The Intimate Relationship between High Energy Emission & Relativistic

Jets in Radio-Loud AGNs

Alan MarscherBoston University

Research Web Page: www.bu.edu/blazars

Main Collaborators

Svetlana Jorstad (Boston University) Margo Aller (University of Michigan)Ian McHardy (University of Southampton, UK)Thomas Balonek (Colgate University)Josè Luis Gómez (Instituto de Astrofísica de Andalucia)H. R. Miller, Kevin Marshall, John McFarland (Georgia St. Univ.)Massimo Villata & Claudia Raiteri (Torino Observatory, Italy)Gino Tosti (University of Perugia Observatory, Italy)Valeri Larionov (St. Petersburg State University, Russia)Omar Kurtanidze (Abastumani Observatory, Rep. of Georgia)A. Lähteenmäki, M. Tornikoski, H. Teräsranta (Metsähovi Radio

Observatory, Finland)Martin Gaskell, Elizabeth Klimer (University of Nebraska)Ian Robson (U. Central Lancashire)Andrei Sokolov (Boston U. → U. Central Lancashire, UK)Funded by NASA & NSF

Emission Regions in an AGN

→ Unbeamed optical/uv/X-ray emission from accretion disk + corona→ Unbeamed broad & “narrow” optical emission lines from clouds→ Optical absorption lines from outflows (funnel-shaped?)→ Radio core & jet emission lie outside broad emission-line region→ Beamed X-rays & gamma-rays come from jet

Unbeamed

Beamed

Some Questions on the Nature of AGNsthat can be addressed by multi-waveband variability & VLBI

Observational• How are high-frequency events related to superluminal

radio knots?• Where does the higher frequency emission occur relative

to the radio jet imaged by mm-wave VLBI?• Is there a relation between emission from the central

engine & events seen in the jets?Theoretical• How are jets made by accreting black holes?• Where & how are jets accelerated to high s?• How/where are jets focused into narrow cones?• Where & how are relativistic electrons accelerated?• What is the physics of the jet flow (shocks, turbulence,

interaction with gas, matter content, instabilities, magnetic field pattern, bending, precession)?

Multi-waveband Tools for Addressing Questions

1. Variability (good for all wavebands)• Related to size of emission region• Cross-frequency time lags help to locate sites of

emission

Multi-waveband Tools for Addressing Questions

2. Radio/mm-wave imaging: great resolution• Picture of the jet (with polarization)• Multi-epoch images allow us to watch variations

Multi-waveband Tools for Addressing Questions

3. Polarization• Probes direction & degree of order of magnetic field

Limitations of Tools

• Timescale of variability is related to size but does not give location relative to central engine or core of jet

• Opacity (even at 7 mm wavelength) & resolution limit how close to the central engine VLBI images can explore

Overcoming Limitations: Combine tools

• Variability of SED (can separate variable & non-variable components)

• Variability & VLBI (can see what’s changing directly or see the effects of variations when the disturbance propagates downstream)

• Multi-waveband polarization (can use polarization signatures to identify high-frequency features on VLBI images)

Evidence that High & Low- Nonthermal Emission is Related:

Correlated Light Curves & Polarization Properties

The Quasar PKS 1510-089 (z=0.361)Radio/X-ray Correlation

(Optical not well sampled)

Arrows: Times of superluminal ejections at 43 GHz

The Quasar PKS 1510-089 (cont.)X-ray variations correlated with high- radio, with radio variations leading

X-ray by an average of 6 days

> 0: Radio Leads X-rays originate from radio-emitting part of the jet

Flares occur as superluminal knots emerge from the core

The Quasar 3C 279 (z=0.538)Extreme variations in flux at optical & X-ray (& γ-ray)

Variations in X-ray flux usually well-correlated with changes in optical brightness

. . . but there is not a one-to-one correspondence on time scales of days

X-ray

Optical

Radio

The Quasar 3C 279 (cont.)Variations in X-ray flux usually well-correlated with changes in optical brightness

→ time lag of 15±15 days, with optical leading

Discrete Cross-Correlation Coefficient (X-ray vs. optical)

Positive delay: optical leads

X-ray/8 mm correlation also quite strong, with radio lagging by 14040 days

Either:

X-ray/optical emission comes from ~ 40 pc upstream of core of jet

or:

X-ray/optical emission comes from core & flux of knot peaks 140 days (~0.2 mas) after it leaves core (in favor: superluminal ejections tend to occur near beginning of X-ray flares; against: flux of many knots already decreasing by 0.15 mas from core)

BL Lac (z=0.069)Usually, X-ray continuum spectrum is too flat to be continuation of synchrotron spectrum

Probably synchrotron self-Compton

X-ray

Optical

X-ray spectral index

Strong X-ray/optical correlation of long-term light curve with ~ zero lag

[Correlation not so good on short timescales (Böttcher et al. 2003)]

Period of strongest activity corresponds to steep X-ray spectral index probably high-E tail of synchrotron emission

Superluminal ejection occurred during period of strongest activity

Why Not One-to-One Correspondence between Radio or Optical/IR & X-ray?

