Physics / Reactor Physics Nuclear Engineering & Reactor Physics 100% Free Open Access
Chapter 5 • Theory & Derivations

Neutron Diffusion Theory & Fermi Age Slowing Down

Mathematical formulation of neutron transport and spatial diffusion theory: six-dimensional phase-space angular flux and rigorous derivation of the steady-state Boltzmann integro-differential transport equation; spherical harmonics P1 expansion, transport mean free path, and mathematical derivation of Fick's law of diffusion; one-speed steady-state neutron continuity and diffusion equations; physical boundary conditions, interface flux/current continuity, and the exact extrapolated boundary distance d = 0.7104 lambda_tr from Milne's transport theory; canonical solutions for point, line, and planar neutron sources in infinite moderating media, and thermal diffusion length L = sqrt(D/Sigma_a); Fermi age theory for continuous slowing down, the Fourier-Laplace solution of the age equation nabla^2 q = dq/d tau, and the Gaussian spatial slowing down kernel; migration area M^2 = L^2 + tau, migration length M, and spatial dispersion of fission neutrons.

§5.1 Fundamentals of Neutron Transport: Phase-Space Angular Flux & Boltzmann Equation

1. Phase Space and Angular Flux

A complete statistical description of a neutron gas requires a six-dimensional phase space: position $\vec{r} = (x, y, z)$ and velocity $\vec{v} = v \vec{\Omega}$ (or energy $E = \frac{1}{2} m v^2$ and unit direction vector $\vec{\Omega} \in S^2$). The angular neutron density $n(\vec{r}, \vec{\Omega}, E, t)$ is defined such that: $$n(\vec{r}, \vec{\Omega}, E, t) \, d^3 r \, d\Omega \, dE$$ is the expected number of neutrons in spatial volume $d^3 r$ around $\vec{r}$, with kinetic energies in $(E, E+dE)$, traveling in direction cone $d\Omega$ around $\vec{\Omega}$ at time $t$. The fundamental quantity of transport theory is the angular neutron flux $\psi$: $$\psi(\vec{r}, \vec{\Omega}, E, t) \equiv v \, n(\vec{r}, \vec{\Omega}, E, t)$$ Integrating over all $4\pi$ steradians yields the scalar neutron flux $\phi$: $$\phi(\vec{r}, E, t) = \int_{4\pi} \psi(\vec{r}, \vec{\Omega}, E, t) \, d\Omega = v \, n(\vec{r}, E, t)$$ The net neutron current density vector $\vec{J}$ is the first directional moment: $$\vec{J}(\vec{r}, E, t) = \int_{4\pi} \vec{\Omega} \, \psi(\vec{r}, \vec{\Omega}, E, t) \, d\Omega$$

2. The Steady-State Boltzmann Neutron Transport Equation

Balancing neutron gains and losses in an infinitesimal phase-space cell $d^3 r \, d\Omega \, dE$: $$\text{Streaming Loss} + \text{Collision Removal} = \text{In-Scattering Gain} + \text{Fission / External Source}$$ $$\mathbf{\vec{\Omega} \cdot \nabla \psi(\vec{r}, \vec{\Omega}, E) + \Sigma_t(\vec{r}, E) \psi(\vec{r}, \vec{\Omega}, E) = \int_{0}^{\infty} dE' \int_{4\pi} d\Omega' \, \Sigma_s(E' \to E, \vec{\Omega}' \to \vec{\Omega}) \psi(\vec{r}, \vec{\Omega}', E') + S(\vec{r}, \vec{\Omega}, E)}$$ This is the linear Boltzmann Integro-Differential Transport Equation. Because exact analytical solutions exist only for idealized infinite half-spaces, reactor physics employs systematic angular approximations—principally the Diffusion ($P_1$) Approximation.

§5.2 The P1 Approximation, Transport Mean Free Path & Derivation of Fick's Law

1. The Spherical Harmonics $P_1$ Expansion

In media where scattering dominates absorption ($\Sigma_s \gg \Sigma_a$) and locations are several mean free paths away from boundaries and localized sources, the angular flux is nearly isotropic with a small directional bias: $$\psi(\vec{r}, \vec{\Omega}) \approx \frac{1}{4\pi} \phi(\vec{r}) + \frac{3}{4\pi} \vec{\Omega} \cdot \vec{J}(\vec{r})$$ This two-term expansion in Legendre polynomials ($P_0$ and $P_1$) is the $P_1$ approximation.

2. Derivation of Fick's Law of Diffusion

Substitute the $P_1$ expansion into the monoenergetic transport equation: $$\vec{\Omega} \cdot \nabla \left[ \frac{1}{4\pi} \phi + \frac{3}{4\pi} \vec{\Omega} \cdot \vec{J} \right] + \Sigma_t \left[ \frac{1}{4\pi} \phi + \frac{3}{4\pi} \vec{\Omega} \cdot \vec{J} \right] = \frac{1}{4\pi} \Sigma_s \phi + \frac{3}{4\pi} \Sigma_s \bar{\mu}_0 \vec{\Omega} \cdot \vec{J} + \frac{S}{4\pi}$$ To extract the vector current equation, multiply the entire transport equation by $\vec{\Omega}$ and integrate over all solid angles $\int_{4\pi} d\Omega$: Using the solid angle identities: $$\int_{4\pi} \vec{\Omega} \, d\Omega = 0, \qquad \int_{4\pi} \Omega_i \Omega_j \, d\Omega = \frac{4\pi}{3} \delta_{ij}$$ The streaming term becomes: $$\int_{4\pi} \vec{\Omega} (\vec{\Omega} \cdot \nabla \phi) \, d\Omega = \frac{4\pi}{3} \nabla \phi$$ The collision and scattering terms become: $$\frac{3}{4\pi} (\Sigma_t - \bar{\mu}_0 \Sigma_s) \int_{4\pi} \vec{\Omega} (\vec{\Omega} \cdot \vec{J}) \, d\Omega = (\Sigma_t - \bar{\mu}_0 \Sigma_s) \vec{J} \equiv \Sigma_{\text{tr}} \vec{J}$$ Equating both sides yields: $$\frac{1}{3} \nabla \phi + \Sigma_{\text{tr}} \vec{J} = 0 \implies \mathbf{\vec{J}(\vec{r}) = - \frac{1}{3 \Sigma_{\text{tr}}} \nabla \phi(\vec{r}) = - D \nabla \phi(\vec{r})}$$ This is Fick's Law of Neutron Diffusion! The diffusion coefficient $D$ is: $$\mathbf{D = \frac{1}{3 \Sigma_{\text{tr}}} = \frac{1}{3 \Sigma_s (1 - \bar{\mu}_0)} = \frac{\lambda_{\text{tr}}}{3}}$$

3. Validity Conditions for Fick's Law

Fick's law is valid under four strict physical conditions:

  1. The medium is weakly absorbing: $\Sigma_a \ll \Sigma_s$.
  2. Scattering is linearly anisotropic in the LAB frame ($\bar{\mu}_0 = 2/(3A)$).
  3. The point of observation is at least $2\text{ to }3$ mean free paths away from strong localized sources.
  4. The point of observation is at least $2\text{ to }3$ mean free paths away from vacuum boundaries or material interfaces.

§5.3 One-Speed Steady-State Neutron Continuity and Diffusion Equation

1. The Continuity Equation

Consider an arbitrary volume $V$ bounded by surface $S$. At steady state, the rate of neutron loss must balance the rate of neutron production: $$\int_S \vec{J} \cdot d\vec{A} + \int_V \Sigma_a \phi \, dV = \int_V S \, dV$$ Applying Gauss's divergence theorem to the surface leakage integral: $$\int_V \left( \nabla \cdot \vec{J} + \Sigma_a \phi - S \right) dV = 0$$ Since this holds for any arbitrary volume, the differential neutron continuity equation is: $$\mathbf{\nabla \cdot \vec{J}(\vec{r}) + \Sigma_a \phi(\vec{r}) = S(\vec{r})}$$

2. The Steady-State Diffusion Equation

Substituting Fick's Law $\vec{J} = -D \nabla\phi$ into the continuity equation (assuming uniform diffusion coefficient $D$): $$\nabla \cdot (-D \nabla \phi) + \Sigma_a \phi = S(\vec{r})$$ $$\mathbf{-D \nabla^2 \phi(\vec{r}) + \Sigma_a \phi(\vec{r}) = S(\vec{r})}$$ Dividing through by $D$: $$\mathbf{-\nabla^2 \phi(\vec{r}) + \frac{1}{L^2} \phi(\vec{r}) = \frac{S(\vec{r})}{D}}$$ where $L$ is the fundamental thermal neutron diffusion length: $$\mathbf{L \equiv \sqrt{\frac{D}{\Sigma_a}} = \sqrt{\frac{1}{3 \Sigma_{\text{tr}} \Sigma_a}}}$$ In source-free regions ($S = 0$), the homogeneous diffusion equation reduces to the screened Poisson / Helmholtz form: $$\mathbf{\nabla^2 \phi(\vec{r}) - \frac{1}{L^2} \phi(\vec{r}) = 0}$$

§5.4 Physical Boundary Conditions & Extrapolated Boundary Distance d = 0.71 lambda_tr

1. Interface Continuity Conditions

At an interface between two distinct multiplying or moderating media (Medium 1 and Medium 2):

  1. Continuity of Neutron Flux: $$\phi_1(\vec{r}_{\text{int}}) = \phi_2(\vec{r}_{\text{int}})$$ (Neutron density cannot jump discontinuously across a mathematical plane).
  2. Continuity of Normal Current Density: $$-D_1 \left( \nabla \phi_1 \cdot \hat{n} \right) = -D_2 \left( \nabla \phi_2 \cdot \hat{n} \right)$$ (Neutrons are conserved; no neutrons can accumulate on a boundary of zero thickness).

2. Vacuum Boundary and Extrapolated Distance

At a vacuum boundary (outer surface of a reactor core adjacent to vacuum or air), no neutrons can enter the reactor from the outside: $$J_-(\vec{r}_s) = \int_{\vec{\Omega} \cdot \hat{n} < 0} |\vec{\Omega} \cdot \hat{n}| \psi(\vec{r}_s, \vec{\Omega}) \, d\Omega = 0$$ Using the $P_1$ partial currents formula: $$J_{\pm} = \frac{\phi}{4} \mp \frac{D}{2} \frac{d\phi}{dn} = 0 \implies \frac{\phi}{4} + \frac{D}{2} \frac{d\phi}{dn} = 0 \implies \frac{\phi}{|d\phi/dn|} = 2D = \frac{2}{3} \lambda_{\text{tr}}$$ Linear extrapolation predicts that the asymptotic flux vanishes at a distance $d$ outside the physical surface: $$d = \frac{2}{3} \lambda_{\text{tr}} \approx 0.667 \, \lambda_{\text{tr}} \quad (P_1\text{ approximation})$$

3. Exact Transport Result: Milne's Integral Problem

A rigorous transport-theoretic solution of the Milne problem (Wiener-Hopf method) demonstrates that the true asymptotic neutron flux vanishes outside the physical boundary at: $$\mathbf{d = 0.710446 \, \lambda_{\text{tr}} \approx 0.71 \, \lambda_{\text{tr}} = \frac{0.71}{\Sigma_{\text{tr}}}}$$ The extrapolated boundary $\tilde{R} = R + d$ allows reactor physicists to apply the simple Dirichlet boundary condition: $$\mathbf{\phi(\tilde{R}) = \phi(R + d) = 0}$$

§5.5 Fundamental Solutions of Diffusion Equation & Thermal Diffusion Length

1. Point Isotropic Source in an Infinite Medium

Consider a point source emitting $S$ monoenergetic neutrons per second at the origin $\vec{r} = 0$ in an infinite homogeneous moderator. In spherical coordinates with radial symmetry, the source-free diffusion equation for $r > 0$ is: $$\frac{1}{r^2} \frac{d}{dr}\left( r^2 \frac{d\phi}{dr} \right) - \frac{1}{L^2} \phi(r) = 0$$ Making the substitution $w(r) = r \phi(r)$: $$\frac{d^2 w}{dr^2} - \frac{1}{L^2} w = 0 \implies w(r) = A e^{-r/L} + C e^{+r/L}$$ Since physical flux must remain bounded as $r \to \infty$, $C = 0$: $$\phi(r) = \frac{A}{r} e^{-r/L}$$ To evaluate $A$, enforce source conservation as $r \to 0$: $$\lim_{r \to 0} 4\pi r^2 J(r) = \lim_{r \to 0} \left[ -4\pi r^2 D \frac{d\phi}{dr} \right] = 4\pi D A = S \implies A = \frac{S}{4\pi D}$$ The exact Green's function for a point source in 3D diffusion theory is: $$\mathbf{\phi(r) = \frac{S}{4\pi D r} e^{-r/L}}$$

2. Planar Infinite Sheet Source

For an infinite planar sheet source at $x = 0$ emitting $S_{\text{plane}}$ neutrons/$\text{cm}^2\cdot\text{s}$ into an infinite medium: $$\frac{d^2\phi}{dx^2} - \frac{1}{L^2} \phi = 0 \implies \mathbf{\phi(x) = \frac{S_{\text{plane}} L}{2 D} e^{-|x|/L}}$$

3. Physical Interpretation of Diffusion Length $L$

The mean square distance $\langle r^2 \rangle$ that a thermal neutron travels from its point of birth (thermalization) to its point of ultimate absorption is: $$\langle r^2 \rangle = \frac{\int_0^\infty r^2 \, \Sigma_a \phi(r) \cdot 4\pi r^2 dr}{\int_0^\infty \Sigma_a \phi(r) \cdot 4\pi r^2 dr} = \frac{\int_0^\infty r^3 e^{-r/L} dr}{\int_0^\infty r e^{-r/L} dr} = \frac{3! L^4}{1! L^2} = 6 L^2$$ Therefore: $$\mathbf{L^2 = \frac{1}{6} \langle r^2 \rangle \implies L = \sqrt{\frac{\langle r^2 \rangle}{6}}}$$ $L$ is directly proportional to the root-mean-square net displacement of a thermal neutron before absorption!

§5.6 Fast Neutron Slowing Down and Fermi Age Theory: Age Equation & Gaussian Kernel

1. The Fermi Continuous Slowing Down Model

In intermediate and heavy moderators, neutrons undergo many collisions with small fractional energy loss, behaving as a continuous fluid slowing down from birth energy $E_0$ toward thermal energy $E_{\text{th}}$. Combining the slowing down density $q(\vec{r}, E) = \xi \Sigma_s E \phi(\vec{r}, E)$ with the spatial leakage $-D(E) \nabla^2 \phi$: $$\nabla \cdot \vec{J}(\vec{r}, E) + \frac{\partial q(\vec{r}, E)}{\partial u} = 0 \implies -D(E) \nabla^2 \phi + \frac{\partial q}{\partial u} = 0$$ Since $\phi = \frac{q}{\xi \Sigma_s E}$, we obtain: $$\frac{D(E)}{\xi \Sigma_s(E)} \nabla^2 q(\vec{r}, E) = -E \frac{\partial q}{\partial E}$$

2. Definition of Fermi Age $\tau$

Enrico Fermi defined the age parameter $\tau(E)$ (having dimensions of $\text{length}^2 = \text{cm}^2$): $$\mathbf{\tau(E) \equiv \int_{E}^{E_0} \frac{D(E')}{\xi \Sigma_s(E')} \frac{dE'}{E'} = \int_{0}^{u} \frac{D(u')}{\xi \Sigma_s(u')} du'}$$ By construction, $d\tau = - \frac{D(E)}{\xi \Sigma_s E} dE$. Substituting into the slowing down equation yields the celebrated Fermi Age Equation: $$\mathbf{\nabla^2 q(\vec{r}, \tau) = \frac{\partial q(\vec{r}, \tau)}{\partial \tau}}$$ This equation is mathematically isomorphic to Fourier's classical heat conduction equation $\nabla^2 T = \frac{1}{\alpha} \frac{\partial T}{\partial t}$, where Fermi age $\tau$ plays the role of time!

3. Gaussian Spatial Slowing Down Kernel

For a point burst of $S$ fission neutrons born at the origin at age $\tau = 0$ ($q(\vec{r}, 0) = S \delta(\vec{r})$), the solution of the Fermi age equation is the 3D Gaussian distribution: $$\mathbf{q(r, \tau) = \frac{S}{(4\pi \tau)^{3/2}} \exp\left( - \frac{r^2}{4\tau} \right)}$$ The mean square slowing-down distance from fission to thermal energy $\tau_{\text{th}}$ is: $$\langle r_s^2 \rangle = \frac{\int_0^\infty r^2 q(r, \tau) 4\pi r^2 dr}{\int_0^\infty q(r, \tau) 4\pi r^2 dr} = 6 \tau_{\text{th}}$$ Hence: $$\mathbf{\tau_{\text{th}} = \frac{1}{6} \langle r_s^2 \rangle \equiv L_s^2}$$ where $L_s$ is the slowing-down length of the moderator!

§5.7 Migration Area M^2 = L^2 + tau, Migration Length & Spatial Dispersion

1. The Migration Area $M^2$

A neutron born in fission travels a distance while slowing down to thermal energy (characterized by Fermi age $\tau$), and then travels an additional distance as a thermal neutron before being absorbed (characterized by diffusion area $L^2$). The total mean square displacement from fission birth to ultimate absorption is the sum of the two independent random walks: $$\langle r_{\text{total}}^2 \rangle = \langle r_{\text{slowing}}^2 \rangle + \langle r_{\text{diffusion}}^2 \rangle = 6 \tau + 6 L^2$$ We define the Migration Area $M^2$: $$\mathbf{M^2 \equiv L^2 + \tau = \frac{1}{6} \langle r_{\text{total}}^2 \rangle}$$ The Migration Length $M$ is the square root: $$\mathbf{M \equiv \sqrt{M^2} = \sqrt{L^2 + \tau}}$$

2. Moderating Material Comparison

Moderator Diffusion Length $L$ (cm) Diffusion Area $L^2\ (\text{cm}^2)$ Fermi Age $\tau\ (\text{cm}^2)$ Migration Area $M^2\ (\text{cm}^2)$ Migration Length $M$ (cm)
Light Water ($\text{H}_2\text{O}$) $2.85$ $8.1$ $27.0$ $35.1$ $\mathbf{5.92}$
Heavy Water ($\text{D}_2\text{O}$) $171$ $29{,}240$ $131$ $29{,}371$ $\mathbf{171.4}$
Beryllium ($\text{Be}$) $21$ $441$ $102$ $543$ $\mathbf{23.3}$
Graphite ($\text{C}$) $59$ $3481$ $368$ $3849$ $\mathbf{62.0}$

Key Physical Distinction: In Light Water, slowing down dominates migration ($M^2 \approx \tau = 27\text{ cm}^2$ vs $L^2 = 8\text{ cm}^2$). In contrast, in Heavy Water, thermal diffusion overwhelmingly dominates migration ($L^2 = 29{,}240\text{ cm}^2 \gg \tau = 131\text{ cm}^2$), because thermal neutrons wander for nearly two meters before being absorbed!

ADVANCED UNIVERSITY HONORS PROBLEMS

Step-by-Step Solved Examination Problems

Comprehensive analytical derivations, quantitative calculations, and step-by-step examination solutions for Unit 1.

Solved Problem Example 5.1: Point Neutron Source in an Infinite Moderating Medium and Radial Flux Profile

An isotopic Americium-Beryllium ($\text{Am-Be}$) source emits $S = 2.5 \times 10^7\text{ neutrons/second}$ isotropically at the center of a large tank of pure light water. For thermal neutrons in light water at room temperature:

  • Diffusion coefficient $D = 0.16\text{ cm}$
  • Macroscopic absorption cross section $\Sigma_a = 0.0197\text{ cm}^{-1}$

(a) Calculate the thermal diffusion length $L$ in $\text{cm}$. (b) Calculate the thermal neutron flux $\phi(r)$ at radial distances $r = 5.0\text{ cm}$, $r = 15.0\text{ cm}$, and $r = 30.0\text{ cm}$ from the source. (c) Calculate the net neutron current density vector $\vec{J}(r)$ and evaluate the total leakage rate of neutrons passing outward through a spherical surface of radius $r = 10.0\text{ cm}$.

