With some creative thinking, plastic scintillator material may actually be reasonably accurate. Since we are not going for a general-purpose detector and have a good idea of what radiation is coming out of the fusor, some steps might be taken to find sources of background noise and increase accuracy.
Plastic scintillators (and organic scintillators in general) are useful for counting fast neutrons because of proton recoil reactions in the plastic matrix or the scintillator material. Some researchers have taken this to the extreme and used crystals of pure stilbene (an organic scintillator) as fast neutron scintillators, with reasonable efficiency (9.5%). Another researcher has reported 30% efficiency for 10 Mev neutrons using a 15 cm thick scintillator of polystyrene and terphenyl.
The downside for a lot of applications is that these scintillators do a fairly poor job of discriminating between proton recoil-induced pulses, and those due to electrons and gamma radiation.
However, in our case, we are not exactly working with the output of a particle accellerator. The energy we are using to initiate the fusion reactions is very low, so that most of the radiation generated by the fusor is in the form of relatively low energy x-rays, most of which are absorbed in the chamber walls. The ones that get out are not anywhere near as efficient at exciting the scintillator as the recoil protons.
No electrons get out, period.
There is a possibility of some gamma radiation due to the 3 Mev protons and 1 Mev tritons and 3He hitting the chamber walls. This could probably be verified using a NaI(Tl)scintillator, which responds very well to gamma radiation, but not well at all to fast neutrons. Some of this could be fitered out with a lead sheet.
There is a possibility of multiple counting due to multiple neutron-proton scattering events. This can be controlled to a certain extent by not going too overboard on the size (or thickness) of the scintillator. Marion and Fowles, vol.1 gives enough information so that this effect can be estimated and dealt with.
Other noise sources include the photomultiplier/scintillator combination, due to either internal noise from the PMT itself or outside interference. These sources can be dealt with by proper shielding techniques and use of a discriminator. Since a discriminator is nothing more than a comparator with a variable bias, this need not be very expensive. Mr. Zambelli reported a fairly high base count rate from his raw PMT/scintillator combination. I suspect that a lot of these counts were low level could be filtered out using a discriminator between the PMT and counter.
It is also interesting to note that the U. of Wisconsin researchers cite a high noise level from the PMT/plastic scintillator they used as a proton detector, and reported that they did not get good correlation between the scintillator and a semiconductor detector until they achieved sufficient reactions rates in their fusor to push the proton count above the noise floor of the scintillation detector. This statement takes on additional weight for neutron detection considering that proton dectection is much more efficient in a plastic scintillator (direct detection) than neutron detection (indirect reaction due to scattering).
In summary, some suggestions for improving the count accuracy of a plastic scintillator neutron detector:
1) Use a discriminator between the detector and counter to filter out low level garbage.
2) Make proper use of RF shielding.
3) Use a lead foil shield to screen out low energy gamma and X-radiation.
4) Optimze the size of the scintillator as a compromise between detection efficiency and inaccuracy due to multiple scattering.
A lot of the information necessary to make these choices can be found in Marion and Fowles's Fast Neutron Physics, Part 1, and William J. Price, Nuclear Radiation Detection, 2nd Edition.
Created on Tuesday, January 16, 2001 2:14 AM EDT by Richard L. Hester