AMEGO: Exploring the Extreme Multimessenger Universe
AAMEGO: Exploring the Extreme Multimessenger Universe
Carolyn A. Kierans a and the AMEGO Team ba NASA Goddard Space Flight Center, 8800 Greenbelt Rd, Greenbelt, Md, USA b https://asd.gsfc.nasa.gov/amego/team.html ABSTRACT
The All-sky Medium Energy Gamma-ray Observatory (AMEGO) is a Probe-class mission concept that willprovide essential contributions to multimessenger astrophysics in the next decade. AMEGO operates bothas a Compton and pair telescope to achieve unprecedented sensitivity between 200 keV and > Keywords:
Gamma-ray astrophysics, multimessenger astrophysics, Compton telescope, pair conversion tele-scope
1. INTRODUCTION
The All-sky Medium Energy Gamma-ray Observatory (AMEGO) is a Probe-class mission concept that will pro-vide groundbreaking new capabilities for multimessenger astrophysics. AMEGO will cover the energy range from200 keV to > GW170817, and TXS 0506+056 have not only provided theelectromagnetic counterpart for identification, but analysis of the gamma-ray signal has led to unique insightsnot attainable in other energy bands. The next generation of gravitational wave (GW) and neutrino detectorswill significantly increase the number of source detections with a sensitivity beyond what the current gamma-raymissions can match; there needs to be a similar advancement in gamma-ray missions to provide the essentialelectromagnetic signature of these sources.AMEGO will be a key contributor to the next-generation revolution in multimessenger astrophysics by study-ing the astrophysical objects that produce GW and neutrinos to answer compelling science questions in theextreme Universe. The three main science goals of AMEGO are to: • Understand the physical processes in the extreme conditions around compact objects involved in gravita-tional wave events and other energetic phenomena
Further author information:E-mail: [email protected], Telephone: (301) 286-7628 a r X i v : . [ a s t r o - ph . I M ] J a n able 1: AMEGO has been optimized for excellent flux sensitivity, broad energy range, and large field of viewto enable the study of multimessenger sources. Energy Range
200 keV to > Angular Resolution per Photon ◦ (1 MeV), 2 ◦ (100 MeV), 1 ◦ (1 GeV) Energy Resolution
1% (1 MeV, FWHM/E), ∼
10% (1 GeV, FWHM/E)
Field of View
Continuum Sensitivity × − (1 MeV), 3 . × − (100 MeV) erg cm − s − in 5 years Line Sensitivity × − ph cm − s − for the 1.8 MeV Al line in 5 years
Polarization Sensitivity
4% MDP for a 100 mCrab flux, observed for 10 s • Resolve the processes of element formation in extreme environments, such as kilonovae and supernovae • Decipher the operating processes of jets in extreme environments such as gamma-ray bursts and activegalactic nucleiIn the past few decades, technological developments have significantly improved the performance of recent MeVgamma-ray missions. Table 1 summarizes the expected performance of a few key telescope parameters forAMEGO. These will be further discussed in Sec. 3.In addition to the groundbreaking multimessenger astrophysics achievable with AMEGO, the mission willprovide unprecedented sensitivity in the historically under-explored MeV gamma-ray regime. AMEGO is ageneral-purpose observatory that will provide new discovery capabilities across four orders-of-magnitude in en-ergy, covering the range from X-ray to GeV missions. Combining exceptional continuum sensitivity, advancednuclear line spectroscopy, and gamma-ray polarimetry, AMEGO will contribute unique measurements of theextreme universe and provide excellent synergies with observations at other wavelengths.The AMEGO mission concept has been supported by over 200 international scientists and was submittedto the Astro2020 Decadal Survey.
4, 5
For a more detailed discussion of AMEGO’s science capabilities, refer toRefs. 4, 5, and the list of relevant Astro2020 White Papers ( https://asd.gsfc.nasa.gov/amego/science.html ).In this paper we will give a detailed description of the AMEGO telescope in Sec. 2. Detailed simulationshave been used to predict the performance of AMEGO, and these are presented in Sec. 3. Finally, in Sec. 4, weintroduce the work being done to build a prototype AMEGO instrument and give an update on the development.
2. THE AMEGO INSTRUMENT
Compton scattering is the dominant photon-matter interaction from a few hundred keV to ∼
10 MeV (dependingon the scattering material), while pair conversion dominates at higher energies. AMEGO takes advantage ofthese two interaction mechanisms to detect gamma rays with unprecedented sensitivity across a wide field ofview (FOV) and four orders of magnitude in energy. To motivate the AMEGO instrument design, we will presentan overview the design considerations for the detection of both relevant interaction types.The interaction cross-section for photons and matter depends on the atomic number (Z) of the material. Toenhance the Compton scattering cross-section, and to minimize the effect of Doppler broadening on the angularresolution, a low-Z detector material should be used for the first Compton scatter interaction. However, thephoton must be fully absorbed to measure the total energy of the event, so a high-Z material should then be usedaround the scattering detector to enhance the chance of photoabsorption for the Compton-scattered photon. Inthe pair regime, high-Z materials promote pair-conversion, while thin layers of a low-Z detector material areideal for tracking the pair-products. Typically this is handled by combining layers of high-Z conversion foilswith position sensitive detectors, similar to the design on the Fermi -LAT. However, these high-Z foils limitthe angular resolution and energy measurements at lower energies due to multiple scattering; therefore, a pairtelescope optimized for the MeV range should use the tracker detector material itself as the converter. The pairevent must be fully contained to determine the energy, so a thick high-Z absorber detector should be used underthe tracking layers to measure the initial gamma-ray energy.2n a Compton scattering event, a gamma ray deposits its energy in two or more discreet interaction points thatmust be measured with high precision to reconstruct the origin of the photon. In pair-conversion, the gamma rayconverts into an electron-positron pair whose tracks through the detector material can be used to determine theinitial gamma-ray direction. Both interactions require fine position resolution (typically ∼ mm) in the scattereror tracker, while the resolution elements of the absorber detector are not as critical to the reconstruction.Additionally, Compton event reconstruction necessitates a precise measure of the energy deposited (typically∆ E/E of a few %) in the scatterer, and both event types require a spectrally-sensitive absorber.When these design considerations are combined, a hybrid telescope that is capable of detecting both inter-action types that consists of a low-Z Tracker, which serves as the scattering element for Compton events andconverter for pair events, and high-Z Calorimeter, to measure the energy, is ideal. Figure 1 illustrates how aTracker and Calorimeter geometry work to detect both event types. At energies less than ∼
10 MeV, a photon willpredominantly Compton scatter in the Tracker layers, and the Compton-scattered photon will then be absorbedin the Calorimeter. By a precise measure of the interaction position and energy deposited in the Tracker andCalorimeter, the standard Compton equation can be used to determine the Compton scatter angle θ of the firstinteraction, which constrains the initial photon direction to a circle on the sky called the “event circle.” An addedadvantage of using a tracking detector is that the direction of the Compton-scattered electron can be measuredif it subtends multiple layers, and with this additional kinematic information, the initial direction of the gammaray can be constrained to an arc on the sky, as opposed to a circle. At energies greater than ∼
10 MeV, a gammaray will predominantly undergo pair conversion in the Tracker, and the path of the electron-positron pair can betracked to determine the initial direction of the photon. The Calorimeter contains the electromagnetic showerthat develops as the electron-positron pair enter the high-Z detector volume to give the total energy in eachevent. Additionally, the position sensitive absorber allows for the shower profile to be imaged, which providesan important background discriminator and allows for reconstruction of events that are not fully contained.Observations in the MeV regime are background dominated; therefore, any reduction in background candramatically increase the sensitivity of the instrument. The first line of defence again the background in orbitis an anti-coincidence detector (ACD) which vetoes interactions in the detector from cosmic-rays. At the lowerenergies, one of the dominant background sources is activation of the instrument from the bombardment of thesecosmic rays. An additional challenge in the MeV range is the adverse effects of passive material; interactions inFigure 1: AMEGO detects gamma rays through both Compton scattering and pair production. At energiesless than ∼
10 MeV, a photon will predominantly Compton scatter, and a precise measure of the position andenergy of each interaction can constrain the initial photon direction to a circle on the sky. If the direction ofthe Compton-scattered electron is measured, this circle can be reduced to an arc, which reduces the backgroundcontribution. At energies greater than ∼
10 MeV, a gamma ray will undergo pair conversion, and the track ofthe electron-positron pairs in the instrument determine the initial photon direction. The energy is measured inthe calorimeter. 3igure 2: The AMEGO instrument consists of four subsystems: the silicon Tracker, the CZT Low EnergyCalorimeter, and the CsI High Energy Calorimeter, which are all surrounded by the ACD.
Left
The ACD andthe micrometeoroid shield (MMS) are cut away to expose the tower structure of AMEGO. The full instrumentmeasures 1.6 × × Right
An exploded view shows the three inner detector subsystems, where the Low EnergyCalorimeter surrounds the bottom third of the Tracker.passive material render Compton events useless and affect the energy and angular measurements for pair events.By reducing the amount of passive material near the detector, one can minimize these effects.With the above considerations, AMEGO has been designed to consist of four main subsystems, as shown inFig. 2. Photons will first interact in the 60 layers of double-sided silicon strip detectors (DSSDs), which formthe silicon Tracker, and acts as the scatterer and converter for Compton and pair events, respectively. DSSDsare chosen as the Tracker detector element because they provide excellent 3D position resolution, good energyresolution, and have a high technology readiness level (TRL). Surrounding the bottom of the Tracker sits thecadmium zinc telluride (CZT) Low Energy Calorimeter, which has excellent spatial and spectral resolution tomeasure Compton-scattered photons and low-energy pair showers. This novel subsystem also operates as a stand-alone Compton detector, not requiring a first interaction in the Tracker, which dramatically increases detectionefficiency of AMEGO at the lowest energies. At the bottom of the instrument, the cesium iodide (CsI) HighEnergy Calorimeter measures the electromagnetic shower from high-energy pair events. Finally, these detectorsystems are surrounded by a plastic scintillator ACD that rejects the charge-particle background in orbit. ThisTracker/Calorimeter geometry using silicon and CsI, which is similar to the
Fermi -LAT but optimized for lowerenergies, has been proposed for various MeV telescope concepts over the past two decades and is a well-understood design.For ease of integration and assembly, the AMEGO instrument is composed of four identical towers that makeup the four quadrants of the instrument. The electronics readout for each detector system is placed at the outeredges of these towers to reduce the amount of passive material within the detector volume. Furthermore, eachtower has a modular design that consists of identical detector elements to save cost and simplify construction andassembly. All primary structural elements that support the detectors are designed using low-Z carbon compositematerials to reduce activation of the instrument. The four AMEGO subsystems and the observatory operationswill be described in further detail in the following sections.
The AMEGO Tracker subsystem needs to have good spectral resolution and excellent position resolution tomeasure the first Compton scatter interaction and the track of pair-conversion products. Unlike the
Fermi -LAT, which uses alternating layers of single-sided silicon strip detectors to track particles, AMEGO needs to4igure 3: A single layer of the AMEGO Tracker con-sists of a 4 × µ m strip pitch for the 190 orthogonal strips on the front and back side of the detector. The wafers aredaisy-chained with wire bonds in a 4 × Fermi -LAT and other gamma-ray and cosmic-ray missions. DSSDshave been previously flown in AMS-02, Astro-H and Pamela, and, with a similar silicon tower structure,5uch of the integration and testing will be based off of the Fermi -LAT instrument. We have assessed theAMEGO Tracker to have TRL 6.
