AGN populations in X-ray surveys Contents Advantage of X-ray - PowerPoint PPT Presentation

AGN populations in X-ray surveys

Contents • Advantage of X-ray surveys • What they find • How many AGN, over cosmic history • X-ray spectra • Winds and outflows • Spectral energy distributions • Star formation in their hosts I’ve been asked to concentrate on Type 1 AGN

Why are X-ray surveys special for AGN? In reality only a few % of an AGN’s power emerges in the X-ray. So why are X-rays so useful? • X-ray emission is ubiquitous in AGN – no X-ray quiet AGN – though the X-rays can be absorbed • For most objects, particularly stars and galaxies, far less than 1% of the power emerges as X-rays. – The contrast between AGN power and other forms of power highest in the X-rays.* *Maybe that’s not true for radio galaxies, but powerful radio emission isn’t universal in AGN.

Brandt & Alexander 2015, A&ARv, 23, 1 Loaring et al. 2005, MNRAS, 362, 1371 Large sky density. Contrast Most sources are AGN

AGN in X-ray surveys • Quasars identified as Uhuru sources. • Seyferts confirmed as X-ray sources with Ariel 5. • Elvis et al. 1978 even made an X-ray luminosity function! Elvis et al. 1978, MNRAS 183, 129

Cosmological evolution • X-ray imaging took AGN luminosity functions to z>2 • Cosmic evolution of AGN population similar in X-rays to optical. • 420 AGN in the Einstein EMSS. Maccacaro et al. 1991, ApJ, 374,117

Into the Rosat era X-rays told us that this break is real Page et al. 1997, MNRAS 291, 324: Boyle, Shanks & Peterson 1988, MNRAS 235, 935: – optical L.F. – X-ray L.F. – peak at z~2 – peak at z~2

Chandra and XMM-Newton era • Evolution models somewhat more complex • Basic shape and evolutionary pattern close to those in Rosat era • Studies try to constrain luminosity function and absorption properties simultaneously • 0.5-2 keV and 2-10 keV bands have different sensitivity to absorption Aird et al. 2015, MNRAS, 451, 1892

Clustering • Clustering analyses tell us (assuming LCDM) about the masses dark matter halos in which AGN reside. • You need a lot of pairs to get a good measurement. • In the 90s X-rays won out at low redshift Boyle & Mo 1993, MNRAS, 260, 925 • But it’s hard to get excited about 90s quality data.

Clustering • Cosmos has good geometry and size for AGN clustering study. • AGN live in > 3x10 10 solar mass galaxies • Findings not radically different to SDSS, or 2Qz though, and constraints much weaker. Gilli et al. 2009, A&A 494, 33

Charting the AGN population: where do X-ray surveys win? • Raw statistical power of optical surveys like SDSS is superior to current X-ray surveys • In clustering studies, I think optical surveys are currently way out in front – I’ll be delighted if you can convince me that this is honestly not the case • Optical luminosity functions better constrained but limited by systematics, particularly towards low luminosities, and in relation to obscuration. – For low luminosity AGN, X-ray surveys win . – We need X-rays to understand how obscuration/absorption affects the luminosity function and its evolution.

AGN X-ray spectra • X-rays come from some kind of corona around the inner part of the accretion disc, very close to the black hole • They carry information about the conditions and geometry of the very innermost part – Which is the most interesting part in physics terms • X-ray spectra may also be related to (and so diagnostic of) the Eddington ratios of AGN.

Soft spectrum AGN, Narrow line Seyfert 1s • When you pick out X-ray sources with very steep spectra, you find a lot of narrow-line Seyfert 1s. • Same when you select in a very soft band (Rosat WFC, PSPC) • This really was a key finding in working out that Puchnarewicz et al. 1991, MNRAS 256, 589 these are high Eddington ratio Seyferts.

Hard spectrum (X-ray absorbed) QSOs Page, Mittaz & Carrera 2001, MNRAS 325, 575 Found using Rosat. Note the absorption lines in the restframe UV. • How come these QSOs are absorbed in X-ray but not in the optical/UV? • XMM-Newton EPIC spectra show that they have ionised winds. (Page et al. 2011, MNRAS 416, 2792)

Similar story for X-ray weak AGN • Searching for AGN which are unusually weak in X-rays shows up an overlapping but related population of absorbed QSOs with ionised winds. • BALQSOs are the extreme end of the distribution. Brandt, Laor & Wills 2000, ApJ 528, 637

AGN X-ray spectra • AGN spectra are roughly power law shaped • Small dispersion in slopes Mateos et al. 2010, A&A 510, A35

AGN X-ray spectra • Stacking (i.e. averaging) AGN spectra has a long heritage. • The Ginga 12 here show that AGN typically have Fe lines and reflection features Pounds et al. 1991, Nature

X-ray spectra • XMM-Newton’s large throughput and lare serendipitous surveys let us do this at cosmological distances.* • Typical AGN at z=1 look rather similar to present day AGN. Corral et al. 2008, A&A 492, 71 *though there are some pitfalls to watch out for in doing this.

AGN spectral energy distributions • Pioneering work by Martin Elvis et al. Elvis et al. 1994, ApJS 95, 1

Star formation in z=1-3 AGN • The black hole/bulge mass relation tells us that the formation of spheroids and black holes are intimately linked. • QSOs had their heyday at z~2. – Most vigorous period of black hole growth. – If black holes and stars grow together, QSOs should also be forming stars rapidly. • Peak of star formation rate also at 1< z < 3.

Energy release from black holes and stars Black holes growing The most rapidly star- by accretion are best forming galaxies are found by X-ray often highly obscured, emission emitting the bulk of their energy in the far infrared Arp 220

Star formation and AGN in the No 250 micron detections: prolific Chandra deep field North star formation is rare in powerful AGN. Significant 250 micron detection fraction at z>1, at moderate AGN luminosities. Many moderate luminosity AGN (~25%) lived in ULIRGs between redshifts of 1 and 3. Page et al. 2012 Nature 485, 213

Rather different picture from stacking analyses • AGN luminosity and star formation appear unrelated at low L. • Star formation evolves with redshift • Star formation and AGN luminosity correlated at high redshift Rosario et al. 2012, A&A 545, A45

The SED is really rather key for disentangling star formation and dust heated by the AGN. • PAHs in mid-IR can be used to give an independent estimate on the star formation rate • Otherwise population properties e.g. covariance between different parts of AGN SEDs the only tool. Symeonidis et al., submitted to MNRAS

So where is the frontier today?

Puchnarewicz et al. 1991, MNRAS 256, 589 Position errors Mason et al. 1995, MNRAS 274, 1194 Association of X-ray sources with optical/ Multiwavelength counterparts is so easy now compared to the old days. Elvis et al. 1978, MNRAS 183, 129

Today’s key questions that X-ray surveys might address • How does the AGN luminosity function behave at low luminosity and high redshift? • What really are the star formation properties of AGN?

Progress with the luminosity function • Low L, high z, i.e. where X-ray sources are faint. • Detecting the X-ray sources is not the (only) limiting problem. • Need high completeness • Must distinguish X-rays from star formation and AGN Aird et al. 2015, MNRAS, 451, 1892

Star formation in AGN hosts • Opposite end of the spectrum, opposite problem. • Distinguish AGN emission from star formation in the infrared. • Current AGN template + SF template approach only gives you what you put in. • One route to a solution might be through statistical analysis of the SEDs. Page et al. 2012, Nature 485, 213

Conclusions • We know an awful lot about AGN through X-ray surveys • The limiting factor right now in understanding AGN through X-ray surveys is how well we can understand the combination of X-ray and multiwavelength data.