Hello:
The conventional activation neutron detection methods tend to use a moderating system, to thermalize the neutrons, for better target isotope absorption. This got me thinking about some more direct method, which takes advantage of a suitable fast neutron reaction, to produce the betas, gammas etc. While the fast neutron cross sections are much lower, the elimination of the thermalization step can compensate for the lower cross section of the fast neutron reactions. Cu, and Mg is used for this purpose, but the half life of the products formed are to long.
I looked at the tables for fast neutron reactions, and one which stood out was the As + n > gamma reaction. The isotope formed has a half life of only 16ms using D-D neutrons, with a reasonable cross section. The emitted gammas would be easy to detect with either a NaI(Tl) or even a plastic scintillator. The advantage of NaI(Tl) would be a lower sensitivity for direct neutrons, thus it would not be hard to discriminate for only the gammas from the As emission.
An aprox. 1" layer of As should be plenty, I also found a cheap easy source of As at the hardware store. Several vegetation killers contained up to 37% by weight elemental arsenic, in the form an organic herbicide compound, named Cacodylic acid:(CH3)2AsO2H. If the water were to be carefully removed from this product via evaporation etc, then it should contain enough As to do the job. The best part is that the As would still remain in a lower toxicity form, since apparently the As is tightly bound up in the organic compound, so that most of the As does not get absorbed via the digestive tract. First I thought of a way to pyrolyticaly decompose this material to yield a denser As material, but why risk the amplified toxicity, unless you obtained and used elemental As in epoxy, which would seem to be ideal?
The advantage of this method if it works, would be to enable real time fast neutron flux measurements, due to the extreme short half life of the As product. Plus your trusty old NaI(Tl) unit could be used as is, with just a slip on As cover. Maybe someone else could add if this is realistic or not, on the surface it does look interesting though. I am sure there may be better methods, but this approach could offer some unique advantages.