The performance in the Compton regime can be enhanced by a calorimeter with excellent spectral and spatialresolution. The function of this subsystem is two fold: to precisely measure the Compton-scattered photon frominteractions in the silicon, and to increase the effective area for the low energy Compton events which only interactin the calorimeter and have superior spectral information. This high-precision detector needs to be thick ( ∼ cm)to provide stopping power for MeV photons, and the coverage of this calorimeter up the sides of the Trackeris a balance between detecting large Compton scatter angles and maintaining a large FOV. The AMEGO LowEnergy Calorimeter has been optimized for narrow-line sensitivity and effective area in the Compton regime.The instrument requirements are satisfied with a thick segmented semiconductor detector. BrookhavenNational Laboratory (BNL) has recently developed thick CZT detectors that achieve excellent spatial and spectralresolution,
14, 15 while operating at room temperature. By using a large geometric ratio detector and placing agrounded electrode near the anode, the virtual Frisch-grid effect shields the anode from the holes. Analogous toa drift chamber, the relative anode and cathode signals give the depth of interaction within the thick detectorvolume. The novel aspect of the BNL work has been the segmentation of this virtual Frisch-grid electrode intoa four sensing electrode; using the relative amplitude of the induced charge on each side pad, the X-Y positionof interaction within the detector can be measured. This additional position information can in turn be usedto make correction for deformities in the CZT crystal structure and improve upon the energy measurement.Laboratory measurements have shown <
1% energy resolution at 662 keV, and a sub-mm position resolution inall 3 dimensions. Compared to pixelated CZT, these virtual Frisch-grid detectors can be made thicker, requirefewer electronic channels, and can utilize lower quality detector material. Since these virtual Frish-grid detectorshave been developed at BNL for ground-based research, there are a number of design considerations necessary fora space-based Compton-telescope application, such as detector packaging and high-voltage risks. The AMEGOteam is currently developing and optimizing these detectors for space applications and this work is discussedfurther in Sec. 4.2.The AMEGO Low Energy Calorimeter design uses 8 × ×
40 mm virtual Frisch-grid CZT bars. The bars arepacked in a 4 × . × . × ×
10 CZT base arraysplugged into a motherboard. These modules surround approximately the bottom one-third of the Tracker; twomodules are placed below the active area of the Tracker tower and two are rotated to cover each corner of theinstrument quadrants, as is seen in Fig. 2.The mechanical design for the Low Energy Calorimeter must minimize passive material, much like the Tracker.The CZT base array is supported by an interlocking structure fabricated of thin printed circuit board (PCB)material, which, in addition to mechanical support, provides electrical connections for the individual CZT bars.Small springs embedded in the walls of this structure guarantee a connection with the side sensing electrodes,which can be seen on the side walls in Fig. 4. The support structure and CZT module design is based onsignificant developments made with a prototype detector, which is described further in Sec. 4.2.The anode and cathode signals are unipolar; however, the induced signal on the side sensing electrodes isbipolar and wave-front sampling is required to get an accurate measure of the pad signals to determine the 3Dinteraction position within the detector. The 16 bars in the CZT base array, with 6 channels per bar, are readout with a single IDEAS IDE3421 digital ASIC located on the anode side of the array. The cathode and padsignals are routed through the side walls of the PCB array structure. The 4 kV high voltage required to biasthese 4 cm thick detectors is distributed on the top detector board, as seen in Fig. 4. Each module plugs intothe motherboard, which has a control FPGA for each CZT module (5 ×
10 CZT base arrays).The CZT Low Energy Calorimeter is a novel subsystem which enhances AMEGO’s low-energy response; how-ever, many of the components have flight heritage. CZT pixel detectors have flown on
Swift -BAT, AstroSat, however, the virtual Frisch-grid detectors have not yet been space qualified. The commercially avail-able IDE3421 ASIC is from a family of IDEAS ASICs with flight heritage. We asses the TRL of the Low EnergyCalorimeter to be 4, which is the lowest TRL system in AMEGO. To extend the sensitivity up to 1 GeV, the High Energy Calorimeter must provide approximately five radiationlengths to fully contain the highest energy pair-conversion events. Many of the subsystem requirements can beextracted from the
Fermi -LAT calorimeter, since it has the same function; however, the AMEGO calorimetermust be optimized for lower energy pair events that extend beyond the Low Energy Calorimeter. This low-energy optimization can be achieved by lowering the energy threshold and improving the energy resolution ofthe calorimeter. A modest position resolution is needed ( ∼ cm) to image the large electromagnetic showers thatare created as the pair products go through the calorimeter. The AMEGO High Energy Calorimeter has beenoptimized to extend the energy response of AMEGO beyond ∼
100 MeV.The AMEGO High Energy Calorimeter design is modeled after the
Fermi -LAT calorimeter and uses CsIscintillators with SiPM readout to enhance the low energy response of the calorimeter. Thallium doped CsI barsthat are each 1 . × . ×
38 cm are arranged in 6 layers in a hodoscopic pattern, with each alternating layeroriented orthogonal to the one above. A single layer of the High Energy Calorimeter contains 26 bars to coverthe total area beneath the Tracker, as shown in Fig. 5. The bars are wrapped in reflective material to maximizelight collection efficiency.Each High Energy Calorimeter layer of 26 bars is supported in a tray made of composite materials. Whilepassive material can still have adverse affects, it is not as critical in the High Energy Calorimeter since the energyresolution requirements for high energy pair events is not strict. However, low-Z material is still important toreduce the activation of the instrument.The CsI scintillation light is collected on both ends of the CsI bars by an array of SensL-J series SiPMs.The energy of the interaction is given by the geometric mean of the pedestal-subtracted pulse heights of thetwo signals. The relative amplitude of signals on each end determines the depth of interaction along the bar to σ = 1 cm resolution at 1 MeV. The SiPM signals are processed with IDEAS VA32TA6 ASICs located on theFEE boards at the perimeter of the layer, and the corner of each layer contains a digital backend board with anFPGA that controls the ASICs for a single layer and communication with the rest of the instrument With theFigure 5: A single layer of the AMEGO High EnergyCalorimeter has 26 1 . × . ×
38 cm CsI bars. Thescintillation light is collected by SiPMs on either end,and the signals are fed to the FEE boards on the outeredges of the layer. The bars are supported by a com-posite tower tray and bar retainers. Figure 6: The AMEGO ACD consists of five plas-tic scintillator panels with wavelength shifting (WLS)bars on each edge. The WLS bars are read out bySiPMs. Each panel is supported by the compositesturctural panels and frame, as shown in the explodedview of the ACD subsystem here.7se of the SiPM readout, the AMEGO calorimeter has significant improvement over
Fermi -LAT at low energieswith a threshold of 60 keV, compared to 600 keV for the
Fermi -LAT PIN diode readout. The CsI High Energy Calorimeter has been modeled off of the
Fermi -LAT calorimeter and is being designedand built by the same team at the Naval Research Laboratory (NRL). The
Fermi -LAT has given hodoscopic arraysof CsI scintillating bars flight heritage. The SensL-J SiPMs used in the AMEGO readout have been recentlyflight qualified on SIRI, and will soon be flown on BurstCube. The IDEAS VA32TA6 ASIC have flightheritage on Astro-H, eXTP, and CALET. We asses the TRL of the High Energy Calorimeter subsystemto be 6.
The dominant source of background for gamma-ray telescopes in orbit comes from the abundance of cosmic rays.The shear number of cosmic rays is orders of magnitude more than gamma rays from astrophysical sources, andthus an effective rejection of these background events is required to keep event rates down and not overwhelm thesystem with unwanted data. All modern high-energy telescopes use an ACD to reject the cosmic-ray background,and the subsystem design is well understood. For example, the
Fermi -LAT plastic scintillator ACD surroundsthe LAT Tracker and is segmented to avoid self-vetoing from the backsplash in high-energy pair events. AsAMEGO is optimized for lower energies, this level of segmentation is not necessary.The AMEGO ACD consists of five plastic scintillator panels that surround the Tracker and Low EnergyCalorimeter. The ACD does not extend to the sides of the High Energy Calorimeter to avoid self-vetoes fromthe electromagnetic shower in pair events. The AMEGO ACD side panels are 134 × × . Fermi -LAT ACD; however, single panels of scintillating plasticare used instead of the segmented ACD of the LAT. The SensL-J series SiPMs, which are also being used forthe High Energy Calorimeter readout, have been recently flight qualified on SIRI. The VATA64HDR16 ASICis part of a family of ASICs that has flight heritage on Astro-H, eXTP and CALET. We have assessed theAMEGO ACD to be TRL 6.
The AMEGO mission will fly in a low-inclination (6 ◦ ), low-earth orbit (600 km) and the prime mission will be5 years. The low-inclination angle is important to minimize the transit through the South Atlantic Anomaly(SAA) which increases the background due to activation of the instrument. With its wide FOV, AMEGOoperates predominantly in a survey mode, scanning the full sky every 3 hours (2 orbits). AMEGO will also havethe capabilities to perform inertial target pointing for targets of particular science interest.With an estimated data volume of 45 GB per day, AMEGO will use a high bandwidth Ka-band communi-cations subsystem for data downlink. The NASA Space Network will provide the primary space-to-ground link,and with a 5 min TDRSS contact every orbit, AMEGO has significant margin on the downlink capability.
3. EXPECTED PERFORMANCE
We have performed detailed simulations to study the expected performance of AMEGO using the Medium EnergyGamma-ray Astronomy Library (MEGAlib) software package. MEGAlib can perform Monte-Carlo simulationsof an instrument’s response to point sources and the background radiation environment in orbit. An accuratemass model of AMEGO, which includes all detector material and properties, as well as an approximation forthe passive material, has been built to determine a realistic performance. Simulations were done to determine8he energy and angular resolution, the effective area, and ultimately the sensitivity to sources of continuum andnarrow-line emission.To better understand the telescope response, we define three event classifications for AMEGO: • Untracked Compton events , generally with the lowest incident energy ( (cid:46) • Tracked Compton events , generally with an intermediate incident energy ( ∼ • Pair events , generally with the highest incident energy ( (cid:38)
10 MeV), are events where the first interactionis the conversion a gamma ray into an electron-positron pair.We look at each of these event classifications separately for the instrument performance parameters, but thesensitivity calculations use a combined response when more than one event classification is relevant. Figure 7shows the effective area, angular resolution, and energy resolution that were obtained from MEGAlib simulations,where we have simulated mono-energetic point sources at the instrument’s zenith for a sampling of energies acrossthe AMEGO band. These results will be discussed in detail in the following section prior to the presentationand discussion of the sensitivity plots.It is worth noting that MEGAlib was first developed for the Medium Energy Gamma-Ray Astronomy (MEGA)instrument built by the Max Plank Institute in the early 2000’s, which had a very similar design to AMEGO.Over the past twenty years, MEGAlib has been further developed through work with the COSI collaboration with a focus on the Compton regime. The MEGAlib pair event identification and reconstruction tools do notyet take into account the state-of-the-art algorithms developed for the Fermi -LAT, and thus we can expect animprovement in the pair regime relative to what is shown here.
The effective area is a measure of the instrument’s detection efficiency. It is given in units of cm since it isdefined as the area that an ideal absorber needs to detect an equivalent number of incident photons. Here,we defined the effective area as the number of valid events qualified by MEGAlib’s reconstruction tool, revan ,divided by the number of simulated incident events entering the telescope, and scaled by the simulated areasurrounding the mass model. The events are separated based on the three classification defined above and thereis no further selection on the reconstructed energy or photon origin.Figure 7 (a) shows the simulated effective area for the three event classifications in AMEGO. By combiningthe detection of Compton and pair events, AMEGO has an effective area that ranges from 500 to 1000 cm acrossfour decades of energy. The untracked Compton events are further separated into two categories. “UntrackedCompton in Silicon” events require their first interaction to be in the silicon Tracker. If this restriction is not madeand we also allow events which interact first in the CZT Low Energy Calorimeter, the effective area increasesdramatically. Although the Low Energy Calorimeter cannot track the direction of the scattered electron, with theexcellent energy and position resolution of the CZT, these events can still be reconstructed and used for imaging,particularly for sources of narrow-line emission. The effective area is not directly an instrument requirement,but it feeds into the sensitivity, as discussion in Sec. 3.4.9 -1 Energy (MeV)10 E ff e c t i v e A r e a ( c m ) Tracked ComptonUntracked ComptonUntracked Compton in SiliconPair (a) Effective Area Energy (MeV) A n g u l a r R e s o l u t i o n ( ◦ ) Tracked ComptonUntracked ComptonPair (b) Angular Resolution Energy (MeV)0.000.010.020.030.040.050.060.070.08 E n e r g y R e s o l u t i o n ( F W H M / E n e r g y ) Tracked ComptonUntracked Compton (c) Energy Resolution
