reactor design, Bethe argues that 
resonances, in principle, could occur in 
either thermal or fast reactors. How- 
ever, he also points out that the condi- 
tions for a resonance are more likely to 
be satisfied in an unmoderated small- 
core reactor than in a moderated 
large-core reactor. Since fast reactors 
generally fit the former description and 
thermal reactors the latter, resonances 
might be expected to crop up more fre- 
quently in fast systems. 
Bethe feels confident that in either 
kind of reactor system resonances can 
be avoided by proper reactor design. 
Further observations planned by the 
AEC on EBR-I should provide valu- 
able information for such designs. 
Supercriticality upon meltdown. 
The final criticism of fast reactors is 
that the large amount of fissionable 
material contained in a fast-reactor 
core might through some accident re- 
distribute itself to form a highly super- 
critical mass. For instance, if all the 
enriched material in the EFFBR were 
assembled in a solid ball and then sur- 
rounded by a good neutron reflector, 
the resulting mass would be six times 
critical. 
This hazard is distinctly a character- 
istic of fast reactors. Thermal reactor 
fuels always more diluted because of 
lower U2%5 enrichment or the presence 
of a moderator. In fast reactors, only 
the geometric dispersion of the fuel for 
cooling purposes prevents the reactor 
from being supercritical. For this rea- 
son, it is essential to make sure that 
the fuel in a fast reactor can never 
assemble in one large mass. 
Such a situation might conceivably 
result if the core of the reactor were to 
melt during a severe power excursion. 
The fact that a partial meltdown did 
occur in the core of EBR-I (4) might 
lead to concern for the safety of all fast 
reactors in this respect. 
Two important aspects of the EBR-I 
incident argue against this: 
First, the meltdown accident oc- 
curred when the reactor was deliber- 
ately used in an unsafe way to test the 
positive temperature coefficient. 
A full-scale power reactor would 
not have a positive temperature coeffi- 
cient and would be designed so that 
operation under “experimental cir- 
cumstances’? would be impossible. 
Second, the EBR-I meltdown did not 
result in the formation of a critical 
mass. Boiling sodium entrapped in 
the core is believed to have forced the 
128 
molten uranium apart, preventing the 
formation of a critical mass. Ap- 
parently the melting of uranium in the 
presence of sodium, which can vapor- 
ize, does not lead to a supercritical 
assembly but tends instead to shut 
down the reactor. 
The most serious potential hazard 
associated with a meltdown arises from 
a situation not present in the EBR-I 
incident. At the moment, as far as 
safety is concerned, the fast-reactor 
concept’s biggest worry is the possi- 
bility that a supercritical meltdown 
might occur following the loss of the 
sodium coolant. Even if the reactor 
were immediately shut down with the 
safety rods, in the absence of coolant 
the fission-product afterheat would 
melt the core in a few minutes. Since 
there would be no sodium vapor to 
exert pressure, the molten core would 
collapse under gravity and might form 
a critical mass in the bottom of the 
reactor container. 
Dr. Bethe and Dr. Tait of the British 
Atomic Research Establishment at 
Harwell, have made rough upper-limit 
calculations of the energy that might be 
released by such an accident. The 
most realistic value obtained so far is 
a total energy release equal to 1,000 lb 
of TNT. Although Bethe points out 
that this figure may be a gross over- 
estimate, the magnitude of this result 
has stimulated a large effort to obtain 
more information. 
The AEC has initiated a comprehen- 
sive program at Argonne designed to 
arrive at a basic understanding of melt- 
down phenomena. 
In addition APDA has asked Nuclear 
Metals, Inc., Cambridge, Mass., to do 
experiments similar to those planned by 
ANL and has also undertaken further 
theoretical studies. 
Meanwhile fast-reactor designers are 
trying very hard to minimize the 
chances of sodium-coolant loss and are 
also building in features to prevent a 
nuclear event if a coolant-loss melt- 
down should occur. 
In the EFFBR secondary contain- 
ment is provided around all primary- 
coolant-system piping and vessels so 
that a leak in the primary system can- 
not cause excessive loss of sodium. In 
the event of loss of pumping power the 
design allows sufficient natural-convec- 
tion cooling for the decay heat. A 
large reservoir of sodium above the core 
appreciably increases the time required 
for a leak to drain the core. Other 
design allow for direct 
emergency action to restore cooling if 
somehow the core were to be completely 
drained. 
To limit the consequence of a melt- 
down a special meltdown container has 
been built into the lower sodium ple- 
num. Ifa meltdown should occur the 
molten fuel is expected to collect in the 
container in a noncritical geometry. 
However, there is some concern that 
the molten fuel might refreeze in the 
coolant channels in the blanket region 
beneath the core and never reach the 
meltdown container. Depending on 
results of additional analysis PRDC 
may abandon the lower blanket in the 
interests of safety. 
In summary, the question of fast- 
reactor safety is most likely one that 
can be eliminated by proper design. 
What constitutes proper design, how- 
ever, is not yet completely understood. 
Outlook 
We may conclude that, as things 
stand today, the feasibility of the fast- 
reactor concept as a basic power pro- 
ducer suffers from both technical and 
economic difficulties. Hopefully, the 
purely technical problems may quickly 
be eliminated; the economic questions, 
however, are likely to remain. 
The combined efforts of the AEC and 
PRDC can be expected to demonstrate 
in the near future that properly de- 
signed fast reactors are as safe to oper- 
ate as thermal reactors. Also experi- 
ence with the large sodium systems 
now being constructed may prove be- 
fore long that the industrial use of 
sodium coolant is technically and even 
economically within reach. On the 
other hand, it is more doubtful that the 
fast reactor can escape the economic 
penalty of the extra fuel reprocessing. 
Higher expenses for core fabrication 
and fuel inventory appear to be perma- 
nent economic disadvantages. 
In its favor the fast reactor has the 
one fundamental advantage of a breed- 
ing ratio substantially greater than 1.0. 
Thus, the value of the extra Pu pro- 
duced by a fast breeder must exceed 
the additional expenses of building and 
operating a fast breeder if this kind of 
reactor is to compete with thermal re- 
actors for the job of large-scale com- 
mercial power production. 
BIBLIO GRAPHY 
W. B. Lewis, Nucueonrcs 14, No. 10, 28 (1956) 
. Nucueonics 15, No. 2, R2 (1957) 
. H. A. Bethe. Reactor safety and oscillator 
tests, APDA-117 (1956) 
. Nuctronics 15, No. 1, 84 (1957) 
provisions 
ph wer 
