Comptonthick AGN in the CDFN I Georgantopoulos NOA

Compton-thick AGN in the CDFN I. Georgantopoulos NOA A. Akylas NOA A. Georgakakis NOA M. Rovilos MPE M. Rowan-Robinson Imperial College

Why are Compton thick AGN important ? § X-ray surveys are extremely efficient in detecting AGN: the sky density of AGN is 5000 deg-2 (Bauer+04) orders of magnitude higher than that in the optical (eg Wolf+ 03) § However, at very high obscurations (NH~1024 cm-2 or AV>>100) even the hard 2 -10 ke. V X-rays cannot penetrate these columns (Compton-thick AGN). § C-T are extremely important for the peak of the X-ray background at 40 ke. V (Gilli et al. 2007) § Spitzer mid-IR surveys are claiming the detection of AGN which are not detected in X-ray § The French/Italian Simbol-X, Japanese Ne. XT have as their main science goals the detection of these sources at very high energies.

Plot adapted from Gilli et al. Simbol-X Present sample

Talk Outline A) X-ray spectral analysis in the CDF-N B) X-ray Stacking analysis of mid-IR selected AGN not detected individually in the X-ray

X-ray spectral analysis

The sample Selected all CDFN sources in the 2 -10 ke. V band with flux > 1. e-15 cgs ie an order of mag above the flux limit Two reasons: i) very good photon statistics spectra ii) Large number of redshifts (spec and photo) 222 s�ources of which 190 have redshift (107 spect) The photo-z have been derived using the code of Babbedge et al. using the IRAC Mags where available. Median number of photons ~200

Method C-T AGN can be detected either: a) Absorption turn-over below 10 ke. V which shifts at lower energies at high redshifts because of K-correction b) a flat spectrum Γ~1 (eg Matt 1996)

Spectral fits a) single power-law model (WA*PO) with the NH free and ��Γ=1. 8 b) single power-law model with the photon index free and the NH fixed to the Galactic. Then one can compare the values of chi-2 for the two models We require Γ<1. 2 (at the 90% upper limit) and Δχ2 > 5 corresponding to a probability value of about 2σ that the ABSORBED model is better than the FLAT one [Likelihood ratio test Mushotzky 1982]

9 Compton thick sources 8 flat spectrum C-T candidates directly at z=2 PLUS 1 for which we can see the spectral turnover 1 has no redshift (not in GOODS), 6/8 spectroscopic Redshifts Mean redshift z=2. 0 log. Lx=42 -43 (observed) Four of them are sub-mm sources (Alexander+2006)

ACS z-band cutouts of the C-T AGN

Fe. Kα

The C-T log. N-log. S Our points Gilli predictions

Mid-IR selection

The Stern diagram: non X-ray detections Is there a population of mid-IR AGN not detected in Xray ? Georgantopoulos+08 astro-ph/ GOODS IRAC sources AGN

X-ray stacking analysis One can increase the effective exposure time by adding (stacking) the photons in areas of non-detected sources. 126 mid-IR ‘AGN’ at mean redshift z=1. 4 Detection in both the soft (0. 3 -1. 5 ke. V) and hard (1. 5 -4 ke. V) band but with a soft spectrum Γ~2. 1 Moreover, - Lx/LIR ~ 10 -5 - Lx ~ 6 x 1040 cgs (soft band) further supporting the galaxy scenario.

Optically-faint & red sources (a la Fiore selection) Criteria: R-[3. 6] > 3. 7 RED (EROs) 24 μm/R > 1000 optically FAINT 103 sources of which 20 detected in X-rays The stacking analysis of the 83 gives Γ~0. 8 or NH= 8 x 1022 cm-2 (@z=2) Lx/LIR ~ 10 -5 These are either C-T or low luminosity absorbed AGN

Optically faint/ red sources: mid-IR diagram Many sources outside the Stern wedge where the galaxy and Sy 2 templates converge at z>2 X = x-ray detection • = non-detection

ACS z-band cutouts of the X-ray detections in the selection ‘a la Fiore Green circle = 3. 6 micron position

X-ray Absorbed sources are associated with earlytype systems at z~1. Does the same hold at higher redshifts ? X-ray obscured sources are RED at z~1 Rovilos & Georgantopoulos 07 (ECDFS) Georgakakis+08 (AEGIS) See also Silverman+07 is there another tendency at higher z ?

Remarks Compton-thick candidate AGN can be easily detected through X -ray spectroscopy. XMM observations could then play the major role Advertisement of the 1. 3 CDFS XMM observation (PI Comastri) of the CDFS which will provide many bona-fide C-T AGN at faint fluxes At brighter fluxes the fraction of C-T AGN is >>1/100. Analysis of the whole 2 XMM catalogue should produce at least a few tens In the future all the pieces of the puzzle will be filled at faint 210 ke. V fluxes with XEUS and observations at hard energies with Simbol-X and NEXT

THE END

Stern diagram : X-ray detections Stern+05 wedge is a good way to sift mid-IR AGN But not all X-ray AGN fall into the Stern wedge see also Barmby+06