Thick-target neutron yields are, evidently, of great interest for hybrid reactor designs. As mentioned above, most measurements resort to a measure of the number of neutrons exiting from a large piece of material. The recent measurements of Hilscher et al. [110] use a large, gadolinium loaded, liquid scintillator. Such detectors allow an event by event measurement of the neutron multiplicity, with a very high detection efficiency. However, the

neutron energy is not measured and the detection efficiency does depend on the neutron energies, especially above 10 MeV. In the Hilscher work an average neutron detection efficiency of 0.85 was assumed.

Independently of the difficulties mentioned above for making a direct comparison between these measurements and the results of a calculation, it is not trivial to define an optimal size for the target. If the target is too small, the initial cascade may not be contained. This is exemplified in figure 6.8 which shows that the multiplicity measured[40] depends on the length and diameter of the target. It is possible to find a target thickness and diameter such that the multiplicity saturates. However, this does not guarantee that the measured multiplicity will be exactly the number of neutrons emitted per incident particle. This is due to two opposite effects: (a) neutrons may be absorbed in the target, even after being reflected from the detector, and (b) neutron multiplication may already be effective in the target. Lead targets should be largely immune from both effects, although Hilscher et al. find that, using a very pure lead target, as compared with

1.22 GeV p + Pb

their standard one, the neutron multiplicity is increased by approximately 2.5%. Figure 6.9 shows the ‘asymptotic’ results obtained by Hilscher et al. Uranium, on the other hand, is a multiplying medium whose кж can be estimated from equation (3.75) and table 3.2, with a value of v = 2.3: k1nat = 0.29 corresponding to a total multiplication in an infinite medium of 1/(1 — 0.29) = 1.4. This effect is probably responsible for the much higher neutron multiplicities observed with uranium as compared with lead, as can be seen by comparing figures 6.9 and 6.10. Another difficulty in interpreting the data, in view of an application to accelerator driven systems, is that the energy of the neutrons and thus their ‘importance’[41] is not known. Therefore, the results of direct thick-target multiplicity measure­ments should not be taken as source data to be input into a neutron propa­gation code, but rather as benchmarks for selecting valid INC calculations. Such calculations have been done by Hilscher et alJ who find that, as a very good agreement between their measured value for their lead target and a HERMES calculation is obtained, the same calculation yields

30.5 n/p reaction/GeV for a target 100 cm long and 150 cm in diameter.