Award Date


Degree Type


Degree Name

Master of Science in Engineering (MSE)


Mechanical Engineering

First Committee Member

Alexander Barzilov

Second Committee Member

William Culbreth

Third Committee Member

Yi-Tung Chen

Fourth Committee Member

Gary Cerefice

Number of Pages



Existing nuclear stockpiles and weapons-making capabilities imperil the global community. Current nonproliferation efforts involve the research and development of newer, more efficient detection systems that can be deployed for the interdiction and monitoring of special nuclear materials (SNM). Spontaneous and induced fission events in SNM produce neutrons and gamma rays, which can be detected and analyzed, in particular, using scintillator detectors. Various electronic data acquisition systems and data analysis methods have been employed to record and characterize neutron and photon signatures. The goal of this thesis is to develop a new method of discrimination between neutrons and photons in the CLYC elpasolite scintillator detector. Because neutrons and photons interact uniquely with scintillator materials, they generate scintillation light decay signals of different time profiles. Several conventional and digital pulse shape discrimination (PSD) methods exist to exploit the different features of detector signal waveforms caused by the different time profiles of the scintillation decay. They can be categorized on the basis of their implementation: time domain only or time and frequency domain. In this study, wavelet analysis is implemented in the time domain. When the discrete wavelet transform is applied to each pulse, the Haar wavelet is sampled over the signal to generate a set of coefficients, which are then further analyzed using numerical integration. The wavelet-based signal analysis code was written in Matlab. The code processes a single detector waveform at a time. It first applies the discrete wavelet transform to smooth the waveform, and then calculates the power of this signal. After performing partial integrations on different parts of the coefficients’ curve, it calculates the radiation identification (RID) value that serves as a threshold for neutron-gamma discrimination. Beyond the identifying threshold, the signal is categorized as a neutron event; below it the signal is categorized as a photon event. Plots of the total integral under the power of the waveform versus RID values and counts versus RID values were generated to visualize the PSD properties; the figure of merit (FOM) of the PSD method was calculated to quantify the quality of neutron-gamma discrimination. The PSD algorithm and the Matlab code were developed and tested on the set of neutron and photon waveforms experimentally measured with the CLYC detector and excellent neutron-gamma separation was observed.


gamma ray; neutron; pulse shape discrimination; radiation detection; wavelet


Mechanical Engineering | Nuclear | Nuclear Engineering