No abstract

No abstract

We have performed a pulse-shape analysis of signals from Si detectors to identify low-energy charged particles with neural networks (NNs). We acquired pulse shapes of proton, deuteron, triton, 3 He and 4 He from a CH2 target bombarded by α particles. We trained the NNs using the pulse shapes for known particles and evaluated their particle-identification ability of the NNs. The NNs successfully distinguished helium isotopes from hydrogen isotopes, but could not separate the helium isotopes into 4 He and 3 He.