IN the early hours of 26 April 1986, two explosions rocked the Chernobyl
nuclear power plant in Ukraine. The blasts destroyed one of four reactors at the
site and ripped through the building that housed it. As pieces of hot,
radioactive debris rained down, fires broke out on the roofs of neighbouring
buildings. More serious, however, was the inferno that engulfed the graphite
core of the reactor and spewed radioactive gas and particles into the atmosphere
for 10 days and nights. In the immediate aftermath, 31 people died of acute
radiation sickness, most of them firefighters who had battled to put out the
blaze.
The Chernobyl disaster was the world鈥檚 worst nuclear accident. It
contaminated crops, livestock and land over vast areas of Europe, triggered an
epidemic of thyroid cancer in children living near the plant, and destroyed
public confidence in the nuclear industry. But while its consequences are well
known, what happened on that fateful Saturday morning 10 years ago is still a
mystery. Physicists know that an uncontrollable power surge triggered the
disaster, but they also know that this alone could not have caused the
devastation that followed.
The mystery deepened when researchers began to hunt for the 190 tonnes of
uranium oxide fuel that had been in the reactor core. In RBMK reactors, like
those at Chernobyl, the core sits in a concrete shaft some 20 metres high and 17
metres in diameter, and scientists had hoped that much of the fuel would be at
the base of this shaft. But cameras peering through holes in the side of the
shaft show that it is virtually empty. Stranger still, the steel sleeve that
lines the shaft is almost unmarked. How can this be when much of the reactor
building was destroyed?
Advertisement
鈥淎lmost every nuclear group in the world has run computer models of the
Chernobyl incident,鈥� says Milt Levenson, a nuclear engineer based in California,
who in 1989 brought together scientists from East and West to discuss Chernobyl.
But none of these models produced a convincing account of what happened. Perhaps
this is not surprising. 鈥淚f a plane crashes you don鈥檛 build a computer model of
the plane, you look at the physical evidence at the crash site,鈥� says Levenson.
鈥淧eople are pretty good at reconstructing explosions by carefully examining the
debris.鈥� Only recently, have engineers carried out forensic-type examinations of
the site. The problem is that these investigations have led to two conflicting
pictures of what happened in the moments after unit 4 at Chernobyl ran out of
control.
The first of these analyses came from Edward Purvis, a nuclear engineer who
headed the US Department of Energy鈥檚 investigation of the accident in 1986. In
1993, he travelled to Chernobyl as a consultant for the Intersectorial
Scientific and Technical Centre of the Ukrainian Academy of Sciences. His
controversial report, published last year, concluded that the explosion which
destroyed the building did not detonate in the reactor shaft鈥攁s most
people had thought鈥攂ut in the vast refuelling hall above it. He believes
that a powerful build-up of steam shot the core out of its shaft and at least 14
metres into the air, where it exploded. Purvis says that this is the only way to
explain the two explosions and the pattern of devastation.
The events that led up to the accident have been well established. In the
core of an RBMK reactor, neutrons crash into uranium nuclei and split them
apart. In the process they generate heat and release two or three further
neutrons, which maintain the chain reaction. The heat turns water circulating
through the core to steam, which drives the turbines. The best way to control
the fission reaction is to use a neutron-absorbing material such as boron. For
this purpose RBMKs and most other reactors have boron control rods that can be
lowered into the core to slow the reaction, or raised to speed it up.
But other materials also absorb neutrons. Water, for example, is a good
neutron absorber, while steam being less dense, is not. A quirk of the RBMK
design is that when water turns to steam in the core fewer neutrons are absorbed
so more of them collide with uranium nuclei. This phenomenon, called a positive
void coefficient, increases the temperature, generates more steam, and can lead
to sudden power surges.
In normal operation, however, these surges are held in check by another, more
dominant effect: as the core temperature increases, uranium becomes better at
absorbing neutrons and so acts as a natural brake on power surges. Only when the
reactor is running at low power鈥攕uch as when it is being shut
down鈥攄oes the positive void coefficient dominate and, even then, power
surges can be quashed by ensuring that at least 30 control rods are inside the
core.
That night, the reactor at unit 4 was being shut down for routine
maintenance. At the same time, operators intended to test the reactor鈥檚 safety
procedures. Some of the electricity produced by unit 4 was used to keep its own
cooling pumps and other equipment running. While the reactor was being shut
down, power for these pumps normally came from the grid: but what would happen
if the supply from the grid was cut? The safety test was designed to check that
the plant鈥檚 slowing turbines would produce enough power to keep the pumps
running until an emergency diesel generator switched in.
