ORNL-MIT-117.txt

 

 

OAK RIDGE NATIONAL LABORATORY
OPEFRATED BY
UNIOH CARBIDE CORPORATION
NUCLEAR DIVISION f

 

 

POST OFFICE BOX X
CAK RIDGE, TENNESSEE 137830
ORNL-MIT-117

R

COPY NO.

 

 

 

November 18, 1970

DATE:
SUBJECT: Removal of Tritium from the Molten Salt Breeder Reactor Fuel
AUTHOR: M.D. Shapiro and C.M. Reed

Consultant: R.B. Korsmeyer

ABSTRACT

Molten Salt Breeder Reactors will produce large quantities of

tritium which can permeate most metals at elevated temperatures and
In this project it was assumed

 

 

thereby contaminate the environment.
that the tritium can be removed from the salt stream by a hydrogen-
helium purge and that the helium can be separated for recycle with a =
palladium membrane. Several systems for concentrating and storing i
the tritium were conceptualized, designed, and economically evaluated. b
Cryogenic distillation of liquid hydrogen appears to be the most eco- &
nomical system. A cryogenic system with a capacity of 4630 gmoles of gg
hydrogen per hour at a 1000-fold tritium enrichment has an estimated ”
. A - - ¥
capital cost of $328,000 and an estimated annual operating cost of E%
$81,000 (excluding depreciation}. n
......... o e
NOTICE -
This report was prepared as an account of work CTJ
spenscred by the United States Goevernment. Neither &3
the United States nor the United States Atomic Energy et
Co::nmxssxon, ner any of their employees, nor any of
their contractors, subcontractors, or their employees,
makes any warranty, express or implied, or assumes any i
fegal liability or respensibility for the aceuracy, come- .
pleteness or usefuiness of any information, apparatus, by
product or process disclosed, or represents that its use “fiw
would not infringe privately owned rights. ™~y q'

 

 

 

Qak Ridge Station
School of Chemical Engineering Practice
Massachusetts Institute of Technology

NOTICE This document contains informatien of a preliminary
’ nature and was prepared primarily for internal use ot the Oak
Ridge Nafiona! Laboratory. !t is subject to revision or cor-

rection and therefore does not represent a final repert. The in-

formation is only for cfficial use and no release to the public
shall be made without the approval of the Law Department of

Union Carbide Carporation, Nuclear Bivision.
SISTRIBUTIOR OF THIS DOCURMENT 1S URLIMITED

   

 
Contents
Page
T. SUMMAYY ¢vveiavncencorssnsancess Ciee et ssransaret s Meeecsne esee A
2. Introduction ....e.e. P re i ebteret ettt an e e ebressaatenn 4
3. Design and Evaluation of A]ternéte Separation Systems ......... o
3.7 Approach ...viviiieiiiiiecencnnennn i aeean e .. B
3.2 Feed Pretreatment ...... eessesssrsestesatrtasetearesiarns b
3.3 Storage of Tritiated Water ............... et ceceseneans 7
_3.4 Water Distillation cevvvvieicecnrinonrrconnnnconrns reeees 7
3.5 Thermal Diffusion ...cicvivernnienconsrinnnens seesvesananes 9
3.6 Cryogenic Distillation .....cvivviiiveennnn vernenca ceenae 9
4, Discussion of Separation Systems ........ovevvevevnnonse S 16
5. Conclusions and Recommendations ............. esareansaenesnons 16
6. Acknowledgement ..... eereserarenes ettt eeircca st o as s 17
7. ApPPendiX ..cciiiiiiiiiiii ittt cc e e 18
7.1 Basis for Water Distillation Costs ....... N 18
7.2 Thermal Diffusion System ............;.,.,.,.,.,,,..,,..‘. 21
7.3 Cryogenic Distillation System ....ovvecveriirorsrreocnoons 22
7.4 Computer Codes ....... ceerean esssserecasan Cetecssoosasecs 26
7.5 Nomenclature ...... eercaasaaaes e asecenen e veeenan wes 29
7.6 Literature References ...,....cvivseennrenenrecennnsnnnoen 30
. SUMMARY

One characteristic of Molten Salt Breeder Reactors (MSBR} is the rela-
tively large quantity of tritium which would be produced in the salt fuel
stream. Tritium, like hydrogen, can permeate most metals at elevated temp-
eratures, and thereby contaminate the environment. An efficient means of
removing and concentrating tritium from the fuel stream is essential to the
development of MSBR.

In this project it was assumed that the tritium can be removed from
the fuel stream by a hydrogen-heiijum purge and that the helium can be
separated from the hydrogen for recycle via a paliadium membrane. Four
systems were conceptualized, designed, and economically evaluated toc con-
centrate or store the hydrogen and tritium: storage of unconcentrated
tritiated water, water distillation, gaseous thermal diffusion and cryo-
genic distillation of 1iquid hydrogen. On the basis of this evaluation
the most economical system, cryogenic distillation, would provide a 1000-
fold tritium enrichment at an estimated capital cost of $328,000 and an
annual operating cost of $81,000,

2. INTRODUCTION

There is presently interest at ORNL in the development of Molten Sait
Breeder Reactors {MSBR}. A characteristic of these reactors, however, is
the generation of large quantities of tritium (half 1ife 12.36 yr). Tritium,
1ike hydrogen, has a very high permeability through most metals at tempera-
ture and concentration levels of the moiten salt; therefore if 1t is not
removed, it will escape from the reactor and contaminate the environment.

Tritium is a weak beta emitter (18.6 kev), but it exchanges readily
with hydrogen and as tritiated water can enter the body by penetrating the
skin. The effect of radiation in a very localized area and the transmuta-
tion of tritium to helium within the body may be of biological significance

(1.

One proposed method of removing the tritium from the MSBR fuel stream
is by means of a helium-hydrogen purge (2). The hydrcgen stream would
then be separated from the helium and the tritium would be concentrated
and stored as tritiated water (HTO). Since tritium is an isotope, its
concentration will depend mainly on physical separation processes.

