internally Cooled Molteg-Salt Reactors, ORNL-CF-59-6-89
(1959)-

APPENDICES
_73..

APPENDIX A. DESIGN CALCULATIONS FOR FUEL
SALT FLUORINATOR AND COOLING TANK

The unit operation of fluorinstion requires temperature contrcl at
about 55000 for the salt being fluorinated. The fact that this salt is
highly radicactive introduces a problem because 1t is necesgsary to extract
this decay heat through the walls of a flucorinator whose size is fixed by
other process requirements, such as throughput and residence time. If
there is insufficient heat transfer surface available for thlis purpose,
then the salt must be allowed to "cool" bvefore entering the fluorinator.
The solution to this problem is to insert a cooling tank immediately
upstream of the luorinator.

The following calculations pertain to the thermal design and size of
the fuel -salt fluorinator and the size of the cooling tank. It was
determined that the maximum permissible heat fiux for the fluorinator is
5.31 x J_olL Btu hr ftf3 and that the size is L.75 in. in diameter by
10.3 £t high. The cooling tank reguires a volume of 22.5 ftg, A Turther
result of this calculation is a graph of heatl generation rate at the
cooling tank exit as a function of elapsed time since discharge from the
reactor.

- The fluorinator design criteria are:

Fuel -salt flow rate 15 £ /aay
Uranium rate T7.3 g-moles/day
Fluorine utilization 33-1/3%
Maximum mass flow rate of gas 0.277 slpm/in.
Maximum heat flux through the frozen wall 1.5 kw/f{
Heat of reaction (UF) + Fp = UFg) -162 kcal/g-mole
Residence time of salt 2 hr
The Fg flow rate through the column is:
F=77.-3x 3= 231.9 g-moles,
= L.83 slpm .
Fluorinator cross-sectional area is:
.83 .2
A = 57T - i7.5 in.  ,

or the column diameter is:

D =/ .__?_‘-':%_Til;'é_: 4. 75 in.
The total column volume needed is:

15 %2 3
: V= = 125 067 .

Therefore, the column helght is:

1.25 x lhh
175

The area of the frozen wall per linear fodt of column is:

- 10.3 Tt .

h =

The open volume per foot of column 1s
| VO'_ 0. 120& ft3

.Therefore, the maxXimum heat removal rate 1n the frozenawall column is:

. 1.25 x 1.5 x 3&13 | u a4 -3
"_;_But, the heat of reactlon contrlbutes at the follow1ng rate:
77.3 x 162 -1 ft—g.'

 BT¥ 1.5 X 0858 - 1620 Bfi“ ar

_Thds, allowable fission product heating is:

5.15 x :LolP Btu hr“l ft“3 .

The surge tank must have sufficient capacity to allow the fuel salt to
L £e73. If the tank is assumed

to be well -mixed, and assuming that the heat generation rate of the salt

cool to a heat rate of 5.15 x lOLP Btu hr

can be expressed as an exponential function of time, then,Bu
H(t) = Keukt (heat generation rate), (A-1)
CE(t) = % ewt/T (age distribution function), (A-2)
where
B T = V/F, average residence time,
V= fluid volume in tank,
F = volumetric Tlow rate,
t = time since exit from reactor core,
SIK,k = constants. |

'The heat generation rate of molten salt from the exit of the well-mixed

surge tank can be expressed as:

Hy = wfm%(t) H(t) dtlg | :  (A-3)

O
_75_

The data on heat generation rate with respect to time elapsed since
the salt has been removed from the reactor can be expressed as a series of
equations of the form of Eq. (A-l). B8ix intervals were chosen from the
data in Figs. 2a and b (see text). In each time interval, the constants
of the exponential equation approximation were determined by coupling the
equation at the two ends of the time interval. This resulted in a repre-
sentation of the heat generation data, which was always equal to or some-
what greater than the calculated heat generation rate. The values of the

constants in the approximate equations are:

Time Interval

No. (hr) K k
1 0-0.0167 5.98 x 10° 24,2
2 0.0167-0.167 4.28 x 107 3.08
3 0.167 -1 .46 x 107 0.651
L 1-10 1.41 x 107 0.098
5 10 -100 5.85 x 10lL 0.0104
6 100 — 2.43 x 10LL 0.0015k4

After substitution of the constants, integration of Eq. (A-3) by time

segments gives:

.{ 5.98 x 10” [1 ) e-0.01667 (%-+ 2&.2)}
(= + 2h.2)

s

HE:

, 428 x 107 [§~0.01667 (% +3.98) __-0.1667 (3 + 3-98)]
T
(= + 3.98)

1
N ziu6 x 107 e-0.1667 (¥-+ 0.651) _ e—(%-+ 0.651)]
5 R 1
. 1iu1 x 10 _(z+0.098) _ _-l0(z+ 0.098)]
(?-+ 0.098) =

5.85 x 10“ [e-lo(%-+ 0.010k4) e~100(%-+ 0'01ou)]
(%-+ 0.010k)

2.43 x 10” [ -1oo(£-+ 0.0015A)]:}
+ il e T »
(F + 0.00154)
- 76 -

This equation was used to determine heat generation rate of the
molten salt at the cooling-tank exit in terms of average residence time
(Fig. A-1). At the design heat generation time, the average residence
time was 36 hr. ‘Therefore, the volume of the cooling tank must be:

_ 36 x 15 _ 3
V = —= = 22.5 1.
HE' HEAT GENERATING RATE AT COOLING TANK EXIT (Btu hr—1ff3)

6,0 x 10

ORNL DWG 65-7779 R1

@

2 T
/DESIGN RESIDENCE TIME | - e
10 20 30 40 50 60 70 80 90 100

T, AVERAGE RESIDENCE TIME (hr)

Fig. A-1. Heat Generation Rate of Molten Salt at the Cooling Tank
Exit. The salt is given 36 hours of decay time to facilitate temperature
control in the fluorinator.

- .LL?
;_78.; o

APPENDIX B. FISSION-PRODUCT HEAT CENERATION RATES
IN THE MOVABLE -BED SORBERS AND NaF WASTE TANKS

Heat generation rates for the following components are presented here:
MovableéBed Sorber | L
Sodium Fluoride Waste Tanks

ShortJIerm Coollng-Statlon
- for Waste Sodium Fluoride

Interlm-Storage of Waste

FlSSlon products, which are volatlllzed in the fluorlnator, accumulate

B in the movable4bed sorbers and create a heat source that must be considered

in the de81gn of these units. Excess heat must be removed so that it does
not 1nterfere with control of bed temperature Actual removal of the heat
1S:not_the problem_here.slnce_Sorbers,:whleh_accomplish this end, have been
designed_and used in the ORNL Velatilifiy Pilot Plant. :

p_. The essential problem is to estimate the heat generation rate due to
radioactive decay of the fission pro&uets present in the system prejected_
here. This was done by first assuming that half the fissionwprOQUCtfiheet__
generating capacity which reached the fluorinator would exit to the sorber.
Further, all the fission-product heat is concentrated in the lower zone in
the sorber, and it was removed in the NaF waste stream. As noted before,
one -fifth of the lower zone is exhausted per day, and there are two parallel
sorbers alternately operating for 12 hr each. This results in an accumula -
tion of decaying fission products or of decreasing heat sources in the
sorber, in the NaF waste tanks, and finally in the interim waste-storage’
Tacility. i

| The heat generation rate for various process components was approximated

by determlnlng the average heat generatlon rate during a specific time
perlod and by assumlng that this rate decayed as the total flSSlonmproduct
‘heat rate decayed as shown in Figs. 2a and b. The accumulated heat
generatlon rate could then be expressed as a rate characterlstle of tfission
-f*produets having an -average age_lntermedlate_between phe_oldest and the

'moet_reeently sorbed.
- 79 -
Movable -Bed Sorber

In the movable-bed sorber, the heat load can be approximated by
assuming that the volatile fission products from the fluorinator accumulate
for 5 days and that those first accumulated will decay or cool as additional
accumulation occurs. A table can be prepared for accumulated heat genera -
tion rate, derived from the residence time of fission products in the

sorber:

Reegidence Time of Fission

Products in Sorber Heat Generation Rate
~ (days) (Btu/hr)
1 144,000
2 120, 000
3 106, 000
L 96, 000
5 86, 000

Steady-state heat generation
rate in movable -bed sorber 552,000 Btu/hr

Sodium Fluoride Wagte Containers

The Nal' waste tanks accumulate NaF and fission products from the
sorbers. FEach tank holds two complete bottom sorber zones from each of
two sorbers {10 days' accumulation of fission producte in one-day incre -
ments ). These zones exhaust to the NaF waste tank each day for 10 days.
According to the slope of the fission-product decay heat curve (Figs. 2a
and b), the average residence time of the fission products, and the average
heat generation rate of the Na¥ bed material as it leaves the sorber, the
following heat generation rate exists in the NaF waste tank at the end of
10 days, at which time it is full.
Residence Time of NaF 3 . Average Heat

in Container Generation Rate
(days) (Btu/hr)
1 110, 300
2 183,400
3 1,500 -
4 62, 600
5 56,600
6 .53,600
I 50,700
8 k7,700
9 L1, 700
10 38, 700
- Heat generation rate in NaF .
waste tank when filled 616, 800 Btu/hr

Short-Term Cocoling Station for Waste Sodium Fluoride

The NaF waste containers are to be cooled for 80 days, within the
processing area. This calls for a cooling station with the capablility of
cooling eight NaF waste containers whose average age varies from about

7 days to 87 days. The following heat generation rates apply to the cooling

station:
Average Heat
Tdentification Number Generation Rate
of Containers (Btu/hr)
1 (about 7 days old) 616,800
2 | 380, 900
3 _ 261, 200
L | | | o - 210, 400
5 | | 163,600
5 145,100
I 119,700
8 (about 87 days old) 104, 000

Total heat generation at :
‘cooling station | 2,001, 700 Btu/hr
- 81 -
Interim-Storage of Waste

As mentioned above, the waste containers are to be cooled for a
minimum of 80 days by forced air convection prior to their transfer to
the interim waste-storage facility. The average heat generation rate of
one container at the end of this time (based on the slope of the fission-
product decay heat for this period) will be 104,000 Btu/hr.

