Calibration strategy for 3-D position sensitive CdZnTe spectrometer arrays

A new detector system having two layers of three by three 3-D position sensitive CdZnTe gamma-ray spectrometers is being developed. Each detector module consists of a 1.5times1.5times1.0 cm³ pixellated (11times11 pixels) CdZnTe detector and a VAS_UM2.3/TAT4 ASIC front-end board. Using the 3-D position sensing technique, each detector can be electronically divided into 11times11times40 (40 depth slices) small voxels. The detector response within each voxel is reasonably uniform (variations in weighting potential and electron trapping are minimized) and the energy resolution can be very good (~1% FWHM at 662 keV). A calibration matrix can be derived based on the voxel response so that the performance of entire detector can be very close to the performance of small voxels. However, each voxel needs a reasonable amount of events for the calibration, thus the number of calibration events needed for the whole detector is considerably large - ~50 million for each detector. This requirement poses a challenge to the calibration of the system especially for large arrays. Simply reducing the readout dead time for each event may not be the ultimate solution due to other bottlenecks in the system such as data transfer rate and data processing speed. This paper will analyze the bottlenecks in the current dual-array system and discuss possible strategies to speed up the calibration process.