The Physics of Heavy Water Moderation

Heavy water, or deuterium oxide (D₂O), replaces the hydrogen in ordinary water with deuterium, an isotope with one proton and one neutron. This extra neutron nearly doubles the molecular weight, but the critical difference is in nuclear physics. When a fast neutron from fission collides with a nucleus, it loses kinetic energy most efficiently when the target nucleus has a mass close to the neutron itself. Ordinary hydrogen (¹H) is an excellent moderator because a neutron can be slowed in a few collisions, but it also has a moderate neutron-capture cross-section—about 0.33 barns—which removes neutrons from the chain reaction. Deuterium, with a mass of 2, is still light enough to slow neutrons quickly, yet its capture cross-section is roughly 0.0005 barns, more than 600 times smaller. This combination of effective slowing power and extremely low parasitic absorption makes heavy water a “transparent” moderator: neutrons can travel long distances without being absorbed, allowing a chain reaction to be sustained with natural uranium fuel.

The implications are profound. In a light-water reactor (LWR), the moderator’s absorption forces operators to enrich uranium to 3–5% uranium-235. A CANDU reactor, by using heavy water, can achieve criticality with the 0.7% uranium-235 found in natural uranium, eliminating the need for enrichment facilities and the associated energy, capital, and proliferation risks. The same physics also enables higher neutron flux at the fuel surface, which drives efficient conversion of fertile uranium-238 into fissile plutonium-239. Roughly half of the energy produced in a CANDU core comes from plutonium bred and fissioned in situ, a cycle that extracts nearly 30% more energy per tonne of mined uranium than a typical once-through LWR cycle.

Heavy water’s moderating ratio—the ratio of slowing-down power to absorption cross-section—is significantly higher than that of light water or graphite. This metric quantifies how effectively a moderator slows neutrons without absorbing them. Heavy water's moderating ratio is approximately 4,000, compared to about 70 for light water and 70 for graphite. This superiority stems directly from deuterium's low absorption, making D₂O the most efficient moderator used in commercial nuclear reactors.

CANDU Core Design: Built Around Heavy Water

The CANDU reactor is a pressurized heavy-water reactor (PHWR) with a unique horizontal configuration. The core centres on a large cylindrical vessel called a calandria, which holds the bulk heavy water moderator at low temperature (about 60–70°C) and near-atmospheric pressure. The calandria is pierced by several hundred horizontal pressure tubes, each containing fuel bundles and cooled by separate heavy water coolant at high temperature (around 310°C) and pressure (about 10 MPa). This separation of moderator and coolant circuits is a fundamental design feature. The moderator remains cool and low-pressure, providing a passive heat sink and allowing online refueling: fresh fuel bundles can be pushed into one end of a channel while spent fuel exits the opposite end without shutting down the reactor.

Fuel bundles are short (roughly 0.5 metres long) assemblies of natural uranium dioxide pellets clad in zirconium alloy. Because natural uranium requires a well-moderated neutron spectrum, the core is generously sized. The heavy water moderator occupies about 80% of the core volume, surrounding each pressure tube with a blanket of D₂O that thermalizes neutrons efficiently. Control systems include light-water absorber rods, adjuster rods for flux shaping, and liquid poison injection (gadolinium nitrate) for rapid shutdown. The entire design relies on the moderating and low-absorption properties of heavy water, and nowhere is that dependence more evident than during the startup sequence.

The calandria is typically made of stainless steel or low-alloy carbon steel, with wall thicknesses designed to withstand the large low-pressure load. Each pressure tube is made of a zirconium-niobium alloy (Zr-2.5Nb) chosen for its low neutron absorption and excellent corrosion resistance at high temperatures. The gap between the pressure tube and the calandria tube is filled with an insulating gas, minimizing heat transfer from the coolant to the moderator. This thermal separation is essential to maintain the moderator at low temperature, preserving its moderating efficiency and providing a safety margin during transients.

