On the origin of the satellites peaks in alpha particle spectra

Abstract

The widespread study on silicon diodes performance in the spectrometry of charged particles is due to the possibility of their use as a research tool both in laboratory and in industrial applications as the measurement of the isotopic ratios between chemical elements. In order to employ this technique in studies related to reactor fuel elements and their properties after irradiation, we have been studying the response of silicon diodes for alpha spectrometry. The devices studied during this work were: a $3590-06 PIN photodiode (Hamamatsu) and an implanted silicon diode (type Al/n*/p/n/Al) manufactured at CERN. The diodes were housed inside a stainless steel vacuum chamber and its electric leads were connected to the charge sensitive preamplifier (based on an integrated circuit A250 from Amptek) whose output was further amplified and shaped and finally fed to a multichannel analyzer. In order to verify the performance of the $3590-06 photodiode (which bears two guard rings) for heavy charged particle spectrometry, several energy spectra were recorded using a 5.5kBq mixed alpha source of 29 Pu, 21! Am, and ?HCm. The experimental results showed that, even at room temperature and without reverse bias, the alpha particles of the principal group of each isotope were observed. Even though the good influence of the bias voltage on photodiode energy resolution was evidenced (FWHM = 17.4keV for the 5.486Mev line from 2 Am), the results revealed some unexpected low intensity peaks spaced about 200keV below each of the three main peaks. The origin of these satellites peaks were firstly thought to be due to incomplete charge collection in weak electric fields around the edges of the diode, near the guard ring region. So, one should expect that the relative intensity of these satellites peaks would increase with the number of guard rings of the diode. Having in mind this pers we carried on some alpha spectra using a new silicon diode, manufactured at CERN, which enabled us to x from one to ten guard rings. Surprisingly, even when ten guard rings were connected to the ground, the relative intensity of these peaks were about 1% compared to the true peaks. This behavior indicates that the true cause of these spurious satellites pulses might be not only associated with the weak electric fields around the guard ring region but also with changes on the entrance window absorption near the edge of the diodes. Further steps are needed to clear up this problem.