SSC: All synchrotron photons that are in the right frequency range to scatter up to a given X-ray energy will contribute equally to the SSC X-ray/γ-ray flux

McHardy et al. 1999 MNRAS→ Relationship between SSC X-ray/γ-ray & synchrotron emission at any given frequency is simple only if source is uniform

→ Frequency-dependent time delays result from(1) light-travel delays (Sokolov & Marscher 2004)(2) frequency stratification (e.g., Marscher & Gear 1985): e-s accelerated at shock front, highest-E e-s “die” quickly, lower-E ones live longer lower ’s emitted over larger volume

How about TeV Blazars?

Thus far, it appears that the radio jets of TeV blazars are rather slow & not highly variable on parsec scales (Marscher 1999, Astroparticle Phys.; Piner 2003 , ApJ)

Maybe they are e--e+ jets with flat electron energy distributions such that the jet loses momentum & decelerates as the electrons cool

Correlated Multiwaveband Polarization → suggests that mm-wave core region = site of optical emission

Blazars: optical & mm-wave linear polarization often has similar E-vector direction (see also Gabuzda & Sitko 1994)

Degree of optical polarization is often higher

Lister & Smith (2000): quasars with low optical polarization also have low core polarization at 43 GHz

Purple: optical pol.

Green: ~ 1 mm pol.

Identifying Knots by Polarization

Many superluminal knots have stable electric vector position angles (EVPAs) of linear polarization

Might be able to identify knots responsible for optical/near-IR variations by similar EVPA as knots on VLBA images

Scale: 1 mas = 4.8 pc = 16 lt-yr (Ho=70)

High γ-Ray Flux Tends to Occur after Ejections of Superluminal Radio Knots (Jorstad et al. 2001 ApJ)

23/28 γ-ray flares are contemporaneous with our VLBA data

11 superluminal jet components are associated with γ-ray flares (≤6 expected by chance)

Example: 3C 273

Increase in polarized radio flux accompanies high γ-ray flux + ejection of superluminal knot (Jorstad et al. 2001 ApJ)

Example: 3C 273

Proposed Method for Identifying Jet Features Responsible for High-frequency Variations

Look for correlated polarization variability in optical, IR, submm, & mm-wave

Objective: Locate sites of variable high-frequency emission on mm-wave VLBA images → relate optical/near-IR light curves to images

Relate to X-ray & -ray by correlating light curves (RXTE, Swift, AGILE, GLAST) with optical/near-IR, & mm-wave light curves

Observational Connection between Jet & Accretion Disk

Microquasars: Change in X-ray state precedes appearance of knots in radio jet (Mirabel et al., Fender et al.)

→ Might expect something similar in AGNs

Need AGNs with blazar-like jets in radio but main X-ray component from accretion disk/corona

→ Blazar-like radio galaxies are good candidates

The FR I Radio Galaxy 3C 120 (z=0.033)

HST image (Harris & Cheung)

Scale: 1 mas =

0.64 pc = 2.1 lt-yr

(Ho=70)

• Superluminal apparent motion, 4-6c (1.8-2.8 milliarcsec/yr)

• X-ray spectrum similar to BH binaries & Seyferts

• Mass of central black hole ~ 3x107 solar masses (Marshall, Miller, & Marscher 2004; Wandel et al. 1999)

Sequence of VLBA images (Marscher et al. 2002)

Superluminal Ejections in 2002-04

Scale: 1 mas =

0.64 pc = 2.1 lt-yr

(Ho=70)

5 bright superluminal knots, apparent motion = 4.5-5c

All other knots are less than 10% of core brightness beyond 0.5 milliarcsec from core

Sequence of VLBA images

Superluminal ejections follow X-ray dips Similar to microquasar GRS 1915+105

Radio core must lie at least 1 lt-yr from black hole to produce the observed X-ray dip/superluminal ejection delay of ~ 60 days

60 days

X-Ray Dips in 3C 120

3C 120 Movie (by J.L. Gómez)Red ellipse: inner accretion disk Contours: 43 GHz intensity

Inset: X-ray light curve

FR II Radio Galaxy 3C 111 (z=0.0485) Seems to Do the Same

Similar to 3C 120 in radio & X-ray 5c (1.5 milliarcsec/yr)

Scale: 1 mas = 0.92 pc = 3.0 lt-yr (Ho=70)May 2004

July 2004

X-Ray Light Curve of 3C 273: 2001-2005(A quasi-blazar, or “quasar”)

Somewhat similar to 3C 120; we will see whether X-ray dips are followed by superluminal ejections. We need first to separate jet & central engine X-rays

Thus far, we see no signs of correlation between radio/mm and X-rays

Previously, McHardy et al. found good K-band/X-ray correlation with ~ 1 day time delay, X-ray lagging IR

Evidence for Collimation of Jets Well Outside Central Engine

• VLBA observations of M87: jet appears broad near core→ Flow may be collimated on scales ~1000 Rs

Junor et al. 2000 Nature

Cygnus A (Bach et al. 2004, 2005)FR II radio galaxy, jet at large angle to l.o.s.

