Particle acceleration in active galaxies – the X-ray view

Martin Hardcastle (U. Herts)

Thanks to many co-authors including Ralph Kraft (CfA), Judith Croston (Herts), Diana Worrall (Bristol)

X-ray Universe, Granada, 28th May 08

Overview

• Motivation

• Types of radio galaxy

• Key role of X-ray synchrotron

• Results from low-power radio galaxies

• Results from high-power radio galaxies

• Particle acceleration mechanisms

Motivation (1)

We observe synchrotron radiation, implying acceleration of high-energy leptons: we want to relate what we observe to energy transport mechanisms.

Motivation (2)

We observe high-energy cosmic rays implying high-energy baryonic acceleration: this must always be accompanied by acceleration of leptons so by studying one we can understand the other.

Types of radio galaxy

• In the radio, FRIs have centre-brightened structures often dominated by bright jets

• FRIIs have edge-brightened structures often with prominent hotspots.

• FRIs have low radio luminosity. FRIIs have high radio luminosity.

• FRI/FRII difference implies different jet physics.

Hotspot

Core

Jet

Hotspot

Lobe

FRIFRII

Plume

X-ray synchrotron

• For radio and even optical synchrotron radiation we cannot distinguish between particles that have been accelerated upstream and advected to where we see them, & particles genuinely accelerated where we are looking.

• For X-ray synchrotron the loss timescales are so short that particles can travel only a few pc from their sites of acceleration: so effectively for the distances involved X-ray synchrotron emission tells us where particle acceleration is happening now.

FRI jets

• FRI jets started showing up in large numbers soon after the launch of Chandra (Worrall et al 01; MJH et al 01).

• X-ray spectra mostly consistent with extrapolation of radio-optical => synchrotron origin generally assumed.

FRI jets

• X-ray emission is diffuse => can no longer sustain a picture of a single acceleration location.

• X-ray (+ optical) spectrum is steep, Γ>2.0 – not consistent with the Heavens & Meisenheimer continuous injection model used for hotspots.

• High-energy particle acceleration appears to be associated with bulk jet deceleration.

Jet deceleration

Laing et al 2002a, 2002b; MJH et al 2002

Particle acceleration process

• Derives its energy from the jet deceleration process (no problem with energetics).

• Distributed throughout the jet• Averaged over the jet, produces a flat radio

spectrum and steep X-ray spectrum, with a break in the IR/optical.

• We need to be able to resolve the particle acceleration process on the loss spatial scale to see whether it is genuinely diffuse or just distributed. Only possible in the nearest FRI, Cen A.

Cen A (Chandra)

720 ks of Chandra data, including a Chandra VLP (PI Ralph Kraft). See Kraft et al 2002, MJH et al 03, MJH et al 06, Kataoka et al 06, MJH et al 07, Jordán et al 08, Sivakoff et al 08, Worrall et al 08, Kraft et al 08 for some Chandra results.

Key results on Cen A jet

1) Strong point-to-point radio/X-ray ratio variation – particle acceleration efficiency varies spatially

2) Compact X-ray emitting knots are stationary in the radio maps – could be shocks?

MJH et al 2003

Key results on Cen A jet

3) Diffuse X-ray emission comes to dominate at large distances from the nucleus.

4) X-ray spectra of knots are flat: X-ray spectrum of diffuse emission gets progressively steeper ending at very high values: X-ray surface brightness falls off faster than radio.

MJH et al 2007 ApJL

Particle acceleration processes

• We suggest that the spatial and spectral differences between the compact ‘knots’ and diffuse emission means that there are two acceleration processes going on in Cen A.

• The compact knots may be shocks producing X-ray-emitting electrons by first-order Fermi.

• The diffuse emission surrounding them is probably something else! – return to this later.

FRII radio galaxies

• Hotspots in FRII radio galaxies are the physical manifestations of the jet-termination shock, so we expect first-order Fermi acceleration at the hotspot. Consistent with early work on broad-band SEDs.

Meisenheimer et al 1989 Blandford & Rees 1974

X-rays from FRII hotspots

• Early work on X-ray detections of hotspots focussed on objects that radiate by the synchrotron-self-Compton (SSC) mechanism (e.g. Harris et al 1994, MJH et al 01). Won’t discuss this here.

• Increasingly it’s become clear (MJH et al 04; Kraft et al 05) that some hotspots’ X-ray emission can’t be explained by an inverse-Compton model but must be synchrotron instead.

X-rays from FRII hotspots

Colours are radio emission, green contours show X-rays

FRII hotspots

Colours are radio emission, green contours show X-rays

Problems with the standard picture

So we can use X-ray synchrotron to locate particle acceleration in FRII hotspots too.

But the results deviate from our expectations in three ways:

1) When we see X-rays coincident with radio/optical hotspots, although we sometimes see the sort of smoothly steepening spectra that we expect from radio-through-optical observations, we often don’t:

3C33

Kraft et al 2006

3C33

Problems with the standard picture

1) (continued) i.e. if we want this emission process to be synchrotron emission, we have either to have an electron energy spectrum that turns up at high energies, or we have to abandon our one-zone model of the electron population.

2) We often see spatial offsets between the peaks of the radio/optical/X-ray emission:

3C227

MJH, Croston & Kraft 2007

Problems with the standard picture

2) (continued) – this more or less requires us to abandon the one-zone picture, but what do we put in its place?

3) We often see diffuse X-ray emission, implying distributed particle acceleration, throughout bright hotspots – completely inconsistent with the idea that particle acceleration is taking place at localized shocks.

3C390.3

MJH, Croston & Kraft 2007

Hotspot consequences

• Problem #1 means that either the hotspots are not homogeneous, or they are not accelerating particles in the expected way, contrary to what the radio/optical spectra told us.

• Problem #2 means that if particle acceleration is localized at shocks, the shocks are, at least some of the time, not where the peak of the radio emission is, contrary to what everyone has always assumed.

• Problem #3 means that at least some of the particle acceleration is not localized at shocks anyway, contrary to the standard picture.

Acceleration mechanisms

• Requirement for a diffuse acceleration mechanism (#3) is interesting because it may be related to the one seen in FRIs.

• What diffuse acceleration processes are there?• Turbulence (second-order Fermi acceleration) and shear

provide possible sources for particle acceleration in jets (e.g. Stawarz & Ostrowski 2002). Unfortunately in resolved FRI jets like Cen A there is no evidence for systematic edge-brightening or flatter spectra in the diffuse emission at the edge.

• Magnetic field line reconnection (e.g. Birk & Lesch 2000) is promising but makes few testable predictions about the spectrum of high-energy particles.

Summary

• The traditional picture that all particle acceleration in radio-loud AGN happens via a first-order Fermi process at shocks needs major revision.– In FRIs, shocks may contribute but don’t dominate– In FRIIs, where we know shocks should be present,

the model fails to explain much of what we see.

• In particular we need a distributed particle acceleration mechanism which can extend over tens of kpc.

• We need testable predictions from existing models!