Startup Procedures: A Delicate Balance of Reactivity

Pre-Startup Verification and Heavy Water Conditions

Startup of a CANDU unit begins long before the first neutron multiplication. Operators confirm the integrity of both moderator and coolant heavy water circuits, checking inventory levels, isotopic purity (typically ≥99.75% D₂O), and chemistry parameters (conductivity, pH, and impurity concentrations). Even trace amounts of light water (H₂O) can degrade neutron economy by introducing hydrogen, a strong neutron absorber. Purification systems using fractional distillation or electrolysis are activated to maintain required purity. Once verified, heavy water is circulated through the calandria and the primary heat transport loop. The moderator is kept at a controlled temperature by dedicated coolers, while the coolant is preheated to avoid thermal shock during power ascension.

Achieving Initial Criticality

With stable heavy water conditions, a neutron source is inserted into or near the core. Most CANDU stations use antimony-beryllium (Sb-Be) or californium-252 (Cf-252) sources to provide a steady neutron flux measurable by startup instrumentation. Initially, the core is subcritical, with an effective multiplication factor (keff) well below 1.0. Reactivity is increased by adjusting the moderator level or by diluting a soluble neutron poison (e.g., gadolinium) that had been added during the outage to ensure deep subcriticality. As the poison concentration drops or the moderator level rises, more neutrons are moderated and reflected back into the fuel, increasing the neutron population.

Operators monitor log-power channels and wide-range neutron detectors while approaching criticality. The standard CANDU method brings the reactor supercritical on a very slow doubling time (typically 50–100 seconds). This cautious pace allows the digital control computers (DCC) to track neutron flux in real time and initiate protective trips if power rises too quickly. The DCCs receive signals from multiple ion chambers and self-powered neutron detectors distributed throughout the core. Once a self-sustaining chain reaction is confirmed, the reactor is held at low power, often a few percent of full power, for thermal stabilization of the primary and secondary systems. This period allows temperatures and pressures to equilibrate across the heat transport system, ensuring uniform conditions before further power increases.

Low-Power Physics Testing and Power Ascension

After initial criticality, a suite of low-power physics tests validates the core’s behaviour: control-rod worth, temperature and void reactivity coefficients, and flux distribution measurements. These tests confirm that the heavy water moderator is performing as predicted and that no anomalies have developed during the outage. Only when test results align with the operational safety report does the operator begin controlled power ascension. Because heavy water’s low absorption means small reactivity changes can cause relatively large power swings, power increases are managed through careful boron-dilution steps and control-absorber movements. The entire startup sequence—from first circulation to full power—can stretch several days, with heavy water chemistry and temperature continuously monitored as the plant moves through hollow-cone and full-core flux patterns.

A key aspect of low-power testing is the measurement of adjuster rod worth. Adjuster rods, made of stainless steel and containing minor amounts of neutron-absorbing elements, shape the neutron flux and compensate for burnout. Their worth in dollars must be known precisely to ensure that the reactor remains within safe reactivity margins during power maneuvers. Similarly, the coolant void reactivity coefficient is measured by comparing neutron flux with and without coolant in selected channels. In CANDU, this coefficient is typically negative, providing inherent feedback that shut down the reactor if coolant is lost—a safety characteristic that distinguishes it from some LWR designs. The low-power tests also verify the effectiveness of the moderator level control system, which can be used to fine-tune reactivity during startup.

Efficiency Gains Rooted in Heavy Water

Superior Neutron Economy

The defining efficiency metric of any nuclear cycle is neutron economy—the balance between neutrons produced and those lost to absorption or leakage. In LWRs, roughly 10–15% of neutrons are absorbed by the moderator. In CANDU, heavy water’s capture cross-section is so low that parasitic losses shrink to a few percent. The saved neutrons are available to convert fertile uranium-238 into plutonium-239, which then fissions and contributes roughly half the total energy output. This in-situ breeding boosts the energy yield per tonne of mined uranium by about 30% compared to a typical LWR fuel cycle. The Direct Use of Spent Pressurised Water Reactor Fuel in CANDU (DUPIC) process exploits this same economy: recycled fuel from LWRs, containing residual 0.9% uranium-235 and plutonium, can be fabricated into CANDU fuel bundles without reprocessing.