(a) Thermal Diffusion Length $L$:

$$L = \sqrt{\frac{D}{\Sigma_a}} = \sqrt{\frac{0.16\text{ cm}}{0.0197\text{ cm}^{-1}}} = \sqrt{8.1218\text{ cm}^2} \approx \mathbf{2.850\text{ cm}}$$

(b) Radial Flux Distribution: The Green's function for a point source in infinite diffusion theory is:

$$\phi(r) = \frac{S}{4\pi D r} e^{-r/L}$$

The prefactor is:

$$\frac{S}{4\pi D} = \frac{2.5 \times 10^7}{4\pi \times 0.16} = \frac{2.5 \times 10^7}{2.0106} \approx 1.2434 \times 10^7\text{ cm}^{-1}\cdot\text{s}^{-1}$$
  • At $r = 5.0\text{ cm}$:
$$r/L = 5.0 / 2.850 \approx 1.7544 \implies e^{-1.7544} \approx 0.17301$$
$$\phi(5.0) = \frac{1.2434 \times 10^7}{5.0} \times 0.17301 \approx \mathbf{4.30 \times 10^5\text{ neutrons/cm}^2\cdot\text{s}}$$
  • At $r = 15.0\text{ cm}$:
$$r/L = 15.0 / 2.850 \approx 5.2632 \implies e^{-5.2632} \approx 0.0051786$$
$$\phi(15.0) = \frac{1.2434 \times 10^7}{15.0} \times 0.0051786 \approx \mathbf{4.29 \times 10^3\text{ neutrons/cm}^2\cdot\text{s}}$$
  • At $r = 30.0\text{ cm}$:
$$r/L = 30.0 / 2.850 \approx 10.5263 \implies e^{-10.5263} \approx 2.682 \times 10^{-5}$$
$$\phi(30.0) = \frac{1.2434 \times 10^7}{30.0} \times (2.682 \times 10^{-5}) \approx \mathbf{11.1\text{ neutrons/cm}^2\cdot\text{s}}$$

(c) Net Leakage Through Spherical Surface of Radius $r = 10.0\text{ cm}$: Using Fick's Law:

$$J(r) = -D \frac{d\phi}{dr} = -D \frac{d}{dr}\left[ \frac{S}{4\pi D} \frac{e^{-r/L}}{r} \right] = \frac{S}{4\pi r^2} e^{-r/L} \left( 1 + \frac{r}{L} \right)$$

The total outward neutron current crossing the sphere of area $4\pi r^2$ is:

$$\text{Leakage}(r) = 4\pi r^2 J(r) = S e^{-r/L} \left( 1 + \frac{r}{L} \right)$$

At $r = 10.0\text{ cm}$, $r/L = 10.0 / 2.850 = 3.5088$:

$$e^{-3.5088} \approx 0.02993, \qquad 1 + \frac{r}{L} = 4.5088$$
$$\text{Leakage}(10.0) = (2.5 \times 10^7) \times 0.02993 \times 4.5088 \approx \mathbf{3.374 \times 10^6\text{ neutrons/second}}$$

Of the original $2.5 \times 10^7\text{ n/s}$ emitted, only $13.5\%$ cross radius $10\text{ cm}$; the remaining $86.5\%$ are absorbed within the inner sphere.

Solved Problem Example 5.2: Derivation of the Extrapolated Boundary Distance d = 0.7104 lambda_tr via Milne Transport

Consider Milne's classic half-space transport problem: a semi-infinite non-absorbing medium occupies $z \ge 0$, bounded by vacuum at $z = 0$. (a) In elementary diffusion theory, the angular flux is approximated as $\psi(z, \mu) = \frac{1}{2} \phi(z) + \frac{3}{2} \mu J(z)$. Show that the zero incoming vacuum boundary condition:

$$J_-(0) = \int_{-1}^{0} (-\mu) \psi(0, \mu) \, 2\pi d\mu = 0$$

yields the simple linear extrapolation distance $d = \frac{2}{3} \lambda_{\text{tr}}$. (b) Explain why the exact integral transport theory solution by Placzek and Seidel yields $d = 0.710446 \lambda_{\text{tr}}$. (c) For reactor-grade graphite with transport cross section $\Sigma_{\text{tr}} = 0.385\text{ cm}^{-1}$, compute the numerical value of $d$ under both approximations and calculate the absolute difference.

