Digital analysis of scintillator pulses generated by high-energy neutrons. Jan Novák, Mitja Majerle, Pavel Bém, Z. Matěj1, František Cvachovec2, 1 Faculty of Informatics, Masaryk University in Brno, Czech Republic 2 University of Defence, Brno, Czech Republic High energy neutron detection • High energy neutron detection is generally based on the nuclear interaction basis. • Charged particles produced in the n reaction on the nuclei are detected by usual charged particle detection methods. • For this purpose, well known (n,p) elastic scattering is frequently used. pn2 p p2 p n1 2 2 2 p n 1 p n 2 p p 2 p p 2 p n 1 p n 2 p n 1 p n 2 cos E n1 E n 2 E p 2 ( Ei mic ) pi c mi c 2 2 2 2 2 4 Proton Recoil Telescope for Neutron Spectroscopy 209,0 mm ■ The Proton Recoil Telescope (PRT) is based on the detection of protons in the n+p scattering on a thin hydrogenated target. ■ By correcting for the energy dependent efficiency, the energy spectrum for the neutrons emitted by the source is determined from the detected proton spectrum, on the kinematics relations basis. ■ Protons are detected at defined angle by telescope of Si detectors. CH2 Φ 14,0 mm m 31.103 mg 12C Tube, SS, T 2 mm 44,4o 7Li 456,5 mm CH2(B) 30x30x60 mm3 Fe 30x30x60 mm3 Beam spot on Li target Φ(FWHM) 4 mm mm Si-D Φ 7,0 mm The PRT is proposed to detect neutron spectrum at short distances from the source where TOF method is not applicable 3 Simplified reconstruction of neutron spectrum ■ A code for the reconstruction of neutron spectrum from mesured proton recoil spectrum was developed. In the code, the response function is based on the n+p differential cross section knowledge (LA-150h library) and the n+p scattering kinematics. ■ The model of response functions was calculated including the realistic geometry arrangement of the neutron source-radiator-detector set. As a simplified approach to the reconstruction procedure, the effects of secondary scattered neutrons were neglected. En=15 MeV Response Function n/MeV n/MeV En=10 MeV Response Function 4,00E-008 4,00E-008 3,50E-008 3,50E-008 3,00E-008 3,00E-008 2,50E-008 2,50E-008 2,00E-008 2,00E-008 1,50E-008 1,50E-008 1,00E-008 1,00E-008 5,00E-009 5,00E-009 0,00E+000 0,00E+000 -5,00E-009 -5,00E-009 1 2 3 4 5 6 Ep (MeV) 4 5 6 7 8 9 Ep (MeV) ■ Response functions were superposed by optimization procedure to fit measured proton-recoil spectrum. Multiplication factors pertaining to response functions create resulting n spectrum. 4 NE 213 scintillator • • • In the NE 213 scintillator, protons are scattered by neutrons. Moreover, at higher neutron energy (over 8 MeV), number of another processes producing charged particles contributes to overall scintillator response. Protons and another charged particles excite atomic levels, whose number is proportional to energy loss of the particle. Therefore, ray caused by the deexcitation has an intensity proportional to energy lost in the scintillator. The light ray is converted to an electric pulse and amplified by the photomultiplier base assembly Ortec 265. Light gain dependency on the proton energy is empirically determined: Lp(Ep)=0.07269+0.11237 E2 Lp(Ep)= -0.20570 + 0.35260 E + 0.01343 E2 + 0.00250 E3 Lp(Ep)= -0.25999 + 0.34141 E + 0.3303 E2 + 0.0092 E3 Lp(Ep)= -1.43180 + 0.69325 E The response function shape is determined mainly by the (n,p) angular distribution and by the scintillator geometry. dN dE E n Max p p (E p ) 0 dw p dE p (En , E p ) dN n dE n ( E n ) dE n 0< E < 1.5 MeV 1.5 < E <3.5 MeV 3.5< E < 8.0 MeV 8.0 < E < 20 MeV Neutron-gamma discrimination • • • In the real experiment, gamma photons interact with the scintillator also. It is necessary distinguish neutron and gamma events. Gamma photons scatter electrons, therefore the atomic levels are excited by a different way. n-gamma discrimination using pulse shape analysis is feasible. 0.2 We have two possibilities of the neutrongamma discrimination: 1. Pulse rise time analysis. Pulses originated from neutrons and gammas are distinguished on the rise time duration. 2. Pulse shape analysis based on the comparison of the overall pulse area and the pulse part area beginning from a fixed sample. The ratio of two areas: r A0 A1 U[V] 0 -0.2 Sample -10 -8 -6 -4 -2 0 2 4 6 8 10 12 14 -0.4 -0.6 -0.8 Area of the pulse part A1 -1 -1.2 -1.4 -1.6 -1.8 Overall pulse area A0 Rise time 16 18 20 1. Pulse rise time analysis Ne 213 Converter PC Start HV PSA TAC Stop In the past, we performed two-parametrical data processing. Besides the pulse amplitude data collection, the pulse was processed by a second way. The PSA generates start and stop pulses, which are derived from 10% and 90% of the pulse amplitude. The TAC converts the time delay between start and stop pulse to amplitude. Hence, amplified NE 213 pulse amplitude and the amplitude containing the pulse rise time are together saved in the time coincidence. In the optimal conditions, we distinguished neutron pulses for neutron energies from 0.5 MeV. 2. Pulse shape analysis based on the areas comparison Present level of the digital signal processing enables us to process very short pulses by the digital oscilloscope technique. Our oscilloscopic card has 10 bits AD converter with maximal sampling frequency 500 MHz. The pulse information is given by some tenths of samples, because the maximal pulse duration is 100 ns. Ne 213 HV PC with card At the present time, all samples are recorded to the PC hard disk. Therefore, we are able to collect pulse samples with relatively low dead time at pulse frequency up to 1000 pulses/s. But now we try preprocess pulse samples before their transfer by PCI bus and saving on the hard disk. This way will lead to saving of the markedly lower parameter number. Advantage of this way will be achievement of the markedly higher pulse frequency. Moreover, the PC hard disc volume requirements will be decreased. Neutron-gamma discrimination results On the figure, the neutron and gamma pulses of the 252Cf neutron source are separated very well. Our NE 213 set-up was used, pulses were digitized by digitizer of Faculty of Informatics, Masaryk University in Brno. In our experimental set-up, the PuBe neutron source was used. The neutron and gamma pulses are clear separated. The A1 area is integrated from 8 th channel after pulse maximum, what is 16 ns after maximum. However, this method is extremely sensitive to signal-noise distance mainly due the A1 area integration. Count 1600 1400 1200 1000 800 600 400 200 0 0 10 20 30 40 50 Integration ratio Conclusions and visions Advantages of the pulse sampling: • • • • • High event frequency (in the mode of the pulse processing before saving). Possibility of the on-line data analysis for the experimental set-up correction. Possibility of the processing algorithm development on the saved samples. Absolute freedom in the processing algorithm choice. Portability. Challenges: • • • To solve the problem with the high neutron-gamma sensitivity to signal-noise distance. To implement the on-line pulse processing before the data PCI transfer and saving on the hard disc. To choose and to handle the optimal method for the neutron pulse spectra deconvolution in order to neutron spectra obtaining.