Blazar Variability & the Radio Galaxy/Cosmology Interface

1. Blazar Variability &the Radio Galaxy/Cosmology Interface Paul J. Wiita Georgia State University, Atlanta, USA Winter School on Black Hole Astrophysics APCTP, Pohang, January 17-20, 2006

2. OUTLINE • Blazar Basics • Accretion Disks in AGN Recent Evidence for their Presence Basic Timescales A Few Important Instabilities Spiral Shocks • Aspects of Jet Produced Variations Coherent Emission Slow Knot Speeds vs. Ultrarelativistic Jets • Radio Galaxies Trigger Extensive Star Formation Spread Magnetic Fields and Metals into IGM

3. Blazar Characteristics • Rapid variability at all wavelengths • Radio-loud AGN • BL Lacs show extremely weak emission lines • Optical polarization  synchrotron domination • Double humped SEDs: RBL vs XBL? • Core dominated quasars (optically violently variable and high polarization quasars) clubbed w/ BL Lacs to form the blazar class • Population statistics indicate that BL Lacs are FR I RGs viewed close to jet direction (Padovani & Urry 1992)

4. Long-term Blazar Lightcurve(Optical monitoring at Colgate U.- Balonek)

5. Long-term Radio MonitoringAller & Aller, U Michigan

6. Microvariability & Intraday Variability too!Romero, Cellone & Combi; Quirrenbach et al (2000)

7. Blazar SED: 3C 279 (Moderski et al. 2003) Left hump: peak in mm or FIR, from synchrotron Right hump: peak in gamma-rays, from Inverse Compton off seed photons: From disk, from jet itself or from broad line clouds

8. Orientation Based Unification Picture

9. Evidence for Accretion Disks in BlazarsBig blue bump in AO 0235+164(Raiteri et al. astro-ph/0503312)

10. More New Evidence for Accretion Disks Optically thick: hidden Balmer edge now claimed to be seen in several quasars. • Ton 202 polarized flux with face-on Kerr disk model fitted to it (Kishimoto et al. 2004)

11. Why Quasi-Keplerian and Disk-like?Quasi-black body fits to disk spectraBroad K lines for NLS1sVariable Double peaked lines [here H lines: Strateva et al, AJ (2003)]Jets probably require disks as launching pads

12. Accretion Flow Geometries • Quasi-accepted picture: L/LE determines disk thickness and extent toward BH: very high L/LE geometrically & optically thick intermediate L/LE  cold optically thick, geometrically thin low L/LE  optically thin hot flow interior to some transition radius.

13. Key Timescales for Accretion • With R = r/3RS, a quasi-Keplerian flow, h the thickness and  the viscosity parameter, the fastest expected direct variations are on dynamical times of hours for SMBHs (e.g. Czerny 2004). M8=MBH/108M Radial sound transmission time Thermal and viscous timescales For thin disks, h0.1r

14. How fast can the cold disk be removed? • Transition radius changes, either by evaporation or substantial outflow • Either way, disk T must go up to about virial T and enough energy to do this must be stored • For an -disk, tevap=tvisc, but more generally, For AGN > 103yr, so if disk appears to disappear quickly, probably from suppression of energy dissipation (I.e., MRI instability turned off, perhaps by some ordered B field.)

15. Longest Timescale? • Governed by rate at which outer disk is fed • Probably the rate at which gas is injected into the core of a galaxy (bars within bars to drive inward?) • Dominated by galactic mergers (probably major) and timescales > 107 years; can exceed 108 yr Does harassment (mere passage) work? • Does the AGN self-regulate, with its energy injection halting the inflow of gas? (Hopkins et al. 2005a,b,c) • Most likely depends on whether quasi-isotropic winds & star-burst supernovae OR narrow jets carry off most kinetic energy from AGN.

16. DISK INSTABILITIES •  many of them. How many are important, especially for blazars? • Radiation pressure instability • Magneto-rotational instabilities • Flares from Coronae • Internal oscillatory modes (diskoseismology) • Avalanches or Self-Organized Criticality • Spiral shocks induced by companions or interlopers • Key point: even if blazar emission dominated by jets, disk instabilities may feed into jets

17. Radiation Pressure Instability Long known that -disks are unstable if radiation pressure dominated (Shakura & Sunyaev 1976) • AGN models should be Prad dominated over a wide range of accretion rates and radii • Computed variations are on tvisc(~100RS) (Janiuk et al. 2000; Teresi et al 2004) • May have been seen in the microquasar GRS 1915+105 (over 100’s of sec). • Scaled to AGN masses: significant outbursts, but over years to decades all the way from X-rays through IR.

