Gamma-Ray Particle Astrophysics: the first year of the Fermi Gamma-ray Space Telescope

1. Gamma-Ray Particle Astrophysics: the first year of the Fermi Gamma-ray Space Telescope Tsunefumi Mizuno Hiroshima Univ. on behalf of the Fermi Collaboration September 02, 2009, Kobe, Japan

2. Plan of the Talk • Review of the high energy gamma-ray missions • Highlights of the Fermi’s first year results: • Gamma-ray bursts • implication on fundamental physics and UHECRs • properties of jets with highest G • Galactic cosmic-rays and dark matter • Direct measurement of Galactic cosmic-rays • Galactic diffuse gamma-rays as an indirect probe of Galactic CRs • Selected Galactic/extragalactic gamma-ray objects • focus on the relation to Galactic CRs and UHECRs

3. Review of High-Energy Gamma-ray Astrophysics Missions

4. GeV Gamma-ray Astrophysics (Eg = a few 10s MeV to ~100 GeV) • 1967 to 1968 -- OSO-3 : First detection of g-rays from the Gal. plane • 1972 to 1973 -- SAS-2 : Crab, Vela, and Geminga • 1975 to 1982 -- COS-B : >=20 g-ray sources EGRET: (on the Compton Gamma Ray Observatory) 271 (>5s) g-ray sources + detailed map of the Galaxy 1991 -- 2000 • 2007 to present -- AGILE • 2008 to present -- Fermi A new gamma-ray satellite every 10 or 15 years

5. TeV g-ray Astrophysics withAtmospheric Cherenkov Imager Arrays (Eg >= 100 GeV) HESS galactic survey Nearly 100 sources under study. CANGAROO III, Australia CTA (2013~) MAGIC II, Canary islands, Spain H.E.S.S., Namibia VERITAS, Arizona, USA Very important but not covered by this talk. See, e.g., talk by Schwanke in PIC 08

6. Fermi Launch • Launched from Cape Canaveral Air Station on June 11, 2008 • Science Operation on Aug 4, 2009 • Orbit: 565 km, 26.5o (low BG)

7. Large Area Telescope (LAT) on Fermi 20 MeV to >= 300 GeV FOV: 2.4 sr • Tracker: Si-strip detectors & W converters • Identification and direction measurement of g-rays • Calorimeter: hodoscopic CsI scintillators • Energy measurement • ACD: segmented plastic scintillators • BG rejection • Technology developed through HEP experiments • See Atwood et al. (ApJ 697, 1071, 2009) for detail

8. Gamma-ray Burst Monitor (GBM) on Fermi Views entire unocculted sky with • 12 NaI detectors: 8 keV - 1 MeV • 2 BGO detectors: 150 keV - 40 MeV LAT+GBM=> more than 7 decades of energy OK, let’s start with GRBs

9. Highlights from Fermi’s 1st year (1): Gamma-ray Bursts

10. Gamma-Ray Bursts Overview (1) • Discovered in 1967 • Cosmological origin (BeppoSAX, BATSE) • Large apparent energy release: Eiso ~ 1052 - 1054 erg • Large Lorentz factor of jet: G >= 100 (a few for m-QSO and ~10 for AGN) • Energetics may be consistent with origin of UHECRs • Peak in ~ MeV gamma-rays • Band function: smoothly joins two power-laws • Synchrotron radiation of ultra-relativistic electrons in jet? 0.01 0.1 1 10 100 MeV

11. Gamma-Ray Bursts Overview (2) 2 s • Bimodal distribution of duration time • Short (<2 s) GRB: progenitor unknown • Merger of NSs or BHs? • Long (>2 s) GRB: association with supernova • Core-collapse supernovae • Gamma-ray emission mechanism not fully understood yet • Fermi observation of GRBs is expected to • constrain the emission mechanism • constrain the bulk Lorentz factor of jet • limit on Lorentz invariance violation • search for the clue of UHECRs • probe the extragalactic background light (star formation in early universe) T90 (duration) in seconds

