Demonstrating a Novel, Hidden Source of Spectral Distortion in X-ray Photon Counting Detectors and Assessing Novel Trigger Schemes Proposed to Avoid It

Abstract

X-ray photon counting spectral imaging (x-CSI) determines a detected photon’s energy by comparing the charge it induces with several thresholds, counting how many times each is crossed (the standard method, STD). This paper is the first to demonstrate that this approach can unexpectedly delete counts from the recorded energy spectrum under some clinically relevant conditions: a process we call negative counting. Four alternative counting schemes are proposed and simulated for a wide range of sensor geometries (pixel pitch 100–600 µm, sensor thickness 1–3 mm), number of thresholds (3, 5, 8, 24 and 130) and medically relevant X-ray fluxes (106–109 photons mm−2 s−1). Spectral efficiency and counting efficiency are calculated for each simulation. Performance gains are explained mechanistically and correlated well with the improved suppression of “negative counting”. The best performing scheme (Shift Register, SR) entirely eliminates negative counting, remaining close to an ideal scheme at fluxes of up to 108 photons mm−2 s−1. At the highest fluxes considered, the deviation from ideal behaviour is reduced by 2/3 in SR compared with STD. The results have significant implications both for generally improving spectral fidelity and as a possible path toward the 109 photons mm−2 s−1 goal in photon-counting CT.