A spectroscopic Monte-Carlo model to simulate the response of pixelated CdTe based detectors

Abstract

A fully spectroscopic Monte Carlo model has been developed to predict the spectroscopic performance of pixelated CdTe based detectors. The model incorporates photon attenuation by the photoelectric effect, Compton scattering and Rayleigh scattering. Charge transport equations are used to simulate the size of the electron cloud, approximated by a symmetrical two-dimensional Gaussian distribution, as it drifts to be read out at the detector anode. Direct comparisons are made between simulated data and an experimentally acquired spectra from a 1 mm thick CdTe sensor coupled to the HEXITEC detector ASIC. The probability of an absorbed photon leading to charge sharing across pixels as a function of incoming photon energy is investigated. The charge cloud size was found to be dominated by cloud growth during drift for photon energies < 100 keV.Furthermore, the portion of charge sharing events due to fluorescence from within the CdTe sensor is calculated -these events are distinguished from regular charge sharing events since their energy response differs. The model described is shown to give a good estimate of the total probability of charge sharing for energies up to 140 keV. CdTe sensor thickness, bias voltage, pixel size and electronic noise threshold can be adjusted to model a range of detector architectures.

Charge Sharing and Charge Loss in High-Flux Capable Pixelated CdZnTe Detectors

“…Flood images were taken with the same sources listed in Table 1 , providing data at the same photopeak energies. In a previous study [ 26 ], additional data was collected with the CdTe detector using an X-ray tube which provided data at three more peaks (32.5, 38.0 and 45.8 keV). Frame occupancy rates for all the CdTe detector data was kept below 1%.…”

“…A Monte-Carlo model has been developed to simulate the spectroscopic performance in pixelated CdTe detectors and has shown good agreement with experiment [ 26 ]. This model has been extended to work with CdZnTe and was used in this study to simulate the response of a 2 mm CdZnTe detector to various photon energies covering the range given by the photopeaks in Table 1 .…”

“…If fluorescence photons are generated due to the photoelectric absorption of the incoming photon, there is a high chance of the fluorescence escaping the incident pixel and stopping in an adjacent pixel, resulting in a fluorescence shared event. For the CdTe detector, at an incoming photon energy of 59.5 keV, ∼12% of all adjacent bipixels and ∼65% of all tripixels contain a shared fluorescence photon [ 26 ]. It is useful to distinguish between shared events that result from a single cloud centre and those due to fluorescence as their energy response can be quite different—with fluorescence shared events typically showing better energy resolution as they are less affected by charge loss [ 37 ].…”

“…The CdTe results are from [26,36] after re-analysis to include diagonal bipixels-joined by a dotted line to help visualisation.…”

“…These serve as a reference to compare performance with a typical CdTe detector used in many of the applications also intended for the HF-CdZnTe detector. Furthermore, a Monte-Carlo detector model [ 26 ] was used to complement the experimentally-determined charge sharing rates and to estimate the contribution of charge sharing in an ideal noise free scenario.…”