under quite pessimistic conditions the 
total energy released during a prompt- 
critical power surge would probably 
not be enough to rupture the reactor 
vessel. 
Moderator accident. Another 
characteristic of fast reactors that 
might influence safety is that the reac- 
tivity of a fast reactor will increase if 
a moderator is somehow introduced 
For instance, if the 
sodium in a fast-reactor coolant system 
were partially replaced by some mate- 
rial containing hydrogen—say oil or 
into the system. 
water—the reactivity would increase 
and might conceivably go beyond 
prompt-critical. Designers believe, 
however, that fast-reactor systems can 
be constructed in ways that will elimi- 
nate this possibility. 
The EFFBR designers, for instance, 
have severely restricted the use of 
hydrogenous materials anywhere in the 
building. Since it is difficult to see 
how even the small amount of material 
that will be available could get into 
the sodium system, the EFFBR de- 
signers believe that for their machine 
the moderator hazard is negligible. 
Positive temperature coefficient. 
The observation that the EBR-I had 
a prompt positive temperature coeffi- 
cient is sometimes interpreted to mean 
that all fast reactors will have prompt 
positive temperature coefficients. 
Those familiar with the physics and the 
history of the EBR-I believe that the 
opposite is true. They contend that 
the observed positive coefficient most 
likely originates from the bowing of the 
fuel elements caused by radial tempera- 
ture gradients present during reactor 
operation. One can estimate the 
amount of bowing that should occur at 
a given reactor power, and this turns 
out to be about the right amount to 
explain the observed temperature co- 
efficient. Since no one can suggest any 
alternative effects to explain the obser- 
vations,* it is generally believed that 
bowing was the sole cause of the EBR-I 
positive temperature coefficient. Such 
an effect clearly depends on particular 
details of the core design and has noth- 
ing to do with characteristics of fast 
reactors in general. 
Resonance instability. The EBR-I 
experiments also revealed another, less 
well known type of instability, called 
resonance instability. This kind of 
reactor instability is characterized by 
spontaneous oscillations of the reactor 
power at certain discrete reactor oper- 
Strategies for Exploiting Our Atomic Resources 
From a technical point of view there are two possible strategies that could be used 
to extract energy from the world’s resources of uranium and thorium. One could use 
fast reactors exclusively or one could use mostly thermal reactors with a few fast 
reactors for breeding purposes. Economic feasibility will determine which scheme will 
eventually be used. It is too soon to say which system will ultimately prove superior. 
At the moment, however, for reasons presented elsewhere in this article thermal reactors 
seem to have the edge. 
If the thermal machines do indeed win out, the atomic power industry may develop 
along the following lines: In the beginning all commercial power will be generated 
by thermal converters using U** or thorium as raw materials. Since thermal reactors 
can breed successfully on a U**-Th cycle and there would be presumably no economic 
advantage to using fast reactors for this purpose, thermal reactors alone will continue to 
be used to exploit thorium energy resources. Thorium deposits are believed to represent 
about half of the total atomic-energy potential; therefore on this basis, thermal thorium 
converters will account for roughly half of the reactor population. Thermal reactors 
converting natural uranium cannot breed and so require periodic additions of fissionable 
material. At first U** obtained by enriching natural uranium will be used for this 
purpose. 
Later on Pu produced by a fast reactor will become cheaper than U** produced by 
enrichment. At this point the fast breeder will take over the function of supplying 
fissionable material to the uranium-converting thermal power reactors. The breeding 
ratio of the thermal converter is close enough to 1.0 so that a single fast breeder can 
supply a number of thermal reactors. Thus if the converter has a breeding ratio of 0.9 
and the fast reactor has a ratio of 1.5 the fast reactor can keep five thermal con- 
verters in business, each at a power level equal to that of the fast reactor. Thus the 
primary function of a fast-breeder power reactor of the future will be to feed fiissionable 
material to thermal reactors which in turn will generate practically all the power. 
In this way the total energy available in the world’s resources of U** can be com- 
pletely utilized. Although in such a scheme the number of fast reactors would be in 
the minority, their presence is absolutely essential. Without them most of the energy 
stored in U** would be inaccessible. 
For this reason, regardless of which reactor type 
proves more economical, the development of satisfactory fast breeder power reactors 
is a job which must be done. 
ating conditions. Again, since experi- 
ence with this kind of behavior is 
limited to the EBR-I, there has been a 
tendency to associate resonance insta- 
bility with fast reactors. However, 
recent studies indicate that the resonant 
behavior of the EBR-I, like the positive 
temperature coefficient, is caused by 
specific design features and would not 
be a general characteristic of fast 
reactors. 
In general, resonance behavior of a 
physical system depends on the exist- 
ence of a feed-back relationship be- 
tween two dynamic variables associated 
with the system. In a reactor the 
power influences the reactivity, and the 
reactivity in turn determines the power 
so that these two variables form a feed- 
back loop. Because of this feed-back 
mechanism reactor systems may have 
resonances at certain well-defined oper- 
ating conditions—when the reactor is 
operated at certain combinations of 
* At one time the Doppler effect in the 
reactor fuel was considered as a possible 
source of a prompt positive temperature 
coefficient. However enough information 
is now available to discount this possibility. 
For the EBR-I at best only 5% of the ob- 
served positive temperature coefficient can 
be explained by Doppler effect. In a large 
power reactor like the EFFBR, because of 
the higher ratio of U2%8 to U2%5, the Doppler 
effect will give a negative temperature 
coefficient. 
power level and coolant flow, the reac- 
tor power will begin to oscillate at a 
definite frequency without any external 
cause. If the reactor power is in- 
creased beyond the resonance value the 
amplitude of the power oscillations will 
increase exponentially with time. The 
reactor would then be intrinsically un- 
stable in this region. However, the re- 
actor would be completely safe as long 
as its power level is kept below the 
resonance value. 
Bethe incorporated in his PRDC 
testimony a thorough analysis of the 
theory and detection of resonances in 
reactorsystems (3). Heconcludes that 
a reactor resonance always requires the 
existence of a large delayed negative 
temperature coefficient of reactivity. 
Such coefficients would probably be 
associated with the heating of struc- 
tural components outside the core by 
convection. Bethe points out that this 
kind of difficulty would bé hard to fore- 
see in the design stage and might not 
be uncovered until the reactor is built 
and tested. (For instance, no one has 
yet explained the physical origin of a 
delayed negative temperature coeffi- 
cient in the EBR-I.) 
Because the existence of a delayed 
negative temperature coefficient de- 
pends on the details of the individual 
127 
