1: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
2: %% ws-procs9x6.tex : 10 October 2003
3: %% Text file to use with ws-procs9x6.cls written in Latex2E.
4: %% The content, structure, format and layout of this style file is the
5: %% property of World Scientific Publishing Co. Pte. Ltd.
6: %% Copyright 1995, 2002 by World Scientific Publishing Co.
7: %% All rights are reserved.
8: %%
9: %% Proceedings Trim Size: 9in x 6in
10: %% Text Area: 7.35in (include runningheads) x 4.5in
11: %% Main Text is 10/13pt
12: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
13:
14: %\documentclass[draft]{ws-procs9x6}
15: \documentclass{ws-procs9x6}
16:
17: \def\ls{{_<\atop^{\sim}}}
18: \def\gs{{_>\atop^{\sim}}}
19: \def\cgs{ ${\rm erg~cm}^{-2}~{\rm s}^{-1}$ }
20:
21:
22: \begin{document}
23:
24: \title{High Energy Large Area Surveys:
25: optically obscured AGN and the history of accretion
26: \footnote{\uppercase{T}his work is partially supported by \uppercase{ASI}
27: grant I/R/057/02, by \uppercase{INAF} grant \# 270/2003 and \uppercase{MIUR}
28: grant Cofin--03--02--23}}
29:
30: \author{F. Fiore$^1$ and the HELLAS2XMM collaboration}
31:
32: \address{$^1$ INAF-Osservatorio Astronomico di Roma\\
33: via Frascati, 33, I00040, Monteporzio, Italy\\
34: E-mail: fiore@mporzio.astro.it}
35:
36: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
37: % You may repeat \author \address as often as necessary %
38: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
39:
40: \maketitle
41:
42: \abstracts{Hard X-ray, large area surveys are a fundamental complement
43: of ultra-deep, pencil beam surveys in obtaining a more complete
44: coverage of the L--z plane, allowing to find luminous QSO in
45: wide z ranges. Furthermore, results from these surveys can be
46: used to make reliable predictions about the luminosity (and
47: hence the redshift) of the sources in the deep surveys which have
48: optical counterparts too faint to be observed with the present
49: generation of optical telescopes. This allows us to obtain
50: accurate luminosity functions on wide luminosity and redshift
51: intervals.
52: }
53:
54: \section{Introduction}
55:
56: Hard X-ray surveys are the most direct probe of super-massive black
57: hole accretion activity, which is recorded in the cosmic X-ray
58: background (CXB) spectral intensity. Deep, pencil beam, Chandra and
59: XMM-Newton surveys have resolved most of the CXB below 4-6 keV, and
60: $\approx50\%$ of the CXB at 10 keV (Worsley et al. 2004). The optical
61: spectroscopic follow-up of the Chandra Deep Field North/South (CDFN,
62: CDFS) and of the Lockman Hole (LH) surveys proved to be very efficient
63: at identifying a large population of Seyfert-like objects up to
64: z=2--3, and a few QSOs up to z=4, see Fig. \ref{surveys}a and
65: references in Table 1). Shallower, but larger area surveys are
66: therefore fundamental to: a) complement the pencil-beam survey to
67: obtain a more complete coverage of the L--z plane, to find QSO with
68: logL(2-10keV)$>$44, i.e. close to AGN luminosity
69: function L$^*$, in a large z range; b) to obtain reliable spectroscopic
70: redshifts of the faint optical counterpats of sources with X-ray to
71: optical flux ratio (X/O) much higher than that of typical broad line
72: AGN (X/O$\sim1$), which make 20--30\% of the full samples. At the
73: 10-100 times fainter fluxes reached by the CDFN, CDFS and LH surveys
74: most sources with high X/O have optical counterparts too faint for
75: even 8--10 m class telescopes. The best hope to obtain information on
76: this elusive X-ray source population is to make use of the information
77: gained at higher fluxes to make reliable predictions about their
78: luminosities and redshifts.
79:
80: \subsection{Source samples}
81:
82: Figure \ref{surveys}b) compares the fluxes and area covered by a
83: number of hard X-ray surveys. In this paper we use the source samples
84: given in Table 1, which include the deepest surveys performed with
85: Chandra and the shallower, but much larger area HELLAS2XMM survey. As
86: of today we have obtained spectroscopic redshifts for more than 150
87: sources, and, most important, we were able to obtain spectroscopic
88: redshifts and classification of many sources with X/O$>10$; finding
89: about ten type 2 QSO at z=0.7--2, to be compared with the similar
90: number obtained from the combination of the CDFN and CDFS, at the
91: expenses of a huge investment of VLT and Keck observing time. For
92: other 8 X/O$>10$ sources a redshift estimate was obtained from their
93: observed R--K colors (Mignoli et al. 2004).
