The DArk Matter Particle Explorer mission
J. Chang, G. Ambrosi, Q. An, R. Asfandiyarov, P. Azzarello, P. Bernardini, B. Bertucci, M. S. Cai, M. Caragiulo, D. Y. Chen, H. F. Chen, J. L. Chen, W. Chen, M. Y. Cui, T. S. Cui, A. D'Amone, A. De Benedittis, I. De Mitri, M. Di Santo, J. N. Dong, T. K. Dong, Y. F. Dong, Z. X. Dong, G. Donvito, D. Droz, K. K. Duan, J. L. Duan, M. Duranti, D. D'Urso, R. R. Fan, Y. Z. Fan, F. Fang, C. Q. Feng, L. Feng, P. Fusco, V. Gallo, F. J. Gan, W. Q. Gan, M. Gao, S. S. Gao, F. Gargano, K. Gong, Y. Z. Gong, J. H. Guo, Y. M. Hu, G. S. Huang, Y. Y. Huang, M. Ionica, D. Jiang, W. Jiang, X. Jin, J. Kong, S. J. Lei, S. Li, X. Li, W. L. Li, Y. Li, Y. F. Liang, Y. M. Liang, N. H. Liao, Q. Z. Liu, H. Liu, J. Liu, S. B. Liu, Q. Z. Liu, W. Q. Liu, Y. Liu, F. Loparco, J. Lü, M. Ma, P. X. Ma, S. Y. Ma, T. Ma, X. Q. Ma, X. Y. Ma, G. Marsella, M.N. Mazziotta, D. Mo, T. T. Miao, X. Y. Niu, M. Pohl, X. Y. Peng, W. X. Peng, R. Qiao, J. N. Rao, M. M. Salinas, G. Z. Shang, W. H. Shen, Z. Q. Shen, Z. T. Shen, J. X. Song, H. Su, M. Su, Z. Y. Sun, A. Surdo, X. J. Teng, X. B. Tian, A. Tykhonov, V. Vagelli, S. Vitillo, et al. (64 additional authors not shown)
TThe DArk Matter Particle Explorer mission
DAMPE collaboration: J. Chang a , G. Ambrosi b , Q. An c , R. Asfandiyarov d ,P. Azzarello d , P. Bernardini e,f , B. Bertucci g,b , M. S. Cai a , M. Caragiulo h , D.Y. Chen a,i , H. F. Chen c , J. L. Chen j , W. Chen a,i , M. Y. Cui a , T. S. Cui k , A.D’Amone e,f , A. De Benedittis e,f , I. De Mitri e,f , M. Di Santo f , J. N. Dong c , T.K. Dong a , Y. F. Dong l , Z. X. Dong k , G. Donvito h , D. Droz d , K. K. Duan a,i ,J. L. Duan j , M. Duranti g,b , D. D’Urso b,m , R. R. Fan l , Y. Z. Fan a , F. Fang j , C.Q. Feng c , L. Feng a , P. Fusco h,n , V. Gallo d , F. J. Gan c , W. Q. Gan a , M. Gao l ,S. S. Gao c , F. Gargano h , K. Gong l , Y. Z. Gong a , J. H. Guo a , Y. M. Hu a,i , G.S. Huang c , Y. Y. Huang a,i , M. Ionica b , D. Jiang c , W. Jiang a,i , X. Jin c , J.Kong j , S. J. Lei a , S. Li a,i , X. Li a , W. L. Li k , Y. Li j , Y. F. Liang a,i , Y. M.Liang k , N. H. Liao a , Q. Z. Liu a , H. Liu a , J. Liu j , S. B. Liu c , Q. Z. Liu a , W. Q.Liu j , Y. Liu a , F. Loparco h,n , J. L¨u k , M. Ma k , P. X. Ma a,i , S. Y. Ma c , T. Ma a ,X. Q. Ma k , X. Y. Ma k , G. Marsella e,f , M.N. Mazziotta h , D. Mo j , T. T.Miao a , X. Y. Niu j , M. Pohl d , X. Y. Peng a , W. X. Peng l , R. Qiao l , J. N. Rao k ,M. M. Salinas d , G. Z. Shang k , W. H. Shen k , Z. Q. Shen a,i , Z. T. Shen c , J. X.Song k , H. Su j , M. Su a , Z. Y. Sun j , A. Surdo f , X. J. Teng k , X. B. Tian k , A.Tykhonov d , V. Vagelli g,b , S. Vitillo d , C. Wang c , Chi Wang k , H. Wang k , H. Y.Wang l , J. Z. Wang l , L. G. Wang k , Q. Wang c , S. Wang a,i , X. H. Wang j , X. L.Wang c , Y. F. Wang c , Y. P. Wang a,i , Y. Z. Wang a,i , S. C. Wen a,i , Z. M.Wang j , D. M. Wei a,o , J. J. Wei a , Y. F. Wei c , D. Wu l , J. Wu a,o , S. S. Wu k , X.Wu d , K. Xi j , Z. Q. Xia a,o , Y. L. Xin a,i , H. T. Xu k , Z. L. Xu a,i , Z. Z. Xu c , G.F. Xue k , H. B. Yang j , J. Yang a , P. Yang j , Y. Q. Yang j , Z. L. Yang j , H. J.Yao j , Y. H. Yu j , Q. Yuan a,o , C. Yue a,i , J. J. Zang a , C. Zhang a , D. L. Zhang c ,F. Zhang l , J. B. Zhang c , J. Y. Zhang l , J. Z. Zhang j , L. Zhang a,i , P. F. Zhang a ,S. X. Zhang j , W. Z. Zhang k , Y. Zhang a,i , Y. J. Zhang j , Y. Q. Zhang a,i , Y. L.Zhang c , Y. P. Zhang j , Z. Zhang a , Z. Y. Zhang c , H. Zhao l , H. Y. Zhao j , X. F.Zhao k , C. Y. Zhou k , Y. Zhou j , X. Zhu c , Y. Zhu k , and S. Zimmer d a Key Laboratory of Dark Matter and Space Astronomy, Purple Mountain Observatory,Chinese Academy of Sciences, Nanjing 210008, China b Istituto Nazionale di Fisica Nucleare Sezione di Perugia, I-06123 Perugia, Italy c State Key Laboratory of Particle Detection and Electronics, University of Science andTechnology of China, Hefei 230026, China d Department of Nuclear and Particle Physics, University of Geneva, CH-1211, Switzerland e Universit`a del Salento - Dipartimento di Matematica e Fisica ”E. De Giorgi”, I-73100,Lecce, Italy f Istituto Nazionale di Fisica Nucleare (INFN) - Sezione di Lecce , I-73100 , Lecce, Italy g Dipartimento di Fisica e Geologia, Universit`a degli Studi di Perugia, I-06123 Perugia, Italy h Istituto Nazionale di Fisica Nucleare Sezione di Bari, I-70125, Bari, Italy i University of Chinese Academy of Sciences, Yuquan Road 19, Beijing 100049, China j Institute of Modern Physics, Chinese Academy of Sciences, Nanchang Road 59, Lanzhou730000, China Corresponding author (email: [email protected]) Also at Department of Physics and Laboratory for Space Research, The University ofHong Kong, Pokfulam Road, Hong Kong
Preprint submitted to Astroparticle Physics September 15, 2017 a r X i v : . [ a s t r o - ph . I M ] S e p National Space Science Center, Chinese Academy of Sciences, Nanertiao 1,Zhongguancun, Haidian district, Beijing 100190, China l Institute of High Energy Physics, Chinese Academy of Sciences, YuquanLu 19B, Beijing100049, China m ASI Science Data Center (ASDC), I-00133 Roma, Italy n Dipartimento di Fisica ”M.Merlin” dell’Univerisity e del Politecnico di Bari, I-70126,Bari, Italy o School of Astronomy and Space Science, University of Science and Technology of China,Hefei, Anhui 230026, China
Abstract
The DArk Matter Particle Explorer (DAMPE), one of the four scientific spacescience missions within the framework of the Strategic Pioneer Program onSpace Science of the Chinese Academy of Sciences, is a general purpose high en-ergy cosmic-ray and gamma-ray observatory, which was successfully launched onDecember 17th, 2015 from the Jiuquan Satellite Launch Center. The DAMPEscientific objectives include the study of galactic cosmic rays up to ∼
10 TeVand hundreds of TeV for electrons/gammas and nuclei respectively, and thesearch for dark matter signatures in their spectra. In this paper we illustratethe layout of the DAMPE instrument, and discuss the results of beam tests andcalibrations performed on ground. Finally we present the expected performancein space and give an overview of the mission key scientific goals.
1. introduction
The interest in space-borne particle/astroparticle physics experiments isgrowing. The achievements of the early space-borne particle detectors suchas IMP [1], HEAO-3 [2], ACE [3] lead to more advanced experiments, namelyEGRET [4] , AMS-01 [5], PAMELA [6], AGILE [7], Fermi [8], AMS-02 [9] andCALET [10]. Additionally, there have been many balloon and ground based ex-periments including BESS [11], IMAX [12], HEAT [13], ATIC [14], CAPRICE[15], CREAM [16], WIZARD [17], Fly’s Eye [18], H.E.S.S [19], MAGIC [20],ARGO-YBJ experiment [21], VERITAS [22], Pierre Auger Observatory [23],HAWC [24] etc. Our understanding of the high-energy universe has been revo-lutionized thanks to the successful operation of these experiments.The DArk Matter Particle Explorer (DAMPE [25]), initially named TAN-SUO [26, 27, 28], was successfully launched into a sun-synchronous orbit atthe altitude of 500 km on 2015 December 17 th from the Jiuquan launch base.DAMPE offers a new opportunity for advancing our knowledge of cosmic rays,dark matter, and gamma-ray astronomy. In this paper a detailed overview ofthe DAMPE instrument is provided, the expected instrumental performancebased on extensive GEANT4 simulations are presented, and the key scientificobjectives are outlined and discussed. 2AMPE is able to detect electrons/positrons, gamma rays, protons, heliumnuclei and other heavy ions in a wide energy range with much improved energyresolution and large acceptance (see Table 1 for summary of the instrument pa-rameters). The primary observing mode is the sky survey in a sun-synchronousorbit at the altitude of 500 km, and it is expected to cover the full sky at leastfour times in two years. The main scientific objectives addressed by DAMPEinclude: (1) understanding the mechanisms of particle acceleration operatingin astrophysical sources, and the propagation of cosmic rays in the the MilkyWay; (2) probing the nature of dark matter; and (3) studying the gamma-rayemission from Galactic and extragalactic sources. Table 1: Summary of the design parameters and expected performance of DAMPE instrument
Parameter ValueEnergy range of γ -rays/electrons 5 GeV −
10 TeVEnergy resolution a of γ -rays/electrons ≤ .
5% at 800 GeVEnergy range of protons/heavy nuclei 50 GeV −
100 TeVEnergy resolution a of protons ≤
40% at 800 GeVEffective area at normal incidence ( γ -rays) 1100 cm at 100 GeVGeometric factor for electrons 0 . sr above 30 GeVPhoton angular resolution b ≤ . ◦ at 100 GeVField of View (FoV) ∼ a σ E /E assuming Gaussian distribution of energies. b The 68% contain-ment radius.
