1. X-ray and gamma-ray observations of active galaxies as probes of their structure Greg Madejski Stanford Linear Accelerator Center and Kavli Institute for Particle Astrophysics and Cosmology Outline: Two classes of active galaxies - (1) Isotropic emission dominant (2) Relativistic jet emission dominant Isotropic emitter (black hole and the accretion disk) as the source of the jet Emission processes and content the relativistic jet Key questions about the nature and the origin of the jet Future observational prospects in the high energy regime towards the answers: GLAST and the future X-ray missions – AstroE2, NuSTAR

2. Compton Gamma-ray Observatory Featured instruments sensitive from ~ 40 keV (OSSE) up to nearly 100 GeV (EGRET)

3. Global observational differences between: Radio-quiet and jet-dominated active galaxies active galaxies(a. k. a. blazars) nostrongMeV – GeV emission Strong signatures of Only weak signatures… circumnuclear matter (symmetric emission lines) No strong compact radio Radio, optical, X-ray cores and jets structure Both classes are rapidly variable, requiring simultaneous observations => spectra and variability patterns can reveal structure and physical processes responsible for emission X-rays and  -rays vary most rapidly – presumably originate the closest to the central engine (?) Study of the isotropic emission as well as the jet should reveal the details of formation, acceleration, and collimation of the relativistic jets

4. Radio, optical and X-ray images of the jet in M 87 * Jets are common in AGN – and radiate in radio, optical and X-ray wavelengths * Blazars are the objects where jet is pointing close to the line of sight * In many (but not all) blazars, the jet emission dominates the observed spectrum

5. Unified picture of active galaxies Presumably all AGN have the same basic ingredients: a black hole accreting via disk- like structure In blazars the jet is most likely relativistically boosted and thus so bright that its emission masks the isotropically emitting “central engine” But… the nature of the isotropically emitting AGN should hold the clue to the nature of the conversion of the gravitational energy to light Again, X-ray and  -ray emission varies most rapidly – potentially best probe of the “close-in” region Diagram from Padovani and Urry

6. Radio galaxy M87 (Virgo-A) studied with the HST Black holes are a common ingredient of galaxies When fed by galaxian matter, they shine – or produce jets – or both The BH mass is very important to know L & the accretion rate in Eddington units Weighing the central black hole Seyfert galaxy NGC 4258 studied using H 2 O megamaser data

7. High-energy spectra of isotropically-emitting AGN: Example is an X-ray bright Seyfert 1 galaxy IC 4329a General description of the broad-band intrinsic X-ray spectrum of a “non-jet” (isotropically-emitting) AGN is a power law, photon index ~ 2, with exponential-like cutoff at ~ 200 keV Asca, XTE, OSSE data for IC 4329a (from Done, GM, Zycki 2003) Average OSSE / Ginga spectrum of ~20 AGN looks essentially the same

8. Effects of the orientation to the line of sight in AGN The spectrum of the object depends on the orientation with respect to the line of sight – soft X-rays are (photo-electrically) absorbed by the surrounding material

9. X-ray Background Spectrum (from G. Hasinger) => E 5 keV still lots of work... Revnivtsev et al., 2003 RXTE XMM LH resolved Worsley et al. 2004 from Gilli 2003

10. Heavily obscured AGN “hiding in the dust”: Important ingredient of the Cosmic X-ray Background? Absorbed (“Seyfert 2”) active galaxy NGC 4945 - Anonymous in radio, optical, but the X-ray spectrum taken by us (Done, GM, Smith) reveals one of the brightest 100 keV AGN in the sky - Sources similar to NGC 4945 at a range of Lx, z and absorption can make up the CXB - BUT – for this, one needs most of the AGN to be heavily absorbed… What’s the absorption geometry? The origin of the diffuse Cosmic X-ray Background is one of the key questions of high energy astrophysics research Most likely it is due to a superposition of individual AGN, at a range of L x, z Spectrum of the CXB is hard, cannot be due to unobscured AGN (“Seyfert 1s”) -> but it can be due AGN with a broad range of absorption in addition to a range of L x, z RXTE PCA + HEXTE

11. Astrophysical jets and blazars: what are blazars? Radio-loud quasars, with compact, flat-spectrum radio cores which also reveal some structure The structures often show superluminal expansion Radio, IR and optical emission is polarized Blazars are commonly observed as MeV – GeV  -ray emitters (~ 60 detected by EGRET) In a few objects, emission extends to the TeV range Rapidly variable in all bands including  -rays Variability of  -rays implies compact source size, where the opacity of GeV  -rays against keV X-rays to e + /e - pair production would be large - opaque to their own emission! Entire electromagnetic emission most likely arises in a relativisitc jet with Lorentz factor  j ~ 10, pointing close to our line of sight

