Pique the Geek 20091101: A Primer on Nuclear Electricity

(10 am. – promoted by ek hornbeck)

We shall get away from food for this installment of Pique the Geek and talk about something more, well, geeky. The concept of nuclear power is widely known, but the actual way that is works is mysterious to some because people think that it is hard. Actually, the basic science behind nuclear power is very simple, but the technology to contain and make it practical is complex.

This complexity is due to several reasons, not the least of which is safety. Whilst the nuclear fuel to power commercial reactors is not very malignant, after that fuel has been used a while it becomes extremely radioactive due to a large number of complex nuclear interactions. It is the spent reactor fuel that is the real problem. However, there is a completely different technology used to generate electricity that does not involve a nuclear reactor, and we shall discuss this one first.

Everyone is familiar with the solar arrays on satellites, the space station, and even here on rooftops. These panels take the energy in sunlight and covert it directly to electricity. This works fine, IF the panels are close enough to the sun such that the radiant energy is sufficient to provide enough power. As one gets further from the sun, the energy falls off rapidly (the radiant energy follows an inverse-square relation, meaning that when you double the distance away from the sun, the energy per unit area decreases by a factor of four. Since the best solar panels are only around 15% efficient overall, it turns out that Mars is about the practical limit for solar energy to be useful. For deep space applications, some other source of power is necessary.

The answer is a radioisotope thermoelectric generator. The principal behind it is very simple: the Seebeck Effect. This is the concept behind thermocouples, commonly used to measure temperature. Your kitchen range uses thermocouples to measure oven and burner temperatures, and the new electronic fever thermometers also use them. It turns out that when two different metals are put into electrical contact, a tiny voltage is produced, and the higher the temperature (for most materials), the greater the voltage. For thermometers, the small voltage is amplified by the electronics and converted to a temperature readout. Single thermocouples rarely produce more than 100 microvolts per degree C (technically, degree K but degrees C and K are the same size, just with different zero points). So, to produce practical amounts of power, you hook up lots of thermocouples in series and use an external heat source to keep them hot.

In a generator, the heat source is a radioactive material that provides enough heat to overcome the inefficiency of the energy conversion (thermocouples have a problem shared with photovoltaic cells, they convert only 10% to 15% of the heat to electricity) and at the same time have enough lifetime to be practical. Remember, all radioactive materials have a half-life, defined at the amount of time required for half of its atoms to decay to other atoms. Now, this very decay is the basis of the energy produced, so things with extremely long half-lives produce little energy per unit mass. On the other hand, things with very short half-lives do not last long enough to be practical. There are other considerations as well, such as ease of production, cost, and most of all, radiation hazards. Even on unmanned craft, care has to be taken to protect other components from radiation damage. There are only two isotopes that have been used to any great extent for radioisotope generators, ⁹⁰Sr (strontium-90) and ²³⁸Pu (plutonium-238).

The Soviets used ⁹⁰Sr for several designs of generators, even ones here on earth for lighthouses in remote and cold areas. This isotope has several advantages, since it emits essentially only beta radiation (high speed electrons) and very little gamma radiation, (very penetrating photons) which is very difficult against which to shield, and it is cheap to make. Remember, shielding means weight, and weight is the cost killer for spacecraft. However, there are two distinct disadvantages: the beta radiation, when it impacts matter, produce X-rays that are hard to shield (but not as hard as gamma), and the half life of ⁹⁰Sr is only 28.8 years. Thus, in less than 30 years, half of the energy is depleted. Thermocouples also degrade over time, further reducing efficiency.

In the case of ²³⁸Pu, the radiation is mostly alpha radiation (fast moving helium nuclei) with very little gamma. Alpha radiation does not produce X-rays upon impact with matter, so shielding for generators powered by ²³⁸Pu require little shielding (about 1/10 inch of lead is sufficient, compared to ten times that for ⁹⁰Sr). In addition, the half-life of ²³⁸Pu is 87.74 years, almost three times that of ⁹⁰Sr. The main disadvantage is that it is much most costly to produce, and the US capacity is extremely limited. We have been buying the bulk of our supply from the Russians for some time.

This is not the material used in nuclear weapons. It is ²³⁹Pu that is fissile, not the 238 one, so the only danger is from the radiation itself, not any risk of explosion. Casings for generators are designed to survive reentry into the atmosphere without failure, and the generator from Apollo 13 actually is deep in the Pacific Ocean. Air and water monitoring showed that the casing was not compromised.

