In which I’m going to do my damndest to get the West’s security services interested in my blog. Hi guys!
Making a nuclear bomb is harder than you might think. When you stuff nuclear fuel into a reactor the fission chain reaction is carefully controlled via neutron moderators and absorbers so that it occurs in a stable, steady-state fashion. This nevertheless liberates a tremendous amount of energy every second, so much so that we need dedicated cooling systems just to deal with the aftermath of the chain reaction when the reactor core is shut down. However, if you took away all the control rods and failsafes and dumped all of the coolant out of the reactor and basically tried to make the thing explode, you’d run into two problems.
1) The percentage of nuclear fuel that is fissile u-235 (3%) is far too small to create much of a bang.
2) While nuclear reactors are deliberately built to be inefficient in terms of the rate of neutron reactions (after all, we don’t really want melted down reactor cores dotting the planet) just taking the neutrons off the leash isn’t particularly efficient either.
The efficiency (or rate) of your nuclear reaction largely on how densely your critical mass is packed together; if you have a lot of u-235 atoms crammed into a small volume you dramatically increase your chances of scoring a hit with a fast1 neutron, as well as increasing the chances of child neutrons inducing subsequent fission events. A critical mass that is unintentionally triggered into undergoing a chain reaction goes on the big list of criticality accidents. Wikipedia says there’s been about sixty of these since nuclear power first became a thing, but you’ll notice there isn’t a corresponding big list of sixty accidental nuclear explosions. This is because criticality accidents are inherently self-correcting in the same way that the light water moderator I talked about in the PWR post is; inducing a chain reaction in a critical mass heats it up, which causes it to expand, which spreads out the fissile atoms, which reduces the critical mass below the density required to sustain a chain reaction. As a result of this most criticality accidents are like the ones caused by the demon core; they release a burst of hard neutrons that usually kill whoever was fiddling with the critical mass, but they don’t proceed any further than that. In order to get our critical mass to actually explode we’re going to have to get creative.
The first step is in packing as much fissile material into the critical mass as possible. That 3% u-235 is far too weedy and the other 97% u-238 is effectively useless for our purposes; if we want a really big bang we’re going to have to enrich our uranium to the point where 90%+ of it is u-235. To do this we need industrial scale isotope separation facilities like the Manhattan Project’s three plants at Oak Ridge, which work in a sequential cascade process. The way the separation process works is that after you “enrich” raw uranium that you get out of the ground you end up with two separate collections of material: one with a slightly higher concentration of u-235, and one with a slightly lower concentration of u-235. The material with more u-235 is sent up to the next step in the chain which separates it again, producing another two sets of material with more and less u-235 respectively. The material with more u-235 is sent further up the production chain where it progressively has more and more of its u-238 content squeezed out of it. The leftover material with less u-235 isn’t wasted either; instead it’s sent back down the chain to start the process again and get as much pure u-235 out of the raw uranium as possible.
The separation process takes a lot of time and effort and is pretty much the reason why only nation-states have the resources to make nuclear weapons from scratch. Iran keeps trying to establish the industrial base from which they can perform sustained isotope separation, which is why the US and Israel are investing quite a lot of time and effort in fucking with the Iranian centrifuges. Assuming we can build the facilities and that the US doesn’t come in and drop a JDAM on it, though, we’ll eventually end up with a lump of weapons-grade uranium. Or rather, several small lumps none of which are large enough to constitute a critical mass. The next step is to find some way of making them go critical on demand.
The simplest way of doing this is the gun-type fission bomb, which is what was dropped on Hiroshima. Here you have two sub-critical pieces of fissile material separated by a short distance, and when you want your bomb to explode you detonate an explosive inside the bomb which “fires” one piece of material at the other, smooshing them together and making them go critical. Cue one big bang. This design works, but it’s rather inefficient in terms of both the total amount of fissile material used to make the bang and the amount of that fissile material which is consumed before the bomb explodes. A bomb which uses 100% of its fissile material in the chain reaction before releasing its energy and exploding is one which is operating at maximum efficiency; however this is very much the ideal scenario, whereas in reality the chain reaction inside a simple critical mass doesn’t occur fast enough to ensure all of the atoms in the fissile material get to react with a neutron. If the reaction is so slow that the fissile material explodes before all of it has had a chance to react, then you’ve effectively wasted a certain percentage of your fissile material, and this is bad for two reasons. One: the bang the bomb will make isn’t as big as it could be. Two: given the trouble you went through to produce the fissile material in the first place, you really don’t want to be wasting it.
To make the bomb as efficient as possible involves taking advantage of a few of the quirks of the critical mass. The critical mass isn’t set in stone; it’s simply the mass of fissile material required to sustain a chain reaction, and so it can change depending on the presence of various conditions that make things more or less favourable for the neutron pinball going on inside the bomb core. The three main ones we’re concerned with are:
Density. We’ve saturated our bomb core with as many fissile atoms as possible by enriching the uranium to 90% u-235, and this makes the core a target-rich environment for any neutrons looking to split an atomic nucleus. However, we can go further than this. The density of u-235 at room temperature is comparatively low. If we were to somehow artificially compress the bomb core we could jam the u-235 atoms into a much smaller space, increasing the efficiency of the chain reaction yet further and ensuring most of them are used to power the explosion.
Presence of a neutron reflector. The most efficient shape for a critical mass is a sphere, as this has the lowest surface area to volume ratio of all possible shapes and ensures the majority of u-235 atoms are surrounded on all sides by other u-235 atoms. This makes child neutrons produced as part of the chain reaction far more likely to strike and react with other u-235 atoms. However, a sphere of fissile material will still lose neutrons from the outer layers of the sphere, decreasing the efficiency of the reaction. This can be solved by surrounding the bomb core with a neutron reflecting material such as beryllium, which bounces escaping neutrons back into the core and ensures as many of them react with u-235 atoms as possible.
Presence of a tamper. When a critical mass undergoes a chain reaction, it will get hot. When things get hot, they expand. When things expand, their density decreases. This is undesirable when we’re trying to make the bomb core as dense as possible during the detonation process, so bomb cores are also surrounded by a very dense material called a tamper that contains this expansion to a degree. It cannot stop the expansion because no material can withstand that stresses and forces contained with an exploding nuclear weapon, but it will slow it down. Early fission weapon designs combined the tamper and the neutron reflector into one component – usually u-238 since this had the added bonus of occasionally reacting with the fast neutrons emitted by the chain reaction and boosting the yield of the explosion even more.
This is why the second type of nuclear weapon design – the implosion design – looks like this:
It’s a core of subcritical fissile material surrounded by a tamper/reflector and a series of high-explosive lenses. When you want to detonate the bomb you trigger the high explosive. This creates a shockwave focused inwards towards the core which compresses it very very quickly, increasing its density and causing it to go critical. The chain reaction starts, and escaping neutrons are reflected back into the core by the tamper. The tamper itself only delays the explosion itself by a fraction of a millisecond, but this is enough time for several additional steps in the chain reaction to take place. Given the exponential nature of the chain reaction (1, 2, 4, 8, 16, 32 etc.) that additional fraction of a millisecond is enough time for the energy of the explosion to increase tenfold2.
Implosion weapons are more complex and advanced than gun-type nuclear bombs, but they’re also far more efficient. The Little Boy bomb dropped on Hiroshima was a gun-type fission weapon containing 141 pounds of 80% enriched uranium, and had a yield of 16 kilotons. The Fat Man bomb dropped on Nagasaki had a 13.6 pound core of plutonium, and yet the nuclear reaction catalysts described above increased its yield to 21 kilotons, producing a bigger bang with less than 10% of the fissile material. If you want economy in your fission weapon, then the implosion design is the way to go.
Fission weapons aren’t the whole story, though. In fact, they’re only the first page. There will be a sequel to this post at some point, and it will talk about the current cutting edge in nuclear weapon design: the thermonuclear bomb.
- Note: there is no moderator present inside a nuclear weapon to produce low energy thermal neutrons, so we have to rely on fast neutrons to sustain the chain reactions
- I can’t find a source for the exact number so I’ve made a guesstimate based on the prompt lifetime of a neutron in a fast fission reaction (10-7 seconds). If anything I’ve lowballed it, though.