It is very difficult to predict the future of scientific developments, and few would even dare to make predictions extending beyond the next 50 years. However, based on everything we know now, one can make a strong case for the thesis that nuclear fission reactors will be providing a large fraction of our energy needs for the next million years. If that should come to pass, a history of energy production written at that remote date may well record that the worst reactor accident of all time occurred at Chernobyl, USSR, in April of 1986.
In that accident, a substantial fraction of all of the radioactivity in the reactor was dispersed into the environment as airborne dust — its most dangerous form. It is difficult to imagine how anything worse could happen to a reactor from the standpoint of harming the public outside.
In the wake of the Chernobyl accident, the primary question on American minds was — can it happen here? Let us try to answer that question.
We have just seen how extremely improbable an accident of that magnitude should be. But if it is so extremely improbable, how could it have happened so early in the history of nuclear power? The response to that question is that there are very major differences between the Chernobyl reactor and the American reactors on which our previous discussion was based.
In order to understand these differences, we must delve much deeper into the details of how reactors work. This discussion may also be useful to those with an interest in the basic science behind nuclear power.
In an ordinary furnace, energy is produced in the form of heat by chemical reactions between the fuel and oxygen in the air. A chemical reaction is actually a collision between atoms in which their orbiting electrons interact. The other constituent of an atom is the nucleus. If two nuclei collide and interact we have a nuclear reaction. However, unlike atoms, which are electrically neutral, nuclei have a positive electric charge and therefore strongly repel one another. Hence nuclear reactions do not normally occur in our familiar world.
An exception to this situation is the neutron, one of the two constituents of nuclei (the other is the proton), which does not have an electric charge. It can therefore approach a nucleus without being repelled and induce a nuclear reaction. Because this happens so easily, a neutron can move about freely for only about 0.0001 seconds before it collides with a nucleus and becomes involved in a nuclear reaction. Since free neutrons last for such a short time, they must be produced as they are used. Neutrons can only be produced in nuclear reactions, so what is needed is a nuclear reaction induced by a neutron which releases more than one neutron. These can then induce further reactions which produce more neutrons, and so forth, in a self-sustaining chain reaction. Such a reaction is available in the interaction of a neutron with a uranium-235 (U-235) nucleus. This is the basis for a nuclear reactor.
When a U-235 nucleus is struck by a neutron, it often splits into two nuclei of roughly half the size and mass in a process called “fission.” Since all nuclei have a positive electrical charge, these two newly formed nuclei repel one another very strongly. As a result they end up traveling in opposite directions at very high speeds, which means that their motion contains lots of energy. As they travel through the surrounding material, whatever it may be, they strike other atoms, giving them some of their energy, until, after about a million such collisions over a few thousandths of an inch of travel, all of their energy is dissipated, and they come to rest. The atoms they strike or their orbiting electrons are given additional motion and have collisions with other atoms, sharing their energy with them. By these processes, the energy released in the fission process is eventually shared by all of the atoms in the vicinity. It increases the speed of their normal random motion and our senses interpret this as increased temperature. Thus, the fission reaction releases heat, 50 million times as much heat as is released in the chemical reaction between a carbon atom from coal and oxygen atoms from the air in the coal-burning process. The purpose of a nuclear power plant is to convert this heat into electricity, as we described in Chapter 6.
The two original fragments from the fission process also have a substantial excess of internal energy which they largely dissipate by shooting off neutrons, typically two or three neutrons from each fission reaction. It is these neutrons that sustain the chain reaction. In order for it to be self-sustaining, at least one of them must strike another U-235 nucleus and cause a fission reaction. Some neutrons get past the surrounding U-235 and are lost to the process. If enough neutrons are lost, the chain reaction will stop. These losses are reduced as the thickness of the U-235 that the neutron must traverse increases. This means that for the chain reaction to be self-sustaining, there must by some minimum amount of U-235. This is called the critical mass. To generate energy, one need only assemble a critical mass of U-235, which is about the size of a cantaloupe, and introduce a few neutrons to start the process. There are simple and readily available ways of providing these start-up neutrons.