So, how have they held up in service? Voyagers 1 and 2, launched in the summer of 1977, are still talking to us, powered only by their radioisotope generators. Thus, after 32 years they are still going strong. Voyager 1 is the most distant human technology from earth, at over 10 billion miles, or almost 0.002 light years. By the way, at that distance, it takes over 14 hours for radio transmissions to reach the earth from Voyager 1. When it says “Hello”, to us, our response back, even if we answer the message immediately, comes over a full day later. It is expected that the generators will have enough reserve power for us to continue communication until 2025 or so, over a half-life of the radioisotope. By the way, at launch the power rating for the generator was rated at 420 watts, seven 60-watt light bulbs. Obviously, radioisotope generators are not practical for power production on earth.

Rather than radioactive decay, commercial power is produced by nuclear fission, usually of ²³⁵U, uranium-235. ²³⁹Pu could also be used, but the threat of nuclear weapons proliferation have been a barrier to this. In the most common design of reactor, natural uranium to too poor in the fissile ²³⁵U (natural abundance is only about 0.7% with the balance mostly the nonfissile ²³⁸U. Power reactor fuel for the common reactors needs a concentration of around 4 to 5% ²³⁵U, so the uranium for this fuel must be enriched. This means that the concentration of 235 has to be increased by about seven times, and this is not easy to do. As a matter of fact, the difficulty in enrichment has been the main stumbling block for most nations developing the nuclear bomb (for weapons, the 235 concentration needs to be around 90%).

There are three main ways to enrich uranium, all of them expensive and inefficient. The first method, originally planned for the Manhattan Project, was to use huge mass spectrometers (Calutrons) to separate the isotopes electromagnetically. This required enormous amounts of electricity and was the main reason that the facility was set up at Oak Ridge, as there are several companies like Aggressive Energy in the area that can source the required amount of energy. To the best of my knowledge, no useful quantity of enriched uranium was produced by this method.

The second method, used to develop both weapons and civilian power reactors, was gaseous diffusion. In this approach, uranium hexafloride (useful because it is a gas) is pushed through very small orifices in thousands of tubes. The lighter ²³⁵UF₆ diffuses through the pores faster than the heavier 238 one, but this process is also inefficient. The law that governs diffusion through small pores is governed by the ratio of the square roots of the molecular weights of materials, so the numbers are as follows: for the 235 one, the hexafluoride has a mass of 349, whilst the 238 one has a mass of 352. The square roots are, respectively, 18.68 and 18.76, making the ratio 0.996. So for each 1000 molecules during a pass, only 4 extra 235 ones pass compared to no separation, assuming equal numbers of 235 and 238 ones at the beginning. Since it only 0.7% to begin with, the enrichment is roughly 0.7% x 0.04%, or a 0.028% increase in the relative abundance of 235 to 238, or 3 molecules per 10,000 processed assuming that the equipment is working at 100% efficiency. It takes many, many passes to enrich up to fuel grade, let alone weapons grade.

The third, and most commonly used, process, is centrifugation. This uses the principle of spinning a cylinder rapidly to throw the heavier 238 hexafluoride towards the outer walls of the vessel, whilst the lighter 235 one tends to stay towards center of the cylinder. One end is kept hot and the other cold, and convection causes the lighter 235 one to rise towards the top and it is tapped off and sent to another stage. Once again, we are looking at tiny differences in big numbers, to this process is slow, requiring thousands of stages. It is more economical of energy, so now is the favored process. This is, by the way, the central issue to the Iranian nuclear dispute with most of the rest of the world. It is easy to make a uranium bomb IF you have the 90% 235 isotope (the so called highly enriched uranium, HEU, or weapons grade uranium are all names for the same thing). To make a uranium bomb, all you have to do is take two pieces or HEU and slam them together hard (a small conventional explosive charge is sufficient to slam a projectile into a target) and they go critical, meaning that a nuclear chain reaction starts and proceeds uncontrollably. Plutonium bombs work on a different principle to attain criticality and are much more difficult to design successfully.