Figure 1: Schematic view of the DAMPE detector. . The DAMPE instrument Fig.1 shows a schematic view of the DAMPE detector. It consists of a PlasticScintillator strip Detector (PSD), a Silicon-Tungsten tracKer-converter (STK),a BGO imaging calorimeter and a NeUtron Detector (NUD). The PSD pro-vides charged-particle background rejection for gamma rays (anti-coincidencedetector) and measures the charge of incident particles; the STK measures thecharges and the trajectories of charged particles, and allows to reconstruct thedirections of incident photons converting into e + e − pairs; the hodoscopic BGOcalorimeter, with a total depth of about 32 radiation lengths, allows to measurethe energy of incident particles with high resolution and to provide efficientelectron/hadron identification; finally, the NUD provides a independent mea-surement and further improvement of the electron/hadron identification. The main purpose of the PSD is to provide charged-particle backgroundrejection for the gamma ray detection and to measure the absolute value ofthe charge (hereafter Z ) of incident high-energy particles in a wide range (i.e., Z ≤ Table 2: Summary of the designed parameters and expected performance of PSD.
Parameter ValueActive area ≥
82 cm ×
82 cmNumber of layers 2Dynamic range Electrons, ions ( Z ≤ a ≤
25% for Z = 1Detector efficiency of single module ≥ .
95 for MIPsPosition resolution b ≤ a σ Z /Z assuming Gaussian distribution. b Geometry size of the PSD bar.A schematic view of the PSD is shown in Fig. 2. The PSD has an active areaof 82 . × . , that is larger than the on-axis cross section of other sub-detectors of DAMPE [30]. The PSD consists of 82 plastic scintillator (EJ-200produced by Eljen [29]) bars arranged in two planes, each with a double layerconfiguration. Each bar is 88.4 cm long with a 2 . × . ≥ .
95, the PSD provides an overall efficiency ≥ . igure 2: Schematic view of the PSD. veto signals due to the “backsplash effect”, which can lead to a misidentificationof gamma rays as charged particles. This phenomenon was observed in EGRETand was found to be significant for photon energies in the GeV region andabove. A similar choice of the segmented design was adopted in the AGILE [7]and the Large Area Telescope onboard the Fermi telescope (Fermi-LAT) [8],both equipped with anti-coincidence detectors consisting of plastic scintillatortiles.Since the PSD is used to identify cosmic-ray nuclei from helium to iron( Z = 26), a wide dynamic range extending up to ∼ is required. To cover sucha broad range with good energy resolution, a double dynode readout schemefor each PMT has been implemented. Signals from the dynode with high gaincover the range from 0.1 MIPs to 40 MIPs, while those from the dynode withlow gain cover the range from 4 MIPs to 1600 MIPs; the overlap region can beused for cross calibration [30, 31, 32].The dynode signals are coupled to VA160 ASIC chip developed by IDEAS [34].This chip integrates the charge sensitive preamplifier, the shaper and the hold-ing circuit for 32 channels. Four groups of front-end electronics (FEE) chips areplaced at all the sides of the PSD, and each FEE processes 82 signal channelsfrom 41 PMTs in each side. With each group of FEE, there is also a high-voltagefan-out board, which supplies the high-voltages to all the 41 PMTs in the same A singly charged MIP at normal incidence, which is assumed as reference, deposits onaverage about 2 MeV in a single PSD bar.
Energy Deposit (MeV)0 2 4 6 8 10 12 14 E n t r i e s ( N o r m a li z ed w i t h T r i gge r C oun t s ) pion 10GeVmuon 150GeVelectron 150GeVproton 8GeV Figure 3: Energy deposited in the PSD as measured on beam tests for different species of Z = 1 particles. In 2014 and 2015, the Engineering Qualification Model (EQM) of DAMPEhas been extensively tested on different particle beams, namely high energygamma-rays (0 . −
150 GeV), electrons (0 . −
250 GeV), protons (3 . − π − (3 −
10 GeV), π + (10 −
100 GeV), muons (150 GeV) and variousnuclei produced by fragmentation of Argon (30 −
75 GeV/n) and Lead (30GeV/n) in the European Organization for Nuclear Research (CERN).Fig. 3 shows the energy deposited in the PSD for different species of chargedparticles with Z = 1. We find that the peaks can be well described by Lan-dau distribution due to the limited number of photons collected by the PMTs.Despite their very different mass and energy, the energy deposits for leptons(electrons, muons) and hadrons (pions, protons) are nearly the same. For asingly charged incident particle, the energy resolution is ∼
10% which can beregarded as the charge resolution of PSD.As mentioned above, in order to effectively separate gamma rays from chargedparticles, the PSD should have a high detection efficiency for Z = 1 particles.Such a performance was checked with electron beams of different energies. Fig. 46 aximum Energy Deposit in Beam Area (MeV)0 1 2 3 4 5 6 7 8 9 E n t r i e s ( N o r m a li z ed ) - - - - Layer X
Electron (20 GeV)
Maximum Energy Deposit in Beam Area (MeV)0 1 2 3 4 5 6 7 8 9 E n t r i e s ( N o r m a li z ed ) - - - - Layer Y
Electron (20 GeV)
Figure 4: Energy deposit for 20 GeV electrons in the PSD modules lying in the beam spotregion (see text). shows the spectra of deposited energy of 20 GeV electron beam in both X and Ylayers. To minimize the influence of the backsplash effect, only modules withinthe beam spot area have been considered. By setting the threshold at 1 MeV,which corresponds to about 0.5 MIP, an efficiency higher than 0.994 has beenachieved for each layer.The performance of the PSD has been also tested with the relativistic heavyion beams at CERN. In this test, the primary Argon beam of 40 GeV/n was sentonto a 40 mm polyethylene target, and the secondary fragments with
A/Z = 2were selected by beam magnets, thus allowing to study the PSD response to allthe stable nuclei with Z = 2 ÷
18. Fig. 5 shows the reconstructed charge spectrafor different ions (
Z >
2) from one PSD module within the beam spot. In thisfigure the Helium contribution has been removed for clarity (the He fraction ismuch higher than that of other ion species). The signals from both sides of eachmodule are used (geometric mean) and the quenching effect has been correctedbased on the ion response from the same test.It can be seen that all the elements from Lithium ( Z = 3) to Argon ( Z = 18)can be identified clearly. By applying a multi-Gaussian fit to the spectrum, weget the charge resolution of PSD for all ion species with the typical value of 0.21for Helium and 0.48 for Argon. The charge resolution is expected to be betterin space, because of much lower ion rates with respect to the case of beam tests.The results show that the position of the Ar peak in the raw Analog-DigitalConversion (ADC) spectrum for different PSD modules is only ∼
20% of thefull dynamic range. By simple extrapolation using the Birks-Chou law [65], thisvalidates that the PSD can cover ion species up to Iron ( Z = 26).7 (Charge Estimation of PSD) C o un t s Li B
C O Ne Mg Si S ArN F Na Al
Figure 5: Reconstructed charge spectra of PSD for nuclei with
A/Z = 2, generated by a 40GeV/n Ar beam. The helium peak has been removed for clarity.
The DAMPE STK is designed to accomplish the following tasks: preciseparticle track reconstruction with a resolution better than 80 µ m for most ofthe incident angles, measurement of the electrical charge of incoming cosmicrays, and photon conversion to electron-positron pairs [66, 67]. The DAMPEtracker-converter system combines the main features of the previous successfulmissions including AGILE [7], Fermi-LAT [8] and AMS-02 [9]. It is composed ofsix position-sensitive double (X and Y) planes of silicon detectors with a totalarea of about 7 m , comparable with the total silicon surface of the AMS-02tracker. Multiple thin tungsten layers have been inserted in the tracker structurein order to enhance the photon conversion rate while keeping negligible multiplescattering of electron/positron pairs (above ∼ RBRadiatorSSD
Figure 6: Exploded view of the STK.Table 3: Summary of designed parameters of STK.
Parameter ValueActive area of silicon detectors 0.55 m ×
12 layersThickness of each silicon layer 320 µ mSilicon strip pitch 121 µ mThickness of tungsten layers 3 × )Spatial resolution a < µ m within 60 ◦ incidencePower consumption 82.7 WTotal mass 154.8 kgNote: a
68% extension range.in Fig. 7. The total strip length along a ladder is about 37 cm. The laddersare glued on the seven support trays to form the 12 STK silicon layers. Eachsilicon layer consists of 16 ladders, as shown in Fig. 6. The two sides of the fivecentral trays are both equipped with 16 ladders each, while for the top and thebottom plane only one side is equipped with the silicon ladders. All the planesare piled up together to form the full tracker system. The silicon ladders on thebottom surface of each tray are placed orthogonal with respect to the ones ofthe top surface of the lower tray, in order to measure the X-Y coordinates ofthe incident particles. The inter-distance between two consecutive silicon layersis ∼ × × . and each SSD is segmented in 768 strips. The strips are 48 µ m wide and93.196 mm long with a pitch of 121 µ m. The bulk resistivity is > · cm with9 ront-endelectronics Sensor Figure 7: The STK single ladder, made by four SSDs. cluster charge [ADC]0 50 100 150 200 250 300 E v en t s ( N o r m a li z ed ) pion 10 GeVelectron 150 GeVproton 400 GeV cluster charge [ADC]0 50 100 150 200 250 300 E v en t s ( N o r m a li z ed ) pion 10 GeVelectron 150 GeVproton 400 GeV Figure 8: (Left)
STK cluster charge response in terms of ADC counts for different singlycharged particles with an incident angle of 0 ◦ (i.e. the particle beam is orthogonal withrespect to the silicon layers). (Right) Cluster charge distributions after the correction. Alldistributions are normalized to unit for shape comparison. C oun t s He Li Be B C N O
Figure 9: STK signal mean distribution for nuclei produced by a lead beam on target, afterremoving Z=1 particles. The signal mean, with current reconstruction procedures, allowsfor the identification of ions until Oxygen. A dedicated Z dependent charge calibration iscurrently being set up (see text). a full depletion voltage of 55 V maximum. The average total leakage currentis of 116 nA at 150 V, well below the specification of 900 nA. The SSDs areglued on the flex part of the Tracker Front-end Hybrid (TFH) board to form aladder, as shown in Fig. 7. The TFH serves as mechanical support for the SSDsand for the collection and amplification of the signals output from the strips.The readout is done one every other strip (corresponding to 384 channels perladder), in order to keep a good performance in terms of spatial resolution, andat the same time reduce the number of readout channels. The signal shaping andamplification is performed by six VA140 ASIC chips (produced by IDEAS [34])mounted on the TFH. The chip design is an updated version of the VA64HDR9Achips used in AMS-02 [70]. Each VA140 chip reads 64 channels.The readout and power supply electronics of the Tracker Readout Boards(TRB) have been mounted on the sides of the trays as shown in Fig. 6. EachTRB module reads 24 ladders and is made of three electronics boards: the powerboard, the control board, and the ADC board. The ladders are connected to theADC board which provides the conversion of the signal from analog to digital,while the voltage to the front-end electronics and the silicon bias voltage aresupplied by the power board. The control board is equipped with two field-programmable gate arrays (FPGAs) which handle not only the communicationwith the DAMPE DAQ system, but also the reduction of the data size, thanksto a zero-suppression and a cluster finding algorithm. More details of the TRBboards can be found in [71, 72]. 11 igure 10: Spatial resolution for different STK planes as a function of particle incident anglefor cosmic rays data at ground. The results obtained from a test beam campaign on singleladder are also shown as reference [67, 73].
