Fermi results on AGN: Insights from multi-band observations Greg Madejski (KIPAC/SLAC) With Masaaki Hayashida, Marek Sikora, Jun Kataoka, Lukasz Stawarz for the Fermi-LAT collaboration

Projection in Galactic coordinates Fermi sky after two years: Projection in Galactic coordinates As expected, Fermi is detecting many blazars A number (all?) of those are showing rapid variability, flare-like behavior A number of papers on individual sources is in the press First list of gamma-ray bright AGN is in press; spectral and variability papers published -------------------------- Blazars are sources of radiation in nearly all measured bands: many are known for their strong TeV gamma-ray emission What is the nature/structure of blazars? Approach to their study is like peeling an onion: study the broad-band spectrum and variability, understand the radiation process, deduce the process that energizes the radiating particles, infer the ultimate energy source -> relativistic jet pointing close to our line of sight powered by accretion of matter onto a black hole in the center of the host galaxy

Multi-band observations Multi-band observations are essential: variability! Participating observatories : more than 20 instruments Gamma-ray: > 200 MeV X-ray Fermi-LAT (uninterrupted monitoring) * Includes optical polarization data from Kanata (in Japan) and KVA (in La Palma) * TeV observatories RXTE-PCA Swift-XRT 3 3

Consensus regarding the emission processes in all blazars Pre-Fermi broad-band spectrum of blazar 3C454.3 (more on it later!) General spectral behavior: two broad peaks: - one in radio – to optical bands - another in X-ray to gamma-ray bands Best picture for the emission processes: - Low energy peak emission is via synchrotron process High energy peak is produced by inverse Compton process Big questions: what is the structure of the jet? Location of radiation region?

Population properties: spectral diversity Gamma-ray spectra in the Fermi band are quite diverse (based on the 3-month “LAT Bright AGN Survey” (LBAS): Abdo et al. 2009) <- High luminosity sources low luminosity sources -> soft spectra hard spectra Clear correlation of g-ray spectrum with blazar sub-class This has strong implications on contribution of blazars to the diffuse extragalactic gamma-ray background: still work in progress, but we expect both types would contribute at some level

Fermi LAT spectra of blazars in the context of broad-band spectra - Clear spectral diversity! FSRQs all FSRQs Log nFn Log n BLLacs Number of sources <G>= 2.33+/-0.01 BL Lacs Those are commonly detected as TeV sources <G>=1.99+/-0.01 Photon index

Origin of spectral diversity of blazars / implication on their structure Fermi+broad-band data confirm the “blazar sequence” higher luminosity sources have respective synchrotron and Compton peaks at lower energies and vice versa What does it tell us about radiating particles? Electrons producing spectral peaks of HBL blazars have higher energy than those in FSRQ blazars One hint: * HBLs have relatively lower luminosity than FSRQs * Both classes have high black hole mass, ~ 109 Mo -> HBLs have lower luminosity in Eddington units -> lower accretion rate in Eddington units “Blazar sequence” (Fossatti et al. 1998)

Origin of spectral diversity of blazars / connection to the accretion properties If HBLs have indeed lower luminosity in Eddington units -thus lower accretion rate in Eddington units – this implies: * Luminous, quasar-type, FSRQ blazars have cold, luminous accretion disks * Low-luminosity, HBL-type blazars have hot, advective, underluminous accretion flows This scenario can explain the difference in electron energies in the two classes: * Maximum electron energy is determined by the competition of acceleration and cooling of electrons * In FSRQs, luminous accretion disk provides ample external photons that are in turn seeds for Compton-upscattering by energetic electrons -> Maximum electron energy is limited * In HBL blazars, such external photon field is much weaker – cooling less effective - -> electrons can attain higher energies

Gamma-ray spectra of BL Lac type blazars Examples of broad-band g-ray spectra of HBL-type blazars: Mkn 421, PKS 1553 * Those are often TeV sources Spectra hard, roughly connect to that measured by TeV instruments – but a sharp break! - highly variable - need to measure spectra simultaneously Gamma-ray spectrum of PKS 1553 (above) and gamma-ray spectrum & MAXI light curve of Mkn 421 (right)

Gamma-ray variability properties of blazars PKS 1502+106 Flares generally symmetrical, occasionally faster rise time than decay time Nearly-continuous, well-sampled g-ray light curves allow a robust measurement of Power Density Spectra of blazars PDS well-described as a power law with an index 1.5 +/- 0.2

