Degenerate Pile-up Correction in Pulse Height Spectra from Gamma-ray Spectrometers

Abstract

One of the primary obstacles faced by spectrometers operating under high counting rates is pile-up, which occurs when two or more events are detected within a timelapse short enough to result in a superposition of the events waveforms. These can not hence be integrated separately in order to get their amplitudes. Piled-up events are typically identified using pile-up rejection or recovery algorithms. In the latter case, the constituent single waveforms and their amplitudes are also restored. However, there are instances in which the pulses overlap so closely that it is impossible to identify the occurrence of pile-up, resulting in the integration of these pulses into a single spurious event. This phenomenon is known as degenerate pile-up. A method to rectify the incorrect reconstruction of degenerate pile-up was developed, based on a statistical approach, which can be directly applied to the pulse height spectra distributions. The approach was tested on a number of synthetic spectra, with counting rates ranging from 20 kHz up to 1 MHz. The recovered spectra were compared to those purely analysed with a pile-up recovery algorithm, demonstrating an improvement of the reconstructed spectrum of several tens of percent when compared to the true synthetic counterpart.