X822 RNL
TR Y T Ja'\j‘
OAK RIDGE NATIONAL LABORATORY MASTER 0]
Operated by OR_N(/IC/F" 5%’12' }fl
UNION CARBIDE NUCL EAR COMPANY .
Division of Union Carbide Corporation 0 R N L

= CENTRAL FILES NUMBER
58-19-17

Oak Ridge, Tennassee
"External Transmittal Authorized”
DATE:  December 15, 1958 copyno. Z 3
SUBJECT: The Need for U235 Breeding .

 

 

 

 

TO: Idisted Distribution

FROM: W. K. Ergen, E. D. Arnold, E. Guth,
S. Jaye, A. Saver, J. W, Ullmann

0F- (&S
PROPERTY OF
NOTICE
This document contains information of pretiminary WASTE MANAGEMENT

nature and was prepared primarily for internal use

at the Oak Ridge Notional Laoboratory, 1t is subject

to revision or correction and therefore does not DGCUMENT
represent a final report.

The information is not to be abstracted, LIBHARY
reprinted or otherwisa given pubiie dig eminntion

without the approval of the ORNI. patent brased, UCN-16212

Legal and lnforwation Control Department. @ 28
 

 

 

LEGAL NOTICE

This report was preporad as an account of Governmant sponsored work. Neither the Unitad States,

nor the Commisslon, nor any peraon acting on bshalf of the Commission:

A, Mokes any warranty or teprasantation, express o implied, with raspect to the aceuracy,
completeness, or usefulness of tha infermation contoined in this report, or that the vie of
ony Information, apparatus, mathod, or process disclosed in this repeort may not infringe
privately owned rights; or

B. Asaumes any liabilities with respact te the use of, or for damages resulting from the use of
any infermation, apparatus, method, or process disclased in this teport,

As used In the above, "persan acting on behalf of the Commission' Includes any employse or

cortractor of the Commission to the extent thot auch smoloyess or contractar prepares, handles

or distributes, or provides access to, any Infoematien pursuant to his employment o contract
with the Commission,

 

 

 
233 *
THE NEED FOR U~”- BREEDING

W. K. Ergen, E. D. Arnold, E. Guth, S. Jaye,
A. Sauver, J. W. Ulimann

Oak Ridge National laboratory
Cak Ridge, Tennessee
ABSTRACT

If the fissionable and fertile materials recoverable at approximately
today's éost are 4o constitute an energy reserve at least as large es fossile-
fuel energy reserves, about 20% of the fertile mafieriai hes to be burned. This
means reectors need & conversion ratio of aboub 96% in the average. Some reac%ors
will of necessity be burmers, hence the 96% average can only be obtained if other
reactors are breeders. |

Cdmparison of nuclear-flssion energy reserves with anticipated pover demands
indicates that breeding will not be necessary until 1980. Whether it will be-
come a necessity between 1980 and 2000 depends on vhich of & mumber of reasonsble
estimates are chosen. TIn order to keep up with fhe demand, breeders must have
a doubling time equal to or shorter than the doubling time of the demand ror nuclesy
power production. The la{ter doubling time is estimated to Be 5 to 10 years.

Such short doubling times will probably be achieved more easily with U253
breeders than with plutonium breeders. Thorium, the raw material for the 0235
breeder, is available in sulficient quantity in economically recoverable deposits
on the North American cortinent, but the raw material .ifor plutbnium breeders
(U258} 1s available in larger amounts.

IAmong U233 breeders the aqueous homogeneous reactor with its higily thermal
neutron spectrum and'consequently high n, its high specific pover, easy fission-
product removal, and short reprocessing time, will probably reach the shortest

doubling tines.

 

* Presented by W. K. Ergen at the American MNuclear Society Meeting in Detroit,
December 1958.
-Da
THE NEED FOR U253 BREEDING

The Oak Ridge National lLaboratory has conducted Q study regarding breeding
on the oo 227 cycle. Object of the study was, on one hand, the importance
of breeding on this cycle and, on the other, a comparisén of the various reactor
types with respect to their suitability as 0255 breeders. The importance of
breeding on the Th252-U233 cycle depends, in turn, on the‘importance of breeding
in general, and secondly on the comparison of the U238-Pu239 breeding cycle with

the Theaa-U2§5 cycle.

I. THE NECESSITY FOR BREEDING - GENERAL REMARKS

The fuel burnup cost in a straight burner, with prgsent prices, is about
% mills/kwvh. Thus a d-ifference of 10% in conversion ratio amounts to about
0.% mills/kwh, since a reactor of conversion ratio B could buy fuel amoufiting
to 10% of its burnup for 0.3 mills/kwh and end up with the seme amount of
fissionable materisl as a reactor of breeding ratio § + O.1l. A b;éeder and
8 convertef*éf reasonably high conversion ratio will not differ inicbnversion
ratio by more than a small multiple of 10%, and the difference in fuel-burnufi
cost will thus be smaller than the uncertainty in the estimated power cost of
a nuclear reactor. Fuel burnup cost on the basis of present prices will thus
not offer a strong reason in favor of breeding. |

A Justification for 5reeding thus involves an element,of.planning'for ‘the
future, a cofisi&eration of the time when the fissionable material recoverable
at reasonable cost will be exhauétéd and the nuclear-power economy depends on
| tapping the energy content of fertile material.
The justification for breeding is then analogous to the justification of

nuclear-power production in general - nuclear-power production is Justified

 

% TIn this connection, any reactor which produces less fissionable material than
it consumes i1s called a converter. '

 

 

 
with a view to future depletion of fossile fuel, rather than with a view to
present prices. The long-range planning is needed in the nuclear-power field
because of the long development and design time - estimated at 15 years - and
long life of power plants, estimated at 25 years. Thus, if breeding will be
hecessary 15 + 25 = 40 years hence, that 1s about by the year 2000, it iz not
too early to proceed with the developmenfi nov. Otherwise, we will have,
40 years hence, a large installed capacity which sti1ll could be used except for
the fact that it burns fissionable material which we can no longer afford to
burn; If it is the intention to scrap these reactors before they are worn out,
they would have to be burdened by larger depreciation costs during their usess

Any estimate of future supply and demand of fissionable materisl is very
uncertain. Estimate of how much fissionable material will be aveilable, and
at what price; depends on guesses as to future discoveries of deposits and also
on how much fissionable material the U.S8. will be able to import from ebroad,
or will export‘td other countries. Demand depends not only on the extremely
uncerbain requireménts of the power economy itself, but to a large extent on
the demand for nuclear-powered naval vessels, aircraft, rockets and weapons.
Concelvably the latter could even become a source rather than s sink 6f fission-
able material, as within the time periods considered nuclear disarmenment and
release of stock-piled material could become g reality. On the 6ther hand, some
of the uses of nuclear energy could be extremely wasteful of figssionable material.
An example for this is the "bomb rocket" intended to propel a large welght into
outer space by a large number of "small" nuclear-bomb explosion behifid the weight
to be lifted.

Thelimpact of fusion on fission reactors is iikewise very uncertain,

Concelvably, fusion could produce power cheaper than fission and put fission
L

power reactors out of business, or fusion based on the D-D reaction could be
& source of neutrons and hence of fissionable material. On the other hand,
lerge-scale power generation by fusion may be uneconomical, or unfeasible, or
dependent on outside supply of tritium and hence on fisslon reactors with
good neutron economy.

An accurate prediction éf the supply and demand situation with respect
to fissionable material is obviously impossible, but it 1s also unnecessary
for the purpose of deciding on the development of a breeder veactor. If there
1s & reasonable probability of breeding being attractive during the next 40 years,
such development would be indicated. In fact, it 1s quite likely that applications
of nuclear enexrgy will be propdsed vhich consume large amounts of fissionable |
material. The bomb rocket is an éxample. If there is a prospect of fissionable
material becoming scarce, the decision regarding such proposals mey very well
depend on the feasifiility of a sultable breeder. In that case, any effort spent
on development of a breeder would pey off in terms of hard information reg&rding
the feasibility of thé breeder, and 1n a firmer basis for the above decision.

Ever if breeding were of little interest fqi the near future in the United
States, it may well be impoftant in foreign countries with less native supply
of flssionable material. The potential need of foreign countries for power is
one of the main justificuations for development of nuclesr-power resctors.- An
analogous argfifiéné'could Justlify the development of breeders.

It appears that,'fé; a breeder, the doubling time is the more important
concept than the breeding ratio. In part this is duwe to the sofiewhat philec-
sophical point that breeding ratic is not always easy to define. Breeding ratio
is the ratio of the amount of fissionable material produced during a fuel cycle

to the amount of fissionable material burned during the cycle. If different
-5-

parts of the fissionable materiallhave different histories, the ”cyéle“ is a
somewhat controversial concept. On the other hand, the doubling time, that is
the time at which the amount of fissionable material has doubled, is clearly
defined.

More important than the sbove philosophical polnt is the fact that the
doubling time of the reactor can be compared directly with the doubling time
Of the demand of the fission-power economy. If the reactor doubling time is
longe: than the doubling time of the demend, then the reactors cannot keep up
with demsnd. A future shortage of the supply of fissionable material will be
reflected back to earlier dates; |

Doubling time has to be defined as the time in ézhich the whole fissionable
in#entory of a reactor is doubled. This lnventory includes fissionable material
contained in the reactor core, the blanket, the reprocessing plent, ete. Re-
processing losses have to bte taken into account.

In copsideringhthe reactor doubling time one should really consider the
average over the whole economy. Since there will be a lerge number of reactors
which will not breed (mobile resctors, for instance), the dncentive for ghort
doufiling time will be high in thosereactors which can be made to breed.

Short doubling time is, of course, only one parsmeter by which to judge a
reactor, High thermal efficiency (which means high operating temperature) is
‘another important parameter. A reactor with high themmsl efficiency, vhich does
no£ breed, uses a relatively small amount of fissionable material, and, though
it does not convert sufficient fertile into fissionable material, it leaves the
energy cohtent of same fertile material untouched, to be availsble for future
users who are ingenious enough to extract it. A low-thermal-efficiency breeder

replaces the fissionable material it uses, but it uses a relatively large amount
-6

of fissionable and hence fertile atoms, and whatever is wasted ls gone forever.
In this respect, high temperature reactors, like the lliquid-metal fuel reactor

and the molten-salt reactor, are desirable even 1if they are no breeders.

II. THE NECESSITY OF BREEDING - QUANTITATIVE CONSIDERATIONS
The following discussion of the reserves of uranium and thorium is based
on an AEC staff paper.l As in this reference, the regerves willl be gquoted as

their equivalent in U508 or Thoe, in units of short tons. The known US uranium

reserves recoverable at approximetely the present cost of $10/1v U308 are
230,000 tons, to which 350,000 tons should be added on the basis of specific
geological evidence in the areas known to contain deposits. This gives & total
of 580,006 tons. Known reserves in the United States recoverable at 550 to

$50/1b of U Og are 6,000,000 tons.*® The energy content of 1 ton of U Ogs if all

3 3
the urenium is used, amounts o 5.9 x 100 Btu. If only the U->° is burned, the

energy content of 1 ton of U308 would be 3.6 x 10T Btu ™t Thus, 580,000 tons of

reasonably certaln US reserves of high-grade ore would amount to Bl x 1018 Btu,

18 Btu, depending on whether all of the uranium, or only the U235,
¥

is burned. For comparison, the US reserves of oll and gas recoverable at up

or 0.21 x 10

to l.3 times the present cost, plus the US reserves of other fossile fuels re-

lS.B‘bu.2 Thus,

coverable at fip to twice the present cost, amount to 6.9 x 10
at least 20% of the U258 has to be burned before the uranium recoverable at ap-
proximately todsy's prices conbributes as much to the energy reserves of the US

as do the fossile fuels recoverable at the above cost. This 20% burnup corre-

sponds to a 96.4% conversion ratio.

 

1. Uranium and Thorium Raw Materials Supplies, Division of Raw Materials,
October 195¢.

2. P. C. Putnam, Energy in the Future, D. Van Nostrand, Inc., Toronto, New York,
London, 1953.

* In addition, there are large reserves recoverable at #50 or more per pound of
U508'

** Tgking into account that for every y22> fission, 0.18 of a U123 atom is lost
by radiastive capture.

 
-7-

Hence, the puclear energy reserves will be greater than thé fossile reserves
only if conversion ratios in excess of 96.4% are obtained. Breeding requires
conversion ratlos of 100% or more. Though it becomes the more difficult to in-
crease the conversion ratlo by l% the higher the conversion ratio alresdy is,
1t does not appear that breeding would be much harder to achieve than s conversion
ratio of 96.4%, |

