Radio-loud AGN energetics with LOFAR

1. Radio-loud AGN energetics with LOFAR Judith Croston LOFAR Surveys Meeting 17/6/09

2. Understanding radio-galaxy physics is important for galaxy feedback models! • X-ray cavity measurements show energy is available to balance cooling in cluster cores, but timescales uncertain + various detection biases. • When central AGN switches off, up to ¾ of available energy still contained within radio lobes – subsequent evolution of lobe contents & impact on the cluster depend on cavity particle & B content. • FRIs (typical cluster centre sources) and powerful FRIIs have different energetics and particle/field content (e.g. JC et al. 2004, 2005, 2008; Dunn et al. 2004, 2005; Kataoka & Stawarz 2005): understanding the origins of this difference is crucial for relationship between accretion mode, jet production and feedback. JC et al. 2003 Wise et al. 2007

3. Radio-galaxy energetics, particle & field content • k is unknown, and in general B and N0 can’t be disentangled: common to assume minimum energy/equipartition. • The main exception is when inverse-Compton emission from the same electron population can be detected: typically true for FRII radio galaxies and quasars. • Measurements of external pressure/X-ray cavity detections can also constrain ETOT (rule out equipartition in FRIs). • Emin and shape of N(E) below observable radio region are important: low-energy electrons dominate relativistic particle population.

4. The low-energy electron population • Most of the energy density in extragalactic radio sources is at energies below currently observable radio region. • Radio-source properties depend strongly on assumed spectrum below ~ 300 MHz: alow and gmin. • See discussion in Harris (2004, astro-ph/0410485) Figures from Harris (2004)

5. Inverse-Compton emission from FRII radio lobes • IC X-ray emission breaks the ne/B degeneracy of radio synchrotron => direct probe of low-energy electron spectrum and of lobe energetics. • IC useful in jets & hotspots too, but for lobes beaming & other X-ray emission processes unimportant. • In most cases CMB photon field dominates over nuclear photons (e.g. Brunetti et al. 1997) & SSC . • Can now routinely detect IC emission from the lobes of FRII radio galaxies and RL quasars: ~ 30 X-ray detections spanning redshifts of 0.006 – 2. Colour: XMM IC Contours: radio JC et al. 2004

6. Comastri et al. 2003, Hardcastle et al. 2002, Brunetti et al. 2002, Isobe et al. 2002, Hardcastle & JC 2005, JC et al. 2004

7. IC/CMB from FRII lobes: results for large samples X-ray detected lobes • X-ray detection in at least one lobe in 70% of X-ray observed 3C FRIIs • Consistent with IC/CMB with B = (0.3 – 1.5) Beq • > 75% of sources at equipartition or slightly electron dominated => magnetic domination must occur rarely, if at all. • Unlikely that relativistic protons dominate source energetics. • Total internal energy in FRII radio sources is typically within a factor of 2 of minimum energy (see also Kataoka & Stawarz 2005) • But assumptions about the low-energy electron population introduce significant uncertainty in these results... Lower limits for non-detected lobes JC et al. 2005 ApJ 626 733

8. Low-energy electron distribution • Assume cut-off frequency, gmin = 10 • in hotspots, gmin ~ 100 – 1000 required (e.g. Carilli et al. 1991) • adiabatic expansion => lower energy electrons in lobes • Assume spectral index, alow = 0.5 (flattening) • shock acceleration models predict d = 2 – 2.3 (a = 0.5 – 0.7) • + hotspot observations (e.g. Carilli et al. 1991, Meisenheimer et al. 1997) • If alow = aobs: • increase in Utot of up to factor of 20 • Bobs/Beqdecreases by up to 60%, • IC/nuclear++ • If gmin = 1000 (instead of 10): • Utot and Bobs/Beq unchanged • IC/nuclear -- • conclusions not affected • If gmin = 1: • increase in Utot by ~25% • small decrease in Bobs/Beq • IC/nuclear ++

9. Spatially resolved IC studies • Chandra & XMM allow us to investigate spatial variation of N(E) and B in lobes. • Lack of correlation between radio and X-ray structure indicates N(E) changes alone can’t explain radio structure; changes in B alone can’t explain relation to radio spectral structure => both are required. • Also relies heavily on assumptions about low-n spectrum... Isobe et al. 2002 Hardcastle & JC 2005 Goodger et al. 2008 & in prep.

10. X-ray environments & cluster cavities • FRI radio lobes at equipartition are under-pressured relative to their environments (e.g. Morganti 1988, Killeen et al. 1988, Feretti et al. 1990, Taylor et al. 1990, Böhringer et al. 1993, Worrall et al. 1995, Hardcastle et al. 1998, Worrall & Birkinshaw 2000, JC et al. 2003, Dunn & Fabian 2004, JC et al. 2008, Birzan et al. 2008) • Either radiating particles & field are NOT at equipartition or some other particle population dominates the source energetics. Dunn & Fabian 2004 MNRAS 355 862 Worrall & Birkinshaw 2000 ApJ 530 719

11. Combined X-ray & radio constraints favour entrainment of ICM • Fraction of energy in radiating particles decreases dramatically with distance:. • These constraints rules out relativistic proton domination, electron dominance and simple B-dominated models (e.g. Nakamura et al. 2006, Diehl et al. 2008) • Consistent with entrained, heated ICM dominating radio-lobe energetics. • Good constraints for models of FRI entrainment, but this relies on assumptions about low-energy electron population... 3C 31: required entrainment rates Hydra A: “missing” pressure as a function of distance 1.4 keV 5 keV 10 keV 50 keV model (1+r/rc) -2.0 (1 + r/rc) -1.0 r const. Comparison with theoretical expectations JC et al. in prep

12. Calibrating radio-loud (FRI) feedback • X-ray cavities provide direct measurement of energy input to ICM: Ekin >> Esynch (e.g. Bîrzan et al. 2004, Dunn & Fabian 2004, Dunn & Fabian 2008) • Cavity detection only possible for modest sample sizes at low/moderate z and is subject to incompleteness problems: depends on angle to l-o-s, X-ray data quality, cluster luminosity, etc. • Feedback models require radio surveys of FRIs to high z to relate direct measurements of energy input to RL AGN population statistics. • Low-n radio spectrum promising for reducing large scatter in cavity scaling relations (Bîrzan et al. 2008) Bîrzan et al. 2008

13. What LOFAR will do • 10-200 MHz observations of large samples of radio-loud AGN will determine distributions of low-n spectral index (& cut-off in some cases) for different radio-loud AGN populations. • Low-n spectra for large samples of FRIIs with X-ray coverage (100+ FRIIs): • determine electron energy distribution for the energetically dominant population below g ~ 105 via X-ray IC; constraints on particle acceleration • remove factor ~ 20 uncertainty in ETOT , factor ~2 uncertainty in B assuming CMB dominates IC photon seed field in most cases, and uncertainty about the role of nuclear IC scattering • Low-n spectra for very large samples of FRIs, including cavity sources, will: • Remove > order of magnitude uncertainty in energetics of radiating particles & field in FRIs/cluster cavities: important to determine entrainment and heating rates. • Allow detailed calibration of AGN heating relations via low-n observations of cavity samples at low-z • Apply new calibrations to comprehensive FRI samples for tightly constrained AGN feedback models