Alexis Poitrasson-Riviere

Dual-particle imager for standoff detection of special nuclear material

An advanced dual-particle imaging system is being developed for standoff, passive detection of special nuclear material. This system consists of three detector planes and will be capable of imaging both photons and fast neutrons. The ability of the system to detect fast neutrons makes it more difficult to effectively shield a threat source. This feature has an advantage over the commonly used Compton-camera systems, which are only sensitive to photons. Additionally, the detection of fast neutrons will allow for increased performance in regions with high levels of photon background radiation. The first two planes of the system consist of EJ-309 liquid scintillators and the third plane consists of NaI scintillators. This detector/plane combination allows image reconstruction using both neutrons and photons. In the liquid scintillators, neutron interactions are distinguished from photon interactions using an optimized pulse shape discrimination technique. The Monte Carlo transport code MCNPX-PoliMi has been used for the initial studies of this system due to its ability to track detailed information on interactions of interest and time-correlated particle production. This information has been used to optimize system parameters and has also allowed for investigation of image reconstruction techniques including simple backprojection and maximum likelihood expectation maximization (MLEM). A small-scale prototype is being developed for testing and validation of the simulations. This paper will analyze preliminary measurements and will also discuss simulations of several shielded source scenarios.

Digital data acquisition and processing for a neutron-gamma-ray imaging system

A digital, data-acquisition system for use with a large number of detectors was set up at the University of Michigan. Fast waveform digitizers from CAEN Technologies with 8 channels were synchronized to create a fully scalable system, with a current set-up of 32 channels. While some of the systems limitations are still being investigated, the excellent time resolution of the system enabled accurate time-of-flight measurements at the Los Alamos Neutron Science Center (LANSCE).

Monte Carlo investigation of a high-sensitivity two-plane Compton camera for long-range detection of SNM

Compton cameras have been used for astronomical and medical imaging applications as early as the 1970s. Recent interest in their potential for the detection and localization of special nuclear material (SNM) has led to increasing investigations. In this work, a specialized algorithm was developed for the optimization of a two-plane Compton camera. The MCNP-PoliMi code was utilized to simulate photon interactions within the detectors and coupled with an analytical technique used to estimate angular uncertainties. Using our specialized algorithm, a large area (approximately 1 m2) Compton camera consisting of two planar arrays of photon detectors was evaluated for several scintillators: LaBr3, CaF2, NaI, and plastic (C9H10). Optimization of plane thickness and voxel size were conducted for Compton camera efficiency and angular uncertainty at source energies of 100-400 keV.

Image reconstruction of shielded mixed-oxide fuel using a dual-particle imaging system

The dual-particle imaging system being developed at the University of Michigan was used at the Joint Research Centre in Ispra, Italy for measurements on samples of special nuclear material. A 1,150-g mixed-oxide (MOX) fuel sample was measured with various shielding configurations to determine how the presence of lead and/or polyethylene shielding degrades the systems ability to localize a source by simultaneous neutron and photon imaging. Three two-hour measurements were taken with the source shielded by an: a) 8-mm lead sheath and 5.1-cm lead bricks, b) 8-mm lead sheath and 6.5-cm polyethylene bricks, and c) 8-mm lead sheath, 5.1-cm lead bricks, and 6.5-cm polyethylene bricks. The 8-mm lead sheath was used in all cases to reduce the measured photon count rate. A “bare” measurement was also made by using only the 8-mm lead sheath, but the unexpected presence of additional sources has rendered the measurement unsuitable for comparison. The resulting images show that the dual-particle imaging system is able to accurately localize the MOX canister in the presence of intervening material.

Time-of-flight neutron spectrum unfolding for mixed-oxide nuclear fuel and plutonium metal using a dual-particle imager

A dual-particle imager with sensitivity to both neutrons and photons has been developed for the detection of special nuclear material. This system has the capability to provide spectral information on detected sources that is useful for safeguard applications. In nuclear facilities with high photon background rates, neutron detection can aid in identification and verification capabilities. The dual-particle imager creates a reconstructed neutron spectrum by correlating neutron counts in pairs of liquid scintillators. The reconstructed spectrum, a combination of energy deposited and time-of-flight, contains resolution effects that blur the true, emitted source-spectrum. To reduce the impact of resolution effects, the reconstructed spectrum can be unfolded to create an estimated source spectrum. A statistical technique, maximum-likelihood expectation-maximization, has been shown to produce an unfolded result for Cf-252 and Am-Be. This technique requires a system response-matrix that contains reconstructed spectra for single neutron energies. Since measuring mono-energetic neutrons is difficult, the system-response-matrix is simulated. Monte Carlo modeling with MCNPX-PoliMi has been show to produce an accurate system response. This work expands on previous neutron unfolding results by showing results for mixed-oxide fuel and plutonium metal measured with the dual-particle imager. These results also demonstrate the ability of the dual-particle imager to discrimination between spontaneous fission neutrons and (α,n) reaction neutrons.