12. Example of radio map of a blazar 3C66B

13. EGRET All Sky Map (>100 MeV) Cygnus Region 3C279 Geminga Vela Cosmic Ray Interactions With ISM LMC PKS 0528+134 PKS 0208-512 Crab PSR B1706-44

14. Broad-band spectrum of the archetypal GeV blazar 3C279 Data from Wehrle et al. 1998

15. Example: broad-band spectrum of the TeV-emitting blazar Mkn 421 Data from Macomb et al. 1995

16. Blazars are variable in all observable bands Example: X-ray and GeV  -ray light curves from the 1996 campaign to observe 3C279

17. The “blazar sequence” * Work by G. Fossati, G. Ghisellini, L. Maraschi, others (1998 and on) * Multi-frequency data on blazars reveals a “progression” – * As the radio luminosity increases: * Location of the first and second peaks moves to lower frequencies * Ratio of the luminosities between the high and low frequency components increases * Strength of emission lines increases

18. * What do we infer? We have some ideas about the radiative processes… –Polarization and the non-thermal spectral shape of the low energy component are best explained via the synchrotron process –The high-energy component is most likely due to the inverse Compton process by the same relativistic particles that produce the synchrotron emission –Relative intensity of the synchrotron vs. Compton processes depends on the relative energy density of the magnetic field vs. the ambient “soft” photon field –The source of the “seed” photons for the up-scattering process is diverse - it depends on the environment of the jet BUT – WE STILL DON'T KNOW HOW THE JETS ARE LAUNCHED, ACCELERATED AND COLLIMATED Radiative processes in blazars

19. From Sikora, Begelman, and Rees 1994 Source of the “seed” photons for inverse Compton scattering can depend on the environment It can be the synchrotron photons internal to the jet (the “synchrotron self- Compton” model - This is probably applicable to BL Lac objects such as Mkn 421 Alternatively, the photons can be external to the jet (“External Radiation Compton” model) - This is probably applicable to blazars hosted in quasars such as 3C279

20. Example of an object where ERC may dominate: 3C279 (data from Wehrle et al. 1998) Example of an object where SSC may dominate: Mkn 421 (data from Macomb et al. 1995) SSC or ERC?

21. Moderski, Sikora, GM 2003; Blazejowski et al. 2004 For the External Radiation Compton models, the ultraviolet flux – from Broad Emission Line regions – is not the only game in town… Infrared radiation – specifically, AGN light reprocessed by dust - might also be important, especially in the MeV-peaked blazars (Collmar et al.) Sensitive hard X-ray through soft  -ray observations will be crucial to resolve this, since IR should be Compton-upscattered to energies less than GeV

22. Modelling of radiative processes in blazars In the context of the synchrotron models, emitted photon frequency is s = 1.3 x 10 6 B x  el 2 Hz where B is the magnetic field in Gauss and  el is the electron Lorentz factor The best models have B ~ 1 Gauss, and  el for electrons radiating at the peak of the synchrotron spectral component of ~ 10 3 – 10 6, depending on the particular source Degeneracy between B and  el is “broken” by spectral variability + spectral curvature, at least for HBLs (Perlman et al. 2005) The high energy (Compton) component is produced by the same electrons as the synchrotron peak and compton = seed x  el 2 Hz Still, the jet Lorentz factor  j is ~ 10, while Lorentz factors of radiating electrons are  el ~ 10 3 – 10 6 Thus, one of the central questions in blazar research is: HOW ARE THE RADIATING PARTICLES ACCELERATED?

23. Interpretation of the observational data for blazars PARTICLE ACCELERATION * The most popular models invoke the Fermi acceleration process in shocks forming via collision of inhomogeneities or distinct plasma clouds in the jet (“internal shock” model, also invoked for GRBs) * This can work reasonably well: the acceleration time scale  acc to get electron up to a Lorentz factor  el can be as short as ~ 10 -6  el B -1 seconds, while the cooling time (due to synchrotron losses) is ~ 5 x 10 8  el -1 B -2 seconds, perhaps up to 10 times faster for Compton cooling, so accelerating electrons to  el up to ~ 10 6 via this process is viable (but by no means unique!) INTERNAL SHOCK SCENARIO MODEL * This model assumes that the central source produces multiple clouds of plasma and ejects them with various relativistic speeds: those clouds collide with each other, and the collision results in shock formation which leads to particle acceleration * A simple "toy model" that reproduces observations well assumes two clouds of equal masses, with Lorentz factors  1 and  2 with  1 > 1) * From  2 and  1 one can infer the efficiency (fraction of kinetic power available for particle acceleration) * Recent simulations reproducing well the X-ray light curves of Mkn 421 (and applicable to other objects) (Tanihata et al. 2003) imply that the dispersion of  cannot be too large * However, the small dispersion of  implies a low efficiency – (< 0.1%) so there might be a problem - as huge kinetic luminosities of particles are required… * MY OWN PREJUDICE IS THAT THE JETS ARE LAUNCHED AS MHD OUTFLOWS, AND ARE INITIALLY DOMINATED BY POYNTING FLUX * WE NEED TO UNDERSTAND DISSIPATION/PARTICLE ACCELERATION AS WELL AS THE DISK – JET CONNECTION