There are, es previously mentioned, applications of nuclear energy other
than for civi;ian‘power prodfiction. Many of these spplications Bave to burn the
fissionable meterial, without being able to pay attention to high conversion ratio
or to breeding.. Assume that the use of nuclear fuel is divided in such & manner
that for every megawatt-day produced there is a fraction of x megawatt deys
produced in burners ahd & fractlon y produced in converters of conversion ratio C.
Then the burners use approximately x gram.of-U235 and the converters y(1-C) grem,
where x +y = 1. (In *%his semiqualitative conslderation we assume 1 g of U2 4o
be equivalent to 1 Mwd.) The emount of natural uranium needed is

140 [x + y(1 - c)] = 140(1 - y ¢).
If 20% of this is to be burned in the pfocess‘of producing the above mentioned
1 Mwd, then |
0.2 x 140(L -y C) =

or

(A
y'C - 'é‘g == 0-9614‘1

Thus, 1f 3.6% or more of the nuclear energy produced is to be made in burners,
(x> 0,036, y < 0.964), the conversion ratio of the civilian power producers would
have to be greater than one, that is the civilien power producers would have to

be breeders or else the useful nuclear energy reserves are smaller than the fossile
-8-

reserves. The exact value x = 0.036 depends of course scmewhat on the ground
rules used (and stated above), but qualitatively the results remein the same
under somewhat different rules.

We also can compare the supply of nuclear energy reserves with the estimated
demand for nuclear power production. Of course these estimates vary considerably,
as indicated in Table 1, which gives the nuclear power production in 1980 and
2000, as well as the doubling time of the nuclear power demand between 1980 and
2000. The 1980 estimates were taken from a table in Nucleonics,5 the "minimm"
being the J. A. Lane estimate, the "average" the arithmetic mean of the McKinney
report estimates, and the "maximum” the Davis and Roddis estimate. These
“minimum,"."average," and "maximum" estifiates given in the Nucleonics table also
for the years up to 1980 were plotted snd extrapolated in an admittedly somewhat
arbvitrary manner to the year 2000. In this manner the estimates for the year

2000 and the doubling times were obtained for Table l.

TABIE 1. NUCLEAR POWER PRODUCTION IN
THOUSANDS OF MEGAWATTS

 

‘ Doubling Time
1980 2000 in Years
Minimum Yo 168 10
Average 93 740 6.7
Maximum 227 3000 5ot

If the power demsnd is sgtisfied by reactors with a conversion ratio of

one, that ls by reactors which just barely miss being breeders, there is no

3\

 

3. DNucleonics 15, No. 4, p. 18 (1957).
-9~

consumption of fissionable material by burnup. The demand of fisgionable
material is solely determined by the inventory requirements. ITf the sbove
‘supply of 580,000 tons of 0308 are taken as a basis, and 100% recovery of
the contained U255 as assumed, the follofiing inventories of fissionsble

material per electrical megawait produced would be permissible.

TABLE 2. PERMISSIBLE FISSIONABLE INVENTORY FER Mwe:

 

(in kg/Mwe)
Nuclear Power Production
Egtimate , }2§9| 2000
Mindmum 15 19
Average 3 4.3
Maziimum 1k 1.06

Estimgtes of the fuel 'inventory, per electrical megevwatt produced, run
between 1 and 10 kg/Mwe for future reactors. Thus, 1if 8 fair fractlion of the
US uranium supply were used for other purposes than power produgtion, and iff
the highgst estimates of power production and the highest ffiel infentory per
Mwe are appllied, breeding msy became a necessity by 1980. If the lowest
eséimates of nuclear power production are used, breeding will be wanecessary
until sometime past the year 2000y in fact, the low-laventory reactors could
get by without breeding unhll 2000 even for the highest power production
estimates.

Uhforfignately, that Lewves the question unanswered whether breeding will
be a necessity within the next Lo years and, hence, whether breeder development
nov 1s timely. The only thing that can be said is that there is a strong
possibility of this being the case, and that breeder develogment should be

pursued as an insurance ageinst this possibility.
-10-

On the other hand, the uranium reserves recoverable at $30 to #50/1b of
U308 are so vast that only price increase, but not shortage of power, would
occur until well past the year 2000, even if the fossile fuel supply would

run out and breeding would not be available in time.

