2. Critical Facilities for Basic Physics 
By H. C. PAXTON, Los Alamos Scientific Laboratory, Los Alamos, New Mesxico 
CritTIcaL assemblies that are intended 
to supply basic reactor physics data are 
distinguished from their reactor-devel- 
opment counterparts primarily in de- 
gree of complexity. The types of 
measurements—for example, determin- 
ing critical configurations and obsery- 
ing related neutron and fission charac- 
teristics, time behavior and effects of 
perturbations—cannot differ  signifi- 
cantly. Thus the general facility re- 
quirements are the same. There must 
be provision for safe operation without 
sacrifice of convenience. For satisfac- 
tory measurements, assemblies must be 
reproducible, while reasonably flexible, 
and instrumentation must be reliable. 
The ‘‘clean-critical”’ assembly is the 
basic source of reactor-physics data, 
whether the fissionable material is pres- 
ent in solution or slurry, as a lattice, or 
as a single piece of metal. The object 
is a complete understanding of ele- 
mentary assemblies, to serve as a 
springboard for the design of practical, 
more complex reactor systems, either 
steady power producers or nuclear ex- 
plosives, or to provide nuclear safety 
guidance. 
A complete experimental survey of a 
certain class of clean criticals involves 
an overwhelming number of parame- 
ters; these describe such things as the 
nature and distribution of the fission- 
able materials, moderators and other 
diluents, and the type, thickness and 
shape of the reflector. In practice, ex- 
perimental surveys are limited in scope 
and depend upon parallel computing 
programs to complete the survey by 
means of interpolation and extrapola- 
tion. Indeed, computing schemes are 
potentially so valuable for reactor de- 
sign and for filling in basic nuclear 
safety data that verifying them and 
providing input parameters for them 
have become major preoccupations of 
critical-assembly organizations. 
Although the same critical facility 
can be used for all types of basic physics 
measurements, one can distinguish 
three major categories: 
Critical surveys. There is no ques- 
tion that the simplified analytical 
schemes—such as diffusion theory for 
thermal reactors and the extrapolated- 
end-point method for fast systems— 
110 
are the work horses of reactor design. 
However, these schemes depend for 
their success on the proper choice of 
certain parameters (e.g., Fermi age, 
resonance escape, etc.) that are best 
obtained from critical surveys of simple 
geometries. 
Experimental critical-mass surveys 
useful for this purpose are illustrated 
by compilations intended for nuclear 
safety guidance. Even more pertinent 
to reactor development are moderated 
lattices of fissionable material, which 
are customarily interpreted in terms of 
simplified calculations. In this class 
“JEZEBEL,” a bare Pu sphere, is typical 
of “simple-geometry” critical assem- 
blies used at Los Alamos to accumulate 
physics data 
are the Brookhaven and Westinghouse 
water-moderated lattices of slightly 
enriched uranium and the Argonne 
heavy-water moderated lattices of U25 
and various diluents. 
Generally, assemblies for this pur- 
pose should have a simple unambiguous 
description at critical, instrumentation 
for measuring neutron multiplication 
during approach to critical, and instru- 
mentation for period measurement to 
calibrate controls so that correction can 
be made for their influence. 
Detailed computation checks. Sim- 
ple critical-mass data are insufficient 
to verify detailed analytical schemes 
such as multigroup, transport or Monte 
Carlo programs; these schemes and 
their input data require checking 
against all available properties of clean 
critical assemblies they are 
ready for use with practical reactors. 
Some satisfying confirmations of de- 
tailed calculations, for instance, have 
been made for the S, transport method 
by observing the properties of simple 
fast-neutron assemblies of U?*®, Pu?39, 
and U**s metal. 
In addition to requirements stated 
earlier for good critical-mass data, as- 
semblies for checking detailed compu- 
tations must provide internal neutron 
detectors to map flux and neutron 
spectra. Other useful measurements 
such as the Rossi alpha, pulsed-source 
data and the reactivity effects of 
changes in core density make further 
demands on the assembly and _ its 
instrumentation. 
Input-parameter measurements. 
Frequently reactor-calculation parame- 
ters may best be had by direct observa- 
tion on clean critical assemblies. Usu- 
ally the critical system is modified so 
that the resulting change in critical 
conditions is a measure of the parame- 
ter in question. For instance ORNL 
measured 7 for U2#3 by comparing criti- 
cal dilute aqueous solutions of U?*? and 
U**5 in spherical containers of the same 
size. 
Since the ratio of the just-critical 
fuel concentrations depends primarily 
upon 7(U23%), 7(U?%5) and the thermal- 
absorption cross sections of U#%, U?5, 
and H, (U2) could be obtained in 
terms of the other better-known 
parameters. 
Occasionally the critical facility will 
be used solely as a radiation source for 
an input-parameter experiment in 
much the same way as a test reactor is 
used to feed a thermal column. Thus 
the radiation may be studied on one 
day as an important property of the 
assembly, then used the next day for 
effects measurements. An example is, 
first, measurement of the spectrum of 
leakage neutrons from Godiva (the 
bare U?%5 metal assembly) and, next, 
application of the neutrons to biological 
measurements. Similarly, the study 
of dynamic behavior of Godiva through 
prompt critical led to its use as a pulsed 
neutron source. 
before 