An average of 28 containers are to be sent to this waste-storage
facility per year at one-month intervals — an average of 2.33 containers
per month. When the cell 1s opened for transfer, the most recently
filled container has 80-day-old material in it, whereas, the first filled
is about 110 days cld. On the average, the heat generating rate of the
transferred containers is characteristic of one that is 95 days old whose
rate is 87,500 Btu/hrg The average heat generation rate of containers sent
to the interim storage facility is:

2.33 x 87,500 = 203,900 Btu/hr .

The heat generation rate of these containers decays with storage

time, and the average rate in the interim facility for one year's

acceunulation is:

Residence Time of NaF Waste Average Heat
Container in Interim Facility Generation Rate

(months ) (Btu/hr)

1 203, 900

2 150,200

3 102, 000

Iy 83, 700

5 69, 800

6 62,200

Il 23,'(00

8 42,900

9 37,600

10 33, 300

11 29,000

12 26, 800

Average heat generation rate of a
one -year accumulation of containers 895,100 Btu/ar
lee -

For five years of wasfie.accumulation, the totai“héat géneration rate_::
in the interim facility is: e -
B '. | L Average Héat_

Age of NaF Waste Tanks ~Generation Rate
(years) - (Btu/hr)

 -?895;100.
© 381, 300
198, 900

| 132,600

5 116,000

o N

Heat generation rate from NaF _ o S e
waste in interim facility . 1,723,900 Btu/nr
- 83 -

APPENDIX C. ESTIMATION OF DISTILIATION
RATE IN VACUUM STILL

Here, calculations are given whereby the distillation rate of the
vacuum still is estimated. The configuration of the still is shown in
Fig. 9 of the text. The calculational method used is a modification of
the procedure for calculating the rate in molecular distillation as given
by Burrows.l6 For our étill, it is estimated that at lOOOOC and at a
pressure of 1 mm of Hg, the distillation rates for LiF and BeF
3 and 2.02 x 107 2,

For molecular distillation, a still is designed so that its condensing

o 2T,

respectively, 3.32 x 10~ g Sec"l cm
surface 1s located quite close to its evaporating surface, thereby
minimizing the transport distance for the vapor. I the separation
distance is small enough and 1f the pressure i1s low enough, a molecule
leaving the liguid surface has a very high probability of reaching the
condenser without ccolliding with another molecule. This is the essence
of molecular distillation. Our still cannot be described precisely as a
molecular still because 1ts operating pressure is too high; however, the
pressure is low enough that conditions for molecular distillation are
approached. TFor this reason, the calculational methods of molecular
distillation, modified to apply to pressures slightly higher than those
for true molecular distillation, are used here to estimate the distillation
rate for our still.

The theoretical rate of distillation of a single substance for such
conditions can be derived from the kinetic theory of gases, shown by Burrows
to be;

-2 -1 -2
5.83 x 10 P ,M?T g sec  cm (Cc-1)

where pmm is the equilibrium vapor pressure in mm of Hg, M is the molecular
welght in grams, and T is the absolute temperature in °K. The resulting
distillation rate is expressed as g sec™ cm ™.

The actual rate of distillation at low pressure will be less than the
theoretical rate because there will be collisions in the vapor space.
Burrows develcoped an expression for the factor by which the theoretical

distillation rate should be multiplied to get the actual rate. Summarizing,
-"8_4'-'

his treatment considers three events that can occur to the moleeule in the

vapor space:

Wtdy :

1. Some molecules can reaeh the condenser without a collision; this
fraction is e4K, where K 1s a dimensionless .factor depending upon
the distance between evaporator and condenser, the equilibrium .
mean free path of the molecules, and the shape of the evaporating
suriace. _ ' . T

2. The fraction of molecules that collide is (1 - e ™ }, and the % XK

o fraction of these molecules that reach the condenser is (l -e e
approx1mately ' '

3. From purely geometrical considerations, the probablllty of a
molecule's striking the condenser after many collisions, which
result in random motion, is a factor ¥, the ratio of condenser
area to condenser area plus evaporator area. The Tfraction of
these molecules that reach the condenser after meny colllslons
is F(1 - eX)(1 - eX). -

The total fraction L of wvaporized molecules that reach the condenser

- is given by the sum of the three fractions above:
f=e™ 4 (1 - e—K)e—K + F(1 - e—K)(l - e_K) ,
=7 o+ (1 - 1) (e Ty |
=1 -(1L -F)(1 - e‘K)2 - - (c-2)
When the theoretical rate is multiplied by f, defined to be the
evaporation coefficient, the rate of distillation for a single component
becomes :
- - 2
5.83 x 1070 20, [T 7T g osec s oem™ . (¢-3)

The evaporation coefficient I is not readily calculated because of

the difficulty in determining the proper value of K, which is defined to
be:

K = k—%E—— (C-4)
where

gap distance between evaporator and condenser,.

mean free path of molecules-in-eqflilibrifim vapef,

suitable consfiant, which is 1 or 1arger,_used tb relate actual
.conditions to average eonditions in the gap. _Burrows16 reporfie

ezperimental values of k in the range 3.5_£o 30.
gL

However, as K increases, the value of f approaches the value of F, a number
that is known from the configuration of the still. For the still shown in
Fig. 9 (see text), F >~ 0.58. An estimate of the size of K for this still

can be obtained for the two values of k given above and from the fact that

KE _ 2.3 x-1022 T , em (¢-5)
B ©
where
o = molecular diameter, cm,
T = temperature, OK,
P = equilibrium pressure, mm Hg.
Substituting into Eq. (C-4),
d pmm 02
K = : (c-6)

For LiF at T 1273%K,
c X 3.26 x 10'8 cm,

I

pm = l.O mi Hg)

and for this still, d ¥ 10 cm. Using these values, one finds that

K= 12.1 for k = 30 ,

and
K= 103.7 for k = 3.5 .
In either case, K is sufficiently large that Eq. (C-2) reduces to
f~FZ0.58.

i

In processing the fuel stream of the MSER, distillation must treat a
multicomponent mixture of LiFéBeFE-UFu—fission products. However, LiF and
BeF2 constitute more than 99 mole % of the mixture, as shown in Table C-1,
making it possible to treat the mixture as a binsry solution for purposes
of calculating the distillation rate.

Burrows has shown that for a binary mixture, Eq. (C-3), which gives

the distillation rate, becomes:

5.83 x 107 fPleAqffii7T g/sec , (c-7)
86 -

Whefe _ _
X = mole fraction of component 1 in the liquid mixture,
ST | : - 2
A = area of evaporating surface, cm_, _
P, = equilibrium vapor pressure of component 1 at temperature T, mm Hg,

. M, = gram-molecular weight of-componént 1.

A more rigorous representation of Eq. (¢-7) would include the act1v1ty
| -coeff1c1ent 71 of component 1 1nstead of' the evaporatlon coefflclent T to
account for_deVLat;ons from_ldeallty. _However,_the activity chfflclentS
afé fiot knofin.fbr these.éalt SOlutions, 80 fdr.this Stfidy it will Dbe. aséfimed
.that Eq (C-?) gives a reasonably'valld estimate of the distillation rate.

'~ Since the still operates at constant volume, a material balsnce
fequlres that _BeF2 and LiF:d;Stlll ‘at the same rate at which they enter
the still. The vapor composition is therefore about 69-31 mole % LiF -BeF,,
_the same ag it 1s in, the stream enterlng the stlll If the total pressure
:1n the vapor space of the stlll is kept at 1 mm Hg, the partial pressures

of LlF and BdF are O 69 and O 31 mm Hg, respectively, assuming that there

2

-Table.C—l.- Reactor-Data and Approximate Composition
of Fuel Stream at Equilibrium

Reactor Data

Fuel volume - 671 T3
Cycle time = 58 days
Power = 2160 Mw (thermal)

Approximate Composition (mole fraction)

 '233U 0.0029
-235U - | .Négligible
Other U __' - | 0.0002
S e S SR O
BeF, - |  0.318
Fission products | ~ 0.0008%

8. ' . '

~Gaseous fission products purged in the reactor circulating locp, and noble
fisslon products that are removed on a very short cycle by attaching them-

'selves_tq_the:INOReS walls, do not contribute to this value.
87

are no other volatile compounds in the still. This agsumption is wvalid
because the quantities of volatile fission products are quite small. If
we assume that Raoult's law applies, the partial pressure of BeF2 is:

p = P
BeF

at 1000°C is about T4 mm Heg; therefore the mole

BeF2 XBeFE

The vapor pressure of BeF2
fractions in the still liquid are approximately:

xBng = 0.31/Thk = 0.00k42 ,

and
Lar T 7 *per,, T 0.9958 .

By Eq. (C-7), the rates at which LiF and BeF,, evaporate from a

surface having area A can be determined;

Wi = 5.83 x 1072 A [ xP [M7T:]L1F g/sec , (c-8)
and
2
Voen = 5.83 x 107° fa[xp JW/T ]LiF g/sec . (c-9)

2
The vapor pressure of LiF at lOOOOC is 0.61 mm Hg. When this wvalue and
the values of the other quantities are inserted in Egs. (C-8) and (C-9),

the specific evaporation rates are found to be:

3.32 x 10”3 g sec ™ cm_2 for LiF,
and

0.02 x 107 g sec™ an™ for BeF, .

Adding these last two equations and solving for the area of the evaporating

surface:
"LaF " VBer,
A = s) ’ (c-10)
5.83x107 £[(xP [M/T) . + (@ /M7T)BQF2]
(436,600 + 359,000) g
A = (254 )(3600) > sec
N 2
(5.83xlo'2)(o.58)[(0.9958)(0.69)/2671273~P(o.ooue)(7u{[E77j§7§g)Secgcme
= 1720 cm2 5

~ 1.85 £6° .
88 -

 Now, the area of the pOOl surface lh-the efiill (see. Fig.-9 inetexfi)'is
_h 9 ft However, 1t is not safe-to say that the stlll is overde51gned
_by a“factor of about 2. 5. Better data on, vapor pressures and partlcularly

.3-act1V1ty coefflclents mlght make a con51derable dlfference 1n the calculated

' _dlstlllatlon rate. It is Strongly suggested that the stlll con81dered for

e'thls study be Vlewed only as an approx1mate de51gn that w1ll probably change

_as mere . is learned about the dlstlllatlon process. :
- 89 -

APPENDIX D. FISSION-PRODUCT ACCUMULATION AND HEAT GENERATION
RATE IN LITHIUM FLUORIDE POOL IN VACUUM STILL

Nonvolatile fission products accumulate in the vacuum still as the
LiF--BeF2 fuel carrier and velatile fission products are distilled away.
The still is initially charged with 4 Tt° of makeup 'LiF, evacuated to &
pressure of 1 mm Hg or less and adjusted to a temperature of about lOOOOC.
Fluorinated fuel salt, contalning fission products, is allowed to flow
continuously into the pool of LiF, and the still is operated so that the
rate of distillation is exactly equal to the feed rate. There is no bottom
discharge, and the volume remains constant. Kelly's work25 showed that the
initial Li¥F and BeF2

1000 from rare earths. Operation 1s continued at the above pressure and

distillate is decontaminated by a factor of 100 to

temperature until one of two phenomena forces termination: (1) Either the

3

solubility of fission products in the 4 ft” of LiF is exceeded and trouble-
some precipitation occurs, or (2) the accumulated heat generation rate from
fission=-product decay begins to tax the capacity of the cooling system.
When the distillation is terminated, the LiF~fission-product residue in the
still is drained to a waste receiver and eventually to permanent storage.
The operating cycle is then repeated.

The aim of the following calculations is to determine the operating
cycle for the still and the limiting conditions for the design shown in
Fig. 9. Since one design criterion is to process the fuel stream with
minimum out-of -reactor holdup, and since the solubility of the rare earths
in LiF at 1000°C is about 50 mole %, it is apparent that condition (2)
above will be controlling. Decay-heat removal will be a serious problem
long before solubility limits are approached. It has been determined that
the h»ft3 still can operate continuously for about 6T.h days at a distilla-
tion rate of 15 fts/day, processing fuel that has the heat generation
characteristics shown in Figs. 2a and b of the texi. The significance of
this number is that it represents the rate of 7Li discard to waste - 116 kg
every 67.4% days. The heat generation rate at this time is 31 x 106 Btu/nr,
the maximum that can be removed by the cooling system.

The calculations below are believed to give a conservative estimate

of the still performance. The basic data on fission-product heat generation
- 90 -

versus decay time represent gross values and hence do not exclude the
contributions of those fission products removed by the reacltor gas sparge,
by deposition on metal surfaces in the reactor system, or by the fluorina-
tion step in chemical processing. Furthermore, the 15-ft3/day design rate
is probably excessive for a 1000 -Mw (electrical) MSER; the economic optimum
rate is perhaps nearer 12 ft3/day. Also, no credit was taken for periodic
interruptions in processing due to the reactor's operating at less than
lOO% planfi'faétor. All these factors tended to shorten the still operating
cyéle and increase the discard cost. |

The calculations are arranged as follows: TFirst, an analytical
expression is derived for the heat generation rate of fission products
in the still as a function of elapsed time since discharge from the

reactor. Second, this equation 1s then used to evaluate the still design.
Analytical Expression for Heatl Generation Rate

When irradiated fuel salt is discharged from the fluorinator, it has
been out of the reactor about 38 hr. A cooling tank downstream from the
fluorinator adds another 24 hr of holdup, so that the salt is about 62 hr
old when it reaches the still. The heat generation characteristics of
the salt durling this periocd were calculated by assuming an infinitely
mixed system; these calculations are described in Appendix A, and the
results are shown graphically in Fig. A-1. Continuous operation is
assumed. throughout the system, and, by the time the fuel reaches the
still, its heat generation (Fig. A-1) is aboub 4.45 x 10° Btu hr~t ££73.

The still is a sink for fission products that exhibit decay behavior
like that shown in Fig. 2a and b. Initially at t = 62 hr = 2.58 days,
the 4 £t of LiF in the still is fresh material having a zero heat genera-
tion rate, but this condition changes rapidly when the still is put on-

stream. Flow into the still is continuous, and the rate of healt generation
 Wil1 rise until flow is stopped. The magnitude of thié rate at any future
time T is therefore an integrated quantity over the accumulation period:
to'z 2.58 days to T. To describe the behavior of the still during this

~period, define the quantities
o -

o} = gpecific heat generation rate of fuel entering still,
- b5 x 107 Bt hrt £,
q(t) = specific heat generstion rate at time t, Btu hr — £t75,
¥ = flow rate of fuel into still,
- 15 £t3/day,
v(t) = fuel volume processed at time t,

T = time that still has been operating, days,

<5
i

volume fuel pool in still,

= L £g3,

The heat generation in the still pool at any time t is V, qs(t), expressed
as Bubt/hr, which is the difference between what the rate would have been
if there had been no decay and the amount the rate has decreased because

of decay. That is,

v, ay(6) = [ o) av(s) - [ a(s) avls) - (D-1)
v v
Noting that
dv = F at , (D-2)
there results
v, oa(t) = qufdt -qu_(t) dt . (D-3)
t T

The limits of integration extend from to = 2.58 days to t = T days, where
T denotes the time after discharge from the reactor at which the heat
generation rate is desired.

To treat BEg. (D-3) analytically, the function q(t) is obtained from
Figs. 2a and b by representing the curve on this graph by four straight-
line segments in the range 2 to 400 days. The procedure is diagrammed in

Fig. D-1. The general form of the equations is:

a(t) = kt" .
The slope n of each segment is determined from values of q(t) read directly
from Figs. 2a and b. The initial conditions of the fuel entering the still,
namely, t = 2.58 days, q = Lk.U5 x :{_olL Btu hr
the equation for the first segment to determine the constant k.. The

1
constant k2 was determined similarly by using the end condition of the

ft"3, were introduced into
HEAT GENERATION RATE (Btu hr"]fi"3)

 ORNL DWG 65-3023R]

i
|
|
l -
|
|
|
|
!
|

|
%
L L
10 20 100 400
TIME AFTER DISCHARGE FROM REACTOR - (days)

_ Fig. D-1. Schqmatic Logarithmic Curve Showing Approximations to
Fission~-Product Decay. In the mathematical model for computing the heat
generation rate, 1t is convenlent to divide the time scale into four parts.

-6 -
_93_

first segment at t = 10 days. This stepwise procedure was followed for

segments two and three to evaluate k., and Kh' The four equations so

3

determined are:

-0.43

q; = 6.69 x 10” ty 2 <%, =10 days , (D-L)
a, = 9.686 x 10" £, 70" 27 10 < %, < 20 days , (D-5)
ay = 2L.2 x 10" t3“0'85” 20 < ¢, < 100 days (D-6)
q, = 194.8 x 10" ), 7" 992 100 < ) = 400 days . (D-7)

In these equations, g is in Btu hr"l ft_3, and t is in days.
The function q(t) in Eq. (D-3) is replaced by the four separate

functions of Egs. (D-4 to (D-7), and the integration is carried out to

give:
v_q () = 15 x 10% [4.h5 (T - 2.58) - 11.7k (th.57 - 2.589-27)
) 23.62(t20‘”l _ loo.ul) _ 1&5.2(t30'1”6 ) 200.1&6)
+ 581.5(tu-0°335 - 100'0'335)] . (D-8)

The units of T, © t.,, and t, are days; the units of V q are
by 58

1’ o 3
Btu's/hr. The restrictions on the several t's are those specified for

Egs. (D-4) to (D-7). 1In solving Eq. (D-8) for v, qS(T), a value is chosen
t

for T, and this value is assigned to either t or th in the

1 B by
fellowing way: For *
2.58 < T < 10 days, let b= T
10 < T < 20 days, let t2 = T ;
20 < T = 100 days, let t3 = T ;
100 < T < 400 days, let th =T .

All ti < T assume their maximum values; all ti > T, of course, are not

A

considered. Filgure D-2 is a plot of Eq. (D-8); it gives the integrated
heat generation rate for times between 2.58 and LO0O days after discharge
from the reactor. Note that the accumulation time for fission products

in the still is (T - 2.58) days.

- oL -

ORNL DWG 65-1821 R2

108

[ e

o o o~ 0w
:
‘\

S

f—
o}
~!
-
-
el
o
[2.0)
b

o L I

HEAT GENERATION RATE (Btu/hr)
|
{

(=]
o

o~ D

X 10-1"

511 | ! -

1 2 ; 4‘ 5'878910 2 3 45678L|02 2 3 456785103
0 at

2.55 d TIME AFTER DISCHARGE FROM REACTOR (days)

10

Fig. D-2. Heat Generation Rate in the LiF Pool Resulting from
Pission-Product Accumulation in the Still. The still is charged with
L £t3 of fresh, molten LiF at the beginning of the distillation cycle.
Barren fuel-carrier (2.58 days old) flows into this LiF at a rate of
15 £t3/day, and LiF -BeF, distills at the same rate, keeping the volume
constant. Accumulatlng fisslon products cause a rapid increase in the
- heat generation rate.
- 05 -
Evaluation of Vacuum-Still Design

Two principal conditions had to be satisfied in the still design:
(1) the evaporating surface had to be sufficient for obtaining the required
15—ft3/day distillation rate, and (2) maximum heat transfer surface had to
be provided to minimize the frequency with which the still is drained to
the waste tank. The choice of dimensions was somewhat arbitrary; a tube-
sheet diameter of 2.5 ft was chosen because 1t gave an evaporating surface
that was about 2.5 times the calculated areca, and a closely spaced arrange=-
ment of 2.5-ft-long by l/2-in.udiam tubes was used for high cooling-surface-
to-volume ratio. The primary unknown operating condition is the length of
time that fission products can be accumulated before the integrated heat
generation taxes the capacity of the cooling system; an estimate of this
time is determined in the following calculations. The schematic diagram
in Fig. D-3 depicts the still operation and summarizes calculated performance
and physical data.

The calculations which are described %below indicate that the still
can accumulate fission products for about 67.4 days in a h—ft3 LiF pool
when the gtill 1s operated continuously at & feed rate of 15 ft3/day. The
discard rate for LiF is therefore a very small fraction of the processing
rate, being about O.h%. As mentioned above, the heat generation rate
predicted by Eq. (D-8) is probably excessive because gross instead of net
fission-product data were used; therefore this accumulation time can be

treasted as & lower limit.

Physical Data for Still

Applicable data are given on Fig. D-3.

Heat Transfer Characteristics

The 1000°C temperature of the IiF pool in the still was chosen to
achieve adequate distillation rate and fission-product solubility. The
pool transfers heat to the tube walls by natural convection, and this heat
is picked up on the outside of the tubes by NaK (22.3-77.7 mole % Na-K

eutectic) under forced convection. The NaK coolant enters the still at
. STILL
BISTILLATE
15 ff3/dc|y

0 .
VACUUM
SYSTEM

FLUORINATOR
~550°C

NaK
750°C

LiF-BeFz-FP

STILL FEED

..96.,.

CONDENSER
COOLANT
(AIR OR NaK)

b e e

| LiF-BeF3
DISTILLATE
15§32

FEED TO
UFg—eUFy4
REDUCTION

COLUMN

Fig. D-3.

(5§13 /day

NaK
750°C

WASTE
RECEIVER

4§13
67 DAY
HOLD-UP

(900-1000°C)___

ORNL DWG 65-3040

LiF FROM
OUTSIDE
MAKE-UP

LiF
MAKE-UP
413

AIR OR Nak

NaK (550° C)
COOLANT
i153 gpm (max)

NaK (550°C)
COOLANT
1153 ngrn {max)

DRAIN TO

UNDERGROUND

STORAGE
LiIF&FP

STILL AND WASTE RECEIVER DATA

817-Ypin. x 16ga. INOR-8 TUBES

TUBE SURFACE =267 §12

TUBE SHEET =2.5ft Dio.x2.5¢f High
. STILL CONDENSER =69 fi2

HEAT LOAD {max):31x108 piyshr

EVAPORATING SURFACE = 4.9 §12

TUBE SPACING = %5 in. A CENTERS

Schematic Diasgram of Vacuum Still Operation.
- 97 -

SSOOC and exits at 75000, giving a logarithmic mean temperature difference
of 340°C (612°F). Heat transfer characteristics of LiF were calculated by
using the Nusselt eguation for natural convection, in conjunction with the
physical-property data given in Appendix F. A Nusselt-type equation for
flow normal to banks of tubes was used for calculating the heat transfer
properties of the coolant. Since the total heat dissipation requirement
of the still was unknown (because fission-product accumulation time was
not known), the coolant heat transfer coefficient was expressed as a
function of the heat generation rate. It was determined that the still
could dissipate about 31 x 106 Btu/hr for the NaK flow conditions shown

in Fig. D=3. The overall heat transfer coefficient.for maximum heat flux
is about 190 Btu hr ™~ ££™= °%F L. From Fig. D-2, the time after discharge
from the reactor corresponding to the above integrated heal generation is
70 days. The fission-product accumulation time is 2.58 days less or 67.4

days because of the time lag before fuel reaches the distillation step.

Li¥F, BeF,, UF,, and Fission-Product Discard Rates. — At the end of

the 67.4-day cycle, the still contents are primarily LiF and fission
products. The equilibrium BeF2 concentration was estimated above to be
only 0.4 mole % because of its relatively high vapor pressure atl 1000°C.
If it is assumed that the fluorination step is 99.7% efficient for uranium
removal, then in 67.4 days about 3.64 kg of U will have entered the still.
Uranium tetrafluoride has a vapor pressure at 1000°C about 2.3 times that
of LiF, so a portion of the UF)Jr will be recovered in the distillate. The
amount recovered cannot be calculated until more is known about the vapor-
liguid edquilibria of multicomponent melten salt mixtures. Filssion-product
accumulation during this period is aspproximstely 190.1 kg; most of the RbF
and CsF of this inventory will distill becsuse of relatively high vapor
pressures. *

An estimate of the inventory relegated to waste every 67.4 days is
given in Teble D-1. A reactor plant-factor of 100% was assumed in the
calculation. These values are based on ideal-solution behavior, particularly
with regard to LiF and BeFE; this is almost surely not the case for this
mixture. As more is learned about the activities of the components, it can
be expécted that the compositions of Table D-1 will be different from those

shown.
=98 -

Table D-l. Estimated Comp081tlon of the Vacuum Still
After 67.k Days' Operation &t 15 £t 3 /Day

Still volume = H £

Weight Mole

(kg) Fraction
@3By 33 . <0.002
235U RO o <05 S
ey | T g
wr 19509 0.850
BeF, 2.1 0.005

 Fission products®™ - 190.1 0.143

_aMolecular weight.= l50,~a$sumed.

Heat Removed by Stlll Condenser. m'The load on the still condenser

s relatlvely small, con81st1ng only of the latent heat in 15 ftB/day of
"a 69-31 mole % LanBéFg mlxture plus some radiation from the pool surface.
 The latent heat of vaporlzatlon of LiF was estimated from vapor pressure

‘data to be MM,OOO cal/g-mole; an experimental valuell of 50,100 cal/g-mole

was used for_BeFe. Radiative heat transfer amounted to about 62, 700 Btu/hr,

giving a total condenser duty of 198,300 Btu/hr.

' _For smooth, nonflashing disfiillation, the condensate is the LiF-Bng

eutectic having a melting point_about 5OOOC, and. a condenser temperature

| slightly_higher than this is éati#factory. - IT there are deviations from

.ideal-opération so that a higher@melting composition distills, the con-

-_défiéer tempera£ure could be adjusted TO temperatures slightly above the

end 845°C for LiF.

| melting poi@ts of the pure components, 803°C for BeF,
2 99 -
APPENDiX E. DESIGN CALCULATIONS FOR WASTE-STORAGE SYSTEM

Separate storage is provided for fuel and fertile stream wastes as
fTluorides in underground tanks designed for bulk accumulation over a 30-
year period. Waste management in the post-~-30-year period was not con-
sidered in this study. However, as mentioned in the text (see Table 2),
it will probably be desirable to reprocess the fertile-stream waste, which
is only mildly radiocactive, at some future time for recovery of thorium,

233

lithium, and uranium values. About 116 kg of U will be present. Fission
products, separated in this recovery, could be stored longer, depending on
the activity.

- In the fuel-stream waste, the only significant value, other than the
possible future value of individual fission products, is 7Li. However, as
explained below, it is necessary to add a mixture of NaF -KF to this waste
to facillitate heat transfer. Since these compounds are chemically similar
to LiF, recovery of the lithium is difficult. In any event, at the end of
the 30-year period, the desirability of recovery would have to be analyzed
in light of the prevailling costs.

Two basic problems must be solved in designing the waste storage
system: (1) the integrated rate of heat generation by fission ~-product
decay must be determined, and (2), using the results of (1), the most
economlc deslgn for the prevailing conditions must be found. The heat
generation rate is computed Trom the fission-product decay behavior
exhibited in Figs. 2a and b (see text), and the results are shown in
Figs. E-1L and E~-2. A previous study by Carter and Ruc:hll!L examined a
similar waste-storage problem, and, in accord with their recommendation
for economic waste management, bulk storage in large, heat-exchanger tanks
is adopted for our wastes. Conceptual designs of the waste facilities are
shown on Dwgs. 58080 D and 58081 ¢ in Appendix F.

Over the 30-year period, 784 £
This volume includes 520 ft3 of LiF—fission-product mixture drained from

the processing cell, plus 284 ft3 of NaF -KF' diluent. The storage tank is

of fuel -stream waste are collected.