24. Diagram for the internal shock scenario – colliding shells model:  2 >  1, shell 2 collides with shell 1 Accretion disk and black hole Broad line region providing the ambient UV Time ->

25. Interpretation of the observational data for blazars PARTICLE ACCELERATION * The most popular models invoke the Fermi acceleration process in shocks forming via collision of inhomogeneities or distinct plasma clouds in the jet (“internal shock” model, also invoked for GRBs) * This can work reasonably well: the acceleration time scale  acc to get electron up to a Lorentz factor  el can be as short as ~ 10 -6  el B -1 seconds, while the cooling time (due to synchrotron losses) is ~ 5 x 10 8  el -1 B -2 seconds, perhaps up to 10 times faster for Compton cooling, so accelerating electrons to  el up to ~ 10 6 via this process is viable (but by no means unique!) INTERNAL SHOCK SCENARIO MODEL * This model assumes that the central source produces multiple clouds of plasma and ejects them with various relativistic speeds * The clouds collide with each other, and the collision results in shock formation which leads to particle acceleration * A simple "toy model" that reproduces observations well assumes two clouds of equal masses, with Lorentz factors  1 and  2 with  1 > 1) * From  2 and  1 one can infer the efficiency (fraction of kinetic power available for particle acceleration) * Recent simulations reproducing well the X-ray light curves of Mkn 421 (and applicable to other objects) (Tanihata et al. 2003) imply that the dispersion of  cannot be too large * However, the small dispersion of  implies a low efficiency – (< 0.1%) so there might be a problem - as huge kinetic luminosities of particles are required… * MY OWN PREJUDICE IS THAT THE JETS ARE LAUNCHED AS MHD OUTFLOWS, AND ARE INITIALLY DOMINATED BY POYNTING FLUX – * WE NEED TO UNDERSTAND DISSIPATION/PARTICLE ACCELERATION AS WELL AS THE DISK – JET CONNECTION

26. Content of the jet Are blazar jets dominated by kinetic energy of particles from the start, or are they initially dominated by magnetic field (Poynting flux)? ( Blandford, Vlahakis, Wiita, Meier, Hardee, …) There is a critical test of this hypothesis, at least for quasar-type (“EGRET”) blazars: If the kinetic energy is carried by particles, the radiation environment of the AGN should be bulk-Compton- upscattered to X-ray energies by the bulk motion of the jet If  jet = 10, the ~10 eV, the H Ly  photons should appear bulk-upscattered to 10 2 x 10 eV ~ 1 keV X-ray flare should precede the  -ray flare (“precursor”) X-ray monitoring concurrent with GLAST observations is crucial to settle this A lack of X-ray precursors would imply that the jet is “particle-poor” and may be dominated by Poynting flux

27. Future of  -ray observations: GLAST EGRET all-sky survey (galactic coordinates) E>100 MeV Features of the MeV/GeV  -ray sky: * Diffuse extra-galactic background (flux ~ 1.5x10 -5 cm -2 s -1 sr -1 ) * Galactic diffuse and galactic sources (pulsars etc.) * High latitude (extragalactic) point sources – blazars and new sources? - typical flux from EGRET sources 10 -7 - 10 -6 cm -2 s -1 * Need an instrument with a good sensitivity and a wide field of view * GLAST – currently under construction - will be launched in 2007

28. GLAST LAT instrument overview e+e+ e–e–  Si Tracker pitch = 228 µm 8.8 10 5 channels 12 layers × 3% X 0 + 4 layers × 18% X 0 + 2 layers Data acquisition Grid (& Thermal Radiators) Flight Hardware & Spares 16 Tracker Flight Modules + 2 spares 16 Calorimeter Modules + 2 spares 1 Flight Anticoincidence Detector Data Acquisition Electronics + Flight Software 3000 kg, 650 W (allocation) 1.8 m  1.8 m  1.0 m 20 MeV – >300 GeV CsI Calorimeter Hodoscopic array 8.4 X 0 8 × 12 bars 2.0 × 2.7 × 33.6 cm  cosmic-ray rejection  shower leakage correction ACD Segmented scintillator tiles 0.9997 efficiency  minimize self-veto LAT managed at SLAC