As discussed in the previous section, several test beam campaigns of theDAMPE EQM have been conducted at CERN in 2014 and 2015. Moreover, inorder to better characterize the key constituent of the STK, dedicated tests havebeen conducted on single ladder units at the CERN Super Proton Synchrotronfacility (SPS). As in the PSD case, the response of the detector is the samefor different singly charged particles and different energies, as shown in Fig. 8,while it changes in case of particles with higher charge numbers (Z > ◦ , i.e. orthogonallywith respect to the silicon surface, almost all the released charge is collected bya single readout strip (higher charge peak). On the contrary, when the particlehits a floating strip, only about 65% of the original charge is collected by thetwo nearby readout strips, which produces the lower charge peak of the ADCdistribution. This charge collection loss reduces as the incidence angle increases,and it could be recovered with a dedicated correction as function of particleincident angle and impact position (more details can be found in [67, 73]). Theright panel of Fig. 8 shows the cluster charge distribution after such correctionis applied.The ions charge identification power of STK was evaluated with a dedicatedtest conducted on single ladder units at CERN with a lead beam. The particlecharge can be identified by looking at the mean value of the signal associated tothe track. The signal mean S = (cid:112)(cid:80) ( ADC i /M IP/N ) is shown in Fig. 9. In thisformula N corresponds to the number of clusters composing the track, ADC i to the signal charge in the i -th cluster, and M IP to the cluster charge in ADCcounts of a minimum ionizing particle. This value is proportional to the particle12harge and allows a straightforward identification of ions up to Oxygen. Due tothe non-linearity of the VAs above a signal of 200 fC, the identification of ionsabove Oxygen with the STK becomes non-trivial and on-going work is underpreparation to improve the charge identification power. Moreover, in order toequalize the signal collected by each ladder and to make it independent from theincidence angle and the particle hit position on the ladder, a comprehensive andcharge dependent STK signal calibration is in progress. Further improvementof the STK charge resolution is expected in the future.Thanks to a dedicated campaign of extensive cosmic ray data collected onground, the STK detector has been aligned before launch, in order to correctfor displacement and rotation of the SSDs with respect to the nominal position.The alignment procedure will be the subject of a dedicated paper. Here we onlyreport the spatial resolution as a function of incident angle after alignment,shown in Fig. 10. As a result of the alignment, the spatial resolution is below80 µ m within the angular acceptance of the STK (i.e. incidence angle < ◦ ) andbelow 60 µ m for particle incidence angles within 40 ◦ . This result is in agreementwith the spatial resolution measurements obtained in test beam campaigns atCERN SPS on a single ladder [67, 73]. The BGO calorimeter onboard DAMPE has three primary purposes: (1)measuring the energy deposition of incident particles; (2) imaging the 3D pro-file (both longitudinal and transverse) of the shower development, and provideelectron/hadron discrimination; (3) providing the level 0 trigger for the DAMPEdata acquisition system [25, 26, 27, 50, 51, 52, 53, 55]. A summary of the keyparameters of the BGO calorimeter is given in Table 4. Fig. 11 shows the layoutof the BGO calorimeter.
Table 4: Summary of designed parameters and expected performance of the BGO calorimeter.
Parameter ValueActive area 60 cm ×
60 cm (on-axis)Depth (radiation lengths) 32Sampling ≥ (cid:39) ∼ × attenuation factorwith respect to the one on S0. The high gain readout channels (dy8) cover the13 igure 11: Schematic view of the DAMPE BGO calorimeter. ADCVA ChipPMT (S0 end) BGO PMT (S1 end)muonLow gain (Dy2)Medium gain (Dy5)High gain (Dy8) filter_S1Filter_S0
Figure 12: The schematic graph of BGO calorimeter readout. The filter on S1 has a 5 × attenuation factor with respect to the one on S0. range 2 MeV −
500 MeV (S0 end) and 10 MeV − −
20 GeV (S0 end) and 400 MeV −
100 GeV (S1 end); the low gain channels (dy2) cover the range 3.2 GeV − − nergy (GeV)0 5 10 15 20 C oun t s _1-20GeV - (a) eCorEnergyRawEnergy Energy (GeV)50 100 150 200 250 C oun t s _50-243GeV - (b) eCorEnergyRawEnergy Figure 13: The electron energy distribution from beam test data, before and after correction,as measured in the BGO (see text). the corrected energy resolution is 1.85% for 10 GeVelectrons and 0.80 % for 100 GeV electrons.
The ground calibration of BGO has been performed using both the datacollected in a beam test campaign at CERN and cosmic ray data collected fromground. The calibration procedure includes the measurement of the pedestals,the evaluation of the calibration constants from the MIP peaks, the evaluationof the dynode ratios, and the measurement of the bar attenuation lengths. Thefull details of the calibration procedure are provided in Refs. [56, 57]. Fig. 13summarizes the performance of energy reconstruction of the BGO calorimeterfor electrons with different energies up to ∼
250 GeV. The data shown in the fig-ure was obtained during the beam test campaigns performed at CERN. Detailson the energy reconstruction and the electron/proton separation are discussedin Section 3.2.1 and Section 3.2.4. The linearity of reconstructed energy isbetter than 1%, as shown in the Fig. 14. The energy resolution is better than1 .
2% at the energies above 100 GeV (see Fig. 23).
The main purpose of the NUD is to perform electron/hadron identificationusing the neutrons produced in hadronic showers initiated in the BGO calorime-ter. In fact, for a given initial particle energy, the neutron content of a hadronicshower is expected to be one order of magnitude larger than that of an electro-magnetic shower. Once the neutrons are created, they are quickly thermalizedin the BGO calorimeter, and the total neutron activity over a few microsecondsis measured by NUD. Table 5 summarizes the key parameters of the NUD.Fig. 15 shows the detailed structure of NUD. It consists of four 30 cm ×
30 cm × . Babundance of 20% [60]. Each scintillator is wrapped with a layer of aluminumfilm for photon reflection, anchored in aluminum alloy framework by siliconerubber, and readout by a PMT. The space between plastic scintillators and15 (GeV)
Beam E -
10 1 10 ( G e V ) R e c E - Beamtest DataSimulation
Figure 14: Energy reconstructed as a function of the incident energy of electron beam. Redtriangles shows the beam test data, and the open blue circles shows the simulation. aluminum alloy framework is 1 mm on each side, and is filled with siliconerubber to relieve the vibration during the launch.The scintillators are embedded with wavelength shift fibers for optical trans-mission in order to reduce the fluorescence attenuation and increase photoncollection efficiency, and then the signals are readout by corner-on HamamatsuR5610A-01 PMTs. The R5610A-01 is a 0.75 inches diameter head-on, 10-dynodePMT with a maximum gain of 2 × , and a spectral response ranging from 300nm to 650 nm, which is a good match to EJ-254’s 425 nm maximum emissionwavelength.Neutron captures are the dominant source of photon generation in the NUDafter ∼ µ s from the initial calorimeter shower. Neutrons entering the boron-loaded scintillator can in fact undergo the capture process B + n → Li + α + γ with a probability inversely proportional to their speed, and a time constant forcapture inversely proportional to the B loading. About 600 optical photonsare produced in each capture [41].A block diagram of the readout electronics is shown in Fig. 16. There are foursignal channels provided in one data processing board. Each channel containsa fast pre-amplifier, a gating circuit (GC), a shaping circuit (SC) and a mainamplifier with peak holding chip (PHC). The GC and PHC are controlled bythe data control unit of the DAMPE satellite. The GC is designed to preventany early signal entering the SC, and is switched-on 1 . µs after the triggeringsignal produced by BGO. Then the delayed neutron signal could be shaped andamplified to the PHC. After the ADC finishes the acquisition of all four signals,a release signal will be sent to the PHC and GC to shut off the signal channeland wait for the next trigger. 16 igure 15: The structure of NeUtron Detector (NUD).Figure 16: A block diagram of the NUD’s Readout Electronics.Figure 17: NUD signals for protons and electrons with an energy of ∼
150 GeV deposited inthe BGO calorimeter (the distributions are normalized to unit area). able 5: NUD designed parameters. Parameter 4 Plastic Scintillators ( B)Active area 61 cm ×
61 cmEnergy range 2 −
60 MeV for single detectorEnergy resolution a ≤
10% at 30 MeVPower 0.5 WMass 12 kgNote: a σ E /E assuming Gaussian distribution.The electron and proton data collected during the beam test has been used tostudy the particle identification power of the NUD. Since protons deposit in theBGO is about 1 / ≈
150 GeV in the BGO calorimeter).In Fig. 17, the NUD signals of electrons and protons are compared. The electronsignals are always less than 30 channels, and in most cases are below 2 channels,while the proton signals are remarkably larger.The PMTs of the NUD and the bottom BGO layer share the same highvoltage module to save electric power and reduce payload weight. As a result,the NUD works in the high gain mode during on-orbit operation, which gives amore powerful capability for electron-proton identification. Detailed GEANT4simulations suggest a proton rejection power for NUD (in its full performance) ofa factor of ∼
10, assuming an electron detection efficiency of 0.95. Preliminaryestimates, based on on-orbit calibration data, show that a rejection power is ∼ . The data acquisition system (DAQ) receives the commands from the satellitecomputer, implements trigger decision logic, collects science and housekeepingdata from the detectors, and transfers them to the ground. Fig. 18 shows thearchitecture of the DAQ system. The DAMPE DAQ system [54] consists of twoelectronics crates, including the Payload Data Process Unit (PDPU) and thePayload Management Unit (PMU).The DAQ system is implemented with dual modular redundancy. The PMUis the control center of DAMPE and it is equipped with a 16 GB flash mem-ory for data storage. The central processing unit (CPU) board of the PMUreceives commands from satellite computer through 1553B bus (1 Mbps). ThePMU decodes the commands and distributes them to the PDPU or the FEE of − X / − Y sides directly (see Fig. 1). When the PMU receives a trigger signalfrom the trigger board in the PDPU, it begins to collect science data from theFEE on the − X / − Y sides, while data from the FEE on the +X / + Y sidesare collected from the PDPU. All collected data are finally stored in the 16 GBmass memory. The PMU also collects housekeeping data of DAMPE periodi-cally and sends them to the satellite computer. All the science and housekeeping18 AMPE DAQ system (Payload Data Handling Unit)DAMPE Satellite computer(On-board Data Handling Unit)
Control signal(OC) Analog monitor Temperature monitorSlow control and housekeeping(1553B bus)
Mass mem.
