1. Deciphering the Gamma-Ray Sky: GeV Astronomy in the Era of GLAST Vasiliki Pavlidou Collaborators: Carolyn Brown Tonia Venters Jennifer Siegal-Gaskins Angela Olinto @ The University of Chicago Brian Fields @ University of Illinois

2. Vasiliki Pavlidou 26March2007 Particle Astrophysics Seminar, Fermilab GeV Astronomy: a Play in Two Acts Act I: G-rated (where it all comes together) EGRET, diffuse emission form the Milky Way, blazars, extragalactic diffuse emission, GLAST Act II: R-rated (where it all falls apart) GeV excess, unidentified sources, the whats whos and how muches of the extragalactic diffuse emisison

3. Vasiliki Pavlidou 26March2007 Particle Astrophysics Seminar, Fermilab Act I: The G-rated version of the story

4. Vasiliki Pavlidou 26March2007 Particle Astrophysics Seminar, Fermilab There was once a great satellite known as the Compton Gamma-Ray Observatory, and aboard it lived EGRET, the GeV gamma-ray telescope. CGRO Credit: NASA EGRET Credit: NASA

5. Vasiliki Pavlidou 26March2007 Particle Astrophysics Seminar, Fermilab EGRET observed the sky for almost 10 years, and made many important discoveries. The GeV Sky: all EGRET events with Energies > 100 MeV Galactic Coordinates Credit: S. Digel Galactic poles Galactic plane Galactic center

6. Vasiliki Pavlidou 26March2007 Particle Astrophysics Seminar, Fermilab EGRET observed diffuse emission from the Galactic plane. Diffuse emission from the Galactic plane It comes from the Galactic plane because this is where all the gas lives. This emission is produced through interactions of cosmic rays with interstellar gas.

7. Vasiliki Pavlidou 26March2007 Particle Astrophysics Seminar, Fermilab EGRET also observed many point sources. Some of the ~ 270 point sources resolved by EGRET Of the sources which were identified with some known astrophysical high-energy source, the vast majority were found to be blazars.

8. Vasiliki Pavlidou 26March2007 Particle Astrophysics Seminar, Fermilab Blazars are Active Galactic Nuclei with their jets aligned with our line of sight. Credit: J. Buckley 1998 (Science), illustration: K. Sutliff

9. Vasiliki Pavlidou 26March2007 Particle Astrophysics Seminar, Fermilab Besides the blazars resolved by EGRET, there are also other blazars too faint to be detected. Faint blazars unresolved by EGRET as wil be seen by GLAST Simulation: S. Digel Bright sources detected by EGRET as will be seen by GLAST Simulation: S. Digel These unresolved sources would look to EGRET just like an isotropic, diffuse emission

10. Vasiliki Pavlidou 26March2007 Particle Astrophysics Seminar, Fermilab EGRET indeed detected isotropic, diffuse emission It looks airtight: (a) CR-produced gamma-rays where we expect them (b) resolved point sources (c) isotropic background from unresolved point sources

11. Vasiliki Pavlidou 26March2007 Particle Astrophysics Seminar, Fermilab The future: GLAST (Gamma-ray Large Area Space Telescope: a bigger, better EGRET) is scheduled for launch in fall 2007. Artist’s impression of GLAST Credit: NASA It will again open up the GeV window for observations and will be a unique opportunity to follow up on the on the discoveries of EGRET EGRETGLAST Peak effective area 1500 cm 2 8000 cm 2 Field of view 0.5 sr2.5 sr Angular resolution 5.8° (100 MeV)3.4° (100 MeV) Source location 15 arcmin<0.4arcmin Sensitivity 10 -7 cm -2 s -1 3  10 -9 cm -2 s -1

12. Vasiliki Pavlidou 26March2007 Particle Astrophysics Seminar, Fermilab And they all lived happily ever after The End (?)

13. Vasiliki Pavlidou 26March2007 Particle Astrophysics Seminar, Fermilab Act II: The R-rated version of the story

14. Vasiliki Pavlidou 26March2007 Particle Astrophysics Seminar, Fermilab Credit: Strong, Moskalenko & Reimer 2004 Diffuse emission from the MW: a closer look EGRET data GeV Excess Model Ingredients for building a MW diffuse emission model: Locally measured CR spectrum Observed distribution of interstellar matter + light Theoretical/experimental cross sections for relevant interactions Add them all up, calculate energy spectrum of gamma-ray intensity

15. Vasiliki Pavlidou 26March2007 Particle Astrophysics Seminar, Fermilab What causes the GeV excess? 1.The CR spectrum elsewhere in the Galaxy is not the same as the CR spectrum in the Solar System (Strong, Moskalenko, Reimer, Mori - but, why? and needs to tweak electron, proton spectra separately to fully fix problem) 2.Point Sources: pulsars, gas clouds, SNRs, bright stars ( Büsching, Pohl, Schlickeiser, Strong, Mori - but GeV excess persists in all directions and sources live mostly in the plane ) 3.Obsolete proton-proton cross section ( Kamae, Abe, Koi - but also need to tweak proton spectrum to fully account for the problem ) 4.Annihilating dark matter [ de Boer - but resulting MW dark matter halo shape counter-intuitive and would make too many anti- protons (Bergstrom et al 2006) and what about the TeV excess (Prodanovic et al 2006)? ]