Large-scale Compton-camera simulations, validation experiments, and image reconstruction

Recent efforts in nonproliferation and homeland security areas have focused on designing systems to accurately detect and locate radioactive material. This task is especially challenging when trying to locate material at large distances. One solution that has been under investigation is the Compton camera. In this work, the performance of a large-scale, two-plane Compton camera is investigated using different imaging reconstruction methods on simulated data. The simulation methodology is being validated by measurements conducted using a small-scale prototype of the system in the laboratory environment. The geometry consists of two planar arrays of scintillation detectors.

Image reconstruction using a three-plane, dual-particle imager for standoff detection of special nuclear material

An advanced, dual-particle, imaging system is being developed for standoff, passive detection of special nuclear material. This system consists of three planes of detectors and will be capable of imaging photons and fast neutrons simultaneously. The first two planes of the system consist of EJ309 liquid scintillators, while the third plane consists of NaI detectors. The EJ-309 detectors have excellent pulse shape discrimination capabilities, which allows the imaging system to distinguish incident photons from neutrons. Preliminary investigations of the dual-particle imager focused primarily on the use of simple-backprojection methods for image reconstruction. However, efforts have been made toward implementing the use of a maximum-likelihood expectation-maximization algorithm for image reconstruction. MCNPX-PoliMi has been used to simulate system response matrices for the dual-particle imager. Results have been promising for image reconstruction using measurement-specific system response matrices. However, field use requires the ability to detect and locate unknown sources in unknown shielding configurations. This has resulted in the development of an algorithm capable of generating a measurement-specific system response matrix on the fly.

Stochastic image reconstruction for non-proliferation applications

The stochastic origin ensembles (SOE) method has been implemented for the reconstruction of neutron and gamma-ray images from a dual-particle imager (DPI). SOE image reconstruction is an iterative stochastic process that creates images using Markov-chain Monte-Carlo sampling with event origins from the reconstructed backprojection cones. This method has been shown to produce image quality comparable to that of the widely implemented maximum-likelihood expectation-maximization (MLEM) algorithm. SOE can be advantageous because the method is simple to implement and thorough definition of spatial efficiency or resolution is not required. The SOE algorithm also reaches a steady-state solution after iterating, as opposed to MLEM solutions, which can degrade with over-iteration. This paper shows photon and neutron images images reconstructed using the SOE algorithm for an experiment with multiple neutron and gamma-ray sources.

Simulations of the cosmic-ray-induced neutron background

A method for simulating the cosmic-ray-induced neutron background at ground level has been developed. This method was used to obtain the expected pulse height distributions, net neutron rates, and angular distribution using EJ-309 liquid scintillators, and has been validated using outdoor measurements. The simulations were carried out using the Monte Carlo transport code MCNPX-PoliMi combined with the source subroutine CRY. The absolute count rates for neutrons were then obtained from the simulations using an empirical scaling formula from literature. Preliminary measurements of the angular distribution of background neutrons were done using a collimated EJ-309 liquid scintillator, showing fair agreement with simulations.

Monte Carlo study of Compton-camera detection sensitivity

The primary issue regarding the proliferation of radioactive materials is their possible ill-intentioned use. Depending on the material, it could enable the construction of dirty bombs or even nuclear devices. Several detection systems have been engineered to help control the transport of these materials and to provide efficient detection capabilities. Compton cameras have been used in fields such as medical or astronomical imaging for nearly 40 years. The existing research on the Compton-camera concept is unfortunately not applicable to these applications: the energies and distances of interest are very different. We have designed a new way to simulate a two-plane Compton camera for nuclear nonproliferation applications using the MCNP-PoliMi code. The simulations include accurate background models and detector properties such as time and energy resolutions, and pulse-generation time. Energy spectra can be obtained for both planes, along with the back-projection images. In this work, we present a study on the sensitivity of various large-scale Compton-camera configurations using this simulation tool. The Compton-camera materials investigated are PVT and CaF2 for the scatter plane, and NaI and LaBr3 for the absorption plane. The planes optimized in previous work; the voxels of 2 inch × 2 inch were used throughout this work. Results are presented for the weakest detectable source (1.4 mCi for CaF2/NaI) at a 100-m standoff in a 60-s measurement for the various Compton-camera configurations.