What is charge sharing and how is it addressed in Direct Conversion photon counting detectors?
When an X-ray photon interacts with the active layer of a photon counting detector, electron-hole pairs are created via the photoelectric effect. To be able to read out a corresponding signal, a high-voltage bias is applied to the active layer which accelerates the charges towards the read-out electronics. This electric field and the electrical repulsion of identically charged particles, together with their thermic energy, lead to a diffusion. This diffusion causes a lateral spread of the initial charge distribution over time.
Why is charge sharing a problem?
In photon counting detectors, the energy of an incident photon is correlated with the net sum of the charge in the primary charge cloud. An illustrative depiction of an initial charge cloud is shown in figure 1a).
Figure 1: Illustrative example of charge sharing. a) Initial charge distribution in three pixels A, B and C after X-ray photon interaction. b) Charge cloud after diffusion. c) Corresponding signal levels in the three pixels A, B and C. Due to charge sharing, the signal in pixels B and C does not exceed the threshold (dotted red line) and the signal is discarded. The initial charge cloud would have led to a valid signal over the threshold (dotted blue line).
Thresholds can be used to distinguish a signal over a certain noise level, but also to discriminate incident photons of different energies. If a certain part of the charge cloud is diffusing to the read-out electronics of a neighbouring pixel, this results in the detection of two events with lower energy than the primary photon, as depicted in figure 1b). Furthermore, if the resulting charge in one of the affected pixels is smaller than the threshold, the event is discarded as noise, as shown in figure 1c). In general, this leads to the underestimation of the energy of incident photons. The registration of one incident photon in several pixels degrades spatial resolution, as the information about the primary interaction is smeared out. Furthermore, this effect leads to degradation of energy resolution due to the general underestimation. Especially in medical applications, charge sharing reduces the dose efficiency, meaning that the useful proportion of the incident dose for imaging applications is reduced.
How can charge sharing be reduced?
There are several approaches to correct for the degradations that are introduced by charge sharing. One possible correction approach is performing a deconvolution in the signal domain, taking into account a per-pixel measured and calibrated detector response function.
Direct Conversion detectors have embedded hardware-integrated charge sharing correction that is able to record correlation signals from neighbouring pixels in real-time and assign back to the pixel which received the initial maximum interaction. This allows our detectors to deliver improved energy resolution especially in dual-energy imaging.