16 f+ in diameter and 6.33 4 high. The corresponding volume of fertile-
stream waste is 1783 ft3; this is stored in a tank 13.5 ft in diameter by
13.5 £% high. ‘
. -100 -

Most of the following calculations are concernéd with the design of

t"tne waste tank for fuelustream effluent because tne most dlffloult deszgn:

- problems ooour for thls waste After the . deoay eharaoterlstlcs of the

j flSSlon produots and 1ntegrated heat generatlon rate are determlned the
-lcaloulatlons examlne, 1n the follow1ng order, the ba51c features of tank _,]fl
”de51gn, ‘maximum allowable heat generatlon rate, volume of. dlluent waste-
'tank deelgn, and the. underground storage fac111ty : Once these calculatlons-

are. made, oomputatlons for the fertlle-stream waste are almost 1no1dental

_.:Fuel¢StreandWaete Systen

'Decay'Characteristicsdof'the Fission.Products

The 1n1t1al problem 1n deszgnlng the waste tank is to determlne the

'z'deoay in terms of the tlme-related oharacterlstlcs of the salt belng added

"e'ito the tank That lS to say} tne stlll bottoms, which represent an accumula~

| '--tlon of flSSlon products hav1ng ages in the range 2.58 to 70 days, generate

d,neat at a rate onaracterlstlc of flSSlon produots hav1ng an average 'age

=somewnere between these values Tne tlme-related benav1or of tne 1ntegrated
.f.speclflc heat generatlon rate is shown in Flgf L1l of the text for the case
of'no inert salt dilution. The graph covers a 5-year collection period of

h-ft3 batches every 67.4 days " The initial point on the graph begins at

-.about lhO days, this 1is the accumulated time in the processing plant with a

-reference time of zero taken as the day the fuel is dlsoharged from the
_nreactor An 1n-cell delay perlod of 67.54 days after removal from the still

1s 1ncluded in this time to permlt some initial coollng before dralnlng to

E 'the underground tank

The decay curves of Flgs Eadand b oan_be.used to'determine tne.decay
:__behav1or when.thls average age'has been fOund.d'Tne volume of fuel salt

'ffrom Whlch flSSlon produots have accumulated 1s

15 ft3/day X 67 u days f 1011 ft3

'~”uIt was determlned in Appendlx D that the heat generatlon rate at age equal

o 6 -
-:to 10 daye out of the reactor is 31 x-10 Btu/hr Therefore on the ba51s
d_of unlt volume of core salt the average heat generatlon rate 15 o

6
7f9$@;1£f%%L.w31fi6x;&%%uhr f@3,-~
- 101 -

Define the average heat generation,

/\q dv

Qg = (2-1)

where v is the volume. The processing rate F, in ftB/hr, is steady, so in
the time interval dt,

dv = F dt . | (B-2)
It was shown in Appendix D that the instantanecus heat generation rate is
represented by an equation of the form:

-1 1

, Btuhr™ £t (E-3)

Combining Egs. (E-2) and (E-3) in (E-1):

tavg
k /\ £+ gt

O

qavg B t ?

avg
[%

O

. [tu-n) o) |

B avg 0
Yavg T (1 - n)(t,  -t,)

q =kt

: (B-4)
g
where k and n are characteristic constants for the decay curve; to (= 2.58
days ) ig the age of the fission products at the beginning of distillation.
The desired average age is tavg’ the time corresponding to qavg'

It will be recalled from the discussion of Appendix D (in particular
Fig. D-1) that the q-versus-t curve is best represented by four segments
over the range of interest. The four equations are:

4 t~o.u3

q; = 6.69 x 10 1 2 < tl < 10 days, (E-5)
a, = 9.686 x 1ou t50'59 10 < t, < 20 days, (E-6)
a3 = 21.2 x ].Ol‘L t§O.85h 20 < t3 < 100 days, (BE-7)

q, = 194.8 x 10* t£1'335 100 < t, < 40O days. (E-8)
- 102 -

If t ' lles 1n the rauge 2 to lO days, then 1ts value ean be found dlrectly
.L_from Eq (E &) by utilizing k = 6 69 x th and n O 43 from Eq. (E~5) |
. However,_lf % fg'ls in the range lO to 20 days, Eq. (E-M) coitalns a second -
term in the numerator 1nvolv1ng the constants k 9. 686 x 10" and n = 0. 59

| of Eq. (E~6) | Slmllarly for t ve in the range 20 to 100 daye, the. numerator
. of Eq. (E-h) contalns a thlrd term 1nvolv1ng the constants of Eq; (E T)

"t'The solutlon is ea51ly determlned by trlal, and for thls case the proper

:;Vform of Eq_ (E—h) contalns the constants k and n from Eqs (E—B) end (E-6) T'

ne o L ___ "'k". : . .
o "1, 1-n 4n 2 1-n " ~n
SR 1 - l 2 2
o avg e
6 69x10“ | 0. 57 | o 57 9 686xlolL .. 041 o;ul -
*w-————-(10 2 58 ) (2 71070
3 066 lou 0. 57 .41 ave
avg |
.The solutlon To: thls equatlon is: _tévg = ll 6 days An average age near

“the 1ower end of the time ecale would be expected because of the 1arger

..t'contrlbutlou by‘the younger flSSlon products to heat generatlon  The

quantity tlo':_lo days used in Eq. (E-9) denotes the upper limit of

'Eq. (E-5) and the lower limit of Eq. (E-6).

.:_Eutegrated Heat Generation in Waste Tank

3

~When the vacuum stlll is taken of T stream and dralned its L-ft
volume is generatlng heat at 31 X 106 Btu/hr, but this rate is decreasing

1n 8 manner characterlstlc of ll 6~day~old fission products, as shown on

- Fig. Eb At this point, the change in rate is rather rapid, 1nd1cat1ng

"that a, short delay tlme 1n the cell before dralnlng to underground storage

w1ll apprec1ably alleviate de51gn requlrements for the 1arge tank.

3

'Installatlon of a second T vessel qulte 51mllar +0 the vacuum stlll

:-fllS 8, eouvenleut way of pTOVldlflg a 67 Mnday holdup, durlng this tlme the

't;_7heat generatlon rate decreaees by'a factor of about h 5.

A decay curve was. calculated for a typlcal batch of stlll re31due by

'toeiaesumlng that the curve was parallel to the decay curve of Flg. 2b for MSER

flSSlon producte after ll 6 days. The 67 h-day in cell delay makes “the

& accumulated tlme from reactor dlecharge to underground storage equal to -
- 103 -

137.4 days; at this time the heat generation rate of the h-rt3 batch is

6.8 x 106 Btu/hr. The curves of Fig. E-L are plotted to show the thermal
history of the underground waste tank over a 5-year collection period
during which 27 batches are added; alsc the decay in the post-b-year
collection period is shown. The curve with positive slope represents the
buildup of heat generation rate from all batches accumulated up to the
indicated time. This curve is actually a stepped curve, but for convenience
it has been drawn smooth through the maximum point of each step. The waste
tank is designed for 30 years' collection, and the accumulated heat genera -
tion at this time can be obtained with little error by extrapolating the
positive-sloped curve of Fig. E-l.

Figure E-2 has been included to show the integrated heat generation
rate in the underground waste tank when the still residue is drained into
the tank immediately upon completion of the distillation cycle.

Figures E-1 and E-2 represent an upper limit for fission-product heat
generation. As pointed out earlier in this report, the values are for
gross fission-product decay, which includes nuclides that have been removed
prior to wvacuum distillation. Also the curves were calculated for a

reactor plant factor of 100%.

Basic Features of Tank Design

A previous studylLL showed that bulk storage in large tanks is the most
economical management for fluoride wastes. A 30-year periocd was chosen
because it coincided with the amortization pericd of the reactor plant, and
g single tank is sufficient because the overall waste volume is small. For
economy and reliability, cooling by forced air draft was adopted, and an
upper limit of 60 fi/sec was assigned to the velocity. Waste temperature

in the tank is not to exceed 750°C.

Meximum Allowable Heat Generation Rate

Under the above ground rules it is apparent that there is an upper
limit that can be tolerated for the volumetric heat production rate, and
that the waste will have to be diluted to wet sufficient cooling surface.

In 30 years at 80% reactor plant factor there will be 520 ft3 of radiocactive
o - : _ ORNL DWG 65-1817 Rl
A S E I B B N l T T T

I

— ' _ e ACCUMULATED SRR
' ' ' : e HEAT GENERATION :

.27 BATCHES ACCUMULATED
~OVER 5-YEAR PERIOD AT
. 67.4-DAY INTERVALS =~
4 BY/BATCH

HEAT GENERATION RATE (Bt/hr)
. _

1 YEAR 2 5.0 a0
4 ! ) Il Pl !]| | | 'i ! ll } |. 1|[ |
10 | L _ . S

102 B 103 - | RN 104 S 10

TIME AFTER DISCHARGE FROM REACTOR {days)

' . Fig. E-1. Decay Curves for Fuel-3tream Fission Products Cooled
67.4 Days Before Draining to Waste Tank. The total heat generation rate
in the underground waste tank is significantly decreased by the 67.4-day
~ holdup in the processing cell. Compare with Fig. E-2. Waste is from an
~ MSBR operating on a 58-day cycle at 2160 Mw (thermal). T
- 105

ORNL DWG 65-1808

RN

F T 71Tl

0000004

| I I B |

ACCUMULATED
—%AT GENERATION
i i ii 0000 XK

! T TTL

T T TT

108

T

HEAT GENERATION RATE (Btu/hr)

q

P T R T

N

BATCH {4 #1% cach)

27

T 11T

¥ YEAR

2
£ !tfi%%i!l?

E J

G 5 24

Figa E—Eu
Tank Without Prior Cooling.

103

TIME AFTER DISCHARGE FROM REACTOR {days}

Decay Curves for Fuel-Stream Fission Products Sent to Waste
When the still bottoms are drained directly to

the underground waste tank, the total heat generation rate is about 2.8

times that shown on Fig. E-1, in which a 67.4-day cooling is allowed.

cycle time and power of the reactor is the same as for Fig. E-l.

The
n_lo6_q

waste salt accumulated whose heat generation rate 1s about L. 76 X 107 Btu/hr,
obtained by extrapolating the upper curve of Fig. E-L. Allow1ng the coolant

to undergo a temperature rlse of 125 C, the ailr regquirement is:

Btu 1o F X. 1

This 1is equlvalent to 85,190 ftB/mln at an average temperature of 87 C.

