Quality Assurance Challenges in X-ray Emission Based Analyses, the Advantage of Digital Signal Processing

Abstract

There is a large scatter in the results of X-ray analysis with solid-state detectors suggesting methodological origin. In order to improve the methodology, detector response functions have been investigated by many researchers and analysts. This was necessary as the departure of the response function of some detector-signal processing electronics from the normally assumed Gaussian line shape can exceed 100% in area. Several detector models have been proposed to improve understanding and establish a firm basis for quantitative work. After reviewing some contradictory results, we describe a signal processor that offers quality assurance, by producing two spectra for each measurement. One is the normal spectrum of accepted events, while the second spectrum contains all of the rejected events. For each measurement, therefore, all X-ray events are recorded, enabling quality control. In addition to this improvement, the digital signal processor of Cambridge Scientific, Canada, delivers a high throughput rate, excellent resolution, decreased low energy tailing and a line shape justified by the physics of the detector. Comparative measurements are presented to demonstrate the improved rejection of background from gamma rays as well as a significant improvement in pile-up recognition. The rejected events spectrum gives insight into the origin of the response function, which suggests that the flat plateau of the frequently used Hypermet function, normally attributed to detector dead layers, originates from pile-up with the low energy noise events. A detailed analysis demonstrates how the relative intensities of the X-ray lines can change in a varying noise environment, thus potentially explaining the unacceptable large scatter in the experimental data currently found in the literature. The comparison of the accepted and rejected events adds the possibility of monitoring the electronic efficiency of signal recognition that has generally been ignored in this field.