18. SPH simulation of Shakura-Sunyaev instability (Teresi, Molteni & Toscano, MNRAS 2004)

19. MRI Induced Variations • Magneto-Rotational Instabilities (e.g. Balbus & Hawley ApJ, 1991) are commonly agreed to be present • Probably produce effective disk  ~ 0.01-0.10 Total (solid), magnetic stress (dashed) and fluid (dotted) viscosities at a disk center (Armitage 1998, ApJL)  Also produce changes in dissipation and accretion rate  Some disk clumping, but not destruction (profile changes?)

20. Turbulence in a Magnetized Disk Distant views of inner disk @ inclinations of 55 and 80o • Integrated flux for inclinations of (top to bottom) 1, 20, 40, 80O for a “hot” simulation using Zeus and pseudo-Newtonian potential • (Armitage & Reynolds, MNRAS 2003) • Significant fluctuations develop on a few rotational timescales (hours for 108M).

21. Spiral Shocks in Disks • Perturbation by smaller BH can drive spiral shocks • Significant flux variations ensue on orbital timescales of the perturber (Chakrabarti & Wiita, ApJ, 1993) Perturbers w/ 0.1 and 0.001 MBH

22. Spiral Shocks and Line Variations • This type of shock provides the best fits to changes in double hump line profiles seen in about 10% of AGN(Chakrabarti & Wiita 1994) Model vs. data for 3C 390.3 H broad lines in 1976 & 1980. Expected variations.

23. Flares and Coronae • Plenty of debate over the relative contribution of disk coronal flares to X-ray (predominantly) and other band (secondarily) emission and variability. • Clearly an important piece of the Seyfert variability but probably usually a small piece of blazar emission. • Total energy releasable from low density coronal flares is probably too small unless “avalanche” or self-organized criticality process is triggered, perhaps propagating inward within a disk (Mineshige et al. 1994; Yang et al. 2000); easily produces “correct” PSD. • But flares can provide low level X-ray variations visible when other activity is minimal; maybe produce a bit of optical variability too.

24. Jet Variations in Blazars • This is the dominant idea, but it still is not well modeled. SOMETHING changes: outflow rate, velocity, B-field structure. Waves can steepen into shocks. • Relativistic shocks propagating down jets can explain much of the gross radio through optical variations via boosted synchrotron emission. Accretion disk fluctuations could drive them. • Turbulence, instabilities, magnetic inhomogeneities can probably explain the bulk of rapid variations. • Inverse Compton models: SSC, External Compton, Mirror Model , Decelerating Jets, can explain particular high energy variations wrt low energy ones, though no model seems able to cover all observations (multiple IC photon sources?)

25. Shock-in-jet model: new components(Aller, Aller & Hughes 1991)

26. Turbulence in a Jet  Rapid Variations(Marscher & Travis 1996)

27. Synchrotron vs. Coherent Emission • Do any compact radio sources show intrinsic TB>1011K? (More realistic self-absorbed source equipartition inverse Compton catastrophe limit ~3x1010; Singal & Gopal Krishna 1985; Readhead 1994) • IDV at cm  big Lorentz factor is necessary (if intrinsic) as simple measurements often give TB~1021K • To avoid it, a size ~ larger is allowed if plasma approaches us with >> 1. So solid angle up ~2. • TB intrinsic boosted by  wrt source frame so total help of 3 available: BUT still need ~103 for enough help • Such huge ’s prevent too many X-rays, but at the cost of low synchrotron radiative efficiencies and thus demand very high jet energy fluxes (Begelman et al. 1994)

28. But what really produces radio IDV? • It seems most IDV is due to refractive interstellar scintillation(e.g., Kedziora-Chudczer et al. 2001) • Then TB,intrinsic~1013K, so 30 solves this problem • However, space VLBI couldn’t resolve many of these sources, so TBcould be much higher (Kovalev) • A recent claim that the blazar J1819+3845 shows diffractive scintillation    10as and TB,intrinsic>(>)1014K (Macquart & de Bruyn 2005) • If true, it demands >103 if incoherent synchrotron emission is the radiation mechanism, and the energy problem returns