1.76 x 10/ = 3.13 x 107 1o/hr .
| The requlred Cross sectlon for alr flow is:

85 100 £63  sec oo
e e - 23.7 £t .

 ThlS area can be ebtalned w1th 89M tubes 2.5 in. in diameter, with a 9-gage
wall, in the flnal &e51gn, Draw1ng 58080 D, Appendix ¥, 937 tubes were used.
These tubes, arranged in a Uatube eonflguratlon, can be accommodated in a
'l6wftmd1am tank |

| Waste salt is stored on the shell side in the tank. The available

- volume per foot of tank lS:

2

LE . (No. tubes)dt e
tank

H

C@0e)® [ (e =325 | ar p a3
o [l (16 x 12)° ] e/

The heat transfer surface per foot of tank height is:

£° >
= (2)(937)(0.577), F—= 108L.3, £t7/ft .

- = S 2.
Therefore, the surface avallable to each cubic foot of waste 1s 7.88 £t /ftB.

“The'mean temperature difference between coolant and salt is 65100, and

the overall heat transfer coefflelent is estimated to be 6.2 Btu hr -1 ftmg

O, =1

TFT The maximum allowable heat generatlon is then:
6.2 gti 7 38 gfi (651)(1 8) °F = 5.75 x 1oh Btu hr“l;ft‘3 .
hr ft F ffi ' D

_'rAt all tlmes during waste aocumulatlon, there must be suff1c1ent volume
o present so that the integrated heat generation (Btu/hr) divided by the
total volume (ft3) does not exceed this figure. |
- 107 -

Volume of Diluent

The total volume of fluid in the waste tank at any time is the summation
of individual batch volumes from the process plus the required diluent.

Expressed mathematically,

N
v(t) = % (h+7V )
n=1
where Vdn ig the reguired diluent volume for the nth batch, and the summation

is carried out over N batches. The above paragraph shows that the total

volume at any time is the quotient of':

integrated heat generation rate (Btu/hr)

allowable volumetric heat generation (Btu nr ft—S)

a(t)
= I
5.75 x 10

The required diluent volume for each batch can now be determined by solving

alt) - 3 (h+v, ) .
5.75 x 10 n=1
The value of q(t) corresponding to the nth batch is read from Fig. E-1.
It is apparent that the largest diluent volume is required when the
first batch is drained to the waste tank. Eventually the tank contains
sufficient volume so that no further inert diluent is required. This
behavior is shown on Fig. E-3. The total volume of diluent is 264 ft3,
meking the 30-year volume of waste plus diluent equal to 78L fts. In
actual practice, the total diluent volume would probably be added at the
beginning of waste collection rather than in discrete steps, as shown in

the figure.

Waste Tank Design

The 937 U-tubes are installed in the tank with one end open to the
interior of the vault, and the other end welded into an exhaust duct
leading to the stack. Air forced into the vault passes over the outside
of the tank before entering the U~tubes. This design provides about

6100 ft2 of tube cocoling surface, which is about 2.5 times the calculated
108 -

"ORNL DWG 65-3042

400 =

" ACCUMULATED 3
TOTAL VOLUME

300

i £ ACCUMULATED
5 DILUENT VOLUME

200

VOLUME (")

100 e

o ' TIME AFTER DISCHARGE FROM REACTOR (days)
Fig. E-3. Proportions of Diluent (NaF -KF') Required to Tnsure Suitable

 H Heat Transfer from Waste Tank. First batch of fuel -stream waste requlres
*'largest proportion; no dilution requlred after Thk days . . |
- 109 -

requirement. The relatively large number of tubes and the small waste
volume lead to a tank that has a low height ~to-diameter ratio. The tank
is 16 £t in diameter and 6.33 ft high and has a storage volume of about

86O(ft3. Monel was chosen asg the structural material.

Underground Storage Facility

The underground storage area is shown on Drawing 58080-D, Appendix F.
In addition to the vault for the fuel -stream waste tank, the area contains
a storage vault for solld Na¥ and Mg’F2 wastes from the UF6 sorption step.

The design of this portion of the waste system is discussed in Appendix B.
Fertile Stream Waste System

Design bases used for fertile-stream waste-storage were:

Thirty -year accumulation in a single waste tank.

Fertile stream power of 62 Mw (thermal).

Only one blanket volume (1783 £t3) discarded in 30 years.
Fission-product heat generation as shown on Figs. 3a and b.
Cooling by natural air convection.

V1w O

It was estimated that the integrated heat generation rate at the end of the
30 -year filling period would be only 5.9 x th Btu/hr, and, if the storage
temperature is allowed to be as high as 900°F, this heat can be dissipated
by about L0 ft2 of cooling surface. Therefore 1t 1s only necessary to
place cooling surfaces over the tank cross section in locations that shorten
the path for heat conduction thrcugh the salt. Twelve L-in.-diam pipes
equally spaced over the cross section are provided to remove internal heat.
Regions of salt most distant from a cooling surface might be molten during
some period in tank lifetime, but this will not present a corrosion problem
because cooling surfaces will always be covered with a Ifrozen salt layer.
The storage tank is 13.5 ft in diameter and 13.5 It high, providing
about 1900 ft3 of storage volume. ©Stainless steel can be used in the
construction because only occasionally will molten LiFJThFhufission product
mixture contact a metal surface. The tank is contalned in an underground

concrete vault as shown on Drawing 58081-C, Appendix F.
- 110 -

APPENDIX F. PHYSICAL-PROPERTY DATA AND DRAWINGS

This Appendix contains the following information:

Table F-1 Thernal Date for LiF, BeF,, Na, K, and NeK
Figure F -1 Calculated Density of MSBR Fuel Salt and LiF
Figure F -2 Calculated Density of MSBR Fertile Salt
Figure F-3a and b =~ Vapor Pressure-Temperature Curves for Several Metal
| ' Fluorides '
Figure Tk Vapor Pressure of Nak
Figure F-S. 'ViSéésity.and'Thermal Conductivity of LiF
Figure F-6 ViscosityTemperature Curve for NeK (22.3-77.7 wt %)
: Alloy : :
Figure F -7 Properties of NeK (22.3-77.7 wt %) Alloy
Figure F -8 | Procesé_Flowsheet for Fuel and Fertile Streams
Figure F-9 Uhdergrbfind Storage System for Fuel -Stream Waste
Figure F-10 'Undergrdund Storage System for Fertile Stream Waste
Figure F-11 Arrangement of Processing Equipment for Fuel and
Fertile Streams
Table I'-1. Thermal Data for LiF, Bng, Na, K, and Nak
Latent Heat of Veporization (cal/g)
LiF 1690 (ref 38; calculated from vapor pressure data)
BeF', 1070 (ref 11)
Na. 1038 (ref 6)
K 496.5 (ref 6)

NaK (22.3-77-7 wt %):

LiF

617 (calculated)

Heat Capacity (cal/g °¢)

0.598 (ref 37)

DENSITY (gm/c:mg)

ORNL DWG 65-3045 Ri

TIT -

1.96
\ “a,
o mE 2,00 ~
< NFUEL SALY

1.92 ~ § 1.98 \

: \ 17800 500 500
1.88 o

< ~ t (°C)
S
_ N i
\
3084' \‘
N N
_ N |
N MP = 8450
1,80 AY
\\ .
%,
_ < -
. xf
1.76
~N
1,72 N <
FUEL SALT COMPOSITION \
i LiF ~ 68,5 mol % 7
BeFy - 31.2 mol %

1.68 UF, - 0.31 mol %

i MELTING POINT = 500°C 7
1.64 <
1.60

400 500 600 700 800 900 1000 1100 1200

Fig. F-1.

TEMPERATURE (°C)

Calculated Density of MSBR Fuel Salt and LiF.

1300
 DENSITY (gm/em")

410

4.60

4,55
»
| _4_.;45.
‘4.-_45
;-4;35

4,30

4,25

4.20

4,15

4,05

4,00

500

-2 -

= . o .
/MP 565°C

- (1049°F)

ORNL DWG 65-3046

\

it =

= 4,993 - 0000775 +
_- °C .

Lif

ThF4
233y 0,012
233pq 0,022 "

71 mol % {17.11 wi %)
(82,79
{0,035
(0,064

29 i

600

700

800

TEMPERATURE (°C)

200

1100

1200

- Fig. F-2. Calculated Density of MSBR Fertile Salt.