But where do we get the U-235? Uranium occurs in nature as a mixture of 99.3% uranium-238 (U-238) with 0.7% U-235. When a neutron strikes U-238, that nucleus does not undergo fission. If we assemble a large mass of natural uranium, we do not get a self-sustaining chain reaction because the great majority of neutrons are lost by striking U-238 nuclei. As one possible solution to this problem we can separate the U-235 out of natural uranium; we do this for making bombs, but it is a very difficult and expensive process.
However, an alternative and much better approach is available. If the neutrons can be slowed down to very low speeds — one ten-thousandth of the velocity with which they originally emerge — due to the quirks of quantum physics, their inherent probability for striking a U-235 nucleus becomes 200 times greater than for striking a U-238 nucleus. In this situation, even with natural uranium most neutrons would strike U-235 nuclei, and we could get a chain reaction.
The method for slowing down neutrons is to arrange for them to strike and bounce off lightweight nuclei, giving the struck nuclei some of their energy. Materials introduced for this purpose are called “moderators” since they moderate the speed of the neutrons. When a neutron strikes any nucleus, there is some chance that it will be absorbed, but the probability varies by large factors for different nuclei. Since we cannot afford to lose many neutrons, a moderator is only suitable if it has a low probability for neutron absorption. This leaves very few options. One of these is very high purity carbon in the form of graphite. It is such a good moderator that natural uranium dispersed in very high purity graphite can provide a chain reaction. That is how the first chain reaction was achieved in the famous experiments directed by Enrico Fermi under the stands of the University of Chicago football stadium in 1942. (That reactor may be seen at the Smithsonian Museum in Washington.) Another possible candidate for a moderator is ordinary water, but its propensity for capturing neutrons is not as low as one would like. A chain reaction cannot be achieved from a mixture of natural uranium and water. (Actually, this is fortunate because if it could be achieved, reactors would be very easy to make and Hitler would have had nuclear bombs during World War II.) However, if the uranium is enriched in U-235 up to 3% (from its normal 0.7%), then water becomes a good moderator. It turns out that providing this relatively low enrichment is not prohibitively expensive.
One further problem in operating a reactor is controlling the rate at which the chain reaction proceeds, which determines the rate at which heat is produced. This is done with “control rods”, rods made of a material which strongly absorbs neutrons. Pushing control rods in absorbs more neutrons to slow down the chain reaction, while pulling them out allows more neutrons to strike uranium nuclei, which speeds up the chain reaction.
From the foregoing discussion, we see that two of the principal options for reactor design are:
Another safety advantage of the U.S. approach is that if, for any reason, the chain reaction speeds up, releasing more energy and thus causing the temperature to rise, the water acts as a buffer. The increased temperature will cause more boiling. This will reduce the amount of moderator, which will slow down the chain reaction and thereby reduce the temperature. The reactor is, therefore, stable against a temperature change; that is, an increase in temperature automatically causes things to happen which will reduce the temperature. No human action or equipment failure can interfere with this natural process.
In a Chernobyl-type reactor, on the other hand, an increase in the speed of the chain reaction causes the temperature to increase, which causes more water boiling. This reduces the amount of “poison,” which causes the chain reaction to accelerate and increases the temperature even further. This process, therefore, tends to make the reactor unstable against a temperature change; an increase in temperature automatically causes things to happen which lead to further increases in the temperature. Something must be done by some person or equipment to prevent the situation from escalating to a disaster. Actually, under normal operating conditions, other factors would contribute to overcome this instability, but in low-power operation, where the infamous accident occurred, this instability represented an extremely dangerous safety problem.
With these two very clear safety advantages for the U.S.-type reactors, one might ask why anyone would build a Chernobyl-type reactor. The reason is that Chernobyl-type reactors are designed to produce plutonium for bombs while they generate electricity. This type of reactor has two big advantages for this application.1 One is that the quantity of plutonium produced varies inversely with the ratio of U-238 to U-235, which means that much more plutonium is produced in Chernobyl-type reactors than in U.S. reactors. The other is that in producing plutonium for bombs, it is important that the fuel be left in the reactor no more than 30 days, and a Chernobyl-type reactor is much better adapted for that purpose.