Figure 7: (a) An effective area of 500–1000 cm isachieved by combining three different event classifica-tions across four orders of magnitude. (b) For Comptonand pair events, the angular resolution is 2.5 ◦ at 1 MeVand 1 ◦ at 1 GeV, respectively. (c) An energy resolutionof 1% FWHM/E is achieved at 1 MeV. See the text forfurther discussion of each instrument parameter. The angular resolution is defined in different ways for Compton and pair events. The angular resolution for aCompton telescope is described by the FWHM of the 1D point spread function (PSF), also known as the ARMdistribution. The ARM, or angular resolution measure, is the smallest angular distance between the nominalsource location and the event circle for each event; the distribution of all ARM values from a sample of Comptonevents gives effective 1D PSF of a Compton telescope. In the pair regime, the angular resolution is defined asthe 68% containment radius of the PSF. The pair PSF is given by the angular distance between the nominal andreconstructed photon direction for events for a point source.Figure 7 (b) shows the simulated angular resolution for the three event classifications in AMEGO. An angularresolution of 2.5 ◦ at 1 MeV and 1 ◦ at 1 GeV, as demonstrated here, satisfy the requirements needed to addressAMEGO’s science goals.Measuring the track of a Compton-scattered electron is not expected improve the ARM for an event since thedirection of the Compton-scattered electron is so poorly measured. In fact, as shown in Fig. 7 (b), the angularresolution for tracked events is slightly worse than untracked events for the same energies. This, however, is dueto a selection effect: for an event to be tracked at 1 MeV, for example, a significant fraction of energy must betransferred to the Compton-scattered electron, which, by the Compton equation, will result in a large Comptonscatter angle. Events with a larger Compton scatter angle inherently have a worse angular resolution due to thepropagation of error in the Compton equation. 10 -2 -1 Energy (MeV) -8 -7 -6 -5 -4 -3 -2 σ C o n t i nuu m S e n s i t i v i t y × E ( γ M e V c m − s − ) Fermi-LATEGRETSPI COMPTELNuSTAR
AMEGO
Energy (MeV) - - - ) - s - c m g N a rr o w L i ne S en s i t i v i t y ( s s) SPI (10 s) COMPTEL (10s) AMEGO (10AMEGO (5 yr survey)
Figure 8:
Left
The 3 σ on-axis point source continuum sensitivity for a 5 year AMEGO mission compared withthe Fermi -LAT (same incident angle and efficiency over 5 years), COMPTEL and EGRET (40% efficiencyover two weeks), and NuSTAR and SPI (exposure of 10 seconds). We assumed a 5-year mission with a 20%observation efficiency (due to field of view and South Atlantic Anomaly). Right
The 3 σ narrow-line sensitivityfor AMEGO is compared to INTEGRAL/SPI and COMPTEL. The energy resolution is given by the FWHM of the reconstructed photopeak reported as a percentage of theincident energy ∆
E/E . The energy resolution for pair events, which was found to be ∼
10% at 1 GeV, isnot shown here since it is not an instrument requirement. However, as discussed above, we expect the energyresolution in the pair regime to improve once the
Fermi -LAT reconstruction tools are implemented.Figure 7 (c) shows the energy resolution for Compton events. An energy resolution of 1% FWHM/E isachieved at 1 MeV. The energy resolution for Untracked Compton events is better than that seen for trackedCompton events for two reasons. First, the Low Energy Calorimeter dominates the Untracked Compton eventclassification and the CZT has better energy resolution than the DSSDs in the Tracker. Second, the energyresolution for tracked events will often be worse since more interactions are recorded (at least two in the tracker,by definition), and the errors add up for each measurement.
The sensitivity of a telescope is a measure of its capability to detect faint a sources; a lower sensitivity is better.For gamma-ray telescope, the sensitivity can be calculated based on the background rate, the effective area, theangular resolution, and, in the case of the narrow-line sensitivity, the energy resolution.The sensitivity has been calculated differently for the two regimes of the AMEGO telescope. In the Comptonregime ( (cid:46)
10 MeV), where the background is dominated by activation in the instrument and surrounding passivematerial, we have performed full background simulations in MEGAlib which include activation. We have thenused MEGAlib’s
SensitivityOptimizer program to determine the continuum sensitivity for this range. In the pairregime ( (cid:38)
10 MeV), where the backgrounds are well understood and modeled from
Fermi -LAT observations, wehave calculated the sensitivity analytically by I src = EA eff T obs × n sig (cid:115) n sig n sig A Eff T obs N B d Ω E , (1)where E is the energy, A eff is the effective area, T obs is the observation time, n sig is the significance (3 σ is usedhere), and N B is the background. The parameter d Ω is defined as 2 π (1 − cos(2 × P SF )), with
P SF given by theangular resolution. The background models used for both the input to the low energy MEGAlib simulations andthe high energy analytical calculation include Galactic, extra-galactic, and diffuse emission, while the activationsimulations also include models of cosmic-ray particles in low-earth orbit.11igure 9: The AMEGO prototype instrument is being built to advance and validate the hardware and softwaretools to be used in the AMEGO mission. Each prototype detector, housed in its own instrument box, fills a10 ×
10 cm area and is stacked to minimize the distance between the separate subsystems. The Tracker consistsof 10 layers of 10 ×
10 cm DSSDs, each supported and enclosed in an aluminum tray. Directly below theTracker is a single layer of 2 cm thick virtual Frisch-grid CZT detectors, and then finally, 5 layers of CsI barshodoscopically arranged. Not pictured here is the ACD detector which surrounds the full prototype instrumentfor the technology demonstration balloon flight in 2022.Figure 8 left shows the continuum sensitivity for the AMEGO five year mission compared to other x-rayand gamma-ray telescopes in the neighboring energy bands. AMEGO will provide over an order of magnitudeimprovement in continuum sensitivity over previous instruments, which will enable the next generation of multi-messenger astrophysics. With AMEGO’s prime operation in survey mode, an observation efficiency of 20% isassumed for the sensitivity.The “W” shape of the sensitivity curve is a result of the competing interaction processes in AMEGO. Wherethese competing processes overlap ∼
10 MeV, the current reconstruction algorithms in MEGAlib struggle withevent classification; however, there are current efforts to use machine learning approaches to improve this. Weexpect this bump to be reduced and a smooth line between the response at 1 MeV and at 100 MeV should beattained.The narrow-line sensitivity demonstrates the sensitivity of a telescope to sources of line emission. This is doneby taking the energy resolution into account in the sensitivity calculation. Since the requirements for narrow-lineemission are limited to the Compton regime, we only calculate the narrow-line sensitivity between 200 keV and10 MeV. For these calculations, we have used the full background simulation which takes into account activationof the instrument and the mono-energetic simulations discussed in the previous section.Figure 8 right shows the narrow-line sensitivity for AMEGO. A sensitivity of close to an order of magnitudeis achieved relative to INTEGRAL/SPI, which will allow for progress in resolving the processes of elementformation in extreme environments, one of AMEGO’s main science goals. It is important to note that SPI hasa 16 ◦ FOV and performs pointed observations, while the AMEGO sensitivity is for all-sky exposures.
4. HARDWARE PROTOTYPE
The AMEGO team has been developing a prototype instrument over the past few years to advance and validatethe hardware and software tools used in AMEGO. The prototype, often referred to as
ComPair in the literature,is funded through several NASA Astrophysics Research and Analysis (APRA) grants. The four subsystems ofAMEGO are being developed for the prototype instrument, but on a much smaller scale; see Fig. 9. The goalof the AMEGO prototype is to verify the AMEGO subsystems work together to detect and reconstruct bothCompton and pair conversion events in a relevant environment. The team is working towards a gamma-ray beamtest of the prototype instrument at the High Intensity Gamma-ray Source (HIGS) at Duke University, and a12igure 10: The AMEGO prototype Tracker consists of 10 layers of DSSDs. An exploded view of the customcarrier that uses elastomeric connectors, as opposed to wirebonds, is shown on the left. The photo on the rightshows a single fully integrated layer of the Tracker. The DSSD mounted in the custom carrier assembly is seenin the front quadrant, two AFEs which each contain 6 VATA460.3 ASICs are connected to the carrier signalboards (one is flipped upside down so the boards could be identical in design), and the back quadrant housesthe DBE which handles the data and communication for the layer.technology demonstration balloon flight in Fall 2022 (delayed due to COVID-19) from Fort Sumner, NM. Anoverview of the subsystems for the AMEGO prototype, description and current status, as well as the integrationand testing plan is given in the following sections.
The AMEGO prototype Tracker has been designed, built and tested by GSFC, with firmware support from LosAlamos National Laboratory (LANL). The subsystem consists of 10 layers of 10 ×
10 cm DSSDs fabricatedby Micron Semiconductor stacked with a 1.9 cm separation. The silicon wafers are 500 µ m thick with 510 µ mstrip pitch, giving 192 AC-coupled strips per side. There have been three revisions of the AMEGO prototypesilicon wafers with slight modifications of the strip width and resistor placement. The GSFC team is currentlyworking with Micron on a fourth revision that aims to optimize the detectors to have low leakage current whileminimizing interstrip capacitance so as to achieve the energy resolution goal of the subsystem. The wafershave been characterized with bulk leakage current and capacitance measurements, in addition to single stripmeasurements of the leakage current, interstrip capacitance, resistance, and coupling capacitance. See Sec. 2.1for an overview of the Tracker requirements and general operating principle.The wafers are mounted in a custom carrier, as shown in Fig. 10, which uses elastomeric connections insteadof wirebonds to route the signals from each strip to the front-end electronics (FEE). This carrier offers a flexibledesign, with the option of testing with different readout systems or daisy-chaining detectors to employ the ladderconfiguration of 4 detectors that the AMEGO design is based on. Capacitance, and therefore noise, scales as thelength of the chained strip, so understanding the performance of the ladder configuration is important for theAMEGO Tracker development. The team has also integrated one wire-bonded carrier to understand the possiblenoise contribution of the elastomeric connector. The wafers integrated in the custom carriers have been testedwith bench-top electronics to confirm functionality.The custom FEE are based on the IDEAS VATA460.3 ASIC. The ASIC has 32 channels and thus the FEEboards have 6 ASICs per detector side. The analog front end (AFE) boards are designed to handle both positiveand negative signals to readout the front (junction side) and back (ohmic side) of the detector. A digital back-end board (DBE) with an FPGA handles data from the 12 ASICs, telemetery, and communication with TriggerModule (see Sec. 4.5) for each layer. Each layer operates entirely independently from the other Tracker layers.A fully integrated silicon wafer in the custom carrier with the AFE and DBE boards mounted in the aluminumtray is shown in Fig. 10 right . The AFE and DBE have undergone extensive testing and are ready to mate withthe detector. The prototype Tracker, while delayed by COVID-19, is expected to undergo integration and testingin the coming months. See Ref. 31 for a more detailed description of the AMEGO prototype Tracker.13igure 11: The AMEGO prototype Low Energy Calorimeter consists of 16 modules that are each approximatelyone cubic inch and contain an array of 4 × The Low Energy Calorimeter is based on position-sensitive virtual Frisch-grid detectors that have been developedat Brookhaven National Laboratory (BNL). The CZT detectors, which are each 6 × ×
20 mm , have goldcontact on the anode (bottom) and cathode (top), and the relative drift time of the signal gives a measure of theZ coordinate of each interaction. To make the detector sensitive in three dimensions, four metal pads are placedon the side of the crystal near the anode. With these virtually grounded pads isolated from the detector, theyshield the anode as if a real Frisch-grid were placed inside the detector, and the amplitude of the signals read outfrom these electrodes are used to evaluate the X and Y coordinates of an interaction. Position resolution testswith a pulse laser have found a single crystal position resolution of σ =0.2-0.3 mm. Significant groundwork hasbeen completed by the BNL group in developing a 3D voxel correction for the spectral response of these bars,which can achieve excellent energy resolution for single crystals of 0.9% FWHM at 662 keV.
15, 33
The CZT bars are packaged in a module that encloses a 4 × × × and is shown inFig. 11, is the base element of the Low Energy Calorimeter detector plane. With each module plugging into amotherboard, one can make an array of any desired size with practically no “dead space.” For the AMEGOprototype, the subsystem consists of 16 modules which fill the 10 ×
10 cm area subtended by the Tracker DSSDs.The unique custom-made module design has undergone a number of revisions to resolve initial issues with leakagedown the edges of the module structure and has since performed well in extensive testing. In October 2020, asingle module was tested with a charged particle beam at the NASA Space Radiation Laboratory at BNL tomeasure activation and detector polarization.The module is read out by a single low-noise custom ASIC designed for the Frisch-grid detector that isconnected to the anode (bottom) side of the module. The top board of the module distributes an external2500 V high-voltage source to bias the detectors. Figure 11 right shows 10 modules mounted onto the motherPCB with the FPGA board, shown on the left of the photo, which handles data from all modules, distributes thelow-voltage power, and communicates with the Trigger Module. The CZT Low Energy Calorimeter subsystemwith 10 operating modules has undergone extensive testing and calibration. See Ref. 35 for a more detaileddescription of the AMEGO prototype CZT Low Energy Calorimeter. The High Energy Calorimeter prototype has been designed, built, and tested at the Naval Research Laboratory(NRL). It consists of 5 layers of six 1.7 × ×
