In 1975, a nuclear reactor at the Leningrad Nuclear Power Plant in the Soviet Union was undergoing maintenance. Leningrad Unit 1 was the first of a new generation of Soviet nuclear power plants to be built - the first prototype of a class of reactor known as RBMK.
The reactor's operators were attempting to restart the reactor from a low power level. To their surprise, they found the reactor would not restart - not without the operators withdrawing nearly all of the control rods that regulate the reactor.
As power slowly climbed, more and more control rods were removed from the reactor. The operators lost the ability to properly regulate and distribute power throughout the reactor. Some sections of the reactor core were barely ticking over. Others were rapidly overheating.
Deep in the core, a single channel filled with fuel and water overheated and ruptured, flooding the reactor vessel with radioactive steam. Water inside the reactor boiled rapidly, more and more of it turning to steam. Power in the reactor accelerated with each passing second, boiling more water, making more steam.. Alarms blared in the control room.
There is a button on the reactor operator's control panel, hidden behind a wax-sealed guard labelled AZ-5. AZ-5, in this case, being translated as Emergency Protection System 5. It is one of a number of emergency power reduction and shutdown modes available to the reactor engineer. This button immediately forces all control rods to be inserted into the reactor at once to shut the reactor down as fast as possible.
The operators at Leningrad push the button. It has to be held in place or the control rods will stop. For tense moments, power continues to climb out of control, threatening catastrophe.
For several long seconds, the reactor doesn't do what it's supposed to. It doesn't stop.
Eventually, to everyone's relief, the control rods begin to take effect. The reactor finally grinds to a halt.
Radioactive steam is vented from the plant. The core is saved. There is no Leningrad disaster.
No information about this accident is shared. Instead, it is classified as a State secret, known only to the reactor's designers and those who happened to be in the control room at the time. Meanwhile, Leningrad Unit 1 will be repaired, and continue in service until 2018.
At this time, Reactor 1 at Chernobyl Nuclear Power Plant is still 2 years from completion. The foundations for the largest nuclear accident in history, have just been poured.
--------
How does an RBMK reactor explode?
This is one possible narrative. One which assumes there are no villains in Chernobyl. There are just people who went to work one evening, working as the system required of them, and the reactor, working as physics required of it.
The result is catastrophe.
---
To figure out why an RBMK reactor can explode, first we have to understand how an RBMK reactor works. It's not actually that complicated. A uranium atom splits - small particles called neutrons fly out from the atomic shrapnel and they each find another uranium atom. When they collide with it, they break it apart too, letting more neutrons find more Uranium atoms to break in a chain reaction. Out of each split we get energy. Once split, the fragments snap apart as the energy holding them together is released - like cutting an elastic band. These fragments explode away, run into the atoms next to them at high speed. Much like the brakes in your car, that speed is turned into heat as the fragments stop. This Heat is used to boil water to steam. This steam turns a turbine. This turbine turns an electric generator and, eventually you get electric power out of it.
Now, it's not quite that simple. Big, heavy atoms like Uranium come in multiple different versions - called Isotopes - sort of like different models of the same car. They're all Uranium, but they're all slightly different at the same time. The most well known, are Uranium 235, and Uranium 238. U238 is by far and way the most common - on the order of 99% or more of Uranium on Earth is U238. It's a big, heavy fat atom that doesn't really like to split - it takes a very fast neutron to break it apart - and the neutrons released by U238 fission aren't fast enough in turn to cause a further fission.
On the other hand, U235 is much happier to break apart - taking a lot less energy to do so. Unfortunately, it's about as rare as common sense. Out of a large block of Uranium, only a small amount is actually useful in a reactor.
Out of the ground, less than 1% of your Uranium fuel is actually fuel.
More than that, in order to have a chain reaction, the neutrons released by one atom splitting have to be able to cause further atoms to split. Otherwise everything just runs down. It turns out, that the neutrons released by fission are extremely fast - too fast to split U235 but still not fast enough to split U238. For a fast neutron, the probability that it will cause another fission is really low.To keep the chain reaction going, each fission event needs to cause at least one further fission. Either you need a lot more U235 around it to get the probability of another fission up, or you need to find someway of slowing it down. If you slow it down - the probability of fission goes way up. This is done by a Moderator.
For most reactors, this moderator is water. Ordinary - albeit extremely pure - water. Water is a good moderator, but it has one slight drawback - it absorbs neutrons. Absorbed neutrons do not get to go and make another fission happen - they just turn the water radioactive. Using water to moderate a reactor absorbs so many neutrons, that the quantity of U235 in the reactor fuel has to be increased beyond that which is in natural Uranium. This process is called enrichment. It's expensive and energy intensive - and it turns out if you enrich Uranium to about 80% U235 it can be used to make an atomic bomb.
Which, naturally, is why Israel got so pissy about Iran having the ability to enrich Uranium. The difference between safe reactor fuel and weapons-grade bomb fuel, is time in the centrifugal oven to bake.
Now, doesn't it seem eminently sensible to find a way to build a reactor that will be happy on regular, non-enriched Uranium?
Canada did it with the CANDU reactor. Instead of regular water, a CANDU reactor uses 'heavy' water. 'Heavy' water is like ordinary 'Light' water, except the Hydrogen atom that's the H in H2O is a little different. It has one neutron and one proton, rather than just a single proton. It absorbs less neutrons, which means more neutrons are free to cause more fission. In fact, Heavy Water is so effective that CANDU reactor doesn't need enriched fuel. The Nazis appearred to have tried a similar appoach, and the destruction of their Norwegian Heavy Water factories stalled their weapons program. On the other hand, heavy water - while common - is still fairly expensive.