16. Vasiliki Pavlidou 26March2007 Particle Astrophysics Seminar, Fermilab But, if we don’t understand how to model the Galaxy, how can we subtract it from the sky map to obtain the extragalactic diffuse background? With very large uncertainties. Conclusion: we don’t really understand how to model the GeV emission from the Galaxy Credit: Strong et al 2004 Credit: Strong et al. 2004 Dar & De Rujula (2001) and Keshet, Waxman & Loeb (2004) would prefer NO extragalactic background

17. Vasiliki Pavlidou 26March2007 Particle Astrophysics Seminar, Fermilab However: certain source classes are guaranteed to make some contribution to the extragalactic diffuse background - e.g. blazars. Contribution of these sources must place a lower limit to extragalactic diffuse background intensity! Complications: Blazar contribution to the extragalactic background extensively studied, [e.g. Padovani et al. 1993; Stecker et al. 1993; Salamon & Stecker 1994; Chiang et al. 1995; Stecker & Salamon 1996; Chiang & Mukherjee 1998; Mukherjee & Chiang 1999; Mücke & Pohl 2000; Dermer 2006; Giommi et al. 2006; Narumoto & Totani 2006] but no consensus on result Estimates of unresolved blazar flux vary by ~ 2 orders of magnitude Conclusion: we don’t really know how many EGRET photons come from the extragalactic diffuse background.

18. Vasiliki Pavlidou 26March2007 Particle Astrophysics Seminar, Fermilab Conclusion: we don’t really know how much blazars contribute to the extragalactic diffuse gamma-ray background. Therefore the extragalactic diffuse emission is very poorly understood. Do we at least understand the nature of the resolved point sources? Of the 270 EGRET point sources: - 93 are blazars or “maybe blazars” - 170 were originally unidentified - most of those are still unidentified - see http://GeVSky.org

19. Vasiliki Pavlidou 26March2007 Particle Astrophysics Seminar, Fermilab Summary of Puzzles: EGRET detected diffuse emission from the Galactic plane, but we don’t really understand how it is produced. EGRET resolved point sources, and we don’t know what most of those are. EGRET detected some isotropic diffuse emission, but we don’t know how much exactly, and we don’t know what its spectrum should look like. We don’t know what the origin of the isotropic diffuse emission is. Conclusion: we don’t really understand the EGRET observations at all.

20. Vasiliki Pavlidou 26March2007 Particle Astrophysics Seminar, Fermilab How can progress be made? Ask the right questions identify potentially important effects Global approach problems interconnected Model-independent techniques too detailed models won’t necessarily improve understanding Explicitly account for uncertainties measurement uncertainties high and unequal

21. Vasiliki Pavlidou 26March2007 Particle Astrophysics Seminar, Fermilab Unidentified Sources: how can they hurt you? If members of an already known class (e.g. blazars): –affect normalization of bright-end of class luminosity function –affect contribution of class to the diffuse background If members of an unknown Galactic class: –unresolved members contribute to Galactic diffuse emission  GeV excess? –counterparts in external galaxies enhance emission of hosts If members of an unknown extragalactic class: –unresolved members contribute to extragalactic background - may even be dominant component!

22. Vasiliki Pavlidou 26March2007 Particle Astrophysics Seminar, Fermilab Are most EGRET unIDed sources Galactic or extragalactic? Can isotropy help? Is there another test?

23. Vasiliki Pavlidou 26March2007 Particle Astrophysics Seminar, Fermilab Are most EGRET unIDed sources Galactic? A new approach If EGRET unIDed sources represent a Galactic population, galaxies similar to the MW must have similar such populations In this scenario: if we place all unIDs at distance of M31, their summed flux should not exceed the M31 gamma-ray flux

24. Vasiliki Pavlidou 26March2007 Particle Astrophysics Seminar, Fermilab Are most EGRET unIDed sources Galactic? using M31 to test For this exercise to work: –need unIDed source luminosities. –but EGRET only measured fluxes & angular positions –so we need way to assign distances Simple first tests: –place all sources at same distance from Earth –place all sources at same distance from Galactic Center –assume all sources have similar luminosities More sophisticated technique: –perform Monte Carlo simulations according to mass distribution in Galaxy

25. Vasiliki Pavlidou 26March2007 Particle Astrophysics Seminar, Fermilab Monte-Carlo Method of Assigning Distances For each source: –probability of source found within distance r  integrated mass out to r One realization of the model: –assign a distance to each unidentified source –calculate total luminosity for each realization Hypothesis testing: –generate many realizations of the model –Calculate how probable it is to get a total luminosity less than the M31 upper limit Assume unidentified source population follows MW mass distribution.