In a U.S. reactor, all of the fuel is inside a single large vessel, and it is a major effort, requiring about a month’s time, to shut down the reactor, open the vessel, and change the fuel. Therefore, this operation is undertaken no more than once a year, which makes these reactors unsuitable for producing weapons-grade plutonium. In a Chernobyl-type reactor, each of the 1,700 fuel rods is enclosed in a single tube through which the water flows. It is relatively easy to open one of these tubes at a time, change the fuel rod, and replace it, without having to shut down the reactor. This makes these reactors excellent facilities for producing bomb-grade plutonium as they generate electricity. In fact, some of the U.S. government reactors designed only to produce plutonium for bombs are somewhat like the Chernobyl-type reactor. After the Chernobyl accident, there were serious questions raised about safety hazards in these U.S. production reactors, but it was eventually concluded that they contain design features that assure their safety.
However, there is one further price in safety that must be paid for the capability to change fuel easily. The fuel-changing operation requires a lot of space and activity by operators. This makes it impractical to enclose the reactor in the type of containment used for U. S. reactors (as described in Chapter 6). The containment used in a Chernobyl-type reactor is designed only to protect against rupture of one of the 1,700 tubes, rather than against a major accident that may rupture hundreds of tubes. All of the added safety obtained from containments in U.S. reactors was, therefore, not available at Chernobyl. In fact, post accident analyses indicate that if there had been a U.S.-style containment, none of the radioactivity would have escaped, and there would have been no injuries or deaths.
In April 1986, it was decided to use the Chernobyl power plant for an electrical engineering experiment on its turbine-generator, the machinery used to convert the energy of steam into electricity. The purpose was to develop a system for utilizing the rotational inertia of the turbine-generator to operate water pumps if electric power should be lost. The only function of the reactor was to get the rotation of the turbine and generator up to speed before beginning the experiment. Since no experimentation with the reactor was involved, no reactor experts were on hand. Electrical engineers supervised the experimental work while the reactor was run by the regular operators.
The experiment was set to start at 1:00 P.M. on April 25, but a need for the plant’s electrical output developed unexpectedly, delaying the experiment until 11:00 P.M. At that time, the power level of the reactor was reduced to the level desired for the experiment, but in the operators’ rush to make up for lost time, they reduced the power too rapidly.
Reactors have a peculiar characteristic: if they are shut down, a neutron-absorbing “poison” develops that prevents them from being restarted for many hours. The overly rapid reduction in power led to a buildup of this poison that made it difficult to get anywhere near the desired power level, 25% of full power. In order to get as much power out of the reactor as possible, many of the control rods had to be withdrawn, but still, the power level was only about 6% of full power (one-fourth of the power level planned for the experiment). At this low power level the temperature instability becomes very pronounced, and it was, therefore, strictly against the plant rules to operate under those conditions.
Nevertheless, at 1:00 A.M. (April 26) the supervisors decided to go ahead with the experiment. At 1:05 A.M., additional water pumps were turned on as part of the experiment; these were the pumps to be driven by the rotational inertia of the turbine-generator following a loss of electricity. No one seemed to notice that this action was providing too much water flow for the reactor at this lower power level. In fact, that quantity of water flow at such a low power level was forbidden by the rules. Coincidentally, a normal operating situation came up which, by a quirk in the reactor design, caused a further increase in water flow at 1:19 A.M. Since water acts as a poison, this additional water flow required withdrawal of the manual control rods. That put the reactor into a condition such that a loss of water would make it “prompt critical”, which means that the power would escalate very rapidly, doubling every second or so. Operating any reactor in that condition is strictly prohibited, but apparently ignoring rules was not considered to be a serious transgression at the Chernobyl plant.