12. Fermi GRB Skymap (as of Jun. 29, 2009) • 7 long + 2 short GRB by GBM+LAT, from 8 keV to tens of GeV • Short & long GRBs: similar phenomenology at high energy? Abdo et al. Sci.323, 1688 (2009) 241 GBM GRBs 9 LAT GRBs 129 In Field-of-view of LAT Abdo et al., submitted to Nature (arXiv:0908.1832)

13. GRB080916C Prompt Emission (<=100s) • z=4.35 +/- 0.15 (GROND; GCN8257) • More than 3000 LAT photons, 145 above 100 MeV and 14 above 1 GeV • Delayed HE onset (1st peak not seen > 100 MeV) • Opacity effect (gg->e+e-)? But no evidence of spectral cutoff • Single Band-function dominant for 6 decades of energy • Lack of prominent SSC component implies high magnetic field or high ge 8-260 keV 260 keV-5 MeV LAT (all) >100 MeV >1 GeV 0 20 40 60 80s

14. Long-Lived HE Emission • HE (>100 MeV) emission shows different temporal behavior • Temporal break in LE emission while no break in HE emission • Cascades induced by ultra-relativistic ions? • Angle-dependent scattering effects? Flux in LAT/GBM bands • E>100 MeV • index = -1.2 +/- 0.2 • E= 50 -300 keV • index: ~-0.6 => ~-3.3 • (at ~T0+55s) Photon Index (LAT only) no significant evolution (Epeak gradually decreases)

15. Constraints on Bulk Lorentz Factor • Large luminosity and short variability time imply large optical depth due to gg -> e+e- (compactness problem) • Small emission region: R ~ cDt • tgg(E) ~ (11/180)sTN>1/E/4pR2 • tgg(1 GeV) ~ 7x1011for a typical GRB of fluence=10-6erg/cm2, z=1, Dt=1 s • Relativistic motion (G >> 1) can reduce optical depth • Lager emission region: R ~ G2cDt • Reduced photon # densities: • N>1/E∝ G2b+2(note: b ~ -2.2) • Blue shift of energy threshold: Eth∝ G • Blue shift of spectrum: N(E) = (GE)b+1 • Overall reduction of optical depth: G2b+2 /G4=G2b-2 ~G-6.4(<=10-12 for G=100) • Limit from GRB 080916C: G ≳ 890±21 • (Largest ever observed as of May 2009)

16. Limits on Lorentz Invariance Violation (LIV) 16.5 s • Some QG models violate Lorentz invariance. A high-energy photon would arrive after a low-energy one emitted simultaneously. (Jacob & Piran 2008. n=1 for linear LIV) • GRB080916C: • 13.2 GeV @ T0+16.5 s • MQG, 1 > (1.5±0.2) x 1018 GeV/c2,1/10 of the Plank mass and the highest as of May 9, 2009. GRB080916C Planck mass Pulsar(Kaaret 99) GRB(Ellis 06) AGN(Biller 98) GRB(Boggs 04) AGN(Aharonian 08) min MQG (GeV) 1015 1016 1017 1018 1019

17. GRB090510 (1) Abdo et al. 2009 Submitted to Nature (arXiv:0908.1832)

18. GRB095010 (2) • Time vs. photon energy • LAT all events • E>100 MeV • Short GRB with > 150 photons above 100 MeV • 31 GeV @ ~T0+0.83s • Solid and dotted line are LIV for n=1 and 2, respectively • Several assumptions of tstart indicated by different colors • Even the conservative case (black line) implies MQG, 1 > 1.19 MPlanck • Other important findings • deviation from Band function • highest Epeak: 5.1 MeV (Band+PL model fit) • delayed onset of LAT emission by 0.1-0.2 s • highest Gmin (~1200) 10 1GeV 0.1 0.01 Preliminary 0 0.5 1 1.5 2 (s)

19. Highlights from Fermi’s 1st year (2): Direct Measurements of Galactic CR Electrons

20. Introduction (1):What Can We Learn from HE e-/e+ and p/p ? • Inclusive spectra: e- + e+ • Electrons, unlike protons, lose energy rapidly by Synchrotron and Inverse Compton: at very high energy they probe the nearby sources • Charge composition: e+/(e- + e+) and p/(p + p) ratios • e+ and p are produced by the interactions of high-energy cosmic rays with the interstellar matter (secondary production) • There might be signals from additional (astrophysical or exotic) sources • Different measurements provide complementary information of the origin, acceleration and propagation of cosmic rays • All available data must be interpreted in a coherent scenario

21. Introduction (2):Positron and Antiproton Fraction: 2008-09 • PAMELA positron and antiproton • Nature 458, 607 (2009) • PRL 102, 051101 (2009) 1 GeV 10 100 • Antiproton fraction consistent with secondary production • Anomalous rise in the positron fraction above 10 GeV • Several different viable interpretations (>200 papers over the last year) See also Nature 456, 362 (2008) and PRL 101, 261104 (2008) for pre-Fermi CRE spectrum by ATIC and HESS.

22. Fermi-LAT Capability for CR Electrons • Candidate electrons pass through 12.5 X0 on average ( Tracker and Calorimeteradded together) • Simulated residual hadron contamination (5-21% increasing with the energy) is deducted from resulting flux of electron candidates • Effective geometric factor (Gf) exceeds 2.5 [m2 sr] for 30 GeV to 200 GeV, and decreases to ~1 [m2 sr] at 1 TeV. Gf times exposure has already reached several x 107 [m2 sr s]. (very high statistics) • Full power of all LAT subsystems is in use: Tracker, Calorimeter and ACD act together Geometric Factor (Gf) Residual hadron contamination 20 GeV 100 GeV 1 TeV

23. Fermi-LAT Electron Spectrum • Abdo et al. Phys. Rev. Let. 102, 181101 (2009) • Cited 38 times within a month • APS Viewpoint Harder spectrum (spectral index: -3.04) than previously thought • Total statistics collected for 6 months of Fermi LAT observations • ~4.5 million candidate electrons above 20 GeV • > 400 candidate electrons in last energy bin (770-1000 GeV)

24. Implication from Fermi-LAT CRE (1) for detail, see D. Grasso et al. arXiv:0905.0636 (accepted by Astroparticle Physics) Old “conventional” CRE Model g0=2.54 New “conventional” CRE models g0=2.42 g0=2.33 Fermi CRE spectrum can be reproducedby the “conventional” Galactic cosmic-ray source model, with harder injection spectral index (-2.42) than in a pre-Fermi conventional model (-2.54). All that within our current uncertainties, both statistical and systematic.

25. Implication from Fermi-LAT CRE (2) • Now include recent PAMELA result on positron fraction • Qualitative approach: the harder primary CRE spectrum is, the steeper secondary-to-primary e+/e- ratio should be. PAMELA shows the opposite. New “conventional” CRE models Old “conventional” CRE Model Precise Fermi measurement increases the discrepancy between a purely secondary origin for positrons, and the positron fraction measured by PAMELA.

26. Implication from Fermi-LAT CRE (3) • It is becoming clear that we are dealing with at least 3 distinct origins of HE e-/e+ • Uniformly distributed distant sources, likely SNRs. • Unavoidable e+e- production by CRs and the ISM • And those that create positron excess at high energies. Nearby (d<1 kpc) and Mature (104 - 106 yr) pulsars? “conventional” sources Anexample of the fit to both Fermi and PAMELA data with Monogem and Geminga with a nominal choice for the e+/e- injection parameter (blue lines). Works well. (Discrepancy in positron fraction in low energy can be understood as the charge-sign effect of solar modulation)

27. Dark Matter Interpretation • What said about pulsars is applicable to dark matter as sources of e- and e+. • PAMELA and Fermi data tighten the DM constraints, favoring pure e+e-, lepto-philic, or super-heavy DM models. preferred likely excluded sv [cm3/s] pure e+e- Models lepto-philic Super-heavy DM 10-22 10-19 10-24 10-21 10-26 10-23 100 GeV 1 TeV DM mass • We need local sources (astrophysical or exotic). The origin is still unclear but is strongly constrained by Fermi data (+ others) • More results from Fermi-LAT are coming. Extending energy range to 5 GeV – 2 TeV and searching for the CRE anisotropy at a 1 % level.