In the MSBR concept the primary salf stream is comprised of molten
salts of uranium, tithium, beryilium, and thorium. The primary salt cir-
culates through the reactor where a critical mass is achieved and fission
occurs. The sensible heat generated by fission is transferred to a steam
cycle by means of a secondary salt stream. The flow plan is illustrated
in Fig. 1,
 

 

 

TO TRITIUM RECOVERY

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

H _ 6
"0
PURGE
(Ho-He) ¥
l it iy ] o
-~ —
HEAT
MSER EXCHANGERS STEAM
CYCIE
[ ) ~ SRCONDARY
Bl SALT
PRIM\RY GALT

 

MASSACHUSETTS INSTITUTE OF TECHNOLOGY
SCHOOL OF CHEMICAL ENGINEERING PRACTICE

AT
OAK RIDGE NATIONAL LABORATORY

 

MSBR FLOW PLAN

 

 

DATE

 

 

DRAWN BY FIL

30 0cr 70| _ALS CE

p

E

S

N

0
X

-117

 

FIG,

 

 
Tritium is produced in the primary salt stream by neutron absorption.
The reactions producing tritium and the estimated production for a 1000
Mw(e) reactor are listed in Table 1.

Table 1. Tritium Production in a 1000 Mw(e) MSBR (3)

 

Ternary Fission 31 curies/day
6Li(n, o} T 1210
Li(n, an) T 1170
19¢(n, 10) T 9

2420 curies/day

v 0.25 gm tritium/day

 

3. DESIGN AND EVALUATION OF ALTERNATE SEPARATION SYSTEMS
3.1 Approach

In this study it was assumed that the tritium could be remcved from
the fuel stream by a mixed helium and hydrogen purge. The hydrogen and
tritium would then be removed from the purge stream and concentrated. The
selection of the most feasible system for effecting the desired concentra-
tion was based on a preliminary design and cost estimate for each system.
The systems studied were storage of the unconcentrated tritium as tritiated
water, water distillation, thermal diffusicon, and cryogenic distillation of
Tiguid hydrogen.

The design for all the systems was based on 111,000 gmoies of hydrogen
per day at an H/T = 106, A 100- to 1000-fold enrichment was desired (i.e.,
H/T = 103 to 10% in the product stream} with a 99 to 99.9% recovery of the
tritium. In all cases the product tritium is to be stored as water on the
MSBR site (2). For all processes the separation equipment will be en-
closed in a separate building to isolate any possible tritium leak.

3.2 Feed Pretreatment

The purge stream will contain helium, hydrogen, and tritium as well as
gaseous fission products such as krypton, xenon, icdine, and hydrogen fluo-
ride. It is proposed to pass the purge stream through a charcoal bed to g
adsorb some of the gaseous fission products. To separate the helium for
recycle from the hydrogen and tritium, & palladium "kidney" would be
employed. A palladium membrane which passes 15 scfh of Hp costs approxi-
mately $5000 (4). When the six-tenths power formula is applied to scale
to the capacity required for the MSBR, an estimated purchase cost of
$136,000 is realized.

It is estimated that the installed cost of the palladium kidney is
four times the purchase cost of the kidney, or approximately $544,000.
The same cost will be associated with each of the four alternate systems.
A second item which is common to the four processes is the oxidation
equipment and its installed cost is estimated to be $136,000.

3.3 Storage of Tritiated Water

The hydrogen and tritium would be oxidized after passing through the
palladium kidney and the resulting tritiated water condensed and sent to
a storage tank. Storage of tritiated water will require steel tanks en-
cased in a concrete tank. Should a leak develop, the liguid would be con-
tained, but an additional tank would be required to effect a transfer
before final repairs could be made (5). The tanks were sized to hold
30 years production of tritium, the expected lifetime of the reactor. The
Tiquid will have to be stored until the activity has decreased to less than
1% (approximately 110 yr). At a production rate of 2000 Titers/day, a tank
capacity of approximately 5.8 million gallons is required. With an esti-
mated capital cost of $1/gal (5), the two-tank system would have a capital
cost of $11.6 million. Annual operating cost for this sytem would be the
cost of the hydrogen and oxygen burned to form the water ($143,000) and the
maintenance cost [2% of the capital cost (13)], $232,000. (See Appendix
7.1 for details.}

3.4 Water Distiilation

During World War II the United States built and successfully operated
several water distillation plants to produce heavy water. The low vaiue of
the relative volatility (o), however, required the use of high reflux ratios
and a large number of plates in the distillation column.

Distillation to separate tritiated water (HTO) is not as difficult as

that for heavy water, since the value for o is several percent higher. A
plot of relative volatility versus pressure indicates that such & system
should be operated under vacuum to take advantage of the higher value of o
(see Fig. 2). A computer program was written to size the distillation
column. The design of the column is based on the use of a high efficiency
packing such as Sulzer CY (designed for use in heavy water systems). This
packing was found to have 21 theoretical plates/meter and a pressure drop

| of 0.19 torr/theoretical plate (6) for heavy water separations at a liguid

g Toading of 2000 kg/M2-hr and a column head pressure of 120 mm Hg. Move
favorable conditions might be achieved with the tritium system by lowering
the head pressure of the column.
 

i.08Q—

1.07¢

1.064

1.05¢

 

1 . 04 (Proom
30

 

Relative Volatility, o

75

    

100 125 150

l | |

175 200 225

 

Pressure, mm Hg

MASSACHUSETTS INSTITUTE OF TECHNOL.OGY
SCHOOL OF CHEMICAL ENGINEERING PRACTICE .
AT

OAK RIDGE NATIONAL LABORATORY

 

 

H,0
RELATIVE VOLATILITY (a - £5-) (1)
ATE DRAWN BY FILE NQ. FIG.

NGV 7g

CEPS-X-117 2

 

 

 

 

 

AT
9

As shown in Fig. 3 the number of theoretical plates is a sharp function
ocf the reflux ratio. Figure 4 is a schematic diagram of the propcsed water
distillation design. The optimum systems and operating conditions were
determined by varying the reflux ratic for different enrichment and recovery
factors (see Tables 4 and 5 in Appendix 7.1). Table 2 shows the major de-
sign specifications and the cost estimates for the optimized systems. Cap-
ital _costs for 99% recovery are $422,700 at H/T = 104 and $362,600 at H/T
= 103. For 99.9% recovery, capital costs are $536,600 at H/T = 104 and
$484,600 at H/T = 103. Although the cotumn packing and building costs are
hiflher for K/T = 10°, the overall cost is less than for H/T =
10% because of the associated storage costs. The cost of recovering 99.9%
of the tritium for H/T = 103 is 33% higher than the cost for 99% recovery.
Operating costs in all four cases are essentially the same, $220,000 annually.
The cost of Ho is the major operating expense, $128,000 annually. A break-
down of the column costs for other reflux ratics is in Appendix 7.1.

3.5 Thermal Diffusion

Thermal diffusion is based cn a temperature gradient in a mixture of
gases which gives rise to a concentration gradient, thereby effecting a
partial separation. Jones and Furry (9) have presented a detailed discus-
sion on the theory and design of thermal diffusion systems for binary sepa-
rations. The thermal diffusion constant between two species with masses
m] and m2 is equal to (mp - m])/(mz + m), and for a hydrogen-tritium
system this ratio is 1/3 which is considered high.

In a thermal diffusion column the separation rate is fixed by the
temperature and pressure of the system. Theory requires that the rate of
production of each column be small compared with the rate of thermal djf-
fusion., The production rate of tritium in an MSBR is so large that 103 to
10% thermal diffusion columns operated in parallel would be necessary.
Based on the theory of Jones and Furry, Verhagen and Sellschop (11) designed
and operated a thermal diffusion system for tritium enrichment. A scaleup
of their apparatus would require 5150 parallel systems for a 1000-fold en-
richment with a power load of 100,000 kwh. The power consumption at
$0.004/kwh would cost $2.9 million per year. (See Appendix 7.2 for appa-
ratus details and operating conditions.)

3.6 Cryogenic Distillation

Due to recent advances in cryogenic engineering, several plants have
been constructed which separate deuterium from hydrogen by cryogenic dis-
tillation of the liquid hydrogen feed. The relative volatility of H2 to HT
is not available, but it is believed to be equal to, if not greater, than
the relative volatility for the Ho-HD system (o ® 1.6 at 1.5 atm). This
is considerably higher than the relative volatility for the water-tritiated
water system (o ¥ 1.05). A second advantage is the Tower consumption of
Ho and 02. For the separation of tritium by water distillation, all the
Ho from the purge stream is oxidized to water, but in the cryogenic system
10

 

(NTP)

Number of Theoretical Plstes

 

     

 

 

 

 

 

1 ' g
600 =
550 —
500 ]
450 -
H/T=103
400 99.9% Recovery -
-/TwlO4
350 —
1/T=10>
300 —
99% Recovery
H/T=}O4
] i ] 1 | ] | j | | ]
26 27 28 29 3¢ 11 32 33 34 35 36 37
Reflux Ratio
MASSACHUSETTS INSTITUTE OF TECHNOLOGY
SCHOOL OF CHEMICAL ENGINEERING PRACTICE
AT
OAK RIDGE NATIONAL LABORATORY
NUMBER CF TEEQRETICAL PLATES
VERSUS REFLUX RATIO
DATE DRAWN BY FILE HO. FIG,
4NOV 70 | MRS CEPS-X-117] 3 i

 

 

 

 

 

 
11

 

 

 

 

 

 

Steam
Ejector

 

    
 

Condenser

 

Y

Distillate
————————
Feed

Packed Column

 

 

 

 

 

Tritium

f!?] Recovery

Reboiler

 

 

 

Steam

 

MASSACHUSETTS INSTITUTE OF TECHNOLOGY
SCHOOL OF CHEMICAL ENGINEERING PRACTICE

AT
OAK RIDGE NATIONAL LABORATORY

 

WATER DISTILLATION SYSTEM

 

DATE DRAWN BY FILE NO. FIG,

4 NOV 70 | M8 CEPS-X-117 4

 

 

 

 

 

 

 

 
12

 

 

 

Table 2. Cost Estimate of Optimum Water Distillation Systems
(See Appendix 7.1 for Details)

Recovery 99% 59% 99,9 99. 9%
Product H/T 104 103 104 103
Column Diameter 1.32 M 1.32 M 1.32 M 1.32 M
Number Theoretical Plates 275 323 362 412
Designed No. Plates = 1.1 (NTP) 303 355 398 453
Column Height 14.4 M 16.9 M 19.0 M 21.6 M
Reflux Ratio 32 32 35 35
Installation Column Cost $90,600 $106,200 $138,000 $157,200
Flow Distributors 1,500 1,200 1,200 1,200
Packing ' 139,000 163,000 200,000 228,000
Building 19,600 23,000 25,900 29,500
Covering 3,060 3,400 4,100 4,100
Ejector {installed) 2,000 2,000 2,000 2,000
Site Preparation 1,000 1,060 1,000 1,000
Instruments 50,000 50,000 50,000 50,000

Total 306,700 349,800 422,200 473,000
Storage Tanks 116,000 11,600 116,000 11,600

Total $422,700 $361,400  $538,200  $484,600
Operating Costs, $/yr (Depreciation Not Included)
Steam $10,570  $10,570  $11,530  $11,530
H, Usage 128,160 128,160 128,160 128,160
02 Usage 14,685 14,685 14,685 14,685
Labor 45,000 45,000 45,000 45,000
Maintenance € 5% Investment 21,135 18,070 26,910 24,230

Total $219,550  $216,485  $226,285  $223,605

 
13

only the final product is burned, and the remainder, more than 99% of the
Ho can be recycled.

The tiquid hydrogen distillation system is based on a plant built by
Gebrider Sulzer for heavy water production in DOMAT/EMS, Switzerland {18,
19). A schematic for this plant is shown in Fig. 5. The hydrogen feed
is initially compressed to 3.7 atm, cooled in a series of three heat ex-
changers {Nos. 1, 2, and 3), then liguified and re-evaporated in the feed
liquifier before it enters the column. The vapor from the top of the col-
umn is split into two streams. One stream passes through exchangers 3, 2,
and 1 to cool the feed, and then is recycled to the hydrogen purge stream.
The remainder passes through exchangers 4 through 8, exchanging against the
returning reflux stream. It exits exchanger 8 at ambient temperatures, and
enters the reflux compressor. Because of interstage compressor cocling,
the H2 gas leaves the compressor at 14 atm and 300°K. The stream re-enters
exchanger 8 and the expansion turbines, and finally exits exchanger 5 as
saturated vapor at 4.5 atm. The saturated vapor then passes through the
bottom of the column where it is condensed by boiling the liquid in the
reboiler. The stream passes through exchanger 4, flashes to 1.5 atm, and
enters the column as saturated liquid.

The computer code used in Sect. 3.4 was modified for use with this
system. Calculations showed that a column with 100 theoretical stages op-
erating at a reflux ratio of two would yield a separation of H/T = 103 at
a recovery of 99.9%. Although calculations showed that a reflux ratic of
two was sufficient for the desired separation, the design was based on a
reflux ratio of five to allow for variation of the operating conditions
and to ensure a conservative cost estimate. With a packing material simi-
lar to Sulzer CY, the column would be only 13 ft high at a liguid loading
of 1500 kg/mZ-nr (6). As seen in Table 3 the column cost represents a
small fraction of the total cost; therefore, the less difficult separations
were neglected in the analysis.

Capital cost is estimated at $328,100 and operating costs at $80,600
annually. The building for this system must not only isolate. the system,
but alsc insulate the apparatus for the low temperatures involved. The
distillation column and the low temperature heat exchangers and expansion
turbines are enclosed in steel vacuum bottles to maintain cryogenic temp-
eratures. No cost information was obtainable on the new high efficiency
insulation currently being used on some cryogenic equipment. However, it
is believed that the cost estimate presented is conservative. The cost of
the expansion turbines was estimated from cost information for a 200 ton/
day oxygen plant. This unit has 100 times the capacity reqguired for the
hydrogen liquification unit. Further details on the cryogenic distiltlation
system, including an explanation of the cost estimate, are given in Appen-
dix 7.3.

R
 

 

 

P ,

 

 

 

3-Stage R
Compressor

Expansion

Turbines —~»-

 

eflux I

A

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  

| Packed
Column

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Ex = Heat Exchangers

Product

 

Ex 1

 

 

 

 

Ex 2

 

 

 

 

Ex 3

 

 

 

 

 

 

 

 

  

   

iFeed

o Ry
e raw

HLiquifier

~fpe- 1O H
é]e

Recy

 

Fee

]

 

MASSACHUSETTS INSTITUTE OF TECHMOLOGY
SCHOOL OF CHEMICAL ENGINEERING PRACTICE
AT

OAK RIDGE NATIONAL LABORATORY

 

CRYOGENIC DISTILLATION SYSTEM

 

DATE
15 Vol i

 

DRAWN BY

AL

 

FtiLE NO.

CEPS-X-~117

 

FIG.

 

 

 

VAl
15

Table 3. Cost Evaluation of Cryogenic Distillation System

 

Recovery 99,9%
H/T in Product 103
Designed Number of Stages 100
Reflux Ratio 5
Cotumn Diameter 8 1in,
Column Height 13 ft
| Purchase Cost Factor Installed Cost
Capital Cost
Column $ 1,000
Packing 925
1,925 5 $ 9,600
Feed Compressor 8,700
Reflux Compressor 31,000
39,700 2 79,400
Heat Exchangers 46,000 2 92,000
Expansion Turbines 25,000
Instrumentation 50,000
Insulated Building 24,500
Vacuum System 9,000 4 36,000
Storage Tank 11,600
Total Capital Cost $§ 328,100

Annual Operating Costs (Depreciation not Included)

Electricity
a) Compressors 1,300
b) Vacuum System 200
H2 and 02 1,300
Labor 45,000
Maintenance at 10% Capital Cost 32,800

L Total Operating Cost $ 80,600

 
16

4, DISCUSSION OF SEPARATION SYSTEMS

Storage of tritiated water without any form of concentration reguires
a capital cost twenty times greater than for water distillation and thirty-
five times that of cryogenic distillation. Thermal diffusion alsc represents
an unsatisfactory sclution to the problem of tritium concentration. Moni-
toring over 5000 two-stage thermal diffusion columns and maintaining control
of the feed to each column appears horrendous; and the annual power cost of
$2.9 mitlion certainly makes this system unfeasible.

Water distillation is a technically sound alternative. However, its
capital and operating costs are not competitive with those of cryogenic
distillation. The packing represents the major capital expense, and since
all of the hydrogen is oxidized to water, the major operating expense is
the cost of hydrogen. However, if the hydrogen concentration in the purge
stream is sufficiently high, it might be feasible to oxidize the hydrogen
directly without the use of a palladium kidney and separate it from the
purge stream as water. This would not be sufficient to make water distil-
lation competitive with the cryogenic system based on operating costs.

Cryogenic distillation has the lowest capital cost estimate as well as
the Towest operating costs. Part of the economic advantage is realized by
recycling 99.9% of the hydrogen to the purge stream. It should be noted
that the cryogenic distillation was designed for a reflux ratio of five,
although for 99.9% recovery at H/T = 10°, a reflux ratio of two is sufficient.
Thus the system is capablie of recoveries in excess of 99.9% at concentrations
Tower than H/T = 103,

5. CONCLUSIONS AND RECOMMENDATIONS

1. Cryogenic distillation of liquid hydrogen is the most economical of
the alternatives studied. A cryogenic distillation §ystem which will enrich
4630 gmole/hr of hydrogen from H/T = 106 to H/T = 10° at 99.9% recovery has
an estimated capital cost of $328,100 and an estimated annual operating cost
of $80,600 (excluding depreciation). There is also the associated capital
cost of $680,000 for the palladium pretreatment and oxidizing systems.

2. HWater distillation or storage of unconcentrated tritiated water
represents too great a capital expenditure and annual operating cost.

3. Thermal diffusion is unattractive for concentrating tritium from a
1000 Mw(e) MSBR.
17

6. ACKNOWLEDGEMENT

The authors would like to express their gratitude to R.B. Korsmeyer
for the assistance and insight he provided during this project. His enthu-
siasm was a constant source of encouragement. The assistance of J.T. Corea
is also gratefully acknowledged.
18

el

7. APPENDIX
7.1 Basis for Water Distillation Costs

1. Column Shell thickness 0.5 in., type 304 stainless steel
$1.25/1b fabricated (12)

2. Flow Distributors: 1 approximately every 5 meters
$300 each (instalied) (13)

3. Packing
$200/Ft>  (14)

4, Building Cost

3 (25)

5. Building Insulation

$2.70/ft

$1/Ft% wall (15)

6. Steam Ejector
$2000 instalied (13, 16)

7. Storage Tanks
$1/9a1 (5)

8. Steam
$0,25/106 Btu (Use of waste steam from the MSBR)
9. Raw Materials

H2 = $0.0048/scf

0, Tiquid = $25/ton (17)
10. Labor

1 man/shift at $15,000 yr = $45,000
11. Maintenance at 5% of investment (13)

12. 7000 hr of Operation Per Year s

Tables 4 and 5 reflect the costs of the various H20 distillation systems,
 

Table 4.

Cost of Various Water Distillation Systems

 

For 99% Recovery with H/T = 10°

For 99% Recove

 

 

 

Reflux 25 27 30 32 35 40 25 27 30

Column Diam (M) 1.17 1.22 1.28 1.32 1.38 1.47 1.17 1.22 1.28
NTP 377 330 291 275 258 239 425 379 340

ANP = NTP x 1.1 415 363 320 303 284 263 471 417 374

Column Height (M) 19.8 17.3 15.2 14.4 13.5 12.5 22.4 19.9 17.8

Column Cost $18,300 $16,700 $15,400 $15,700 $14,800 $14,600 $20,800 $19,200 $18,000
Installation (5x) 91,500 83,500 77,000 75,500 74,000 73,000 104,000 96,000 90,000
Flow Distributors 1,500 1,500 1,500 1,500 1,500 1,500 1,500 1,500 1,500
Packing 150,000 143,000 138,000 139,000 143,000 150,000 170,000 164,000 162,000
Building 27,000 23,600 20,700 19,600 18,400 17,000 30,500 27,000 24,200
Covering 3,800 3,400 3,100 3,000 2,800 2,700 4,200 3,900 3,500
Ejector (installed) 2,000 2,000 2,000 2,000 2,000 2,000 2,000 2,000 2,000
Site Preparation 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000
Instrumentation 50,000 50,000 50,000 50,000 50,000 50,000 20,000 50,000 50,000
Subtotal 345,100 324,700 308,700 306,700 307,500 311,800 384,000 364,600 352,200
Storage Tanks 116,000 116,000 116,000 116,000 116,000 116,000 11,600 11,600 11,600
Total Capital Cost $461,100 $440,700 $424,700 $422,700 $423,500 $427,800  $395,600 $376,200 $363,800
Steam Cost/yr 8,330 8,980 9,940 10,570 11,530 13,130 8,330 8,980 9,940

 
Table 5. Cost of Various Water Distillation Systems

4

 

 

 

For 99.9% Recovery with H/T = 10 For 99.9% Recove
Reflux 25 27 30 32 35 40 25 27 30
Column Diam (M) 1.17 1.22 1.28 1.32 1.38 1.47 1.17 1.22 1.28
NTP 614 516 431 397 362 329 660 561 482
ANP = NTP x 1.1 675 570 474 437 398 362 726 617 530
Column Height (M) 32.1 27.1 22.6 20.8 19.0 17.2 34.6 29.4 25.2
Column Cost $35,900  $29,000  $25,400 $24,100 $23,000 $22,200 $36,900 $31,500 $28,300

Installation (5x) 180,000 145,000 127,000 121,000 115,000 111,000 185,000 158,000 142,000

 

 

Flow Distributors 2,100 1,800 1,500 1,200 1,200 1,200 2,100 1,800 1,500
Packing 265,000 224,000 205,000 200,000 200,000 206,000 272,000 243,000 229,000
Building 43,800 37,000 30,800 28,400 25,900 23,400 47,200 40,000 34,400
Covering 6,200 5,000 4,300 4,400 4,100 3,800 6,400 5,300 4,700
Ejector (installed) 2,000 2,000 2,000 2,000 2,000 2,000 2,000 2,000 2,000
Site Preparation 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000 1,000
Instrumentation 50,000 50,000 50,000 20,000 50,000 50,000 50,000 50,000 50,000
Subtotal 596,000 494,800 447,000 432,100 422,200 420,600 602,600 532,600 492,900
Storage Tanks 116,000 116,000 116,000 116,000 116,000 116,000 11,600 11,600 11,600

 

 

Total Capital Cost $712,000 $610,800 ¢563,000 $548,100 $538,200 $536,600 $614,200 $544,200 $504,500
Steam Cost/yr 8,330 8,980 9,940 10,570 11,530 13,130 8,330 8,980 9,940

 
21

7.2 Thermal Diffusion System

The theory of Jones and Furry (9} states that the temperature and pres-

sure fix the operating parameters of a thermal diffusion system.

The calcu-

tations are presented to illustrate that thermal diffusion systems of the

scale required by an MSBR are uneconomical,

From Equations 70-72 of Jones

and Furry, we obtain for concentric columns:

These equations are used subject to the constraint that 5 <K_/
an efficient operation K¢/Ky

3

7 3 2
k= L2 S4E a1/
Kd = 2w p DB

= 10,

< 25, For

The ratio of B/2w = 20 co?regponds to

the value of Jones and Flurry and Verhagen (11).

Inserting the appropriate values for hydrogen,

= 1 atm P=5§

300°K; T, = 600°K T, =
1.18 x 107% poise No=
0.54 x 107% g/cc 5 =
2.99 cmzlsec D =
0.174 o =

= 1.253 cm ow =
15.9 cm B =
0.774 x 107 g/sec H =

= 5.06 x 1072 g-cm/sec K. =

= 5.07 x 107 g-em/sec K, =

g = rate

in g/sec.

of mass transport of the desired

we find that;

atm
300°K; T2 = 600°K
1.18 x 1077 poise
2.71 x 1074 g/cc
0.5225 cmz/sec
0.149
0.574 cm
11.48 cm
8.17 x 107% g/sec
2.66 x 1073 g-cm/sec
2.66 x 107

species, i.e., the production rate
22

McInteer (10} has reported that o/H must be much less than one; and
from Jones and Furry (9) we see that o/H should be on the order of 10-%4 to
10-6 for each stage. MSBR requirements are such that the total flow rate
is 9.55 x 10-4 g/sec. To satisfy theory 104 to 106 columns in paraiiel
will be required. Based on the theory of Jones and Furry, Verhagen and
Sellschop (11) designed and operated a thermal diffusion system for tritium
enrichment.  Their apparatus is schematically shown in Fig. 6. The speci-
fications are given in Table 6.

Table 6. Thermal Diffusion System of Verhagen and Sellschop (11)

 

Stage 1 Stage 2
Type of unit concentric tube hot wire
Radius, hot wall 6.30 cm 0.02 cm
Radius, cold wall 8.76 cm 1.50 cm
Length 7.20 m 2.75 m
Temperature, hot wall 600°K 1200°K
Temperature, cold wall 1 300°K 300°K
Power consumption 18.5 kwh 1.66 kwh
Pressure 1 atm I atm

 

At the required enrichment of 1000-fold, this system has a production
capacity of only 0.125 cc(STP)/min of tritium. Based on this design approx-
imately 5150 systems in parallel are necessary to process the 642 cc(STP)/
min from the MSBR. Aside from the extreme difficulty in maintaining a
uniform feed rate to each column [to which thermal diffusion systems are
very sensitive (9)], the 9ower consumption is extremely high. The overall
system requires 72.7 x 107 kwh/yr at $0.004/kwh or a yearly power cost of
$2.9 million.

7.3 Cryogenic Distillation System

7.3.1 Design of Cryogenic Distillation Column

 

The specifications given in Table 3 were based on an o of 1.6. This
value is for a hydrogen-deuterium system; however, it is believed that the
a for the tritium system would be at least as great if not greater. g
23

e

 

': // HZ Cutlet

1/ /T

‘;;iff ‘{V/’Coo1ing Jacket

  
  

 

 

 

 

 

  
 
   

-f

Mixed Gas //é’

Inlet Heater

 

g 1

 

Cooling Jacket

Product Cutlet

 

AN

 

 

 

 

 

 

MASSACHUSETTS INSTITUTE OF TECHNOLOGY
SCHOOL OF CHEMICAL ENGINEERING PRACTICE
AT

OAK RIDGE NATIONAL LABORATORY

 

TWO-STAGE THERMAL DIFFUSION SYSTEM

 

DATE DRAWN BY FILE RO, FIG.

k- 1§ No¥ de | GRS CEPS-X-1171 6

 

 

 

 

 

 

 
24

Computer calculations were made for the four different separations at
various reflux ratios and for both saturated liquid and saturated vapor
feeds. Selected results are presented in Table 7.

Table 7. Variation of NTP with Operating Parameters
for H2 Distillation Column

 

NTP 53 58 88 93 98 76 49 50
Q' 1 ] 1 1 0 0 1 G
Reflux Ratio 2 2 2 2 2 2.2 3 3
% Recovery 99.0 99.0 99.9 99.9 99.9 99.9 99.9 99.9
/T Product  10%  10° 16t 100 100 108 108 108

 

For a liquid feed and R = 2, it is seen that lowering H/T from 104 to

103 causes an increase of only five stages for either 99.0 or 99.9% recovery.

The difference between 99 and 99.9% recovery is 40 plates at R = 2, but NTP
is still under 100 for 99.9% recovery. If the feed is a saturated vapor,
NTP is slightly higher but the difference is insignificant. NTP drops
rapidly at reflux ratios above 2. A 10% increase in reflux from 2 to 2.2
lowers NTP from 98 to 76, and increasing the reflux from 2 to 3 lowers NTP
from 93 to 49.

In view of these results it was felt that a column with 100 theoretical
stages operating at & reflux ratio of 2 would be adequate for the most dif-
ficuit separation, i.e., recovery = 99.9% and H/T = 103, However, in calcu-

lating the cost of the system a reflux ratio of 5 was used toc provide opera-
tional flexibility, conservative separation capability, and cost estimate.

7.3.2 Basis for H, Distillation Costs

1. Column shell thickness 1/4 in., type 304 stainless steel
$1.25/1b fabricated (12)

2. Packing
$200/Ft° (14)

3. Heat exchangers

$4/ft2 purchase price for plate-fin type (20)

area calculated from Q = UAAT
25

2 Btu/hr-ft2-°F for gas-gas exchange

where U
AT = 5°% = 9°F

Btu )
mole refiux

LD
I

(4630 mole feed/hr)(6 mole reflux/mole feed)(7.44

207,000 Btu/hr

7.44 Btu is the AH per mole of Ho from saturated vapor at 1.5 atm to
295°K at 1.5 atm (21)

area = (207,000)/(2)(9) = 11,500 ft°
4, Compressors

purchase cost = $7740 + ($99.5)(HP) (20)

For 3-stage reflux compressor,
cost = 3(7740) + 99.5(HP)

Calculated compressor size based on

KNRT p
_ 1 2\ (k-1)/k
W, = T k - (pT) ‘] (22)
(adiabatic)

where: k = Cp/CV = 1.405 for normal H2 (23)

From this equation the work of compressing the feed was calculated:
Ty = 300°K, Py = 1 atm, Pp = 3.7 atm, and Wy = 0.947 kcal/mole. For a feed
of 4630 mole/hr,

0.00156 HP)

W = (4385 kca]/hr‘) (m—

0 6.8 HP
A 20% margin of safety was allowed and an overall efficiency of 85% was
assumed, yielding a design of 10 HP.

For the 3-stage reflux compressor with interstage cooling, a Ty of
245°K was used for all three stages. The initial pressure is 1.5 atm and
interstage pressures of 5.5 and 9.5 atm were used in applying the above
equation which yields W, = 1.51 kcal/mole. For a reflux ratio of 5 the
power requirement is 54 HP, and with a 20% margin of safety and 85%
efficiency the design is 77 HP.

5. Expansion Turbines

Lady (24) quotes a capital cost of $25,000 for an expansion turbine
in a 200 ton/day liquid 02 plant. To be on the conservative side the same

 
26

figure is used in this cost estimate for the installed cost of the three
expansion turbines.
6. Insulated building

$500/ft2 floor space (15)
/. Two vacuum tanks, type 304 stainless steel

0.5-in.-thick x 15-ft-high x 3.5- ft -diam at $1.25/1b

fabricated = $9000 (12)

; Installed cost of vacuum bottles including all vacuum equipment is
36,000.

8. Electricity
$0.004/ kwh (g)
9. Raw materials:

H

5 $0.0048/scf (17)

0 $25/ton (liguid)

2
1G. Labor

one man/shift = $45,000
11. Maintenance
10% of investment (13)

12, 7000 hr operation per year
7.4 Computer Codes

A listing of the computer program for the water distlilation column
is shown below. The basis of calculation was a feed of 100 mole.

XF=.999998

55 ACCEPT 3K= $,K; IF(K),57,; ACCEPT KK;IF(KK),58,

ACCEPT $PRT= $,PRT,$NP= $,NP,$DELP= $,DELP

58 ACCEPT $REC= $,REC,$WOT= §$.WpT,$R= $,.R

BHTP=REC*0 . 0002 ; BNAT=BHT@*WpT ; B=BHTf+BUAT ; XB=BWAT/B;D=100. -B
DWAT=99. 9998-BWAT; XD=DWAT/D;RLOV=R/ (R+1.) ; SLV=(R+100. /D)/ (R+1.)
DXD=XD/ (R+1. ) ;BXB= (B/D)*XB/ (R+1.)

60 FPRMAT(15,F15.10,F9.5)
N=1;J=1;d0=1;Y=XD;PR=PRT;A1=RLPV;A2=DXD
27

20 CALL ALPHA(PR,A)
X=Y/(A-(A-1.)*Y);IF
24 TF(N),29, ;s IF(X-X
29 IF(X-XB)40,,
25 Y=A1*X+A2;J=J+1;IF(J-NP)26,27,27
26 PR=PR+DELP;G@ TP 20
27 DISPLAY $STAGES= f,a,$x= $,X;d=1;PR=PRT;JJ=1;68 T¢ 20
22 JJ=JJ+1;WRITE(1,60)J,X,A;G8 TP 24
30 N=0;A1=SL@V;A2=-BXB;G@ TP 25
40 WRITE(1,60)J,X;60 TP 55
57 STPP
END
S'E ALPHA(PR,A)
D'N AL(9),P(9)
DATA AL/1.0775,1.0735,1.07,1.067,1.062,1.056,1.0512,1.0478,1.044/
DATA P/50.,60.,70.,80.,100.,130.,160.,200.,250./
D@ 8 J=1,10
IF(P(J)-PR) 8,9,10
8 CONTINUE
9 A=AL(J); G@ TP 15
10 A=AL (J-1)+(AL(J)-ALE-1))*(PR-P(J-1)}/(P(J)-P(J-1))
15 RETURN
END
*

(J-JJ*K)24,22,22
F)30,30,25

The following list defines the important variables in the program.
XF mole fraction of Hy,0 in the feed
PRT pressure at the top of the column (torr)

NP maximum number of plates in a column; if NP is exceeded the program
continues with a new column, resetting the top pressure to PRT

DELP  pressure drop per theoretical stage (torr)

REC fractional recovery of HTO

WOT ratio of Ho0 to HTO in the bottoms

R reflux ratio = L/D

BHT®  number of moles of HTO in bottoms based on 100 moles feed
BWAT  number of moles of H20 in bottoms

B number of moles in bottoms

D number of moles in distillate

St XB mole fraction HZO in bottoms
28

XD mole fraction HZO in distillate
DWAT  number of moles HZO in distiilate
RLAV  L/V above the feed

SL@Y  L/V below the feed

DXD D « XD/V

BXB B - XB/V

A relative volatility of HZO to HTO

PR pressure at any point in the column (torr)
X Tiguid mole fraction of HZO at any point
Y vapor mole fraction of HZO at any point

The program assumes constant molal overflow which is certainly justi-
fied since the liquid and vapor streams are at least 99.9% H20. The varia-
tion of relative volatility with pressure is introduced according to the
curve in Fig. 2. The feed is assumed to be saturated liquid. |

For the cryogenic Hp distillation the program was modified for the
thermal condition of the feed, and the variable aipna subroutine was re-
placed with a constant alpha. The following printout shows the modified
program.

XF=.999998; A=1.6
55 ACCEPT $K= $,K; IF(K),57,
ACCEPT $REC=$,REC,$WPT= $.WOT,$R= $,R,30= $.Q
BHT@=REC*0.0002 ; BWAT=BHT@*WQT ;B=BHTP+BWAT ; XB=BWAT/B;D=100.-B
DWAT=99.9998-BWAT ; XD=DWAT/D;RLBV=R/ (R+1.) ;VL=D*(R+1.}-100.*(1.-Q)
SL@V=(R*D+Q*100. )/ VL ;BXB=B*XB/VL ;DXD=XD/ (R+1.)
60 FPRMAT(I5,F15.10)
IF(Q.EQ.1.0) GP TP 703XIN=(XF/(Q-1.)+XD/(R+1.))/(Q/(Q-1.)-R/(R+1.))
DISPLAY $XIN= §,XIN; GB T@ 69
70 XIN=XF
69 N=13Jd=1;d0=1;Y=XD;A1=RLAV ;A2=DXD
20 X=Y/(A-(A-1.)*Y);IF(d-d0*K)24,22,22
24 TF(N),29,;IF(X-XIN)20,30,25
29 IF(X-XB)40,,
25 Y=A1*X+A2;d=0+1;60 T¢ 20
22 JJ=JJ+1;WRITE(1,60)J,X;60 T@ 24
30 N=0;A1=SL@V;A2=-BXB;G) TP 25
40 WRITE(1,60)d,X;60 T@ 55
57 ST@P

e
VL
XIN

29

Definitions of the new variables are as follows:

heat to convert one mole feed to a saturated vapor
molar heat of vaporization

For a staurated liquid feed § = 1; for a saturated vapor feed § = 0.
vapor rate below the feed
Tiquid mole fraction of Ho0 at the intersection of the Q-line with

the upper operating line; at this X value the program switches from
the upper operating line to the lower operating line

