TU-C-I-609-10: Preliminary Study of Pixel Pitch and Effect of Divergence On Thick Cadmium Tungstate Detector

Abstract

Purpose: To theoretically study the effects of pixel pitch and beam divergence on 1 cm thick crystals for use in megavoltage cone beam computed tomography (MVCBCT). Method and Material: linear array (80-elements, each 0.275 × 0.8 × 1 cm3 in fan beam) has been proven to be a good candidate for MVCT in our lab. For this study, DOSXYZnrc Monte Carlo code was used to find the point spread function (PSF) as the distribution of energy deposited in a theoretical flat detector (30 × 30 × 10 mm3, 0.01 × 0.01 × 10 mm3 voxels). A pencil-beam of 6 MV photons was incident at a range of angles (0°, 5°, 10°, 15°, 20°) with respect to the normal. In each case, the detector PSF (ignoring optical spread) was convolved with a realistic source function to give the system PSF in the plane of the object. A realistic object to detector magnification of 1.4 was used in the system PSF calculations. The PSF at 0.01 mm pitch was re-binned into several pitches (0.5, 0.7, 1.0 mm) and Fourier transformed to obtain the system modulation transfer function (MTF). Results: This analysis suggests an optimum pitch of 0.5 mm which cannot be fabricated using because of the cleavage plane. However, the pre-sampled detector PSF is further degraded due to small optical leakage through reflective coating and scintillator-photodiode interface, and the scattered radiation from the object. So, in practice a pitch of 1 mm may be sufficient. The degradation of the system MTF with the increasing divergence is significant between 0° and 5°, and becomes large at 10°. Conclusion: Practical pitch of 1 mm is not ideal yet maybe sufficient. The use of flat panel photodiode arrays for MVCBCT is precluded due to divergence; instead a focused detector should be used.