FPGA Implementations and Performance Comparisons of Pulse Shape Discrimination Algorithms for Neutron Detection

In this study, robust pulse shape discrimination (PSD) algorithms for neutron detection are implemented and evaluated with the target to FPGA platforms. The penetration characteristics of neutrons have made them indispensable in many fields, including nuclear physics, radiobiology, material science, and defense applications. However, neutrons are not easily detected due to the inevitable interference of gamma-induced signals. Thus, a PSD technique is commonly required to differentiate neutron and gamma radiation. A wide range of PSD algorithms have been developed yet remain constrained to offline analysis. While FPGA provides the necessary resources for realtime processing, related studies are currently fragmented and insufficiently in-depth. To fulfill that gap, this research presents an experimental evaluation of the performance of four robust PSD algorithms implemented on FPGA. The methods under test include zero-crossing (ZC), charge comparison (CC), falling edge percentage slope (FEPS), and partial charge-to-peak ratio (PCPR). A large and open-access dataset was utilized to test the designs. The performance analysis for each method is based on two criteria: discrimination quality and resource consumption. The results showed that the CC method offers high discrimination efficiency at the expense of resources, while the FEPS method trades off discrimination for lower resource consumption. The ZC and PCPR methods achieve a good balance between the two criteria; however, certain factors need to be carefully considered when deploying practical applications.