CANDU is, in many ways, an ideal reactor concept—not because it is minimal or superficially elegant, but because it aligns naturally with the underlying physics and avoids imposing constraints that other reactor types must navigate.
The use of heavy water as a moderator offers an exceptionally strong neutron economy, and this single choice drives many of the reactor’s advantages. It allows operation on natural uranium, and even when enrichment is used, very low levels of enrichment are sufficient, far below what light water reactors require. This relaxation of fuel supply constraints reduces dependence on enrichment infrastructure and enhances flexibility in terms of fuel cycle options.
The significant neutron margin means that the reactor is not restricted to a single optimized fuel form. Slightly enriched uranium can be used to improve burnup, recovered uranium can be recycled effectively, and thorium-based approaches can be implemented without necessitating a fundamentally different reactor design. The CANDU system accommodates these variations within the same framework.
The pressure tube configuration is another defining feature of CANDU, representing a strategic choice rather than just a mechanical one. By separating the moderator from the coolant, CANDU decouples functions that are tightly linked in pressure vessel reactors. The moderator remains at low temperature and low pressure, providing a large, stable thermal mass, while the coolant operates under pressure only within the fuel channels where heat removal is required.
This separation introduces structural and thermal resilience. The moderator acts as a passive heat sink, the system is inherently distributed rather than centralized, and there is no single massive pressure boundary whose failure would dominate the design basis. Transients that would tightly couple neutronics and thermal hydraulics in a single-vessel reactor are softened by this layout.
Online refueling directly results from this geometry. With pressure tubes and small fuel bundles, fuel can be inserted and removed during operation, avoiding shutdowns tied to fuel management. This approach also eliminates large excess reactivity swings associated with batch refueling. As a result, reactivity control becomes continuous, and power distribution across the core can be managed more smoothly.
Even the commonly cited drawbacks represent conscious trade-offs rather than unresolved weaknesses. While heavy water is expensive, its cost is increasingly contextual. Its production is energy-intensive rather than resource-constrained, and with abundant low-cost electricity—such as excess wind generation—it becomes much less of a barrier.
The positive void coefficient observed in some configurations is similarly a well-understood characteristic that is explicitly addressed through design choices and operating strategies.
CANDU does not rely on clever features layered on top of one another. Its strengths stem from selecting a favorable physical regime and building a reactor that works harmoniously within that context.
***
Heavy water is a better moderator than light water for reasons that go beyond its very low parasitic neutron absorption. Paradoxically, one of its key advantages is that it moderates less per unit volume.
Light water is such an efficient slowing medium that, in a thermal reactor, it often forces a compromise. Neutronics tends to favor a tighter lattice with less moderator in the fuel region, while thermal-hydraulics would benefit from more water—more flow area, more inventory, more margin in heat removal. Heavy water softens that conflict. Because each unit volume moderates less strongly, you can afford to place more water in the core without overmoderating. The lattice can be opened, coolant fraction increased, and geometry made more forgiving while still maintaining good neutron economy.
So the advantage is not only that fewer neutrons are absorbed, but that the designer is no longer forced to minimize water where it is most useful for heat transfer.
This becomes particularly relevant after shutdown. Decay heat persists while active systems may be degraded or unavailable. At that point, the design is judged by how robustly it can remove heat with whatever remains—inventory, geometry, and natural circulation. A core that was allowed to be “wetter” from the outset simply has more physical capacity to carry heat away. More water in the right places means more thermal inertia, more flow paths, and more tolerance to off-normal conditions without relying on precise control.
The usual shorthand—“heavy water has better neutron economy”—captures only part of this. It also enables a core layout where thermal considerations are not constrained as tightly by neutronics.
The traditional drawback is size. If both moderator and coolant are heavy water inside a single pressure vessel, the vessel becomes large. Ågesta nuclear power plant illustrates this: a relatively small unit at about 65 MW thermal (roughly 10–12 MW electric), yet already requiring a vessel on the order of 4.7 m in diameter and about 5.6 m in internal height. That scaling becomes increasingly unfavorable at higher powers.
The pressure-tube approach avoids this. The moderator sits in a low-pressure calandria, while the coolant flows in separate pressure tubes. The moderator can remain heavy water to preserve neutron economy, while the coolant in the tubes can even be light water to reduce heavy-water inventory and losses. The functions separate cleanly: heavy water provides the neutron environment, and the tube coolant handles heat removal.
In that sense, heavy water is not just a more “transparent” moderator. It allows a fundamentally different compromise—one where adding water to improve heat removal is no longer in tension with the neutronics.
***
Comparing RBMK and CANDU reactors involves more than just reactivity; it also concerns what happens after shutdown when decay heat must still be managed. Neutron physics plays a crucial role in this thermal behavior.
In CANDU reactors, heavy water has a long neutron mean free path, resulting in a need for a large moderator volume. This feature significantly benefits CANDU’s design: a spacious, cool heavy water moderator surrounds the fuel channels.
The calandria is not merely a requirement for neutronics; it serves as an essential heat sink. Even without forced coolant flow, heat can transfer from the fuel to the pressure tube and then into the moderator. This heat can further dissipate into surrounding structures, allowing for passive heat removal that is inherently integrated into the reactor’s geometry.
Conversely, the RBMK reactor has a different approach. Although graphite moderation also necessitates a large volume, the moderator is solid and hot, making it an ineffective heat sink. Consequently, cooling in an RBMK reactor is closely linked to the fuel channels and relies on forced circulation. There is no equivalent thermal reservoir efficiently coupled to the fuel.
Therefore, when cooling flow is lost, the RBMK design depends on its stored heat capacity, whereas the CANDU design utilizes established heat transfer pathways.
Both reactor types feature pressure tubes surrounded by a large moderator, but they lead to vastly different outcomes.
In reactor design, the implications of physics extend beyond simply neutronic considerations; they significantly influence thermal behavior when it is most critical.