It is a case of swings and roundabouts. Either suffer the increased fuel costs, and international oversights of enriched Uranium, or suck up the costs of seperating heavy water from seawater. Both are expensive.
What if you could build a reactor that ran on natural uranium, that didn't need heavy water?
The very first self-sustaining nuclear reactor - Chicago Pile 1 - used blocks of Graphite as a moderator. It ran at so low a power, it didn't require cooling.
The first Hanford Reactors also used graphite. They also used regular, light water as a coolant to keep the reactor from melting down with its own heat. This hot water was dumped merrily into the local river along with whatever contamination it picked up along the way. Of course, somebody had worked out that if the reactor lost cooling water, it would very quickly begin to run out of control so the Hanford reactors were built miles from anywhere inhabited. They never generated a watt of electricity- but they did create the Plutonium for your nuclear weapons.
The British Government, aware of this risk of a runaway reaction, built the Windscale Piles to be Air cooled - with giant fans blowing air over hot graphite and metal. These then went and caught fire. In the end, the solution was to use graphite and an inert gas, like carbon dioxide, to cool the core. This was still extremely expensive, an inefficient due to the limitations of the materials used.
The Soviet Union looked at this and thought; We can build a graphite moderated reactor, cool it with regular light water and so long as we don't fuck up, we'll have a shitton of free energy.
We know the result already. But that's being a little bit churlish. These people weren't fools.
This why an RBMK reactor is different to every other reactor built anywhere else in the world. It's in the name - Reaktor Bolshoy Moshchnosti Kanalnyy - Big Powerful Channel Reactor. The majority of modern nuclear reactors are basically big pressurised kettles, filled mostly with pressurised water. This water either boils in the kettle- or it is under such a high pressure that it remains liquid and is used to boil water in a second circuit. In an RBMK reactor, the fuel is contained in more than 1600 vertical channels cut through the graphite moderator. Inside these channels, light water flows as coolant. It enters a pair of drums high above the reactor where any steam bubbles in the water seperate and are drawn off to the turbine to generate power - while the liquid water is recirculated, being mixed with cooler water coming back from the turbine. Interspersed within these channels are more than 200 others - each contaning a control rod. These control rods are also cooled by liquid water - but at a much lower temperature.
The main steam circuit on an RBMK reactor operates at somewhere around 270 degrees and 60-odd Bar of pressure. The control rod circuit operated at 70 degrees.
The light water in the channels still absorbs neutrons sure - but because there's so much less of it, the reactor will still run on natural unenriched Uranium. It also means that, since each fuel assembly has its own individual channel, it can be removed, moved and replaced without shutting down the reactor. This is a feature few other reactors have - most reguire a shutdown to open the reactor vessels to refuel. This is good for fuel economy, and good for efficiency. A neutron reflector surrounding the core reflected neutrons that would've otherwise escaped, back into the core where they could go on to cause more fission. Finally, an RBMK reactor did not require high quality materials, or the specialised construction methods used to construct the large pressure vessels used to hold conventional reactor cores. An RBMK reactor could be built and maintained reliably with less-skilled labour.
The people who designed the Chernobyl reactor weren't fools. There were compelling reasons for making the decisions they did. It made a big, powerful reactor cheap and easy to build, while improving the reactor's fuel economy and general uptime. And you could potentially fuel a weapons program with it.
The one clear drawback with this should be obvious. When the cooling water boils, its replaced by a bubble of steam. This steam absorbs far fewer neutrons than liquid water - meaning more neutrons become available for fission, which means more fission, more heat, more steam, more neutrons, and more fission.
This is called a Positive Void Coefficient. It is an example of positive feedback - where an action creates a stronger action in the same direction. Positive feedback is like setting a ball rolling at the top of a hill, it's only going to start rolling faster as it gets further down. Engineers love positive feedback -it usually has entertaining results.
This is potentially a problem for an RBMK reactor specifically because the water does not act as a moderator - more correctly, it provides little to no moderation. In a conventional reactor, the water also provides moderation. If water is boiled away by heat, the moderation in the reactor reduces, neutrons get faster, the probability of fissions gets lower, less fission happens, heat decreases and the problem self-corrects. In an RBMK reactor - even if all the water in the core is somehow removed, the moderator is still present in the form of the graphite to keep the reaction going.
It would be dangerous to have a reactor which behaved like this. The engineers who designed Chernobyl were, of course, aware of this. Real physics is not that simple. As the fuel heats up and gets more energetic, it responds to neutrons differently. The hotter it gets, the harder it is for a neutron to cause a fission. Hotter fuel is less likely to fission - so an increase in power will actually reduce the ability of the fuel to fission and create power - in effect an automatic brake provided free by simple physics. This is called Negative feedback, and is basically the same as you feeling a tug in the steering wheel of your car, and steering the other direction to compensate.
Positive feedback acts to destabilise. Negative feedback acts to stabilise.
If the negative feedback from the fuel heating is stronger than the positive feedback from the steam boiling, the reactor's power level will self stabilise and everything will be fine.
For a large part of the reactor's life this was true. Even if the cooling water boiled off, the reactor would still self-correct as the increase in temperature in the fuel would have a stronger effect.
This changed as the reactor got older.
In general, the higher the proportion of neutrons that are absorbed by the water in the reactor in comparison to all other neutron absorbers in the reactor, the stronger the void coefficient became. Where a reactor has been running for several years- the fuel gets more and more depleted of fissionable attom. In addition, more and more reactor poisons are added, each of which absorbs neutrons differently or introduces additional hazards into the reactor. To compensate for the reduced reactivity of the older fuel, fixed neutron absorbers were removed from the reactor. With less absorbers in the reactor, a higher proportion of neutrons were lost to absorption in the reactor coolant.