Gap between core & counterjet < 0.7 masApparent speed increases with distance from core

CoreCounter-core

Intrinsic Half Opening Angles of Jets (Jorstad et al. 2005, AJ, submitted)

Intrinsic Half Opening Angles of Jets (Jorstad et al. 2005, AJ, submitted)

1/

Agrees with models in which jet is focused as it is

accelerated over an extended region.(HD:

Marscher 1980; MHD: Vlahakis & Königl 2004)

Jet Acceleration over Extended RegionJet Acceleration over Extended Region

HD: Pressure gradient p r-a Lorentz factor increases with cross-sectional radius R:

Γ R p-1/4 ra/4

If a < 4/(3+1) and viewing angle is small, brightest emission is where Γ reaches its asymptotic value

If viewing angle is large, brightest emission is at lowest r where high-E electrons are accelerated

(Marscher 1980 ApJ)

MHD: Models still being developed

Vlahakis & Königl (2004, ApJ) solution appears similar to HD solution, except that Γ decreases away from jet axis & there is no distinct boundary

In either case, energy density at base of jet must exceed Γρc2

Might require a magnetosphere (pulsar or ergosphere of spinning BH)

Theory: A jet with > ~10 cannot propagate out of nuclear region (Phinney 1987)

Conclusions

Combining multi-waveband techniques is a powerful – but painful! - way to explore the most interesting physics of AGNs. The effectiveness is limited mainly by the amount of telescope time + astronomers’ time available

Thus far, each source behaves differently → we need to follow more objects with even better time coverage to probe the jet and its relation to the central engine properly

In PKS 1510-089 the X-rays come from the radio-emitting section of the jet; the same seems to be true for gamma-rays from most EGRET sources.

In 3C 279 the X-rays (& gamma-rays) come from the optically emitting region, which seems to lie well upstream of the VLBI core of the jet.

The core in the blazar-like radio galaxies 3C 120 & 3C 111 lies > 0.4-1 pc from the central engine; drops in X-ray flux occur as disturbances shoot down the jet

We need better radio, IR, & optical monitoring to take full advantage of AGILE, GLAST, VERITAS et al. (Cherenkov telescopes), & other monitoring instruments

End of Talk

Remaining slides are to be shown only if they will be useful for discussion after the talk

Sketch of 3C 120 Nucleus

Eikenberry & van Putin (2003), Livio et al. (2003), Tagger et al. (2004): More efficient outflow into jet when turbulent field becomes more poloidal

→ Gas in disk falls into BH more quickly, more advective, radiates less (dip)

X-rays mainly from corona/wind, some might be from jet as well

Popular model: jet propelled by twisted poloidal magnetic field emanating from disk or ergosphere

Comparison of GRS1915+105 with 3C 120 Light Curves

BH mass of 3C 120 ~2x106 times that of GRS 1915+105, so timescales of hours to months in the former are similar to the scaled-up quasi-periods (0.15 to 10 s) & duration of X-ray dips in the latter.Typical fractional amplitude of dips is also similar Long, deep dips not yet seen in 3C 120

blow-up

← GRS 1915+105 over 3000 s on 9/9/97Light curve (top) & PSD (bottom)(Taken from Markwardt et al. 1999 ApJL)

The “Microquasar” GRS 1915+105: A Faster-than-light Object in Our Galaxy

Mirabel & Rodriguez (1994 Nature)Apparent velocity 2-3c, but jet makes a large angle to line of sight Γ~5 (Fender et al. 1999)

Binary system, giant star + black hole of 14 solar masses

Ejection of superluminal knots follows end of low, hard X-ray state

3 different types of behavior (Belloni et al. 2000)

Evidence for Acceleration of Jet on Parsec Scales

Accelerating Jet

Acceleration of proper motion near core in some jets

A jet with > ~10 cannot propagate out of nuclear region (Phinney 1987)

How then can we have jets with observed apparent motions > 25c, implying > 25 and focusing to within 0.5 degrees or less?

→ Models in which jet accelerates & collimates out to pc-scale radio core (Marscher 1980, ApJ; Vlahakis & Konigl 2004, ApJ)

The Radio Galaxy NGC1052 (z=0.0049)

Apparent speed of radio knots ~ 0.25c

Iron line at 6.4 keV had more pronounced “red” wing prior to radio ejection event than at two other epochs

M. Kadler et al. (2004)

A major radio ejection event may be preceded by enhanced inflow in the relativistic region near the black hole in NGC1052

The iron line in 3C 120 is probably variable based on widely diverse reports of equivalent width; we are analyzing our RXTE observations in an effort to measure variability of the line & search for occasional broad wing

3C 120 ASCA spectrum (Grandi et al. 1997)

BL Lac (z=0.069)

Jet does a hula dance:

~ 2-yr cycle of 24° swing in direction of jet near core

Scale: 1 mas = 1.3 pc = 4.1 lt-yr (Ho=70)