“Few of us want a nuclear plant in our community–we’ve heard about Three Mile Island and Chernobyl and know that accidents can happen anywhere.”
-Greenpeace
I think this is a great opportunity to explain some of the basics of how nuclear power works. Bear with me for a bit of nuclear history.
During the Manhattan Project, three different approaches to atomic bomb design were tried: the uranium gun, the plutonium gun, and the plutonium implosion device. The uranium gun (which was evenually used at Hiroshima) propelled one piece of weapons-grade uranium into another. Weapons-grade uranium–more than 93% uranium-235, as opposed to natural uranium’s 0.71%–was extremely difficult to make. They didn’t have centrifuges or lasers or diffusion systems at the time, and had to pass a gaseous uranium compound through electromagnets that would separate it by weight; as the weight difference was only 1.26%, this obviously used quite a bit of energy and took a lot of time. Since they needed to produce many more atomic bombs than their enrichment capacity would allow, they turned to plutonium.
Weapons-grade plutonium (94%+ plutonium-239), by comparison, is easier to make but much harder to use; it must be produced in a reactor by allowing neutrons to hit uranium-238 atoms to produce plutonium-239. Natural uranium will not support a chain reaction out in the open; the neutrons that hit uranium atoms and cause them to split must be slowed down in order to be effectively absorbed. At the same time, the substance that slows down the neutrons must not absorb large amounts of neutrons itself, preventing them from simply using ordinary water. Heavy water is almost as good as water at slowing down neutrons yet absorbs far fewer neutrons than water does, allowing a chain reaction to occur in natural uranium–but it’s difficult to make, defeating the original purpose (another option, enriching the uranium to about 3% so that it will support a chain reaction in ordinary water, combines the difficult parts of both uranium and plutonium approaches and also defeats the purpose of building a plutonium bomb, which is to not use uranium enrichment). This left them with graphite, which was fairly easy to make; however, it had one problem: the process used at the time left residues of the neutron-absorbing metal boron in the graphite, rendering it useless. Altering the process fixed this problem in the US, but the Nazis never figured it out, consequently had to use heavy water, and ran out of time while trying to produce it. But anyway:
From this point, it’s all reactor design. The first artificial nuclear reactor was a carefully-constructed “pile” of graphite and natural uranium; it was used only as a proof-of-concept. If they were going to produce large amounts of weapons-grade plutonium, they would need bigger reactors, and with bigger reactors came two problems. First, the proof-of-concept was a “zero-power” reactor, meaning that it produced a negligible amount of heat, but larger ones would require cooling systems; water was an excellent coolant, so why not use it? Second, only plutonium-239 works in bombs, and significant amounts of plutonium-240 form after a short time in the reactor. Taking the uranium out before it is even close to being fully used is the only way around this and the only way to do that with any semblance of efficiency is to do it while the reactor is still operating. Thus it made sense to have vertical water pipes running under pressure through a block of graphite; when enough plutonium-239 had been produced, a refueling machine would be attached to the top of a particular pipe, equalize the pressure, remove the uranium rod, and insert a new one. This is where they ran into problems.
Recall the tendency of water to absorb many more neutrons than graphite does. This means that losing the coolant decreases neutron absorption–which increases the reaction rate, technically known as a positive void coefficient. This is the opposite of what should happen; a reduction in coolant levels should mean a reduction in power. For several years, they used these reactors to produce large amounts of weapons-grade plutonium without using heavy water or uranium enrichment, but they always had this problem of a positive void coefficient. If a particular reactor of this type had a sudden reduction in coolant level, depending on the reactor’s specific design, the reactor’s heat output could suddenly spike, causing first the fuel rods’ cladding and then the cooling system to overheat and burst. If there is no sealed concrete dome over the reactor to absorb this pressure wave, some of the light already-split atoms can be ejected from the reactor building by the pressure wave, and if this all happens to a reactor that is near an urban area and the reactor is large enough, many people can be exposed to large radiation doses from the ejected materials. Clearly this is unacceptable; unfortunately, the Soviets had spies in the Manhattan Project who gave them this reactor design but were discovered before they could tell the Soviets about the positive void coefficient. The positive void coefficient was treated as a military secret when it was discovered by Edward Teller, even though the fact of its existence had essentially no relevance to a bomb program but quite a lot of relevance to safety. In retrospect, the US government should have told the Soviets about it and offered to cooperate on the development of civilian nuclear power, but it was kept secret, and the Soviets built their nuclear power program around this specific reactor design, known as the RBMK, by converting their waste heat into electricity. Chernobyl was an RBMK, and experienced the same type of accident as described above during a “safety test” that involved draining the reactor’s cooling water.
It turns out that no other reactor design has this characteristic.
While it is possible to operate an RBMK safely, and there are measures that can be taken to mitigate the problems that an RBMK inherently has, many (like enriching the RBMK’s uranium fuel) defeat the original purpose. But the basic problem is that the RBMK stinks as a power plant; it goes incredibly far out of the way to explicitly produce plutonium-239, but in a power plant, why is that necessary? Nuclear reactors are much less picky about their fuel than nuclear weapons; the requirement that the device go out of control and fly apart at high speeds creating a blast wave and high temperatures is no longer there (and in fact is heavily discouraged!)–and with it goes any need to use or produce weapons-grade materials. So why do it at all? In designing a power reactor–a nuclear reactor specifically designed for electricity production–it is actually desirable to cut a few corners in that department for the sake of cost and safety; at the time, waste was not considered to be a problem, and everything was assumed to be a proliferation hazard and kept secret. The experience with the positive void coefficient scared the hell out of nuclear physicists, however, and further development of nuclear reactors always assumed that they were inherently dangerous, never attempted to honestly assess the impact of an accident (substituting an exaggerated worst-case scenario, an exaggerated best-case scenario, and a set of probabilities for everything in between), and did not allow the technology to develop through trial and error as all others have. Strangely, however, they did not think much towards the future; any future hardship would be accepted so long as a design did not have a negative impact on current operations. Accordingly, the RBMK was banned in the United States before any nuclear power plants were ever built, and the search for a reactor that was safe (meaning it cannot go out of control), relatively simple, cheap, and used as much existing infrastructure as possible, with little regard for waste (it was known to be a smaller problem than coal’s was
te output), proliferation (everything’s secret, so why bother separating anything), or the environment (it was rarely a concern in the 1950s). Under such a system, the Light Water Reactor (LWR) leapt to the fore.