At 1:22, the added water flow started at 1:19 was stopped, but since it takes a minute or two for this to affect the conditions in the reactor, the manual control rods were not immediately re-inserted. The very dangerous operating condition continued. At 1:22:30, a computer printed out a warning that the reactor’s condition was unsafe and that it should be shut down immediately, but the operator decided to ignore it, a very serious rule violation. It is difficult to understand why he ignored it, but we will never know because he died in the accident. At 1:23:04 the experiment began.
One of the effects of the experiment was to cause the water pumps turned on at 1:05 A.M. to slow down — their normal electrical drive was shut off, and they were then being driven by the rotational inertia of the turbine-generator. The slow-down of these pumps reduced the water flow to the reactor, which caused more of the water in the fuel tubes to turn into steam. Another effect of the experiment was to cut off the flow of steam to the turbine — it was to continue rotating only by inertia — which increases the amount of steam in the reactor. All reactors have an interlock which automatically shuts down the chain reaction when the steam supply to the turbine is cut off, but that interlock was disabled for the experiment (as were several others). At about this time, the reduction in water flow at 1:22 A.M. began to have an effect, which further increased the amount of steam in the reactor. Since water is a poison, converting water to steam reduces the amount of poison in the reactor, causing the chain reaction to speed up. In reaction to this, the automatic control rods went all the way in, which gives the maximum effect they can provide to reduce power, but the chain reaction still continued to speed up, accentuated by the instability against temperature increase described above. The increase in temperature caused more water to boil into steam, which further accelerated the chain reaction, and further increased the temperature.
If the manual control rods had not been withdrawn, there would have been no problem, but unfortunately there was not enough time to reinsert them. At 1:23:40, the order was given to insert the emergency shutdown control rods. However, because of a questionable aspect in their design, their insertion is rather slow, only one-fourth as fast as in U.S. reactors. Before they got all the way in, they were blocked by damage that had already occurred. Some of the tubes had burst. The speed of the chain reaction continued to escalate, and there was no way to stop it. In a U.S. reactor, loss of water would have meant loss of the moderator, which would have stopped the chain reaction, but in the Chernobyl reactor the graphite moderator was still in place. The speed of the chain reaction, and hence the rate at which it was producing heat, reached 100 times what the reactor was designed for.
Very high temperatures melt and vaporize things. This builds up pressure that can lead to damaging explosions. Intense heat also causes water to react with metals to form the explosive hydrogen gas we discussed in Chapter 6. No one knows exactly what took place inside the reactor, but at 1:24 A.M., there were two loud explosions, and glowing materials were seen flying out of the top of the reactor building. These explosions were not, of course, nuclear bomb detonations, but rather more like explosions of an overheated furnace or boiler. Nuclear bombs require much more highly enriched uranium than that used in power plants.
The most immediate problem at this point was to put out the fires, especially because there was another reactor in the same building that was in immediate danger. At about 1:30 A.M., firemen arrived from the nearby cities of Pripyat and Chernobyl, and by 3:54 A.M., the most threatening fires were out. By 5:00 A.M., all fires in the building were out and the other reactor was shut down. Two reactors in a contiguous building continued in operation for another 20 hours before receiving permission from Moscow to shut down. They were put back into operation a few months later.
The firemen displayed extraordinary heroism in putting out the fires. They received very high radiation doses, largely from radioactivity sticking to their bodies. In addition, they suffered thermal and chemical burns. Many of them later died. The most serious effects, due to beta rays irradiating their skin, would have been averted if they had worn the protective clothing routinely used in U.S. plants when dealing with radioactive contamination. In fact, it would have been very helpful even if attention had been paid to cleaning the exposed parts of their bodies to remove radioactive materials that were sticking to their skin.
The next problem was that the graphite inside the reactor was afire. Chemically, graphite is not much different from coal and is a fine fuel. Recall that radioactivity from a reactor is dangerous principally when it is dispersed into the environment as a fine dust. It would be difficult to devise a more efficient dispersal method than to engulf the radioactivity in a very hot fire sending a plume of smoke high up into the air. Radioactive dust was spewing out with this smoke. Something had to be done to put out the fire inside the reactor.
