A summary on an investigation of GAGG:Ce afterglow emission in the context of future space applications within the HERMES nanosatellite mission
G. Dilillo, R. Campana, N. Zampa, F. Fuschino, G. Pauletta, I. Rashevskaya, F. Ambrosino, M. Baruzzo, D. Cauz, D. Cirrincione, M. Citossi, G. Della Casa, B. Di Ruzza, G. Galgoczi, C. Labanti, Y. Evangelista, J. Ripa, A. Vacchi, F. Tommasino, E. Verroi, F. Fiore
AA summary on an investigation of GAGG:Ce afterglowemission in the context of future space applications withinthe HERMES nanosatellite mission.
Giuseppe Dilillo a,b,c , Riccardo Campana d,e , Nicola Zampa c , Fabio Fuschino d,e ,Giovanni Pauletta a,c , Irina Rashevskaya h , Filippo Ambrosino f , Marco Baruzzo a,c ,Diego Cauz a,c , Daniela Cirrincione a,c , Marco Citossi a,c , Giovanni Della Casa a,c ,Benedetto Di Ruzza h , G´abor Galg´oczi m,n , Claudio Labanti d,e , Yuri Evangelista f,g ,Jakub Ripa m,l,o , Andrea Vacchi a,c , Francesco Tommasino i,h , Enrico Verroi h , and Fabrizio Fiore ba University of Udine, Via delle Scienze 206, I-33100 Udine, Italy b INAF-OATs Via G.B. Tiepolo, 11, I-34143 Trieste c INFN sez. Trieste, Padriciano 99, I-34127 Trieste, Italy d INAF-OAS Bologna, Via Gobetti 101, I-40129 Bologna, Italy e INFN sez. Bologna, Viale Berti-Pichat 6/2, I-40127 Bologna, Italy f INAF-IAPS,Via del Fosso del Cavaliere 100, I-00133 Rome, Italy g INFN sez. Roma 2, Via della Ricerca Scientifica 1, I-00133 Rome, Italy h TIFPA-INFN, Via Sommarive 14, I-38123 Trento, Italy i Department of Physics, University of Trento, via Sommarive, 14 38123 Trento l Department of Theoretical Physics and Astrophysics, Faculty of Science, Masaryk University,Brno, Czech Republic m E¨otv¨os Lor´and University, Egyetem t´er 1-3, 1053 Budapest, Hungary n Hungarian Academy of Sciences, Wigner Research Centre for Physics,1525 Budapest 114,Hungary o Astronomical Institute of Charles University, V Holeˇsoviˇck´ach 747/2, CZ-18000 Prague 8,Czech Republic
ABSTRACT
GAGG:Ce (Cerium-doped Gadolinium Aluminium Gallium Garnet) is a promising new scintillator crystal. Awide array of interesting features — such as high light output, fast decay times, almost non-existent intrinsicbackground and robustness — make GAGG:Ce an interesting candidate as a component of new space-basedgamma-ray detectors. As a consequence of its novelty, literature on GAGG:Ce is still lacking on points crucial toits applicability in space missions. In particular, GAGG:Ce is characterized by unusually high and long-lastingdelayed luminescence. This afterglow emission can be stimulated by the interactions between the scintillatorand the particles of the near-Earth radiation environment. By contributing to the noise, it will impact thedetector performance to some degree. In this manuscript we summarize the results of an irradiation campaignof GAGG:Ce crystals with protons, conducted in the framework of the HERMES-TP/SP (High Energy RapidModular Ensemble of Satellites - Technological and Scientific Pathfinder) mission. A GAGG:Ce sample wasirradiated with 70 MeV protons, at doses equivalent to those expected in equatorial and sun-synchronous Low-Earth orbits over orbital periods spanning 6 months to 10 years, time lapses representative of satellite lifetimes.We introduce a new model of GAGG:Ce afterglow emission able to fully capture our observations. Resultsare applied to the HERMES-TP/SP scenario, aiming at an upper-bound estimate of the detector performancedegradation due to the afterglow emission expected from the interaction between the scintillator and the near-Earth radiation environment.
Keywords:
Scintillators, GAGG:Ce, afterglow, space missions, nanosatellites, Near-Earth radiation environ-ment a r X i v : . [ a s t r o - ph . I M ] J a n . INTRODUCTION HERMES-Technologic and Scientific Pathfinder (HERMES-TP/SP) is a constellation of six 3U nanosatelliteshosting simple but innovative X-ray detectors for the monitoring of Cosmic High Energy transients such asGamma Ray Bursts and the electromagnetic counterparts of Gravitational Wave Events. The main objectiveof HERMES-TP/SP is to prove that accurate position of high energy cosmic transients can be obtained usingminiaturized hardware, with cost at least one order of magnitude smaller than that of conventional scientificspace observatories and development time as short as a few years. The main goals of the projects are: 1)join the multimessenger revolution by providing a first mini-constellation for GRB localizations with a total ofsix units, performing a first experiment of GRB triangulation with miniaturized instrumentation; 2) developminiaturized payload technology for breakthrough science; 3) demonstrate COTS applicability to challengingmissions, contribute to Space 4.0 goals; push and prepare for high reliability large constellations.The HERMES-TP project is funded by the Italian Ministry for education, university and research and theItalian Space Agency. The HERMES-SP project is funded by the European Union’s Horizon 2020 Researchand Innovation Programme under Grant Agreement No. 821896. The constellation should be tested in orbitin 2022. HERMES-TP/SP is intrinsically a modular experiment that can be naturally expanded to provide aglobal, sensitive all sky monitor for high energy transients.The foreseen detector for this experiment employ the solid-state Silicon Drift Detectors (SDD) developed byINFN and FBK in the framework of the ReDSoX Collaboration ∗ . These devices, being sensitive to both X-rayand optical photons, and characterised by a very low intrinsic electronic noise, can be exploited both as directX-ray detectors and as photo detectors for the scintillation light produced by the absorption of a gamma-ray inan inorganic scintillator crystal. This allows for the realisation of a single, compact experiment with a sensitivityband from a few keV to a few MeVs for X and gamma-rays, and with a high temporal resolution ( < µ s). Figure 1 shows a sketch of the proposed payload (which will fit in a 3U CubeSat platform).Since the nanosatellites will be launched as secondary payloads in a low-Earth orbit, they are subject to arelatively large amount of high-energy radiation fluxes (mostly cosmic-rays and geomagnetically trapped protons).This results in a degradation of the performance of the scintillator crystals (due to e.g. afterglow, activation andcreation of additional luminescence centers) and of the silicon detectors (mainly due to the Non-Ionizing EnergyLoss radiation damage, leading to an increase of the leakage current).The aim of the beam test campaign at the Protontherapy centre was to gain insight on the behaviour of boththe proposed scintillator crystal, when exposed to proton doses representative of typical values encountered onorbit during the whole operative life (at least 1–2 years).
2. BACKGROUND AND AIMS
For the HERMES detectors, the choice of the optimal scintillator material required a careful evaluation of severalfactors: • Maximisation of the light output (photons per unit of absorbed energy). • Non-hygroscopicity of the crystal. • High density and average atomic number (stopping power). • Good radiation-resistance properties. • Low light emission characteristic time.Therefore, the choice has fallen on a relatively recent material, the Cerium-doped Gadolinium-Aluminium-Gallium Garnet (Gd Al Ga O or GAGG:Ce), developed firstly in Japan around 2010, and commercially avail-able since 2014. This material has a high intrinsic light output ( ∼ ∼
90 ns, a high density (6 .
63 g/cm ), a peak light emission at ∗ http://redsox.iasfbo.inaf.it igure 1. Exploded view of the HERMES payload
520 nm and an effective mean atomic number of 54 .
4. All these characteristics make this material very suitablefor the HERMES application.Since GAGG is a relatively new material, it has not yet been extensively investigated with respect to radia-tion resistance and performance after irradiation, although the published results are very encouraging. Theseliterature results (performed with fixed proton doses up to 10–100 krad and Co gamma-ray doses of 100 krad)showed that, compared to other scintillator materials largely used in the past years in space-borne experimentsfor gamma-ray astronomy (e.g. BGO or CsI), GAGG has a very good performance, i.e. a very low activationbackground (down to 2 orders of magnitude lower than BGO), and an inferior light output degradation withaccumulated dose. However, an issue is the non-negligible amount of afterglow observed after a 100 krad Coirradiation: the crystals exhibits a phenomenon of phosphorescence on long time scale, that will lead to a slighttransitory worsening of the energy resolution and light output.The first objective of this campaign was therefore to further investigate these results by a careful evaluationof the afterglow phenomenon and the variations in optical properties of the scintillator with proton doses andat increasing dose steps, representative of the actual in-orbit radiation environment foreseen for HERMES. Inthe sections that follow we will outline the experimental procedure and summarize the results from the dataanalysis. A more complete paper on the subject is currently in preparation.
3. EXPERIMENTAL PROCEDURE AND SETUP
The characterization was carried on at the TIFPA-APSS experimental area of the Trento Proton Therapy Center,with irradiated doses chosen to be representative of typical fluxes encountered in the foreseen HERMES low-Earth orbit, at a beam energy of 70 MeV and using several irradiation steps (cf. Table 1). A teflon-wrappedGAGG:Ce scintillator crystal—dimensions 3 × × —was housed in a lightproof metal case and opticallycoupled to a pair of photomultiplier tubes (PMT) through an optical waveguide a few cm long, to ensure thePMT shielding from the proton beam. The afterglow effects were measured immediately after (from 60 s up to ∼ easurementsCrystal Irr. Duration [s] Dose [1] [p] E.O.P. [2] Current Counts Waveform2019.01.30 19:59 • J2 90 s (1 . ± . ×
1y EQ (cid:88) (cid:88) (cid:88) • J2 90 s (1 . ± . ×
2y EQ (cid:88) (cid:88) (cid:88) • J2 270 s (3 . ± . ×
5y EQ (cid:88) (cid:88) (cid:88) • J2 100 s (1 . ± . ×
10y EQ (cid:88) (cid:88) (cid:88) • J2 144 s (1 . ± . ×
2y SSO (cid:88) (cid:88) (cid:88) • J2 115 s (1 . ± . ×
10y SSO (cid:88) (cid:88) (cid:88)
Table 1. Detailed table of the irradiation runs. Each row represents an irradiation step.Runs are identified by current log start time and color coded as in article body. Temperatures in range 21 ± . ◦ C.[1] : Estimated from GEANT4 simulation of a proton beam irradiation of GAGG:Ce crystal. The 70 MeV proton beamis modelled with a Gaussian shape, having a width of σ = 6 . The crystal is placed at the beam center, and possible positioning errors have been taken into account. The totalnumber of protons simulated has been generated according to the flux measured by the beam monitor thus calculatingthe total energy deposit and resulting dose.[2] : Equivalent Orbital Period. The reported values equal the flux of proton with energies > . logger. Scintillator optical properties such as characteristic emission time, light output and energy resolutionwere then evaluated by acquiring spectra of single proton events by means of a waveform digitizer, and comparingthe results with those obtained before the irradiation. The anode signal of one of the two PMTs was measured bya Keithley 6487 picoammeter. The signals from the last dynode of both PMTs were brought to a discriminatorunit where they were split and redirected to a counter and a multi-channel digitizer for waveform acquisition.
4. RESULTS4.1 Afterglow model and irradiation data analysis
Long-lived afterglow emission in scintillators is due to presence of impurity sites or defects within the crystallattice. Arriving at the impurity sites, charge carriers (electrons, holes) can create excited configurations whosetransition to ground state is forbidden. At later times, charge carriers can escape these sites by differentprocesses (e.g. to the conduction band by thermal energy absorption or to nearby recombination centers by director thermally assisted tunneling ). Ultimately all of the charge carriers recombine, giving rise to luminescence.Different scintillators displays different afterglow characteristics, varying from almost non-existent to a verylong and intense emission. GAGG:Ce has been known to belong to the end of the spectrum showing intenseemission lasting up to several days. Afterglow mitigation techniques, such as Mg-codoping, proved successfulat cost of diminished light-yield. We propose a semi-empirical model in which i. afterglow emission results from the release of charge carrierstrapped in metastable states with few different, discrete lifetimes; ii. the density of a particular trap species canchange linearly with the irradiation dose. Making use of such a model one can determine the number of chargecarriers in a particular trap species during each irradiation step. Once the irradiation ceases, the number oftrapped charge carriers decreases exponentially with time.The PMT anode current we observe will be proportional to the time derivative of the total number of trappedcharges, I ( t ) ∝ dNdt , the proportionality constant depending from a number of quantities measurable throughcalibration, namely the PMT gain and quantum efficiency, and the photon transport efficiency from the crystalto the photocathode. Through a fitting procedure, we derive the model parameters i.e., τ i , trap emission timeconstants, n i , the density of charge carriers trapped for particle of incident radiation and ∆ n i , the the totalvariation in trap species density at the end of the last irradiation. In this procedure we start by considering onlythe data from the first few irradiation and measurement runs, then progressively add subsequent measurements.At each step we use the partial model to check the accuracy of the predictions for the following irradiations. Themodel is then progressively adjusted to successfully fit the whole dataset. i [ s ] σ τ [ s ] n i [ cm − ] σ n [ cm − ] ∆ n [ cm − ] σ ∆ n [ cm − ]25 .
32 0 .
41 196 . . . . .
97 1 .
00 129 . . − . . . . . . − . . . . . . − . . . . . . . . − . . . · . . . Table 2. Traps properties as estimated from fit to the model of Section 4.1.Afterglow data from a GAGG:Ce sample at temperature 21 ± . ◦ C. We find that the smallest number of trap species needed to resolve patterns in fit residuals of the lastmeasurement run is 7. Models with more trap species results in worse estimates of best fit parameters whileproviding no enhancements in the residuals.Fit results are plotted in Figure 2. Corresponding best parameters estimates are reported in Table 2. For twoof the trap species we find the probability of capture to grow with the accumulated dose, starting from initiallynegligible values. At variance, one of the trap species ceases to capture carriers during the last irradiation. Forthe latter, an emission component is still found in the last measurement. We believe to be reasonable to supposethat such an effect could stem from saturation of the trap species: at some point during irradiation, the traps ofthis species are all occupied and can no longer capture new charges. Beside trap saturation, other processes thatcould explain such phenomena are variations in the traps average lifetime, formations of new traps or reductionin the number of existing traps, dose dependent formation of non-radiative centers and other dose-dependentnon-linear effects. From the present data we cannot take the investigation of these phenomena further or evenexclude the possibility that such effects could arise as a consequence of the experimental procedure.
Figure 2. The dataset fitted to the model outlined in Section 4.1. Measurements, best fit curves and residuals are shown.Parameters estimates as in Table 2. Temperatures are in range 21 ± . ◦ C .2 Impact of GAGG:Ce afterglow on silicon drift detectors The first HERMES spacecrafts are expected to fly on a LEO, near-equatorial orbit. These units are expected tobe operative for a minimum of 2 years. We expect afterglow emission to be stimulated in orbit by interactionbetween the scintillator and particles trapped in the Van Allen radiation belt. Having no intrinsic gain, SDDcells coupled to a scintillator will not be able to resolve the dim afterglow photons into full-blown signals. Still,the afterglow will contribute to noise in a way similar to the detector leakage current causing a degradation ofthe detector performance.Through