At the time of the accident, the reactor in Chernobyl had been running for about three years. After this amount of time, the void coefficient had become so strong, that the negative feedback from the fuel heating was no longer strong enough to counterbalance the effect of the void coefficient.
By the time of the accident, under certain conditions, the reactor operated in a positive feedback loop.
An increase in power, left unchecked would create a further increase in power. Only the reactor's control rods then kept the reactor under control, The majority of these control rods inserted from the top of the reactor. Some inserted from the bottom. They served to absorb any excess neutrons in the core and act as the final brake on the reactor to keep it in control, to keep the reactor critical.
Fission reactors are at their happiest when they're critical. A critical reactor is a reactor running in a balanced steady state at a constant power. It's the desired state of being. Every fission is creating one further fission and that's it. A reactor that is supercritical, is a reactor that's accelerating - each fission creates more than one further fission. A reactor that is subcritical, is a reactor that's decelerating - each fission creates less than one further fission.
The reactor is moved from state to state by adding or removing reactivity. Reactivity is like the throttle and brake on the reactor. It's not really the current power level - it's close to the potential change in power level. Positive reactivty means fission is more likely to happen than it is now - which will cause an increase in power. Negative reactivity, means making fission less likely to happen - which will cause a decrease in power.
In theory, there is no limit to the amount of negative reactivity you can add - all it does is stop the reactor faster. There is a limit to the positive reactivity.
When an atom fissions, the vast majority of neutrons are released instantaneously - at the moment of fission. The neutrons fly away, get themselves moderated, and in the space of microseconds find more atoms to collide with and split. The scientists of the Manhattan project called this a 'Shake' and it is an extremely short interval of time - from nanoseconds to microseconds. These are called Prompt neutrons.
Fission with prompt neutrons happens so quickly, that there is little to no mechanical process capable of controlling and regulating it. If the universe had been created in such a way that there were only prompt neutrons - controllable fission power would likely be impossible.
Had this been the case, the Chicago Pile 1 experiment could have had a far more amusing result.
Luckily for the citizens of Chicago, a very small fraction of the neutrons released by a fission event are delayed - they happen seconds, to minutes later. It is this miniscule fraction of delayed neutrons which enables every nuclear fission reactor to be controlled. It is possible therefore, to have a reactor which is critical on the combination of the Prompt, and the Delayed neutrons. This is how things normally are. Even a supercritical reactor will take seconds to minutes, to change power output. There's time there for the process machinery of the reactor to respond to changes and stabilize.
However, if the reactor is pushed to the point that it is capable of achieving criticality on the prompt neutrons alone - before any delayed neutrons are emitted - things get interesting. Instead of a power increase that happens on the order of seconds to minutes - now the only limiting factor the the reactor's power increase is the time it takes for one neutron to find the next atom to fission and however long the reactor manages to withstand the energies that are being very rapidly liberated. A reactor which has gone prompt critical, has become, in effect, a really, really shit nuclear bomb. The big difference being that bombs take advantage of physics, inertia and a dozen other things to keep the reaction going that few nanoseconds longer it takes to go from 5 tons of TNT, to 15 Kilotons of TNT.
Scientists at the Manhattan project, for whatever reason, called this interval a 'Dollar' of reactivity. Once you get a reactor past that point - unless it's a type specifically designed to go there and self recover - the reactor will be destroyed. Importantly, this does not have to happen within the entire reactor - it can be limited to a very small part of the core where conditions align like the stars over R'lyeh.
At the Chernobyl reactor, Reactivity was added by fresh fuel, by removing control rods and by boiling water to make steam. Reactivity was removed by increasing water flow, adding control rods, heating up the moderator and fuel, and by another factor.
The shrapnel left over from fission creates what's known as 'fission products'. Most of these are hideously radioactive. Many of these are effective at absorbing neutrons, and hence are called poisons. Absorbed neutrons reduce reactivity, which has to be compensated for either by withdrawing control rods, or by removing the used fuel and replacing it with fresh fuel. One of the most effective neutron absorbers is an isotope of Xenon, called Xenon-135.
A product of radioactive decay, it starts to appear about six hours after the fission events that effectively 'created' it. The amount of it that's created, is in direct proportion to the quantity of fission that happened six hours ago. So if a reactor is run at full power for a long time, and then throttled down, Xenon will continue to appear according to that fuel burned six hours previously. It's a bit like the exhaust from your car's engine magically taking longer to form after the combustion in the cylinders. Normally, with the reactor in a steady-state, Xenon is created as quickly as it is consumed - the physics balances out. It can make it very difficult to increase or reduce power - if power is reduced too quickly, and the Xenon continues to build, the reactor might even be stalled by it.
It can also mean that, if the fuel in the reactor has been active for a long time - there may not be sufficient reactivity in the remaining fuel to overcome this Xenon pit. The reactor will stalled and effectively impossible to start, no matter how many control rods are withdrawn.
This is important. After a few more hours, the Xenon goes away. More than that, Xenon which absorbs a neutron also 'goes away' - it's no longer Xenon-135 and it's massive ability ot hoover up neutrons is suddenly gone.
Keeping all of these positive and negative reactivities in balance is the job of the Senior Reactor Engineer, who manipulates the reactor core's systems and control rods to achieve the required stable power output. The Engineer at the reactor's controls has only so much control as the rods will give them.
The Control rods of an RBMK reactor are manufactured from boron. Boron absorbs neutrons, which reduces reactivity, and causes a power drop. The deeper they go into the reactor - the more neutrons are absorbed - the further reactivity decreases. They can also be moved independently of each other - which changes where and how power is produced throughout the reactor, to compensate and balance for old and fresh fuels and how they're distributed through the reactor.