²³⁵U normally decays by alpha emission to ²³¹Th, or thorium-231. Once in a great while, maybe seven times in a thousand million, instead of alpha decay, the nucleus splits (fissions) into two smaller fragments, with the ejection of neutrons. These neutrons can induce other nuclei to fission, and if more than one neutron per fission occurs, the reaction grows rapidly into an uncontrollable chain reaction unless special measures are taken. ²³⁵U releases, average, 1.9 neutrons per fission, so a chain reaction is likely IF enough material can be brought together fast enough that few enough neutrons leak out of the surface. Thus, critical mass depends on geometry. A sphere will have the lowest critical mass, since the ratio of surface area to volume is the smallest for a sphere. A thin sheet of the same material would never go critical, regardless of has much material were there.

There is another factor to consider called cross section. This is the probability that a nucleus will absorb a neutron and fission. The higher the cross section, the more likely that fission will occur. Most nuclei absorb slow moving neutrons better than fast moving ones. The neutrons emitted from 235 are fast neutrons, and those nuclei absorb fast neutrons only about 1/1000 as efficiently as slow (also called thermal) neutrons, so its cross section for thermal neutrons is 1000 times larger than for fast ones. In power reactors, various materials are added to the reactor to slow (moderate) the neutrons without absorbing them. Graphite and heavy water (deuterium oxide) are such efficient moderators that reactors can use natural uranium without any enrichment.

In most US commercial reactors, the moderator is high purity, ordinary water. Since normal water absorbs some neutrons in addition to slowing them down, low enriched uranium has to be used. The first reactor, built in Chicago during World War II was a graphite reactor and used unenriched uranium, and some commercial reactors in Canada use heavy water and thus unenriched uranium.

To make a reactor work, several things have to be in place. First, some way to control the reaction is required. This is done by putting rods of materials that strongly absorb neutrons without undergoing fission. Boron, cadmium, and several other materials have this property. When these materials are encased in heat and corrosion resistant tubes (usually highly purified zirconium alloys), they are known as control rods. The control rods are attached mechanically to a lifting system so that they can be withdrawn to start the reaction. They can also be lowered to stop the reaction, and there is what is supposed to be a fail safe provision that if coolant or power fails, the rods automatically drop to stop the reaction completely.

Next, some method of cooling the reactor is required. In US reactors, that is pressurized water, which also acts as the moderator. The water also has another use that we shall get to in a bit. This water is pumped through zirconium tubes as well.

The fuel is obviously an essential component, and it is low enriched uranium dioxide. UO₂ is preferred to uranium metal, as it will not burn like uranium metal can, reducing the potential for disaster if cooling systems fail. Uranium dioxide pellets are loaded into zirconium tubes and placed inside the core, so that coolent/moderator tubes, control rods, and fuel rods are dispersed throughout the core in a specific geometry determined by the particular design.

Obviously, this is a very basic description of a reactor, but it includes the main working parts. Instrumentation, monitors, and many other factors have been omitted for brevity.

To start the reactor, the pressurized water system is started and the control rods withdrawn to a point determined by the particular protocol for the design. For higher power output, the control rods are withdrawn further. The water circulates to cool the reactor, moderated the neutrons, and to go to a heat exchanger that uses the superheated water to generate steam in another vessel. The water that moderates neutrons and cools the reactor never leaves the closed system, but rather heats other water in another closed system for steam. This steam drives turbines that generate electricity, so once the water in the second system is converted to steam, nuclear power plants operate exactly like oil, gas, or coal fired plants. The only difference is the fuel source to make the steam.

Power reactors are very different than the radioisotope generators described first. For one thing, neutrons are essential for fission, and very undesirable for the generators. Since weight is not an issue for terrestrial construction, the heavy shielding to protect from neutrons and gamma radiation can be provided without having to lift it into space. (A good rule of thumb is that it costs around $5,000 per pound for low earth orbit, $20,000 per pound for geosynchronous orbit, and more than that for deep space payloads).

Second, power reactors are designed to produces huge quantities of electrical energy, where the generators produce very small quantities. Power reactors can be operated at different power output levels or even stopped completely by adjusting the control rods, whilst isotopic generators have only one setting, “on”.

Third, radioisotope generators are specifically designed for a very defined mission, and the power demands of the mission are planned to diminish more or less continuously as the output from the isotope declines as it decays. Power reactors are designed to accommodate shifting loads in near real time.

