Speaker
Description
Gamma-ray spectroscopy of radiation fields is constrained by pulse pileup. Conventional acquisition systems rely on pileup rejection (PUR), which maintains energy resolution while introducing significant dead-time. This is unacceptable when measurement times are limited and statistical accuracy is crucial. To enable spectroscopy of high activity radiation fields, we developed a pulse pileup correction algorithm that restores pulse heights from pileup events. The Real-Time Segmentation (RTS) algorithm isolates individual events by selecting samples around a trigger, based on sampling rate and rise-time of the pulse. Pulse heights are determined by taking the difference between samples on the amplitude and baseline. RTS was derived on synthesized pileup with a waveform simulator, and was optimized for LaBr$_3$(Ce) detectors for their energy resolution and stable rise-times. Application of the RTS to synthetic pileup pulses showed a significant improvement against the trapezoidal shaping approach. To establish performance of RTS, waveforms were acquired from $^{60}\text{Co}$, and $^{137}\text{Cs}$ sources using a commercial digitizer at count-rates up to $713$ kcps. Spectra generated by the digitizer using trapezoidal shaping and PUR were compared to spectra generated by an offline implementation of RTS. The spectra generated by RTS increased counts in comparison to the commercial system, up to $20\times$ for $^{137}\text{Cs}$ during the $713$ kcps measurement. At $713$ kcps, the $^{60}\text{Co}$ peaks were suppressed by the commercial system due to PUR, but were resolved by RTS. RTS was then implemented for an Field Programmable Gate Array on a custom acquisition module. Preliminary spectra generated by the RTS algorithm in real-time from a $^{137}\text{Cs}$ source at count-rates up to $40$ kcps show that the RTS algorithm can run in real-time with minimal degradation to resolution. Ongoing work focuses on further optimizing for improved resolution, and performing higher rate measurements for validation against the results from offline analysis.