7.5 Nomenclature

heat transfer area, cm2

actual number of plates

mean circumference = w(r1 - rz), cm
specific heat at constant pressure, cal/g-°C
specific heat at constant volume, cal/g-°C
coefficient of self diffusion, cmzfsec

heat exchanger

acceleration of gravity, cm/sec2
transport coefficient of thermal diffusion, g/sec
ratio of specific heats (Cp/CV)

transport coefficient of mixing due to convection
transport coefficient of mixing due to diffusion
atomic mass

number of gram moles

number of theoretical plates

system pressure, atm

heat transferred per unit time, Btu/hr
30

Q' ratio of molar enthalpy change in converting feed to saturated vapor
to the molar heat of vaporization

R reflux ratio, gas constant

T radius (r1 > rz), cm

STP standard temperature and pressure

T temperature, °C

AT temperature driving force, °C

U overall coefficient of heat transfer

W half the annular distance = %{r} - rz), chi

W, work of adiabatic compression per mole, Hp

o relative volatility, thermal diffusion constant
M viscosity. poise

D density, g/cc

o mass transport rate of desired species, g/sec

Subscripts

1,2 atomic mass of components 1 and Z respectively, or initial and final
conditions of temperature or pressure

7.6 Literature References
1. Jacobs, D.G., "Sources of Tritium and Its Behavior Upon Release to
the Environment, p. 1, USAEC Div. of Technical Information (1968).

2. Korsmeyer, R.B., personal communication, ORNL, November 20, 1970.

3. Briggs, R.B., and R.B. Korsmeyer, "Distribution of Tritium in an
MSBR," ORNL-4548, p. 54 (Feb. 28, 1970).

4, McNeese, L.E., personal communication, ORNL, November 11, 1970.
5. Yarbro, 0.0., personal communication, ORNL, November 12, 1970.
6. Huber, M., and W. Meier, "Measurement of the Number of Theoretical

Plates in Packed Column with Artifically Produced Maldistribution," Sulzer
Technical Review, Achema 1970, Sulzer Brothers Limited, Winterthur, Switzer-

Tand.
el

i
..........

31

7. Benedict, M., and T.H. Pigford, "Nuciear Chemical Engineering,"
443-453, McGraw-Hill, New York (1957).

8. Murphy. G.M., ed., "Production of Heavy Water," 9-52, McGraw-Hill,

New York (1955).

G. Jenes, R.C., and W.H. Furry, "The Separation of Isotopes by Thermal
Diffusion," Rev. Mod. Phys., 18(2), 151 (1946).

 

10. McInteer, B.B., L.T. Aldrich, and A.0. Nier, "Concentration of He3
by Thermal Diffusion," Phys. Rev., 74(8}, 946 (1948).

11. Verhagen, B.T., and J.P.F. Sellschop, "Enrichment of Low Level
Tritium by Thermal Diffusion for Hydroiogical Applications," Proc. Third
Int. Conf. on Peaceful Uses of Atomic Energy, Geneva, 1964, Vol. 12, 398-
408, United Nations, New York (1965).

12. Russell, W., personal communication, Knoxville Sheet Metal Co.,
Knoxvilie, Tenn., November 4, 1970.

13. Peters, M.S., and K.P. Timmerhaus, "Piant Design and Economics for
Chemical Engineers, 2nd ed., 655, McGraw-Hill, New York (1968).

14, Bragg, E.J., personal communication, Packed Column Corp., Spring-
field, N.J., November 4, 1970.

15. Corea, J.T., personal communication, ORNL, November 1970.

16. Chilton, C.H., "Cost Engineering in the Process Industries," p.
112, McGraw-Hi11, New York (1960?

17. Ellis, E.C., personal communication, Purchasing Div., ORNL,
November 11, 1970.

18. Becker, E.W., "Heavy Water Production," Review Series No. 21,
Int. AEA, Vienna (1962).

19. Hanny, J., "Eine Tieftemperaturanlage Zur Gewinnung Von Schwerem
Vasser," Kaltetechnic, 12(6), 158-169 (1960).

20. Perry, J.H., "Chemical Engineers' Handbook," 4th ed., Chapt. 11,
p. 14, McGraw-Hi1l, New York (1963).

' 21. Barron, R.F., "Cryogenic Systems," pp. 644-645, McGraw-Hi11, New
York- (1966).

22. Weber, H.C., and H.P. Meissner, "Thermodynamics for Chemical
Engineers,” 2nd ed., p. 114, Wiley & Sons, New York (1957).

23. Airco Rare and Speciality Gases Catalog, p. 48, 1968.
32

24, Lady, E.R., "Cryogenic Engineering,"” Univ. of Mich. Engineering
Summer Conferences, Chapt. XVII, "Expansion Engines and Turbines,” p. 2
(May 1965).

25. "Building Construction Cost Data - 1971," R.S. Means Co., p. 148
{Factories), 1971.
33

GRNL-MIT-117

Internal

.S. Bettis
.B. Briggs
.N. Haubenreich
.R. Kasten
.B. Korsmeyer
1. Lundin
L.E. McNeese
ewis Nelson
M.W. Rosenthal
12, J.5. Watson
13. M.E. Whatley
14-15. Central Research Library
16. Document Reference Section
17-19. Laboratory Records
20. Laboratory Records, ORNL R.C.
21. ORNL Patent Office
22-36. M.1.T. Practice School

E.S
R.B
P.N
P.R
R.B
M.

E
Lew

External

37. R.F. Baddour, MIT
38. L.A. Clomburg, MIT
39, S.M. Fleming, MIT