Science data(LVDS)
FEEPSD (+X/+Y)
TRBSTK (+X/+Y)
FEEBGO CAL. (+X/+Y)
NUD
Trigger Board
Payload Data process unit (PDPU)
Payload Manag-ement Unit (PMU)
FEEPSD (-X/-Y)
TRBSTK (-X/-Y)
FEEBGO CAL. (-X/-Y)
Science dataTriggerSci. DataTrigger
Event process boardEvent process Board
CPU Board (AT697)
Pulse per-second
NUD Signal process Board
To Ground
Figure 18: The Data Acquisition System (DAQ) of DAMPE. data are finally relayed to ground with the timestamp of 1 ms precision. ThePMU calibrates its timer with the clock of the Global Positioning System (GPS)spacecrafts with one pulse per second.The PDPU is responsible for collecting the science data from FEE of +X / + Ydirection, collecting housekeeping data from FEE, generating global trigger sig-nal for DAMPE and distributing the commands from PMU. The trigger boardof the PDPU receives signals from the BGO calorimeter and makes a triggerdecision within 1 µ s [55]. The trigger is sent to the FEE and to the PMU,while at the same time the PDPU prevents further events to be collected untilall science data is stored, which is collected by the event process board of thePDPU and sent to the PMU.Only the signals from eight out of fourteen BGO layers are sent to thetrigger board. The trigger board implements the trigger decision logic with aflash memory based FPGA chip. Four different triggers have been implemented:Unbiased trigger, MIP trigger, High Energy trigger and Low Energy trigger.They are “OR-ed” to generate the global trigger signal for the detector (seeFig. 19). The Unbiased trigger requires signals in the two top BGO layersexceeding a low threshold of ∼ . ∼ ∼ ± ◦ ). At high latitudes, the MIP trigger is disabled and the pre-scalerratios of Unbiased and Low Energy triggers are set to 2048 : 1 and 64 : 1,respectively. The expected average rate of global triggers is about 70 Hz inflight (the rate of High Energy triggers is 50 Hz, the rate of Unbiased triggersis about 2 . Hits From BGO CAL. CounterExternal Trigger Counter pre-scalerunbias trigger Counter pre-scalerMIPS trigger CounterHigh energy trigger Counter pre-scalerlow energy trigger
Periodical trigger (fixed number)Periodical trigger (contiunous) Trigger information record
Global Trigger
Counter CounterEnable/Disable Enable/DisableEnable/DisableEnable/DisableEnable/DisableEnable/DisableHits From BGO CAL.
Figure 19: The trigger decision logic of DAMPE (see text).
3. Instrument modeling and event reconstruction
A full Monte Carlo (MC) simulation has been developed to accurately eval-uate the detector response to incident particles. The simulation is central bothin the design/optimization phase and in demonstrating the possible achieve-ments in terms of dynamic ranges, resolutions and background rejection power.The simulation procedure mimics the real data taking condition of the instru-ment during both ground tests and in-flight observations, by using proper inputparticle fluxes and fully modeling the detector geometry and readout chain.20 light data
Trigger
Digi Data
BGO digi NUD digiSTK digiPSD digi
Digitization
PSD recoBGO recoSTK recoNUD reco
REC Data
Reconstructionglobalreco
SimulationGenerationOrbit SimuG4 Simu
GDML Geometry&QGSP physics list
IGRF Data
Figure 20: General scheme for DAMPE data handling and simulation.
Fig. 20 shows the flow chart of the data processing for DAMPE, which in-cludes simulation, digitization and reconstruction. The DAMPE simulation isbased on the GEANT4 toolkit [42, 43], a software widely used in high energyphysics experiments to handle particle generation, propagation and interactions.The information on the DAMPE geometry, including the position and the ma-terials of all the detector elements (both active and passive), is stored in gdmland xml files which are used by GEANT4 to build a detailed model. Thewhole simulation procedure is implemented in a GAUDI-like software frame-work [44, 45, 46], which produces collections of energy hits for each sensitivedetector element. A digitization algorithm has been developed to convert en-ergy hits into ADC counts, with the same format as real data, including thecalibration constants (i.e., pedestal noises, PMT gains). In this way, the MCdata can be processed by the same reconstruction algorithms and the simulationcan provide an accurate representation of the instrument response for analysis.Also, for the orbit simulation the same trigger conditions as for real data havebeen implemented to simulate the final data stream.
The first step of the energy reconstruction algorithm is the conversion ofthe ADC counts into energy based on the calibration constants, once pedestalshave been removed, and choosing the signals from the proper readout dynodes(dy8/dy5/dy2). The total deposited energy is then calculated by summing upthe energies of all BGO crystal elements. The typical pedestal width is about 8fC, corresponding to 0.32 MeV (S0) and 1.6 MeV (S1) for dy8, 12.8 MeV (S0)and 64 MeV (S1) for dy5, 512 MeV (S0) and 2560 MeV (S1) for dy2, respectively.21n orbit, cosmic-ray proton MIP events will be selected to calibrate the energyresponse of ADC for each BGO crystal. The resulting ADC distribution of eachindividual BGO crystal will be fitted with a Landau function convolved witha Gaussian distribution. The most probable value (MPV) corresponds to theMPV in energy units taken from the simulation ( ≈ . Thanks to the multi-dynode readout design, the BGO calorimeterenables a measurement of the energy of electrons or gamma rays upto at least 10 TeV without saturation. The measurable energies for asingle bar range from 0.5 MIPs ( ∼ MeV) to MIPs ( ∼ TeV),covering a dynamic range of × . From the simulation we find that,for a 10 TeV electromagnetic shower, typically the maximum energydeposit in one BGO bar does not exceed ∼ TeV, which is within thelinear region of dy2.
The energy deposited in the BGO calorimeter underestimates the true energyof incident particles. Electrons and photons can in fact lose a significant fractionof their energy in the dead materials of calorimeter, such as the carbon fibers andrubber used for the support structure. For incident electron and photon energiesabove hundreds of GeV, the energy leakage should be taken into account. Inaddition, the energy deposited in the STK and in the PSD cannot be neglected,in particular for low-energy incident particles. The true energy of electronsand photons is evaluated by properly modeling the transversal and longitudinaldevelopment of electromagnetic showers in the calorimeter.Two methods are used to calculate the corrected energy starting from aset of reconstructed variables, exploiting their dependence on the depositedenergy. In the first case the correction is performed starting from the ratiobetween the sum of the maximum energies in each layer and the total depositedenergy, which was found to be sensitive to the energy loss in dead material ofBGO calorimeter. In the second case the correction is performed starting fromthe depth of the shower maximum obtained by fitting the longitudinal profilewith the Gamma-distribution, which shows a good correlation with the energyleakage. The correction parameters for different incident energies and differentincidence angles are obtained from the simulations and are checked with beamtest data (see Fig. 13). The details of these procedures can be found in ref. [47].The energy measurements for cosmic-ray protons and nuclei are much morecomplicated than that for electrons or gamma rays, as hadronic showers gen-erally are not fully contained in the BGO. Moreover hadronic showers includean electromagnetic and a hadronic component with large event-by-event fluc-tuations, which brings relatively large uncertainties in the energy deposition.An unfolding algorithm based on the Bayes theorem [48] will be implementedto estimate the primary energy spectra of cosmic-ray nuclei.
DAMPE canmeasure hadronic cosmic rays to an energy of ∼ TeV withoutsignificant saturation. For such high energy events, the maximumdeposit energy in one BGO bar is typically a few TeV, within thelinear region of dy2. We are developing a correction method usingthe adjacent non-saturated bars for a few events which may exceedthe linear region of the readout dynodes. .2.2. Track reconstructionBGO Track reconstruction. Despite its limited spatial resolution, the BGOcalorimeter can also be used for the track reconstruction. The track recon-struction procedure starts by searching for the “clusters” of fired bars in eachBGO layer. A cluster is built starting from the bar with the maximum energydeposit and associating to it all the neighboring bars on both sides with de-creasing energy deposits. The cluster construction is terminated when one ofthe following conditions is met: (1) the side of BGO is reached; (2) a non-firedbar is found; (3) a bar with increasing energy deposit is found. Finally, we makeclusters symmetric about the maximum energy bar.Therefore, if the left (right) tail of the fired bar cluster has more bars thanthe right (left) tail, the bars in excess are removed. We allow one cluster perlayer at most, and then perform a linear fit starting from the positions of thebars in the clusters, and each bar is weighted with the corresponding energydeposit. The fitting result, however, is found to bear some systematic bias forinclined incident particles. To minimize this bias, we rotate the coordinate toalign the X axis with the track direction obtained from the first fit. A second fitis then performed in the new coordinate system, and the final result is obtainedby converting back into the original coordinate system. The direction found bythe BGO track reconstruction (if available) is used as a seed for the STK trackreconstruction. STK Track reconstruction.
The raw data of STK are ADC values as the outputof data reduction algorithm on board of the satellite [49]. Preliminary clus-terization of signal is performed on board of the satellite, where cluster seedsare found from the channels which have a signal-to-noise ratio
S/N > . ADC >
5. A refined hit reconstruction is then performedoffline from ground, as outlined below. The ADC values are grouped into ar-rays of 384 channels per ladder. Channels which did not pass the on-board datareduction are assigned to zero. The offline clustering algorithm looks for seedswhich are defined as local signal maxima with
S/N >
4, and then form thecluster by collecting all the neighboring strips with
S/N > . e + e − pairs, where each peak correspond to its own parti-cle) the cluster reconstruction terminates if a strip signal fulfills the condition S n /N − S n − /N >
5, where S n and S n − are the signals in the current stripin the cluster and the next strip respectively. The hits in X and Y projec-tions in same tracking plane are then combined in all possible ways to formthree-dimensional hits. Since quarter-planes of STK are readout by separateelectronic boards, only X-Y hit combinations coming from the same quarterplane are allowed, thus reducing significantly the number of candidate hits.Track reconstruction is done as follows. The direction found in the BGOis projected onto the closest layer of the STK with the corresponding errormatrix, either infinite, or the one evaluated from the shower position and angularresolution as a function of energy. If the hit is found within a reasonable windowaround the projected position, a seed is formed and the track is reconstructed23sing the Kalman filter. If the resulting track is of insufficient quality (i.e. the χ -test or the number of hits in the track does not fulfill the correspondingthreshold values), the procedure is repeated with other hits in that layer. Ifa track is not found afterward, it is repeated with the hits in the second andthird closest layer to the calorimeter. If a track is found, the whole procedure isrepeated again with the first point of previous track being removed from the listof available points. The same iterations are repeated from beginning until allseed points are exhausted. Finally, the procedure is repeated also with the threefurthermost layers of the tracker in the opposite direction (towards calorimeter).Once a set of tracks is formed, the ghost tracks are eliminated by looping over alltracks and removing those with lover quality crossed by the other tracks. Thetrack forks (two tracks starting from the same point) in the direction towards thecalorimeter are allowed, while those which point toward the opposite direction,are considered as a track crossings and treated correspondingly. The measurement of the energy spectra of cosmic-ray nuclei ( Z = 1 − Z and to the path length. Thefirst step of charge reconstruction is to find the candidate track, which allows tofind the PSD strips crossed by the particle, and to evaluate the path lengths andthe positions in which the tracks intersect the strips. Since each PSD strip isreadout by two PMTs mounted at each end, two signals per strip are obtained.From each signal an energy deposition value is calculated, correcting for thepath length and the position of the track along the strip to account for lightattenuation. Since a track can intersect a maximum of four PSD strips, a totalof eight energy values per event can be used for charge reconstruction, whichare then combined to provide an accurate estimate of Z .The STK, with its 12 layers of silicon strip detectors, can also be used tomeasure the charge of incident particles, starting from the energy depositionpoints for the clusters along the track. The energy deposition for a cluster canbe deduced from the impact point and incidence angle. The impact point can beestimated by the ADC values of the readout strips in the cluster [59]. The chargenumber can be estimated by combining all those measurements. Furthermore,in case of fragmentation of an incoming nucleus due to interaction with materialof the instrument (for example with the tungsten plates), the charge number isexpected to change along the path of the track towards the calorimeter. ThePSD and STK will be combined to further improve the measurement of Z . The measurement of the total spectrum of cosmic ray electrons/positrons is amajor goal of DAMPE. Therefore, besides the track and energy reconstruction, ahigh identification and discrimination power of protons from electron/positrons24s required. The basic approach for electron/proton identification is an image-based pattern recognition method, mainly inherited from the one used in theATIC experiment [61, 62, 63].Since the BGO has a radiation length of 1 .