But, at the tip of the control rod on a telescoping extension, is a single slug of graphite. The graphite tip of the control rod acted as a displacer. Its purpose was to push water out of the control rod channel, to remove it and its neutron absorption effect after the rod was withdrawn. In effect, instead of giving the control rod an action of -3,-1 - they are something like -3,+0. They graphite displacer gives the control rod a stronger control action. It makes it more powerful by adding reactivity after the rod is withdrawn. There was about 1.5 metres of water between the control rod proper, and the displacer rod.
This is the first generation RBMK reactor, as built at Leningrad Unit 1.
Shortly after these first reactors were completed, computer models had begun to show some unexpected tendancies in the reactor core. When running at low power, the reactor began to behave strangely. It became less and less stable, and harder to control. Attempting to balance power in the bottom part of the reactor by adjusting the control rod positions could cause momentary power-surges and hot spots - especially towards the edge of the reactor close to the neutron reflector. This could potentially damage the reactor, damaging fuel elements, or potentially rupturing the fuel channel
In one possible worst case scenario, the fuel would superheat to the point where it began to melt, triggering a chemical reaction between the molten fuel, the zirconium cladding of the fuel, and the now suddenly superheated steam in the channel. Hydrogen gas would be released, along with quantities of oxygen - potentially leading to explosive results.
Finally, there is the concept of the Reactivity Margin. This is an estimate of the amount of reactivity the reactor operator can still add to the reactor to increase the power or move power around the reactor. In an RBMK, the reactivity margin was measured in numbers of control rods, where One Control Rod of reactivity, was the equivelant reactivity introduced by removing one average control rod fully from the reactor. A reactivity margin of Thirty Control Rods did not necessarily mean there were Thirty control rods inserted fully into the core - this would be spread out amongst the different rods, each inserted at a different height in the reactor to properly balance power.
In general, the higher the reactivity margin, the further the control rods are inserted in the core and the more reactivity can be added to the core by removing them . A reactor with fresh fuel will have a very high reactivity margin. A reactor with old fuel, or with xenon poisoning, will have a low reactivity margin. It may seem that a low reactivity margin might be 'safer', because now there's less reactivity that can be added by the control rods (which are already out of the reactor at low margin). At a high reactivity margin, more control rods are inside the reactor. More of them can be withdrawn, to push the reactor further into the supercritical - more positive reactivity can be added to the reactor.
The Chernobyl reactor was happy around about 30 Rods of reactivity. The equivalent of thirty full control rods were left in the reactor. The official limit, was somewhere around 15 Rods of reactivity.
This was not considered a safety limit. Running at low reactivity margins could be a mark of skill of the reactor operator. A low reactivity margin meant less scope for moving power around the reactor and keeping things in balance. It increased the risk that sections of the core could be overloaded.
It was never thought of as a safety limit. In November 1975, Leningrad Unit-1 an accident occurred. A fuel channel rupture while operators were trying to increase power after a shutdown. A similar incident happened at Chernobyl Reactor 1. In both cases, the reactor's safety systems operated correctly. Radiation was still released through the plant's ventilation systems. Neither incident was reported in the media.
Reactor 4 at Chernobyl is slightly different. Reactor 4 was a second generation reactor. Over the years, a few improvements to economise the reactor and try to stabilise the reactor's void coefficient are made. There're changes in fuel enrichment, control rod design, and rod configuration. Nothing seemed major. As part of these changes, the control rod tips were shortened by approximately a meter.
These modifications led to an unsettling side-effect when tested in practice. The control rods, when fully withdrawn, allowed a 1.5 metre high column of water to remain in the bottom of the reactor. This water would absorb neutrons in the bottom of the reactor, acting as a gentle brake on fission.
When a newly-built RBMK reactor at Ignalina in Lithuania was being run through its commissioning tests in 1984, it was discovered that, where control rods had been pulled all of the way out of the reactor and where most of the fission in the reactor was happening at the very bottom of the core - an attempt to slow the reactor down by inserting the control rods could cause a momentary increase in power before the boron control rods travelled the full height of the core to quence the reactor.
The water slugs at the bottom of the control rod channels were replaced with graphite for a few seconds. The graphite moderated neutrons, where the water had absorbed them. Consequently, an increase in reactivity occured at the bottom of the core.
Instead of decreasing, power momentarily increased.
The engineers at Ignalina wrote a letter to the reactor's designers, advising them of the deviation. This was not initially thought to be much of a concern - power changes in the reactor after all, take longer than it takes for the rod to travel.
A similar effect was noted at an RBMK in Smolensk, at Chernobyl reactor 3, and finally, at Chernobyl reactor 4 when it was completed.
It gradually became clear to the designers that accidents with the RBMK were not only possible - but even likely in some circumstances. Still - it would reflect badly on the designers and the ministry they worked for if the reactor they had overseen was found to have a potential flaw, especially after millions of rubles had been spent on building them. Careers, influence and bureaucratic prestige were at stake. It was quietly buried as a footnote in the reactor documentation. Not as a threat, or a risk - just as a deviation of the reactor from its design specifications. This effect was something that happened from time-to-time in very specific circumstances, but one which had never really caused a problem.
Even when the Soviet government found out about the cover-ups, no overt action was taken, no matter how furious they were in private. The world at large couldn't know. Even so, nature could only be be fooled for so long. A series of modifications were proposed - changes to the sequencing of the control rods especially - which should've reduced the probability of this 'tip-effect' causing the reactor to run up. The plan was to quietly fix the reactors, saving face before the world. These fixes were to be implemented at Chernobyl Reactor No. 4 at it's next scheduled maintenance shutdown. Before that, however, one further test of the reactor's safety systems was required.