94:
95:
96: \begin{table}[ph]
97: \caption{\bf 2-10 keV surveys}
98: {\footnotesize
99: \begin{tabular}{lcccccc}
100: \hline
101: Sample & Tot. Area & Flux limit & \# sour. & \% z-spec & refs & symbol\\
102: & deg$^2$ & $10^{-15}$ cgs & & & &\\
103: \hline
104: HELLAS2XMM & 1.6 & 8.0 & 231 & 66\% & 1,2 & open circ.\\
105: CDFN faint$^a$ & 0.0369& 1.0 & 88 & 59\% & 3,4 & filled squar.\\
106: CDFN bright$^b$ & 0.0504& 3.0 & 44 & 65\% & 3,4 & filled squar.\\
107: CDFS faint$^a$ & 0.0369& 1.0 & 68 & 62\% & 5,6 & stars \\
108: CDFS bright$^b$ & 0.0504& 3.0 & 55 & 58\% & 5,6 & stars \\
109: Lockman Hole$^c$& 0.126 & 4.0 & 55 & 75\% & 7 & filled triang.\\
110: SSA13$^d$ & 0.0177& 3.8 & 20 & 65\% & 8 & filled circ.\\
111: \hline
112: Total & & & 561 & 65\% & & \\
113: \hline
114: \end{tabular}
115: \begin{tabnote}
116: $^a$ Inner 6.5 arcmin radius; $^b$ outer 6.5--10 arcmin annulus; $^c$ inner 12 arcmin
117: radius; $^d$ inner 4.5 arcmin radius; (1) Fiore et al. 2003; (2) Cocchia et al. in
118: prep.; (3) Alexander et al. 2003; (4) Barger et al. 2003; (5) Giacconi et al.
119: 2002; (6) Szokoly et al. 2004; (7) Mainieri et al. 2002; (8) Barger et al. 2001.
120: \end{tabnote}
121: }
122: \label{table1}
123: \vspace*{-13pt}
124: \end{table}
125:
126:
127: \begin{figure}[ht]
128: %\epsfxsize=10cm %width of figure - will enlarge/reduce the figures
129: %\epsfbox{fig3.eps}
130: %\figurebox{2cm}{3cm}{} %to have a box alone
131: \centerline{
132: \hbox{
133: \epsfxsize=2.3in\epsfbox{zlx_lin.ps}
134: \epsfxsize=2.3in\epsfbox{agnsurveys.ps}
135: }
136: }
137: \caption{
138: \footnotesize{
139: Left: the L(2-10keV)--z plane for the source samples in Table
140: 1 (symbols as in Table 1). The lower (upper) solid (dashed) lines
141: represent the flux limits of $10^{-15}$ \cgs (i.e. Chandra deep
142: surveys) and $10^{-14}$ \cgs (i.e. HELLAS2XMM survey)
143: respectively. Right: the flux-area diagram for several 2-10 keV
144: surveys. Open triangles = Chandra surveys, filled triangles = XMM-Newton
145: surveys, stars = Chandra or XMM-Newton ``contiguous'' mosaics}
146: }
147: \label{surveys}
148: \end{figure}
149:
150:
151: \section{Optically obscured AGN: a robust method to estimate
152: their luminosity}
153:
154: Fiore et al. (2003) discovered a striking correlation between X/O and
155: the 2--10 keV luminosity for the sources with the nucleus strongly
156: obscured in the optical band, i.e. not showing broad emission lines
157: (Fig. \ref{zz}a).
158: %The optical R band light of these objects is dominated by the host
159: %galaxy light, which spans a luminosity range much smaller than the
160: %X-ray luminosity, giving rise to the above correlation.
161: %Indeed, the
162: %steep correlations between SMBH mass and bulge light and velocity
163: %dispersion implies that powerful AGN live only in large and luminous galaxies.
164: %On the other hand, the X-ray luminosity depends on both the SMBH mass and
165: %the accretion rate. Its large spread in comparison with the host galaxy
166: %luminosity and the consequent correlation between X/O and the 2-10 keV
167: %luminosity is
168: The dispersion along the correlation in Fig. \ref{zz}a) is of 0.40 dex
169: in luminosity. Note as at X/O$\gs10$ most of the sources are from the
170: HELLAS2XMM sample, i.e. it is mostly thank to this sample that it is
171: possible to extend and validate the correlation at high X/O values,
172: and therefore at high luminosities. From X/O, and hence the
173: luminosity, it is possible to estimate the redshift of optically
174: obscured sources. We call the resulting redshift estimates
175: ``X-photo-z''. Fig. \ref{zz}b) plots our X-photo-z against the
176: spectroscopic redshifts. The correlation is again rather good and
177: $\sigma(\Delta z/(1+z)\sim0.2)$. This can be compared with the value
178: of $\sim0.1$, typical of accurate ``photometric'' redshift estimates.