12 cm and a nuclear interactionlength of 22 . , while keeping at least a 90% electronidentification efficiency. Electrons and protons depositing the same amount ofenergy in the BGO calorimeter can be separated by means of the reconstructed3D images of the showers. An electron/proton rejection power close to 2 × while keeping a 94% electron identification efficiency has been achieved usingBGO only beam test data.For the DAMPE calorimeter, almost all electrons deposit more than 90% oftheir energy into the calorimeter while protons usually just deposit ∼ /
3. Sincethe cosmic ray proton spectrum is approximately proportional E − . , the on-orbit rejection power will be improved of a factor ≈ . ≈ ∼
3. Finally, the NUD can be used to further increase therejection power by a factor of ∼
4. Performance and Operation
The expected instrument performance is summarized in Figs. 21 −
24 for elec-trons/photons, and Figs. 25-26 for protons. These results are based on simula-tions of DAMPE instrument performance from the event reconstruction and se-lection algorithms, which includes trigger filter, track reconstruction, geometryconstraints, charge reconstruction, particle identification and energy reconstruc-tion. The efficiency of each step has been carefully studied with MC simulationsand checked with beam test data and cosmic-ray muon data at ground. Theperformance parameters (in particular for gamma ray detection efficiency) areexpected to improve in the future with improved algorithms, as the event re-construction and selection algorithms will be further optimized after a betterunderstanding of the on-orbit performance.Fig. 21 shows the effective area as a function of energy for gamma ray de-tection at normal incidence and at 30 ◦ off-axis angle, respectively. The adoptedevent selection algorithm for gamma rays is the following. Firstly events withshower well contained in the calorimeter are selected, then a first hadronic back-ground rejection is performed by using information from the BGO only (see25ec. 3.2.4). Candidate electron/gamma-ray events with a track in the STK arethen selected. Finally the PSD is used as an anti-coincidence detector to rejectcharged particle events. The drop of effective area above 100 GeV (shown inFig. 21) is due to the backsplash effect, which has not been taken into accountin the present gamma-rays event selection. The same cuts without the PSDanti-coincidence veto can be used to select electrons/positrons. Starting fromevents with the High Energy trigger, the resulting acceptance for electrons islarger than 0 . sr above 50 GeV, as shown in Fig. 22. The energy resolutionfor electromagnetic showers is shown in Fig. 23. The angular resolution (i.e. thecorresponding 68% containment angle) for gamma rays converted in the STKis shown in Fig. 24 for normal and 30 ◦ incidence angles, respectively. Energy [GeV]1 10 ] E ff e c t i v e A r ea [ c m normal incidence30 deg incidence Figure 21: Effective area as a function of energy for gamma rays at normal incidence (solidline) and at 30 ◦ off-axis angle (dashed line). For hadronic cosmic rays, the acceptance is about . sr for en-ergies above ∼
100 GeV , which varies for different nuclei species dueto different trigger efficiency. The energy measurement of cosmic raynuclei is more complicated than that of electrons/photons, becauseof the energy leakage due to limited nuclear interaction thickness ofthe calorimeter ( ∼ . nuclear interaction length) and fluctuation ofthe hardonic shower development. To convert the measured energyspectrum to the primary energy spectrum, an unfolding algorithmcould be used to reconstruct the nucleus energy spectrum, by us-ing the MC detector response matrix. Figure 25 shows the deposit(blue) and reconstructed (red) energy distributions for on-axis inci-dent proton beams with momenta of 5, 10, 150, and 400 GeV /c . Thereconstructed procedure allows to recovery the incident beam energy Energy [GeV]0500100015002000250030003500 A cc e p t a n c e [ c m s r ] Figure 22: Acceptance for electrons/positrons as a function of energy. Energy [GeV]0.000.010.020.030.040.050.06 E n e r g y r e s o l u t i o n [ % c o n t a i n m e n t ] normal incidence30 deg incidencenormal beam test Figure 23: Energy resolution for gamma rays and electrons/positrons at normal incidence(solid line) and at 30 ◦ off-axis angle (dashed line). DAMPE beam test results (with electrons)are over-plotted as reported in Fig. 13. as well.The energy resolution ( σ E /E ) of on-axis incident protons (after igure 24: Angular resolution at 68% containment angle for gamma rays at normal incidence(solid curve) and at 30 ◦ off-axis angle (dashed curve). the unfolding), estimated from the simulation data, is shown by thedotted line in Fig. 26. As a comparison, the results for the beam testdata at four energies are overplotted. It is shown that the energyresolution for protons varies from ∼ at several GeV to ∼ at TeV. Above 10 TeV, the uncertainties on the hadronic interactionmodel as implemented in Geant4 are expected to be non-negligible.While a detailed treatment is currently being undertaken in the col-laboration, we expect these uncertainties to yield uncertainties inthe reconstructed spectrum of about 10%. We are also investigatingthe use of alternative simulation packages that incorporate hadronicinteractions at the TeV scale better (e.g. Fluka ). As discussed above, the verification of the estimated performance was carriedout using the data from the beam test campaign, as well as a set of data collectedwith cosmic-ray muons at sea level. In particular, several cosmic-ray muon testswere performed during different stages of the DAMPE assembly, especially in theenvironmental testing phase and in the pre-launch preparation of the satellite.In these tests, a proper trigger logic was adopted to select cosmic-ray muons.We were able to collect a large amount of muon events, which has been usedto perform a full calibration of the energy response for MIPs and to implementthe alignment procedure for the STK. After launch, the spacecraft entered the nergy (GeV)10 C oun t s (a) proton_5GVEnergy MeasuredEnergy Unfolded Energy (GeV)10 C oun t s (b) proton_10GVEnergy MeasuredEnergy Unfolded Energy (GeV) C oun t s (c) proton_150GVEnergy MeasuredEnergy Unfolded Energy (GeV) C oun t s (d) proton_400GVEnergy MeasuredEnergy Unfolded Figure 25: Distributions of deposited energies (blue) and unfolded ones (red) for beam testprotons at incident momenta of 5, 10, 150, and 400 GeV/ c . sky-survey mode immediately, and a dedicated calibration of the detector wasperformed in the first 15 days, including pedestals, MIP responses (protons),alignments, and timing etc. Comparison between on-orbit data with simulationsand ground cosmic-ray data demonstrates the excellent working condition ofDAMPE detectors. Details of the on-orbit calibration and performanceevaluation will be published elsewhere [64].
Since December 17 th eam Energy (GeV)1 10 E ne r g y R e s o l u t i on ( % ) MonteCarlo ProtonBeamTest Proton
Figure 26: The energy resolution for on-axis protons. The dotted line represents the energyresolution of MC simulated protons after spectral unfolding while the red points represent thebeam test data.
On ground the data are processed by the Ground Support System (GSS) andthe Scientific Application System (SAS). Binary raw data (housekeeping andscience data) transmitted to ground are first received by three ground stationslocated in the south, west and north of China at early morning and afternoonof each day respectively, when the satellite passes China’s borderline. Then allbinary data are automatically transmitted to the GSS located in Beijing, and aretagged as level-0 data. On average, about 12 GB level-0 data are produced perday. Upon arrival of the level-0 data at the GSS, they are immediately processedand several operations are performed, including data merging, overlap skippingand cyclic redundancy check (CRC) which is an error-detecting code based onthe protocol CRC-16/CCITT.The level-0 data are daily processed into level-1 data, which includes 13 kindsof completed telemetry source packages, one for science data and 12 for house-keeping data. Daily level-1 data will then be processed by the GSS within 1 hour.The SAS located at the Purple Mountain Observatory of Chinese Academy ofSciences in Nanjing monitors the level-1 data production 24 hours a day con-tinuously. The new level-1 data will be synchronized to the mass storage at thePurple Mountain Observatory immediately. Then 12 housekeeping data pack-30ges are parsed and inserted into the housekeeping database, which allows tomonitor the conditions of the DAMPE payload and the satellite platform. Afterprocessing the housekeeping data, routine checks on key engineering parametersare performed to guarantee the proper data taking conditions.The processing pipeline of science data includes the Raw Data Conversion,Pre-Reconstruction and Reconstruction algorithms implemented in the DAMPEsoftware framework (DAMPESW). The Raw Data Conversion algorithm splitsraw science data into about 30 calibration files and 30 observation files, and con-verts them into ROOT data files [58]. During this procedure, key housekeepingdata required by science analysis are also stored into in the ROOT data files.Calibration files are used to extract calibration constants which are used in Pre-Reconstruction and Reconstruction algorithms. Reconstructed data from allsub-detectors are then merged to generate level-2 science data products. Thesetwo procedures increases the raw science data volume by approximately a factorof five.The processing pipeline of science data is designed to run on a cluster ofbatch processors. The SAS hosts more than 1400 computing cores at the PurpleMountain Observatory, which can reprocess three years of DAMPE data withintwo weeks. In addition, INFN and University of Geneva computing resourcesare also used, which are mainly dedicated to MC data production and couldalso be used as backup reprocessing sites if needed.
5. Key scientific objectives
DAMPE is a high energy cosmic-ray and gamma-ray observatory with abroad range of scientific objectives. The data sets provided by DAMPE couldbe used to study cosmic-ray physics, to probe the nature of dark matter, and toreveal the nature of high energy gamma-ray phenomena. The large field of viewof DAMPE provides the opportunity to monitor the violent GeV-TeV transientsfor various purposes.
Cosmic rays impinging the Earth with energies below ∼ eV are be-lieved to be mainly produced through energetic astrophysical processes withinthe Milky Way. Their interactions with interstellar medium, interstellar radi-ation fields, and Galactic magnetic fields are the main source of the detectedGalactic diffuse gamma-ray emissions. Moreover, cosmic rays are the only sam-ple of matter originated from distant regions of the Galaxy that can be directlymeasured with spaceborne experiments. Therefore, understanding the origin,acceleration, and propagation of cosmic rays is a crucial subject on the under-standing of the Universe.With more than three years of operation, DAMPE will be able to observeelectrons/positrons or photons from GeV to 10TeV, and protons, helium orheavier nuclei from 10 GeV to 100 TeV. The measurement of energy spectra withunprecedented precision and energy coverage at higher energies, together with31patial distribution of these particles are expected to significantly enhance ourunderstanding of the origin of cosmic rays. Below we outline the key scientificoutputs regarding cosmic-ray studies potentially achievable with DAMPE. • The proton and helium are the most abundant components of cosmicrays. The standard paradigm for particle acceleration and propagationpredicts single power-law spectra up to the so-called “knee” at ∼ eV. Surprisingly, the spectra of cosmic-ray nuclei measured by ATIC [77],CREAM [78], PAMELA [79] and AMS-02 [80, 145] all showed remarkablehardening at the magnetic rigidity of several hundred GV. Such a resulttriggered various modifications of the standard, simple picture of Galacticcosmic rays. Interesting possibilities include the superposition of injectionspectra of the ensemble of sources [81, 82], the effect of local source(s)[83, 84, 85], the complicated acceleration of particles [86, 89], or a non-uniform diffusion coefficient [87, 88]. Current spectral measurements are,however, uncertain for energies above TeV/n. DAMPE will be able toclearly measure the spectral changes and precisely determine the highenergy spectral indices of various nuclei species, as shown in Fig. 27 forproton and helium. Furthermore, the DAMPE data will be able to testwhether there are additional structures on the high energy cosmic-rayspectra, as may be expected from nearby sources [104]. Recently, theproton and helium spectra from the CREAM-III flight have been publishedand tentative breaks at ∼ −
20 TeV are displayed [105]. With theenergy resolution of ∼
20% at such energies (see Fig.26; which is betterthan that of CREAM-III) and an expected exposure of ∼ . sr yr,DAMPE will reliably test such a possibility. The cosmic-ray spectra upto 100 TeV by DAMPE will overlap with that measured by the ground-based air shower experiments (e.g. [106]), which can provide us with afull picture of the cosmic-ray spectra up to above the knee. DAMPEcan also measure the Boron-to-Carbon ratio, to about 5 TeV/n,which can effectively constrain the propagation parameters. • Electrons/positrons contribute ∼
1% of the total amount of cosmic rays.Unlike the nuclei, electrons/positrons lose their energies efficiently duringthe propagation in the Galaxy. This is particularly true for ∼ TeV elec-trons/positrons which are expected to reach the Earth only if the sourceis relatively nearby ( (cid:46) (cid:46) yr) [90, 91]. With anacceptance of ∼ . sr at TeV energies (see Fig.22), DAMPE willprecisely measure the trans-TeV behavior of the energy spectra of elec-trons/positrons, and determine the spectral structures e.g. spectral cut-off[94, 95, 96, 97, 98, 101, 102, 103]. As a consequence, DAMPE will be ableto directly test a long-standing hypothesis that nearby pulsars or SNRs(e.g., Vela) are efficient TeV electron accelerators [92, 93] by measuringthe spectrum and/or the spatial anisotropy of TeV electrons/positrons(see Fig.28 for an illustration). • DAMPE can also measure gamma-rays from Galactic and extra-galactic32osmic ray accelerators such as SNRs, pulsars, quasars [107] etc. Althoughthe effective acceptance of DAMPE is smaller than that of Fermi-LAT,DAMPE may play an auxiliary role in deep observations of these sources,especially in connection with ground-based measurements at hundreds ofGeV.
Log(E(GeV/n))1 1.5 2 2.5 3 3.5 4 ) - s r - s - m . J ( E ) (( G e V / n ) · . E sr yr) DAMPE (exposure - 0.3 mPamelaCREAMAMS-02ATIC protonhelium
Log(E(GeV/n))1 1.5 2 2.5 3 3.5 4 ) - s r - s - m . J ( E ) (( G e V / n ) · . E sr yr) DAMPE (exposure - 0.3 mPamelaCREAMAMS-02ATIC p helium Figure 27: Expected spectra of protons (left) and helium (right) that can be obtained byDAMPE, assuming the AMS-02 fluxes and their extrapolations, with an exposure of 0.3 m sr yr, compared with current measurements [77, 78, 79, 80, 145]. Log(E(GeV))0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5 s s r) / m J ( E )( G e V · E DAMPE - 3 yearsAMS-02 (2014)ATIC (2008)FERMI-LAT (2010)FERMI-LAT (2017)HESS (2009)VERITAS (2016)PAMELA (2011) - electrons onlyT.Kobayashi et al. 2004 ApJ 601:340-351
Monogem CygnusLoop Vela
Figure 28: Current measurement [94, 98, 96, 99, 100] and the expected spectrum of cosmicray electrons (and positrons) for three years operation of DAMPE, assuming the AMS-02intensity, a cut-off and the contribution of Vela as calculated in [146]. Note that some nearbyyoung/middle-aged supernova remnants may give rise to additional TeV bump(s) in the spec-trum.
As early as the 1930s, it was recognized that some matter in the Universeis invisible [108]. The existence of this so-called dark matter was graduallyand firmly established since the early 1970s [109]. In the standard model of33osmology, the ordinary matter, dark matter and dark energy share 4.9%, 26.6%and 68.5% of today’s total energy density of the Universe. Compelling evidenceshows that the commonly existing dark matter is non-baryonic; however, thephysical nature of the dark matter particle is still poorly known [110, 111]. Manytheoretical models have been proposed, and the suggested candidates span over awide range of masses, mechanisms, and interaction strengths [110, 111]. Amongvarious candidates of dark matter particles, one of the most attractive modelsis the weakly interacting massive particle (WIMP), which is widely predictedin extensions of the standard model of particle physics. The annihilation ordecay of WIMPs can give electromagnetic signals, primarily in the gamma-rayband, as well as standard model particle products such as electrons/positrons,neutrinos/anti-neutrinos and protons/antiprotons [110, 111, 112, 113].Anomalous peaks or structures in the energy spectra of cosmic rays (inparticular for electrons/positrons and antiprotons) and/or gamma rays fromparticular directions with accumulated dark matter distribution could indicatethe existence of dark matter particles. In the past few years, several anomalousexcesses had been reported in different cosmic-ray and gamma-ray data sets,including the electron/positron excesses [94, 95, 96, 97, 98, 101, 102, 103], theGalactic center GeV excess [114, 115, 116, 117, 118], the possible excesses in afew dwarf galaxies [119, 120], and the tentative ∼
130 GeV gamma-ray line [121,122]. Recently, another line-like structure around 43 GeV from a number ofgalaxy clusters was reported with the Fermi-LAT Pass 8 data [123]. Thesecandidate signals are either too weak to be claimed as a firm detection, or canbe interpreted with astrophysical models or potential instrument systematics(e.g., [124, 125, 126, 113]).With its much improved energy resolution (see Fig. 23), DAMPE is suitablefor the search of gamma-ray line emission which can be expected in the annihi-lation channel of γX , where X = ( γ, Z , H ) or other new neutral particle. Theenergies of the monochromatic gamma-rays are given by E γ = m χ [1 − m X / m χ ],where m χ is the mass of dark matter particle [121]. The firm detection ofgamma-ray line(s) is believed to be a smoking-gun signature of new physics,because no known astrophysical process is expected to be able to produce suchspectral feature(s). The high resolution is also crucial to identify multiple lineswith energies close to each other [128, 127]. A set of gamma-ray lines wouldfurther provide convincing evidence of dark matter particles, and could pro-vide more information of physical properties of dark matter particle, such astheir couplings with standard model particles. Theoretically the line emissionis typically suppressed due to particle interactions through loop process, otherscenarios e.g. the internal bremsstrahlung from dark matter annihilating into apair of charged particles, might dominate the potential line signal [129]. Axionsor axion-like particles (ALPs), if produced non-thermally, could be candidate ofcold dark matter [147, 148, 149], which produce spectral fine structures due tothe photon-ALP oscillation [150, 151, 152]. DAMPE will enhance our capabilityto search for monochromatic and/or sharp spectral structures of gamma-rays inthe GeV-TeV range.For illustration purpose, we take into account two types of dark matter den-34ity profile, including a contracted Navarro-Frenk-White profile with γ = 1 . α = 0 .
17 [131]. Following [122], theRegions of Interest (ROIs) have been taken as a 3 ◦ (16 ◦ ) circle centered on theGalactic center, respectively. For the Einasto profile, we also mask the galacticplane region with | l | > ◦ and | b | < ◦ . The Galactic diffuse emission model gll iem v06.fits [132], the isotropic diffuse model iso P8R2 SOURCE V6 v06.txt [133],and 3FGL point sources of Fermi [134] have been combined to model the gamma-ray background. The projected sensitivities of DAMPE in 3 years in case oftargeted observations towards the Galactic center and in 5 years of sky-surveyobservations are presented in Fig. 29. m (GeV) -30 -29 -28 -27 -26 -25 < v > % C LL i m i t ( c m s - ) NFWc (5yr survey) Einasto (5yr survey) NFWc (3yr pointing) Einasto (3yr pointing)
Figure 29: Expected sensitivity of the gamma-ray line search by DAMPE in 3 years of targetedobservations of the Galactic center, and in 5 years of sky-survey observations.