We do not yet know how an RBMK reactor explodes. But we know what we need to know.
The reactor's operators were attempting to restart the reactor from a low power level. To their surprise, they found the reactor would not restart - not without the operators withdrawing nearly all of the control rods that regulate the reactor.
As power slowly climbed, more and more control rods were removed from the reactor. The operators lost the ability to properly regulate and distribute power throughout the reactor. Some sections of the reactor core were barely ticking over. Others were rapidly overheating.
Deep in the core, a single channel filled with fuel and water overheated and ruptured, flooding the reactor vessel with radioactive steam. Water inside the reactor boiled rapidly, more and more of it turning to steam. Power in the reactor accelerated with each passing second, boiling more water, making more steam.. Alarms blared in the control room.
There is a button on the reactor operator's control panel, hidden behind a wax-sealed guard labelled AZ-5. AZ-5, in this case, being translated as Emergency Protection System 5. It is one of a number of emergency power reduction and shutdown modes available to the reactor engineer. This button immediately forces all control rods to be inserted into the reactor at once to shut the reactor down as fast as possible.
The operators at Leningrad push the button. It has to be held in place or the control rods will stop. For tense moments, power continues to climb out of control, threatening catastrophe.
For several long seconds, the reactor doesn't do what it's supposed to. It doesn't stop.
Eventually, to everyone's relief, the control rods begin to take effect. The reactor finally grinds to a halt.
Radioactive steam is vented from the plant. The core is saved. There is no Leningrad disaster.
No information about this accident is shared. Instead, it is classified as a State secret, known only to the reactor's designers and those who happened to be in the control room at the time. Meanwhile, Leningrad Unit 1 will be repaired, and continue in service until 2018.
At this time, Reactor 1 at Chernobyl Nuclear Power Plant is still 2 years from completion. The foundations for the largest nuclear accident in history, have just been poured.
--------
How does an RBMK reactor explode?
This is one possible narrative. One which assumes there are no villains in Chernobyl. There are just people who went to work one evening, working as the system required of them, and the reactor, working as physics required of it.
The result is catastrophe.
---
To figure out why an RBMK reactor can explode, first we have to understand how an RBMK reactor works. It's not actually that complicated. A uranium atom splits - small particles called neutrons fly out from the atomic shrapnel and they each find another uranium atom. When they collide with it, they break it apart too, letting more neutrons find more Uranium atoms to break in a chain reaction. Out of each split we get energy. Once split, the fragments snap apart as the energy holding them together is released - like cutting an elastic band. These fragments explode away, run into the atoms next to them at high speed. Much like the brakes in your car, that speed is turned into heat as the fragments stop. This Heat is used to boil water to steam. This steam turns a turbine. This turbine turns an electric generator and, eventually you get electric power out of it.
Now, it's not quite that simple. Big, heavy atoms like Uranium come in multiple different versions - called Isotopes - sort of like different models of the same car. They're all Uranium, but they're all slightly different at the same time. The most well known, are Uranium 235, and Uranium 238. U238 is by far and way the most common - on the order of 99% or more of Uranium on Earth is U238. It's a big, heavy fat atom that doesn't really like to split - it takes a very fast neutron to break it apart - and the neutrons released by U238 fission aren't fast enough in turn to cause a further fission.
On the other hand, U235 is much happier to break apart - taking a lot less energy to do so. Unfortunately, it's about as rare as common sense. Out of a large block of Uranium, only a small amount is actually useful in a reactor.
Out of the ground, less than 1% of your Uranium fuel is actually fuel.
More than that, in order to have a chain reaction, the neutrons released by one atom splitting have to be able to cause further atoms to split. Otherwise everything just runs down. It turns out, that the neutrons released by fission are extremely fast - too fast to split U235 but still not fast enough to split U238. For a fast neutron, the probability that it will cause another fission is really low.To keep the chain reaction going, each fission event needs to cause at least one further fission. Either you need a lot more U235 around it to get the probability of another fission up, or you need to find someway of slowing it down. If you slow it down - the probability of fission goes way up. This is done by a Moderator.
For most reactors, this moderator is water. Ordinary - albeit extremely pure - water. Water is a good moderator, but it has one slight drawback - it absorbs neutrons. Absorbed neutrons do not get to go and make another fission happen - they just turn the water radioactive. Using water to moderate a reactor absorbs so many neutrons, that the quantity of U235 in the reactor fuel has to be increased beyond that which is in natural Uranium. This process is called enrichment. It's expensive and energy intensive - and it turns out if you enrich Uranium to about 80% U235 it can be used to make an atomic bomb.
Which, naturally, is why Israel got so pissy about Iran having the ability to enrich Uranium. The difference between safe reactor fuel and weapons-grade bomb fuel, is time in the centrifugal oven to bake.
Now, doesn't it seem eminently sensible to find a way to build a reactor that will be happy on regular, non-enriched Uranium?
Canada did it with the CANDU reactor. Instead of regular water, a CANDU reactor uses 'heavy' water. 'Heavy' water is like ordinary 'Light' water, except the Hydrogen atom that's the H in H2O is a little different. It has one neutron and one proton, rather than just a single proton. It absorbs less neutrons, which means more neutrons are free to cause more fission. In fact, Heavy Water is so effective that CANDU reactor doesn't need enriched fuel. The Nazis appearred to have tried a similar appoach, and the destruction of their Norwegian Heavy Water factories stalled their weapons program. On the other hand, heavy water - while common - is still fairly expensive.
It is a case of swings and roundabouts. Either suffer the increased fuel costs, and international oversights of enriched Uranium, or suck up the costs of seperating heavy water from seawater. Both are expensive.
