KOR

e-Article

Energy and DOI Calibrations for High Spatial Resolution CZT Detectors
Document Type
Periodical
Source
IEEE Transactions on Nuclear Science IEEE Trans. Nucl. Sci. Nuclear Science, IEEE Transactions on. 61(1):518-527 Feb, 2014
Subject
Nuclear Engineering
Bioengineering
Application specific integrated circuits
Detectors
Calibration
Cathodes
Noise
Polynomials
Histograms
ASIC
charge losses
charge sharing
CZT
depth of interaction (DOI)
energy calibration
gamma radiation
Language
ISSN
0018-9499
1558-1578
Abstract
In this paper we investigate the energy calibration methods of a high spatial resolution CdZnTe pixel detector, across a wide energy range, along with a fine estimation of the depth of interaction (DOI). A $24 \times 24$ CdZnTe detector ($160~\mu\hbox{m}$ pixel size, $300~\mu\hbox{m}$ pitch) is successfully calibrated in energy, from 32 keV to 662 keV, dealing with three effects, which have a strong impact on the acquired data: charge sharing, dependence with the DOI and incomplete charge induction. The CdZnTe detector is bump-bonded to a VATA210P ASIC, which outputs the induced charge on the square of nine pixels and on the common cathode. The ASIC noise is characterized by an automatic adjustment of the internal threshold available for every pixel. The channels are linearized in charge units in order to improve the resolution when the complete induced charge is computed as the sum of the adjacent pixel contributions. For energy calibration, the variation of the photopeak histogram with the cathode-to-pixel ($C/P$) ratio is analyzed. Two calibration profiles are generated for 32 and 662 keV for a complete ${}^{137}{\rm Cs}$ histogram. The obtained resolution, full width at half maximum (FWHM), is 24% for 32 keV (7.9 keV) and 4% for 662 keV (30 keV). The higher charge sharing for 662 keV events into many pixels explains the different energy resolutions. The DOI calibration is performed using simulation tools. A linear function with an exponential term is found suitable to characterize the DOI. Finally, the parametric fit obtained by simulation is applied to the experimental data with good results.