Monte Carlo simulations as a feasibility tool for non-metallic land-mine detection by thermal-neutron backscattering

The use of Monte Carlo simulations is presented for modelling a simplified land-mine detector system with thermal neutron backscattering (TNB) analysis based on a Cf-252-neutron source. Different aspects and a variety of external conditions, related to localisation and identification of a buried object have been investigated. In particular, the influence of moisture in a formation has been assessed, as moisture can be a serious interference for hydrogen as an indicator for explosives. The results of sensitivity calculations confirm that land-mine detection methods, based on an analysis of TNB may be applicable in homogeneous formations with low porosity provided that pore-water remains <5% by weight. In dry limestone. the TNT-based explosives can be well distinguished from other hydrogen-rich materials, except wood. However, in dry siliciclastic sands TNT explosives and wood are distinguishable. (C) 2002 Elsevier Science Ltd. All rights reserved.