What if you could build a reactor that ran on natural uranium, that didn't need heavy water?
The very first self-sustaining nuclear reactor - Chicago Pile 1 - used blocks of Graphite as a moderator. It ran at so low a power, it didn't require cooling.
The first Hanford Reactors also used graphite. They also used regular, light water as a coolant to keep the reactor from melting down with its own heat. This hot water was dumped merrily into the local river along with whatever contamination it picked up along the way. Of course, somebody had worked out that if the reactor lost cooling water, it would very quickly begin to run out of control so the Hanford reactors were built miles from anywhere inhabited. They never generated a watt of electricity- but they did create the Plutonium for your nuclear weapons.
The British Government, aware of this risk of a runaway reaction, built the Windscale Piles to be Air cooled - with giant fans blowing air over hot graphite and metal. These then went and caught fire. In the end, the solution was to use graphite and an inert gas, like carbon dioxide, to cool the core. This was still extremely expensive, an inefficient due to the limitations of the materials used.
The Soviet Union looked at this and thought; We can build a graphite moderated reactor, cool it with regular light water and so long as we don't fuck up, we'll have a shitton of free energy.
We know the result already. But that's being a little bit churlish. These people weren't fools.
This why an RBMK reactor is different to every other reactor built anywhere else in the world. It's in the name - Reaktor Bolshoy Moshchnosti Kanalnyy - Big Powerful Channel Reactor. The majority of modern nuclear reactors are basically big pressurised kettles, filled mostly with pressurised water. This water either boils in the kettle- or it is under such a high pressure that it remains liquid and is used to boil water in a second circuit. In an RBMK reactor, the fuel is contained in more than 1600 vertical channels cut through the graphite moderator. Inside these channels, light water flows as coolant. It enters a pair of drums high above the reactor where any steam bubbles in the water seperate and are drawn off to the turbine to generate power - while the liquid water is recirculated, being mixed with cooler water coming back from the turbine. Interspersed within these channels are more than 200 others - each contaning a control rod. These control rods are also cooled by liquid water - but at a much lower temperature.
The main steam circuit on an RBMK reactor operates at somewhere around 270 degrees and 60-odd Bar of pressure. The control rod circuit operated at 70 degrees.
The light water in the channels still absorbs neutrons sure - but because there's so much less of it, the reactor will still run on natural unenriched Uranium. It also means that, since each fuel assembly has its own individual channel, it can be removed, moved and replaced without shutting down the reactor. This is a feature few other reactors have - most reguire a shutdown to open the reactor vessels to refuel. This is good for fuel economy, and good for efficiency. A neutron reflector surrounding the core reflected neutrons that would've otherwise escaped, back into the core where they could go on to cause more fission. Finally, an RBMK reactor did not require high quality materials, or the specialised construction methods used to construct the large pressure vessels used to hold conventional reactor cores. An RBMK reactor could be built and maintained reliably with less-skilled labour.
The people who designed the Chernobyl reactor weren't fools. There were compelling reasons for making the decisions they did. It made a big, powerful reactor cheap and easy to build, while improving the reactor's fuel economy and general uptime. And you could potentially fuel a weapons program with it.
The one clear drawback with this should be obvious. When the cooling water boils, its replaced by a bubble of steam. This steam absorbs far fewer neutrons than liquid water - meaning more neutrons become available for fission, which means more fission, more heat, more steam, more neutrons, and more fission.
This is called a Positive Void Coefficient. It is an example of positive feedback - where an action creates a stronger action in the same direction. Positive feedback is like setting a ball rolling at the top of a hill, it's only going to start rolling faster as it gets further down. Engineers love positive feedback -it usually has entertaining results.
This is potentially a problem for an RBMK reactor specifically because the water does not act as a moderator - more correctly, it provides little to no moderation. In a conventional reactor, the water also provides moderation. If water is boiled away by heat, the moderation in the reactor reduces, neutrons get faster, the probability of fissions gets lower, less fission happens, heat decreases and the problem self-corrects. In an RBMK reactor - even if all the water in the core is somehow removed, the moderator is still present in the form of the graphite to keep the reaction going.
It would be dangerous to have a reactor which behaved like this. The engineers who designed Chernobyl were, of course, aware of this. Real physics is not that simple. As the fuel heats up and gets more energetic, it responds to neutrons differently. The hotter it gets, the harder it is for a neutron to cause a fission. Hotter fuel is less likely to fission - so an increase in power will actually reduce the ability of the fuel to fission and create power - in effect an automatic brake provided free by simple physics. This is called Negative feedback, and is basically the same as you feeling a tug in the steering wheel of your car, and steering the other direction to compensate.
Positive feedback acts to destabilise. Negative feedback acts to stabilise.
If the negative feedback from the fuel heating is stronger than the positive feedback from the steam boiling, the reactor's power level will self stabilise and everything will be fine.
For a large part of the reactor's life this was true. Even if the cooling water boiled off, the reactor would still self-correct as the increase in temperature in the fuel would have a stronger effect.
This changed as the reactor got older.
In general, the higher the proportion of neutrons that are absorbed by the water in the reactor in comparison to all other neutron absorbers in the reactor, the stronger the void coefficient became. Where a reactor has been running for several years- the fuel gets more and more depleted of fissionable attom. In addition, more and more reactor poisons are added, each of which absorbs neutrons differently or introduces additional hazards into the reactor. To compensate for the reduced reactivity of the older fuel, fixed neutron absorbers were removed from the reactor. With less absorbers in the reactor, a higher proportion of neutrons were lost to absorption in the reactor coolant.
At the time of the accident, the reactor in Chernobyl had been running for about three years. After this amount of time, the void coefficient had become so strong, that the negative feedback from the fuel heating was no longer strong enough to counterbalance the effect of the void coefficient.
By the time of the accident, under certain conditions, the reactor operated in a positive feedback loop.
An increase in power, left unchecked would create a further increase in power. Only the reactor's control rods then kept the reactor under control, The majority of these control rods inserted from the top of the reactor. Some inserted from the bottom. They served to absorb any excess neutrons in the core and act as the final brake on the reactor to keep it in control, to keep the reactor critical.
Fission reactors are at their happiest when they're critical. A critical reactor is a reactor running in a balanced steady state at a constant power. It's the desired state of being. Every fission is creating one further fission and that's it. A reactor that is supercritical, is a reactor that's accelerating - each fission creates more than one further fission. A reactor that is subcritical, is a reactor that's decelerating - each fission creates less than one further fission.
The reactor is moved from state to state by adding or removing reactivity. Reactivity is like the throttle and brake on the reactor. It's not really the current power level - it's close to the potential change in power level. Positive reactivty means fission is more likely to happen than it is now - which will cause an increase in power. Negative reactivity, means making fission less likely to happen - which will cause a decrease in power.
In theory, there is no limit to the amount of negative reactivity you can add - all it does is stop the reactor faster. There is a limit to the positive reactivity.
When an atom fissions, the vast majority of neutrons are released instantaneously - at the moment of fission. The neutrons fly away, get themselves moderated, and in the space of microseconds find more atoms to collide with and split. The scientists of the Manhattan project called this a 'Shake' and it is an extremely short interval of time - from nanoseconds to microseconds. These are called Prompt neutrons.
Fission with prompt neutrons happens so quickly, that there is little to no mechanical process capable of controlling and regulating it. If the universe had been created in such a way that there were only prompt neutrons - controllable fission power would likely be impossible.
Had this been the case, the Chicago Pile 1 experiment could have had a far more amusing result.
Luckily for the citizens of Chicago, a very small fraction of the neutrons released by a fission event are delayed - they happen seconds, to minutes later. It is this miniscule fraction of delayed neutrons which enables every nuclear fission reactor to be controlled. It is possible therefore, to have a reactor which is critical on the combination of the Prompt, and the Delayed neutrons. This is how things normally are. Even a supercritical reactor will take seconds to minutes, to change power output. There's time there for the process machinery of the reactor to respond to changes and stabilize.
However, if the reactor is pushed to the point that it is capable of achieving criticality on the prompt neutrons alone - before any delayed neutrons are emitted - things get interesting. Instead of a power increase that happens on the order of seconds to minutes - now the only limiting factor the the reactor's power increase is the time it takes for one neutron to find the next atom to fission and however long the reactor manages to withstand the energies that are being very rapidly liberated. A reactor which has gone prompt critical, has become, in effect, a really, really shit nuclear bomb. The big difference being that bombs take advantage of physics, inertia and a dozen other things to keep the reaction going that few nanoseconds longer it takes to go from 5 tons of TNT, to 15 Kilotons of TNT.
Scientists at the Manhattan project, for whatever reason, called this interval a 'Dollar' of reactivity. Once you get a reactor past that point - unless it's a type specifically designed to go there and self recover - the reactor will be destroyed. Importantly, this does not have to happen within the entire reactor - it can be limited to a very small part of the core where conditions align like the stars over R'lyeh.
At the Chernobyl reactor, Reactivity was added by fresh fuel, by removing control rods and by boiling water to make steam. Reactivity was removed by increasing water flow, adding control rods, heating up the moderator and fuel, and by another factor.
The shrapnel left over from fission creates what's known as 'fission products'. Most of these are hideously radioactive. Many of these are effective at absorbing neutrons, and hence are called poisons. Absorbed neutrons reduce reactivity, which has to be compensated for either by withdrawing control rods, or by removing the used fuel and replacing it with fresh fuel. One of the most effective neutron absorbers is an isotope of Xenon, called Xenon-135.
A product of radioactive decay, it starts to appear about six hours after the fission events that effectively 'created' it. The amount of it that's created, is in direct proportion to the quantity of fission that happened six hours ago. So if a reactor is run at full power for a long time, and then throttled down, Xenon will continue to appear according to that fuel burned six hours previously. It's a bit like the exhaust from your car's engine magically taking longer to form after the combustion in the cylinders. Normally, with the reactor in a steady-state, Xenon is created as quickly as it is consumed - the physics balances out. It can make it very difficult to increase or reduce power - if power is reduced too quickly, and the Xenon continues to build, the reactor might even be stalled by it.
It can also mean that, if the fuel in the reactor has been active for a long time - there may not be sufficient reactivity in the remaining fuel to overcome this Xenon pit. The reactor will stalled and effectively impossible to start, no matter how many control rods are withdrawn.
This is important. After a few more hours, the Xenon goes away. More than that, Xenon which absorbs a neutron also 'goes away' - it's no longer Xenon-135 and it's massive ability ot hoover up neutrons is suddenly gone.
Keeping all of these positive and negative reactivities in balance is the job of the Senior Reactor Engineer, who manipulates the reactor core's systems and control rods to achieve the required stable power output. The Engineer at the reactor's controls has only so much control as the rods will give them.
The Control rods of an RBMK reactor are manufactured from boron. Boron absorbs neutrons, which reduces reactivity, and causes a power drop. The deeper they go into the reactor - the more neutrons are absorbed - the further reactivity decreases. They can also be moved independently of each other - which changes where and how power is produced throughout the reactor, to compensate and balance for old and fresh fuels and how they're distributed through the reactor.