(a) Derivation of $d = \frac{2}{3}\lambda_{\text{tr}}$ in Elementary Diffusion Theory: The inward partial current across the surface at $z = 0$ is:

$$J_-(0) = 2\pi \int_{-1}^0 (-\mu) \left[ \frac{1}{4\pi} \phi(0) + \frac{3}{4\pi} \mu J(0) \right] d\mu$$

Evaluating the angular integrals:

$$\int_{-1}^0 (-\mu) d\mu = \left[ -\frac{\mu^2}{2} \right]_{-1}^0 = 0 - \left(-\frac{1}{2}\right) = \frac{1}{2}$$
$$\int_{-1}^0 (-\mu^2) d\mu = \left[ -\frac{\mu^3}{3} \right]_{-1}^0 = 0 - \left(-\frac{1}{3}\right) = \frac{1}{3}$$

Multiplying by $2\pi$:

$$J_-(0) = \frac{1}{4} \phi(0) + \frac{1}{2} J(0)$$

Setting $J_-(0) = 0$ at the vacuum interface:

$$\frac{1}{4} \phi(0) + \frac{1}{2} J(0) = 0 \implies \phi(0) = -2 J(0)$$

Substituting Fick's Law $J(0) = -D \left.\frac{d\phi}{dz}\right|_{z=0}$:

$$\phi(0) = 2 D \left.\frac{d\phi}{dz}\right|_{z=0}$$

The linear extrapolation distance $d$ is defined where the tangent line reaches zero: $\phi(-d) = \phi(0) - d \left.\frac{d\phi}{dz}\right|_{z=0} = 0$:

$$d = \frac{\phi(0)}{\left.\frac{d\phi}{dz}\right|_{z=0}} = 2 D$$

Since $D = \frac{1}{3} \lambda_{\text{tr}}$:

$$\mathbf{d = \frac{2}{3} \lambda_{\text{tr}} \approx 0.6667 \lambda_{\text{tr}}} \quad \text{(Q.E.D.)}$$

(b) Origin of the Exact Transport Correction $0.7104 \lambda_{\text{tr}}$: In reality, within $1\text{ to }2$ mean free paths of a vacuum boundary, the angular flux becomes severely anisotropic (peaked grazing along the boundary) because no neutrons arrive from the vacuum hemisphere. The $P_1$ two-term expansion breaks down in this boundary layer (the Knudsen transport boundary layer). Solving the exact Fredholm integral equation of transport theory via Wiener-Hopf contour integration yields the asymptotic linear profile whose zero-crossing is:

$$d = c \cdot \lambda_{\text{tr}} = \mathbf{0.710446 \lambda_{\text{tr}}}$$

This exact factor represents a $+6.57\%$ correction over the crude $P_1$ approximation.

(c) Numerical Calculation for Graphite: Given $\Sigma_{\text{tr}} = 0.385\text{ cm}^{-1}$:

$$\lambda_{\text{tr}} = \frac{1}{\Sigma_{\text{tr}}} = \frac{1}{0.385\text{ cm}^{-1}} \approx 2.5974\text{ cm}$$
  • Under $P_1$ approximation:
$$d_{P_1} = \frac{2}{3} \times 2.5974\text{ cm} \approx \mathbf{1.7316\text{ cm}}$$
  • Under exact Milne transport theory:
$$d_{\text{exact}} = 0.710446 \times 2.5974\text{ cm} \approx \mathbf{1.8453\text{ cm}}$$

Difference:

$$\Delta d = 1.8453 - 1.7316 = \mathbf{0.1137\text{ cm}} \approx \mathbf{1.14\text{ mm}}$$

In precision reactor criticality calculations, this millimetric difference in extrapolated boundary shifts the calculated eigenvalue $k_{\text{eff}}$ by tens of pcm!

