Pulse classification and waveform reconstruction of saturated signals in a LaBr_3(Ce) detection system at the SLEGS gamma beamline

摘要: To address pulse saturation and pile-up in the high-energy region of output pulses from a 1''×1'' LaBr_3(Ce) detection system under intense 3MeV and 4MeV γ-ray beams at the Shanghai Laser Electron Gamma Source (SLEGS), this work proposes an offline data-processing method integrating abnormal-event fine classification and saturated waveform reconstruction. Saturated-event candidates with downward-drifting characteristics in the pulse shape discrimination (PSD)-Energy distributions were first preselected, accounting for 28.70% and 25.32% of the total raw datasets, respectively. A secondary fine classification based on first-derivative time-domain features further removed residual pile-up events, yielding single saturated pulses that constitute 99.81% and 99.86% of the candidates for subsequent waveform reconstruction. Then, 21.20% and 23.85% of normal pulses were extracted for the average waveform modeling, and the double-exponential (DE) model was identified as the optimal pulse model. An adaptive piecewise weighted nonlinear fitting method was developed to reconstruct the saturated region using undistorted rising-edge and tail samples, enabling accurate recovery of waveform shape and peak amplitude for mildly to moderately saturated pulses and wide flat-top pulses above 11.4MeV. Experimental results from the 3MeV and 4MeV beams indicated that 74.58% and 74.47% of reconstructed events, respectively, maintained low corrected chi-square errors, and more than 99.9% of events within 4.3-11.1MeV were stably reconstructed. The proposed method extends the effective measurable energy range of the LaBr_3(Ce) detection system to approximately 27MeV, corrects the energy-dependent downward drift of the original PSD values, and restores the PSD values to a normal pure γ band centered at 0.343 with a range of 0.329-0.364. The equivalent energy resolutions (ER_eq) of the recovered spectra are 44.18% and 40.03%, respectively, markedly better than those obtained from a wide-range unsaturated NaI(Tl) detection system under the same beam conditions. Without hardware modification or digitizer-range upgrading, this method effectively recovers waveform, spectral, and PSD information from high-energy saturated events, providing a reliable offline data-recovery solution for large-scale high-flux and high-energy nuclear physics experiments.