The firemen first attempted to pump water into the reactor, but that was unsuccessful. They then decided to drop materials from helicopters. From April 28 to May 2, 5,000 tons of boron compounds, dolomite, sand, clay, and lead were dropped onto the reactor. Boron was used because it strongly absorbs neutrons and thus stops a chain reaction. Lead melts easily, and they hoped it would flow over the top to keep air out and thus stop the burning. The dolomite, sand, and clay were to smother the fire. The helicopter pilots had to fly into the rising plume of radioactive dust, exposing many of them to heavy doses of radiation that later proved fatal. As a result of their heroism, the discharge of radioactive dust dropped sharply by May 6.
Many of the firemen and helicopter pilots, as well as some of the workers inside the plant, received radiation doses of more than a million millirems. In all, 31 men died, two of them killed immediately by the explosions, and the rest as a consequence of burns and radiation sickness (see Chapter 5). According to the group of Soviet physicians who toured the United States, none was saved by the bone marrow transplants carried out by a visiting American physician, accompanied by widespread publicity.
While these deaths among workers at the plant were horribly tragic, it is perhaps worth noting that an average of 50 deaths occur every day due to occupational accidents in the United States, and single accidents that kill more than 31 workers occur frequently in coal mines.
There has been no direct evidence of injury to any member of the public as a result of the Chernobyl accident, but there were substantial doses of radiation. The city of Pripyat, with a population of 45,000, mostly families of plant workers, extends from near the plant to 2 miles away. However, exposure in that area averaged only 3,300 mrem, because radioactive materials projected high into the air did not descend rapidly enough to affect those close by. The largest exposures, averaging 50,000 mrem, were received by the 16,000 people who lived from 2 to 6 miles away. The 8,200 people living from 6 to 9 miles away also received substantial doses, averaging 35,000 mrem. The 65,000 people living from 9 to 18 miles away received only about 5,000 mrem. All of these 135,000 people were evacuated over the first few days to avoid further exposure from radioactive material deposited on the ground as well as from that still being released from the reactor.
For the most-exposed 16,000 who averaged 50,000 mrem, their risk of dying from cancer as a result is about 4%, raising their total risk of dying of cancer from the normal 20% to 24%. This is less than some of the variation in cancer risk from living in different U.S. states. A Soviet scientific team has announced plans to carefully keep track of these highly exposed people to determine how many cancers actually do appear.
For the first 2 days after the accident, the winds carried the radioactive dust over Finland and Sweden. On the third and fourth day, the wind shifted to bring it toward Poland, Czechoslovakia, Austria, and Northern Italy. It then shifted further southward to deposit the material over Rumania and Bulgaria.
People all over the world were exposed to external radiation from radioactive gases and dust suspended in the air and settled on the ground. They were also exposed internally by inhaling these materials or eating foods contaminated with them. The average radiation doses to the public3 in millirems during the first year after the accident were 76 in Bulgaria, 67 in Austria, 40 to 60 in Greece, Rumania, and Finland, 30 to 40 in Yugoslavia, Czechoslovakia, and Italy, 20 to 30 in the USSR, Poland, Switzerland, Hungary, Norway, and East Germany, 10 to 20 in Sweden, West Germany, Turkey, and Ireland, and less than 10 elsewhere (0.15 in United States and Canada). Note that in no country was the exposure higher than one-fourth of that due to natural radiation during that year.
Some of the material on the ground will continue to be radioactive for many years, exposing people externally and internally through the food supply. The estimated average total exposure in millirems after the first year3 will be 120 in southeastern Europe, 95 in North and Central Europe, 81 in the USSR, 15 to 19 in Western Europe and Southwest Asia, 8 in North Africa, and less than 2 elsewhere (0.4 in North America). The sum of exposures to people all over the world will eventually, after about 50 years, reach 60 billion mrem,3 enough to cause about 16,000 deaths. Note that this is still less than the number of deaths caused every year by air pollution from coal-burning power plants in the United States.