Critical error
The plan was to start the test with the reactor operating at about 30 per
cent of its maximum power. Just beforehand, however, the reactor鈥檚 output fell
to around 10 per cent, a level where the positive void coefficient dominates. At
this point, the operators made a critical error: they tried to increase the
output by withdrawing the control rods, so that at the time of the accident the
core contained fewer than 10. In the space of about four seconds at exactly
1.23.40 am, the rate of fission reactions in the core rose to 100 times its
design limit. Then the trouble really started.
The design of the core played a crucial part in what followed, says Purvis.
In RBMK reactors the core is about the size of a small house. It is made of a
cylindrical stack of graphite blocks weighing about 200 tonnes. Running
vertically through the stack are 1880 tubular zirconium channels that carry
water at pressures of up to 70 atmospheres. Inside most of these channels are
fuel rods鈥攍ong, thin zirconium canisters full of peanut-sized pellets of
uranium oxide. The rods can be replaced while the reactor is running with the
help of a 14-metre-high crane-like device called the fuel handling machine,
which operates in the hall above the core. People working in this hall are
shielded from radiation by a concrete and steel disc 17 metres wide and 3 metres
thick, weighing more than 450 tonnes.
Where the zirconium cooling channels leave the core they connect to steel
pipes. Purvis believes that these transition joints are a weak point in the
design. 鈥淭he coefficients of thermal expansion for these materials differ by a
factor of three,鈥� he says. Prior to the accident operators had switched on a
reserve pump that fed cold water into the reactor and he believes this weakened
the joints.
Steam catapult
When the disastrous power surge took place, the temperature of the fuel
increased rapidly and ruptured some of the fuel rods, says Purvis. Water reached
the hot fuel pellets and evaporated. 鈥淭his sudden pressure pulse led to the
failure of a small number of transition joints at the bottom of the reactor,鈥�
Purvis says. Water flashed into steam beneath the graphite core, causing the
other transition joints to fail. 鈥淚t was like an unzippering and it created a
massive increase in pressure beneath the reactor.鈥�
He calculates that the steam exerted enough pressure to force the core,
biological shield and the fuel handling machine鈥攎ore than 3000 tonnes in
all鈥攐ut of the shaft. 鈥淚t was like a steam catapult,鈥� he says. 鈥淎nd we
know that it leapt high into the air because the fuel handling machine is now on
a ledge 14 metres above its original position. How else could it have got
迟丑别谤别?鈥�
Purvis reckons that this explosion was followed by a second, while the
reactor was still in the air. He believes that water would have been rapidly
ejected from the core, massively reducing the amount of neutron-absorbing
material and allowing the fission reaction to run away. In a fraction of a
second, the temperature of the fuel rose to 7000 掳C. 鈥淭he second explosion
was a blast wave of vaporised nuclear fuel,鈥� he says. 鈥淏ecause it occurred in
the reactor hall, the lining of the reactor shaft is unmarked.鈥�
Purvis backs up his theory with evidence from his examination of the
radioactive debris that lies around the site, which included many tiny spheres
of ruthenium鈥攁n element produced by the fission of uranium. These can only
have condensed from a vapour, he says.
Shortly after the accident, Soviet scientists argued that hydrogen had built
up and ignited to cause the second blast. Purvis admits that hydrogen could have
been created when water came into contact with unusually hot zirconium, but he
says that a hydrogen explosion would not have generated temperatures high enough
to vaporise ruthenium, which has a boiling point of more than 4000 掳C.
Levenson is not convinced on this point. The ruthenium spheres could have formed
within the fuel rods during normal operation of the reactor, he says.
Purvis also notes that the few surviving beams in the roof of the reactor
hall are deformed in a way that is consistent with a spherical blast wave. This
would have been produced by an explosion in the reactor hall rather than the
shaft. But in the northwestern corner, the damage is less extensive. This
suggests that it was protected by the huge biological shield.
鈥淭he second explosion sent the shield bouncing around the reactor hall like a
tossed coin,鈥� says Purvis. Again, the evidence to back up this claim lies in the
debris. The shield is now propped up almost vertically at the mouth of the
shaft, at least 5 metres above its original position. One slab of the north wall
of the reactor hall is pinned beneath it, and steel sheets that lined the floor
of the hall are inside the shaft. 鈥淭he shield must have crashed against the
north wall and bounced off the floor of the hall before coming to rest,鈥� he
says.
