Neutron Leakage Probability

Neutron leakage probability P_L is the likelihood that a neutron generated by fission escapes from the reactor core without being absorbed or causing further fission. It is defined as P_L = 1 - 1/k, where k is the effective multiplication factor. It is a key parameter because leakage represents a loss of neutrons that could otherwise sustain the chain reaction. For a critical reactor (k = 1), P_L = 0, meaning no leakage (in the theoretical infinite reactor). In a finite core, leakage is always positive, and its magnitude influences the required fuel enrichment, core size, and reflector design to achieve criticality.

Neutron leakage probability is inversely related to the core size because the surface-to-volume ratio increases as the core shrinks. A smaller core has a larger fraction of neutrons that are born near the surface and can escape before being scattered back. Mathematically, the leakage probability depends on the geometric buckling (B²) and the diffusion length; smaller cores have larger buckling, which increases leakage. This is why small reactors (like those in space or naval applications) require higher enrichment to compensate for the higher leakage.

The relationship is P_L = 1 - 1/k. For a critical reactor (k = 1), P_L = 0, indicating no leakage (this is the infinite reactor approximation). For k > 1 (supercritical), P_L is positive but less than 1; for k < 1 (subcritical), P_L is negative? Actually, the formula gives P_L = 1 - 1/k. If k > 1, P_L is between 0 and 1 (e.g., for k = 1.05, P_L = 0.0476). If k = 1, P_L = 0. If k < 1, P_L is negative, which is non-physical; this indicates that the simple formula applies only when k > 1 (since leakage must be positive). In practice, for a finite reactor, the leakage probability is positive and is given by P_L = 1 - (Σ_a / (νΣ_f)) in diffusion theory, which is equivalent to 1 - 1/k when k > 1.

A reflector (such as graphite, heavy water, or beryllium) surrounds the core to scatter neutrons back into the core, reducing the number that escape. This effectively decreases the leakage probability by providing an additional scattering medium that redirects escaping neutrons back into the fuel region. The reflector acts as a 'mirror' for neutrons, increasing the effective multiplication factor and allowing a smaller core to achieve criticality. The leakage probability in a reflected core is lower than in a bare core of the same dimensions, which is why reflectors are used in research reactors and some power reactors to improve neutron economy.

Fast neutrons (born from fission at ~2 MeV) have a longer mean free path and are more likely to leak out of the core before being moderated to thermal energies. Thermal neutrons, after slowing down, have a shorter mean free path due to higher scattering cross-sections, so their leakage probability is lower. The total leakage probability is a combination: some fast neutrons leak before slowing down, and some thermal neutrons leak after moderation. In thermal reactors, the fast leakage is often the dominant component. The six-factor formula (which includes fast and thermal non-leakage probabilities) separates these contributions to model the neutron balance accurately.

In one-group diffusion theory, the leakage probability is related to the buckling B² by P_L = (D B²) / (Σ_a + D B²), where D is the diffusion coefficient and Σ_a is the macroscopic absorption cross-section. The buckling depends on the core shape and dimensions: for a cylinder, B² = (2.405/R)² + (π/H)². Larger B² (smaller core) means higher leakage probability. The criticality condition is k_eff = k∞ / (1 + L² B²), where L is the diffusion length; this shows that leakage reduces k_eff and is accounted for via B².

The non-leakage probability (PNL) is the fraction of neutrons that do not leak out of the core. It is simply PNL = 1 - P_L. In the six-factor formula, the fast and thermal non-leakage probabilities are multiplied together: k_eff = k∞ × P_FNL × P_TNL. The total leakage probability is the sum of fast and thermal leakage contributions, but typically PNL is expressed as PNL = 1 - (leakage fraction). In a critical reactor, PNL = 1/k (since k = k∞ × PNL). Thus, leakage probability P_L = 1 - 1/k is directly tied to the effective multiplication factor.

The moderator's scattering cross-section and atomic mass influence the slowing-down length and the diffusion length. A good moderator (like H₂O) has a high scattering cross-section and small diffusion length, meaning neutrons are slowed down and absorbed locally, reducing leakage. In contrast, a poor moderator (like D₂O or graphite) allows neutrons to travel farther before being thermalized, increasing the diffusion length and thus the leakage probability. However, these materials also have lower absorption, which can compensate. The leakage probability is optimized by choosing a moderator that balances moderation efficiency with absorption and diffusion length.

In a fast reactor, the neutrons are not moderated; they remain at high energies (typically >100 keV) and have much longer mean free paths (because scattering cross-sections are lower and the neutron speed is higher). This increases the diffusion length and makes it easier for neutrons to escape the core before interacting. To compensate, fast reactors must have a compact core with a high fissile concentration (higher enrichment) and often a reflector (like depleted uranium or stainless steel) to reduce leakage. The leakage probability in a fast reactor is a key design parameter that influences the core's criticality and breeding ratio.

Control rods absorb neutrons, reducing the neutron population and changing the flux distribution. When a control rod is inserted, it creates a local flux depression, which can increase the leakage probability because the flux gradient near the rod increases the diffusion current out of the core. However, the primary effect of a rod is to reduce the multiplication factor by absorption, which increases the leakage fraction relative to the reduced absorption. The leakage probability itself is a property of the core geometry and material composition; control rods change the absorption cross-section, which in turn changes k and thus the calculated P_L = 1 - 1/k. A positive reactivity insertion (rod withdrawal) increases k and decreases P_L.

In a large commercial PWR or BWR, the total leakage probability is relatively small, typically around 2-5%. This means that 95-98% of the neutrons are absorbed in the core (either in fuel or in non-fuel materials) before leaking out. The fast leakage component is usually about 1-2%, and the thermal leakage is slightly lower. Larger cores have lower leakage; smaller cores (like those in research reactors or naval reactors) may have leakage probabilities of 10-20% or more, requiring higher enrichment to maintain criticality.

As fuel burns up, the fissile material decreases and fission products (like Xe-135) build up, reducing the multiplication factor k. According to P_L = 1 - 1/k, if k decreases, the leakage probability P_L increases. However, the actual leakage fraction may not change significantly because the flux distribution and core geometry remain the same. The increase in P_L is a mathematical consequence of the reduced k; physically, the leakage rate (neutrons per second escaping) might not change, but the fraction of neutrons that leak relative to absorption increases. This is one reason why the reactor needs to be refueled—as k drops, the leakage fraction becomes too large to sustain criticality.

In storage and transport, fuel assemblies are subcritical (k < 1). The leakage probability is high because the fuel is not in a reactor core with a reflector; it is in air or water with a large surface-to-volume ratio. In these configurations, the leakage fraction is significant (often >50%), and the multiplication factor is well below 1. Criticality safety analyses use leakage probability to ensure that even under optimal moderation conditions (e.g., water flooding), the system remains subcritical. The leakage probability is a key factor in determining the safe number of assemblies that can be stored together, with spacing and neutron absorbers used to increase leakage.

Temperature affects the leakage probability through changes in the neutron cross-sections and density. As the moderator (water) heats up, its density decreases, reducing the scattering cross-section (Σ_s) and increasing the diffusion length. This makes it easier for neutrons to leak out, increasing the leakage probability. The fuel temperature also affects the spectrum (Doppler broadening), which can change the absorption-to-scattering ratio. In general, the leakage probability increases with temperature, which is one component of the negative temperature coefficient of reactivity. This is an important safety feedback: as the core heats up, leakage increases, reducing reactivity.

In the one-group diffusion equation, the leakage term is -D∇²φ, where D is the diffusion coefficient. The leakage probability is the ratio of the leakage rate (D B² φ) to the total neutron source (νΣ_f φ). This gives P_L = (D B²) / (νΣ_f). Since k = νΣ_f / Σ_a and k = 1/(1 + L² B²) (with L² = D/Σ_a), one can show that P_L = 1 - 1/k. Thus, the leakage probability is directly related to the diffusion coefficient and the buckling. A larger D (longer diffusion length) leads to higher leakage probability, which is why materials with high D (like graphite) are used as reflectors, not as core moderators in thermal reactors.