But, at the tip of the control rod on a telescoping extension, is a single slug of graphite. The graphite tip of the control rod acted as a displacer. Its purpose was to push water out of the control rod channel, to remove it and its neutron absorption effect after the rod was withdrawn. In effect, instead of giving the control rod an action of -3,-1 - they are something like -3,+0. They graphite displacer gives the control rod a stronger control action. It makes it more powerful by adding reactivity after the rod is withdrawn. There was about 1.5 metres of water between the control rod proper, and the displacer rod.
This is the first generation RBMK reactor, as built at Leningrad Unit 1.
Shortly after these first reactors were completed, computer models had begun to show some unexpected tendancies in the reactor core. When running at low power, the reactor began to behave strangely. It became less and less stable, and harder to control. Attempting to balance power in the bottom part of the reactor by adjusting the control rod positions could cause momentary power-surges and hot spots - especially towards the edge of the reactor close to the neutron reflector. This could potentially damage the reactor, damaging fuel elements, or potentially rupturing the fuel channel
In one possible worst case scenario, the fuel would superheat to the point where it began to melt, triggering a chemical reaction between the molten fuel, the zirconium cladding of the fuel, and the now suddenly superheated steam in the channel. Hydrogen gas would be released, along with quantities of oxygen - potentially leading to explosive results.
Finally, there is the concept of the Reactivity Margin. This is an estimate of the amount of reactivity the reactor operator can still add to the reactor to increase the power or move power around the reactor. In an RBMK, the reactivity margin was measured in numbers of control rods, where One Control Rod of reactivity, was the equivelant reactivity introduced by removing one average control rod fully from the reactor. A reactivity margin of Thirty Control Rods did not necessarily mean there were Thirty control rods inserted fully into the core - this would be spread out amongst the different rods, each inserted at a different height in the reactor to properly balance power.
In general, the higher the reactivity margin, the further the control rods are inserted in the core and the more reactivity can be added to the core by removing them . A reactor with fresh fuel will have a very high reactivity margin. A reactor with old fuel, or with xenon poisoning, will have a low reactivity margin. It may seem that a low reactivity margin might be 'safer', because now there's less reactivity that can be added by the control rods (which are already out of the reactor at low margin). At a high reactivity margin, more control rods are inside the reactor. More of them can be withdrawn, to push the reactor further into the supercritical - more positive reactivity can be added to the reactor.
The Chernobyl reactor was happy around about 30 Rods of reactivity. The equivalent of thirty full control rods were left in the reactor. The official limit, was somewhere around 15 Rods of reactivity.
This was not considered a safety limit. Running at low reactivity margins could be a mark of skill of the reactor operator. A low reactivity margin meant less scope for moving power around the reactor and keeping things in balance. It increased the risk that sections of the core could be overloaded.
It was never thought of as a safety limit. In November 1975, Leningrad Unit-1 an accident occurred. A fuel channel rupture while operators were trying to increase power after a shutdown. A similar incident happened at Chernobyl Reactor 1. In both cases, the reactor's safety systems operated correctly. Radiation was still released through the plant's ventilation systems. Neither incident was reported in the media.
Reactor 4 at Chernobyl is slightly different. Reactor 4 was a second generation reactor. Over the years, a few improvements to economise the reactor and try to stabilise the reactor's void coefficient are made. There're changes in fuel enrichment, control rod design, and rod configuration. Nothing seemed major. As part of these changes, the control rod tips were shortened by approximately a meter.
These modifications led to an unsettling side-effect when tested in practice. The control rods, when fully withdrawn, allowed a 1.5 metre high column of water to remain in the bottom of the reactor. This water would absorb neutrons in the bottom of the reactor, acting as a gentle brake on fission.
When a newly-built RBMK reactor at Ignalina in Lithuania was being run through its commissioning tests in 1984, it was discovered that, where control rods had been pulled all of the way out of the reactor and where most of the fission in the reactor was happening at the very bottom of the core - an attempt to slow the reactor down by inserting the control rods could cause a momentary increase in power before the boron control rods travelled the full height of the core to quence the reactor.
The water slugs at the bottom of the control rod channels were replaced with graphite for a few seconds. The graphite moderated neutrons, where the water had absorbed them. Consequently, an increase in reactivity occured at the bottom of the core.
Instead of decreasing, power momentarily increased.
The engineers at Ignalina wrote a letter to the reactor's designers, advising them of the deviation. This was not initially thought to be much of a concern - power changes in the reactor after all, take longer than it takes for the rod to travel.
A similar effect was noted at an RBMK in Smolensk, at Chernobyl reactor 3, and finally, at Chernobyl reactor 4 when it was completed.
It gradually became clear to the designers that accidents with the RBMK were not only possible - but even likely in some circumstances. Still - it would reflect badly on the designers and the ministry they worked for if the reactor they had overseen was found to have a potential flaw, especially after millions of rubles had been spent on building them. Careers, influence and bureaucratic prestige were at stake. It was quietly buried as a footnote in the reactor documentation. Not as a threat, or a risk - just as a deviation of the reactor from its design specifications. This effect was something that happened from time-to-time in very specific circumstances, but one which had never really caused a problem.
Even when the Soviet government found out about the cover-ups, no overt action was taken, no matter how furious they were in private. The world at large couldn't know. Even so, nature could only be be fooled for so long. A series of modifications were proposed - changes to the sequencing of the control rods especially - which should've reduced the probability of this 'tip-effect' causing the reactor to run up. The plan was to quietly fix the reactors, saving face before the world. These fixes were to be implemented at Chernobyl Reactor No. 4 at it's next scheduled maintenance shutdown. Before that, however, one further test of the reactor's safety systems was required.
We do not yet know how an RBMK reactor explodes. But we know what we need to know.
Last edited:
