The objective of this project was to develop and demonstrate a prototype field-deployable device to non-destructively identify the fillers of previously excavated unexploded ordnance (UXO) as high explosive or benign using the Associated Particle neutron Time-of-Flight (APnTOF) technique. This technique employs an electronically collimated tagged neutron beam from a novel compact, portable 14 MeV neutron generator system that can simultaneously provide three-dimensional (3-D) imaging of objects and register their elemental composition.

Neutron interactions with unknown material show promise for the non-destructive elemental characterization of surface- and subsurface-UXO. Given the highpenetrating power of neutrons, such neutron-based techniques have been actively pursued for more than two decades for the non-intrusive inspection of concealed explosives in objects ranging from small airline bags to large shipping containers. This methodology rests on the principle that explosives can be distinguished from each other and from innocuous materials by analyzing the quantities and ratios of carbon, nitrogen, and oxygen in the material. Neutron interrogation methods for fill material exploit one or both of two types of neutron interactions with nuclei, inelastic scattering and neutron captures, and then detect the induced element-specific high-energy prompt gamma rays. While these systems have some capability to characterize larger munitions (shells 90 mm and larger), several unresolved issues hinder the wide application of this technique. The most important one is that neutrons interact with all surrounding matter, not only with the interrogated material, leading to a very high gamma ray background in the detector. Systems requiring bulky shielding and having poor signal-to-noise ratio (SNR) for measuring elements are unsuitable. In particular, it was pointed out in a 2004 Phase I demonstration report of the Pulsed Elemental Analysis with Neutrons (PELAN) system that because soil contains many of the same elements (likely to be dominated by oxygen) that are inside a shell, the SNR is severely affected by the background. Further, these techniques afford no information on the position sensitivity of the detected elements, so reducing the background from clutter will be difficult, if not impossible.

The APnTOF technique applied in this project was based on using mono-energetic neutrons produced by accelerating deuterium ions into the tritium target of a neutron generator. This reaction produces neutrons and alpha particles of 14.1 and 3.5 MeV, respectively, which are generated nearly back-to-back relative to  the production site in the tritium target; this correlation is used to tag a specific fraction of the emitted neutrons. The neutrons within this cone, defined by the detection of correlated alpha particles by a built-in alpha detector in the neutron generator, interact with the nuclei of the interrogated object (often via inelastic scattering reactions), and emit element-specific prompt gamma rays. Gamma detectors, in turn, uncover these rays. Measuring the time difference between detecting the alpha particles and the gamma radiation gives the distance traveled by the neutron before it scattered from a nucleus in the interrogated object (14 MeV neutrons travel at 5 cm/ns, while gamma rays move at 30 cm/ns). The energy spectrum of the gamma rays provides a means for identifying the element that scattered the neutrons, and the time delay yields the position along the cone where the reaction occurred. This information, along with directional data from the position-sensitive alpha detector, enables the investigator to determine the site of origin of the particular gamma ray (carbon, nitrogen, or oxygen) within the interrogated volume. Since the energy spectra are obtained from alpha-gamma coincidences, the data acquisition system does not register gamma rays that do not fit into this correlation; thus, it excludes gamma rays from the soil surrounding a munition casing, those from clutter, and those from reactions with neutrons that have lost their energy in multiple collisions with surrounding matter (thermalized neutrons). The SNR for measuring the carbon, nitrogen, and oxygen signatures of filler material also simultaneously improve. The availability of compact sealed neutron generators with embedded alpha-particle detectors makes it promising for field deployment.

A laboratory system was set up to demonstrate the efficacy of the APnTOF technique for elemental analysis with high SNR. Using the neutron time-of-flight information, an object of interest was located by means of its characteristic prompt gamma ray emission signals without interference from nearby clutter signals. The following key conclusions were reached upon completion of the tasks:

• The API120 neutron generator system allows easy integration with data acquisition electronics and has a stable neutron output (<1% fluctuation).
• The peak-to-background ratios obtained with the APnTOF technique for carbon (4.43 MeV) and oxygen (6.13 MeV) were 22 and 10 respectively. These values are vastly superior to the ratio of 1.3 obtained for both the elements when the nanosecond time restrictions on the data acquisition system were removed as is the case with conventional neutron interrogation techniques.
• The UXO time slice correctly revealed the carbon signal from graphite in the three cases of non-overlapping and overlapping clutter without interference from the clutter signals. The system geometry and event timing is thus optimized.
• A 5-mm thick iron casing will reduce the intensities of gamma ray signals in the energy range 4-7 MeV by less than 10%. 

The proof-of-concept work was performed for run times between 0.5 to 1 hour. Long run times were required for adequate data representation because a single gamma ray detection system was used. Also, a large solid angle was necessary to record time and energy data from multiple objects in the defined beam. As a result, the object-of-interest, (the graphite block) was positioned 43 cm from the neutron generator and the gamma ray detector to center of graphite distance was 69 cm. Based on the MCNPX/MODAR neutron and gamma ray transport simulations, a compact system could be implemented to considerably reduce the interrogation time.

This advance will allow for the inspection of a specific element of volume, thus improving the selectivity, and hence, SNR over present neutron interrogation methods. As a result, the identification of small ammunition (20-40 mm shell diameters) will be enhanced. It will also offer elemental distribution maps of the selected voxel to better identify the type of munition. The APnTOF’s main advantage is its ability to suppress the background signal that is unrelated to the inspected area by imposing several conditions on the data acquisition system.