In high flux nuclear experiments, radiation detectors often face a signal "traffic jam" when multiple events arrive almost at the same time and their pulses overlap. This pile up distorts the energy spectrum, undermines neutron counting accuracy and increases the risk of misinterpreting experimental data.
A team from the Shanghai Institute of Applied Physics, Chinese Academy of Sciences, has addressed this problem with a bipolar cusp like digital pulse shaping algorithm implemented on a field programmable gate array. The approach converts raw detector outputs into precisely shaped pulses in real time, allowing the system to maintain energy resolution and count rate performance even under intense radiation fields.
Instead of relying on conventional trapezoidal or triangular shaping, the new method uses an unfolding synthesis technique to generate ultra narrow pulses. The narrower width helps slice through overlapping signals so the contribution from individual neutron or gamma events can still be reconstructed.
The algorithm is bipolar, which allows it to suppress baseline drift that typically appears at high count rates and can corrupt amplitude measurements. By stabilizing the baseline, the system recovers pulse heights more faithfully, which is essential for accurate energy determination and for separating different kinds of radiation.
Architecturally, the pulse shaping filter mainly uses adders and multipliers, making it well suited for deployment on FPGA hardware with limited logic resources. With this implementation, the system can process millions of events per second, enabling continuous, low latency radiation monitoring.
To adapt the shaping parameters to different detectors and experimental conditions, the researchers employed a multi objective evolutionary optimization scheme. Rather than manually tuning decay time constants, the algorithm searches parameter space automatically to find settings that balance energy resolution against pulse shape discrimination performance.
The optimization targets both clear separation of radiation types and robust operation at high count rates. This data driven tuning allows the same digital architecture to be retargeted for various scintillators or measurement scenarios without extensive redesign.
Experimental validation used a NaIL scintillator, a lithium doped sodium iodide detector that responds to both neutrons and gamma rays. In tests with a 241Am Be neutron source, the system achieved a figure of merit of 2.11 for neutron gamma discrimination.
This figure of merit indicates that neutron signatures form a distinct fingerprint in pulse shape space, well separated from the gamma ray background. The system can therefore classify events reliably while updating energy spectra event by event in real time.
The researchers emphasize that the work translates mathematically complex shaping concepts into an efficient hardware algorithm for practical nuclear measurements. By embedding the entire processing chain on FPGA, they show that high precision reconstruction and discrimination of pile up neutron and gamma signals is achievable under demanding count rate conditions.
Potential applications include nuclear security screening, where border monitoring systems must operate reliably in variable radiation fields. The technique could also enhance real time safeguards and diagnostics in nuclear power plants and improve certain advanced medical imaging modalities that rely on fast, high throughput radiation detection.
Research Report:Real-time reconstruction and discrimination of pile-up neutron and gamma signals via bipolar cusp-like pulse shaping in NaIL scintillators
