I worked for George Miley (referenced in "The World's Simplest Fusion Reactor: And How to Make It Work") in grad school and have a fair amount of experience working with these things. Of course I haven't touched one in ten years since I bailed out and went into EE, so take all this with a grain of salt. With that disclaimer, a few points.
First, the dangerous output of these things is not neutrons, but x-rays. An ungodly amount of x-rays get pumped out in these things, so if you have a window you ought to shield it (the vacuum vessel did a good job of stopping most of the x-rays, only those headed toward the window needed to be shielded). We used something like a 1/4" of leaded glass.
Second, the fusion that occurs in these devices, at least the ones we built, are beam-target interactions where the target is the background deuterium gas. What one would like to have is beam-beam interactions where the fast deuterons interact with each other rather than the background gas. This would be good for a few reasons, first the resulting fusion output would depend on the square of the input current, not linearly as is it does with beam target. This means that as you increase input power, you would approach and eventually pass break-even (assuming your grids didn't melt or somesuch). Second with beam-beam interactions, you can evacuate the device more thouroughly which helps avoid some types of loss, particularly charge exchange. Third , beam-beam interactions occur at up to four times the energy of beam-beam interactions, which is particuluar attractive for the exotic fuel combinations (D-He3, p-B11, etc). The problem is that it's easy to arrange for a fair amount of background gas to stay in the chamber, but to get high enough denisities for signifigant beam-beam interactions to occur you need some combination of very high input power, very high recirculation rates (see below) and very good focusing at the grid center. To the best of my outdated knowledge, no one has achieved this yet.
Third there are two issues of loss in this type of device. First let's talk about break even. In order to break even, you need to be able to extract as much power out of the device as you put in. Assuming that you can convert about 50% of the energy coming out of the device into power (this may well be optimistic), your output power needs to be twice your input power, since half your power is lost as heat. In other words, you need to produce as much additional power from fusion as you put in as electricity. When we were working on this we were produncing something like.00000000000001 times as much power[1] from fusion as we put in as electricity. Far from break even. Going to D-T instead of D-D would probably up this by a factor of 10 or so, but still far from break even.
The second loss factor involves losses of the recirculating D+ ions. One is grid losses, the star mode referenced in the above article helps a fair amount here. In our experiements, the grid was about 95% transparent, but because the discharge avoids the grid, we ended up with an effective transparancy of 99.5% or so. The background gas presents another big source of loss. A real killer is charge exchange: you invest 30 keV in some ion and then it grabs an electron from a D2 molecule in the background gas. Now your fast particle is not constrained, being neutral, and it goes crashing into the vacuum chamber wall and is lost.
Even if you won't be able tomake one of these into a powerplant in the near future, there are some applications for a relatively simple neutron generator. One that I believe has been commercialized already is neutron activation analysis. In other words bombard some object with neutrons to make it radioactive (activate it), and analyze the type of radiation that comes out to see what the object is made of. Sounds scary, but your only making the object a tiny bit radioactive. Really.
There's some slides from a relatively recent IEC confe
While D-He3 fusion should indeed be somewhat easier to achieve than D-D fusion, it's still singificantly harder to achieve than D-T fusion (the easiest reaction to achieve). And, we still don't have a working D-T reactor design, and that has nothing to do with either a lack of Deuterium or Tritium.
The real beauty of D-He3 fusion is that it is almost aneutronic. That is, the reaction produces very little neutron radiation, and the radiation it does produce is at a lower energy than that produced by D-T fusion. This means that less shielding is needed for the reactor, and long term damage to reactor structure do to transmutation is less of an issue.[D-He3 produces no neutrons itself, but once you toss D and He3 into a plasma, there are also neutron producing D-D reactions, albeit at a slower rate.]
[Note that the P-B11 reaction is really aneutronic, but its even harder to make work]
So, while it's nice to know there's He3 around it does us no good till someone irons out the considerable kinks in fusion reactor design.
Slashdot Mirror
I worked for George Miley (referenced in "The World's Simplest Fusion Reactor: And How to Make It Work") in grad school and have a fair amount of experience working with these things. Of course I haven't touched one in ten years since I bailed out and went into EE, so take all this with a grain of salt. With that disclaimer, a few points.
.00000000000001 times as much power[1] from fusion as we put in as electricity. Far from break even. Going to D-T instead of D-D would probably up this by a factor of 10 or so, but still far from break even.
First, the dangerous output of these things is not neutrons, but x-rays. An ungodly amount of x-rays get pumped out in these things, so if you have a window you ought to shield it (the vacuum vessel did a good job of stopping most of the x-rays, only those headed toward the window needed to be shielded). We used something like a 1/4" of leaded glass.
Second, the fusion that occurs in these devices, at least the ones we built, are beam-target interactions where the target is the background deuterium gas. What one would like to have is beam-beam interactions where the fast deuterons interact with each other rather than the background gas. This would be good for a few reasons, first the resulting fusion output would depend on the square of the input current, not linearly as is it does with beam target. This means that as you increase input power, you would approach and eventually pass break-even (assuming your grids didn't melt or somesuch). Second with beam-beam interactions, you can evacuate the device more thouroughly which helps avoid some types of loss, particularly charge exchange. Third , beam-beam interactions occur at up to four times the energy of beam-beam interactions, which is particuluar attractive for the exotic fuel combinations (D-He3, p-B11, etc). The problem is that it's easy to arrange for a fair amount of background gas to stay in the chamber, but to get high enough denisities for signifigant beam-beam interactions to occur you need some combination of very high input power, very high recirculation rates (see below) and very good focusing at the grid center. To the best of my outdated knowledge, no one has achieved this yet.
Third there are two issues of loss in this type of device. First let's talk about break even. In order to break even, you need to be able to extract as much power out of the device as you put in. Assuming that you can convert about 50% of the energy coming out of the device into power (this may well be optimistic), your output power needs to be twice your input power, since half your power is lost as heat. In other words, you need to produce as much additional power from fusion as you put in as electricity. When we were working on this we were produncing something like
The second loss factor involves losses of the recirculating D+ ions. One is grid losses, the star mode referenced in the above article helps a fair amount here. In our experiements, the grid was about 95% transparent, but because the discharge avoids the grid, we ended up with an effective transparancy of 99.5% or so. The background gas presents another big source of loss. A real killer is charge exchange: you invest 30 keV in some ion and then it grabs an electron from a D2 molecule in the background gas. Now your fast particle is not constrained, being neutral, and it goes crashing into the vacuum chamber wall and is lost.
Even if you won't be able tomake one of these into a powerplant in the near future, there are some applications for a relatively simple neutron generator. One that I believe has been commercialized already is neutron activation analysis. In other words bombard some object with neutrons to make it radioactive (activate it), and analyze the type of radiation that comes out to see what the object is made of. Sounds scary, but your only making the object a tiny bit radioactive. Really.
There's some slides from a relatively recent IEC confe
>P-B11?
p (proton) + B11 (Isotope of boron) -> 3 He4 (standard helium) + 8.7 MeV (bunch o' energy)
Note, no neutrons (good), but you need a ridiculous high energy plasma to see much fusion from this.
While D-He3 fusion should indeed be somewhat easier to achieve than D-D fusion, it's still singificantly harder to achieve than D-T fusion (the easiest reaction to achieve). And, we still don't have a working D-T reactor design, and that has nothing to do with either a lack of Deuterium or Tritium.
The real beauty of D-He3 fusion is that it is almost aneutronic. That is, the reaction produces very little neutron radiation, and the radiation it does produce is at a lower energy than that produced by D-T fusion. This means that less shielding is needed for the reactor, and long term damage to reactor structure do to transmutation is less of an issue.[D-He3 produces no neutrons itself, but once you toss D and He3 into a plasma, there are also neutron producing D-D reactions, albeit at a slower rate.]
[Note that the P-B11 reaction is really aneutronic, but its even harder to make work]
So, while it's nice to know there's He3 around it does us no good till someone irons out the considerable kinks in fusion reactor design.