Master of Science / Department of Mechanical and Nuclear Engineering / William L. Dunn / One of the most pressing threats facing the United States is the increasingly effective use of improvised explosive devices or IEDs. Many commonly used techniques to detect explosives involve imaging. The primary drawback of imaging is that it requires interpretation of one or more images from each target. Human interpretation requires extensive training and is subject to the chance of false-negatives due to human fatigue. To counter the threat posed by IEDs, the signature-based radiation scanning (SBRS) technology has been developed. The goal of this project is to create an automated system, with minimal operator assistance, that is capable of detecting at least a gallon-sized explosive sample from at least one meter away. It is hoped that this can be accomplished quickly, in less than 30 seconds, with high sensitivity and specificity. The SBRS technique is based on the fact that many classes of materials have similar stoichiometries. For example, many common explosives have characteristic concentrations of hydrogen, carbon, nitrogen and oxygen. As neutrons interact with a material, unique gamma rays are created based on the composition of the material. Specifically, in this work, the gamma rays from inelastically scattered neutrons and from thermal neutron capture are investigated. Two neutron detectors are also used, whose responses depend on neutron thermalization in and around the target. Response templates are created based on gamma-ray and neutron responses that are collected from targets that contain explosives,. These templates are developed under different conditions for many different explosive materials to create a library of templates. The collection of responses from an unknown target is compared to a subset of the library of templates using a figure of merit to distinguish benign from explosive targets. Preliminary experiments were performed at Kansas State University. A high-purity germanium detector (HPGe) was used to detect the gamma rays. Two neutron detectors, one covered with cadmium, were used to detect back-streaming neutrons. A 252Cf radioisotope source as well as a Triga Mk III reactor were used as neutron sources.

Master of Science / Department of Mechanical and Nuclear Engineering / William L. Dunn / Two types of high efficiency neutron detector arrays (HENDAs), a 32-channel planar device and two trenched devices, were constructed at Kansas State University (KSU) and characterized. These HENDAs are prototypes for a detector that will be developed for the Spallation Neutron Source (SNS), which is located at Oak Ridge National Laboratory (ORNL). The general design objectives of a proposal from the KSU Semiconductor Materials and Radiological Technologies (SMART) Lab to the National Science Foundation, which led to a grant that funded this research, were reached. A spatial resolution for the HENDA prototypes of approximately 120 μm was achieved. The prototypes had relatively fast response times of approximately 1.2 μs, with rise times of 300 ns for the planar device and 200 ns for the 100-μm deep trenched device. The thermal neutron counting efficiency of one of the trenched devices was measured to be about 12%. It is expected that the goal of a 50% efficient HENDA is attainable by making trenches contained within the trenched device deeper and by stacking modules in a sandwich design. The pulse heights produced by the HENDA prototypes were approximately 0.5 volt with noise levels of 13 mvolt, resulting in a signal to noise ratio of almost 40:1. The response of HENDA, when placed in the neutron beam from the tangential beam port of the KSU TRIGA Mark II was proportional to the reactor power from 2 kW to 512 kW. At 512 kW, the neutron flux is φ = 1.08x10[superscript]7 cm[superscript]-2 s[superscript]-1, and therefore HENDA can operate with negligible dead time at neutron fluxes beyond 107 cm-2 s-1. From the experimental results, HENDA is a valuable linear array detector and can be applied to experiments that are designed to study material properties and structures through methods such as neutron diffraction and imaging.

Master of Science / Department of Mechanical and Nuclear Engineering / William L. Dunn / Improvised explosives devices (IEDs) are the cause of many casualties worldwide. Current methods for detecting IEDs are insufficient. A signature-based scanning technique based upon the fact that explosives consist primarily of hydrogen, oxygen, nitrogen, and carbon is examined as a possible rapid, standoff method for detecting IEDs. Devices employing this method rely on a template-matching technique in which the detector responses acquired through neutron and photon interrogation are compared to responses from a known explosive. A figure-of-merit is calculated to determine how well the template and the unknown match. This thesis explores the feasibility of employing the neutron interrogation aspect of this method.