1300
- 113 -

ORNL DWG 65-2992R2

104 S
9
8
7
8
5
3
2
MELTING  BOILING POINT(°C)
POINT(°C) AT 760 mm
3 CsF 684 1251
10"~ BeF, 803 1170
. RbF 798 1410
, SEE= S LiF 845 1676
P RE=n = UFy 1035 1450
_4_?) ‘-‘.‘ AN CSF - -
g8 b
\E.a A %
S 3 % N
(%) %
n 3 =
(W)
& X SN
o 3 %
Q N Nertrbel2
AY % A
$ X, \ -
\ A
~LiF * c
%
3 S UF 4% %\
102 : .
2 3 % “ R‘.)F 3
8, % s £ -l
7 = 3 =i
6. m = 5 %
5 A3 Y % “ 2
4 a R X 2
3 n > - “
\ LY h:Y
2 Py A
A \
) '8} [§) @) p— = N
4 § 9 TS, b & 51 iy
) & & — VIS N STy 2
5.5 6,0 6.5 7.0 7.5 8,0 8.5 9,0
104/1 €10
Fig. F-3a. Vapor Pressure — Temperature Curves for Several Metal

Fiuvorides.
- 11h -

&4 L'e}
fiD.\ra\n ; .\. n\..-U1
o (D06/9)—
& | -
o ¥ )
T Yiin St il AN
© M I =T -
; ] -
G _F.M% - lfmlm!.f.r 4
= . . L o
0 y =M r > oy o
Ll E » oLzl
m F i 3 i Lt |
L, X 2
O { I i
Afb_ 4 ot 5
2 Ly i ]
i
o - < i -l
i = ML (D008~
” P i il
i &
b
4 [a S
S
y; ,
T (20868}
o
o e
I i Lis
RS h\ Y | ]
L5 : 2
% _.FA;M. nl\ T "
- “a
i al 17 Y e
ST (et06)
R il -
3 Lot 5
I L L -
- i
[ oK : L i 1] .aml kl
g e |- Bl ot
i 5 ) ‘ &
| %" 1.
St (DolL6)®
o i
: o4
= 3er i
| EMM. i % .” % L Ea \“E.
> T
» Hil S (D090 L)
! P
Hits, I
Jof%
pr
: Qo1 )}
: _ - o © =¥ P
Py
W o~ W i} & o Ly
o® % e b o flw tlo nUo
oo g <3 ) [an]

(B ww) JYNSSI¥! YOIVA

104/1“ e

Vapor Pressure - Temperature Curves for Several Metal

Fig. F-3b.

Fluorides.
VAPOR PRESSURE (mm Hg)

Composition: 77.7 wt % K,

SE; 22,3 wt % Na.

? i
" Dgta from Reference 6. i
4 ! I ] v i) j .
109 i
8
7
[
5.
2
9 —] N : P - =

45 6 7 8 9 10 11 12 13 14 15

104/1 (°K)

Fig. F-4. -Vapor Pressure of NaK Butectic.
THEMAL CONDUCTIVITY (Btu/hr t °F)

396

20—

VESCOSETY (cehfi:po_iséé) o

- 116 -

: '4.04 .

ORNL DWG 65-3049

402 G

CALCULATED
" BY METHOD

~ OF REF. 35

39|

' \\

w0 w9
S : ]‘EM?ERAT_URE_';(O

0. 950 __moo'
Q)

1050 -

1100

K

\

uicp) = 0.0586 3936/T
T =
(Ref. 36). __

L6}

AN

N

1.4

DN

: 102: :

- i.?’o_.
800

850

_.900 e

950

~ TEMPERATURE (°C)

1000

1050

_'1 100

 Fig. F-5. Viscosity and Thermal Conductivity of LiF.
- 117 -

ORNL DWG 65-3051

0e54 1

0.50

0.46

0.42

0.38

0.333 743 o/T
\ plcp p0-333 743 o/

0.34 p = g/cm ]
T =

VISCOSITY (centipoises)

0.30

0.26 ' \

02 N

~_

0.18 -
\

0.14
100 200 300 400 500 600 700 800

TEMPERATURE (°C)

‘ Fig. F-6. Viscosity-Temperature Curve for NaK (22.3-77.7 wt %) Alloy.
(Data from Reference 6.)
- 118 -

ORNL DWG 65-3050

16,5

DENSITY

15._.5. ;

| THERMA

L CONDUCTIVITY -

15.0

<

14.51

THERMAL CONDUCTIVITY (Btu/hr ft °F)

o
B

N

14.0—

300 400

. TEMPERA

- 500

TURE (°C)

600

700

800 90

0.23—

D

>

0.1

 HEAT CAPACITY (cal/g °C)

-0.80

10.78

0.76

0.74

0.72
0

- TFig. F-T.
-Referegce_6,)

200

400

3000 400
- TEMPERATURE (°C)

500

600

700 800

Properties of NeK (22.3-77.7 wt %) Alloy. (Data from

Jo.g2

DENSITY (g/emd)

S900
- 119 -

VENT TO OFF GAS

e _Fa RECYCLE Fe RECYCLE
VENT TO 1
z OFF GAS l l\
| | r
{ 1 Tiguio N
|
R ' §
! & NET FiSSION POWER MET FISSiON POWER | Q 2 ud
28 Mwe 972 Mwe | by Q
9 LS e S Q)
% 06’ ? % | u ’g’"‘ 5?
| 5 %————!powm PLANT o292 12008 [power PLANT l—i > 3 l[ Z By o
| FERTILE STREAM FUEL STREAM 157 FF¥d I
5 Volume =1783 Fi® pavg - 1.93 Uzs5 - %ifgzsbfg%d i Y% "
I uU-233 =627Kg Yolume =671 F3 OtherU = 1.002 Kg/d | Fs UF,
| Pa-233 =lIOKg u-233 =735.9 Kg LiF  =33.,8 Kg/d -6 o
| Th =141, 290Ky U-238% =Hi Kg BeF, =277.6 Kg/d L P,
| LiF = 38,765 Kg Other U = 58.1 Kg 7
; . £4,49 ; 218538 K :
E Fos Hoo'E 4 !é;;‘g = 52099 K? E G@;’;fig&
| ' £ Fuel wh =38700 Kg . | indl
| T100°F. |O00°F. 00468 %4
| _CONTINUOUS _ STV /j\ | LiF- 2.26 Kg/d
i w U=2.85 Kg/d _FLUORINATION BeFr 0.017 Kg/d |
; \358% u Rzmovai) Th® G422 Kg/d oGO F SURGE :
LIF=1762 Kg/d TANK
r U-233+1.24 g/ o/ |
: Ps:23372,179g/) U233+ 38,715 |
! h = ] u-235= |
T 27%8g/t 564
| Other - 3.057
- 42,560
A S iR LIF = 1028 FLUORINE
SUPPLY BeR, = 847 suppLY [
st U233 *2.822 %g Al | o e
o163 R¥d | ¥
LiF-3.54Kq/4 gre— 4 n
Th*=12.8 Kg/d OFF-BAS LiF = 336.8Kg/d DISTILLATION
LB2Fn057 o/ MAKE UP DISPOSAL BefF,® 277 .8 Kg/d 1666°¢C
1E2233= 10040/, Th* I575 Kg/d 238+ 0.02% Ke/d £.mm Hg
WASTE LiF= 3,54 Kg/d__ &m;?fiwi?fi"? ‘;3
| [DISTILLATE St
[COLLESTION
[U-238 = 178 g/d |
0.0468 Fi¥d D‘f;&g;&%’_

uF v 2.2¥ o
Befe 0.0IF Kgld |
U-233 = (:018 Kgrd
U-2s5= o.0002 g/
Ofherii=0.000 Kg/d

[ WASTE
COLLECTION

_UFg SALES

| i
U-233 +12. 650 Kg/d
U-235= 0.190 Kg/d

U-233: 2.644 Kg/d|

Otherl= &-999 Kyg/d

REFERENCE DRAWINGS NO.

{ LFg Recycle %o fuel stream

DAK RIDGE NATIONAL LABORATORY
OPERATED BY
UNION CARBIDE NUCLEAR COMPANY
DIVISION OF UNION CARBIDE CORPORATION
CAK RIDGE, TENNESSEE

LTS OK OIMENSioNs uness | MSBR. INTEGRATED PROCESSING PLANTSOC 4500
P PROCESS FLOWSHEET

DECMALS & FUEL & FERTILE STREAMS

AMGLES & o CHEMTIED TEPDIED ! APPROVED

sonsNons oD T Tseom (01

Fig. F-8. Process Flowsheet for Fuel and Fertile Streams.
- 120 -

COQLING AR INLET COOQLING AIR INLET
{ {8'x9" DUCY) y .

{za"x 30" oucy

JffT
7

ol

AlR SUPPLY DIFFUSER
(tye.} ) 3

¥

124 14 ARRAY 10"DIA.
CYLINDER { SEE CYLIHDER
BETAIL)

; 465 0" DIA. % 6'-4" HIGH
I5-6" WF BEAM ON U'-6" SR
18°-0"LG.

: COOLING - U-TUBES
————ToT i {932 TUBEg-~2:1/4"0.0.
—5‘% b = : Of. 3-3/4 " CENTERS )

STD. RING JOINT (METAL)
GE

342710 1 4.3/47 0.0, FoR E.VZHS%A;NG JOINT LMETAL
‘ééoé';ém'"e doi foa' 10" FLANGE
L : " $ ROD
K ' l - vy 1 .
i Lyt TA({:'KpV;’ELD
143
| / KEH:TL% RA Dsp LED)
H - (171 /i3
K | *—*=4‘\~fi--:f~—3“ SCH. 40 PIPE
2 :
WT:-— : , : S T R\ : 10*-150™ FLANGE
. “ - : 10" SCH, 40 PIPE
2.0 . 8 %%g')f’ERS ~COOLING AIR COOLING AIR EXHAUST . i o
i : EXHAUST IO RIO"DUC (28" 30" DUST) ; |
2l.0“ !8'-0“ 2';051 22!_ Ou 43"9“ 1 L
T i : =
L ag.o" HW
]
PLAN 8-0 TOP DETAIL
r-4" REMOVABLE CONCRETE - e e S AT T T O s PR ;
. SLABS — : _
T
e~
! A
N
] N
% L4 Kl
/8" § 7 N
{ 4 i 1
S .
180" T
e =3 / ‘\2
it ; ; BB . 2T IS e s E fi# - _u.}__w_.
/4" THK ST.S. ) SUPPLY DUCT &" WF BEAMS AR i
LINER 3.0 ML (typ.) t2'-G"CENMTERS) ' . :
ELEVATION REFERENCE DRAWINGS - N,
o . OAK RIDGE NATIONAL LABORATORY
- OPERATED BY _
Union QarsiDeE NUCLEAR ComPANY
- DIVISION OF UNION CARBIDE CORPORATION
CYLINDER DETAIL OAK RIDGE, TENMESSEE
LIMIYS ON DIMENSIONS UNLESS ; 4 LBG.
| L SN Quesions MSBR INTEGRATED PROCESSING PLANTIS4500
%0, REVISIONS _ : : ' FUEL STEANM WASTE SYSTEM -
. DRAR ) 3] DECIMALS % o ' L '
foird,G.K. {3-26-65 o o o :
W ANGLES = SUBMITIED APFRENED AFPROVED
i 5-CDS -B5i ——
oH T GATE PP " APHRVED REV]
___Eégmfia_. a0 & 1 [ Tseosd[p[™

Fig. F-G. Undergrouna;Stprage $ystem_for Fuel -3tream Waste.