179: We note that in our case only 2 bands are used, agaist the several
180: O-NIR bands necessary to obtain reliable photo-z. Our present
181: X-photo-z estimates are obtained using a linear fit to the
182: logX/O-logL(2-10keV) relationship. More accurate estimates can be
183: obtained using higher order polinomia (Fiore et al. 2004 in prep.).
184:
185: \begin{figure}[ht]
186: \centerline{
187: \hbox{
188: \epsfxsize=2.1in\epsfbox{ffcorr.ps}
189: \epsfxsize=2.1in\epsfbox{zz.ps}
190: }
191: }
192: \caption{
193: \footnotesize{Left: the X-ray to optical ratio as a function of the X-ray luminosity
194: for 123 optically obscured AGN with secure redshift and classification from
195: the samples in Table 1. The three solid lines are the least squares fit
196: of y=y(x) and x=x(y) and their average.
197: Right: spectroscopic redshift versus the redshifts
198: evaluated using the correlation in the left panel and the X-ray flux.
199: }}
200: \label{zz}
201: \end{figure}
202:
203:
204: \subsection{Optical vs X-ray obscuration: the ``Figure of Merit''}
205:
206: Perola et al. (2004) found that most type 2 AGN in the HELLAS2XMM
207: sample have significant obscuration (rest frame absorbing column
208: $N_H>10^{22}$ cm$^{-2}$) also in the X-rays. Perola et al. (2004) and
209: Comastri \& Fiore (2004) also found a good correlation between X/O and
210: $N_H$. Both facts suggest that X-ray sources log$N_H>22$ are likely
211: to be also optically obscured AGN. However, for many weak sources in
212: the samples in Table 1 $N_H$ proper spectral fits are unfeasible and a
213: rough spectral information can be derived from their softness ratio
214: only. The correlation between the softness ratio and $N_H$ is
215: reasonably good (Perola et al. 2004) and therefore using the former to
216: select X-ray obscured AGN is not a bad approximation. The
217: relationship between X-ray obscured AGN and optically obscured AGN is
218: not one to one. This is clearly seen in Fig. \ref{hrt}b), which shows
219: that $\sim90\%$ of type 1 AGN have (S-H)/(S+H)$>-0.5$ (i.e. most of
220: them have a soft spectrum) and $\sim70\%$ of optically obscured AGN
221: have (S-H)/(S+H)$<-0.5$ (i.e. the majority have a hard, likely
222: obscured spectrum, but $\sim30\%$ have a soft spectrum). This means
223: that using a single (S-H)/(S+H) threshold to decide whether an X-ray
224: source with a faint optical counterpart is an optically obscured AGN
225: has an intrinsically large uncertainty. For this reason, we instead
226: compute for each source a ``figure of merit'' (FoM, Fiore et al. 2004
227: in prep.), using: 1) the X/O ratio and the fraction of
228: obscured/unobscured sources as a function of X/O (see Fiore et
229: al. 2003); 2) the morphology of the optical counterpart, i.e
230: point-like sources are likely to be optically unobscured sources; 3)
231: the probability distribution of the softness ratio for obscured and
232: unobscured AGN, as estimated for the sources with spectroscopic
233: redshift and classification in Table 1. If a source has a FoM
234: qualifying it as optically obscured we use the correlation in
235: fig. \ref {zz}a) to estimate its redshift; if a source qualifies as an
236: optically unobscured AGN we guess its redshift using the loose
237: relation between X-ray flux and redshift for type 1 AGN. These
238: estimates should therefore be considered in a statistical sense only
239: (Fiore et al. 2003). At fluxes lower than $10^{-14}$ \cgs only
240: $\approx30\%$ of the sources qualify as unobscured AGN.