Master of Science / Department of Mechanical and Nuclear Engineering / William L. Dunn / Neutron dosimeters measure neutron doses but portable, real time, high efficiency, and gamma insensitive neutron dosimeters are not commonly available. Characterization of a newly invented neutron dosimeter, based on perforated semiconductor neutron detectors (SNDs) whose perforations are filled with neutron reactive material, was the main purpose of this research study. The characterization procedure was performed by both simulation and experiment. The Monte Carlo N-Particle (MCNP) transport code was used to model a boron-filled dosimeter and to study the responses when the dosimeter was located on the surfaces of a water phantom and an anthropomorphic phantom for parallel beams of neutrons having various energy spectra. A pair of detectors was modeled: one bare and one Cd-filtered. Dosimeter responses were normalized for a beam that would produce 1 mSv ambient dose equivalent if incident on the ICRU sphere phantom. Dosimeter responses were estimated at different positions on the torso and it was found that the responses are relatively insensitive to the placement on the torso. For 100% efficient detectors and for beam with a Watt spectrum incident from front to back of the phantom, the bare detector produces about 140 counts per [Mu]Sv and the Cd-filtered detector produces about 80 counts per [Mu]Sv. The experimental characterization study involves observing SND counts with the dosimeter placed on an anthropomorphic torso phantom and determining the corresponding neutron dose. A TLD pair method was used to determine the neutron dose on the surface of the phantom. The neutron reactive material of the dosimeter was [superscript]6LiF, which is different from that assumed for the modeled dosimeter. A bare dosimeter response collected over 10 min was 25113 [plus or minus] 158 counts and the corresponding neutron dose was measured to be 2.57 mSv. The Cd-filtered dosimeter response collected over 10 min was 23886 [plus or minus] 155 counts and the corresponding neutron dose was measured to be 2.32 mSv. The neutron dosimeters are capable of detecting doses in the [Mu]Sv range and above, and are anticipated to provide direct read-out in dose units in future using count-to-dose conversion factors for bare and Cd-filtered SNDs.

Doctor of Philosophy / Department of Mechanical and Nuclear Engineering / Douglas S. McGregor / A novel design of neutron sensor was investigated and developed. The perforated, or micro-structured, diode neutron sensor is a concept that has the potential to enhance neutron sensitivity of a common solid-state sensor configuration. The common thin-film coated diode neutron sensor is the only semiconductor-based neutron sensor that has proven feasible for commercial use. However, the thin-film coating restricts neutron counting efficiency and severely limits the usefulness of the sensor. This research has shown that the perforated design, when properly implemented, can increase the neutron counting efficiency by greater than a factor of 4. Methods developed in this work enable detectors to be fabricated to meet needs such as miniaturization, portability, ruggedness, and adaptability. The new detectors may be used for unique applications such as neutron imaging or the search for special nuclear materials. The research and developments described in the work include the successful fabrication of variant perforated diode neutron detector designs, general explanations of fundamental radiation detector design (with added focus on neutron detection and compactness), as well as descriptive theory and sensor design modeling useful in predicting performance of these unique solid-state radiation sensors. Several aspects in design, fabrication, and operational performance have been considered and tested including neutron counting efficiency, gamma-ray response, perforation shapes and depths, and silicon processing variations. Finally, the successfully proven technology was applied to a 1-dimensional neutron sensor array system.

Master of Science / Department of Mechanical and Nuclear Engineering / William L. Dunn / The purpose of this research was to develop a technology to quickly identify hidden explosive materials. The developed method needs to be performed at a standoff distance of approximately two meters or more, must have high sensitivity (low false-negative rate) and good specificity (low false-positive rate), and should be able to detect a minimum amount of approximately one gallon (15 lbs) of explosive material. In an effort to meet these goals, a template-matching procedure to aid in the rapid detection of hidden improvised explosive devices was investigated. Multiple photon-scattered responses are being used as a part of a multidimensional signature-based radiation scanning (SBRS) approach in an attempt to detect chemical explosives at safe, standoff distances. The SBRS approach utilizes both neutron and photon interrogation to determine if a target contains explosive material, but the focus of this thesis is on photon interrogation. Beams of photons are used to create back-streamed responses called signatures, which are dependent on the density and the composition of the target. These signatures are compared to templates, which are collections of the same signatures if the interrogated volume contained a significant amount of explosives. The signature analysis produces a single figure-of-merit. A low figure-of-merit indicates an explosive might be present in the target. Experiments have been conducted that show an explosive surrogate (fertilizer) can be distinguished from several inert materials using these photon signatures, proving these signatures to be very useful in this particular method of chemical explosive detection.

Master of Science / Department of Mechanical and Nuclear Engineering / William L. Dunn / This report investigates the feasibility of a standoff interrogation method to identify nitrogen-rich explosive samples shielded by other materials (“clutter”) using neutron beams from Cf-252 and deuterium-tritium (D-T) generator sources. Neutrons from the beams interact with materials in the target to produce inelastic-scatter gamma rays, and, after slowing down to thermal energies, prompt-capture gamma rays. By detection of these gamma rays, a response vector is formed that is used to calculate a figure-of-merit, whose value is dependent upon the contents of the target. Various target configurations, which include an inert-material shield and a sample that may or may not be explosive, were simulated using the MCNP5 code. Both shielding and collimation of 14.1-MeV neutron beams were simulated to produce effective neutron beams for target interrogation purposes and to minimize dose levels. Templates corresponding to particular target scenarios were generated, and their effectiveness at nitrogen-rich explosive identification was explored. Furthermore, methods were proposed yielding more effective templates including grouping target responses by density and composition. The results indicate that neutron-based interrogation has potential to detect shielded nitrogen-rich explosives. The research found that using a tiered filter approach, in which a sample must satisfy several template requirements, achieved the best results for identifying the explosive cyclonite (RDX). A study in which a 14.1-MeV neutron beam irradiated a target containing a shielded sample, which could either be explosive (RDX) or inert, yielded no false negatives and only 2 false positives over a large parameter space of clutter-sample combination.

Master of Science / Department of Mechanical and Nuclear Engineering / Douglas S. McGregor / Semiconductor neutron detector design, fabrication and testing are all performed at Kansas State University (KSU). The most prevalent neutron detectors built by the KSU Semiconductor Materials And Radiological Technologies Laboratory (SMART Lab) are comprised of silicon diodes with [superscript]6LiF as a neutron converter material. Neutron response testing and calibration of the detectors is performed in a neutron detector test facility. The facility utilizes diffraction with a pyrolytic graphite (PG) monochromator to produce a diffracted neutron beam at the northwest beamport of the KSU Training Research Isotope production General Atomics (TRIGA) Mark-II nuclear reactor. A 2-D neutron beam monitor can also be used in conjunction with the test facility for active calibrations. Described in the following work are the design, construction and operation of a neutron detector test facility and a 2-D neutron detection array. The diffracted neutron beam at the detector test facility has been characterized to yield a neutron beam with an average Gaussian energy of 0.0253 eV. The diffracted beam yields a flux of 1.2x10[superscript]4 neutrons/cm[superscript]2/s at 100 kW of reactor power. The PG monochromator is diffracting on the (002) plane that has been positioned at a Bragg angle of 15.5 degrees. The 2-D neutron detection array has been characterized for uniform pixel response and uniform neutron detection efficiency. The 2-D 5x5 array of neutron detectors with a neutron detection efficiency of approximately 0.5 percent has been used as a beam monitor when performing detector testing. The amplifier circuits for the 5x5 array were designed at the KSU Electronics Design Lab (EDL) and were coupled to a LabVIEW field-programmable gate array that is read out by a custom LabVIEW virtual instrument. The virtual instrument has been calibrated to produce a pixel response that varies by less than two percent from pixel to pixel. The array has been used for imaging and active monitoring of the diffracted neutron beam at the detector test facility. The following work is part of on-going research to develop various types of solid state semiconductor neutron detectors.

Doctor of Philosophy / Department of Mechanical and Nuclear Engineering / J. Kenneth Shultis / A method for deterministically calculating the population variances of Monte Carlo particle transport calculations involving weight-dependent variance reduction has been developed. This method solves a set of equations developed by Booth and Cashwell [1979], but extends them to consider the weight-window variance reduction technique. Furthermore, equations that calculate the duration of a single history in an MCNP5 (RSICC version 1.51) calculation have been developed as well. The calculation cost, defined as the inverse figure of merit, of a Monte Carlo calculation can be deterministically minimized from calculations of the expected variance and expected calculation time per history.The method has been applied to one- and two-dimensional multi-group and mixed material problems for optimization of weight-window lower bounds. With the adjoint (importance) function as a basis for optimization, an optimization mesh is superimposed on the geometry. Regions of weight-window lower bounds contained within the same optimization mesh element are optimized together with a scaling parameter. Using this additional optimization mesh restricts the size of the optimization problem, thereby eliminating the need to optimize each individual weight-window lower bound. Application of the optimization method to a one-dimensional problem, designed to replicate the variance reduction iron-window effect, obtains a gain in efficiency by a factor of 2 over standard deterministically generated weight windows. The gain in two dimensional problems varies. For a 2-D block problem and a 2-D two-legged duct problem, the efficiency gain is a factor of about 1.2. The top-hat problem sees an efficiency gain of 1.3, while a 2-D 3-legged duct problem sees an efficiency gain of only 1.05. This work represents the first attempt at deterministic optimization of Monte Carlo calculations with weight-dependent variance reduction. However, the current work is limited in the size of problems that can be run by the amount of computer memory available in computational systems. This limitation results primarily from the added discretization of the Monte Carlo particle weight required to perform the weight-dependent analyses. Alternate discretization methods for the Monte Carlo weight should be a topic of future investigation. Furthermore, the accuracy with which the MCNP5 calculation times can be calculated deterministically merits further study.

Doctor of Philosophy / Department of Mechanical and Nuclear Engineering / Douglas S. McGregor / Strengthening the crystal lattice of lanthanide halides, which are brittle, anisotropic, ionic crystals may increase the availability and ruggedness of these scintillators for room-temperature γ-ray spectroscopy applications. Eight dopants for CeBr[subscript]3, including CaBr[subscript]2, SrBr[subscript]2, BaBr[subscript]2, ZrBr[subscript]4, HfBr[subscript]4, ZnBr[subscript]2, CdBr[subscript]2, and PbBr[subscript]2, were explored at two different doping levels, 500ppm and 1000ppm, in an effort to identify potential aliovalent strengthening agents which do not adversely affect scintillation performance. All dopants and doping levels exhibited improved ingot yields over the undoped case, indicating an improvement in the ease of crystal growth. Scintillation performance was gauged using four key metrics. Scintillation emission spectra or, rather, radioluminescence spectra were recorded using x-ray irradiation. Total light yield was estimated through pulse height comparison with bismuth germanate (BGO) scintillators. Scintillation kinetics were checked by measuring single interaction pulses directly output by a fast response PMT. Finally, light yield proportionality was measured using a Compton coincidence system. Samples from each ingot were harvested to benchmark their performance with the four metrics. Of the eight dopants explored, only BaBr[subscript]2 and PbBr[subscript]2 clearly altered scintillation spectral emission characteristics significantly. The remaining dopants, CaBr[subscript]2, SrBr[subscript]2, ZrBr[subscript]4, HfBr[subscript]4, CdBr[subscript]2 and ZnBr[subscript]2, altered scintillation performance to a lesser degree. No dopant appeared to affect light yield proportionality, nor did any drastically alter the light decay characteristics of CeBr[subscript]3. HfBr[subscript]4 and ZnBr[subscript]2-doped CeBr[subscript]3 exhibited the highest light yields, significantly higher than the undoped CeBr[subscript]3 samples tested. Finally, aliovalent doping appeared to greatly improve CeBr[subscript]3 ingot yields, regardless of the dopant, thus it is a promising method for improving crystal strength while not deleteriously affecting scintillation performance. HfBr[subscript]4 and ZnBr[subscript]2 both demonstrated high performance without any noticeable negative side-effects and are prime candidates for future study.