the Los Alamos molten-plutonium 
reactor experiment. But this reactor 
concept has extreme design require- 
ments and is regarded as something 
that may prove feasible only in the 
distant future. 
At the present time the fast- 
reactor concept seems to be burdened 
with excessive fuel-reprocessing costs 
which may or may not be eventually 
eliminated. In any case the fast 
reactor will always be faced with the 
extra expenses associated with its 
large fuel inventory. 
served to focus attention on the ques- 
tion of the safety of fast-reactor sys- 
tems. The EBR-I, the AEC’s first 
experimental fast breeder, is the only 
fast reactor with which we have had 
extensive operational experience to 
date. It is a matter of record that 
during its recent operation the EBR (a) 
exhibited resonance instability, (b) 
showed definite evidence of a positive 
temperature coefficient, and (c) suffered 
a partial meltdown of its core during a 
power excursion. 
It is easy to infer that these unpleas- 
ant traits are fundamental to all fast 
Safety 
The recent Power Reactor Develop- 
ment Corp. (PRDC) hearings have 
reactors. 
Principles of Fast Breeders 
A fast reactor is one in which the fission events are caused by fast neutrons—neutrons 
with energies close to 1 Mev. Thus, a fast-reactor designer must choose a configuration 
of materials that gives a neutron born in fission a good chance of being captured in 
fissionable material before it can either be captured elsewhere or be slowed down 
through scattering. The neutron cross section of materials in general (and of fissionable 
materials in particular) limit the design choice to one basic type—an enriched, un- 
moderated, core of fuel diluted as little as possible by other materials. 
The absence of a hydrogenous moderator and the careful restriction of the amount 
and kind of materials used in the core are necessary to keep the neutron spectrum fast. 
Thus, the use of light elements in a fast-reactor core is always avoided. Structural and 
coolant materials should be held to a minimum since the addition to the core of any 
material other than fuel tends to soften the neutron spectrum. For instance, in going 
from the EBR-I (a fast experimental reactor) to the Enrico Fermi Fast Breeder Reactor 
(a power reactor) the average neutron energy is depressed from 0.5 Mev to 0.25 Mev 
by the addition to the core of materials necessary for cooling and breeding. 
Since the fast-neutron fission cross section of U** is only a few times greater than 
the absorption cross section of U*", enriched fuel must be used to make a fast reactor 
critical. In addition, the fast reactor, to achieve criticality, meeds a greater total 
amount of U* than does a liquid-cooled thermal reactor. The reason is that the fast 
fission cross section of U*** is roughly 400 times smaller than its thermal fission cross 
section. Thus if the two kinds of reactors were to have the same core volume, the 
critical mass of U2" for the fast reactor would be many times larger than the critical 
mass of the thermal reactor. 
However, a fast-reactor core will always be smaller than a thermal-reactor core pro- 
ducing the same total power. The fast-reactor core is smaller to begin with since it 
has no moderator component. In addition, the designer makes the core still smaller 
to reduce the critical mass and to keep the neutron spectrum fast. The first effect is 
purely geometrical; the second holds because a large core will be more diluted with 
nonfuel materials. The minimum practical core size is fixed by the amount of power 
per unit of core that can be tolerated from heat-transfer considerations. 
It turns out that compared with a liquid-cooled thermal reactor of the same power 
the optimum fast reactor, although it has a much smaller core volume, still has a 
significantly larger critical mass. A fast reactor will then always operate at a higher 
power per unit volume of core and at a lower power per unit weight of invested 
fissionable material. 
A primary reason for building fast power reactors is that breeding of U** is only 
possible in a fast reactor. The breeding ratio (number of fissionable atoms created 
per fissionable atom destroyed) in a thermal reactor is less than 1.0 for the U**-fuel- 
U**-blanket and the Pu*°-fuel-U~*-blanket combinations.* The fast reactor has a 
breeding ratio near 1.2 for U*-U** and of at least 1.5 for Pu*®-U**. 
The fast-reactor breeding ratio is larger in general because of the smaller absorption 
of neutrons in structurals. For Pu?-U** and U?*-U**, in particular, it is larger be- 
cause of the fast fissions in U?* and the high value of a (ratio of the nonfission capture 
to fission cross sections in the fissionable isotope) for thermal neutrons in Pu and 
U*. 
To obtain the breeding ratios quoted above, the average energy of the neutrons in the 
fast reactor must be kept above 0.10 Mev. Below this energy the value of q@ in- 
creases rapidly, reducing the breeding gain. (See figure on p. 112) 
* Jt is true that the breeding ratio in a thermal reactor can be greater than 1.0 where 
U* is the fuel and thorium is the fertile material but we are concerned here with U*. 
126 
However, PRDC testimony 
refutes this guilt by association. In 
his testimony (2) Hans Bethe argued 
that the undesirable features of the 
EBR-I are not basic to fast-reactor de- 
signs and that a small amount of addi- 
tional knowledge in this field would 
point the way to fast-reactor designs 
that are safe. 
There are a number of characteristics 
peculiar, or sometimes believed to be 
peculiar, to fast reactors that are often 
cited by those who argue that this class 
of reactors is, in fact, inherently unsafe. 
They are cited here along with the 
counterarguments. 
Short prompt lifetime. Perhaps the 
most conspicuous of these character- 
istics is that the prompt-neutron life- 
time of the fast reactor is very short, at 
least a hundred times smaller than the 
prompt-neutron lifetime of a typical 
thermal reactor. Thus, if the reactiv- 
ity ever exceeds prompt-critical, the 
power in a fast reactor will begin to rise 
with a period a hundred times faster 
than that in a thermal reactor. 
Although the advocates of fast sys- 
tems admit that the power excursion 
resulting from a prompt-critical reactiv- 
ity would be much more severe for a 
fast reactor, they point out that the 
nature of fast reactors permits designs 
that make a prompt-critical situation 
very unlikely. 
Because fast reactors are insensitive 
to fission-product poisons and have a 
large critical mass, the excess reactivity 
required to operate for a given number 
of megawatt days is much smaller for a 
fast reactor than for a thermal reactor. 
A large thermal power reactor requires 
an excess reactivity of the order of ten 
dollars. The needs of a comparable 
fast power reactor can be satisfied by an 
excess reactivity of less than one dollar, 
which is below the prompt-critical 
region. Thus, unlike thermal-reactor 
designs, the total amount of excess re- 
activity built into a fast reactor is less 
than the amount required to make the 
reactor go prompt-critical. Fast-reac- 
tor designers feel that under these cir- 
cumstances it is difficult to imagine 
how a prompt-critical situation can 
arise without postulating a series of 
very unlikely events. 
Even if by some unforeseen combina- 
tion of circumstances the reactor should 
exceed prompt-critical, fast-reactor de- 
signers predict that nothing too serious 
would happen. NDA (Nuclear De- 
velopment Corporation of America) 
has carried out calculations for APDA 
(Atomic Power Development Associ- 
ates, Inc.) which show that even 
