Pique the Geek 20110403: How Nuclear Reactors Work. Part the Third

(9 pm. – promoted by ek hornbeck)

The past two weeks we have been looking about current nuclear technology for power production.  Almost all of the plants currently in commercial production are the so called Generation II and Generation III plants, and the most advanced ones are sometimes called Generation III+.  This evening we shall examine nuclear technologies that are not yet used for commercial production.

These designs are called Generation IV plants, and may either be prototypes or merely designs that have not yet even had a prototype built, but appear to be feasible to come on line commercially by 2030 or so, give or take.  There is also a Generation V set of concepts, but they are much further out as far as construction of even a prototype in concerned, and we shall not consider them here.

Before we get going on that, the situation in Japan has continued to be bad.  Two workers have been found dead, presumably killed by the tsunami and not by radiation exposure.  One of the concrete retaining walls has a huge hole in it, allowing highly radioactive water to escape into the Pacific Ocean, presumably caused by the earthquake itself.  Some recent reports also suggest that localized critical events are occurring, as evidenced by flashed of a bluish light around the containment building.  This is particularly grave, because extremely high fluxes of gamma rays and neutrons that are produced during those events.

I do not mean this to be a comprehensive account of the events in Japan, and there are a couple of blogs already that do a much more complete job of that coverage.  I do mention it because of the impact that this terrible situation may have on the future of nuclear power, most likely delaying the startup of all planned plants, and also the construction of Generation IV plants.  This is unfortunate, because Generation IV plants, at least several designs of them, would have made the situation seen in Japan much less likely.

Generation III plants have three serious problems.  The first is the requirement that a high pressure vessel (a single forging by the way, no welds or rivets except for the access cover) be used to house the core of the reactor due to the extremely high pressure of the steam produced by the water that acts as moderator, coolant, and working fluid.

The second is the problem of highly radioactive actinides (covered last time) in spent fuel rods.  It is, for the most part, these isotopes that are responsible for the high radiation density and long hazardous lifetime of spent fuel.

Third is the requirement that water be pumped continuously through the reactor core to prevent overheating.  Most of the problems in Japan are due to the failure of the circulation system due the the Diesel engines that were supposed to run the backup generators to supply electricity becoming hydrolocked due the the tsunami.

Many Generation IV reactor designs address one of more of these fundamental problems.

The most well established design is the Very High Temperature Reactor (VHTR), which is actually a Generation IV derivative of the Generation III High Temperature Gas Cooled Reactor (HTGR).  The concept goes back almost as far as Fermi’s very first nuclear reactor.  This is a graphite moderated unit, essentially a requirement because only graphite combines the temperature robustness and moderating capacity required.  In this design, the coolant/working fluid is circulated at near atmospheric pressure through cavities in the moderator.

The coolant/working fluid most often used is helium, but molten inorganic salts can also be used.  The helium cooled ones are designed to operate such that the helium is released at around 1,000 degrees C, whilst the molten salt ones operate at around 1,500 degrees C or so.  One major advantage is that the coolant/working fluid is pretty much at atmospheric pressure, so that an explosion from failure of the coolant containment system is not possible is it is with water cooled reactors.

Another major advantage is that the high temperatures make the production of hydrogen from water, though a rather elaborate stepwise process, possible.  Water cooled reactors do not get the working fluid hot enough to drive this reaction.  This allows the same reactor both to produce electricity and hydrogen for use in internal combustion engines.  If you are a long time reader of this series you will recall that I debunked the hydrogen economy a couple of years ago by pointing out that most commercial hydrogen today is actually produced from fossil fuels, with carbon dioxide the major waste product.  The thermal production of hydrogen from water is more efficient than electrolysis, so could be a real competitive advantage.

The fuel most often mentioned for this type of reactor is either enriched uranium dioxide (or carbide) or MOX (mixed uranium and plutonium oxides) in the form of TRISO pellets, very small beads with fuel in the center, a layer of porous carbon next (to absorb gaseous fission products), then a layer of hard carbon, a layer of silicon carbide, and finally another layer of hard carbon.  These beads are extremely robust and extremely high melting, such that failure of coolant will not cause them to deform and release the actinides and fission products contained therein.  This is huge advancement over the current fuel rod design.

The beads can be used in two different ways.  In one configuration they are loaded into zirconium alloy fuel rods, much as uranium oxide pellets are loaded in Generation III units.  These fuel rods are inserted in a similar manner, and control rods are used much like Generation III reactors.  In the other configuration, the TRISO beads are bonded in hard graphite into spheres a little bigger than a golf ball.  These “pebbles” are simply poured onto a bid of the proper geometry and become critical.  In addition, there is a neutron reflector surrounding the bin that reduces the amount of pebbles needed to maintain criticality, and in most designs the control rods are integrated into the reflector.

This design, called the pebble bed configuration, is very much safer than current Generation III designs.  As a matter of fact, in a prototype model, the core was filled, helium flow started, and the control rods removed until the unit came up to full power.  Then the flow of helium was stopped.  The reactor acutally cooled down, because of Doppler broadening of the neutron flux.  Basically, this results from the thermal energy NOT being removed from the reactor core, causing some of the neutrons to speed up and so be less apt to induce fission in the uranium-235 fuel.  Fast neutrons, however, are apt to be absorbed by uranium-238, which does not result in fission, but rather, ultimately the production of plutonium-239.  This throttles back the heat output of the reactor, making this design much safer than water cooled reactors.  I do not like the term inherently safe, because nothing designed by humans is perfect, but this is a much safer design than Generation III reactors.

The major disadvantage of the VHTR designs is that the nuclear fuel is designed for a single pass, meaning that, like conventional Generation III reactors once the fuel in the TRISO pellets is used, they are no longer useful and must be put into long term storage, a real problem with current reactors.

Another type of thermal neutron reactor is the Molten Salt Reactor (MSR), and these have already been prototyped.  In these reactors, the fuel (generally uranium-233 for some important technical reasons) is dissolved as uranium tetrafluoride in lighter inorganic fluorides, such as lithium fluoride.  At startup, the salts have to have outside heating until they melt, but then can be pumped at low pressure into a graphite moderator chamber, where the thermalization of neutrons cause the fuel to go critical, thus producing its own heat.  Since they are under low pressure, they can be made rather small and light as well.  As a matter of fact, the DoD looked at them in the 1950s as a potential power source for aircraft, but that did not pan out very well.

In addition to being under little pressure, MSRs can be used as slow breeder reactors, producing more fuel than they consume.  It turns out thorium-232, the most common isotope of thorium, and be converted to uranium-233 by thermal neutrons.  In contrast, uranium-238 is converted to plutonium-239 by fast neutrons, making fast breeder reactions more problematic, because of the relative difficulty in controlling fast neutrons.  In a MSR, the thorium can be added to the fuel, or used as a “blanket” around the reactor core.  The latter has some advantage, because not only are neutrons that are normally “wasted” used to produce more fuel, the thorium blanket provides significant neutron shielding for personnel and other plant materials.

However, they are not as simple as the pebble bed reactors described earlier.  Molten fluoride salts are rather treacherous, and water and fluorides do not play well together.  In addition, exotic alloys have to be used to resist the corrosive nature of the molten fluoride salts.  There is one significant advantage, though:  the amounts of long half life products is much, much lower than those produced by Generation III reactors and the VHTRs described earlier.  Thus, spent fuel would need special storage only for hundreds of years, in contrast to many thousands of years for conventional spent nuclear fuel.  Besides producing more fuel than they consume, the reduction in long half life products make this a very viable candidate.  In common with VHTRs, it is thought that MSRs can be operated at high enough temperatures to produce hydrogen, another advantage.

Another type of Generation IV is the Supercritical Water Reactor, SWCR.  I am extremely dubious of this technology.  One of the problems with Generation III reactors is the high pressures in the core, and supercritical water is both very hot and under very high pressure.  A supercritical fluid is a fluid that has been subjected to a temperature that is above its boiling point at a pressure that prevents boiling.  For water, those parameters are 374 degrees C (not that extreme), and 218 atmospheres (THAT is pretty extreme, around 3200 psi, much like that in a compressed oxygen cylinder).  These are minimum values, so they can go higher.  The efficiency of SCWRs should be higher than subcritical ones, because of higher temperatures, although not nearly as hot the reactors mentioned above.  Supercritical “boilers” are already being used in the fossil fuel power generation business, but the technology is much easier to apply when heat is the only hazard.

I used to “own” a supercritical water oxidation unit when I worked for the Army, intended to be used to destroy old or off specification smoke chemical compositions.  From experience, I know that it is not easy to keep those conditions under control, and this was not even considering intense neutron bombardment of the high pressure components.  Considering the problems with high pressures in Generation III reactors, I believe that SCWRs are not viable at present.  Supercritical water is a strange beast anyway, with properties quite unlike ordinary water.  For example, where in ordinary water materials like salt is quite soluble and things like oil are insoluble, the reverse is the case with suprcritical water.  Whilst MSRs have their technological challenges, the challenges for SCWRs are much more daunting, in my opinion.

