The Case for Nuclear Power

Say the words "Three Mile Island" or "Chernobyl" and practically anyone will know what you mean. Chernobyl was the worst nuclear power plant accident ever, killing more than 30 people immediately (and an estimated 4,000 due to delayed effects of radiation) and forcing a 20-mile radius around the plant to be evacuated due to high radiation levels. Three Mile Island did not become such an accident mainly because its design, unlike that of Chernobyl, included inherent safety features, the most notable of which was a "secondary containment" structure around the reactor, which kept the dangerous radiation from escaping even after the reactor core had partially melted down. But the lack of secondary containment was not the only design flaw in the Chernobyl reactor--it also did not have the same control features as the TMI reactor, so that the reactor actually created explosions instead of just melting down as TMI's did.

These differences between TMI and Chernobyl are interesting, but not because they show how important it is to engineer safety features into a reactor. In fact, they are interesting for precisely the opposite reason: because they show how, even if you take the trouble to design and build in all those safety features (secondary containment, guaranteed controllability of the reaction), so that when a practically worst-case accident happens, it is still contained and doesn't harm the general public, that won't save the technology. Not that the people who caused the TMI accident should be exonerated--but that accident was the beginning of the end for nuclear power in the US, even though the reactor's safety features did their job and kept the public safe. And then the Chernobyl accident put the final nail in the coffin, even though its reactor design would never have passed muster in the US or any other country except the Soviet Union during the Cold War.

Perhaps that was a defensible decision in the 1980's, when we didn't yet understand the full environmental impacts of our other energy sources--coal, oil, natural gas, hydroelectric, wind power, solar power, biomass, geothermal etc.--and when we didn't fully realize what would be required for a sustainable future for humanity. But in the last couple of decades we have gotten a much better handle on how much energy is required to give everyone in the world a good standard of living, and we have learned a lot more about the impacts of the energy sources that have been proposed to get us there. We now understand that none of those energy sources are a free lunch; every one of them has downsides. Here's a short list of the biggies:

Greenhouse Gas Emissions: We don't know for sure what the impact of this is going to be, but that's the point (I make this argument more fully in another article in this category). All carbon-based energy sources--coal, oil, and natural gas--have this downside. (Biologically derived energy sources such as wood or biodiesel do emit CO2, but the plants that are used to produce them absorb it, so the net impact is essentially zero.)

Water Usage: This is the big downside of hydroelectric power, which used to be touted as the obvious "green" energy source. Not only does it impact river levels, which in the Southwest US, the area with the largest hydro power capacity, is no small thing, it also impacts fish and other wildlife (see next item).

Wildlife Impact: As just noted, hydroelectric affects fish and other wildlife that depend on river ecosystems. Wind power affects birds (because they fly into the wind generators). Coal, oil, and natural gas, which have to be extracted from the ground, can also affect wildlife who happen to live in the extraction area. And biomass (particularly wood) affects wildlife by altering habitats such as forests.

Land Use: All power plants take up some space, and so do the plants for extracting and refining their fuel (if applicable). But the traditional "green" alternatives, particularly wind power, take up a lot more (and this comparison gets even more lopsided when you factor in generating capacity--see below).

Radiation: You might be surprised to see this one anywhere except nuclear, but in fact coal (the number one energy source in the US and many other countries) and its waste products can produce more radiation over the lifetime of a power plant than a nuclear plant of similar capacity. This is because coal, like everything you dig out of the ground, contains natural radioactivity, and the process of refining and then burning the coal concentrates it. This is important because coal is the easiest to mine, easiest to refine, and easiest to burn of the fossil fuels, which means that it's the most attractive option for developing countries that want to expand their economies--but factoring in this impact can often change the equation vs. nuclear power.

Low Capacity: This is the real kicker for the "greenest" energy sources usually favored by environmentalists--solar, wind power, and geothermal. A typical wind farm might have a capacity of 1 megawatt (MW), which is enough to power a few hundred homes. A typical geothermal plant might be about the same. There is no real "typical" solar power plant, but a typical solar panel that you can install on the roof of your house might supply a kilowatt (kW) for a few hours a day if it's sunny--over the course of a month that might amount to one percent of the electricity your house consumes (depends on the size of your house and what appliances you have), and only if you live in an area that gets enough sunlight (the Southwest would be good, the East or Midwest not so good, and someplace like Seattle very poor indeed). By comparison, a typical coal, oil, natural gas, or nuclear power plant might have a capacity of 1000 MW, enough to power a city, while probably taking up less land area than the wind farm (and about the same as the geothermal plant). So when you talk about generating power on the scale of a country or the world, doing it with the traditional "green" energy sources is going to take several orders of magnitude more plants if it can be done at all. That will make all the other impacts above even worse, and may well make them unmanageable.