28. Highlights from Fermi’s 1st year (3): Galactic Diffuse Gamma-ray Emission (Indirect Probe of Galactic CRs)

29. 1 particle/m2/sec Flux (m2 sr s GeV)-1 Knee 1 particle /m2/yr Ankle 1 particle/km2/yr Energy (eV) Cosmic-Rays Overview • Discovered by V. Hess in 1912 • Globally power-law spectrum with some structures (knee and ankle) • hint of the origin • Large energy density (~1 eV cm-3): comparable to UB and Urad • UHECRs : not covered by this talk in detail • small scale anisotropy V. Hess, 1912 Galactic G or EG? Extragalactic • Auger Collaboration, Sci. 318, 938 (2007) • 18/27 events > 5.6 x 1019 eV correlate with nearby AGNs. See also arXiv:0906.2347

30. SNR RX J1713-3946 B HESS π 0 e e π gas gas + + + - - - CRs and Galactic Diffuse Gamma-Rays HE g-rays are produced via interactions between Galactic cosmic-rays (CRs) and the interstellar medium (or interstellar radiation field) (CR Accelerator) (Interstellar space) (Observer) ISM X,γ synchrotron Chandra, Suzaku, Radio telescopes IC ISRF P He CNO diffusion energy losses reacceleration convection etc. bremss Pulsar, m-QSO HESS, Fermi A powerful probe to study CRs in distant locations

31. Outstanding Question: EGRET GeV Excess • EGRET observations showed excess emission > 1 GeV everywhere in the sky when compared with models based on directly measured CR spectra • Potential explanations • Unexpectedly large variations in cosmic-ray spectra over Galaxy • Dark Matter • Unresolved sources (pulsars, SNRs, …) • Instrumental • Fermi-LAT is able to confirm or deny this phenomena |b|=6°-10° 0.1 1 10 GeV |b|=2°-6° |b|<=2° ~100% difference above 1 GeV Hunter et al. 1997

32. Intermediate Latitude Region seen by LAT |b|=10°-20° EGRET LAT Preliminary Abdo et al. submitted to PRL Porter et al. 2009 (arXiv:0907.0294) 0.1 1 10 GeV • |b|=10°-20°: avoid Gal. plane but still have high statistics • EGRET spectrum extracted for the same region • LAT spectrum is significantly softer and does not confirm the EGRET GeV excess • Strongly constrains the DM interpretation

33. Probing CRs using Gamma-rays from ISM • Correlation with gas column density reveals the CR spectrum • Method go back to SAS-2/COS-B era • Fermi-LAT’s high performance + CR propagation model (e.g. GALPROP) to predict IC • Sensitivity significantly improved ISM (e.g., LAB HI survey) (http://www.astro.uni-bonn.de/~webaiub/english/tools_labsurvey.php) Gamma-ray intensity (Fermi LAT data) • High latitude region: • Detailed study of local CRs (most of the gas is close to solar system) • Galactic plane: • CR gradient in the Galaxy (need to resolve point sources)

34. Accurate Measurements of Local CRs • Mid-high lat. region in 3rd quadrant: • small contamination of IC and molecular gas • correlate g-ray intensity and HI gas column density Abdo et al. 2009, accepeted by ApJ (arXiv:0908.1171) contact author: TM • Best quality g-ray spectrum in 100 MeV-10 GeV (Tp = 1-100 GeV) • Agree with the model prediction from the local interstellar spectrum (LIS) LAT data model from the LIS nucleon-nucleon • Prove that local CR nuclei spectra are close to those directly measured at the Earth electron-bremsstrahlung