Solved Problem Example 5.3: Fermi Age, Diffusion Length, and Migration Area Calculations for Beryllium

A nuclear reactor uses high-density Beryllium metal ($\text{Be}$, $A = 9.012$) as both moderator and reflector. The nuclear parameters at room temperature are:

  • Density $\rho = 1.85\text{ g/cm}^3$
  • Microscopic scattering cross section $\sigma_s = 6.1\text{ b}$
  • Microscopic absorption cross section $\sigma_a = 0.0092\text{ b}$
  • Slowing down length for fission neutrons $L_s = 10.1\text{ cm}$

(a) Calculate the transport cross section $\Sigma_{\text{tr}}$ and diffusion coefficient $D$ in $\text{cm}$. (b) Calculate the thermal diffusion area $L^2$ and diffusion length $L$ in $\text{cm}$. (c) Determine the Fermi age $\tau$ from fission to thermal energy. (d) Calculate the migration area $M^2$ and migration length $M$ of Beryllium.

(a) Atom Density, Transport Cross Section, and Diffusion Coefficient: Atom density of Beryllium:

$$N = \frac{\rho N_A}{M} = \frac{(1.85\text{ g/cm}^3)(6.02214 \times 10^{23}\text{ atoms/mol})}{9.0122\text{ g/mol}} \approx 1.236 \times 10^{23}\text{ atoms/cm}^3 = 0.1236\text{ b}^{-1}\text{cm}^{-1}$$

Macroscopic scattering cross section:

$$\Sigma_s = N \sigma_s = (0.1236\text{ b}^{-1}\text{cm}^{-1})(6.1\text{ b}) \approx 0.7540\text{ cm}^{-1}$$

Average cosine of scattering angle:

$$\bar{\mu}_0 = \frac{2}{3 A} = \frac{2}{3 \times 9.012} = \frac{2}{27.036} \approx 0.07398$$

Transport cross section:

$$\Sigma_{\text{tr}} = \Sigma_s (1 - \bar{\mu}_0) = 0.7540 \times (1 - 0.07398) = 0.7540 \times 0.92602 \approx \mathbf{0.6982\text{ cm}^{-1}}$$

Diffusion coefficient $D$:

$$D = \frac{1}{3 \Sigma_{\text{tr}}} = \frac{1}{3 \times 0.6982\text{ cm}^{-1}} \approx \mathbf{0.4774\text{ cm}}$$

(b) Thermal Diffusion Area $L^2$ and Length $L$: Macroscopic absorption cross section:

$$\Sigma_a = N \sigma_a = (0.1236\text{ b}^{-1}\text{cm}^{-1})(0.0092\text{ b}) \approx 1.1371 \times 10^{-3}\text{ cm}^{-1}$$

Diffusion area $L^2$:

$$L^2 = \frac{D}{\Sigma_a} = \frac{0.4774\text{ cm}}{1.1371 \times 10^{-3}\text{ cm}^{-1}} \approx \mathbf{419.8\text{ cm}^2}$$

Diffusion length $L$:

$$L = \sqrt{419.8\text{ cm}^2} \approx \mathbf{20.49\text{ cm}}$$

(c) Fermi Age $\tau$: By definition, Fermi age to thermal is related to the slowing down length by $\tau = L_s^2$:

$$\tau = (10.1\text{ cm})^2 = \mathbf{102.01\text{ cm}^2}$$

(d) Migration Area $M^2$ and Migration Length $M$:

$$M^2 = L^2 + \tau = 419.8\text{ cm}^2 + 102.01\text{ cm}^2 = \mathbf{521.8\text{ cm}^2}$$

Migration length:

$$M = \sqrt{M^2} = \sqrt{521.8\text{ cm}^2} \approx \mathbf{22.84\text{ cm}}$$

In Beryllium, thermal diffusion accounts for $80.4\%$ of the migration area ($L^2 / M^2 = 419.8 / 521.8$), while fast slowing down accounts for the remaining $19.6\%$.