Development of detectors for the diagnostics of ELI-NP photons beams

The ELI-NP (Extreme Light Infrastructure-Nuclear Physics) facility, currently under construction near Bucharest (Romania), is the pillar of the project ELI dedicated to the generation of high intensity gamma beams for frontier research in nuclear physics. To develop an experimental program at the frontiers of the present-day knowledge, two equipments will be deployed at ELI-NP: a high power laser system consisting of two 10 PW lasers and a high brilliance gamma beam system. The ELI-NP gamma beam will be obtained by collimating the radiation emerging from incoherent inverse Compton scattering of short laser pulses on relativistic electron beam bunches. Using this method it will be possible to obtain a gamma beam with unique characteristics in terms of brilliance, photon flux and energy bandwidth that are necessary to cover the proposed experiments in fundamental physics, nuclear physics and astrophysics, as well as applications in material and life sciences, industrial tomography and nuclear waste management. The system will consist of two energy lines: a low-energy line (LE) delivering gamma rays with energies up to 3.5 MeV and a high-energy line (HE) where the energy of the gamma rays will reach up to 19.5 MeV. Such a gamma beam requires peculiar devices and techniques to measure and monitor the beam parameters during the commissioning and the operational phase. To accomplish this task, a Gamma Beam Characterisation system equipped with four elements has been developed: a Compton spectrometer (CSPEC), to measure and monitor the photon energy spectrum; a nuclear resonant scattering spectrometer, for absolute beam energy calibration and intercalibration of the other detectors; a beam profile imager to be used for alignment and diagnostics purposes and finally a sampling calorimeter (GCAL), for a fast combined measurement of the beam average energy and intensity. The system must be able to cope with the time structure of the beam made by 32 pulses of 10^5 photons each, with a duration of 1-2 ps, separated by 16 ns and delivered at 100 Hz. The combination of the measurements performed by GCAL and CSPEC allows to fully characterize the gamma beam energy distribution and intensity with a precision of about 0.5%, enough to demonstrate the fulfillment of the required parameters. The work described in this thesis concerns the realization and the characterization of these two detectors, which are under construction at the INFN Firenze. The first detector described is the CSPEC, used to reconstruct the energy spectrum of the γ beam with a non-destructive method. The basic idea is to measure energy and position of electrons recoiling at small angles from Compton interactions of the beam, on a thin micro-metric mylar target. A high purity germanium detector (HPGe) will be used to precisely measure the energy of the Compton scattered electron, while a double sided silicon strip detector will determine the impact point of the e^- on the detector. The recoil photon is detected by Barium Fluoride (BaF2) crystals, whose fast response in coincidence with the HPGe signal will provide the trigger. The CSPEC is expected to reconstruct the γ beam energy spectrum with a precision of about the 0.1% on the reconstruction of the beam peak energy and width. The characterization procedure of the HPGe and the BaF2 detectors are presented in this thesis. The resolution on the beam energy measurement critically depends on the accuracy of the electron energy determination, which in turn is correlated to the HPGe energy resolution and to the energy loss in the materials preceding the HPGe active volume. We verified the excellent energy resolution and linearity of the HPGe by exposing the detector to different radioactive γ sources and obtaining a resolution of 0.156% at 1332 keV. In addition, the accuracy of the HPGe MC simulations, in particular of the parameters related to the dead layers preceding the HPGe crystal, has been verified using electrons of definite energy emitted by a 207^Bi source. The measured peak positions are in agreement with the simulated ones with a precision better than 1 keV confirming the correctness of the simulation geometry. The MC simulation describe well also the width of the peaks, for which we measured values that differ less than 0.4 keV from the expected ones. Concerning the CSPEC photon detector we implemented a signal shape identification method that use the ratio between the two light components of the BaF2 detector to discriminate between signals produced by γ and those due to α particles (the intrinsic radioactivity of the crystal) or to those due to thermal noise. We also characterized the crystals response in terms of linearity and energy resolution using different γ sources. In addition we verified that the BaF2 crystals intrinsic radioactivity can be used to infer changes in the energy calibration of the detector. The second detector subject of this thesis is the GCAL, a calorimeter providing a fast combined measurement of the beam average energy and intensity by absorbing the gamma pulses in a longitudinally segmented calorimeter. The intensity of the gamma beam is not exactly known, so the photon energy cannot be simply determined from the total energy released, as usually happens in calorimeters. The basic idea is to use properties of the gamma energy released inside the detector, that depends only on the photon energy and not on the beam intensity. This is obtained by exploiting the monotonic energy dependence of the total photon interaction cross section for low-Z materials in the energy range of interest at the ELI-NP facility. Thus, realizing a sampling calorimeter with low Z absorber, the average energy of the beam can be measured by fitting the longitudinal profile against parametrized distributions, obtained with detailed MC simulations. Once the photon energy is known, assuming a monochromatic beam, the number of impinging photons is obtained from the total energy released. The calorimeter for the LE beamline has been realized as a sampling calorimeter composed by 22 identical layers. Each element consists of a block of Polyethylene absorber (an inexpensive and easily workable low-Z material) followed by a readout board hosting 7 adjacent silicon detectors. The silicon detectors time response has been tested using an infrared laser. Indeed the time response is a critical issue, since the calorimeter has to be able to resolve the 16 ns separated pulses of the ELI-NP beam. This test has shown that the silicon sensors equipped with a fast custom electronics are able to disentangle pulses with the same time structure of the beam, with an accuracy at the level of per mill. The functionality of each sensor composing the calorimeter has been checked with an infrared laser. We verify the signal dependence from the γ impact point and that there are no anomalies on detector response, scanning the sensors horizontally and vertically. The last part of the activity has regarded the optimization of the calorimeter MC simulation. Starting from a simplified simulation used in the early stage of the project were there was no geometry details, new simulations were made considering a thorough description of the microstrip detectors (dead area, aluminum strip and backplane metallization) and including the presence of the aluminum supporting structures and of the acquisition board. The performances of the GCAL has been evaluated executing the energy reconstruction procedure on these new MC samples and indicate a statistical accuracy on the average beam energy and on the number of photons, better than few per mill after collecting data for a few seconds of beam operation. We have checked that the effect of the background particles is negligible given that the energy released from these particles is four order of magnitude smaller than the one released by the γ beam. The effects of some of the main sources of systematic uncertainties in the determination of the beam energy and intensity of the low-energy calorimeter have been investigated. We have studied the variations produced by having a γ beam with a characteristic energy spectrum and spatial distribution or with a random jitter on the beam energy (or intensity) rather than a monochromatic point-like beam and finally the effects related to incorrect inter-calibration of the different detector layers. The energy and intensity beam jitters do not deteriorate the GCAL performance and with a realistic beam, we obtained a reconstructed energy that is the average energy of the beam rather than the peak value in agreement with the calorimeter working principle. Due to the asymmetry on the low-energy side of the energy spectrum, the offset in the total energy deposited translates into an underestimation of the beam intensity. This effects can be accounted for by correctly simulating the beam energy distribution when producing the energy profiles. The main systematic effect turns out to be the miscalibration of the silicon pads that introduces a systematic shift on the values of the beam energy and intensity that amounts to about 0.5% for the energy and 0.7% for the intensity.