Purvis鈥檚 ideas face strong opposition from other researchers who have studied
the accident. They point to the distribution of fuel after the explosion, saying
that it cannot be explained by Purvis鈥檚 theory. Alex Sich, a nuclear engineer
who worked at the Massachusetts Institute of Technology in the early 1990s and
spent two years at Chernobyl, says that most of the fuel is beneath the reactor
shaft. In the days after the accident, he says, the fuel melted through the
reactor鈥檚 lower lid, mixed with the metal and sand-like materials it contained,
and flowed into the rooms and corridors below. As it cooled, this 鈥渓ava鈥�
solidified into strange shapes. 鈥淲e know that most of the fuel is in this lava,鈥�
says Sich.
But estimating the amount of fuel in this lava is no easy task. Researchers
have measured the amount of infrared radiation given off by the fuel and taken
samples of the lava. They have also measured the temperature of air flowing into
and out of the lava-encrusted rooms. The researchers then worked out how much
fuel would produce this temperature rise. According to Sich, these tests show
that the lava contains 70 per cent of the fuel. 鈥淭his one fact challenges
[Purvis鈥檚] entire thesis,鈥� says Sich. 鈥淚t just doesn鈥檛 make sense.鈥�
But Purvis is adamant that Sich and his colleagues are wrong. He claims that
they have not accounted for the heat generated by the temporary lighting that
has been set up inside the remains of unit 4. 鈥淭his has inflated their figures.
And in any case, the volume of the lava is simply not enough to contain 70 per
cent of the fuel.鈥�
He argues that the lava is formed from the remnants of fuel blown back into
the shaft by the second explosion. This could not have been more than 30 per
cent of the total, he says. 鈥淚 think that most of the fuel is on the floor of
the reactor hall,鈥� he says. After the explosion, helicopter pilots identified at
least one hot glowing mass on the floor of the reactor hall. 鈥淭his may have been
a large lump of the core,鈥� he says.
Checking this claim will not be easy. The reactor hall is highly radioactive
and the floor is covered by tonnes of sand and boron that were dropped by
helicopter to extinguish the glow. But more forensic investigation is needed,
says Purvis. First, researchers must look for fragments buried in the walls hit
by the blast. 鈥淭here may by about 2.3 tonnes of fresh fuel embedded in the
eastern wall of the reactor hall,鈥� he says. The type of particle and how deeply
it has penetrated will shed valuable light on the size and nature of the
blast.
Levenson agrees that Purvis鈥檚 ideas deserve more investigation. Analysis of
lava samples obtained by robots does not support the notion that it contains 70
per cent of the fuel. 鈥淓very time they take a new sample, the figure drops,鈥� he
says.
And there are other mysteries that Purvis鈥檚 theory seems to account for.
Pieces of the graphite core and fuel landed all round unit 4. Levenson argues
that these could never have reached these positions if the second explosion had
taken place inside the shaft while the biological shield was in place. 鈥淭he
graphite could not have penetrated the biological shield so how did it get
迟丑别谤别?鈥�
Ukrainian researchers have yet to be convinced. Victor Kashparov, who heads
the department of chemistry at Ukraine鈥檚 Scientific Research Institute for
Agricultural Radioecology, is sceptical about Purvis鈥檚 theory. He says there is
no evidence of an explosive expansion of nuclear fuel as suggested by Purvis.
Such a blast would have bombarded the refuelling hall with neutrons, leaving
behind low-levels of radioactivity in any metallic structures. Purvis鈥檚 theory
would be disproved if this telltale sign is not found, he says.
Sich says that many researchers at the plant are unhappy with Purvis鈥檚
account and angered by his criticism of the measurements they have made of the
lava鈥檚 temperature. A cubic metre of the lava could produce tens of kilowatts of
heat, not the few tens of watts that a light bulb puts out. 鈥淭he heat from the
lighting is insignificant,鈥� says Sich. 鈥淭he researchers at Chernobyl know this.
They are not stupid.鈥�
Lifting the lid
Sich believes that the build-up of steam created by the power surge fractured
a small number of cooling channels and lifted the lid only a centimetre or two.
鈥淭he first `explosion鈥� may have been the noise of this release of steam and the
lid settling back into position,鈥� he says. But this movement sheared still more
of the cooling channels, releasing the pressure and allowing yet more water
inside the reactor to flash into steam. The resulting build-up of pressure sent
the biological shield flying into the air carrying with it just some of the
fuel. Sich鈥檚 version of events will be published this year in the journal
Nuclear Safety.
In the end does it matter how unit 4 was reduced to rubble in such a short
time? 鈥淚f all the RBMKs were going to be closed it wouldn鈥檛 matter,鈥� says
Levenson. But this is not the case. Two RBMKs are still operating at Chernobyl,
and more than a dozen in Russia and Lithuania may survive into the 21st century.
鈥淚f you don鈥檛 really know what the accident scenario was,鈥� says Levenson, 鈥渉ow
can you avoid it in future?鈥�
