Aneutronic fusion

Aneutronic fusion is a (hypothetical) form of fusion power where no more than 1% of the total fusion energy released is carried by neutrons. It has long been a dream of both the conventional and alternative fusion communities because of problems associated with neutrons like radiation damage, biological shielding, remote handling, and safety issues.

There are a few nuclear fusion reactions that have no neutrons as products on any of their branches. Those with the largest cross sections are these:

D	+	3He	→		He	(3.6 MeV)	+		p	(14.7 MeV)
D	+	6Li	→	2	4He	+ 22.4 MeV
p	+	6Li	→		4He	(1.7 MeV)	+		3He	(2.3 MeV)
3He	+	6Li	→	2	4He		+		p	+ 16.9 MeV
3He	+	3He	→		4He		+	2	p
p	+	11B	→	3	4He	+ 8.7 MeV

The first two of these use deuterium as a fuel, and D-D side reactions will produce some neutrons. Although these can be minimized by running hot and deuterium-lean, the fraction of energy released as neutrons will probably be several percent, so that these fuel cycles, although neutron-poor, do not classify as aneutronic.

The middle two reactions are usually treated as a chain in the hope of attaining an enhanced reactivity due to a non-thermal distribution. The product ³He from the first reaction could participate in the second reaction before thermalizing, and the product p from the second reaction could participate in the first reaction before thermalizing. Unfortunately, detailed analyses have not shown sufficient reactivity enhancement.

The second-to-last reaction suffers from a fuel-availability problem. ³He occurs naturally on the Earth in only miniscule amounts, so it would either have to be bred from reactions involving neutrons, so that the potential advantage of aneutronic fusion is largely neutralized, or mined on the Moon or other extraterrestrial bodies, needless to say at great expense.

For the above reasons, most advocates of aneutronic fusion concentrate on the last reaction, p-¹¹B. Even here, though, there are a number of side reactions that will produce neutrons or other radiation, among them the following:

p	+	11B	→	12C + γ
p	+	11B	→	n + 11C
4He	+	11B	→	n + 14N
4He	+	11B	→	p + 14C
4He	+	11B	→	T + 12C
11B	+	11B	→	junk

as well as reactions with a possible ¹⁰B impurity fraction. Detailed calculations (Heindler and Kernbichler, Proc. 5th Intl. Conf. on Emerging Nuclear Energy Systems, 1989, pp. 177-82) show that at least 0.1% of the reactions in a thermal p-¹¹B plasma would produce neutrons. This is still extremely intense radiation, as can be seen by the following simple calculation.

If we assume 0.1% of the energy is carried off by neutrons, even a "kitchen-sized" reactor with 30 kW of fusion power will produce 30 W of neutrons. If there is no significant shielding, a worker in the next room, 10 m away, might intercept (0.5 m^2)/(4 pi (10 m)^2) = 4e-4 of this power, i.e., 0.012 W. With 70 kg body mass and the definition 1 rad = 100 erg/g = 0.01 J/kg, we find a dose rate of 0.017 rad/sec. Using a quality factor of 20 for fast neutrons, this is equivalent to 0.34 rem/sec. The maximum yearly occupational dose of 5 rem will be reached in 15 sec, the fatal (LD-50) dose of 500 rem will be reached in half an hour. For an industrial size (100 MW) reactor under the same assumptions, the dose rate would be thousands of time higher, and anyone standing nearby would be dead in a fraction of a second. The neutrons would also activate the structure so that remote maintenance and radioactive waste disposal would be necessary. Of course, material damage and safety problems would be brought into an easily manageable range.

If we look at where these neutrons come from, they are dominated by the reaction

¹¹B + α → ¹⁴N + n + 157 keV

If we really want to eliminate neutrons, we see that we cannot tolerate fast alphas in the plasma. Usually, the product alphas are relied on to keep the fuel hot. If the alphas have to be extracted with their full energy, we will need very, very efficient processes to collect this power, transfer it, and drive whatever process maintains the plasma energy. The reaction itself produces only 157 keV, but the neutron will carry a large fraction of the alpha energy, which will be close to E_fusion/3 = 2.9 MeV.

Suppose we can do this, so that fast alpha reactions are suppressed by several orders of magnitude. We will always have the fuel ions, protons and borons. Of course, p+p doesn't do much, and boron-boron reactions can probably also be neglected due to the large Coulomb barrier. The species can however react with one another in a number of ways to produce neutrons. These reactions are all endothermic. The smallest barrier is for the reaction

¹¹B + p → ¹¹C + n - 2.8 MeV

In a thermal plasma of a few hundred keV temperature, there is a sufficient number of protons in the high energy tail that this reaction is a significant source of neutrons. If the proton temperature is reduced below about 30 keV, then this process is suppressed, but there is also no longer any significant fusion. The only way around this dilemma is to produce a nearly mono-energetic proton energy distribution, that is, a beam. If the beam energy is chosen to be at the fusion resonance around 600 keV, then the reactivity is also about three times higher than the maximum for a thermal plasma.

Let us assume that we can produce and maintain such a non-Maxwellian distribution, so that (p,n) reactions are suppressed by several orders of magnitude. What is the next most serious source of neutrons? Probably those associated with fuel impurities. If the density of fast alphas and fast protons is controlled to suppress reactions with ¹¹B, then the reactions with impurity ¹⁰B should be similarly suppressed for the same reasons. The impurity deuterium density must be kept low enough to suppress D-D fusion. Since the fusion rate is proportional to the square of the deuterium density, this is presumably not too difficult. More serious is perhaps the reaction

¹¹B + d → ¹²C + n + 13.7 MeV

The cross section for this reaction should be similar to that for p-¹¹B fusion, so that it will be necessary to use very pure hydrogen fuel. Considering the factor 2 mass difference and the small amount of fuel needed, one might assume that it is technically and economically feasible to reduce the deuterium concentration several orders of magnitude below its natural abundance of 1.5e-4.

Let us assume that fuel of sufficient chemical and isotopic purity can be made. Any other elements getting into the plasma, for example through outgassing of the walls, are another potential source of neutrons. Any energetic fuel or product particles striking solid surfaces can also produce neutrons. It is difficult to estimate the severity of these reactions, even with a particular configuration in mind.

If, in addition to the other assumptions above, we assume that interactions between the plasma and the containment device can be adequately controlled, then neutron production will be suppressed by many orders of magnitude. What other types of radiation will be a concern? Bremsstrahlung will produce extremely large quantities of hard x-rays, which must and, presumably, can be shielded by a modest amount of metal. The fusion reaction

¹¹B + p → ¹²C + γ + 16.0 MeV

will produce 4, 12, and 16 MeV gammas with a branching probability relative to the primary fusion reaction of about 10^-4. With no shielding, this would be a tremendous radiation dose. The calculation above would apply if the production rate is decreased a factor of ten and the quality factor is reduced from 20 to 1. Without shielding, the occupational dose from a small (30 kW) reactor would still be reached in about an hour, so enough shielding must be installed to attenuate the hard gamma flux by well over three orders of magnitude. For an industrial reactor, the attenuation should be well over six orders of magnitude. Radiation shielding will be complicated by the possibility of (γ,n) reactions in the shield material, reintroducing the neutron problem. This should be doable, but illustrates the seriousness of the neutron problem, even for "aneutronic" fusion.