Natural Uranium Fuel and Supply Flexibility

Eliminating enrichment frees CANDU operators from dependence on enrichment services. Countries with domestic natural uranium deposits can fuel their reactors entirely from local sources, reducing geopolitical risk and currency exposure. The excellent neutron economy also makes CANDU a platform for alternative fuel cycles—thorium-based fuels have been tested in research reactors, and future designs may use thorium blankets with a fissile driver to produce less long-lived radioactive waste. This flexibility positions heavy-water reactors as key players in closing the nuclear fuel cycle. Additionally, the ability to use reprocessed uranium from spent LWR fuel without enrichment reduces waste volumes and extends uranium resource utilization.

High Capacity Factors and Online Refueling

Online refueling eliminates the periodic, multi-week shutdowns required to replace fuel in LWRs. CANDU stations routinely achieve lifetime capacity factors above 90%, with some units exceeding 95% in annual operation. For instance, Bruce Power's Unit 3 achieved a world record of 594 days of continuous operation without a planned outage. When maintenance is required, the horizontal fuel-channel layout allows individual channel inspections, pressure-tube replacements, and even retubing without rebuilding the entire core—a flexibility that has extended the licensed life of plants like Bruce and Darlington to 60 years and beyond. The separation of moderator and coolant circuits also allows coolant chemistry to be optimized for corrosion control without affecting moderator transparency, supporting net thermal efficiencies near 30%—competitive with other commercial reactor types.

The thermal efficiency of CANDU plants, typically around 30-32%, is slightly lower than that of modern LWRs due to the lower coolant temperature in the primary circuit. However, this is offset by the superior fuel utilization and high capacity factors. Annual refueling operations involve handling thousands of bundles, but the online nature means the reactor never needs to shut down for fuel, contributing to an overall energy output per unit of installed capacity that is often higher than LWRs with similar rated power. The use of natural uranium eliminates the energy penalty associated with enrichment, further improving the total system efficiency.

Safety Integration: Inherent Advantages of Heavy Water

Safety in CANDU design is not an add-on; it is embedded in the physics of heavy water. Because the moderator is physically separated from the coolant, a loss-of-coolant accident (LOCA) does not immediately remove the moderating medium. The calandria remains full of cool, low-pressure heavy water, providing a built-in emergency heat sink and ensuring that the chain reaction shuts down as the coolant void fraction increases. This negative reactivity feedback from coolant voiding gives operators time to respond. The positive void coefficient that can challenge LWRs during certain transients is carefully managed in CANDU through fuel-channel geometry, short fuel bundles, and real-time reactivity control. As detailed by the Canadian Nuclear Safety Commission (CNSC), multiple independent shutdown systems—including shut-off rods and liquid poison injection—are actuated by diverse parameters, never relying on a single neutron-flux signal.

Passive heat removal is another facet of the heavy-water advantage. In a station blackout, the moderator alone can absorb residual heat for hours, providing a generous grace period before active cooling needs to be restored. This feature was demonstrated during the 2011 earthquake in Japan, when operating CANDU units in Ontario rode through grid disturbances without incident. The thermal inertia of the large heavy water inventory (roughly 250–300 tonnes per unit) smooths out transients—whether from grid frequency changes or equipment faults—and helps keep fuel and pressure-tube temperatures within safety limits.

The dual shutdown system is a distinguishing safety feature. Shutdown System 1 (SDS1) uses mechanical shut-off rods that drop into the core by gravity, while Shutdown System 2 (SDS2) injects a concentrated gadolinium nitrate solution into the moderator at high pressure. Both systems are independently triggered by diverse parameters such as high neutron flux, low coolant pressure, or high moderator temperature. This redundancy ensures that no single failure can prevent the reactor from shutting down, and the use of heavy water as the medium for SDS2 injection takes advantage of its low absorption to distribute poison evenly without causing localized power peaking.

Heavy Water Inventory Management and Tritium Control

Operating a CANDU reactor requires a substantial up-front inventory of heavy water—approximately 250 to 300 tonnes for a single 700 MWe unit. Historically, Canada built dedicated heavy-water plants at Glace Bay, Nova Scotia, and later at Bruce, Ontario, where the Bruce Heavy Water Plant operated until the 1990s. Today, strategic reserves and international agreements meet new demand. Because heavy water slowly picks up light-water contamination through leakage, sampling, and maintenance, each station runs continuous purification systems—typically fractional distillation or electrolysis—to restore isotopic purity above 99.75%. This protects neutron economy and is a key operational cost. The cost of heavy water production is higher than light water, with current market prices around $300-$600 per kilogram, but the long-term benefits in fuel savings and operational flexibility justify the investment.

Tritium (³H) builds up unavoidably as deuterium absorbs a neutron: ²H + n → ³H. At concentrations above a few curies per kilogram, tritium becomes a radiological hazard that must be managed through periodic detritiation campaigns. Specialised facilities, such as the Tritium Removal Facility (TRF) at Ontario Power Generation’s Darlington site, strip tritium from the heavy water using vapour-phase catalytic exchange and cryogenic distillation. Clean D₂O is returned to the reactor, while the tritium is immobilized in stable compounds for storage or future use (e.g., as a fusion fuel). This closed-loop management keeps worker dose rates low and ensures public safety even under severe accident scenarios. The TRF at Darlington has removed over 11,000 curies of tritium from the plant's heavy water systems since its commissioning in 2002.

Innovations and the Future of Heavy Water Reactors

Research continues to push the capabilities of heavy-water reactors. Advanced fuel cycles marrying thorium blankets with a central driver fuel region are being modelled for CANDU-derived small modular reactors (SMRs), promising drastically reduced long-lived actinide waste. Digital twin technology now allows operators to simulate the exact neutron-thermohydraulic behaviour of the core during every startup sequence, identifying the fastest yet safest path to full power. International collaboration under the International Atomic Energy Agency’s (IAEA) advanced reactor research umbrella examines whether heavy-water designs could be adapted to run on re-enriched uranium recovered from spent fuel, further boosting resource sustainability.

At Bruce Power’s site, life-extension projects have integrated new heavy-water purification membranes that reduce tritium permeation. Similar approaches are being considered for refurbishments at other stations. The development of accident-tolerant fuel claddings and improved pressure-tube materials aims to increase permissible coolant temperature, lifting electrical output without altering the fundamental heavy-water architecture. These advances reinforce the view that the CANDU model—grounded in the unique properties of deuterium oxide—will remain a reliable contributor to low-carbon electricity grids for decades.

Several CANDU operators are exploring the use of recycled uranium from dismantled nuclear weapons as a fuel source. This initiative, known as the Consolidated Fuel Cycle, would blend weapons-grade highly enriched uranium with depleted uranium to create reactor-grade fuel that can be used in CANDU reactors without further enrichment. Such programs enhance global security while providing a practical use for surplus military materials. Additionally, the development of advanced digital control systems and automated refueling machines is expected to reduce startup times and improve operational flexibility, allowing CANDU plants to load-follow and integrate with intermittent renewable energy sources.

Conclusion

Heavy water is not merely an input to CANDU reactors; it is the enabling substance that defines their operational philosophy. From the meticulously managed startup that coaxes a subcritical assembly into stable self-sustaining power, to the continuous high-capacity-factor operation made possible by online refueling, D₂O’s low neutron absorption and high moderating efficiency are at the heart of every process. The safety margins baked into heavy-water-moderated cores, the economic advantage of natural-uranium fuel, and the flexibility to exploit recycled or thorium-based cycles all flow from the same physical thread. As global energy transitions accelerate, the quiet, scientifically elegant role of heavy water in CANDU technology will continue to deliver reliable, low-emission power while offering a platform adaptable to the fuel cycles of tomorrow.