29. Coherent Radiation Could Solve Problems • If strong Langmuir turbulence develops in AGN jets then coherent mechanisms can produce needed huge TB without requiring extreme Lorentz factors(e.g., Baker et al. 1988, Krishan & Wiita 1990, Benford 1992). • One possibility: a pump field can be scattered off a collective mode of a relativistic electron beam: Stimulated Raman Scattering; for a density n, area A, electron Lorentz factor and bunching fraction  For n~109cm-3, ~103, A~1032 cm2, ~0.5: Lo ~1046 erg/s BUT: problems with absorption of masers hard to solve

30. What Type of Coherent Radiation? • Above models implicitly assumed plasma> cyclotron but some only required mild population inversions. • Begelman, Ergun and Rees (2005) have argued that the opposite, c p is more likely in blazar jets. • Employ small-scale magnetic mirrors, arising from hydromagnetic instabilities, shocks or turbulence: any could provide good conditions for numerous transient cyclotron masers to form • Current into mirror inhibits motion of e’s along flux tube. Maintaining current demands parallel E field and accelerates e’s Accelerated e’s along converging flux tubes  population inversion needed for cyclotron maser Maser pumped by turning kinetic and magnetic energy into jE work • Synchrotron absorption is serious but high TB maser photons can escape from a boundary layer giving TB,obs ~ 2x1015K (/10)4 R

31. Magnetic Mirror & Cyclotron Maser Current carrying magnetic mirror on quasi-force-free flux rope. Parallel E field maintains electron flow through mirror. Parallel potential + magnetic mirror turns initial electron distribution into a horseshoe shaped one (shell in 3-D) Conditions: mirror ratio R=5, Current Jzm=30mA/m2 (Jz0=6mA/m2); Epar=500 keV, consistent w/ Te=100 keV & n=100 cm-3 (Begelman et al. 2005)

32. Modest Superluminal VLBI Speeds • Only semi-direct probe of extragalactic jet speed: VLBI knot apparent motions: > 30% subluminal for TeV blazars (Piner & Edwards 2004; Giroletti et al. 2004) •  low ~2-4 contradict usual blazar estimates & IDV 1ES 1959+650 @ 15 GHz 3 epochs Natural (top) vs. Uniform (bottom) weighting (Piner & Edwards 2004) vapp=-0.1 +/- 0.8 c

33. TeV Blazars want High Doppler factors • To avoid excessive photon-photon losses variable TeV emission demands ultrarelativistic jets(Krawczynski et al. 2002) with 15<  < 100 • Taking into account IR background absorption strongly implies 45 <  needed in “unreddened” emission (e.g. Kazanas; Wagner) • Evidence for TB,intrinsic > 1013K in IDV sources would also imply > 30 • While rare (Lister), some vapp > 25c components are seen (Piner et al.) in EGRET blazars. • Substantial apparent opening angles are seen for some transversely resolved knots. • GRB models usually want  > 100

34. How to Reconcile Fast Variations with Slow Knots? • Spine-sheath type systems: fast core gives variations via IC and slower outer layer seen in radio (Sol et al. 1989; Laing et al. 1999; Ghisellini et al. 2004) • Rapidly decelerating jets between sub-pc (-ray) and pc (VLBI knot) scales(Georganopoulos & Kazanas 2003) • Viewing angles to within ~1o could work in an individual case; but  too many slow knots. • Differential Doppler boosting across jet of finite opening angle can make the weighted probable vapp surprisingly small(Gopal-Krishna, Dhurde & Wiita 2004) • Motions can reflect pattern, not physical, speeds

35. Conical Jets w/ High Lorentz Factors Weighted app vs  for  = 100, 50, 10 and opening angle = 0,1,5 and 10 degrees, with blob 3 boosting Probability of large app can be quite low for high  if opening angle is a few degrees

36. High Gammas Yet Low Betas • app vs  for jet and prob of app >  for opening angles = 0, 1, 5, 10 degrees and  = 50, 10 (continuum 2 boosting) • Despite high  in an effective spine population statistics are OK • Predict transversely resolved jets show different app

37. Finding Jet Parameters • Determining bulk Lorentz factors, , and misalignment angles, , are difficult for all jets • Often just set  =1/ , the most probable value • Flux variability and brightness temperature give estimates: S = change in flux over time obs Tmax= 3x1010K app from VLBI knot speed  is spectral index

38. Conical Jets Also Imply • Inferred Lorentz factors can be well below the actual ones • Inferred viewing angles can be substantially underestimated, implying deprojected lengths are overestimated • Inferred opening angles of < 2o can also be underestimated • IC boosting of AD UV photons by ~10 jets would yield more soft x-rays than seen (“Sikora bump”) but if >50 then this gives hard x-ray fluxes consistent with observations • So ultrarelativistic jets with >30 may well be common