241:
242:
243: \begin{figure}[ht]
244: \centerline{
245: \hbox{
246: \epsfxsize=2.0in\epsfbox{hrthisto.ps}
247: \epsfxsize=2.0in\epsfbox{evol_nn.ps}
248: }
249: }
250: \caption{
251: {\footnotesize
252: Left: the distribution of the softness ratio of optically
253: obscured and unobscured sources, Right: the evolution of the number
254: density of hard X-ray selected AGNs in three luminosity bins: The
255: solid curve represents the evolution of optically selected QSO
256: with M$_B\ls-24$. The three dashed curves are predictions from Menci
257: et al. 2004}
258: }
259: \label{hrt}
260: \end{figure}
261:
262:
263:
264: \section{The evolution of hard X-ray selected sources}
265:
266: Once redshifts and luminosities of the sources of the full sample are
267: known (either spectroscopically measured or photometrically and
268: statistically estimated) we can compute the luminosity function of
269: hard X-ray selected sources and its redshift
270: evolution. Fig.\ref{hrt}b) plots the AGN number density in three
271: luminosity bins as a function of z. This is compared with
272: the number density of luminous optically selected QSOs and with the
273: prediction of the Menci et al. (2004) semi-analytic, hierarchical
274: clustering model (MM). While the density of luminous AGN increases
275: monotonically up to z$\sim3$, following the evolution of optically
276: selected AGN, and in agreement with the MM, the evolution of
277: increasingly lower luminosity AGN peaks at decreasingly lower
278: redshifts. The peak redshifts are also lower than those of the MM
279: predictions. At z$\gs1-2$ the density of logL=$44-44.5$ AGN decreases
280: sharply, and it is inconsistent with the MM predictions, while the
281: density of logL$<44$ AGN is not well constrained by the present
282: data. An extenction of the present analysis to flux limits lower than
283: those in Table 1, and therefore to higher z, is in progress (Fiore et
284: al. 2004 in prep.). The paucity of Seyfert-like objects at z$\gs1-2$
285: can be due to at least two reasons: a) a selection effect, i.e.
286: highly obscured AGN are common at these redshifts (as in the nearby
287: Universe) but are missed (or their luminosity is underestimated) in
288: Chandra and XMM-Newton surveys (La Franca et al. 2004 in prep,; b) a
289: different description of the mechanisms regulating the amount of cool
290: gas in low-mass host galaxies, the physical mechanism at work at small
291: accretion rates and/or the statistics of DM condensations is needed in
292: the MM. To avoid possible selection effect, for an unbiased census of
293: the AGN population making the bulk of the CXB and an unbiased measure
294: of the AGN luminosity function at z=1--2, the "golden epoch'' of galaxy
295: and AGN activity, sensitive observations extending at energies where
296: photoelectic effect no more reduces the observed flux (i.e. E$>10$ keV)
297: are clearly needed. More specifically to resolve $\sim$50\% of the
298: 20--100 keV CXB we need to go down to fluxes of $10^{-14}$ \cgs in
299: this band (see fig. 6 of Menci et al. 2004). This can be achieved
300: only by imaging X-ray telescopes, possibly with multi--layer coatings
301: (see e.g. Pareschi \& Cotroneo 2003). Key issues are: a) high
302: collecting area; b) sharp PSF (15 arcsec or less Half Energy Diameter,
303: HED); c) low detector internal background.
304:
305: \section*{Acknowledgments}
306: {\footnotesize
307: The original matter presented in this paper is the
308: result of the effort of a large number of people, in particular of the
309: {\tt HELLAS2XMM} team (A. Baldi,
310: M. Brusa, F. Cocchia, N. Carangelo, P. Ciliegi, F. Cocchia, A. Comastri,
311: V. D'Elia, C. Feruglio, F. La Franca, R. Maiolino, G. Matt, M. Mignoli,
312: S. Molendi, G. C. Perola, S. Puccetti, C. Vignali), N. Menci, A. Cavaliere,
313: M. Elvis and P. Severgnini.
314: }
315: \begin{thebibliography}{0}
316:
317: \bibitem{}D.M. Alexander et al {\it Astron. J.} {\bf 126}, 539 (2003).
318:
319: \bibitem{}A. Barger et al {\it Astron. J.} {\bf 126}, 632 (2003).
320:
321: \bibitem{}A. Barger et al {\it Astron. J.} {\bf 121}, 662 (2001).
322:
323: \bibitem{}A. Comastri \& F. Fiore astro-ph/0404047 (2004).
324:
325: \bibitem{}F. Fiore et al. {\it Astron. Astroph.} {\bf 409}, 79 (2003).
326:
327: \bibitem{}R. Giacconi et al. {\it Astroph. J. S.} {\bf 139}, 369 (2002).
328:
329: \bibitem{}V. Mainieri et al. {Astron. Astroph} {\bf 393}, 425 (2002).
330:
331: \bibitem{}N. Menci et al. {\it Astroph. J} {\bf 606}, 58 (2004).
332:
333: \bibitem{}M. Mignoli et al. {\it Astron. Astroph.} {\bf 418}, 827 (2004)
334:
335: \bibitem{} G. Pareschi \& V. Cotroneo {\it SPIE Proc.} {\bf 5156} (2003)
336:
337: \bibitem{}G.C. Perola et al. {\it Astron. Astroph.} in press (2004), astro-ph/0404044.
338:
339: \bibitem{}G.P. Szokoly et al. {\it Astroph. J} in press (2004), astro-ph/0312324.
340:
341: \bibitem{}M.A. Worsley et al. {\it MNRAS} submitted (2004), astro-ph/0404273.
342:
343:
344: \end{thebibliography}
345:
346: \end{document}
347: