9th International Workshop on Radiation Imaging Detectors


in Erlangen, Germany 22-26 July 2007


Björn Heismann (Siemens Medical Solutions, Germany):

Signal transport in Computed Tomography detectors

In Computed Tomography (CT) X-ray intensities are measured by large-scale solid-state detectors. The standard signal generation chain comprises a scintillator pixel array attached to a matrix of photo sensors which in turn is read out by analog-to-digital conversion electronics. Both the physics of signal generation and transport as well as the analog-to-digital conversion define the performance of the detectors and have to be studied carefully to optimize it. We have developed and validated a broad system model describing the multi-stage physical process. The first step comprises a Monte Carlo (MC) tracking of the primary X-ray quanta energy deposition, taking into account the relevant fluorescence and scattering processes. The second step models the transport of optical photons in the scintillator pixels formed by a solid-state bulk with surrounding back-scattering TiO2 walls. In a third step the individual events are integrated to a read-out signal and analyzed for their statistical properties. A multi-pixel set-up is used for the simulation to take into account the signal transfer between adjacent pixels.

The system model is verified by a comparison to optical measurements. A scintillator array is excited by a) a flat-field and b) a needle beam X-ray source. The emitted light-field is read out by a high resolution CCD sensor. A good agreement between simulation and experiments is found, with a typical deviation in the range of 5%. We use the model to evaluate several macroscopic figures of merit of a CT detector. This includes signal cross-talk and system modulation transfer function as well as the variance increase due to X-ray scattering and optical transport. We finally calculate the energy weighting function D(E). It is an important figure of merit for dual-energy spectral CT, telling us how much a quantum of energy E contributes to the overall detected signal. Significant deviations from the expected linear behavior are observed. They are explained by and linked to specific properties of the signal transport.