22.

05 Reactor Physics - Part Thirteen

Criticality Conditions

1. Objective:

Thus far, we have calculated the shape of the neutron flux in critical assemblies of
certain geometries. However, we have not addressed the question of how to
select a material composition and/or a geometrical configuration that will result in
the critical condition. We do this by reviewing the definitions of multiplication
factors and the parameters elucidated in the neutron life cycle and then relating
these to the one-velocity, steady-state diffusion equation.

2. K-infinity and K-effective:

Definitions of both K-infinity and K-effective follow from our understanding of

the neutron life cycle. We define K-infinity as “the ratio of the number of
neutrons resulting from fission in the current generation to the number absorbed
in the preceding generation in a system of infinite size.” For an infinite system,
there is no loss from leakage. We define K-effective as “the ratio of the number
of neutrons resulting from fission in the current generation to the total number lost
by absorption and leakage in the preceding generation.”

The ratio of K-effective to K-infinity is therefore:

Fission
K eff Absorption + Leakage
=
K∞ Fission
Absorption

Neutrons Absorbed
=
Neutrons Absorbed + Neutrons Leaking Out

We define this ratio as P which is called the non-leakage probability for the
system. Thus,

K eff = K ∞ P

This relation is quite simple. Yet, it provides insight relative to the nature of
criticality. An assembly of materials will be critical if its value of K-effective is
exactly unity. We see that K-effective is the product of two factors: K-infinity

1
 and the non-leakage probability. The former is in turn a function of the materials
that form the assembly. The latter is a function of the arrangement (i.e., the
geometry) of these materials. Thus, the achievement of criticality depends on
both the choice of the fuel/moderator/coolant and on the size/shape and layout of
each of these. Thus, when designing an actual reactor, two approaches often
arise. One is to select geometry, hold it constant, and vary the material number
densities. For example, vary the fuel loading or amount of moderator. A second
approach is to select the materials, hold them constant, and vary the dimensions.
For example, vary the reflector thickness.

3. One-Group Critical Equation: The steady-state, one-velocity (or one-group)

diffusion equation is:

D∇ 2 φ − Σ a φ + ν Σ f φ = 0

Recall that this relation allows for neutron production from fission, neutron
absorption, and leakage. Hence, this equation does apply to reactors of finite size.
The above relation does not allow for neutron scattering. Hence, its use is limited
to situations where all neutrons have approximately the same energy.

For an actual problem, the one-velocity equation would normally be solved

numerically by first rewriting it as a set of difference equations and then
programming it on a computer. A difference equation is one in which derivatives
such as dφ / dx are approximations ( (φi − φi −1 ) / Δx . This approach has much to
offer. It is, after all, a major computational tool. But, it doesn’t provide physical
understanding. For that we rewrite the one-velocity equation as follows:

D∇ 2 φ − Σ a φ + S = 0

Where S is the source term (νΣ f φ) . For every neutron that is absorbed in the fuel,
there are K ∞ fission neutrons produced. Thus,

S = K∞Σa φ

And we have

D∇ 2 φ − Σ a φ + K ∞ Σ a φ = 0

or

⎡ (K − 1)Σ a ⎤
∇ 2φ + ⎢ ∞ ⎥ φ=0
⎣ D ⎦

2
For one energy group, we have previously defined D/Σa as equaling L2, the square
of the diffusion length. Thus,

∇ 2φ +
(K ∞ − 1) φ = 0
L2

or

∇ 2φ + B2φ = 0

Where B 2 is, also as previously named, the buckling and is equal to

K∞ −1
B 2c =
L2

The subscript c denotes that satisfaction of the above relation is essential for the
achievement of criticality.

The above relation gives us considerable insight into the critical condition. The
buckling, as was shown in part 12 of these notes, depends on the geometry and
dimensions of the reactor. The quantity (K ∞ − 1) / L2 depends on the materials
used to construct the reactor.

So, this equation is telling us that for a given set of materials, adjustment of the
geometry could result in criticality. Conversely, for a given geometry,
modification of the material composition could achieve criticality. An example
(see #6.3, p. 283, Lamarsh) illustrates this. The simplest realistic geometry for a
critical assembly is spherical because the only geometrical parameter that can be
varied is the radius. So, from our previous analysis of a bare sphere, we have;

B 2 = (π / R )2 Geometry Re quirement
K∞ −1
= Materials Re quirement
L2

Hence,

1/ 2
⎡ L2 ⎤
R = π⎢ ⎥
⎢⎣ K ∞ − 1⎥⎦

3
 For a given set of materials, this is readily solved for the critical radius, R. L2 is
the “diffusion area” and equals D/Σa. D is the diffusion length and is given by
1/3ΣTr. Σa is the macroscopic absorption cross-section (fuel and other materials);
ΣTr is the macroscopic transport cross-section. K∞, for a fast reactor, is given by:

K∞ = ηf

Tabulations of η, Σa, f, etc., are given in reactor handbooks.

4. Applicability of One-Velocity Analysis:

The analysis done so far has all been based on the one-velocity diffusion equation.
This use of the equation is acceptable for an unreflected fast reactor.

Can this type of analysis be extended to unreflected thermal reactors? The answer
depends on the extent to which neutron diffusion occurs. If diffusion is
significant during the slowing down process, then the size of the reactor will be
different from what our theory predicts. So, we seek some criterion that will let
us determine the suitability of the approach for a thermal reactor. To do this, we
will write two equations, one for the fast neutrons and one for the thermal ones.
We will then determine under what conditions these two equations give the same
result as the single one-velocity equation. The two equations are linked by their
source terms. Thermal neutrons are created by the scattering of fast ones. Fast
neutrons are created by the thermal fission of uranium-235. This analysis is
called “modified one-group” or “modified one-velocity.”

5. Modified One-Velocity Theory:

The first issue is that of nomenclature. We use the subscript 1 to denote the fast
group and 2 to denote the thermal group. The second issue is that of simplifying
assumptions. These are:

a) Fast Removal: It is assumed that there is no absorption of fast neutrons.

The sole removal mechanism is scattering to thermal energies. The
macroscopic scattering cross-section is Σ s1 .

b) Source Term for Fast Group: The reality is that fast neutrons originate
from both fast and thermal fission. So, we could write two fission terms
(Σ f 1φ1 and Σ f 2 φ 2 ) because both fast and thermal fission produce
neutrons in the fast energy range. Instead, we write the term for thermal
fission and modify it by use of ε , the fast fission factor. Thus, the source
term is:

εηfΣ a 2 φ 2

4
 Where ε is the fast fission factor, η is the thermal reproductive factor, and
f is the thermal utilization. The nomenclature and definitions are those of
the neutron life cycle that was studied earlier. The fast source term is
therefore:

S1 = εηfΣ a 2 φ 2

k ∞ Σ a 2φ2
=
p

Where p is the resonance escape probability.

c) Source Term for Thermal Group: Neutrons enter the thermal range by
scattering out of the fast range. The scattering rate is:

Fast Scattering / cm 3 s = Σ s1φ1

Of those that scatter out, some are absorbed by the U-238 resonances.
Thus, the number attaining thermal energies is:

Thermal Source Term = pΣ s1φ1

The two-group diffusion equations are:

Fast k ∞ Σ a 2φ 2
D1∇ 2φ1 − Σs1φ1 + =0
p

Thermal D 2∇ 2φ2 − Σ a 2φ2 + pΣs1φ1 = 0

The above equations apply to a bare (unreflected) reactor. Under such

conditions, it can be shown that the two group fluxes (φ1 and φ 2 ) have
the same shape and that shape is the one given by the one-velocity
equation. Thus,

φ1 = A1φ

φ2 = A 2φ

5
Where A1 and A2 are constants and φ is given by an equation of the type:

∇ 2φ + B2φ = 0

The thermal and fast fluxes differ in magnitude with A1 and A2 indicating
that difference.

Hence, the two-group equations become:

(
− D1B 2 + Σ s1 A1 + ) k ∞Σa2A 2
p
=0

(
pΣ s1A1 − D 2 B 2 + Σ a 2 A 2 = 0 )
The above relations constitute a set of linear homogeneous equations.
These will have a nontrivial solution if the determinant of the coefficients
of A1 and A2 is zero. Thus,

(
− D1B 2 + Σ s1 ) k ∞Σa2
p
=0
pΣ s1 ( 2
− D2B + Σa2 )
or

( )(
k ∞ Σ a 2 Σ s1 − D1B 2 + Σ s1 D 2 B 2 + Σ a 2 = 0 )
or

K ∞ Σ a 2 Σ s1
(D B1
2
)( 2
+ Σ s1 D 2 B + Σ a 2 ) =1

or

K∞
(
1 + B 2 L2T )(1 + B2 τT ) = 1

6
Where we have divided numerator and denominator by Σ a 2 Σ s1 and where
we define

L2T = D 2 / Σ a 2 and

τ T = D1 / Σ s1

The quantity L2T is the “thermal diffusion area” and τT is termed the
“neutron age.” Both have units of area.

To continue, the quantities in the denominator of the above relation are the
non-leakage probabilities. Thus, the probability that a thermal neutron
will not leak out is:

(
p T = 1 / 1 + B 2 L2T )
And the probability that a fast neutron will not do so is:

(
p F = 1/ 1 + B2 τT )
Here, we can write

K-effective = K∞PTPF

If the reactor is designed so that neutron leakage is small then the product
( )
of the leakage terms B 4 L2T τ T will be very small and can be ignored.
Thus, we obtain:

K∞
(
1 + B 2 L2T + τ T )=1
or

K∞
=1
1 + B 2 M T2

Where

M T2 = L2T + τ T

7
M T2 is termed the “thermal migration area.”

We are now in a position to compare our derived two-group analysis with

the previously analyzed one-group analysis. The one velocity (or one-
group) critical equation for an unreflected reactor is:

K∞
=1
1 + B 2 L2

The two-group result, or as it is more correctly called, the modified one-

group equation, is:

K∞
=1
1 + B 2 M T2

The two equations are of the same form and if τT is much less than L2T ,
the two equations are the same. So a comparison of τT (neutron age) and
L2T (thermal diffusion area) provides a means of determining if one-
velocity theory will be sufficient for estimating critical size and/or
composition:

Comparison of τ T and L2T

Moderator τT L2T
Light Water 27 8.1
Heavy Water 131 9600
Be 102 480
Graphite 368 3500

This shows that the one-velocity approach does suffice for moderators of
heavy water and graphite, that it is marginal for moderators of Be, and it
fails entirely for moderators of light water. Quantities such as the
“neutron age” and the “thermal diffusion area” are no longer widely used
in reactor analysis. What is important to recognize in the above discussion
is that one-velocity theory is only applicable if neutron scattering is not
significant.

Examples of the use of modified one-velocity theory for the determination

of composition are given in Section 6.5 of Lamarsh.

8
6. Relevance of One-Velocity and Modified One-Velocity Theory:

These two approaches combine neutron life cycle analysis with diffusion theory
to provide a quantitative method for calculating flux shape given a geometry and
for estimating both critical dimensions and compositions. This methodology was
the principal design tool for Generation I and some of the smaller Generation II
reactors. It allows the engineer to obtain a hands-on feel for the calculations.
However, it is not a viable approach for the design of reactors that are moderated
by light-water (which are to date the vast majority) or of modern large reactors.
For that, we require a thorough understanding of neutron scattering and the
development of either multi-group methods (cross-sections averaged over small
energy ranges) or Monte Carlo methods (exact cross-sections).