Feasible Energy Calibration for Multi-Threshold Photon-Counting Detectors Based on Reconstructed XRF Spectra

Abstract

Energy calibration is an important procedure for photon-counting detectors. In specific, energy calibration allows the connection between detector output and actual energy, and it aligns the uniformity of thresholds and pixels. Some calibration methods involve radioactive isotope sources or X-ray kilovoltage peaks (kVp), but these methods are limited because of low availability and poor accuracy. In this study, an optimized method that uses X-ray fluorescence (XRF) emitted from metal compounds as sources is proposed. With a CdZnTe multi-threshold linear photon-counting detector (PCD), experimental results from different data ( $\text{K}_{{ {\alpha }}}$ peaks, $\text{K}_{{ {\beta }}}$ peaks, and $\text{K}_{{ {\alpha }}}$ and $\text{K}_{ { {\beta }}}$ peaks) and different methods (kVp, XRF, and isotope) are analyzed and compared. A spectrum reconstruction method known as spectrum stripping is used to verify and reduce the inconsistency problem of the XRF calibration results from the distortion effects in the spectra obtained from the PCDs. Results show that the proposed method shows better accuracy and more uniform calibration than the methods without spectrum reconstruction.