Contents lists available at ScienceDirect Applied Radiation and Isotopes journal homepage: www.elsevier.com/locate/apradiso Analysis of alpha particles spectra of the Radon and Thoron progenies generated by an electrostatic collection detector using new software C. Sabbarese a,c, ⁎ , F. Ambrosino a , R. Buompane a,c , M. Pugliese b,c , V. Roca b,c a Dipartimento di Matematica e Fisica, Seconda Università degli Studi di Napoli, Caserta, Italy b Dipartimento di Fisica “E. Pancini”, Università degli Studi di Napoli “Federico II”, Napoli, Italy c Istituto Nazionale di Fisica Nucleare, Sezione di Napoli, Italy ARTICLE INFO Keywords: Alpha-particle spectrometry Peaks deconvolution Radon-Thoron activity measurement RaMonA collection efficiency ABSTRACT A complete and detailed analysis of alpha spectra from the 222 Rn and 220 Rn progenies was performed by newly developed software. The software identifies the alpha peaks, performs appropriate fits and calculates the net area and its uncertainty, considering the entire background. The deconvolution of the overlapped peaks of 218 Po and 212 Bi allows us also to evaluate their minimum detectable area. The efficiency of the electrostatic detection method was recalculated and new useful considerations on the collected alpha emitters were made. 1. Introduction The measurement of the 222 Rn (Radon) and 220 Rn (Thoron) specific activity in air performed by a device based on electrostatic collection of the ionized descendants of the two gases is routinely utilized, but not all their potentialities are properly exploited and optimized. The spectra, which can be obtained with this apparatus, require accurate analysis to demonstrate these potentialities. This is the case also of the RaMonA system that detects the two gases and furnishes the complete spectrum of the alpha particles emitted by all their daughters deposed on a silicon detector surface and that was used in various studies and applications (Roca et al., 2004a; 2004b; Buompane et al., 2014). A typical alpha spectrum is reported in the Fig. 1, where the progenies of both Radon isotopes are present. The spectrum contains four principal peaks in the range 5–10 MeV. The second peak at 6.7 MeV and the third peak at 7.7 MeV are the completely resolved lines of 216 Po and 214 Po, suitable to determinate Thoron and Radon activity, respectively. The first peak depends on the presence of the non-resolved lines of the 218 Po at 6.02 MeV and of the 212 Bi at 6.09 MeV, due to the Radon and Thoron decay chain, respectively. The more energetic line, at 8.8 MeV, and the wide count distribution at its right side is due to the 212 Po and to the sum of the energies of the random coincidence into the detector of 8.8 MeV alpha particle and the beta particle coming by the second branch of the 212 Bi decay. When this coincidence happens, the signal corresponds to the deposed charge by both the alpha and the beta particle. A consequence of the prolonged use of one detector is the accumulation of an increasing amount of the long lasting 210 Pb (T 1/ 2 =22.3 y), responsible of the production of the alpha emitter 210 Po. In this case a peak at 5.3 MeV is also present in the spectrum. The analysis of such spectrum when just one of the two radioactive gas isotopes is present can be done fixing a ROI (Region of Interest) for each peak and counting the events in each one. In this case the response time of the instrument to the 222 Rn and 220 Rn isotopes would be 20 min and few minutes respectively (Buompane et al., 2014). In presence of both Radon and Thoron, a sharp separation of the overlapping 218 Po and 212 Bi alpha peaks is not possible. The activity of the 222 Rn could be obtained from the 214 Po line, and consequently the response time should increase to about 3 h and the ability instrument to quickly monitor the radon concentration variations would be compromised. Moreover, the background in the spectra due to the tail left of each peak becomes significant when the counts increase and it cannot be evaluated with the use of the sharp ROIs. For these reasons, an appropriate software capable to make an adequate signal processing should allow us to exploit the excellent energy resolution and the temporal fast response of the instrument and, also, to provide more complete and accurate information. Many studies have been carried out to perform the alpha spectrum fitting and the deconvolution of the overlapped peaks. A particular example was the application of an artificial neural network technique (Baeza et al., 2011); some others were based on semi-empirical mathematical functions (Baba, 1978; Wätzig and Westmeier, 1978; Bortels and Collaers, 1987; Martín Sánchez et al., 1996; Deyras, 2002). Several strategies were taken into account when doing the fit; hence different models for different aims were realized. For example, in some models Gaussian modified functions (Baba, 1978; Wätzig and Westmeier, 1978) or more complex functions were used, such as the http://dx.doi.org/10.1016/j.apradiso.2017.01.042 Received 29 June 2016; Received in revised form 20 January 2017; Accepted 27 January 2017 ⁎ Corresponding author. E-mail address: carlo.sabbarese@unina2.it (C. Sabbarese). Applied Radiation and Isotopes 122 (2017) 180–185 Available online 31 January 2017 0969-8043/ © 2017 Elsevier Ltd. All rights reserved. MARK