One advantage that they do have over Generation III reactors, at least in concept, is that they are not moderated as much, due to the lower density of supercritical water as opposed to liquid water, producing more fast neutrons.  Those neutrons can convert uranium-238 into plutonium-239, making them breeder reactors.  In addition, those fast neutrons also reduce the amount of long lived products in spent fuel.  Still, I am quite dubious of this concept.

Thus far, we have concerned ourselves with thermal neutron (or predominately thermal) reactors.  The other major kind are fast neutron reactors, where no moderator is used.  I actually saw a prototype of one of them in person, SEFOR in northwest Arkansas.

This was one of the first Sodium Cooled Fast Reactors (SCFRs), and, as its name indicates, uses sodium metal as the coolant.  Right there some warning bells should go off, loudly.  I mentioned before that fluoride salts do not play well with water, and metallic sodium plays even worse with it.  If you saw the episode of Mythbusters where they tried to blow up the toilet with sodium you know what I mean.  The purpose of fast neutron reactors is to produce power and to breed uranium-238 into plutonium-239.  Thus, like other breeder reactors, they produce more fuel than they consume.  There are some technical challenges, but one real advantage is that the coolant is under low pressure, just about atmospheric, because sodium does not need high pressure to keep from boiling like water does, making this design safer from a pressure standpoint.

In common with some of the other reactors previously mentioned, this design destroys much of the long lived products in the fuel after fission, and as I said a minute ago, actually uses them as part of the energy output from the plant.  However, I believe that LSR’s are more practical since thorium-232 is more abundant than uranium-238.  With that said, this is known technology, for the most part, and the engineering is in many respects easier than that required for SCWRs.  Another disadvantage is that the output temperature, at least with current designs, is not quite high enough to convert water to hydrogen efficiently.

Another fast neutron design is the Lead Cooled Fast Reactor (LCFR).  In this design, lead (or better, the eutectic mixture of lead and bismuth, for reasons to be made clear later) is the coolant/working fluid.  It is a closed loop, low pressure design and the metallic working fluid boils water in a heat exchanger, just like a sodium cooled one.  This design, or an earlier version of it, has actually been used by the Soviet Union for submarines, so the concept is proven.  In addition, they are smaller and simpler than water cooled reactors.  The fuel is similar to that for other fast breeder reactors, and likewise the long half life products are largely destroyed by the fast neutrons, being used as fuel.  This may be a real candidate for many uses, particularly since being a breeder, the fuel lasts for decades rather than years, and the problem with long term storage of the finally spent fuel is reduced.

Another advantage is that the molten metal circulates by convection, not by pumps, so in case of a power failure, coolant supply is not interrupted.  Yet another one is that the metallic coolant is not water reactive, so no explosion hazard exists should it come into contact with water as is the case with sodium cooled reactors.  Finally, if something happens and a leak develops, control rods can be inserted to slow down the reaction.  Then the coolant solidifies, stopping any leak.  This is a double edged sword, since the reactor is essentially locked up then.  However, the lead/bismuth eutectic melts at about 124 degrees C, so relatively minor outside heat (or careful withdrawal of the control rods) can reliquify the coolant, rendering the reactor operable again.  One disadvantage is that current designs output only temperatures of around 550 degrees C, making hydrogen production unfeasible.  However, it may be possible to increase those temperatures.

The last type of fast neutron reactor that I intend to discuss is the Gas Cooled Fast Reactor (GCFR).  These are quite similar in concept to the VHTR described earlier, except there is no moderator.  This makes using carbon coated fuel beads impractical, since the carbon is the moderator.  New fuel pellet designs are being developed, and this type of reactor shows great promise, but it a little further out in the timeline than the other two fast neutron ones.  Since it is also a low pressure core design, and most designs use helium as the coolant/working fluid, no pumps are necessary since convection supplies the coolant movement.  I believe that this will be the fast neutron reactor of choice in future.  It has all of the advantages of other fast neutron reactors, including long fuel life, shorter half life of fuel that is finally spent, and nothing explosive or water reactive in it.

Those are some of the designs coming up in the near future.  There others, and variations of these themes.  I like the breeders in the long term, but for the short term the pebble bed once through fuel use is likely the most practical for power production in the nearer term.

Well, you have done it again!  You have wasted many einsteins of perfectly good photons reading this “hot” material.  And even though Terry Jones (NOT the Python one) regrets burning that book and all of the blood on his hands when he reads me say it, I always learn much more than I could ever possibly hope to teach by writing this series, so keep those comments, questions, corrections (especially), and other feedback coming.  Remember, no scientific or technical subject is off topic here.  I shall remain as long as comments warrant tonight, and shall return around 9:00 PM tomorrow evening for Review Time.

Warmest regards,

Doc