35. CR Distribution in Galaxy • CR distribution is a key to understand their origin and propagation • distribution of SNRs not well measured • Previous Gamma-ray data suggests a flatter distribution than SNR/pulsar distributions (e.g., Strong et al. 2004) Pulsar distribution (Lorimer 2004) SNR distribution (Case & Bhattacharya 1998) CR source distribution from g-rays (Strong & Mattox 1996) sun 0 5 10 15 kpc • Fermi-LAT is able to map out CR distributions in the Galaxy with unprecedented accuracy • Work in progress. (arXiv:0907.0304 and arXiv:0907.0312) Gal. Center Inner Galaxy Outer Galaxy

36. Highlights from Fermi’s results (4): Selected Galactic and Extragalactic Objects as a Key to Understand CRs

37. Introduction: g-ray objects seen by the LAT • Variety of objects in the LAT bright source list (Abdo et al. ApJS 183, 46, 2009) • >=200 sources. More than 80% are identified (EGRET:~30%) • Here I will pickup SNRs, LMC and Blazars and briefly discuss their implications for CRs. • Many other very important objects and topics will not be discussed. (See LAT publications, please)

38. Fermi LAT Study on SNRs • SNRs are the most favored explanations for the origin of Galactic CRs. • Diffusive shock acceleration in SNR shell. Sufficient to supply CRs up to knee. • Significant progress in recent years in keV and TeV observation of young SNRs. • Key issues to be addressed by Fermi-LAT: • Searching for pion signatures & measuring total energy content per SNR • Several possible associations to SNRs in the LAT bright source list including • W44: (T. Tanaka et al. proc. ICRC 2009) • Middle age (2000 yr), Mixed Morphology, 3 kpc • Interactions with Molecular Cloud • EGRET • Fermi-LAT: (0FGL J1855.9+0126: 3 month data yield 39s) • W51C: (Y. Uchiyama et al. proc. ICRC 2009) • Middle age (20000 yr), 6 kpc • Interactions with MC • HESS (Fiasson et al. 2009, no spectrum) • Fermi-LAT: (0FGL J1923.0+1411: 3 month data yield 23s)

39. W51C: The Fermi Source is “Extended” • Mean surface brightness (2-8 GeV) as a function of distance from the SNR center vs. Fermi-LAT PSF => Spatially extended Black contours: ROSAT X-ray (0.1-2.4 keV) Green contours: VLA 1.4 GHz Color: Fermi-LAT count map (2-8 GeV) Preliminary 0.6 deg R (Note) PSF of Fermi LAT depends heavily on energy. The PSF shape above is obtained by taking account of the energy distribution (not presented).

40. Spatial Extent of W44 Profile along the rectangle Contributions form the diffuse backgrounds and nearby sources are subtracted Smoothed Count Map (>1 GeV) Preliminary Red: Observed Counts Black: Expected Profile for a Point Source Black Cross: Pulsar (PSR B1853+01) location • For both W44 and W51C, gamma-rays are spatially “extended” & positionally coincident with SNRs. The luminosity is found to be very large. • Spectral analysis will be presented in a refereed journal

41. Local Group Galaxies EGRET Observation Summary: • LMC detection: CR density is inferred to be similar to MW • SMC non-detection: CR density is smaller than in the MW • M31 non-detection: has to have smaller CR density than the MW (size M31>MW) • First direct evidence that CRs (E<Eknee) are Galactic and not universal • Key issues not fully addressed yet • CR propagation in each Galaxy • detailed comparison of CR densities among galaxies

42. Fermi-LAT Resolved the LMC CRATES J060106-703606 30 Doradus Gal. longitude Preliminary Gal. latitude • 161 days of survey data, ~ 1300 events above 100 MeV • Gamma-ray is clearly extended, with the maximum consistent with the massive star-forming region 30 Doradus Dust map (SFD) Detailed study of spatial and energy distribution is in progress adaptively smoothed 100 MeV - 10 GeV counts map (s.n.r. = 5)

