Encrypted login | home

Program Information

Design and Signal Optimization of Multi-Point Plastic Scintillator Detectors Using a Novel Hyper-Spectral Deconvolution Method

no image available
H Linares Rosales

H Linares Rosales1*, B Wallace2 , R Roy2 , L Archambault3 , S Beddar4 , L Beaulieu3 , (1) CHU de Quebec-Universite Laval, Quebec, Quebec, (2) Universite Laval, Quebec, Quebec, (3) CHUQ Pavillon Hotel-Dieu de Quebec, Quebec, QC, (4) UT MD Anderson Cancer Center, Houston, TX


SU-K-205-6 (Sunday, July 30, 2017) 4:00 PM - 6:00 PM Room: 205

Purpose: This study is devoted to explore and optimize the detection limits related to the number of scintillating elements emitting in different spectral regions for multi-point plastic scintillator detectors (mPSDs).

Methods: The optical chain allowing the scintillator signal collection was simulated with an in-house code. The code takes into account the optical properties of various components: filters, dichroic mirrors, scintillating and coupling optical fibers and photomultipliers tubes. The response of each component was experimentally determined and used as input.The numerical optimization of the signal is then used to obtain the best combinations of sensitive materials and light transport system that will produce maximum light collection. A multivariable linear regression with likelihood maximization method was used to properly de-convolve each scintillator contribution to the total light signal. Theoretical mPSDs composed of 3 up to 30 scintillating elements were explored to establish the limit criteria for precise signal identification. A 3-points mPSD was constructed for experimental validation.

Results: The total number of scintillating elements that can be used on one single collecting fiber depend directly on the emission spectrum FWHM and the spectral separation of the scintillating elements. For 20 nm FWHM and 10 nm spectral separation, the maximum number of points is 30, while for 70 nm FWHM (e.g. usual scintillators) and 20 nm spectral separation decreases to approximately 12. The statistical spectral deconvolution algorithm properly extracted the relative amplitude of individual components for all configurations. Comparison against experimental measurements for the 3-points mPSD showed agreement within 1% difference with the algorithm at all wavelengths (350-750 nm).

Conclusion: This study presented a methodology to explore and optimize mPSD designs as well as establishing a theoretical upper limit of mPSDs that can be used for complete signal identification. An optimized 3-point mPSD was shown to experimentally performed as predicted by the model

Contact Email: