Instruments of RT-2 Experiment onboard CORONAS-PHOTON and their test and evaluation IV: Background Simulations using GEANT-4 Toolkit
Ritabrata Sarkar, Samir Mandal, Dipak Debnath, Tilak B. Kotoch, Anuj Nandi, A. R. Rao, Sandip K. Chakrabarti
aa r X i v : . [ a s t r o - ph . I M ] D ec Noname manuscript No. (will be inserted by the editor)
Instruments of RT-2 Experiment onboard CORONAS-PHOTON and their test and evaluation IV: BackgroundSimulations using GEANT-4 Toolkit
Ritabrata Sarkar · Samir Mandal · DipakDebnath · Tilak B. Kotoch · Anuj Nandi · A. R. Rao · Sandip K. Chakrabarti
Received: date / Accepted: date
Abstract
Hard X-ray detectors in space are prone to background signals due to theubiquitous cosmic rays and cosmic diffuse background radiation that continuously bom-bards the satellites which carry the detectors. In general, the background intensitydepends on the space environment as well as the material surrounding the detectors.Understanding the behavior of the background noise in the detector is very importantto extract the precise source information from the detector data. In this paper, wecarry out Monte Carlo simulations using the GEANT-4 toolkit to estimate the promptbackground noise measured with the detectors of the RT-2 Experiment onboard theCORONAS-PHOTON satellite.
Keywords
Radiation detectors · X- and gamma-ray telescopes and instrumentation · Background radiation, cosmic · Structural and shielding materials · Monte Carlosimulations
PACS · · · · This work was made possible in part from a grant from Indian Space Research Organiza-tion (ISRO). The whole-hearted support from G. Madhavan Nair, Ex-Chairman, ISRO, whoinitiated the RT-2 project, is gratefully acknowledged.Ritabrata Sarkar, Samir Mandal, Dipak Debnath, Tilak B. Kotoch, Anuj Nandi + Indian Centre for Space Physics, 43 Chalantika, Garia Station Rd., Kolkata 700084E-mail: [email protected]; [email protected]; [email protected]; [email protected];[email protected](+: Posted at ICSP by Space Science Division, ISRO Head Quarters)A. R. RaoTata Institute of Fundamental Research, Homi Bhabha Road, Colaba, 400005E-mail: [email protected] K. ChakrabartiS.N. Bose National Centre for Basic Sciences, JD Block, Salt Lake, Kolkata 700097(Also at Indian Centre for Space Physics, 43 Chalantika, Garia Station Rd., Kolkata 700084)E-mail: [email protected]
Observational astronomy in the X-ray and γ -ray bands of the electromagnetic spec-trum is very crucial to explore high-energy physical phenomena in the Universe. X-rayor γ -ray observations from the ground-based instruments are not possible due to theatmospheric attenuation. In the last four decades, huge efforts have been made to-wards the development of space-borne X-ray and γ -ray telescopes. At the same time,these efforts are also limited by the hostile space environment, particularly in hardX-rays and γ -rays. High energy charged particles coming from outer space and fromthe Solar wind become trapped in the Earth’s surrounding magnetic field creating ra-diation belts around the Earth, known as the Van Allen radiation belt. Though thesatellites carrying the X-ray and γ -ray detectors are usually placed below the innerradiation belt (altitude varying from a few 100 km to 10 , km ), still there are somelocalized energetic charged particle regions which may result in severe damages to theinstruments if they are activated while passing through those regions such as the Po-lar Regions and the South Atlantic Anomaly (SAA) region. Apart from these trappedcharged particle regions, there are the cosmic diffuse radiation and the cosmic rays,mainly protons and alpha particles. These cosmic-ray particles, depending on the ge-omagnetic strength at the altitude and position, enter into the Earth’s atmosphereand interact with the atoms and molecules resulting in various secondary particles.These primary and secondary cosmic radiations and charged particles bombard thedetector and satellite materials and may produce secondary or higher order particlesin prompt interactions (through bremsstrahlung, pair creation etc.) or/and may ini-tiate the detector activation. These radiations and particles will increase the detectornoise which may mask the original source signal. The high-energy charge particles canpass through detector and space craft and may deposit a line of charge in the detectorvolume. This may then be detected in the same way as energy deposit produced by the‘real’ X-ray radiations. It is thus very essential to make an accurate estimate of the on-board background noise before designing any space experiment. The background noisein an X-ray instrument is mainly due to the interactions of the cosmic-ray protons,the albedo protons and neutrons due to Earth’s atmosphere, Cosmic Diffused Gamma-Ray Background (CDGRB), secondary gamma rays formed in the detector material,its frame structure and with the satellite (Dean, Lei & Knight 1991). In addition, thelong-term activation of the detector materials by these radiations or particles or by theparticles in the trapped particle region are also responsible for the background noise.The background noise varies over a wide range of energy and actually depends on thedetection capability of the specific detector. These background noise compete with thesignals due to the interactions of source photons with the detectors. It is thereforeimportant to understand the interactions of these background components with thedetector material and to remove them while extracting the source signal.The RT-2 experiment aboard the CORONAS-PHOTON satellite (Kotov et al. 2008,Nandi et al. 2009) consists of 4 payloads: three X-ray detectors (RT-2/S, RT-2/G &RT-2/CZT) and one processing electronic device (RT-2/E). Detailed description of allthe payloads and their functionality are given in Debnath et al. (2010), Kotoch etal. (2010), Nandi et al. (2010) and Sreekumar et al. (2010). The Phoswich detectors(RT-2/S & RT-2/G) are made of NaI (Tl) and CsI (Na) scintillating crystals. Both thePhoswich detectors are sensitive to detect high energy X-rays in the energy range of15 keV to ∼ keV . The RT-2/CZT detector is a solid-state imaging device, which consists of CZT and CMOS detectors. Both the detectors are sensitive in the energyrange of 20 keV to 150 keV .In the present work, we concentrate on predicting/comparing the background inthese detectors during the passage of the space craft through the low-background equa-torial region (away from the SAA and polar regions), to estimate the sensitivity of thedetector. Hence, we make a detailed simulation of the interaction of primary and sec-ondary protons, cosmic diffused gamma rays, secondary gamma rays and secondaryneutrons with the detector volume as well as the whole structure of the satellite car-rying the detectors.In this paper, we carry out Monte Carlo (MC) simulations using the GEANT-4toolkit and highlight the effects of shielding material in calculating the backgroundnoise due to cosmic-ray photons on the detectors of the RT-2 Experiment. In thenext section ( § § §
4, we present the simulation results of the RT-2/S (RT-2/G) andRT-2/CZT payloads. In § § Table 1: Detector specifications of RT-2/S (RT-2/G) and RT-2/CZT payloads.
Payload RT-2/S (RT-2/G) RT-2/CZT
Detector type NaI + CsI CZT & CMOSMaterial composition NaI (Tl activated) Cd . Zn . TeCsI (Na activated) CMOS (Amorphous Si photo-diodeGd O S:Tb)Thickness (mm) 3 + 25 5 & 3Size (mm) 116 dia 40 ×
40 & 24.5 × ) 105.6 48 (3 modules) & 5.7FOV 4 ◦ × ◦ (6 ◦ × ◦ ) 6 ′ − ◦ Readout PMT pixels
For the simulation of the effects of various background components in the detec-tors we consider an approximate mass distribution of the whole CORONAS-PHOTONsatellite hosting various detectors on it. Figure 1 depicts the simplified shape of thewhole satellite including the detectors on it. This shape was used in the simulation.The major contribution of its mass is from the satellite shell structure which is madeup of Aluminum (Al) and the electronics modules inside it, which consist of Aluminum (Al), Silicon (Si) and Copper (Cu) as the major elements. We distribute these materi-als throughout the satellite cavity for simplification. Also for the detector componentsother than RT-2/S, RT-2/G and RT-2/CZT we consider simplified structures consist-ing of the approximate weights of the major components of concerned detectors. Theoverall height of the satellite construction under our consideration is ∼ cm and ra-dius is ∼ cm . For the detectors of our concern i.e., RT-2/S, RT-2/G and RT-2/CZT,we use more detailed geometry described in the following sections. Fig. 1
A simplified 3D view of the approximate mass distribution of the satellite containingthe detectors used for the simulation. γ -ray de-tectors. They are: the CDGRB photons, secondary gamma-ray photons due to Earth’satmosphere, primary Cosmic-Ray (CR) protons, secondary protons due to interactionof CR in Earth’s atmosphere and the albedo neutrons from the Earth’s atmosphere.To simulate the CDGRB photons, we generate the incident photons from the surfaceof a hemisphere (at +ve z-axis) of radius 120 . cm . The center of the whole geometryis at the center of the satellite common mounting plate on which the RT-2 detectorsand some other detectors are mounted (see Figure 1). The position of the incidentphoton generating hemisphere is such that it covers the whole region above the Earth’shorizon. The randomness of the incident photons are ensured by placing their originrandomly on the hemisphere and the directions of the photon momenta are also chosento be random within the solid angle subtended by the dimension of the satellite radiusat the vertex of each incident photon. We are interested in the response of the detector in the incident energy range of 10 keV − MeV . The CDGRB spectrum can berepresented by the equation (Gruber et al. 1999), dNdE = . E − . exp − E/ . if E ≤ . keV . × − (cid:0) E (cid:1) − . + 8 . × − (cid:0) E (cid:1) − . +4 . × − (cid:0) E (cid:1) − . if E ≥ . keV , (1)where, E is incident photon energy and dNdE is in the unit of counts/cm /s/sr/keV .We divide the entire energy range into 500 bins equal in logarithmic scale and simulate100 ,
000 photons in each bin to retain a good statistics in the simulation result. Theincident photon spectrum on the detectors due to CDGRB is shown in Figure 2a.For the simulation of the secondary albedo gamma-ray photons, we consider pho-tons randomly originated from a hemispherical surface of radius 300 cm with the centercoinciding with the center of the satellite common mounting plate. In this case, weconsider the position of the generating hemisphere at the opposite that of the primaryCDGRB (i.e., at -ve z-axis). The direction of the secondary photons has been achievedin the same way as that of the primary CDGRB. In this case also we consider theenergy range of 10 keV − MeV . Atmospheric gamma-ray line emission, such asthe 511 keV emission from positron annihilation, has not been considered, while sim-ulating for secondary albedo gamma-ray photons. The energy spectrum is representedby (Ajello et al. 2008; Mizuno et al. 2004), dNdE = . × − ( E . ) − . + ( E . ) . if E ≤ . keV . × − (cid:0) EMeV (cid:1) − . if 200 . keV ≤ E ≤ . MeV . × − (cid:0) EMeV (cid:1) − . if E ≥ . MeV (2)in units of counts/cm /s/sr/keV . We divide the entire energy range into 500 binsequal in logarithmic scale and in each bin, we inject 100 ,
000 photons. The incidentphoton spectrum on the detectors due to secondary gamma-ray photons is shown inFigure 2b.We also simulate the detector response due to the CR and secondary protons. Weconsider the input differential spectra for the downward and upward going protons.While the upward going protons are mostly secondaries from the Earth’s atmosphere,the downward going component also contains the primary CR proton above the cutoff.We consider these spectra near the equatorial region.For the downward going proton component we consider random protons from aspherical section of radius 300 cm around the satellite with its center coinciding withthe center of the common mounting plate and covering the open region above theEarth’s horizon. The randomization in the direction of the particles is the same asdescribed for the CDGRB or secondary photons. We carry out the simulation for theproton energy range of 100 MeV − GeV . To produce the energy distribution weconsider the spectral data given by Alcaraz et al. (2000) for the low geomagneticlatitude (0 < Θ M < . dNdE = 1 . × − (cid:0) EGeV (cid:1) − a exp − (cid:0) EE cut (cid:1) − a +1 +16 . × − (cid:16) E + ZeφGeV (cid:17) b × ( E + Mc ) − ( Mc ) ( E + Mc + Zeφ ) − ( Mc ) × (cid:0) EE cut (cid:1) − . (3) in units of counts/cm /s/sr/keV , where e is the magnitude of the electron charge, Z isthe atomic number of the particle, a = 0 . E cut = 5 . × keV , Solar modulation φ = 6 . × kV (a value near Solar activity minimum), Mc is the proton mass and E cut = 12 . × keV . In this case, we divide the whole incident energy range in100 bins equal in logarithmic scale and inject 100 ,
000 photons in each bin. Figure 2cpresents the incident energy spectrum for the Primary CR protons.The contribution of the upward going proton spectra from the Earth’s atmosphereis considered as follows. For the positional and directional aspects of the generation ofthese protons we used the same methods as the albedo photons. Here in this simulationwe consider the energy range of 100
MeV − GeV . The spectral form is given as(Alcaraz et al. 2000; Mizuno et al. 2004), dNdE = 1 . × − (cid:16) EGeV (cid:17) − a exp − (cid:16) EE cut (cid:17) − a +1 (4)in units of counts/cm /s/sr/keV , where, a = 0 .
155 and E cut = 5 . × keV . Theincident secondary proton spectrum is depicted in Figure 2d.For the simulation of the secondary neutrons due to the interaction of the CR in theEarth’s atmosphere, we consider the generation of the neutrons from a hemisphericalsurface of radius 300 cm in the same manner as the secondary photon simulation. Theincident particle direction is also randomized in the same way. The energy range ofthe neutron simulation is 10 keV − GeV . The spectral form of the neutron energydistribution is given by (Armstrong et al. 1973), dNdE = . × − (cid:0) EGeV (cid:1) − . if 10 keV ≤ E ≤ MeV . × − (cid:0) EGeV (cid:1) − . if 1 MeV ≤ E ≤ MeV . × − (cid:0) EGeV (cid:1) − . if 100 MeV ≤ E ≤ GeV (5)in units of counts/cm /s/sr/keV . We divide the whole energy range in 500 energybins equal in log scale and inject 100 ,
000 photons in each bin. Figure 2e shows theincident spectrum for the secondary neutron generation.The spectrum of energy deposition on the detectors is calculated by normalizingthe deposited spectrum in the following way. We have N incident energy bins equal inlogarithmic scale, where E i is the width of the i th bin. In each bin we are simulating I s (here 100 , A I and over the solid angle Ω I subtended by the detector to the vertex of the incident photons. Now the total numberof photons in each incident energy bin is I i = R E i R A I R Ω I dNdE dE . Then we calculate the normalization constant C i = I i I s . To calculate the normalized deposition spectrum wedivide the whole deposition energy range in M bins. We have D ij number of photons inthe j th bin of the deposition spectrum due to I s photons in the i th incident bin. Thenwe calculate the normalized photon counts in the j th bin of the deposition spectrumfor the all incident photons as D j = N X i =1 D ij × C i A D × E j (6)where, A D is the area of the crystal (detector) and E j is the width of the j th energyrange.In the present work, we are only dealing with the sources for the prompt backgroundnoise. So we simulate here some of the main sources of the prompt background noise Energy (keV)10 / s / s r / ke V c oun t s / c m −9 −8 −7 −6 −5 −4 −3 −2 −1 Incident Energy Spectrum Due to CDGRB a Energy (keV)10 / s / s r / ke V c oun t s / c m −7 −6 −5 −4 −3 −2 Incident Energy Spectrum Due to Secondary Gamma−Rays b Energy (keV) / s / s r / ke V c oun t s / c m −11 −10 −9 −8 Incident Energy Spectrum Due to Downward Protons c Energy (keV) / s / s r / ke V c oun t s / c m −11 −10 −9 −8 Incident energy Spectrum Due to Upward Protons d Energy (keV)10 / s / s r / ke V c oun t s / c m −9 −8 −7 −6 −5 −4 Incident Energy Spectrum Due to Secondary Neutrons e Fig. 2
The incident energy spectrum due to the (a) CDGRB photons, (b) secondary gamma-ray photons, (c) downward going protons, (d) upward going protons and (e) albedo neutrons. sources. Apart from the prompt background, significant noise will be present due tothe detector material activation as the satellite is in a polar orbit. However, there areconsiderable uncertainties in estimating the contribution due to long term activationto the total background. This is because the relevant package, namely,
Cosima basedon Geant4 is still in the developmental stage and cannot be trusted for predictingbackgrounds due to activation (Zoglauer et al. 2008; Zoglauer, 2009). In the presentpaper, we have deferred the inclusion of the effects of activation on the detector. Thiscan be dealt with in a future work.
To simulate the incident spectrum of a Gamma-Ray Burst (GRB), we consider theBand spectral (Band et al. 1993) form of a bright GRB (GRB 880725), which followsthe equation, dNdE = A (cid:0) E keV (cid:1) α exp (cid:0) − EE (cid:1) if E ≤ ( α − β ) E A (cid:16) ( α − β ) E keV (cid:17) ( α − β ) exp ( β − α ) (cid:0) E keV (cid:1) β if E ≥ ( α − β ) E , (7)in units of counts/cm /s/keV , where, A = 7 .
056 is a constant, α = − .
32 is the lowenergy slope, β = − . E = 67 . keV is the break energy(Strohmayer et al. 1998). The incident photons are generated from a plane of areaequal to that of the collimator of the detectors and is placed 60 cm above the detectorbase plate. We consider parallel photons falling straight into the detector through thecollimator. We also consider the source location (GRB) at an angle of ∼ ◦ to theon-axis of the payload and its effect on resulting photon distribution after interactingwith detector materials. The incident energy spectrum (Eqn. 7) of the GRB is shownin Figure 3. Energy (keV)1 10 / s / ke V c oun t s / c m −4 −3 −2 −1 Incident GRB Energy Spectrum
Fig. 3
The incident photon spectrum due to a typical Gamma-Ray Burst (GRB). km in an orbitwhich is inclined at 82 . ◦ , which results in the passage of the satellite through highenergetic charge particle regions of the SAA, North Cap (NC) and South Cap (SC) for ∼
40% of its orbital time. In these regions, satellite operations will be restricted for theprotection of the detectors from the high radiation dose. Apart from the high energeticcharge particles of SAA, NC and SC regions, the CDGRB, the high energy protonsand other albedo particles and photons will hit the satellite and detector materialsto produce the background noise in the detectors either directly or by creating localspallation background. The hard X-ray solar flares along with some other astrophysicalsources (e.g,. GRBs) would be detected by the RT-2 detectors in the energy range of15 − keV , extendable up to ∼ keV . The underlying physical processes by which photons can interact in any medium depend on particle energy and the materialproperties of that medium.For the simulation of the detectors under this circumstances, we use the Geant4simulation toolkit version 9.1.p03 and the cross section data version G4EMLOW5.1,G4ABLA3.0, G4NDL3.12. For the electromagnetic processes in the simulation we con-sider the low energy electromagnetic physics list. For the photon interactions we areusing the low energy photo-electric effect (activated with Auger electron production),low energy Compton effect, low energy Rayleigh Scattering and low energy Gamma rayconversion. In the above mentioned energy range, the incident photon produces elec-trons as the secondary particles. Electrons lose their energy through the low energy ion-ization, multiple scatterings and low energy bremsstrahlung processes. For positrons weconsider the bremsstrahlung, annihilation, ionization and multiple scattering processes.We are using a production cut-off value of 1 µm for photon, electron and positron whichis the same for all the materials. This implies a threshold energy in the Aluminum as990 eV , 1 . keV and 1 . keV respectively for photons, electrons and positrons. For thehadronic interactions we are considering the “LHEP” physics list defined in the Geant4.In the Geant4 toolkit, above mentioned processes are defined as separate modules inthe “process” category and this module is included in the “Physics list”(Agostinelli etal. 2003). Further information are available at http://geant4.web.cern.ch/geant4/.2.4 Shielding materialTo shield the detectors from X-rays and gamma-rays from off-axis sources and diffuse X-rays and other cosmic particles, materials with high atomic numbers such as Lead andGold etc. are used. These high-Z elements, however, have a deep dip in their absorptioncoefficient just below their K-shell binding energies. Sometimes, materials with loweratomic number like Tin and Copper are used to absorb the characteristic fluorescentX-rays. In the present case, however, the primary objective of the experiment is tomake very sensitive hard X-ray measurement of solar flares (typically below 50 keV )and in particular use the Phoswich technique for background reduction. Since theCsI detector in the Phoswich configuration has a good sensitivity above 100 keV , analternate objective is to use the detector as an open monitor above 100 keV for gammaray bursts and other bright sources. For these dual objectives, we find that Tantalumhas a very good absorption property with atomic number 73, a K-shell binding energyof 67 . keV and high density of 16 . gcm − . For the shielding of the detector we havechosen the Tantalum shield in a way which minimizes mass as well as maximizes thedual objectives of solar X-ray flare spectroscopy and off-axis source observations. The background simulation of the RT-2/S (RT-2/G) detector is carried out with thevirtual detector constructed within the Geant4 toolkit environment with the construc-tion parameters given in Table 2. These parameter values are the same as in the onboarddetector geometry, but to avoid the complexity in the detector geometry constructionin the simulation we simplified some of the detector parts which are not expected tochange the simulation results significantly. These simplifications consist of omitting theribs to hold the collimator and the detector walls, simplification of the construction of the Photo-Multiplier Tube (PMT), etc. A 3D view of the RT-2/S detector whichis considered in the simulation is given in Figure 4. This figure is the direct outputfrom the simulation and agrees closely with the CAD design of the actual detectorconstruction. We simulate the detector response for the background noise due to fivemajor background components and a typical GRB spectrum as has been describedearlier. In the following subsection, we will discuss the response of the detector forthese particular background radiation under our consideration and in the subsequentsubsection we will give the results for the GRB spectrum. Fig. 4
A schematic 3D view of RT-2/S (RT-2/G) payload with different components (e.g.,crystals, collimator, PMT etc.). ∼ keV . Low energy photons areabsorbed by the 0 . mm Al filter used for protection of the crystals. The CsI spectrum,also shows a few counts in the low energy range less than ∼ keV . These counts aremainly due to the partial energy depositions of the higher energy photons. In higherenergy range, the counts are high due to the photons coming through the collimator Detector Parts Dimensions
Bottom plate 25 . × . × . cm Detector Housing 20 . × . × . cm (inner dimension)wall thickness 0 . cm Top Plate 23 . × . × . cm hole radius 6 . cm Collimator 13 . × . × . cm (inner dimension)wall thickness 0 . cm FOV 0 . cm Tantalum plate thickness0 . cm gap between two Ta plates (4 ◦ × ◦ FOV for S)0 . cm gap between two Ta plates (6 ◦ × ◦ FOV for G)NaI Crystal radius 5 . cm , thickness 0 . cm CsI Crystal radius 5 . cm , thickness 2 . cm (gap between two crystals 0 . cm )Optical Coupling (Silicon Oxide) radius 3 . cm thickness 1 . cm PMT Upper part : radius 4 . cm , height 3 . cm Middle part (conic section) : height 1 . cm Lower part : radius 2 . cm , height 12 . cm thickness of the whole PMT 0 . cm (Aluminum)Shielding 0 . cm thick Tantalum strip, height 1 . cm ,around the collimator wall below the FOV to shieldNaI+CsI crystal. Shielding weight ∼ g and not interacting on the NaI crystal and as well as due to those photons entering intothe detector other than the collimator part. The peak in the NaI and CsI spectrumaround 60 keV is due to the Ta K α fluorescent emission. The detector response to thealbedo Gamma-ray photons due to the Earth’s atmosphere is given in Figure 5b. Herein the figure we can see the Ta K α and K β fluorescent emission in the NaI spectrumalong with the photon annihilation peak which is also visible in the CsI spectrum.Figure 5c presents the energy deposition spectrum in the NaI and CsI crystal due tothe downward going protons at low Earth orbit position. We have shown the detectorresponse spectrum in Figure 5d, for the upward going protons due to the interactionof the cosmic rays in the Earth’s atmosphere. Figure 5e depicts the energy responsespectrum in the NaI and CsI crystals for the secondary neutron background spectrum.From Figure 5(a,b,c,d,e) we can conclude that the most important contributor to theprompt background noise are the CDGRB photons, while the secondary gamma-rayphotons also promote a significant portion of the noise.The results from the simulation of RT-2/G are almost identical (except the nor-malization in the low energy range) as that of the RT-2/S, since both of the payloadshave identical configuration except the FOV of the collimator and 2 mm Al shieldingon the top of the collimator of RT-2/G to have higher energy cut-off (below 25 keV ).So, the above results for the RT-2/S are also applicable to RT-2/G payload regardingthe response of the detector.3.2 Response of a GRB spectrum in RT-2/SThe RT-2 detectors are primarily designed for spectroscopic measurements of solarflares in the 10 − keV region. For efficient background rejection, a thick CsI detectorwas used. While designing the experiment, it was realized that above 100 keV , the RT-2detectors will act as omni-directional gamma-ray burst detectors. Here we also simulatea GRB source and give the results below. Energy (keV) / s / ke V c oun t s / c m −5 −4 −3 −2 −1 NaICsI
Energy Spectrum Due to CDGRB a Energy (keV) / s / ke V c oun t s / c m −4 −3 −2 −1 NaICsIEnergy Spectrum Due to Secondary Gamma−Rays b Energy (keV) / s / ke V c oun t s / c m −6 −5 −4 −3 NaICsI
Energy Spectrum Due to Downward Protons c Energy (keV) / s / ke V c oun t s / c m −7 −6 −5 −4 −3 NaICsI
Energy Spectrum Due to Upward Protons d Energy (keV) / s / ke V c oun t s / c m −6 −5 −4 −3 −2 NaICsI
Energy Spectrum Due to Secondary Neutrons e Fig. 5
The energy deposition spectrum in the NaI and CsI crystals due to various backgroundcomponents incident spectrum of (a) CDGRB photon, (b) secondary gamma-ray photons, (c)downward going protons, (d) upward going protons in atmosphere, (e) albedo neutrons.
We compare the response of the detector for a known source over the total noiselevel due to various background components. So we carry out a simulation for theincident GRB spectrum shown in Figure 3. Moreover we consider two source positions:on-axis (0 ◦ ) and off-axis (50 ◦ ) location with respect to the detector axis.In Figure 6(a,b) we have shown the energy deposition spectrum in the NaI and CsIcrystals due to the GRB incident spectrum for the on-axis (0 ◦ ) (in red colour) andoff-axis (50 ◦ ) (in blue colour) source position. Energy (keV) / s / ke V c oun t s / c m −4 −3 −2 −1 on−axis srcoff−axis src a NaI Detector Energy Spectrum Due to GRB
Energy (keV) / s / ke V c oun t s / c m −4 −3 −2 −1 on−axis srcoff−axis src b CsI Detector Energy Spectrum Due to GRB
Fig. 6
The energy deposition spectrum in the (a) NaI and (b) CsI crystals due the GRBincident spectrum (for on-axis (0 ◦ ) and off-axis (50 ◦ ) source position). It is observed that the GRB spectrum in NaI is highly absorbed (due to the FOVwalls and shielding at the collimator wall) in the entire energy range for off-axis positionof the source (Figure 6a) and we can also see the peak around 60 keV is due to the Ta K α fluorescent emission.In case of CsI crystal, the photon energy below ∼ keV is getting absorbedboth for off and on-axis source position. For on-axis, the photons below ∼ keV ismostly detected by NaI crystal and we can also notice a emission peaks around 30 keV due to the K α and K β fluorescence of Iodine, whereas for off-axis the photons below ∼ keV are partially blocked by the shielding material around the lower part of thecollimator. Detection of source photon below ∼ keV is less significant in CsI crystal.This feature is clearly seen for both the off and on-axis source position (Figure 6b).In Table 3, we present the number of photons from the incident GRB which depositenergy in the NaI and CsI detectors. We have subdivided the whole energy range of10 − keV into smaller energy bands. We have given both the results for the on-axisand the off-axis source positions and the counts due to the total estimated background(fourth column). The background counts within parenthesis is from the prompt sources.To obtain the total backgrounds we need to include the contribution from the long-term activations in the detector. In Section 5 we showed that this prompt backgroundonly accounts for up to ∼
40% of the total background in the NaI crystal. So in orderto compare the photon counts due to the GRB with the total background noise, we havemultiplied the background due to prompt emission by a factor of 2 . − keV for on-axis case. However, for the off-axis case, the S/N does not permit todetect the source at all. On the other hand, for the on-axis case, CsI has a quite highS/N value in the energy range of 50 − keV and for the off-axis case this energy rangegets broadened in 20 − keV . So this simulation of the GRB ensures the capabilityof such source detection by the RT-2/S instrument. Based on the orbital and detectorconfigurations a conservative estimate suggests a detection of ∼
20 GRBs per year. ◦ ) and off-axis (50 ◦ ) orientation of the source posi-tion and due to the total estimated background spectrum with the simulated total promptbackground in the bracket (see text for details). Energy range on-axis ( ◦ ) off-axis ( ◦ ) total bkg. (prompt)in (keV) GRB GRBNaI −
20 1348 . . . . −
50 4764 . . . . −
100 3449 . . . . −
150 996 . . . . − . . . . CsI −
20 21 . . . . −
50 167 . . . . −
100 512 . . . . −
150 692 . . . . −
500 848 . . . . − . . . . The parameters used to construct the virtual detector of RT-2/CZT payload withinthe Geant4 toolkit environment is given in Table 4. These parameter values are sameas that used in onboard detector design and we applied some simplifications as wehave already mentioned in section 3. A 3D view of the RT-2/CZT payload along withdifferent components (e.g., detectors, CAM (Coded Aperture Mask), FZP (FresnelZone Plate), collimator etc.) as considered for the simulation is shown in Figure 7.This figure is the direct output from the simulation.
Fig. 7
A schematic 3D view of RT-2/CZT payload with different components (eg. detectors,CAM (Coded Aperture Mask), FZP (Fresnel Zone Plate), collimator etc.).5Table 4: Payload construction parameter specifications of RT-2/CZT.
Detector Parts Dimensions
Bottom plate 25 . × . × . cm Detector Housing 20 . × . × . cm (inner dimension)wall thickness 0 . cm Top Plate 23 . × . × . cm External Collimator 10 . × . × . cm (inner dimension)wall thickness 0 . cm Internal Collimator 10 . × . × . cm (inner dimension)wall thickness 0 . cm FOV 0 . cm Aluminum shielded by 0 . cm Tantalumtwo perpendicular walls to divide the collimator in four quadrants.CZT Crystal 4 . × . × . cm (3 crystals)gap between modules 0 . cm CMOS 2 . × . × . cm CAM 0 . cm thick Tantalum CAM on top of two collimator quadrants.1 mm Al filter used for one CAM and other one is open to sky.Zone Plate 0 . cm thick Ta zone plates at the face and endof the collimator on one CZT and CMOS crystalsShielding 0 . cm Tantalum around the collimator wall.0 . cm Tantalum below the top plate.0 . cm Tantalum around the detector housing side plates.Shielding weight ∼ g . ∼ keV ) due to Ta shielding of thickness 0 . mm on top plate and around theside plates of the payload. The simulated results for the other two CZT modules aremostly identical and hence not discussed here. The energy response spectrum in CZTand CMOS due to the secondary gamma-ray spectrum is shown in Figure 8b. Thesespectra also show more or less the same features as the primary gamma-ray (CDGRB)spectrum. Figure 8c presents the energy deposition spectrum in the two crystals forthe downward going protons, while Figure 8d shows the same for the upward goingprotons. In these two cases we can see a relatively higher energy deposition in theCMOS than in the CZT. The energy response of CZT and CMOS for the secondaryneutron spectrum is given in Figure 8e. Figure 8(a-e) also shows the dominance of theCDGRB and secondary photon spectrum in the noise contribution.4.2 Response of a GRB spectrum in RT-2/CZTWe now simulate the detector for a typical GRB spectrum (Eqn. 7) as the incidentspectrum shown in Figure 3 for two source location as mentioned in section 3.2.In Figure 9(a-d), we plot the energy deposition spectrum separately for the threeCZT modules and CMOS detector both for the on-axis (0 ◦ ) and off-axis (50 ◦ ) GRBposition. Figure 9a shows the energy deposition spectrum in the CZT (CZT1) contain-ing a CAM along with an 1 . mm thick Al filter in front of the collimator facing the Energy (keV) / s / ke V c oun t s / c m −5 −4 −3 −2 −1 CZTCMOS
Energy Spectrum Due to CDGRB a Energy (keV) / s / ke V c oun t s / c m −4 −3 −2 −1 CZTCMOSEnergy Spectrum Due to Secondary Gamma−Rays b Energy (keV) / s / ke V c oun t s / c m −5 −4 −3 −2 CZTCMOS
Energy Spectrum Due to Downward Protons c Energy (keV) / s / ke V c oun t s / c m −7 −6 −5 −4 −3 CZTCMOS
Energy Spectrum Due to Upward Protons d Energy (keV) / s / ke V c oun t s / c m −6 −5 −4 CZTCMOS
Energy Spectrum Due to Secondary Neutrons e Fig. 8
The energy deposition spectrum in the CZT and CMOS due to (a) CDGRB photons,(b) secondary gamma-ray photons, (c) downward going protons, (d) upward going protons and(e) albedo neutrons. sky. The spectrum in the CZT (CZT2) containing a CAM at the face of the collimatoris given in Figure 9b. Figure 9c presents the same in the CZT (CZT3) containing apair of aligned zone plates (FZPs) placed at 32 cm apart in third quadrant of thecollimator. In Figure 9d we have shown the spectrum of CMOS detector which is alsohaving two aligned zone plates (FZPs) placed in the fourth quadrant of collimator.It is evident from Figure 9(a-c) that all the CZT spectra are roughly the same irre-spective of their different configurations (Nandi et al. 2010) of the collimator for the Energy (keV) / s / ke V c oun t s / c m −4 −3 −2 −1 on−axis srcoff−axis src a CZT1 Detector Energy Spectrum Due to GRB Energy (keV) / s / ke V c oun t s / c m −4 −3 −2 −1 on−axis srcoff−axis src b CZT2 Detector Energy Spectrum Due to GRB Energy (keV) / s / ke V c oun t s / c m −4 −3 −2 −1 on−axis srcoff−axis src c CZT3 Detector Energy Spectrum Due to GRB Energy (keV) / s / ke V c oun t s / c m −4 −3 −2 −1 on−axis srcoff−axis src d CMOS Detector Energy Spectrum Due to GRB
Fig. 9
The energy deposition spectrum in the CZT and CMOS crystals due the GRB incidentspectrum for on-axis (0 ◦ ) and off-axis (50 ◦ ) source positions. (a) Spectrum in CZT with a CAMalong with Al sheet, (b) that in CZT with a CAM, (c) that in CZT with a pair of zone platesand (d) that in CMOS with a pair of zone plates. off-axis (50 ◦ ) source position though they differ for the on-axis source position (lowenergy photons are absorbed in CZT1 due to 1 mm Al shielding). This is because forthe off-axis source position most of the photons which are depositing their energy intoCZT or CMOS are coming through the detector side plates having uniform shieldingmaterial. Those photons interacting on the mask pattern are less likely to go to theCZT and CMOS to deposit their energy. All the CZT spectra show a prominent Tafluorescent peak (around 60 keV ), caused by the shielding material.In Table 5, we give the number of photon counts for three CZT modules andfor CMOS detector due to the GRB spectrum for the on-axis (0 ◦ ) and off-axis (50 ◦ )source positions and due to total estimated background spectrum along with the promptbackground noise as discussed in section 3. We have subdivided the total energy rangeof 10 − keV into several energy bands. These results could be useful while analyzingthe data from real observation (GRB) with RT-2 instruments.From the results in Table 5 it appears that with CZT detector, we can detectGRBs quite confidently in the energy range of 10 − − keV for the off-axis case. ◦ ) and off-axis (50 ◦ ) source position and due tototal estimated background spectrum with the simulated total prompt background inside thebracket (see text for detail). Energy range CZT1 CZT2 CZT3 CMOS(in keV) on-axis ( ◦ ) GRB −
20 182 . . . . −
50 1049 . . . −
100 939 . . . . −
150 408 . . . −
200 141 . . . . − . . . off-axis ( ◦ ) GRB −
20 7 . . . . −
50 52 . . . −
100 210 . . . . −
150 173 . . . −
200 98 . . . . − . . . total bkg. (prompt) −
20 2 . .
1) 2 . .
0) 4 . .
7) 2 . . −
50 16 . .
5) 17 . .
1) 16 . . −
100 59 . .
7) 56 . .
6) 59 . .
7) 11 . . −
150 33 . .
1) 34 . .
8) 35 . . −
200 24 . .
8) 21 . .
5) 23 . .
3) 1 . . − . .
9) 35 . .
3) 39 . . We compare the simulation results for the various background components with thereal data measured by the detector in its in-flight operation. We consider one set ofonboard NaI spectral data of RT-2/S near the equatorial region and far from thetrapped charged particle regions. This data refers to the satellite position when theinstruments came out of the polar region, the count rates were found to be droppingslowly. We noticed that the low energy count rates stabilized faster than the high energycount rates. We took the background region near the equatorial region when: (a) thelow energy ( < keV ) count rates were steady (better than a percent stability in 10minutes) and (b) when the high energy count rates ( > keV ) were steady by about5% in 10 minutes.The primary energy range of the instrument for spectroscopic measurements is ∼ − keV in the NaI detector. Hence we have concentrated on predicting backgroundaround this region.The comparison of the data in the energy range of 20 − keV with the simulatedbackground components of primary and secondary gamma rays, downward and upwardgoing protons (consisting primary CR and secondary protons) and secondary neutronsare shown in Figure 10.From Figure 10 we observe that the total simulated background noise due to themain sources of the prompt background components are not sufficient to explain thetotal measured background flux. This prompt background components are able toaddress only about 40% of the measured background. The rest of the background noiseis suspected to be due to the long-term detector activation which is to be consideredin a future work. Among the simulated components the most significant contributor( ∼ Energy (keV)20 30 40 50 60 70 80 90 100 / s / ke V P h / c m −7 −6 −5 −4 −3 −2 −1 Onboard dataTotal simulated CDGRB Secondary Gamma−Ray Downward ProtonUpward ProtonSecondary Neutron
Fig. 10
Onboard RT-2/S NaI background spectrum along with various simulated backgroundcomponents of CDGRB photons, secondary photons, downward going protons, upward goingprotons and albedo neutrons. photons contribute ∼ ∼ .