26. Vasiliki Pavlidou 26March2007 Particle Astrophysics Seminar, Fermilab Case I: Could most EGRET unIDed sources be a Galactic halo population?

27. Vasiliki Pavlidou 26March2007 Particle Astrophysics Seminar, Fermilab Case II: Could most EGRET unIDed sources be a Galactic disk+bulge population?

28. Vasiliki Pavlidou 26March2007 Particle Astrophysics Seminar, Fermilab Conclusion: very likely that EGRET unIDs contain a very significant extragalactic component What would then be the contribution of their unresolved counterparts to the extragalactic diffuse emission? Caution! Question involves daunting uncertainties even for blazars with 67 well-studied members and deep AGN surveys – how can we ever hope to have meaningful result for unidentified sources? (no redshifts / no luminosities / no idea about their cosmic evolution) Instead: try to assess whether it is likely for the unID component to be important/dominant

29. Vasiliki Pavlidou 26March2007 Particle Astrophysics Seminar, Fermilab An empirical model for collective emission from unresolved unIDs Seek to answer 2 questions: 1.Numbers/Fluxes: How plausible is that unresolved unIDs, if extragalactic, have significant contribution to gamma-ray background? 2.Spectral indices: Would collective unresolved emission from unIDs be spectrally consistent with the gamma-ray background? Assessing the unIDed contribution to the extragalactic background: I know one thing, that I know nothing

30. Vasiliki Pavlidou 26March2007 Particle Astrophysics Seminar, Fermilab 1. Exclude likely Galactic sources based on location on the sky 2.Try analysis with 3 different samples (from “generous” to “conservative”) to test sensitivity of results a) all sources with no 3EG ID + all “maybe blazars” b) all sources with no 3EG ID c) all sources with neither 3EG ID nor later proposed possible ID Which resolved sources?

31. Use cumulative flux distribution of resolved objects and extrapolate to lower fluxes –how far into the low-flux regime should it be extrapolated so that contribution is dominant at least at low energies? –is the extrapolation extreme and unlikely or moderate and plausible? Power-law fit to range of fully resolved fluxes UnIDs and the extragalactic diffuse background: are there enough? Cutoff so that EGRB is not exceeded

32. UnIDs and the extragalactic diffuse background: does the spectrum work? Energy spectrum of cumulative emission depends on spectra of individual sources Critical input: spectral index distribution. –Assume spectral index distribution of unresolved objects same as that of resolved objects –BUT - need to account for individual measurement errors! likelihood approach

33. unresolved unID emission Strong, Moskalenko & Reimer (2004) Sreekumar et al (1998) EGRB fit SMR (2004) EGRB SMR (2004) systematics Spectrum of Extragalactic background from UnIDs

34. Vasiliki Pavlidou 26March2007 Particle Astrophysics Seminar, Fermilab EGRET Unidentified Sources - Conclusions M31 constraint excludes with high significance the possibility that unIDs are a Galactic halo population Galactic disk+bulge population allowed by M31 constraint, although sky distribution uncomfortable If most unidentified sources Galactic, most probable total luminosity ~ 50% of upper limit of M31 gamma-ray luminosity. For high-latitude, likely extragalactic, unidentified sources: a moderate extrapolation of the flux distribution for ~ 1 order of magnitude implies unID contribution to extragalactic background dominant at low energies Spectrum of unresolved unID emission in excellent agreement with observed extragalactic diffuse background at energies where possibly dominant, within systematics elsewhere. We can never hope to adequately understand the origin of the diffuse emission without at least some understanding of the nature of unidentified sources

35. Vasiliki Pavlidou 26March2007 Particle Astrophysics Seminar, Fermilab The next steps - GLAST! Whatever the nature of the EGRET unIDs, GLAST will resolve many more –if unresolved unIDs currently responsible for considerable fraction of extragalactic background:  associated reduction of GLAST isotropic background –If unresolved unIDs currently responsible for considerable fraction of Galactic diffuse emission:  associated reduction of GLAST diffuse MW (GeV excess affected?) GLAST will detect M31, all M31-based constraints stronger.

36. Vasiliki Pavlidou 26March2007 Particle Astrophysics Seminar, Fermilab The next steps - theory Unresolved unIDs a potentially significant contribution to diffuse emission  worth pursuing more detailed models. Nature unknown, but there are only so many energy- releasing mechanisms sufficient to generate high-energy emitters –What if they follow AGN evolution? –What if they follow evolution of cosmic structure? –What if they follow cosmic star formation? Depending on nature, GLAST will detect different numbers  population identification with no need of identification of individual members

37. Vasiliki Pavlidou 26March2007 Particle Astrophysics Seminar, Fermilab To be continued… in a preprint server near you

38. Vasiliki Pavlidou 26March2007 Particle Astrophysics Seminar, Fermilab

39. Credit: de Boer et al 2005 Credit: Kamae et a. 2005 Credit: Strong et al. 2004 GeV Excess Explained