-~ 121 -

DUCT OPENING TOBE

13'-8" DIA. x
STAINLESS

.
e =9 ol
2o ____E‘- 16" -0 " em—z"g"
E .
W 2-o"
10"x 10™ COOLANT INLET LOCATE
CENTER 1-C" BELOW TOP PLUG. 3
".o" :
FROM FLOOR-——
136" W, x 3/8"TH -
STEEL TARNK ) 19-0
15-Q"
12-COOLING TUBES {4” SCH.40 PIPE}
EQUALLY SPACED OVER TAMK DIA, -y
4 -WASTE INLETS (t"SCH. 40 PIPE)? ;
¥
2"0" H
k]

i74" S8T. LINER 3-0"
HIGH AND OVER BOTTOM —f-
.

AN

10" 10" COOLING EXIT DUCT
CENTER 1-0" FROM BOTTOM
OF PLUG.

s o0 " "
e
.

CERC

LA

MIAMY
5.
R

° .
LA

138"

>

L

i o

3
o 8o o
o s

©a?

© o

o 0
o e

BT S
3 ©o oo
3

|
LA
oo Tl
6n=,:_‘,C' H
[ A
PRI S |
B i
va o H
P
UL W
- [
LT 20-6
0" g
osas o
. 1A
s 4 16-6
%", .
P ;
'\q°° i
FIN,
o 94
RN
2%, H
RIS
RS H
e o H
Tt e i
ey i
cre® i
o ® o '
as® *
":
ey
«%0 ) 2-0
o

6" WF BEAMS
{5-REQ'D)

FRACTIONS £
NO. REVISIONS | oare [areo] ape0 FERTILE STEAM WASTE STORAGE SYSTEM
DRAWN DATE | SUBMITTED 1 DATE | APPROVED | DATE DECIMALS +
Ford, G, &, ;4-5-656 :
DESIGNED DATE APPROVED DATE APPROVED | DATE ANGLES * SUBMITTED APFROVED
CHECKED DATE | APPROVED | DATE | APPROVED | DATE scaLE: 174"= 10" APFROVED 1 |5808| |C IREV‘

ELEVATION

REFERENCE DRAWINGS NO.

0ak RIDGE NATIONAL LABORATORY
OPERATED BY
UNioN CARBIDE NUCLEAR COMPANY
DIVISION OF UNION CARBIDE CORPORATION
OAK RIDGE, TENNESSEE

LIMITS ON DIMENSIONS UNLESS
OTHERWISE SPECIFIED:

MSBR

INTEGRATED PROCESSING PLANT

BLDG.
No, 4500

Fig. F-10.

Underground Storage System for Fertile Stream Waste.
122 -

00 olg - : 2379 prie .70 P 828 ul
i
. : i
i e AR A AT
et ol R : : i ¥
v I . I [
7 & i3 ¥
-, : ¥ 4
o Ry a;’.,‘ ‘:
2 Hoa o e YR B oo By
i » L
- 4 N
V ? Vo' B LILAEN o i, "‘%‘
¥ ¥ o AR g
‘cg’ f WEF * Fen,
i = g
PROCESSING CELLS— . : S REACTOR CELL
/ 4 5 e R, u
; v o W
AR-NaK ¥ -9
HEAT i = 3 G
EXCHANGER o 4 3
. i -2
4 f‘; / “P‘
i b 0 Pl VI e i Ry e Ve By TeF gk TS FE a’ !’" "
Lo P ' o -
H— - -0 "1 REACTOR HEAT
¥ EXCHANGER
a o B G v, B D B i P e TR 1%k B ”J C
Nak T i 7
SUPPLY | 7 it A
- % i * )
b . v !
: " : PUMPS L] 4
5 1 > = 3§ REACTOR 3
El B ; i 1] ' - 2
2 e B o o0 o0 Ak b
Rl ) A R > [ A £ 4 : {5 :
SUPPLY AND E =3 SO @ \;) 3 : > %
3By -
i 7 o SURGE 1w 2 3
MOAKEUP AREA 14 o / TaNg 17 A S — ;
> 4 { O b - b 2
7 i FLUORINATOR ; i 2
i R
- . I o NOEEN o avee ) f\} B
PRI i N O 7 O TNTANKGT Y b - ; r
ot & Y \o ot A B K,/ y
TE| 151 ClurgTrae ~ soRBERs = 1% 4 g
a . A A e T S S E PN S A Y TP e ey L Al B}
L3 Ry F
t—i ’ i P!
. - g e P e o e el e n R e e ¥ %) b
. CT} \) F : {i_t B B & g 15:é,-_ &
v ooE | 2O OO0 % %
o My O O B “ KoY KOH COLD TRAPS A e
© Dol Z1TANK SCRUBBER STILL ; ;
sPPLYE @ [F 3 L s
DD W WG4 _CONDENSER ==
ook o wiSTe "
i i O S !
OGN K Fp E UFg=LiF g
a0 5 5 HYDRATOR REDUCTEON R ;
LR oF SORB & &
He CH® 5 ? 5 1 i -
SUPPLYD O I, Iy % Pels £
©OF COLD TRAPSZ o =\, P P 2
S Q¢ F | FERTILE ,’ W, s i -
> . ‘ STEAM | VI e e o a S LI
UFg fl: ! MAKE-UP N 7 " s - P
PRODUCT (y [ 7 PRI gTER 2
| & 3 P B
ThF 4~ LiF; > r N - S 3 *
MAKE-UP A 2 o . ; z
o, e e 3 REFERENCE DRAWINGS Ho.
9 [ k4 Lisa N P
N W e S B o R e S T Lo S L B P LS - | n 2 ,.”ZE, OAK RIDGE NATIONAL LABORATORY
: © g E Ry, = Bl s o o g B L) o
-9 e SRRV 5229 _ _ . & Unton CARBIDE NUCLEAR COMPANY
20 12-9 Ny : RS DIVISION OF UMION CARBIDE CORPORATION
¥ I Bde Baiosu pinop viry poars grd B o Bt w0 Fiws s £ e el te S . OAM RIDGE, TENNESSEE -
S LNITS ON DINENSIONS UMLESS | MSER INTEGRATED PROCESSING PLANT - - BilGasog
L SR ‘ OTHERWISE SPECIF e _Ne.
NC. REVSIONS | patE !APPO!A?PD i AR § FRACTIONS ' PRELIMINARY EQUIPMENT
W1 DAIE | SUBMITIED | BATE ] AT ' S MALS + ‘
FARMER §3-15-65] ! ! . S e LAYOUT
~TERGRES | DRE | AHOID | DR | RROVED | OATE _ R ST T P — SUBHTES REFRONED  APPRONVED
14 ! 1 . L
CHECKED | DATE | APPROVED | DATE | APPRGVED | DATE S _ : R FOPROVED ; y TV,
: ! ! , = SOME3/16" =170 [ Tseos0]p [™]

Fig. F-11. Arrangement of Proce531ng Equlpment for Fuel and Fertile
Streams. The highly radioactive operations in fuel-stream processing are
carried out in the smaller cell (upper left). The other cell houses
equipment for the fertlle stream and the cooler fleld stream operatlons.

- 123 -

ORNL-3791
UC-80 — Reactor Technology
TID-4500 (46th ed.)

INTERNAT, DISTRIBUTION

1. Biology Library 65. C. E. Larson
2-4. Central Research Library 66. R. B. Lindauer
5. MSRP Director's Office 67. H. G. MacPherson
6-7. Reactor Division Library 68. W. B. McDonald
8-9. ORNL — Y-12 Technical Library 69. H. F. McDuffie
Document Reference Section 70. L. E. McNeese
10~44. Laboratory Records Department 71. R. P. Milford
45. Laboratory Records, ORNL R.C.  72. R. L. Moore
46. G. M. Adamson 73. F. L. Peishel
47. L. G. Alexander 74. A. M. Perry
48. C. F. Baes 75. J. T. Roberts
49. S. E. Beall 76. C. D. Scott
50. F. F. Blankenship 77. D. Scott
51. 8. Cantor 78. M. J. Skinner
52. W. H. Carr 79. I. Spiewzk
53. W. L. Carter 80. A. Taboada
54. G. I. Cathers 81, J. R. Tallackson
55. F. L. Culler 82. R. E. Thomsa
56. D. E. Ferguson 83. D. B. Trauger
57. H. E. Goeller 84. A. M. Weinberg
58. W. R. Grimes 85. M. E. Whatley
59, C. E. Guthrie 86. H. B. Whetsel
60. P. N. Haubenreich 87. Gale Young
61l. R. W. Horton 88. P. H. Emmett (consultant)
62. P. R. Kasten 89. J. J. Katz (consultant)
63. S. Katz 90. C. W. J. Wende (consultant)
64. M. E. Lackey 91. C. E. Winters (consultant)

EXTERNAL DISTRIBUTION

92. Research and Development Division, AEC, ORO
93-430. Given distribution as shown in TID-4500 (46th ed.) under
Reactor Technology category (75 copies — CFSTI)