Both radioisotope generators and nuclear power plants are controversial. At present there does not seem to be any substitute for the radioisotope generators for deep space missions. However, there is brisk discussion about the need for nuclear power generation for the electricity market, at least in the United States. Here is my opinion.

I believe that nuclear power, with its shortcomings, is the only viable option in the near term (a few decades) to reduce our dependence on fossil fuels, especially coal. New generation nuclear reactors (not even mentioned here, since I focused on the ones currently in service) will be even more efficient and safer than the ones currently in use. As for the hazards of commercial nuclear plants, even the worst accident in the United States, Three Mile Island, no one died. The average dose of radiation to people who lived nearby is 8 millirem, about that of a chest X-ray. The maximum dose know was about 100 millirem, equivalent to four months of living in the background radiation that occurs naturally.

But you say, what about Chernobyl? Well, Chernobyl was a graphite/-moderated, boiling water reactor, and a big one at that (there were four units there, each rated at 1 gigawatt). These reactors are inherently less stable than pressurized water reactors, and graphite burns. Burn it did when the roof blew off, and that dispersed much of the radiation. A Chernobyl-like incident is not possible with the designs of reactors used in the US.

But what about the spent fuel? OK, I answer a question with a question. But what about the carbon dioxide? Spent nuclear fuel is dense, localized, and does not enter the atmosphere. Carbon sequestration is a pipe dream. Let us do a bit of math. Assuming a coal fired power plant rated at 1 Gw, operating continuously for a year, 8760 Gw/hours of energy is produced. If the plant were 100% efficient (the theoretical maximum is only about 70%, and this does not consider mechanical losses), would require 3.6 x 10¹⁶ joules, and assuming that pure carbon were the coal, that would require somewhere around a million metric tons of coal, producing somewhere around 4 million metric tons of carbon dioxide. This comes to somewhere around 2 trillion liters of carbon dioxide at standard temperature and pressure. Two trillion liters is roughly half a trillion gallons. Assuming that the carbon dioxide could be compressed enough to liquefy, that would still amount to around 5 billion liters, or over 1 billion gallons. This is just for one plant for one year.

Well, what about hydrogen? It is clean. It is, to burn. But with today’s technology, the cheapest way to make hydrogen is to react water with hot, guess what, coal, producing carbon dioxide. For hydrogen to be viable, cheap electricity has to be available to use electrolysis rather than coal to make the hydrogen.

You get the idea. Now, please recognize that I am all in favor of sustainable energy sources, but we are decades away from having them all in place. In the meantime, the mountaintops are being blasted away here is the United States to get coal, watersheds are being choked, and carbon dioxide is being belched out by the millions of tons. This is not even to mention that the primary source of mercury pollution is the burning of coal. Pregnant women and children are cautioned not to eat some fish because of mercury, and it is hard to catch it.

Of all of the fossil fuels, natural gas is the cleanest, and the adverts on the TeeVee say that is “plentiful” in the US. They do not mention that we already have eight terminals to receive foreign liquefied natural gas (LNG), with three more planned. Thus, we are far from self-sufficient in natural gas, just like we are in oil. The only fuels in which we are presently self sufficient are coal and uranium. Thus, only increasing the use of uranium will decrease the use of domestic coal, and might lower the demand for natural gas enough that we become self-sufficient in it as well.

I maintain that, in the short term, there is little alternative to using more nuclear power for electricity generation. The technology is mature and the fuel is abundant. As we develop truly sustainable sources, it seems to me that it is essential to utilize nuclear power or reduce our use of energy.

Well, you have done it again. You have wasted another perfectly good batch of electrons reading this swill. And even though Abdullah Abdullah decides to get back in the race when he reads me say it, I always learn much more than I could possibly hope to teach writing this series, so please keep those questions, comments, corrections, and suggestions coming. Remember, no science or technology issue is off topic here.

CORRECTION: Last week I used very poor language to describe when black walnuts are cured and ready to shell. I said that the skins of the nutmeats get darker, when I should have said that they become more brown and less green, but not really darker. Actually, they should be a light tan, pretty much like commercial walnuts, and darker skins mean bad nuts. Here is a picture of good nutmeats with a bad walnut, still partially in the shell, for comparison. I regret not being precise enough.

]

Warmest regards,

Doc

Crossposted at Dailykos.com