10 cm CsI bars where each layer is orthogonal to the one above,resulting in a hodoscopic calorimeter. The bars, which are wrapped in Tetratex to increase light collection, areread out on either end by SensL ArrayJ 6 × quad (2 by 2 array) SiPMs. The relative amplitude of the14igure 12: The AMEGO prototype High Energy Calorimeter consists of 30 1.7 × ×
10 cm CsI bars arrangedin a 5-layer hodoscopic pattern. The bars are read out on each end with a 2 × σ =1.5 cm at 662 keV and an energy resolution of 3.5% at 662 keV.The 30 CsI bars are housed in a 3D printed plastic structure, as shown in Figure 12. The 60 SiPM channels(2 channels per bar) are read out with the new IDEAS 64-channel ROSSPAD (Read Out System for SiliconPhotomultiplier Avalanche Diodes), which contains 4 SIPHRA ASICs. The prototype High Energy Calorimeteralso uses an Arduino Due to interface with the Trigger Module, since the ROSSPAD cannot accept the 8-bit eventID that allows the subsystems to determine coincident triggers across the different subsystems. The prototypeCsI calorimeter has been interfaced and tested with the ROSSPAD readout, and calibration of the CsI bars withcollimated gamma-ray sources and atmospheric muons is underway. See Ref. 20 for a detailed description of theAMEGO prototype High Energy Calorimeter. The prototype ACD, which is being designed, tested, and built at GSFC, consists of 5 panels of 15 mm thickEljen EJ-208 plastic scintillator that surround the four sides and the top of the instrument. Each panel haswavelength shifting bars across two adjacent edges with air gap coupling, and the corner where they meet ismachined flat and read out by 2 × Each of the AMEGO prototype subsystems is able to operate as a stand-alone detector; however, the goal of theAMEGO prototype work is to have the subsystems work together as a Compton and pair-conversion telescope.In the current design, each subsystem has its own data acquisition (DAQ), and the Trigger Module is used todetermine when coincidence conditions between the subsystems are satisfied. The Trigger Module then sendsout a trigger acknowledgement and an 8-bit Event ID to all subsystem DAQ systems. The event ID and clockinformation can then be used for off-line event reconstruction for the full prototype instrument.The Trigger Module, designed and built at GSFC, uses the Xilinx FPGA Evaluation Board (ZC706) withtwo separate custom connector boards to distribute and receive signals from each FEE board; see Fig. 13. The15igure 13: The AMEGO prototype Trigger Module consists of a Xilinx FPGA Evaluation Board and two customconnector boards to send and receive signals from each of the subsystem FEE boards.connector boards allow for 14 separate FEE connections: the ACD and Low Energy Calorimeter each use one,the High Energy Calorimeter uses two, and the 10 layers of the Tracker take up 10 connections.The Trigger Module currently has 16 different trigger conditions that account for all possible coincidenceoutcomes in the four subsystems of the instrument. The ideal trigger condition corresponds to two hits inthe Tracker (coincidence between X-Y (top/bottom) strips in a single detector), and one in either calorimeter;however, other trigger conditions are being explored. The Tracker itself has 3 different coincidence models: 1)only one side of one detector has a hit (very susceptible to noise), 2) coincidence between X-Y on a single Trackerlayer, and 3) any two Tracker layers have a hit on one side of a detector. The variety of coincidence modes,in addition to a pre-scalings option which can be applied to the different trigger modes, allow for a flexiblesystem which can accommodate noisy and dead strips. The Trigger Module has been tested with mock detectorinputs and has been fully integrated with the Low Energy Calorimeter. Tests with the High Energy Calorimeterand Tracker are expected to take place over the next few months. See Ref. 36 for a detailed description of theAMEGO prototype Trigger Module.The software pipeline for the AMEGO prototype is under development on the instrument level and the teamis working towards the development of a single, system-wide DAQ and analysis system for the prototype balloonflight.
Each of the AMEGO prototype subsystems is housed in an individual aluminum box, as shown in Fig. 9. Thedistances between each of the subsystems and the layers of the Tracker were minimized in an attempt to maximizethe effective area and FOV of the instrument. The CsI High Energy Calorimeter sits 45 mm below the CZT LowEnergy Calorimeter, which is 35 mm below the last Si Tracker layer. In mechanical integration, the edges of thesubsystem boxes bolt together, as shown in Fig. 9. The precise alignment needed between the Tracker layersand the CZT calorimeter is achieved with alignment rods which pass through two corners of the Tracker carrierboards and into the CZT Low Energy Calorimeter box. Alignment between the CsI High Energy Calorimeterand the Low Energy Calorimeter does not need to be as accurate due to the poor position resolution of the CsIbars.While the subsystems are currently being developed and tested individually, they will be integrated, tested,and calibrated as a single instrument at GSFC. This includes basic functionality tests and a full-system coin-cidence test with the Trigger Module to confirm timing diagrams and trigger logic. After a calibration of eachsubsystem is performed, an off-line analysis will be done to confirm the sucessful reconstructed Compton eventsand muon tracks. A full calibration of the prototype’s reconstructed energy resolution, angular resolution, andeffective area will follow.The major test of the instrument will be at the free-electron laser HIGS beam facility at Duke University.The instrument will be tested at 5 different energies between 1.5 MeV to 29.5 MeV, which spans the difficult16ompton and pair conversion cross-over regime. The beam has a standard energy resolution of 3-4% and atunable flux that is expected to be set to ∼
5. CONCLUSION
The AMEGO Probe-class mission will enable multimessenger astrophysics in the next decade. Using four sub-systems that together measure Compton and pair events, AMEGO is sensitive across 4 orders-of-magnitude inenergy. A prototype instrument is currently being developed to validate the design of AMEGO, and tests areunderway.
ACKNOWLEDGMENTS
The author would like to acknowledge the significant work from Julie McEnery, Regina Caputo, Jeremy Perkins,and Judy Racusin that went into developing the AMEGO Request for Information (RFI) response to the As-tro2020 Decadal Survey that served as a basis for much of the work presented here. Iker Liceaga Indart createdall mechanical drawings for the AMEGO RFI and these proceedings. The author would also like to thank theComPair hardware team, namely the subsystem leads Rich Wulf, Alex Moiseev, and Jeremy S. Perkins.
REFERENCES [1] S. M. Matz, et al., “Gamma-ray line emission from sn1987a,”
Nature (6155), 416–418 (1988).[2] A. Goldstein, et al., “An Ordinary Short Gamma-Ray Burst with Extraordinary Implications: Fermi-GBMDetection of GRB 170817A,”
ApJL , L14 (Oct. 2017).[3] IceCube Collaboration, et al., “Multimessenger observations of a flaring blazar coincident with high-energyneutrino IceCube-170922A,”
Science , eaat1378 (July 2018).[4] J. McEnery, et al., “All-sky Medium Energy Gamma-ray Observatory: Exploring the Extreme Multimes-senger Universe,” in [
Bulletin of the American Astronomical Society ], , 245 (Sept. 2019).[5] J. McEnery, et al., “AMEGO: A Multimessenger Mission for the Extreme Universe - Response to Astro2020Decadal Request for Information.” https://asd.gsfc.nasa.gov/amego/files/AMEGO_Decadal_RFI.pdf (2019).[6] A. Zoglauer and G. Kanbach, “Doppler Broadening as a Lower Limit to the Angular Resolution of NextGeneration Compton Telescopes,” Proceedings of SPIE (2003).[7] W. B. Atwood, et al., “The Large Area Telescope on the Fermi Gamma-Ray Space Telescope Mission,”
ApJ , 1071–1102 (June 2009).[8] G. Kanbach, R. Andritschke, F. Schopper, V. Sch¨onfelder, A. Zoglauer, et al. , “The MEGA project,”
NewAstronomy Reviews , 275–280 (Feb 2004).[9] T. J. O’Neill, A. Aky¨uz, D. Bhattacharya, M. Polsen, J. Samimi, and A. Zych, “The TIGRE gamma-raytelescope,” in [ Gamma 2001: Gamma-Ray Astrophysics ], S. Ritz, N. Gehrels, and C. R. Shrader, eds.,
American Institute of Physics Conference Series , 882–886 (Oct. 2001).[10] A. De Angelis, et al., “The e-ASTROGAM mission. Exploring the extreme Universe with gamma rays inthe MeV - GeV range,”
Experimental Astronomy , 25–82 (Oct. 2017).[11] A. Kounine, et al., “The Alpha Magnetic Spectrometer on the International Space Station,” InternationalJournal of Modern Physics E , 1230005 (Aug. 2012).[12] G. Sato, et al., “The Hard X-ray Imager (HXI) for the ASTRO-H Mission,” in [ Space Telescopes andInstrumentation 2014: Ultraviolet to Gamma Ray ], T. Takahashi, J.-W. A. den Herder, and M. Bautz, eds., , 673 – 683, International Society for Optics and Photonics, SPIE (2014).[13] S. Straulino and Pamela Tracker Collaboration, “The PAMELA silicon tracker,”
Nuclear Instruments andMethods in Physics Research A , 168–172 (Sept. 2004).1714] A. E. Bolotnikov, et al., “A 4 × Nuclear Instruments and Methods in Physics Research A , 161036(Feb. 2020).[15] A. E. Bolotnikov, et al., “CdZnTe position-sensitive drift detectors with thicknesses up to 5 cm,”
AppliedPhysics Letters , 093504 (Feb. 2016).[16] A. E. Bolotnikov, et al., “Use of high-granularity position sensing to correct response non-uniformities ofCdZnTe detectors,”
Applied Physics Letters , 263503 (June 2014).[17] S. D. Barthelmy, et al., “The Burst Alert Telescope (BAT) on the SWIFT Midex Mission,”
Space ScienceReviews , 143–164 (Oct. 2005).[18] V. Bhalerao, et al., “The Cadmium Zinc Telluride Imager on AstroSat,”
Journal of Astrophysics andAstronomy , 31 (June 2017).[19] F. A. Harrison, et al., “The Nuclear Spectroscopic Telescope Array (NuSTAR) High-Energy X-Ray Mission,” The Astrophysical Journal , 103 (May 2013).[20] R. S. Woolf, J. E. Grove, B. F. Phlips, and E. A. Wulf, “Development of a CsI:Tl calorimeter subsystemfor the All-Sky Medium-Energy Gamma-Ray Observatory (AMEGO),” in [ ], 1–6 (2018).[21] L. J. Mitchell, B. F. Phlips, J. E. Grove, T. Finne, M. Johnson-Rambert, and W. N. Johnson, “StrontiumIodide Radiation Instrument (SIRI) – Early On-Orbit Results,” in [ ], 1–9 (2019).[22] J. Smith and BurstCube Collaboration, “BurstCube: Mission Concept, Performance, and Status,” in [ ], International Cosmic Ray Conference , 604 (July2019).[23] S. N. Zhang, et al.,, “eXTP: Enhanced X-ray Timing and Polarization mission,” in [ Space Telescopes andInstrumentation 2016: Ultraviolet to Gamma Ray ], J.-W. A. den Herder, T. Takahashi, and M. Bautz, eds., , 505 – 520, International Society for Optics and Photonics, SPIE (2017).[24] Y. Asaoka, et al., “The CALorimetric electron telescope (CALET) on the international space station:Results from the first two years on orbit,”
Journal of Physics: Conference Series , 012003 (feb 2019).[25] A. Zoglauer, R. Andritschke, and F. Schopper, “MEGAlib The Medium Energy Gamma-ray AstronomyLibrary,”
New Astronomy Reviews , 629–632 (Oct 2006).[26] J. Tomsick, et al., “The Compton Spectrometer and Imager,” in [ Bulletin of the American AstronomicalSociety ], , 98 (Sept. 2019).[27] V. Schoenfelder et al. , “Instrument description and performance of the Imaging Gamma-Ray TelescopeCOMPTEL aboard the Compton Gamma-Ray Observatory,” Astrophys. J. Suppl. , 657 (1993).[28] D. J. Thompson, et al., “Calibration of the Energetic Gamma-Ray Experiment Telescope (EGRET) for theCompton Gamma-Ray Observatory,” ApJS , 629 (June 1993).[29] G. Vedrenne, et al., “SPI: The spectrometer aboard INTEGRAL,” A&A , L63–L70 (Nov. 2003).[30] A. Zoglauer, “Using Deep Learning for the Event Reconstruction of Combined Compton-scattering and Pair-creation Telescopes,” in [
American Astronomical Society Meeting Abstracts ], American AstronomicalSociety Meeting Abstracts , 372.21 (Jan. 2020).[31] S. Griffin, C. A. Kierans, L. Parker, A. Schoenwald, P. Shawhan, R. Caputo, J. McEnery, J. S. Perkins,and ComPair Team, “Current Status of the ComPair Silicon Tracker,”
These Proceedings , 11444–323.[32] A. E. Bolotnikov, et al., “Optimization of virtual Frisch-grid CdZnTe detector designs for imaging and spec-troscopy of gamma rays,” in [
Hard X-Ray and Gamma-Ray Detector Physics IX ], R. B. James, A. Burger,and L. A. Franks, eds.,
Society of Photo-Optical Instrumentation Engineers (SPIE) Conference Series ,670603 (Sept. 2007).[33] A. E. Bolotnikov, et al., “Use of high-granularity position sensing to correct response non-uniformities ofCdZnTe detectors,”
Applied Physics Letters , 263503 (June 2014).[34] E. Vernon, et al., “Front-end asic for spectroscopic readout of virtual frisch-grid czt bar sensors,”