Of course, there’s always a big uproar over reprocessing because of its proliferative potential. What if security breaks down and we get lots of HEU into the black market? Again, this is a political problem of the “Doctor, it hurts when I do this,” sort. Answer: Don’t do that. Have decent security. The whole argument is approximately equivalent to worrying about somebody breaking into a nuclear weapons storage facility. People did worry about it, and now nobody views the possibility of a break-in as particularly troublesome. So, do the same thing with your reprocessing facilities.

I like the Lftr. The LFTR is a very simple, efficient, and elegant type of reactor. It can use any kind of nuclear fuel, bomb material, or nuclear waste product to produce very high temperature heat and at the same time breed more fuel in the bargain. This thrifty approach to nuclear energy greatly appeals to me, but I became even more interested in the LFTR when the details of a new patent were revealed by Dr LeBlanc. It opens up the possibility of building a reactor that can run for 30 years without refueling in an unattended mode sited underground while it breeds new fuel within the thorium structure of the reactor itself. In order to get to this U233 that has been produced inside the very walls of the reactor containment vessel, a proliferator must destroy the reactor, chop it up into small pieces while enduring heavy gamma radiation exposure without being detected, then reprocess these reactor pieces using isotopic separation since the U233 is denatured with enough U238 to make chemical separation of bomb grade U233 impossible. Now, this is a tall order for any proliferator and may just be an impossible assignment.

At the end of the service life of the Lftr, the reactor vessel is sent back to the factory where it is reduced to liquid fluoride salts that become the feedstock of a next new Lftr. This feedstock can only be used by the new Lftr and not for bombs. The waste products are held at the factory for a few hundred years to cool down before they are mined for the many precious elements contained within like platinum and iridium. Now that’s what I call a safe, efficient and thrifty mode of operation.

I think you’re both right on the mark, and I can’t tell you how refreshing (and rare) it is to read clear and informed commentary on nuclear power. Let me contribute a bit further on the subject of proliferation resistance.

As TRM says, the if you’re worried about the security of plutonium during reprocessing then you just have to pay for the security you need. But even better would be not to produce any plutonium, or other weapons-usable material, in the first place. This is one big advantage of the thermal thorium breeder cycle that Anon. mentioned earlier. Breeding Th-232 into fissile U-233 and using that as fuel results in essentially no other actinides being created. This means that there’s no americium or curium and so very little long-lived high-level waste; but it also means that there’s no plutonium either. So with the thorium cycle you don’t have to re-process and burn away the plutonium, you never have to worry about it at all.

It is technically true that U-233 can be used to make weapons, but there are many reasons why this is so difficult in practice that it’s not (to my knowledge) generally considered a proliferation risk. A main factor is that U-233 that is bred from thorium will typically contain contamination by U-232, which is highly radioactive with a short lifetime and a high-energy gamma decay in its decay chain. This makes the stuff very difficult to handle outside a facility like a reactor, and it also creates pre-ignition problems that would make the U-233 ineffective for weapons. In his second bomb book _Dark Sun_, Richard Rhodes explains that the US did mange to make fission weapon cores from U-233 in the 1950’s but standardized on plutonium as being much easier and more reliable. And this is the main reason that U-233 is not considered a proliferation risk: since no one regularly uses it to make weapons there are no plans you can readily buy or steal, and there’s no one you can hire who immediately knows how to do it.

So the thermal thorium breeder cycle is really something of a win-win: vastly less long-lived waste, and essentially no proliferation potential. Why you don’t hear about it more often is a longer story, for a longer comment.

Of course, there’s always a big uproar over reprocessing because of its proliferative potential. What if security breaks down and we get lots of HEU into the black market? Again, this is a political problem of the “Doctor, it hurts when I do this,” sort. Answer: Don’t do that. Have decent security. The whole argument is approximately equivalent to worrying about somebody breaking into a nuclear weapons storage facility. People did worry about it, and now nobody views the possibility of a break-in as particularly troublesome. So, do the same thing with your reprocessing facilities.

The IFR concept has BOTH components: the sodium-cooled fast reactor, and reprocessing based on electrorefining of molten metals. Beyond the current system of any extant fuel cycle, it not only separates transuranics but DESTROYS them completely. The very heavy nuclei, having half-lives in thousands of years, are fissioned by high-energy neutrons, yielding short-lived fission products (decades or less). Any remaining, unburned actinides are partitioned out and sent back to the reactor – it’s a closed cycle. In this way, the waste does not need storage beyond a couple of centuries.

