A linear separation method for neutron/gamma discrimination with organic scintillators

A linear separation method (LSM) is proposed to discriminate neutrons from γ-rays. Pulses from an EJ-301 liquid scintillator are used to examine the method. Results show that LSM works well on both the pulse dataset after dimensionality reduction by principal component analysis and the original dataset. Both figure of merit (FoM) and probability distribution can be calculated to evaluate the discrimination performance. FoMs of LSM and charge comparison method (CCM) for the training dataset are 1.576 and 1.366, respectively. LSM also gets higher FoMs than CCM for divided energies. FoM for the testing dataset is similar to the training dataset. For a low trigger threshold dataset, FoMs of LSM and CCM are respectively 1.135 and 0.995. •A linear separation method (LSM) is proposed for n/γ discrimination.•LSM can process both the pulse dataset after dimensionality reduction and the original dataset.•No other algorithms are needed for feature extraction or discrimination.•LSM works better than charge comparison method.