Neural network pulse shape analysis for proportional counters events

In this paper we present a method to select pulses generated in proportional counters by events with point-like or distributed deposition of energy. This has been developed in the context of the GNO experiment, which is measuring the low-energy solar neutrino flux using a 30-ton gallium detector at Laboratori Nazionali del Gran Sasso, Italy, via the inverse β reaction 71Ga(νe,e−)71Ge. As the expected signal rate is very low, an effective background rejection by means of digital pulse shape analysis is of paramount importance. The application of the conventional analysis allowed to reduce the average background rate from 0.5 to 0.06 counts/day; this is further reduced to 0.055 counts/day using this method. Besides the reduction of the background rate and the consequent improvement of the signal-to-noise ratio, this method proves to be more robust and flexible.