You will note, of course, that every alternative to nuclear power has some significant downsides on the above list. There really is no free lunch. So what are the comparable downsides of nuclear power? Let's look at the big ones:

Radiation: To be consistent with the above, I'll confine this to "low-level" radiation from concentrated natural radioactivity. I compare it with coal above. None of the other sources have it. Dealing with it requires safe containment of the fuel, the reactor, and the waste products.

Nuclear Waste: I keep this separate from "radiation" because this is a specific problem of nuclear power (some would say the specific problem of nuclear power). The waste products from fission reactions are highly radioactive, so spent fuel needs to be contained properly; and since some of those products remain dangerously radioactive on a time scale of thousands of years, the containment apparently needs to be very, very reliable (but see below for some alternatives on this issue). Also, these waste products could be stolen by terrorists and used (at least a portion of them) to make weapons (though this isn't a simple process or a very efficient one, so it wouldn't necessarily look like a very attractive option to a terrorist).

Reactor Safety: The problem that was brought to the fore by TMI and Chernobyl. It's worth noting, however, that fossil fuel power plants are also vulnerable to accidents--the major additional risk of nuclear plants is the release of high-level radiation to the environment. In recent years this aspect, like the previous one, has also been considered in the light of the risk of a terrorist attack.

All of these are significant risks--but did you notice something about them? Unlike the risks I listed above for the other energy sources (except for radiation, which as noted is a risk common to nuclear and coal), all of the risks involved with a nuclear power plant are susceptible to engineering solutions. That is, if we're smart enough, we can figure out ways to reduce the nuclear risks to acceptable levels (you can never reduce a risk to zero--refusal to accept this fact is probably the biggest single problem with the more extreme wing of the environmentalist movement). By contrast, no matter how smart we get, we're not going to be able to burn fossil fuels without producing CO2 (we could capture the CO2 instead of letting it into the atmosphere, but even if we got very smart about that we'd still run out of fossil fuels before too long, and we aren't going to get smart enough to replicate the process that made them in less than the millions of years it took to do it the first time). We won't be able to build a hydroelectric plant that doesn't use water, or build a wind farm that doesn't take up a lot of space or has the capacity of a typical fossil fuel or nuclear plant. But we are already smart enough to deal with all of the nuclear energy downsides above, and we're continually getting smarter.

For example, consider the nuclear waste problem. The solution we usually hear about for this is to package the waste in several layers of protective material and then bury it underground in a geologically stable place (a place that is expected to remain stable for hundreds or thousands of years). This solution has taken a long time to get off the ground in the US because of difficulties in getting agreement from all parties to use the Yucca Mountain site in Nevada (formerly a site for underground nuclear testing--the difficulties have mostly been political, not technical), so in the meantime the waste is stored onsite at the power plants, in concrete and steel containment vessels and/or immersed in pools of water to contain the radioactivity. The US government now (December 2004) thinks the waste may be able to start moving into Yucca Mountain by 2010. There are also other geologic disposal ideas being considered, such as burying the waste much deeper down (a few kilometers) to ensure that it's completely removed from the surface biosphere.

Geologic disposal might not seem like the best way to deal with nuclear waste, since it's very hard to convince yourself that we are really capable of finding a site that is guaranteed to remain stable for thousands of years. But we'll continue to get smarter about these things--in a hundred years we may have a much better understanding of how to find such a place, and even the storage methods we have now are good for a few hundred years. But it would be a lot better if we could concentrate the waste at one site (Yucca Mountain, or even a single above ground site until Yucca Mountain is ready) instead of having it spread out over more than 100 plants around the country. The main reason we haven't is that political considerations, combined with an unreasonable insistence on a guarantee of essentially zero risk before taking any action at all, have prevented it. It's simply a refusal to accept that good engineering is good engineering--not a guaranteed cure-all, but a reasonable solution to a problem that will work until we get smart enough to think of a better one.

There are also plenty of other good ideas for dealing with nuclear waste. For example, we now are much better at "reprocessing" nuclear waste. In a typical current nuclear reactor, only about 5 percent of the uranium-235 in the fuel is actually consumed during a fuel cycle; reprocessing removes the rest from the spent fuel so that it can be recycled into new fuel and used. The word "recycling" is not just an attempted nod to environmentalists--it really is what's done to make efficient use of the U-235, as well as to mitigate the waste problem. (This was actually part of the long-term plan for nuclear energy from the beginning, but it took time for all the technical issues to be worked out, and then politics came into play--see below.)

Reprocessing also removes the plutonium from the spent fuel--plutonium is formed from uranium-238, which does not undergo fission but comprises most of the mass of the fuel (enriched uranium fuel for reactors is only a few percent U-235). Plutonium can also be used as a fuel in reactors once it's recovered by reprocessing; this is a much better way of handling it than handling it as waste because without it the waste will remain dangerously radioactive for only a few hundred years instead of thousands (which means that we can adopt waste disposal solutions that only have to work for a much shorter time, and we have plenty of those and will discover more as we get smarter), and the plutonium itself gets consumed in the reactor and is no longer a problem. (Since the plutonium is also the portion of the waste that could be used to make weapons--though as I noted above, that's not a simple process or a very efficient one--removing it and using it as fuel is the best way to prevent it from getting into the hands of terrorists.)

An even better way to do this is to build a "breeder reactor", which is designed to make plutonium as it consumes U-235, and then recycle the plutonium into the reactor to be used as fuel. This makes the fuel virtually inexhaustible (there is enough U-238 on Earth to provide our energy needs for more than a billion years), in addition to simplifying waste management. We'll only continue to get smarter about this as time goes on--a recent article in Scientific American suggested that we may actually find uses for this "waste" material in another 50 or 100 years. (I should note that current US policy does not allow reprocessing to be done in this country; this was a political decision made by the Carter Administration in 1977, for reasons that are explained in the Frontline link below. Carter was a nuclear engineer, but that doesn't necessarily mean that policy was a good idea. No other country that uses nuclear power has this prohibition on reprocessing.) We may or may not discover such uses, but we certainly will continue to get smarter about how to deal with nuclear waste--engineers will continue to discover solutions, some of which may even be better than reprocessing or deep geologic disposal.

Similar remarks apply to the reactor safety issue. As I noted above, we knew from the first how to build a reactor that was safe when operated properly, and would contain an accident even if operated improperly (as happened at TMI--the Soviets simply ignored those safety features when building Chernobyl). But now we know how to design reactors that use the laws of physics as safety features; it's physically impossible for them to melt down or otherwise catastrophically fail. They still require operator expertise to operate properly, but there is simply no way to operate them improperly to cause an accident like TMI (much less like Chernobyl). The IFR (Integral Fast Reactor) and the Pebble Bed Reactor designs are two such advanced designs that have been prototyped, and we'll continue to get smarter about that too.

Once again, I didn't say we are perfectly smart--I didn't say that we are smart enough to know every possible thing that could go wrong. We're not. But that's true of every technology we use. As I noted above, it is impossible to make anything completely risk free. What we can do is apply engineering solutions to reduce risks to acceptable levels, but as I noted above, that works for some risks but not for others (because of inherent physical limitations such as the chemistry of fossil fuel burning or the limited energy available from wind by the very nature of wind). What we can also do is assess relative risk, which is something that is rarely done in the US with energy sources because political or ideological factors get in the way. Relatively speaking, nuclear energy is the safest source known for its capacity: the number of lives lost or people injured by nuclear energy is far less than the number killed in coal mines, for example. (And to put the Chernobyl numbers in perspective, more than 50 people are killed each day in the US in traffic accidents, and a few thousand die from drug-related incidents in the US each month--so this country suffers dozens to hundreds of Chernobyls every year from other causes than energy sources in general.) Also, as the comparison of downsides above makes clear, nuclear energy's environmental impacts can be mitigated by engineering, where those of other energy sources cannot.

Other developed countries have recognized all this and, unhampered by the political and ideological factors at work in the US, have steadily increased their reliance on nuclear energy. If the US really wanted to reduce its dependence on foreign oil, we could, in perhaps 20 years, the time it took France to go from zero nuclear capacity to its current level, get to a similar level of nuclear energy capacity (75% of demand). (Yes, nuclear plants can be built that fast. What takes so long in the US is not the actual construction but the lawsuits that have to be finished before construction can start.) No "renewable" energy source is going to come close to that (claims by environmentalists notwithstanding--see "Low Capacity" under downsides above).