ITT. COMPARISON OF PLUTONTUM AND US2> BREEDING

From & pfa@tical viewvpoint, the main difference between plutonium and
U235 oreeding lies in the inventory of fissionable materisl and in the theo-
retically achiévable breeding gain.

The 1nventory is larger for plutonium breeders than for U255 brenders
This ig mainly a consequence of basic physical facts. because of the energy
dependence of the n of Py 39, plutonium breeders have to oyerate at high neutron
energles where the cross sections are small and where it takes meny plutonium
atoms to‘catchla neutron with sufficient probability before it escapes or slows
down. A contributing cause of the 1a:ge ifiventory is fhe intricate core structure
~of present fast breeder deSigns and the resulting large hold-up of flssionable
ma%erial external to the reactor.

The UE35 breéders, on the other hand, opejafie.best in the thermal region
vhere the cross séctions are large, and fever atoms suf'fice to prevent an
adequate number of neuwtrons from escaping. Morc importént, atoms other thsn
fissionable ones can be used to do a large part of the neubtron scattering and
escaph preventing. Neutron-energy degradation by these "other atoms" does not
have to pe avoided and is, in fact, desired. Thus, the critical mass and in-
ventory in a U253 breeder can be made very low, and the specific power very hilgh.

The theoretically achievable 5reeding_5§§§ (that is conversion ratio minus

one) for fast plutonium breeders is very high, values up to 96% have been
 

11~

cdmputed for small, low power reactors of this type. Breeding gains of 50%
seem to be obtsinable from practical reactors, even allowing for chemical re-
processing losses, ete. For'thermal U233 breedérs, such gains are out of the
question, n-2 amounts to only 0.28 so that practical breeding gains would be
limited to 10 to 15%. | |

The design paremeters of s typicel thermal breeder (a 300 Mwe aqueous
homogeneous feactor station) call for about 4500 kw (thermal) per kilogram of

* With this specific power a breeding gain of 10% wouid

fissionable meterial.
correspond to 4.2 years doubling time.

The Enrico Fermi fast breeder reactor has & ceritical mass of h85'kg for
300 Mw (thermal) output, that is 600 kw (thermal) per kilogrem of fissiomable
materiai. In the first assemblies the holdup of fissionable material in the
blanket and the external reprocessing cycle might be as much as three sdditional
critical masses, which would reduce the specific power to 150 kw (thermsl) per
kilogram of figsionable material. If tfie first core achieves a net breedifig
gain of 10% (taking into account chemical reprocessing losses) the doubling
time would be 100 years, that is, it would be irrelevant if compared to ‘the
doubling time of the electrical pover production.

On the other hand, future fast breeders, in particular those fueled with
plutonium, may héve higher breeder gains, by a factor of 5, as indicated gbove.
Another factor of 2 or o may be obtained by cutting down on kthe holdup in the
external repfioceSsing cycle, increasing tfie fiower density and so on. This
would_give doubling times of about 10 years for the fast breeders. The doubling'

time of the nuclear power production between 1980 and 2000 is 5 to 10 years,

 

4. Computed from "Fluid Tuel Reactors" (J. A. Iane, H. G. MacPherson and
F. Maslen, Editors), Table 9-9, p 508, Addison-Wesley Publishing Co., Inc.,
Reading, Mass., 1958. To the fissionable inventory quoted in the tabile,
16 kg have been added to allow for holdup in the "Chem Plant," etec. This
was done on the basis of oral communicetion from R. 3. Korsmeyer.
depending on which estimate is used. In view 'of the uncertainty in the above
numbers; it is thus possible that both the fast and the thexrmal breeder can
keep up with the nuclear-pover demand, or that neither can. 3Should only the
thermal oreeder, but not the fast breeder, be able to keep up with the demand,
then the thermal system would have a definite advantage. Whether this will be
the case is, on the basils of the above numbers, uncertain.
~ Another important point of comparison for the breeding cycles is the

aveilability of the fertile materials, uranium for the plutonium cycle, and
thorium for.fhe 0235 cycle. There is more thorium than uranium in the earth's
crust, but there is more uranium than thorium in ores recoverable at today's
prices.l At some price for the oxide, the awailability o thorium must equal
that of uranium but it is not knowm whether this price will be anyvhere near
a price at vhich nuclear energy is practical. |