Mis-aligned blazars are also g-ray emitters NGC 1275, a.k.a. 3C84 or Perseus A, is a nearby radio galaxy containing a flat-spectrum, compact (VLBI-scale) blazar-like, variable radio source It has been detected in the Fermi LAT data, at a much higher level than the upper limit from EGRET -> variable (on ~ year time scales) It is located in the Perseus cluster of galaxies, but variability clearly excludes the association of g-rays with the cluster Jet is not pointing close to the line of sight Seyfert-like characteristics of the nucleus * g-ray emission also seen from “mis-aligned” jets Fermi discovered several more, both FR-I and FR-II Luminosity generally low, ~1045 erg/s The jet associated with the active galaxy is most likely responsible for inflating the “cavities’’ seen in the Chandra images of the Perseus cluster Fermi spectra of FR-II Luminosity distribution of gamma-ray emitting AGN Variability of NGC 1275 in the LAT band

FSRQ 3C 279 (z = 0.536) One of the EGRET-brightest AGN g-ray X-ray Apparent luminosity: as high as 1048 erg/s Can vary in g-ray flux by roughly two orders of magnitude > 100 GeV emission detected by MAGIC Black hole mass: ~ 6 x 108Msolar Radio VLBI image shows superluminal motion: Gjet ~ 15, angle to line of sight ~ 2o © NRAO / AUI / NSF Wehrle et al. 2001 Adopted EGRET-era emission model radio- optical : synchrotron X-ray : SSC Gamma-ray : EC (disk+BEL) 1996 flare g-ray X-ray 12 12

g-ray photon index (LAT) g-ray (LAT) g-ray photon index (LAT) Light curves covering ~ a year starting July 2008 (Abdo et al. 2010, Nature 463, 919; more recent data in a few VG) X-ray optical-UV optical polarization degree (PD) optical polarization angle (EVPA) Near-Infrared Lack of mm variability, via synchrotron self-absorption, constrains the size of the g/optical emission region : Rb < 5 x 1016 cm Radio

Gamma-ray flare with polarization change The event lasts for ~ 20 days 20 days flux vs. polarization degree (PD) - g-ray flare coincides with the drop of PD fluxes decrease, followed by a temporary recovery of PD clear correlation: - a highly ordered magnetic field g opt. PD polarization angle (PA or EVPA) during the low PD, it gradually decreases by 208 deg (~12 deg/day) -> non-axisymmetric structure of jet a perpendicular shock moving along the axisymmetric jet and viewed at a small but constant angle to the jet axis EVPA The rotation of optical polarization angle seems to be a common feature of blazars (R. Itoh, M. Uemura posters) – conclusions might be more general than just 3C279

Location of the dissipation region g-ray flare correlated with polarization change : duration ~ 20 days -> constrains on the location of the g-ray/optical emission region 1. helical magnetic field model (Marscher et al. Nature 2008) 2. curved jet model Gjet Gjet location of the emission region : ~ 105 gravitational radii Numerical simulations by McKinney + Blandford 2008) In both scenarios the coherent polarization event is produced by a emission region co-moving along the jet. the emission region propagate as fast as the bulk speed of the jet. In addition, we also consider another scenario, so called “flow-through” scenario, where emission pattern may move much slower than the bulk speed of the jet. This model involves swing (like wobbling) of the jet associated with jet instabilities such that jet boundary moves relative to our line of sight. In this case the time scale for the observed variation is the time scale for the jet motion. 3. “flow-through” (“quivering jet”) scenario (emission pattern may move much slower than the bulk speed of the jet) emission region can be located closer (x G2jet) to BH : revent ~ cDtevent ~ 103 gravitational radii

Best model for the jet structure? Comparison to previous EVPA observations (Larionov et al. 2008) (our results) Scenarios 1 and 3 imply the rotation of EVPA should always rotate the same direction -> But, not in the observations: needs further confirmation (1) helical magnetic field (2) curved jet model (3) flow-through the twist presumably originates in the inner accretion disk. For scenarios 3, it is also necessary to have some more theoretical work to investigate the jet instability by simulation or so on. This is the next step for this topic Gjet © Nature, vol436, 887 (2010) A.Young Favored (with caution)

Spectral Energy Distribution 2.38+/-0.08 (number: photon index) 1.62+/-0.10 2.57+/-0.10 1.70+/-0.13 Red Blue * g-ray emission is dominant and most variable - several x 1047 erg/s at the peak comparable photon indices between two states With emission region “far,” Comptonization of the dust-originating IR radiation is favored (Sikora+ 2009)

Isolated X-ray flare g-ray X-ray a significant, symmetrical flare at 54950 MJD (~60 days after the g-ray flare) the optical and g-ray emission is in a lower state the X-ray spectrum is much harder than the optical spectrum likely to be generated by IC scattering of low energy electrons duration ~ 20 days; similar profile to the g-ray flare X-ray flare cannot be a simple delayed version of the g-ray flare (due to, e.g, particle cooling) Still, the g-ray emission is dominant; 5 times higher the energy flux even during the isolated X-ray flare event challenge to the simple, one-zone emission models

So – what’s been happening to 3C279 lately? Relatively less active than last year – but some activity in g-rays and X-ray band - It is too close to the Sun for optical monitoring, but that is restarting

3C454.3 with LAT Vital statistics: * Well-known radio, EGRET gamma-ray source - an OVV quasar at z = 0.859 * Good multi-epoch VLBI data, superluminal exp., d = 25, Gjet ~15, q ~0.8o * Very active (bright, variable) since 2000 Shows rapid flares, with the risetime on a scale of ~3 days - compact emission region Can’t be optically thick to the escape of g-rays via e+/e- pair production -> another way to determine Doppler factor Here, d > 6 – consistent with the VLBI-measured jet geometry Snapshot of the broad-band spectrum for Aug 7-12 2008

3C454.3 with LAT: g-ray spectrum Fermi LAT data: g-ray spectrum not a simple power law It steepens to higher energy – can be described as a broken power law with a break, G1 ~ 2.3 to G2 ~3.5 at Ebr ~ 2 GeV * Origin of the spectral break? Not a simple “cooling break” (expected Da=0.5) – more likely a signature of intrinsic break in the electron distribution, with the g of electrons radiating at the break of ~ 3000 Similar breaks are seen in many of the high luminosity (FSRQ) blazars The cooling time scales are quite short, much shorter than the source crossing time - imply distributed acceleration throughout the jet volume

Progress on the modelling front Gamma-ray spectra of many bright blazars show relatively sharp spectral break (by Da ~ 1), similar to 3C454.3 (Abdo et al. 2010, paper on blazar spectra) This has been interpreted as possibly due to Klein-Nishina effect via drop of cross-section for Compton scattering (Finke & Dermer 2010) Here the IC “seed” photons are BLR Detailed modelling (applied to 3C454.3 flare) indicates that the break produced by KN is via upscattering of BLR photons not sufficiently sharp (Abdo et al. 2010) Break in the energy spectrum of the radiating particles is still the best scenario 3C454.3 g-ray spectrum with KN model in red (Abdo et al. 2010)

Late-breaking news on the 3C454.3 front 3C454.3 has been flaring again in mid-November * In the Fermi LAT band it was the most luminous source in the sky (besides gamma-ray bursts): isotropic luminosity of almost 1050 erg/s * Active in other bands: currently probably at the historical high in X-rays (Swift) Recent light curves of 3C454.3 in the Fermi band (left) and Swift band (right) Should be detectable by MAXI!

Summary Long-term multi-band observations including Fermi-LAT Fermi sensitivity allows detailed studies of jet-dominated AGN: blazars, mis-oriented jets (radio galaxies) Clear differences between luminous FSRQs and low-luminosity HBLs: differences probably governed by the accretion rate For FSRQ objects as bright as 3C279 or 3C454.3, Fermi-LAT can resolve the g-ray activity on daily scales * In the gamma-ray emitting FSRQs, g-ray band is energetically dominant * 3C279: g -ray flare associated with optical polarization change lasting ~ 20 days evidence for the presence of highly ordered magnetic filed non-axisymmetric trajectory of the emission pattern constraints on the location of the dissipation region Isolated X-ray flare challenges the simple, one-zone emission models Spectral break seen in the g-ray band not due to KN effects 3C454.3 – current “record holder” for most luminous source in the Universe Long-term multi-band observations including Fermi-LAT and optical polarization are essential for investigating the structure of quasar jets