The issue is even more pressing in the developing world. China's coal usage already exceeds that of the US, despite their efforts to expand their use of nuclear energy. If we really wanted to reduce greenhouse gas emissions, we could start encouraging China, India, and the rest of the developing world to build safe nuclear power plants. There's no other "green" way for them to generate the energy they need to give their people the standard of living they want. Telling them that, well, they just can't have all that energy and the standard of living that goes with it (which is basically what the Kyoto Protocol, now apparently all but dead, tried to do) is not a policy that seems to be meeting with much success. (Nor is it a cogent worry that these countries might use plutonium from the reactor waste products to make weapons. China and India already have nuclear weapons.)

So there it is: nuclear power can meet our energy needs and, with proper engineering, do it safely and without significant environmental impact. Every other energy source either (a) can't meet our needs (not enough capacity) or (b) has a significant environmental impact that can't be mitigated by engineering. Please note that this doesn't mean we should use nuclear power exclusively--there is no reason why we can't use other energy sources to supplement our energy needs. We just can't depend on them to meet all of our needs. There's also good reason for us to use energy more efficiently--we just can't depend on that alone to get us there. Perhaps it's just not easy for non-engineers to believe this; the case of TMI certainly indicates that, at least in the US, even good engineering that does its job and protects the public isn't always recognized as it should be. But we're coming to a point where the choice is getting harder and harder to avoid: either recognize that we need nuclear power to meet the world's energy needs--and let the engineers loose to find reasonable solutions to make the risks acceptable--or don't meet them. And the latter is really not an option.


I should recognize that there is also research being done into what is called "solar thermal" power--that is, converting sunlight into electricity by using it to heat up something instead of directly through a photovoltaic cell. The heat source is then used to drive a steam turbine, similar to "conventional" fossil fuel or nuclear power plants. Solar thermal power can have significantly more capacity than what I was referring to as "solar power" above (photovoltaic cells), because of the extremely low efficiency (only a few percent) of photovoltaic cells. Since the substance which is heated up by sunlight (the light is first concentrated by mirrors, similar to the way you may have set a piece of paper on fire using a magnifying glass as a kid) can store the heat as well as use it to drive the steam cycle, a solar thermal plant can continue to provide power if it's cloudy or during the night.

Solar thermal still suffers from one major downside of solar power in general, however--it is still best used in a specific location, because it still needs to have enough sunlight averaged over time. That means you can't build a solar thermal plant close to the center of power demand, as you can with fossil fuel or nuclear plants; you have to build the solar thermal plant where there's enough sun, and then send the power through the power grid to where it's used. So supplying, for example, the entire US with solar thermal power by building enough capacity in the deserts of the Southwest (which could be done, though it would require a very large land area--not impossibly large, but very large) would also require a virtually loss-free power grid throughout the country--which would mean a huge investment in superconducting power cables and fault protection systems for them, since if the superconductivity was lost through a system failure there could be blackouts over the entire country.

Still, we may eventually make such an investment in the grid anyway (because a loss free power grid makes any method of power generation more efficient). In any case, this example illustrates that, unlike the downsides to other non-nuclear energy sources that I discussed above, those of solar thermal power (except for the land area required, which can't be avoided) may be amenable to engineering solutions, which will get better as we get smarter. So it may be that solar thermal power is another potential solution to supplying our future energy needs, either in combination with or (much less likely, in my opinion, because of the land area that would be required for that much capacity) instead of nuclear power. As I said in the above article, I don't think we should adopt nuclear power to the exclusion of all other energy sources. I just think we need to recognize the real energy needs of the world, and their magnitude, and focus on sources that can supply power in that magnitude, without letting politics get in the way.

Solar Thermal Fact Sheet: Some information about solar thermal power.


The following is a collection of links that are relevant to this article. As I find more, I'll continue to add them here.

Nuclear Reactor at Wikipedia: The online free encyclopedia entry. A good general overview of the technology, with lots of useful links to follow.

Nuclear Power at HowStuffWorks: Another good general overview.

World Nuclear Association: A good site for information about all aspects of nuclear power, of which I only discuss some in this article, as well as good comparisons with other energy sources.

Frequently Asked Questions About Nuclear Energy: By John McCarthy, Professor Emeritus of Computer Science at Stanford (for you computer and AI types, yes, this is the John McCarthy). Good information and good arguments for why we need nuclear energy, from someone who isn't part of the nuclear energy community (and so has no vested interest in the growth of nuclear energy).

Chernobyl: Another page from John McCarthy's site. This is a good brief summary of why the Chernobyl accident was as bad as it was.

Overview of Nuclear Energy: A summary page at an Australian web site about nuclear power and how it compares with other energy sources, with links to "briefing papers" giving more details. The "sustainable energy", "economics", and "waste management" briefing papers are particularly interesting.

The Future of Nuclear Power: A study released in 2003 by MIT. It deals with both the advantages and the risks of nuclear power, and also gives balanced comparisons with other energy sources.

One note about this report: you may notice that their position on reprocessing is the opposite of the one I favored above. One of the main reasons given for this is that, in their opinion, reprocessing entails an unacceptable increase in the risk of proliferation (which is essentially the argument that was used in the Carter years to stop it, and in the Clinton years to keep it stopped). This is open to question, since Europe and Japan have been reprocessing for decades (in fact, Japan ships its spent fuel all the way to Europe and then gets the reprocessed fuel back) and there have been no incidents. However, the MIT report does not give proliferation as the only reason not to do reprocessing: it also talks about cost and fuel cycle safety as reasons not to do reprocessing.

I'll leave it to you to sort out these claims (but I would recommend reading the other reprocessing links below before trying to do so), but it should be noted that the MIT report doesn't pitch their arguments as a reason to permanently prevent reprocessing, but only as a reason to delay doing it until nuclear fuel becomes expensive enough to make it economically preferred. This type of consideration was not, as far as I can tell from the sources linked to here, a factor in the Carter administration's decision to not do reprocessing.

A Frontline Interview: Another "Frequently Asked Questions", but this is with a nuclear expert. Good debunking of some common misconceptions.

A Frontline Excerpt on Reprocessing: Includes two opposing viewpoints about the US policy decision not to allow reprocessing.

Political Will and Nuclear Waste Storage: An article at the "Blowhards" web site about the waste issue. Despite the site's name, this is a good discussion of the key points.

Usenet Archive on Reprocessing: A brief discussion thread from Usenet (circa 1994) that gives another interesting look at the forces arrayed against reprocessing in the US.

Usenet Archive on Three Mile Island: Another discussion thread from Usenet (circa 1999) about how the TMI incident got blown way out of proportion.

IFR (Integral Fast Reactor): An advanced breeder reactor design that could reprocess its own spent fuel and re-use the plutonium from it, greatly simplifying waste management. It also had "passive safety" features similar to the pebble bed design. It was canceled by the Clinton Administration in 1994. (This link appears to be broken.)

The Pebble Bed Reactor: A good brief article on the pebble bed design at, including a comparison of the pebble bed design with "conventional" reactor designs.

Pebble Bed Modular Reactor The First Generation IV Reactor To Be Constructed: A paper at the World Nuclear Association's 2003 Annual Symposium. Good summary of the goals, features, and safety aspects of the PBMR design.

Rethinking the Nuclear Option: An article at CNN.COM which talks about the Koeberg pebble bed reactor in South Africa.

Pebble Bed Reactor Technology: A page on the Eskom site (the South African company that is building the Koeberg reactor). It includes a schematic of the design.

Let a Thousand Reactors Bloom: An article in Wired magazine about China's push to adopt pebble bed reactors.

MIT, Tsinghua Collaborate on Development of Pebble Bed Nuclear Reactor: MIT's announcement of its joint project with Tsinghua University in Beijing (mentioned in the Wired article above).

What's Wrong with the Modular Pebble Bed Reactor?: The Three Mile Island Alert website opposes the PBMR design. The issues they raise are addressed at some of the other sites linked to above (the Australian site with the "briefing papers", for example, though these were not written as a response to the specific criticisms given at this site). This page also links to other web sites which give much the same criticisms (and they aren't the only ones, as a Google search will reveal).

Wind and Hydropower Technologies: A page on the US Department of Energy website about wind power and hydropower. Recent (as of June 2006) estimates of the available capacity for wind power in the US seem a lot higher than my pessimistic estimates in the above article, which, if the estimates pan out, would make wind power a significant addition to the mix of energy technologies. (The American Wind Energy Association also has a web site that gives similar estimates of potential capacity.) I'm still skeptical of these estimates because I haven't been able to find any details about how they were done or any actual math behind the numbers. The basic physics of a wind turbine is well understood, including its power production as a function of wind velocity, so it ought to be reasonably simple to calculate available power given enough data on measured wind velocities over time, and presumably that's where these estimates come from. However, it would be nice to see the actual supporting data and the methodology laid out explicitly.