Reactor safety FAQ

Could the fuel in the ILL's reactor melt?


We have already seen that as soon as the reactor shuts down the reactor core is adequately cooled by simple natural convection. Of course, for this to be possible there must be sufficient water in the reactor vessel. Accidents liable to result in the loss of the water in the reactor vessel can therefore lead to the meltdown of the fuel, as natural convection is no longer possible.

The reactor vessel and its associated cooling systems are designed in such a way that there must always be at least two independent failures for the water inventory of the vessel to be lost and the core to be exposed. The frequency of such an occurrence is therefore extremely low, around 10-5 to 10-6 per year (once every hundred thousand to a million years, on average).
This scenario was nevertheless taken into account in the most recent safety review of the installation. Following this review, an emergency water makeup system was installed. This allows water to be reinjected directly into the reactor vessel in this sort of scenario to ensure that there is always enough water in the vessel to cool the core by natural convection. Recovery pumps have been installed in the ‘crypt’ of the reactor to reinject the water that has leaked from the reactor vessel and thus re-establish a “closed” cooling system.
Finally, as part of the continuing process of safety improvement, ILL has recently proposed a new emergency core cooling system that will allow a communication passage to be opened up if necessary between the reactor vessel and the reactor pool and adjacent storage pool, providing a total volume of 1000m3 of water. Water from the reactor pool will then fill the reactor vessel by simple gravity, i.e. without the need for an external power source, allowing core cooling to continue. This new system is in place since 2012. 
With these new emergency systems the probability of a core meltdown will then be extremely low, approaching a frequency of 10-7 per year, the value at which the risk is considered to be “residual” and no longer needs to be taken into account in the installation’s design basis. 

In the Borax-type accident automatically included in the design basis, the core also undergoes partial meltdown but remains covered by the water in the pool. Most of the radioactivity released during the meltdown is therefore retained in the pool, which acts as a filter. A small amount of the radioactivity, predominantly the noble fission gases krypton and xenon, is released immediately into the reactor containment.

What is the most serious accident that could happen at the ILL?

The most serious accident would be a core meltdown following the loss of the water inventory.

It is on the basis of this extremely pessimistic scenario that the Institute's internal emergency plan (PUI) and the off-site emergency response plan (PPI) have been established.

In such a case, the fuel meltdown would result in: 

  • the loss of the fuel cladding, which serves as the first barrier to contain the radioactive fission products;
  • the release of some of the fission products normally stored in the fuel element matrix.

The radioactive fission products would therefore be released into the reactor's primary cooling system.

The worst-case scenario included in the design basis of the ILL reactor involves the extremely pessimistic assumption that the primary cooling system, which is the second barrier to contain radioactive substances, is itself damaged. In this case, the radioactive materials are released directly into the reactor containment - the third and final barrier - and not into the water of the pool surrounding the reactor vessel. This is a particularly unfavourable assumption, as the water in the pool is an extremely efficient filter for retaining a very large proportion of the radioactive fission products.

This worst-case scenario was included in the initial design of the reactor and is taken into account for the improvements made continuously to the safety of the installation. Preventive measures are in place to reduce the probability of its occurrence to as low as is reasonably achievable; damage limitation measures are also in place to reduce to the lowest level reasonably possible the severity of the accident, should it nevertheless occur.

It is on the basis of this extremely pessimistic scenario that the Institute's internal emergency plan (PUI) and the off-site emergency response plan (PPI) have been established. It is this scenario which, despite its improbability, was used to calculate the "300 m danger zone" (evacuation of personnel of companies on the site) and the "500 m danger zone" (“take cover” order for the few hundred people living in the Bastille district of Fontaine).

Would the ILL have to release radioactive gases into the atmosphere if this accident occurred?


In the event of a core meltdown, the reactor containment is immediately isolated. In the hours following the accident, the pressure inside the containment may increase slightly. It stabilises at a maximum of around 0.1 bar, so there is no risk of damage to the containment. However, to guarantee that there are absolutely no unfiltered and unmonitored radioactive releases into the atmosphere, it is preferable to maintain the pressure inside the containment at slightly less than atmospheric pressure. This is done by regularly releasing small quantities of air from the containment via the 45m exhaust stack, through very high efficiency aerosol filters and iodine traps. As these releases are calculated, monitored and controlled, they are known as "planned releases".

In the worst-case scenario taken into account by the ILL, i.e. a core meltdown in air, a fraction of the radioactive fission products is directly released into the containment.

Levels of radioactivity in the containment are continuously monitored by three radiation detectors in the containment itself and by three others in the ventilation system. If readings from any two of three identical detectors exceed a pre-set threshold, the containment is automatically and completely isolated.
However, the total isolation of the containment involves shutting down all the ventilation systems, including the air conditioning, which keeps the air in the containment cool. The pressure in the containment will therefore rise due to the heat given off by various sources inside the containment, in particular the residual power of the fuel. The evaporation of liquid nitrogen and helium used for ILL's research experiments will also lead to a slight increase in the pressure inside the containment.

In the worst possible situation, i.e. on a very hot summer day, the pressure in the containment would rise within a few hours to stabilise at about 0.1 bar.
This explains why the ILL reactor containment is double-walled. It comprises an inner wall made of 40 cm-thick concrete and an outer shell of 11 mm steel. Between the two walls an overpressure of 0.135 bar is constantly maintained, using clean air from the outside. Given the dimensions of the building (60 m wide and 35 m high), there are inevitably some minors leaks. The double-walled containment guarantees that these leaks are from the outside to the inside, and not the other way round. It is clean air from the outside that enters the containment rather than air potentially polluted by radioactive materials that leaves the building.

It is nevertheless preferable to reduce the pressure inside the concrete containment to slightly below atmospheric pressure. This guarantees, this time with absolute certainty, that there are no unmonitored leaks, even if the systems that maintain the overpressure in the annular space between the inner and outer containment walls fail. To achieve the underpressure inside the inner concrete containment, the air to be released is analysed for radioactivity and then directed through the 45m exhaust stack, across very high efficiency filters and iodine traps. The radioactivity released is, of course, measured and monitored before it enters the stack.

What would be the impact of such an accident on Grenoble and the surrounding area?

The impact on people located in the vicinity of a nuclear facility where an accident has occurred is always evaluated in terms of the radiation dose received.
In the ILL’s internal emergency plan (PUI) (which is the responsibility of the nuclear operator) and the off-site emergency response plan (PPI) (which the responsibility of the public authorities and, in particular, the Prefect), there are two danger zones defined around the facility:

  • An inner circle corresponding to the zone that must be evacuated. The reference dose value for defining the radius of this zone is 50 mSv. For the worst-case accident for the ILL reactor, this circle has a radius of 300 m and only concerns employees working for companies in the immediate vicinity of the ILL: the ESRF, EMBL, PSB, LPSC, ST and IBS.
  • An outer circle corresponding to the zone in which “take cover” procedures apply. The reference dose value used to define the radius of this zone is 10 mSv. For the worst-case accident for the ILL reactor, this circle has a radius of 300m. It concerns a small number of the staff at the neighbouring CNRS and CEA sites. The only local people concerned are the 300 residents of the Bastille district of Fontaine on the other side of the river Drac opposite the ILL.

The values used by the ILL to define the danger zones are based on WHO recommendations and have been adopted in most countries.
The dose beyond these zones is not nil, of course (the cloud does not stop at the border), but it decreases with distance. The doses received after one week are given below. They are calculated for a person located in the radioactive plume with no protection (a person outdoors, breathing the air of the plume for a whole week).

  • 3 mSv at a distance of 1 km;
  • 0.9 mSv at a distance of 2 km;
  • 0.15 mSv at a distance of 5 km.

For comparison:

  • The statutory annual dose limit for the general public, excluding natural sources and medical procedures, is 1 mSv;
  • The natural radiation dose received by those living in the Grenoble basin is 2.4 mSv per year;
  • The natural radiation dose received in certain highly populated areas of India or Brazil is 30 mSv per year;
  • The average annual dose received in France for medical purposes is 1.3 mSv, although there are large disparities: an abdominal scan, for example, results in a radiation dose of over 10 mSv.

Basically there are two types of exposure to radiation:

  • External exposure: the radiation source is located outside the body. The radioactive substances which are the source of the radiation decay and emit particles, mainly gammas, which enter into contact with the body. It is relatively easy in this situation to protect oneself, by reducing to a minimum the time spent near the source (TIME), by moving away from the source (DISTANCE), and by placing screens (SHIELDING) between oneself and the source. A house wall, for example, will reduce the dose of radiation received by a factor of 10. For external exposure, the dose rate received per unit of time (expressed in mSv /h) is the relevant value.
  • Internal exposure: the source of radiation is inside the body. This is the case if we inhale contaminated air (by breathing when inside the radioactive plume or “cloud”) or ingest contaminated foodstuffs (drinking water, milk, vegetables, meat, etc ...). The radioactive substances enter the body and are then absorbed by various body organs, depending on their physical and chemical characteristics. The calculations take into account all radionuclides released (see fission products). The organs are irradiated by the particles emitted during decay, only in this case, because the radiation is emitted directly inside the body, the beta and, in particular, alpha particles if any, contribute substantially to the dose received. Obviously, if the source of radiation is inside the body, the simple techniques described above to protect against external exposure can no longer be used. For internal exposure, the amount of radioactivity incorporated (expressed in Becquerel (Bq)) is the relevant value.

In an accident at the ILL reactor, the radiation doses received by those situated within the 300 m zone would primarily be due to external exposure; the source of the radiation would be the reactor building itself. The thickness of the reactor containment would reduce this radiation at its source by a factor of at least 100.

The maximum dose rate would be less than 1 mSv/h at 170 m from the containment, on the edge of the ILL site, and about 0.1 mSv/h at 300 m. These low dose rates would give staff time to take all the necessary action calmly and without haste. This explains why, for example, in such a scenario ILL staff would evacuate the site on foot. This is also the most efficient solution, as it solves the problem of possible traffic jams. 

For those outside the 500 m zone, the dose received would mainly be due to releases of gas from the exhaust stack. With a southerly wind, for example, as is usually the case in the morning, the inhabitants of Grenoble would not receive any dose at all. In the afternoon with a northerly wind, they might receive a fraction of the low doses mentioned above.

How would this accident evolve in kinetic terms?

In an accident involving a core meltdown due to the loss of the water inventory, the kinetics are relatively slow: the number of backup systems available ensures that water levels remain sufficient for the core to continue to be properly cooled by natural convection.

The Basic Safety Rules (RFS - Règles Fondamentales de Sûreté) are based on the assumption that certain emergency backup systems fail after 24 hours of use. It is this additional failure that, in this scenario, would result in the exposure of the core and its fusion in air.
The operator and the authorities would therefore have sufficient time to implement their respective emergency response plans (PUI and PPI) before the radiation accident itself (i.e. the core meltdown) occurs.

In the Borax-type accident automatically included in the design basis the kinetics are extremely fast. A small fraction of the fission products is released immediately into the reactor containment; most of the radioactivity, however, is retained in the reactor pool. Although this type of accident comes under the so-called "reflex response phase" of the PPI, the dose rates generated around the installation would be considerably lower than those resulting from a core meltdown in air. Any measures that might need to be taken (evacuation and / or “take cover” order) to protect the populations in the 300 m and 500 m zones could therefore be carried out calmly and carefully, without haste.

Is it necessary to take iodine tablets if there is an accident at the ILL’s reactor?

Yes and no

Taking into account the nominal efficiency of the iodine traps, calculations made for the worst-case accident scenario (i.e. a core meltdown) show that the equivalent dose in the thyroid gland is about 10mSv for children (the most vulnerable group) within the 500 m zone.
In France the decree of 20 November 2009 fixes the thyroid equivalent dose above which stable iodine must be administered at 50 mSv. The administration of iodine tablets would therefore not be compulsory in our case.
The Prefect may, however, order tablets to be distributed to people in the 300 m and 500 m zones as an added precaution.

It may be useful to note that:

  • Iodine tablets saturate the thyroid gland with the (non-radioactive) stable iodine they contain. If the thyroid is saturated, it cannot absorb any radioactive iodine inhaled.
  • As the protection afforded lasts about 24 hours, there is no point taking the tablet too early. This is why it is important to listen to the instructions issued by the Prefect.

Find out more:

Is the spent fuel in the storage pool as well protected as the reactor core?


The pool used to store spent fuel awaiting reprocessing at La Hague is situated, like the reactor pool itself, inside the reactor containment; its design constraints (resistance to various stresses and possible damage) are exactly the same as those for the reactor pool.

As of the 242nd day after reactor shutdown, a spent fuel element can be cooled properly using only air. As the ILL operates a maximum of four 50-day cycles per year, it therefore produces four spent elements (cores) per year, of which no more than three can have a cooling period (the time that has elapsed since they stopped being used) of less than 242 days. The other fuel elements in interim storage in the pool have all been cooling for over 242 days and can therefore perfectly withstand dry conditions without being damaged or releasing their remaining radioactivity.

The consequences of a total loss of water in the storage pool, an event classed as a “residual” risk, are lower than those of the worst-case accident scenario included in the reactor design basis, namely "a core meltdown in air".

What are the dangers of ionising radiation?

Ionising radiation has two main types of effect:

1. Deterministic effects

These are effects that occur at high doses, typically 1000 mSv and above, and are observed in anyone who receives such a dose. They have three basic characteristics:

  • They are threshold effects. This is fundamental for radiation protection purposes, since it confirms that a dose below this threshold has absolutely no impact on health.
  • The severity of the effect depends on the dose. The higher the dose, the more rapid and acute the clinical effects.
  • They are early effects. Except in special cases such as cataracts, the effects appear rapidly, in the hours or days following exposure.

In the case of whole body exposure to radiation, the severity of the symptoms (fatigue, nausea, vomiting) increases with the dose received. At levels exceeding 2000 mSv, there is a drop in lymphocyte, red blood cell and platelet counts, which explains the haemorrhaging and infections observed.

2. Stochastic effects

These are effects that occur randomly, irrespective of the dose received, to some of the individuals in a population exposed equally to the same dose of radiation. They mainly appear in the form of so-called “radiation-induced” cancers, which in fact are absolutely identical to the various forms of cancer observed throughout the human population when there has been no exposure to anything other than natural radiation. These stochastic effects are exactly the opposite of deterministic effects:

  • They are effects that have no threshold. This assumption, which is applied on the basis of the precautionary principle, is again fundamental for radiation protection purposes, since it implies that any dose, however low it may be, leads to a slight increase in the probability of an individual developing cancer and therefore to a slight increase in the frequency of occurrence of cancers in a uniformly exposed population.
  • The severity of the effect does not depend on the dose. The dose received does not determine the severity of the cancer for any given part of the body.
  • They are late effects. The first cancers to appear are leukaemias, typically 5 years after exposure. “Solid” cancers can take as long as several decades to appear.

These stochastic effects have been demonstrated in humans through epidemiological studies at exposure levels of over 100 mSv. The benchmark study establishing the relation between dose and effect is still the one performed on the populations exposed at Hiroshima and Nagasaki. More than 86000 individuals were or are still being studied (for those who are still alive 66 years after the explosions). These studies have not identified any significant rise in the cancer rate among subgroups with an exposure level of less than 100 mSv. This does not mean there is no effect below this dose, but simply that if such effects exist, there are too few excess cases to be statistically significant. This difficulty, and it is not the only one, is inherent in any epidemiological study, regardless of the parameter being tested. The International Commission on Radiological Protection (ICRP), whose recommendations provide the basis for all radiation protection doctrine and the resulting regulations, has therefore decided to err on the side of caution and extrapolate these dose-effect relations to low doses as a straight-line relationship without a threshold. The risk coefficient for any individual in the population is currently estimated at 0.07 per Sv received. An individual exposed to 10 mSv will therefore see his/her risk of developing a cancer increase by 0.0007 (bearing in mind that the “natural baseline” is 0.25, since one person in four develops cancer during his/her life). Finally, it should be noted that this so-called no-threshold hypothesis is currently the subject of considerable research, particularly in biology.

Many scientists believe that the hypothesis lacks credibility in molecular biology terms, given the complexity of the mechanisms involved in the cancerisation of tissue.

See also the website of the IRSN (Institut de Radioprotection et de Sûreté Nucléaire)