For the OP, there are several ways to make a nuclear weapon, and nuclear power makes, that shouldn't surprise you, use of *similar* technology than in nuclear weapon production. That means, there are similarities, and there are differences.
If you want to learn about nuclear weapons, you can find a lot of information here:
http://nuclearweaponarchive.org/
according to the website author, all this is publicly available information, and I recon that's true, as a lot of the information there I found in other sources.
Point is, in order to make a fission weapon, you need a strongly over-critical mass, because you want to liberate as much fission energy before the weapon destroys itself, which happens on the time scale of a few microseconds. So you only have a few microseconds to do all the energy production, and that means that you need a very fast divergence.
Just barely critical is not going to work, because the energy production rate would increase too slowly: this is what's called a fizzle: your nuke is a dud.
There are a few materials who can potentially serve to make such strongly over-critical masses: there's U-235, there is Pu-239 (and in fact, most mixtures of plutonium isotopes except Pu-238), you can even do it with Neptunium-237 and also with Am-241. You could also do it with U-233.
Happily, none of these materials exist as such, in pure enough form, in nature. Plutonium, neptunium and americium simply don't exist in nature, and U-235 is only 0.7% of natural uranium. U-233 doesn't exist in nature (or if it does, it is in trace amounts). U-238 is an efficient neutron absorber, and if it is too much present, it will stop any supercriticality.
So the first difficulty is to get the right material. And it is here that there is a link with the technology of nuclear power stations.
The second difficulty in a nuclear weapon is the fact that you need to obtain strong supercriticality, but of course before the weapon is used, the material cannot be present in critical form, or it would blow into your face during assembly. So one needs a technique to turn, quickly, a sub-critical set of materials into a strongly supercritical assembly. This part of technology has nothing to do with nuclear power stations, but it is important to know.
Historically, there have been two ways to make the supercritical assembly: by far the simplest one is "gun assembly": with some kind of "gun", two or more sub-critical parts are shot into one another, so that they form a supercritical assembly. The advantage of this technique is that it is robust, relatively easy to make, but it has two strong drawbacks:
- the supercritical mass thus obtained needs a LOT of material
- no matter how fast one thinks a gun can fire, this is, on nuclear scale, a very slow process. In fact, it is so slow, that one cannot use any material that generates a lot of neutrons by radioactive decay, because the assembly "slowly" goes from subcritical to supercritical, and crosses hence the zone of "barely critical", at which point background neutrons can start a chain reaction, and blow the weapon apart with "low energy". We have a fizzle.
It turns out that only pure U-235 and eventually pure U-233 could be used for such a weapon. All other materials have far too high a neutron background.
The other way is "implosion": chemical explosives in a special configuration generate a converging spherical shock wave which compresses a ball of nuclear material. It turns out that criticality rises strongly with material density. So the same amount of material, at normal density, can be sub critical, and at higher (compressed) density, supercritical. It turns out that such a shock wave compression is much much faster than gun assembly. Moreover, one needs much less nuclear material that way (we don't need a supercritical mass at normal densities, while we did with gun assembly). However, this time, the difficulty is that you only have your supercritical assembly during a few microseconds (the time of the implosion). So now you need, exactly at the right time, to generate a pulse of neutrons that will start the chain reaction.
This is much more difficult technology, and this time, the success depends upon the precision and the mastery of the shock wave and the neutron pulse.
Although one can cope with higher neutron backgrounds, nevertheless, also here, the higher the neutron background, the higher the chances to obtain a dud or a low-yield weapon. Implosion works very well with Pu-239 if one masters enough the implosion technology. The higher the contents of Pu-240, the harder it becomes to make it work, although with enough implosion quality, any Pu mixture (except for Pu-238) can be made into a bomb.
How does one get "weapon-grade" materials ?
U-235 is obtained by isotopic separation. From 20% U-235 onward, one can in principle make a bomb, but one needs high purities in order for the needed mass of material to remain reasonable (at 20% one would need several tons, which is impractical). Usually one wants over 80% U-235, and preferentially, more.
There are 2 known "large scale" isotopic separation technologies: diffusion and centrifuges. Centrifuges are better at making small quantities of high-purity material, while diffusion is better at making large quantities of low-enriched uranium, but both can be used for both purposes. It is difficult and large-scale technology.
However, as PWR need low-enriched U-235, this technology is present in the nuclear power sector, although not directly in the form needed for highly enriched U for a bomb. So the technology is present, but the configuration of a plant for power production is different than the configuration for making weapon-grade U-235.
There exist however, other technologies to separate U-235 with high purity, which are not used in the power sector, but which do exist. The most promising example is probably laser based isotope separation (AVLIS
http://en.wikipedia.org/wiki/AVLIS ).
In order to make plutonium, one has to place U-238 in a neutron flux, such as is present in a reactor. A PWR has a certain amount of plutonium in its spend fuel ; however, the isotopic composition of that plutonium is rather bad: it contains a lot (~ 30%) of Pu-240, which gives it a huge neutron background.
In order for this plutonium (which is present on percent-level in spend fuel) to be extracted, one needs a chemical treatment of the spend fuel, in what's called the PUREX process. This chemistry is known now, but the difficulty resides in handling the highly radioactive spend fuel. The high Pu-240 content in spend fuel comes from the fact that fuel elements remain a long time in a PWR. In order to obtain much more pure Pu-239, one needs to irradiate U-238 only slightly, and reprocess it. That's better done in a reactor in which one can change continuously some fuel elements. A PWR is not suited for that, but a graphite reactor is, most research reactors are, and Candu reactors are.
So, "good-quality" weapon plutonium needs fiddling with non-power reactors, and need in any case a reprocessing facility. "poor-grade" plutonium can be obtained from spend fuel of a PWR. One can make an implosion weapon from both, but it is harder with the second kind of material.
In conclusion: if one goes for simple weapon design (gun type), one needs enrichment to a high degree. If one thinks one can handle an implosion weapon design (much harder), one can do with a reactor and some chemistry in a reprocessing facility. Using spend fuel from a PWR gives plutonium of a worse quality than making special-purpose reactors.
One can also irradiate thorium, which converts to U-233. Now, this is an interesting combination for a bomb builder, because he doesn't need isotopic separation (the U-233 can be separated chemically from the thorium in a reprocessing plant, using the THOREX process). Moreover, the activity of this irradiated thorium is lower than that of spend fuel for Pu production. And U-233 can in principle also be used in a gun-type weapon. So this combines the two "easy" approaches in both cases: chemistry instead of isotopic separation, and gun type instead of implosion. Only, this is not the path nuclear nations have followed, and hence more of a "development risk" than proven technologies.
So we see the link between the technology of nuclear power on one hand, and nuclear weapons on the other:
- enrichment, needed for low-enriched uranium for PWR, is a technology which can be modified into making highly enriched uranium for a weapon. However, we're talking about very large installations here, which have to be seriously modified to switch them from producing lowly enriched uranium into highly enriched uranium. However, it is the same technology, applied differently.
- fuel reprocessing which extracts plutonium. The difficulty here is that these are also large facilities, because they have to handle chemistry of very active materials.
- reactors. PWR can only make "bad quality" plutonium (which can be used in a bomb, but with more difficulties). Other reactors can be turned into specific weapon-grade plutonium production reactors.
Clearly, the technological knowhow of nuclear power is interesting for a bomb builder, although it doesn't contain everything. Also, existing facilities for nuclear power have to be seriously modified in order for them to make weapon material ; so as long as they are regularly inspected, this can be verified. Finally, there exist technologies totally outside of the nuclear power sector, which can also lead to nuclear weapons. I mainly think of laser isotope separation.