The largest deposits of thorium are, however, in Brazil and India, and
voth countriés have at present embargoes against the export of thorium. Waether
this is serious for the time periocd under consideration in this Study is de-
batable. The North Amevicar continent, U.S. and Canads, have about 200,0001
tons of high-grade thorium ore, which is a fraction of the high-grade firanium
ore supply but still of the seme order of magnitude and very substantial. If
all converted into energy this supply would correspond to i3 x 1018 2tu which
is quite comparable to the whole fossile fuel supply of the U.S, and Cafiada.
It would cover the anticipated U.S. requirement of electrical energy well beyond
the year 2000. Considering the U.S. alone, the known thorium supply is rela-
tively small, but this is probably largely due to the leck of interest in Cinding

thorium.
~13-

In summary of the supply situation,there are considerably less thorimm
deposits in the US than uranium deposits, but if thorium were needed i+ could
ve found in sufficient quantities either by further expioration or-by import

from Canada, if not from India or =razil.

3

A
As far as price goes, the U7 should at present be cheaper than thorium

decause it is obtainable trom the tailings of UE’?J5 production which is needed

by users other than commercial power pi&nts. However, the amount of thorium
required by Ugf}5 breeders is smaller than the amount of UE:’8 needed by plutordium
breeders.

Both recycled thorium and plutenium are radiation hazards. There Seems
to be no significant difference in the handling of the two substances.

Thus, we fail to see any strong reason for favbring efie of' the two breeding
cycles over the other. A strong case can be made for parallet development of
the plutonium and U255 breeding cyclcs; Heither cycle has been demonstrated
to give breeder reactors of sufficiently lbw'inventory and doubling flimé.
Gambling on one cycle - with the possibility that the other cycle woul& hgve
been the vnly successful one - would be dangerous to bthe oxtent finat Lreeding

is necessary.

V. COMPARISON OF DIFFERENT REACTOR TYPES ¥OR U~~~ BREEDTHG

Perry, Preskiti wund Zuloert investigeied the use of gas-nooled, sraphliie-

&
oA -

moderated reactoré for 577 nreeding. The creeding paia turned out 1o bt
small, if not negative, mainly vecause of the dilenms vetween, cn ohs hang,,
large C:U ratic and laryg: absorption in graphite, and, on the other hand, 2
smaller C:U ratio with inéu?ficient moderation and lover n-values correspagiin
to higher neutron energies. The inventory was of course large. With reocercsd

to breeding, the gas-ccoled, graphite-moderated reactors are not competifive
b

with the aqueous homogeneous reactors.

The same authors are niv investigsting gas-cooled, DEO-moderated rezctors
4ith some misgivings about the absorptions in the zirconium-pressure tubes.
Liquid-nmetal fuel reactors and molten-salt reactors are bound to have laége
inventories and, at best, lov breeding gains, and do not appear to be suitable
as breeders for this reason. Their high-thermal efficienciés speak, howvever,
in their favor, even if conservation of fissionable and fertile material is
made the primary consideration (see Section I).

ilone of our investigations so far considered solid fuel elements with
verylliuwm cladding. If such elements were used with DQO noderator and coolant,
they may concelvably be competitive with agueous homogeneous reactors with
respect to doubling time. The decision would depend essentially on whether the
so far uncertaln poisoning of the agueous homogeneous reactor by soluble cor-
rosion products outweighs the poisoning of the solid fuel elements by fission

fragments.
1.
2.

1{'.
5"l5¢
16.
17,
18.
19.
20.
21,

-15-

DISTRIBUTION

D. Arnold 22,
B-- Bl"iggs 23‘
A. Charpie ol
L. Culler 25.
K. Ergen 26,
Guth 27.
Jaye 28.
H. Jordan 29.
A. Lsne 20-31.
G. MacPherson A2
M. Perry 35

511"'11'8a

C.
A.
M.
Je
Jde
A,
Ce

A, Preskitt
Sauer

Je« Skinner

A. Swartout
We Ullmann

M. Weinberg
E. Winters

Iaboratory Records, ORNIL-RC
Iaboratory Records '
Central Research Library
Document Reference Section
TISE, AEC