Abstract
XRF analysis allows for quick and accurate evaluation of chemical composition of wide range of materials, and is actively used in different fields of science and industry. The accuracy of a measurement depends on number of counts recorded to spectrum, and can be improved either by extending the measurement time, which is unwelcome especially when analyzing a large amount of samples, or by increasing the count rate of spectrometer. The latter however leads to parasitic effects due to pulse pile-up caused by detectors finite resolving time. When acquiring XRF spectrum at high count rates with semiconductor detectors pulse pile-up leads to distortion of recorded spectrum, in particular to appearance of so-called pile-up peaks. Pile-up peaks appear due to detector being unable to separate two or more incident quanta arriving in time interval less than detectors resolving time, and recording them as a single quantum with energy that equals to sum of incident quanta energies. If not accounted for pile-up peaks may lead to quantitative and qualitative errors while making composition calculations based on the recorded spectrum. In this article the statistic of pile-up peaks in XRF spectra, their dependence on spectrum acquisition parameters and ways of improvement of pulse pile-up rejection algorithms are considered. Experimental data was acquired using Elvatech Prospector 3 spectrometer. A high dependency of spectrometers resolving time on quanta energies was observed. Based off experimental data an improvement to pile-up rejection algorithm was proposed.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have