39. Inferred Lorentz Factors inf vs.  for =100, 50 and 10 for =5o P() and < inf>

40. Inferred Projection Angles • Inferred angles can be well below the actual viewing angle if the velocity is high and the opening angle even a few degrees • This means that de-projected jet lengths are overestimated

41. Radio Lobes in the Quasar Era The dramatic rise in both star formation rate and quasar densities back to z > 1 motivates investigation of a possible causal connection. Radio lobes affected a large fraction of the cosmic web in which galaxies were forming at 1.5 < z < 3 • Most powerful radio galaxies (RGs) are only detectable for a short fraction of their total lifetimes, so the volumes filled by old, invisible, lobes are extremely large. • The co-moving density of detected RGs was roughly 1000 times higher at 2 < z < 3 than at z = 0. • These RG lobes need only fill much of the "relevant universe", the denser portion of the filamentary structures containing material that is forming galaxies, not the entire universe; much easier for these ‘rare’ AGN!

42. RGs Suffer Restricted Visibility All recent models of RG evolution (Kaiser et al. 1997; Blundell et al. 1999 -- BRW; Manolakou and Kirk 2002; Barai & Wiita 2006) agree that radio flux declines with increasing source size because of adiabatic losses, and with redshift because of inverse Compton losses off the CMB. Jets of power Q0, through a declining power-law density, n(r); has total linear size D; with a0 the core radius (10 kpc), n0 the central density (0.01 cm-3), and  = 1.5. Many properties of low frequency radio surveys (3C, 6C, 7C) can be fit if typical RG lifetimes are long (up to 500 Myr) and if the jet power distribution goes as Q0-2.6 (BRW). For RGs at z > 2, most observable lifetimes () are only a few Myr, even if the jet lifetimes (T) are 100s of Myr

43. P-D Tracks for Different Models(Barai & Wiita 2005)

44. Radio Luminosity Functions Powerful (FR II) RGs were nearly 1000 times more common between redshifts of 2 and 3 (Willott et al. 2001). RLF is flat for about a decade in radio power P151 > 1025.5 W Hz–1 sr–1 , where FR II sources are most numerous. With the correction factor (T/ ~ 50) we find at z = 2.5 the proper density of of powerful RGs living for T is  ~ 4 x 10–5(1+z)3 T5 Mpc–3 ( log P151)–1 with T5 = T/(5 x 108 yr). Integrate over the peak of the RLF and take into account generations of RGs over the 2 Gyr length of the quasar era. We find (Gopal-Krishna & Wiita 2001) the total proper density of intrinsically powerful RGs is about:  = 8 x 10 –3 Mpc-3

45. Radio Luminosity FunctionWillott et al. 2001:FR II vary much more than FR I

46. Models & Data Agree Adequately(B&W for MK)

47. The Relevant Universe The web of baryons traced by the WHIM at z=0 in a 100 Mpc3 box (Cen 2003) RGs nearly all form in these filaments and so most of the radio lobes will be confined to them

48. Radio Lobes Penetrate the Relevant Universe During the quasar era, only a small fraction of the baryons had yet settled into the proto-galactic cosmic web: roughly 10% of the mass and 3% of the volume (Cen & Ostriker 1999). Thus RG lobes have a big impact if they pervade only this filamentary "relevant universe", with volume fraction ~ 0.03. Assuming BRW parameters and integrating over beam power and z, we find the fraction of the relevant universe filled during the quasar era by radio lobes: = 2.1 T518/7–1 (5/RT)2, is > 0.1 if T > 250 Myr and RT (RG length to width) ~ 5.

49. Overpressured Lobes Can Trigger Extensive Star Formation RG lobes remain significantly supersonic out to D > 1 Mpc. Their bow shocks will compress cooler clouds within the IGM (e.g., Rees 1989; GKW01), triggering extensive star formation. Much of the "alignment effect" (McCarthy et al. 1987) is thus explained. Recent numerical work that includes cooling (Mellema et al. 2002; Fragile et al. 2003, 2004) confirms that RG shocked cloud fragments become dense enough to yield massive star clusters (Choi et al. 2006). Hence, RGs may accelerate the formation of new galaxies and in some cases produce them where they wouldn’t have formed in the standard picture.

50. Jet/Cloud Interaction Simulation When cooling is included powerful shocks leave behind dense clumps that can yield major star clusters (Mellema et al.)