5% and rest from the otherfactors.For onboard calibration of the detectors (RT-2/S and RT-2/G), we used radioactivesource Co-57 (122 keV) that is placed into one of the slats of the collimator. Extensivecalibration of both the Phoswich detectors (NaI/CsI) on the ground reveals that the ∼ keV emission line (Debnath et al. 2010) is the intrinsic detector backgroundfeature (not due to the Tantalum (Ta) shielding around the collimator mesh). Theemission feature of ∼ keV , earlier reported by Gruber et al. (1996) was also observedin the onboard data of RT-2/S and RT-2/G (Nandi et al. 2009). The extra emissionline feature at around 20 keV (cutoff energy of RT-2/S is around 15 keV ) could bedue to the detector’s electronic noise, which was not observed during ground testingand the investigation is still ongoing. These two emission features characteristics willbe taken up in a separate paper, while considering the long-term activation aspectsof the background components in the detector. Also the comparison of the simulatedresult with the real data from the other active volumes are under investigation due tothe calibration concern of the detectors.The CsI detector in the phoswich combination is primarily used for backgroundrejection and we use it only for very bright sources like GRBs. To investigate thebackground in CsI for faint sources, one also needs to simulate the exact instrumentcharacteristics like the pulse shape for partial energy deposition in NaI and CsI. Theseactivities would be taken up in a separate paper. We have, however, looked at the CsIspectra and ensured that the background spectral shape broadly agrees with the modelprediction. Background simulations of space-borne payloads is one of the challenging tasks to un-derstand the space environment as well as the effect of high energy radiation (photons,charged particles, neutrons) on the detectors itself. The effects of the major promptbackground components like cosmic diffused gamma-ray background, secondary gamma-ray photons, primary cosmic-ray protons, secondary protons and albedo neutrons onthe RT-2 payloads (RT-2/S, RT-2/G and RT-2/CZT) are studied in detail, whichhelped the background calibration and source data extraction from the RT-2 Experi-ment. The weight of the material (Ta) that is used for shielding purpose for RT-2/S(RT-2/G) and RT-2/CZT payload is around 35 g and 951 g respectively.The current work of estimating the prompt background noise covers ∼
40% ofthe measured background. The rest part of the background noise is probably due tothe long-term activation of the detector materials due to the CR or trapped chargedparticles which is to be estimated in subsequent work (Zoglauer et al. 2008; Zoglauer,2009)As we have already mentioned that this experiment is primarily designed for thespectroscopic measurement of Solar flares, but it is also capable of detecting GRBs.Since the energy threshold for these GRB detection is about 100 keV (for off-axis sourcepositions) (see, Figure 6b), they will be sensitive for GRBs > − ergs/cm . Theprobability of on-axis detection is less than 1 in 1000 and hence GRBs are not expectedto be detected on-axis. As for the uniqueness of the experiment, if we have a reasonablespectral response above 100 keV , it is possible to constrain the spectral parametersusing the data in conjunction with other contemporaneous spectral measurements. Theexercise of the present paper is to demonstrate that off-axis response can be handledin a reasonable way.On 30th January, 2009, the CORONAS-PHOTON satellite was launched success-fully and all the RT-2 instruments are functioning to our satisfaction. Already severalgamma ray bursts (Rao et al. 2009; Chakrabarti et al. 2009abc) and solar flares havebeen detected by the instrument. Detailed reports on the on-board data quality andbackgrounds would be discussed elsewhere. Detailed results on the observed GRBs andsolar flares are also being submitted for publication elsewhere. Acknowledgements
RS and TBK thank RT-2/SRF fellowship (ISRO) which supported theirresearch work. The authors are thankful to ICSP/TIFR/VSSC/ISRO-HQ for various supportsduring RT2 related experiments. We are thankful to an anonymous referee for his very helpfulcomments which improved the paper substantially.
References
1. Agostinelli, S. et al.: G4-A Simulation Toolkit, Nuclear Instruments and Methods in PhysicsResearch. A 506, 250-303 (2003)2. Ajello, M. et al.: Cosmic X-Ray Background and Earth Albedo Spectra with SWIFT BAT,ApJ. 689, 666-677 (2008)3. Alcaraz, J. et al.: Protons in Near Earth Orbit, Physics Letters B. 472, 215-226 (2000)4. Armstrong, T. W. et al.: Calculation of Neutron Flux Spectra Induced in the Earth’sAtmosphere by Galactic Cosmic Rays, Journal of Geophysical Research. 78, 2715-2726 (1973)5. Band, D. L. et al.: BATSE observations of gamma-ray burst spectra. I - Spectral diversity.ApJ, 413, 281 (1993)6. Chakrabarti, S. K. et al.:GRB 090820: detection of a strong burst by RT-2 on board CORO-NAS PHOTON. GCN Circular No. 9833 (2009a)17. Chakrabarti, S. K. et al.:RT-2 observation of the bright GRB 090926A, GCN Circular No.10009 (2009b)8. Chakrabarti, S. K. et al.:Detection of a short GRB 090929A by RT-2 Experiment. GCNCircular No. 10010 (2009c)9. Dean, A. J., Lei, F., Knight, P. J.: Background in Space-borne Low-energy γγ