the model outlined in Section 4.1 we are now able to predict the impact of GAGG:Ce afterglow emissionin terms of equivalent leakage current. To estimate the trapped particle fluxes along the orbit we use the IRENE(International Radiation Environment Near Earth) AE9/AP9 models. These models allows to compute protonand electron orbital fluxes in space. The previous AE8/AP8 versions were developed by NASA, are regarded asthe industry standard for radiation belt modeling, and are available in the MIN and MAX variants accountingfor minimum or maximum of solar activity. The AE9/AP9 models —which are built upon much more recenttrapped radiation observations— are expected to replace their predecessors in the near future.Assuming the scintillator to be irradiated at constant average rate for periods of times equal to the simulationtimesteps (10 s) and considering the flux values calculated with AE9/AP9 along 30 days of orbit, we can estimatethe number of trapped charges N ( t ) at some orbital time using the model parameterization of Section 4.1.We are interested in an upper-bound estimate of the noise to be expected from afterglow. Hence we analyzethe following worst-case scenario: for each trap species we consider the maximum capture capacity between thevalues expected at mission start and mission end. The expected value of leakage current is I L ( t ) = − ef (cid:15) dNdt ,where N is the number of trapped electrons at a given time, (cid:15) indicates the quantum efficiency of the SDD, f isthe crystal to SDD photon transport efficiency and e is the elementary charge.In Figure 3 we report our worst-case estimate of the leakage current resulting from afterglow emission ofa GAGG:Ce sample with HERMES detector dimensions, as expected over 100 orbits at altitude 550 km andinclination 10 ◦ . The reported minimum, mean and maximum values of leakage current are calculated startingat 24 hours of orbital lifetime. We remark that the data on which the model from Section 4.1 was built weregathered starting from a minute after the end of the irradiations. It follows that the model is expected to failinside and up to one minute following a passage over trapped radiation regions. For our purposes this is not aproblem since in these regions the HERMES instruments will be turned off. For this reason, the current valuesexpected during such transits are omitted in Figure 3.HERMES low-noise front-end electronics (FEE) will be able to grant nominal performance up to ∼
100 pAof leakage current, a value which is well above two order of magnitude from the estimated maximum. Despitethis fact, increases in leakage current will still result in a worsening of the detector energy resolution.Since the first six HERMES Technological Pathfinder units will be launched in near-equatorial orbits, weexpect the afterglow impact on payload performance to be small. However, when the HERMES fleet will beenlarged to host spacecrafts in orbits at higher inclinations, the impact of afterglow on detector performance willneed further, more accurate investigations.In fact, the flux of trapped radiation varies greatly with the inclination of the orbit. As a consequence, weexpect the intensity of the afterglow emission to change accordingly. Repeating our worst-case evaluation foran orbit with altitude 550 km and 50 degrees inclination we find the leakage current from afterglow emission toexceed the FEE’s nominal performance limit of 100 pA. However one must note that producing these results ourassumptions about the afterglow and trapped particles models, shielding, kinetic-ionization conversion insidethe scintillator and SDD quantum efficiency have all been very conservative. Shielding in particular is expectedto play a substantial effect in moderating the afterglow component due to interactions between scintillator andtrapped belt radiation. An accurate estimate of the degradation in detector performances due to GAGG:Ceafterglow emission for spacecrafts in orbits at higher inclination cannot be separated from a careful descriptionof the payload and spacecraft itself, something which goes beyond the scope of this work. igure 3. Estimated worst-case leakage current of a SDD cell with dimensions 6 . × .
05 mm resulting from GAGG:Ceafterglow emission induced by irradiation of a 12 . × . × .
50 mm scintillator at temperature of 21 ± . ◦ C. Thescintillator is completely shielded from radiation on one of the smaller face. Orbital populations of protons and electronsare modelled through AE9/AP9 packages for an orbit with 550 km altitude and 10 ◦ inclination orbit over ∼ ACKNOWLEDGMENTS
This project has received funding from the European Union Horizon 2020 Research and Innovation FrameworkProgramme under grant agreement HERMES-Scientific Pathfinder n. 821896 and from ASI-INAF AccordoAttuativo HERMES Technologic Pathfinder n. n. 2018-10-HH.0. We also thank the TIFPA staff for theirprecious assistance and the ReDSoX collaboration.
