What is the extrapolated boundary in diffusion theory and how is the extrapolation length z_e determined?

• A. The boundary for diffusion is extended by an extrapolation distance z_e beyond the physical boundary; z_e ≈ 0.71 × L_d.
• B. The boundary for diffusion is the same as the physical boundary; no extrapolation.
• C. z_e ≈ 2 × diffusion length.
• D. z_e is the mean free path for scattering.

Answer: (A)

In diffusion theory, the boundary that governs leakage isn’t just the physical surface of the material—the math introduces a small extra distance beyond it, called the extrapolation length. The idea is that the neutron flux would effectively drop to zero not exactly at the physical boundary, but at a point a short distance outside, because leaving the medium can be treated as if there were zero flux beyond that extrapolated surface. This extrapolated boundary makes the diffusion equation with the usual boundary conditions align with the actual leakage behavior.

The scale that controls how far this extrapolated boundary lies is the diffusion length, L_d = sqrt(D/Σ_a), where D is the diffusion coefficient and Σ_a is the absorption cross-section per unit path length. For a vacuum boundary, a standard result from solving the diffusion equation is that the extrapolation length z_e is about 0.7104 times the diffusion length: z_e ≈ 0.71 × L_d. So the effective boundary is located at a distance z_e outside the physical boundary.

This z_e is determined by solving the diffusion equation with the appropriate boundary condition that there is no net current leaving the medium into the vacuum (or by equivalently requiring the flux extrapolates to zero at x = x_boundary + z_e). In short, z_e is a small, material-property–dependent extension of the boundary, and for typical vacuum conditions its value relative to the diffusion length is about 0.71.

That’s why the correct understanding is that the boundary for diffusion is extended by an extrapolation distance, with z_e ≈ 0.71 × L_d. It’s not zero at the physical boundary, it’s not 2 × the diffusion length, and it’s not the mean free path for scattering.