29. Schematic principle of operation of the GLAST Large Area Telescope   -rays interact with the hi-z material in the foils, pair- produce, and are tracked with silicon strip detectors * The instrument “looks” simultaneously into ~ 2 steradians of the sky * Energy range is ~ 30 MeV – 300 GeV, with the peak effective area of ~ 12,000 cm 2 * This allows an overlap with TeV observatories

30. GLAST LAT Science Performance Requirements Summary ParameterRequirement Peak Effective Area (in range 1-10 GeV)>8000 cm 2 Energy Resolution 100 MeV on-axis<10% Energy Resolution 10 GeV on-axis<10% Energy Resolution 10-300 GeV on-axis<20% Energy Resolution 10-300 GeV off-axis (>60º)<6% PSF 68% 100 MeV on-axis<3.5° PSF 68% 10 GeV on-axis<0.15° PSF 95/68 ratio<3 PSF 55º/normal ratio<1.7 Field of View>2sr Background rejection (E>100 MeV)<10% diffuse Point Source Sensitivity(>100MeV)<6x10 -9 cm -2 s -1 Source Location Determination<0.5 arcmin

31. Sensitivity of GLAST LAT

32. GLAST LAT has much higher sensitivity to weak sources, with much better angular resolution GLAST EGRET

33. GLAST LAT’s ability to measure the flux and spectrum of 3C279 for a flare similar to that seen in 1996 (from Seth Digel)

34. * The future is (almost) here: Next high energy astrophysics satellite: Astro-E2 will be launched in ~ June/July 2005 * Astro-E2 will have multiple instruments: * X-ray calorimeter (0.3 – 10 keV) will feature the best energy resolution yet at the Fe K line region, also good resolution for extended sources (gratings can’t do those!) - but the cryogen will last only ~3 years * Four CCD cameras (0.3 – 10 keV, lots of effective area) to monitor X-ray sources when the cryogen expires * Hard X-ray detector, sensitive up to 700 keV NEAR FUTURE: Astro-E2

35. Principal Investigator is Fiona Harrison (Caltech) the NuSTAR team includes Bill Craig, GM, Roger Blandford at SLAC/KIPAC; Steve Thorsett, Stan Woosley at UCSC; Columbia, Danish Space Res.Inst., JPL, LLNL,

36. It’s the first focusing mission above 10 keV (up to 80 keV) brings unparalleled  sensitivity,  angular resolution, and  spectral resolution to the hard x-ray band and opens an entirely new region of the electromagnetic spectrum for sensitive study: it will bring to hard X-ray astrophysics what Einstein brought to soft X-ray astronomy NuSTAR was recently selected for extended study, with the goal for launch in 2009 (Fiona Harrison/Caltech, PI)

37. The three NuSTAR telescopes have direct heritage to the completed HEFT flight optics. The 10m NuSTAR mast is a direct adaptation of the 60m mast successfully flown on SRTM. NuSTAR det- ector modules are the HEFT flight units. Based on the Spectrum Astro SA200-S bus, the NuSTAR spacecraft has extensive heritage. NuSTAR will be launched into an equatorial orbit from Kwajalein. NuSTAR is based on existing hardware developed in the 9 year HEFT program Orbit 525 km 0° inclination Launch vehicle Pegasus XL Launch date 2009 Mission lifetime 3 years Coverage Full sky Hardware details of NuSTAR

38. NuSTAR science – point sources One of the main goals of NuSTAR is to conduct a census of hard X-ray sources over a limited part of the sky What are the hard X-ray properties of AGN? How do they contribute to the “peak” of the CXB? Spectrum of NGC 4945, a heavily obscured active galaxy

39. Spectrum of the blazar PKS 1127-145 Best probe of the content of the jet will be the hard X-ray / soft gamma-ray observations, simultaneous with GLAST Bulk of the radiating particles is actually at low energies, inferred only from hard X-ray observations

40. Extended sources with Astro-E2 Hard X- ray Detector, and with NuSTAR Besides compact sources such as AGN and binaries, diffuse sources are also great targets: In supernova remnants, hard X- rays might point to the origin of cosmic rays Examples: Cas-A, Kepler on the right Hard X-ray emission from clusters is also expected – via energetic electrons (inferred from radio data) by Compton- scattering the CMB (see Abell 2029 on the right)

41. Thank you!