43. LAT Bright AGN Sample (LBAS) • 125 non-pulsar sources at |b|>10o • 106 high-confidence (P>90%) associations with AGNs • 11 lower-confidence (40%<P<90%) associations • 9 unidentified (3EG: 96/181 at |b|>10o) 58 FSRQ 42 BL Lac 4 of Uncertain class 2 Radio Galaxies Only ~30% of the bright Fermi AGNs were detected by EGRET. The Sky changes! Abdo et al. ApJ 700, 597 (2009)

44. Population of the LAT AGNs • 42 BL Lacs and 58 FSRQs (EGRET: 14 and 46) • BL Lac has harder spectrum than FSRQ (1.99 +/- 0.22 vs. 2.40 +/- 0.17) • V/Vmax test (Schmidt 1968) indicates the positive evolution for FSRQ • (more sources or brighter sources at earlier time) • Local emissivity • ℓBL ≥ 1031 W Mpc-3, ℓFSRQ ≈ 1030 (ℓUHECR ≈ 3x1029; Waxman & Bahcall 1999) • BL Lacs are favored as the origin of UHECRs (if AGNs are the sources) S. Razzaque, J. Finke and C. Dermer 3x1029 W Mpc-3

45. Summary • Presented a very biased summary of gamma-ray particle astrophysics • Long history of more than 40 years. Significant progresses in recent years by Air Cherenkov Telescopes and Fermi. • Fermi view of GRBs: • >240 GRBs, 9 detected by LAT (as of June 2009) • GRB080916C & GRB090510 • strongly constrains the bulk Lorentz factor, Lorentz invariance, etc.. • CR electrons by Fermi + PAMELA and other data. • Local sources are required. • Nearby mature pulsars. Constrains on DM scenario • Diffuse gamma-rays as a probe of Galactic CRs • non-GeV-excess. Local CRs close to those measured at the Earch. • Is able to map out CR distribution in the Galaxy • Found extended sources positionally associated with SNR. Resolved LMC for the first time. BL Lacs are favored (than FSRQs) as the origin of UHECR. Thank you for your attention!

46. Backup Slides

47. GRB 080916C Spectrum • No conclusive evidence of extra HE component • Probability of no extra component is ~1% • Effect of EBL • HE absorption • Transparency:0.03–1.0(model dependent) • Single Band-functiondominant for 6 decadesof energy band • Lack of prominent SSC component implies • High magnetic field • εe/εB ≲ 0.1 • Epeak,SSC ≫ 10 GeV (γe ≫ 100) Time bin ‘d’ LAT GBM NaI Band + power law GBM BGO Band function Ep,syn νFν γe2 Ep,SSC ~εe/εB SSC synchrotron ν

48. FOM for CRE Measurement Exposure factor (effectively) determines the # of counts Ef(E) = Gf(E)*Tobs L. Baldini

49. LAT vspre-Fermi Model • Compare with a CR propagation model prediction based on pre-Fermi CR data (Strong et al. 2004, Porter et al. 2008) • π0-decay, e-Brems, Inverse Compton • Source and isotropic (w/ residual BG) component come from fitting the data to the sky above 30 deg latitude with model fixed • Although there is a uniform excess above the model, data is reasonably reproduced by the model LAT model total Preliminary • p0-decay • e-Brems • IC The model is successful considering it is a priori pre-Fermi model

50. Correlation with the HI Column Density • Mask point sources (52 total) and subtract the residual point source contributions. Also subtract the IC contributions. • Correlation from 100 MeV to 10 GeV. The slope gives the g-ray emissivity spectrum of local HI gas produced through interactions with CRs. (error bars are statistical only) 400-560 MeV 1.6-2.3 GeV HI column density (1020 cm-2) 400-566 MeV E2 x g-ray Intensity HI column density (1020 cm-2)