IFR is not the only idea – there’s others, though none have been commercially realized. (Closed fuel cycles are more resource-intensive, hence costly, than conventional “once-through” cycles; there’s no economic incentive to reprocess when mined fuel is cheap.) For example, fast reactors can be based on working fluids other than liquid sodium. Water is not an option, because it is an effective neutron moderator and will thermalize the fast neutrons that you need. A lead-bismuth eutectic is one alternative – it is currently used in the reactors of Russia’s Alfa-class nuclear submarines. Or a gas coolant is possible – the German THTR-300 reactor used helium cooling, which allows very high temperatures to be reached (liquid metals are limited by corrosion issues, I think). And if you switch to a thorium fuel cycle, you don’t even need fast reactors: Th-232 can sustain a breeding cycle (to fissile U-233) in the thermal neutron spectrum, which allows for moderated reactors like pressurized heavy-water reactors (PHWR) under development in India (they are closely related to CANDUs), or a liquid thorium fluoride reactor (LFTR), tested at Oak Ridge in the 1960’s.

So the options are very broad – fast or thermal reactors; chemical or electrochemical reprocessing; uranium, plutonium (MOX), or thorium fuel – all are candidates for a closed fuel cycle. All closed fuel cycles share the key benefits in common: no long-lived transuranic waste, no need for geologic waste storage, and fuel efficiency two orders of magnitude higher than current-generation reactors.

James Hansen wrote a letter to Obama endorsing these closed nuclear fuel cycles. It is hosted on his academic page at Columbia University (I can’t link to it or the spam filter will likely delete my comment, for having links in it).

Virtually all of the innovative work in alternative energy has been performed by individuals with minimal funding. How can we create more of them?

Anon has presented you with the Integral Fast Reactor as one option. The French, I believe, recover uranium and plutonium through a PUREX process, which leaves a much smaller amount of extremely high level waste. Don’t want to reprocess? Then there are plenty of techniques, like vitrification, that can stabilize high-level waste for very long periods of time. Perhaps not for a million years, but a pretty long time. And, if we survive as a civilized species, I’m sure that future generations will be somewhat less recto-cranially inverted than we are currently, and they’ll want to dig up all this “waste” and put it to good use. Lots of things you can do with radioactive stuff, and it’s hard to refine.

But let’s put all that aside for a moment and consider exactly how much high-level waste we’re talking about here. I found some EIA numbers that indicate that, in the US, we used 47,000 metric tonnes of uranium from 1968-2002. But we’re interested in a span of time from the early 1950s through, say, 2200, and there’s other low-level waste, so let’s multiply that number by, oh, I don’t know, 50, and say we’ve got 2,350,000 tonnes of pretty nasty stuff to dispose of. Let’s assume we did absolutely nothing with that waste except to bury it. How much land would we pollute?

Well, the approximate density of uranium is 19.1 grams per cubic centimeter, a million grams to the tonne, which means that we have 123 billion cubic centimeters of waste, which is 123,000 cubic meters of waste at a million cc’s to the cubic meter. In other words, if you took all the waste for 250 years and melted it down into a single, giant cube, that cube would be 50 meters on a side. It would fit comfortably into a football stadium.

Of course, if you put all that spent fuel in one big cube, you’d have a terrible mess, because the whole thing would go critical and melt. So you have to parcel it up into nice little bite-sized chunks before you bury it. Let’s say that each chunk is 0.1 cubic meters. Let’s further say that each chunk has to be buried 2 meters away from every other chunk. By my reckoning (and it’s always scary doing this in front of physicists) that means that you’d need 4.9 million square meters to bury all the waste.

But, at a million square meters to the square kilometer, that’s only 4.9 square kilometers, or a square 2.2 kilometers on a side.

Now, there’s no doubt that, by just burying this stuff, we’ve totally trashed those 5 square kilometers. We have also trashed the surrounding water table, so maybe we’ve actually trashed an area that’s 50 or even 100 square kilometers. And you’d definitely want to avoid having this stuff leech into an aquifer, or blow away in topsoil, or a whole bunch of other things. So let’s all agree that we’d like to a better job than tossing the stuff in a bunch of holes.

But look how tiny the problem is when you actually do the arithmetic. On a grand scale, we’re just not talking about a lot of space to store the stuff. This is why I say that waste disposal is largely a political problem, not a technical one. Sure, it’ll take multiple tens of billions of dollars to address, and we’re really going to piss of a bunch of NIMBYs and BANANAs. But it’s not a technically difficult problem.

I’d like to compare the impact of radioactive pollution to that of semiconductor pollution, or the amount of land taken out of service by new power lines to service wind farms, but I’ve already gone on way too long. Radioactive waste is just another poison, like the arsenic and gallium used in semiconductor manufacture. But it’s not anything to get worked up over.

I’m not opposed to nuclear power in principle; but some of those nuclear wastes need to be isolated from contact with humans for hundreds of thousands of years. All human written history (estimated generously) does not span more than 5,000 years (and I’m including characters on Chinese kettles). Over that period of time, we cannot count on continuity of civilization, culture, language, etc., etc.

So how do we protect future generations against exposure to radwastes over that span of time?