The LWR is, in principle, extremely simple. Uranium is enriched to between 3% and 5% uranium-235 and fabricated into ceramic pellets. The pellets are stacked to make rods, which are then lowered into a tank of water. The tank is closed up and pressurized, and boron rods are then withdrawn. The reaction can control itself from this point; the additional uranium-235 overcomes the water’s tendency to absorb neutrons but is not enough to support a chain reaction without the water, so if all else fails and the reactor overheats, the water boils off and the reaction stops. Sixteen natural nuclear reactors–all LWRs–have been found in ancient uranium deposits that date from approximately two billion years ago, when the level of uranium-235 in natural uranium was higher, and they all operated on this inherent physical principle.
An artificial LWR produces more energy, however, and its fuel would melt if exposed while operating. Accordingly, LWRs are designed to never expose their fuel; even though it would not result in a Chernobyl-style accident in which the reaction rate accelerates, it would result in a puddle of melted uranium at the bottom of the tank, a ruined reactor, and an expensive cleanup job. There are still important physical effects that can be exploited in an artificial LWR, however, including convection, gravity, or the fact that hot metal and ceramic expand, and these in principle allow a passively-safe LWR to be built (one that changes its reaction rate based on pure physics).
Without computers, the early nuclear engineers could not juggle the thousands of variables necessary to design a passively-safe nuclear power plant, but they knew that pumps, valves, and other active systems were good enough for cooling if a small probability of a core melt were accepted and a huge concrete dome were installed to keep everything in if such an accident happened. They had the choice to either do complex physics while designing an LWR or do complex engineering to minimize the failure rate of an active system. They chose complex engineering; they really had very little choice. In doing so, they substituted reliability (the failure rate) for safety (the effects of a failure). All major models assumed that nuclear power was inherently unsafe and that the worst conceivable accident was in fact possible (with no data to back this up), and that overconservatism was ironically the basic flaw in early safety estimates. This assumption led them to over-apply backup systems–and those backup systems have been the source of problems ever since. This culminated in Three Mile Island: a government-ordered “safety valve” stuck open and drained the reactor’s coolant. The uranium fuel rods melted and formed a puddle at the bottom of the reactor, which solidified back into solid ceramic when enough of the short-lived already-split atoms had decayed. Three gases–krypton, xenon, and iodine–got out of the reactor through the stuck-open valve and into the concrete containment structure, and some were released out of a misplaced fear that they were hydrogen and were going to explode. Nobody was killed; the radiation emitted by the released materials was no more than the radiation levels at a coal plant.
Had Three Mile Island been designed with passive safety in mind, this accident would have been physically impossible and would not have happened.
Overheating accidents have been induced in modern inherently-safe reactors such as the waste-eating Integral Fast Reactor (IFR); since the reaction rate decreases with increases in temperature, it stopped running based on pure physics alone. A Chernobyl-style accident cannot happen everywhere; it can only happen at a Chernobyl-style bomb factory. A Three-Mile-Island-style accident cannot happen everywhere; it can only happen at a Three-Mile-Island-style reactor that used 1950s-era engineering assumptions that mistakenly tried to be conservative to the point of causing additional safety problems. And that is the key lesson to learn: honest assessment is better than exaggeration either way.
Statements that lump all nuclear power plants together are incredibly irresponsible, just as irresponsible as statements that lump all chemical power plants together. Each plant design must be evaluated on a case-by-case basis; in a sane society, sloganeering does not take the place of asking:
1. How do these accidents happen?
2. Are the conditions that cause these accidents present “everywhere?”
3. If not, how is an accident supposed to happen?
Remember that it is not the presence of an effect that is important; we can argue forever about whether there are effects that we don’t know about (and there undoubtedly are). The matter that should be discussed is whether the technology’s unintended consequences have less impact than clear and present dangers that the technology will solve. If we know that the technology will solve more problems than it will cause, or that it will minimize problems without solving them per se, we’ve found a way to improve society and we should implement it. Why should we abandon a perfectly good solution to three or four problems because it does not solve six or seven? We’re not taking a risk here; there’s no risk in reducing risk!
One extremely important piece of information can be derived from the quote: whenever you hear something like this–a political slogan about a matter of physical fact that does not employ or refer to a physical analysis–run like hell. This tactic is the same thing that religious fundamentalists do, and that other politically-motivated crackpots (e.g., no-Moonies) also do. The same requirement for action applies: kick their butts, kick ‘em hard, and keep up the pressure. The wackos are coming, and they will destroy the sciences if we let them.
It’s time to get going, people.
Filed under Anti-Nuclear Quote of the Day, Chernobyl, International, Physics, Proliferation, Safety, Three Mile Island
Posted on February 13, 2007 by Stewart Peterson | 5 Comments »
