LNL Annual Report 2004

The neutron spectrometer discussed here consists of a windowless PIN diode coupled with a polyethylene converter (thickness 1 mm). Neutrons are detected through the recoil-protons generated in the converter via elastic scattering. The response functions composing the response matrix for unfolding the spectra of energy Ed deposited in silicon by recoil-protons were determined analytically [1]. The experimental verification of the response functions was performed at the LNL Van De Graaff accelerator by irradiating the device with monoenergetic neutrons. These irradiations also allowed the energy calibration of the detector. The minimum detectable energy for recoil-protons is set by energy deposition in silicon from secondary electrons produced in the detector assembly by gamma rays associated to the neutron field. This causes a tail in the low energy part of the Ed spectrum which overwhelms the distribution of recoil-protons. In particular, it was found [1] that high-energy electrons travelling not perpendicularly to the diode surface are responsible for energy deposition from about 150 keV up to the minimum detectable energy for recoil-protons. The maximum detectable energy is imposed by the thickness of the fully depleted layer and is about 6 MeV for the diodes under study with a full depletion layer about 300 m thick. The limited interval of detectable energies restricts the application of the spectrometer to low-energy neutron fields like the ones which can be produced at facilities hosting low-energy ion-accelerators. The possibility of lowering the minimum detectable energy through pulse-shape discrimination (PSD) was investigated [2] with a device based on a PIN diode in the “reverse-injection” configuration. The present work discusses the performance of this new spectrometer, by focusing on its capability of resolving continuous neutron spectra. The signals from secondary electrons can be discriminated from those due to low-energy recoil-protons by exploiting the different transit times of charge carriers in the silicon detector. The most favourable condition for pulse-shape discrimination is given by the “reverseinjection” configuration, i.e. by placing the polyethylene converter in contact with the N layer. The slower component of the electronic signal (i.e. the holes) induced by low-energy protons (those which have to be discriminated from secondary electrons) is more important in this case, because the charge carriers are initially concentrated far away from the collecting electrode (the P layer). A p-i-n diode 300 m thick, with an N dead layer of about 3 m was set up in the reverse-injection configuration for the experimental verification. The diode sensitive area is 3 mm. The diode was totally depleted with a reverse bias voltage of 30 V. The pulse-shape discrimination is based on the zero-crossing time of bipolar signals from a (CR)-(RC) fast filter. A low-noise preamplifier was realised at the Nuclear Engineering Department of the Politecnico di Milano to improve the discrimination capabilities of the whole chain. The electronic chain is described in details in ref. 2. The output signal from the low-noise preamplifier is sent: (i) to a spectroscopy amplifier connected to an ADC (energy chain); (ii) to a timing chain which provides the time spectrum of the signals and the veto for data acquisition. This set-up allowed to lower the minimum detectable energy to about 900 keV, as also confirmed by irradiations with monoenergetic neutrons at the LNL [2].