Since the mechanism for dispersing radioactivity over long distances was so efficient in the Chernobyl accident and is so inefficient in U.S. reactors, it is almost impossible to believe that an accident in a U.S. reactor can ever cause nearly as much radiation exposure at large distances from the plant.
Any technology is developed and improved by learning from past mistakes, and the nuclear power establishment has always been very active in this regard. A tremendous amount was learned from the Three Mile Island accident. Within weeks, two large new organizations were established by the nuclear industry, the Nuclear Safety Analysis Center (NSAC) in Palo Alto, California, and the Institute of Nuclear Power Operations (INPO) in Atlanta, Georgia. Both have been very active since that time. NSAC carries out very intensive technical analyses, including some referred to in this book. INPO has established an impressive electronic communication network among power plants, reactor vendors, and centers of expertise all over the world. It also conducts inspections of power plants, offering suggestions and criticisms that are taken very seriously. As a result of INPO activities, there have been remarkable improvements in operational performance. The average number of operating hours per year for plants has increased by 12%. Unplanned shutdowns have been reduced by 70%, accident rates for workers has declined more than three-fold, the volume of low-level radioactive waste has declined by 72%, and radiation exposure to workers has been cut in half. INPO represents a very intense effort by the nuclear industry to police itself, as a result of the Three Mile Island accident.
Government agencies were also very active in trying to learn from the Three Mile Island accident. There was a Presidential Commission to investigate it and offer suggestions and criticisms of the industry. The NRC carried out very extensive studies and even produced a thick document entitled “Lessons Learned From The Three Mile Island Accident.” The NRC also developed a set of modifications in nuclear plants to respond directly to some of the problems uncovered in the analysis of the accident; these modifications were quickly implemented at a cost of about $20 million dollars per plant. In addition, the NRC undertook a broad study of a variety of nuclear safety issues. It is not an exaggeration to say that lessons learned from the Three Mile Island accident revolutionized the nuclear power industry.
After the Chernobyl accident, both government agencies and the nuclear industry were eager to investigate and learn from the experience. However, after long and careful study they finally concluded that we had very little to learn from it. The whole episode is now viewed as a vindication of the U.S. approach to nuclear power. (Essentially all nuclear power programs outside of the Soviet bloc use the U.S. approach.)
To understand this, let us review some of the problems that contributed to the accident, which would have been avoided by the U.S. approach:
Actually, it had been recognized for decades that Soviet reactors could not be licensed in the United States, even if they had containments. Items 4 and 5 in the above list further emphasize the fact that reactor safety has always received much lower priority in the Soviet Union than in the United States. An obvious question is why this is so.
There has been substantial contact between Soviet and American reactor safety experts, including many visits in both directions, personal friendships, and lots of informal discussions over cocktails. The above question has been asked and discussed many times, and the Soviet reply runs along the following lines.
The extreme concern about reactor safety in the United States has gone far beyond the bounds of rationality. It is difficult to argue with the Soviets on this point. In Chapter 8 we will show that our reactor safety programs have spent billions of dollars per expected life saved. This is irrational for two reasons. First there are many opportunities for saving lives with medical screening programs, highway safety measures, and the like, at a cost of about $100,000 per life saved, so the money spent to save one life from a nuclear reactor accident could save over 10,000 lives if spent in these other areas. Second, as a result of the cost increase for nuclear power plants, utilities are forced to build coal-fired power plants instead of nuclear plants, and the air pollution from a coal-fired plant is estimated to cause several thousand deaths over its operating lifetime. This irrational attitude toward nuclear reactor safety in the United States is, therefore, leading to thousands of unnecessary deaths every year, and wasting billions of dollars that could be used to save thousands of other lives. Why should the Soviet Union repeat our insanity?
It is difficult to argue with this logic of the Soviet reactor safety experts, but they carried things too far. The Chernobyl-type reactor is very much more dangerous than U.S. power reactors ever were. All the differences we have pointed out apply equally to even the earliest American power reactors. The United States has erred in one direction, but the Soviets erred in the other. They truly gambled, and they lost.