REFERENCES [1] Fiore, F. et al., “The HERMES-Technologic and Scientific Pathfinder,”
Society of Photo-Optical Instrumen-tation Engineers (SPIE) Conference Series , 11444–166, these proceedings (2020).[2] Campana, R. et al., “A compact and modular X and gamma-ray detector with a CsI scintillator and double-readout Silicon Drift Detectors,”
Proc. SPIE Int. Soc. Opt. Eng. , 99056I (2016).[3] Fuschino, F., Campana, R., Labanti, C., Evangelista, Y., Feroci, M., Burderi, L., Fiore, F., Ambrosino, F.,Baldazzi, G., Bellutti, P., Bertacin, R., Bertuccio, G., Borghi, G., Cirrincione, D., Cauz, D., Ficorella, F.,Fiorini, M., Gandola, M., Grassi, M., and Zorzi, N., “HERMES: An ultra-wide band X and gamma-raytransient monitor on board a nano-satellite constellation,”
Nuclear Instruments and Methods in PhysicsResearch Section A: Accelerators, Spectrometers, Detectors and Associated Equipment (11 2018).[4] Yoneyama, M., Kataoka, J., Arimoto, M., Masuda, T., Yoshino, M., Kamada, K., Yoshikawa, A., Sato, H.,and Usuki, Y., “Evaluation of GAGG:ce scintillators for future space applications,”
Journal of Instrumen-tation , P02023–P02023 (feb 2018).5] Dilillo, G. et al., “Investigating GAGG:Ce afterglow and tolerance to space radiation through proton irra-diation,” in preparation (2020).[6] Tommasino, F., Rovituso, M., Fabiano, S., Piffer, S., Manea, C., Lorentini, S., Lanzone, S., Wang, Z., Pasini,M., Burger, W. J., La Tessa, C., Scifoni, E., Schwarz, M., and Durante, M., “Proton beam characterizationin the experimental room of the Trento Proton Therapy facility,” Nuclear Instruments and Methods inPhysics Research A , 15–20 (Oct 2017).[7] Knoll, G. F., [
Radiation Detection and Measurement ], John Wiley and Sons, 4 ed. (2010).[8] Huntley, D., “An explanation of the power-law decay of luminescence,”
Journal of Physics CondensedMatter , 1359–1365 (Feb 2006).[9] Lucchini, M., Babin, V., Bohacek, P., Gundacker, S., Kamada, K., Nikl, M., Petrosyan, A., Yoshikawa,A., and Auffray, E., “Effect of Mg2+ ions co-doping on timing performance and radiation tolerance ofCerium doped Gd3Al2Ga3O12 crystals,” Nuclear Instruments and Methods in Physics Research Section A:Accelerators, Spectrometers, Detectors and Associated Equipment , 176 – 183 (2016).[10] Ginet, G., O’Brien, T., Huston, S., Johnston, W., Guild, T., Friedel, R., Lindstrom, C., Roth, C., Whelan,P., Quinn, R., Madden, D., Morley, S., and Su, Y.-J., “AE9, AP9 and SPM: New Models for Specifying theTrapped Energetic Particle and Space Plasma Environment,”
Space Science Reviews (11 2013).[11] “Ecss-e-st-10-04c - space environment.” https://ecss.nl/standard/ecss-e-st-10-04c-space-environment/ . Accessed: 2020-09-25.[12] Grassi, M., Gandola, M., Mele, F., Bertuccio, G., Malcovati, P., Fuschino, F., Campana, R., Labanti, C.,Fiorini, M., Evangelista, Y., Piazzolla, R., Feroci, M., Zampa, G., Zampa, N., Rachevski, A., Bellutti, P.,Borghi, G., Demenev, E., Ficorella, F., Picciotto, A., Zorzi, N., Rashevskaya, I., Vacchi, A., Fiore, F.,and Burderi, L., “X-/gamma-Ray Detection Instrument for the HERMES Nano-Satellites Based on SDDsRead-Out by the LYRA Mixed-Signal ASIC Chipset,” in [ ], 1–6 (2020).[13] Ripa, J. et al., “A comparison of trapped particle models in Low Earth Orbit,”
Society of Photo-OpticalInstrumentation Engineers (SPIE) Conference Series11444