The electron/positron spectra can also be used to probe dark matter, al-though the discrimination from local astrophysical sources may not be trivial.In general the contribution from astrophysical sources is expected to be non-universal, and may induce multiple features on the total energy spectra [135].DAMPE will accurately measure the energy spectra of electrons/positrons attrans-TeV energies to resolve these potential fine structures, which can be usedto test/constrain dark matter models, in order to consistently explain the elec-tron/positron excesses. With DAMPE’s much improved energy resolution, pos-sible new and fine spectral structures on the electron/positron spectra may berevealed as the existence of dark matter particles [136].35 .3. Studying high energy behaviors of gamma-ray transients and the diffuseemission
DAMPE observes gamma-ray photons in the energy range of 10 GeV to TeVand above with very high energy resolution. Note that with the low energy trig-ger system the threshold can be as low as ∼ . ∼
64. Compared with Fermi-LAT, DAMPE has a smallereffective area and a higher energy threshold. Therefore for stable GeV sources,DAMPE is not expected to be competitive compared with Fermi-LAT due tolimited counting statistics. However, DAMPE may play a complementary, andpossibly crucial role in catching bright GeV-TeV transients, as each gamma-raydetector can only cover part of the sky at the same time.The collected gamma-ray data can be used to study the violent physical pro-cesses behind activities of Active Galactic Nuclei (for instance Mrk 421, 3C279and 3C454.3), the Crab flares, and some bright gamma-ray bursts (GRBs) suchas GRB 130427A [137, 138]. Bright short GRBs with an isotropic GeV γ -rayenergy release of ≥ × erg, if taking place within ∼
400 Mpc, mightbe detectable by DAMPE and could also serve as the electromagnetic coun-terparts [139, 140] of advanced LIGO/Virgo gravitational wave events [141] orIceCube PeV neutrino events [142]. High energy gamma-ray observations canalso be used to probe the extragalactic background light, the intergalactic mag-netic field, and the fundamental physics such as Lorentz invariance violationand quantum gravity.The hadronic interaction of cosmic ray nuclei with the interstellar mediumproduces bright diffuse gamma-ray emission, primarily along the Galactic plane.In some regions (e.g. the Galactic center ridge and the Cygnus region) of theGalactic plane, fresh cosmic ray accelerators may light up surrounding materialswith gamma-ray emission on top of the diffuse background [143, 144]. Thanksto the improved hadron rejection power of DAMPE ( > ), it is possible tomeasure the diffuse gamma-ray emission up to TeV energies without significantcontamination from cosmic rays. DAMPE has the potential to reliably detect >
6. Summary
DAMPE was successfully launched into a sun-synchronous orbit at the alti-tude of 500 km on December 17 th Acknowledgments:
The DAMPE mission was funded by the strategic pri-ority science and technology projects in space science of the Chinese Academyof Sciences (No. XDA04040000 and No. XDA04040400). In China this work36s also supported in part by National Key Research and Development Programof China (No. 2016YFA0400200), the National Basic Research Program (No.2013CB837000), National Natural Science Foundation of China under grantsNo. 11525313 (i.e., Funds for Distinguished Young Scholars), No. 11622327(i.e., Funds for Excellent Young Scholars), No. 11273070, No. 11303096, No.11303105, No. 11303106, No. 11303107, No. 11673075, U1531126, U1631111and the 100 Talents program of Chinese Academy of Sciences. In Europe thework is supported by the Swiss National Science Science Foundation, the Uni-versity of Geneva, the Italian National Institute for Nuclear Physics, and theItalian University and Research Ministry. We also would like to take this op-portunity to thank the scientific laboratories and test facilities in China andEurope (in particular CERN for provision of accelerator beams) that assistedthe DAMPE team during the qualification phases.
References [1] M. Garcia-Munoz, G. Mason and J. Simpson,
Astrophys. J. (1975) 265M. Garcia-Munoz, G. Mason and J. Simpson,
Astrophys. J. (1977) 859.[2] W. R. Binns et al.,
Astrophys. J. (1988) 1106J. Engelmann et al.,
Astron. Astrophys. (1990) 96.[3] A. Labrador et al.,
Proc. 28th Int. Cosmic Ray Conference (Tsukuba, Japan,2003), p. 1773.[4] R. C. Hartman et al.,
Astrophys. J. Suppl. (1999) 79.[5] M. Aguilar et al.,
Phys. Rep. (2002) 331.[6] W. Menn et al.
Advances in Space Research , (2013) 209-218.[7] M. Tavani et al. Nucl. Instr. Meth. Phys. Res. A , (2008) 52.[8] B. Atwood et al., Asptrophys. J. (2009) 1071.[9] A. Kounine,
Int. J. Mod. Phys. (2012) 1230005.[10] S. Torii, et al. J. Japan Soc. Microgravity Applications , (2007) 120.[11] A. Yamamoto et al., Adv. Space Res. (1994) 75.[12] W. Menn et al., Asptrophys. J. (2000) 281.[13] S. W. Barwick et al.,
Asptrophys. J. (1998) 779; S. Coutu et al.,
As-tropart. Phys. (1999) 429.[14] T. G. Guzik et al., Adv. Space Res. (2004) 1763.[15] P. Panini et al., Asptrophys. J. (2004) 259.37 able 6: Abbreviations in the main text.
Abbreviation Full expression3FGL 3rd Fermi Gamma-ray LAT catalogACE Advanced Composition ExplorerADC Analog-Digital ConversionAGILE Astrorivelatore Gamma a Immagini LEggeroALP Axion-Like ParticleAMS Alpha Magnetic SpectrometerASDC ASI Science Data CenterASIC Application Specific Integrated CircuitATIC Advanced Thin Ionization CalorimeterBESS Balloon-borne Experiment with Superconducting SpectrometerBGO Bismuth Germanium OxideCALET CALorimetric Electron TelescopeCAPRICE Cosmic AntiParticle Ring Imaging Cherenkov ExperimentCERN European Organization for Nuclear ResearchCFRP Carbon Fiber Reinforced PlasticsCPU Central Processing UnitCRC Cyclic Redundancy CheckCREAM Cosmic Ray Energetics And MassDAMPE DArk Matter Particle ExplorerDAMPESW DAMPE Software FrameworkDAQ Data AcQuisitionEGRET Energetic Gamma Ray Experiment TelescopeEJ EljenEQM Engineering Qualification ModelFEE Front-End ElectronicsFermi-LAT Fermi Large Area TelescopeFoV Field of ViewFPGA Field-Programmable Gate ArrayGC Gating CircuitGEANT GEometry ANd TrackingGPS Global Positioning SystemGRB Gamma-Ray BurstGSS Ground Support SystemH.E.S.S. High Energy Stereoscopic SystemHAWC High-Altitude Water CherenkovHEAO High Energy Astronomy ObservatoryHEAT High Energy Antimatter TelescopeIMAX Isotope Matter Antimatter eXperimentIMP Interplanetary Monitoring PlatformINFN Istituto Nazionale di Fisica NucleareLIGO Laser Interferometer Gravitational wave Observatory38able 6 — continued.Abbreviation Full expressionMAGIC Major Atmospheric Gamma Imaging Cherenkov TelescopesMC Monte CarloMIP Minimum Ionizing ParticleMPV Most Probable ValueNFW Navarro-Frenk-WhiteNUD NeUtron DetectorPAMELA Payload for Antimatter Matter Exploration and Light-nuclei AstrophysicsPAO Pierre Auger ObservatoryPDPU Payload Data Process UnitPHC Peak Holding ChipPMT PhotoMultiplier TubePMU Payload Management UnitPSD Plastic Scintillator strip DetectorROI Region of InterestSAS Scientific Application SystemSC Shaping CircuitSPS Super Proton SynchrotronSSD Silicon micro-Strip DetectorSTK Silicon-Tungsten tracKer-converterTFH Tracker Front-end HybridTMVA Toolkit for MultiVariate data AnalysisTRB Tracker Readout BoardVERITAS Very Energetic Radiation Imaging Telescope Array SystemWIMP Weakly Interacting Massive Particle[16] E. S. Seo, et al. (The CREAM Collaboration),
Adv. Spac. Res. (2004)1777-1785.[17] G. Grimani et al., Asptrophys. J. (2002) 287.[18] R. M. Baltrusaitis et al.,
Nucl. Instrum. Meth. A (1985) 410.[19] J. A. Hinton (H.E.S.S. Collaboration),
New Astron. Rev. , 331 (2004).[20] Magic collaboration, et al. Nuclear Physics B - Proceedings Supplements , (2003) 247-252.[21] B. Bartoli, et al. [ARGO-YBJ Collaboration], Phys. Rev. D , , 112017(2015).[22] T. Weekes, et al. (VERITAS collaboration). Astroparticle Physics , (2002) 221.[23] J. Abraham et al., Nucl. Instrum. Meth. A , (2004) 50.[24] A. J. Smith (for the HAWC Collaboration), Journal of Physics: ConferenceSeries , (2007) 131. 3925] J. Chang, Chin. J. Spac. Sci. , Proceedings of 31st International Cosmic Ray Con-ference , Vol.1 (2009) 1613; http://icrc2009.uni.lodz.pl/proc/pdf/icrc0823.pdf [27] J. Wu & J. Chang,
Proceedings of 32nd International Cosmic Ray Confer-ence , Vol.5 (2011) 228.[28] J. Chang & Y. Z. Fan.
Bull. Chin. Acad. Sci. , (2011) 161.[29] [30] Y. H. Yu, et al., Astropartic. Phys. , in press, arXiv:1703.00098 (2017).[31] Y. Zhou, Z. Y. Sun, Y. H. Yu, et al.
Nuclear Instruments & Methods inPhysics Research A (2016) 79.[32] Y. Zhou, Z. Y. Sun, Y. H. Yu, et al.
Nucl. Sci. Tech. (2016) 70.[33] H. Bethe, J. Ashkin, Experimental Nuclear Physics, ed. E. Segr´e, J. Wiley,New York , (1953).[34] Integrated Detector Electronics AS, [35] P. Azzarello, et al., Nuclear Instruments & Methods in Physics ResearchA , (2016) 378.[36] X. Wu et al., Proceedings of the 34th International Cosmic Ray Conference,1192 (2015)[37] V. Gallo et al, Proceedings of the 34th International Cosmic Ray Confer-ence, 1199 (2015)[38] Hamamatsu Photonics, [39] B. Alpat, et al., Nuclear Instruments & Methods in Physics Research A(2005), A 540 (2005) 121-130[40] Gamma Medica-Ideas, VA140 Document datasheet (V0R1) (2011)[41] D. M. Drake et al., Nucl. Instr. & Meth. A , 576-582 (1986)[42] S. Agostinelli, et al., Nucl. Instrum. Methods A , , 250 (2003)[43] J. Allison, et al., IEEE Trans. Nucl. Sci. , , 270 (2006)[44] G. Barrand, et al., Computer Physics Communications , , (2001) 45-55[45] A. Tykhonov, et al., Proceedings of the 34th International Cosmic RayConference , 1193 (2015) 4046] C. Wang et al. , arXiv:1604.03219 [astro-ph.IM].[47] C. Yue, et al.
Nucl. Instrum. Methods A , (2017) 11.[48] G. D’Agostini, Nuclear Inst. and Methods in Physics Research A , (1995) 487.[49] F. Zhang, et al., Chin. Phys. C. , (2016) 116101.[50] J. H. Guo, M. S. Cai, Y. M. Hu, & J. Chang, Chin. Astron. Astrphys. , (2012) 318.[51] Y. L. Zhang, B. Li, & C. Q. Feng, et al. Chin. Phys. C , (2012) 71.[52] C. Q. Feng, D. L. Zhang, & S. S. Guo, et al. IEEE Trans. Nuc. Sci. ,(2015), 31175; arXiv:1406.3886.[53] Z. Y. Zhang, Y. L. Zhang, & J. Dong et al. Nuclear Inst. and Methods inPhysics Research A , (2015) 21.[54] J. H. Guo, Z. L. Xu, M. S. Cai, Y. M. Hu & J. Chang, Proceedings of the32nd International Cosmic Ray Conference , Vol.5 (2011) 252.[55] L. Zhang, J. H. Guo, & Y. Q. Zhang,
Chin. Astron. Astrphys. , (2015)380.[56] Z. Y. Zhang, C. Wang, & J. Dong et al. Nuclear Inst. and Methods inPhysics Research A , (2016) 98.[57] Z. Y. Li, Z. Y. Zhang, & Y. F. Wei et al. Chin. Phys. C , (2016) 086202.[58] https://root.cern.ch/ [59] J. Alcaraz, B. Alpat, G. Ambrosi, et al. Nuclear Instruments and Methodsin Physics Research A , (2008) 376-398.[60] M. He, T. Ma, & J. Chang, et al. Acta Astron. Sini. , (2016) 1[61] J. Chang, et al. Proceedings of the 26th International Cosmic Ray Confer-ence, Salt Lake City, Edited by D. Kieda, M. Salamon, and B. Dingus,
Vol.5 (1999) 37.[62] W. K. H. Schmidt, J. Chang, et al. (The ATIC Collaboration),
Proceedingsof 26th International Cosmic Ray Conference, Salt Lake City, Edited by D.Kieda, M. Salamon, and B. Dingus,
Vol.5 (1999) 41.[63] J. Chang, et al.,
Advances in Space Research , , 431 (2008).[64] G. Ambrosi, et al. [The DAMPE collaboration], Astropart. Phys. , to besubmitted (2017).[65] R. Dwyer, D. Z. Zhou,
Nuclear Inst. and Methods in Physics Research A , (1985) 171. 4166] X. Wu et al., Proceedings of the 34th International Cosmic Ray Conference(The Hague, Holland, 2015) PoS(ICRC2015)1192[67] P. Azzarello, et al., Nuclear Instruments and Methods in Physics ResearchA, 831 (2016) 378.[68] [69] G. Barbiellini et al, Nuclear Instruments and Methods in Physics ResearchA 490 (2002) 146-158[70] G. Ambrosi et al., Nuclear Instruments and Methods in Physics ResearchA 435 (1999) 215[71] F. Zhang et al., arXiv:1606.05080[72] D. Y. Fan et al., Chinese Physics C Vol. 39, No. 11 (2015) 116202[73] V. Gallo et al.,Proceedings of the 34th International Cosmic Ray Confer-ence (The Hague, Holland, 2015) PoS(ICRC2015)1199.[74] A. Giuliani, et al. Astrophys. J. , (2011) L30[75] M. Ackermann, et al. (Fermi-LAT collaboration). Science , (2013) 807[76] A. Abramowski et al. (H.E.S.S. collaboration). Nature , (2016) 476[77] J. P. Wefel, et al., ATIC Collaboration, Int. Cosmic Ray Conf. (2008)31.[78] H. S. Ahn, etal., CREAM Collaboration, Astrophys. J. (2010) L89.[79] O. Adriani, et al., PAMELA Collaboration,
Science (2011) 69.[80] L. Accardo, et al. (AMS Collaboration).
Phys. Rev. Lett. (2015)171103.[81] V. I. Zatsepin, N. V. Sokolskaya,
Astrophys. Astron. (2006) 1.[82] Q. Yuan, B. Zhang, X. J. Bi,
Phys. Rev. D (2011) 043002.[83] A. E. Vladimirov, G. J´ohannesson, I. V. Moskalenko, T. A. Porter, Astro-phys. J. (2012) 68.[84] S. Thoudam, J. Hoerandel,
Mon. Not. Roy. Astron. Soc. (2012) 1209.[85] K. Fang , B. B. Wang , X. J. Bi , S. J. Lin , P. F. Yin, Astrophys. J. (2017) 172; K. Fang, X. J. Bi, P. F. Yin, (2017), arXiv:1706.03745.[86] V. Ptuskin, V. Zirakashvili, E. S. Seo,
Astrophys. J. (2013) 47.[87] N. Tomassetti,
Astrophys. J. (2012) L13.4288] Y. Q. Guo, Z. Tian, C. Jin,
Astrophys. J. (2016) 54.[89] Y. Ohira, N. Kawanaka, K. Ioka,
Phys. Rev. D (2016) 083001.[90] Y. Z. Fan, B. Zhang, J. Chang, Int. J. Mod. Phys. D (2010) 2011.[91] M. Ackermann et al. [Fermi-LAT Collaboration], Phys. Rev. D , 092003(2010) doi:10.1103/PhysRevD.82.092003 [arXiv:1008.5119 [astro-ph.HE]].[92] C. S. Shen, Astrophys. J. (1970) L181.[93] X. Li, et al.
Phys. Lett. B (2015) 267.[94] J. Chang, et al., The ATIC Collaboration,
Nature (2008) 362.[95] O. Adriani, et al., The PAMELA Collaboration,
Nature (2009) 607.[96] F. Aharonian, et al., H.E.S.S. Collaboration,
Astron. Astrophys. (2009)561.[97] A. A. Abdo, et al., The Fermi LAT Collaboration,
Phys. Rev. Lett. (2009) 181101.[98] M. Ackermann et al. [Fermi-LAT Collaboration], Phys. Rev. D , 092004(2010) doi:10.1103/PhysRevD.82.092004 [arXiv:1008.3999 [astro-ph.HE]].[99] O. Adriani et al. [PAMELA Collaboration], Phys. Rev. Lett. , 201101(2011) doi:10.1103/PhysRevLett.106.201101 [arXiv:1103.2880 [astro-ph.HE]].[100] S. Abdollahi, et al., The Fermi LAT Collaboration, Phys. Rev. D. , (2017) 082007.[101] M. Aguilar, et al., The AMS-02 Collaboration, Phys. Rev. Lett. (2013)141102.[102] L. Accardo, et al., The AMS-02 Collaboration,
Phys. Rev. Lett. (2014)121101.[103] M. Aguilar, et al., The AMS-02 Collaboration,
Phys. Rev. Lett. (2014)121102.[104] G. Bernard, et al.
Astron. Astrophys. (2013) A48.[105] Y. S. Yoon et al.
Astrophys. J. , (2017) 5.[106] B. Bartoli et al., Phys. Rev. D92(9), 092005 (2015)I. De Mitri et al., EPJ Web of Conferences 121, 03009 (2016)S.M. Mari et al., EPJ Web of Conferences 121, 03008 (2016)[107] Z. L. Xu, et al. The Astronomer’s Telegram , No.9901 (2016). http://adsabs.harvard.edu/abs/2016ATel.9901....1X [108] F. Zwicky,
Helv. Phys. Acta (1933) 11043109] J. A. Peacock, Cosmological Physics (Cambridge University Press, 1999).[110] G. Bertone, D. Hooper, and J. Silk, Phys. Rep. , 279 (2005).[111] J. L. Feng,
Annu. Rev. Astron. Astrophys. , 495 (2010).[112] M. Turner and F. Wilczek, Phys. Rev. D (1990) 1001.[113] J. M. Gaskins, Contempoorary Physics , arXiv:1604.00014 (2016).[114] D. Hooper and L. Goodenough,
Phys. Lett. B , 412 (2011).[115] T. Daylan, D. P. Finkbeiner, D. Hooper, T. Linden, S. K. N. Portillo, N.L. Rodd, and T. R. Slatyer,
Phys. Dark Univis. (2016) 1.[116] B. Zhou, Y.-F. Liang, X. Huang, X. Li, Y.-Z. Fan, L. Feng, and J. Chang, Phys. Rev. D , 123010 (2015).[117] F. Calore, I. Cholis, and C. Weniger, J. Cosmol. Astropart. Phys. (2015) 038.[118] M. Ajello et al. (Fermi-LAT), arXiv:1511.02938.[119] A. Geringer-Sameth et al. Phys. Rev. Lett. , (2015) 081101.[120] S. Li et al. Phys. Rev. D , (2016) 043518.[121] T. Bringmann and C. Weniger, Phys. Dark Univis. , (2012) 194.[122] M. Ackermann et al. [Fermi-LAT Collaboration], Phys. Rev. D , no.12, 122002 (2015) doi:10.1103/PhysRevD.91.122002 [arXiv:1506.00013 [astro-ph.HE]].[123] Y. F. Liang, et al. Phys. Rev. D (2016) 103525.[124] P. D. Serpico, Astropart. Phys. , (2012) 2.[125] X. J. Bi, P. F. Yin, Q. Yuan, Front Phys. , (2013) 794.[126] D. P. Finkbeiner, M. Su, C. Weniger, J. Cosmol. Astropart. Phys. , (2013) 029.[127] Y. Li and Q. Yuan, Phys. Lett. B (2012) 35.[128] M. Su and D. P. Finkbeiner, arXiv:1206.1616 (2012).[129] T. Bringmann, F. Calore, G. Vertongen, C. Weniger,
Phys. Rev. D , (2011) 103525.[130] A. V. Kravtsov, A. A. Klypin, J. S. Bullock, and J. R. Primack, Astrophys.J. (1998) 48.[131] J. F. Navarro et al.
Mon. Not. R. Astron. Soc. (2010) 21.44132] F. Acero et al. [Fermi-LAT Collaboration], Astrophys. J. Suppl. , no. 2,26 (2016) doi:10.3847/0067-0049/223/2/26 [arXiv:1602.07246 [astro-ph.HE]].[133] https://fermi.gsfc.nasa.gov/ssc/data/access/lat/BackgroundModels.html [134] M. Ackermann et al. [Fermi-LAT Collaboration], Astrophys. J. , no. 1,14 (2015) doi:10.1088/0004-637X/810/1/14 [arXiv:1501.06054 [astro-ph.HE]].[135] D. Malyshev, I. Cholis, J. Gelfand,
Phys. Rev. D , (2009) 063005.[136] L. Feng, Q. Yuan, X. Li, Y. Z. Fan, Phys. Lett. B , (2013) 1.[137] M. Ackermann, et al. Science (2014) 42.[138] Y. Z. Fan, et al.
ApJ (2013) 95.[139] B. P. Abbott et al.
Astrophys. J. , no. 1, L13 (2016) doi:10.3847/2041-8205/826/1/L13 [arXiv:1602.08492 [astro-ph.HE]].[140] B. P. Abbott et al.
Astrophys. J. Suppl. , no. 1, 8 (2016)doi:10.3847/0067-0049/225/1/8 [arXiv:1604.07864 [astro-ph.HE]].[141] B. Abbott, et al.
Phys. Rev. Lett. (2016) 061102.[142] M. Aartsen, et al.
Phys. Rev. Lett. (2013) 021103.[143] F. Aharonian, et al.
Nature (2006) 695.[144] A. A. Abdo, et al.
Astrophys. J. (2007) L33.[145] M. Aguilar, et al. (AMS Collaboration).
Phys. Rev. Lett. (2015)211101.[146] D. Grasso et al., Astroparticle Physics 32 (2009) 140-151[147] J. Preskill, M. B. Wise, F. Wilczek,
Phys. Lett. B (1983) 127.[148] L. F. Abbott, P. Sikivie,
Phys. Lett. B (1983) 133.[149] M. Dine, W. Fischler,
Phys. Lett. B (1983) 137.[150] P. Sikivie,
Phys. Rev. Lett. (1983) 1415.[151] D. Hooper, P. D. Serpico, Phys. Rev. Lett. (2007) 231102.[152] M. Ajello, et al. (Fermi-LAT Collaboration). Phys. Rev. Lett.116