1: \documentclass[12pt,preprint]{aastex}
2: %\documentclass[12pt,referee]{aastex}
3: %\documentclass[manuscript]{aastex}
4: \usepackage{emulateapj5}
5: %\oddsidemargin=0pt
6: %\evensidemargin=0pt
7: %\textwidth=6.5truein
8: %\baselineskip=20pt
9: \newcommand{\mincir}{\raise
10: -3.truept\hbox{\rlap{\hbox{$\sim$}}\raise4.truept\hbox{$<$}\ }}
11: \newcommand{\magcir}{\raise
12: -3.truept\hbox{\rlap{\hbox{$\sim$}}\raise4.truept\hbox{$>$}\ }}
13: \newcommand{\minmag}{\raise
14: -3.truept\hbox{\rlap{\hbox{$<$}}\raise5.truept\hbox{$<$}\ }}
15: \newcommand{\be}{\begin{equation}}
16: \newcommand{\ee}{\end{equation}}
17: \newcommand{\ba}{\begin{eqnarray}}
18: \newcommand{\ea}{\end{eqnarray}}
19: \newcommand{\brr}{\begin{array}}
20: \newcommand{\err}{\end{array}}
21: \newcommand{\bc}{\begin{center}}
22: \newcommand{\ec}{\end{center}}
23: \newcommand{\tR}{\mbox{\tiny R}}
24: \newcommand{\tM}{\mbox{\tiny M}}
25: \newcommand{\tN}{\mbox{\tiny N}}
26: \newcommand{\tL}{\mbox{\tiny L}}
27: \newcommand{\lb}{{\left<\right.}}
28: \newcommand{\rb}{{\left.\right>}}
29: \newcommand{\hm}{\,h^{-1}{\rm Mpc}}
30: \newcommand{\vel}{\,{\rm km\,s^{-1}}}
31: \newcommand{\nnabla}{{\bf $\nabla$}}
32: %\renewcommand{\baselinestretch}{1.0}
33:
34: \newenvironment{inlinefigure}{%
35: \def\@captype{inlinefigure}%
36: \noindent\begin{minipage}{\linewidth}\begin{center}}
37: {\end{center}\end{minipage}\smallskip}
38: \makeatother
39:
40: %\renewcommand{\baselinestretch}{1.0}
41:
42: \shorttitle{Luminosity Dependent X-ray AGN Clustering}
43: \shortauthors{M. Plionis et al.}
44:
45:
46: \begin{document}
47:
48: \title{Luminosity Dependent X-ray AGN Clustering ?}
49:
50: \author{M. Plionis,\altaffilmark{1,2}, M. Rovilos\altaffilmark{1,3},
51: S. Basilakos\altaffilmark{4}, I. Georgantopoulos\altaffilmark{1},
52: F. Bauer\altaffilmark{5} }
53: \altaffiltext{1}{Institute of Astronomy \& Astrophysics,
54: National Observatory of Athens, I.Metaxa \& B.Pavlou,
55: P.Penteli 152 36, Athens, Greece}
56: \altaffiltext{2}{Instituto Nacional de Astrofisica,
57: Optica y Electronica (INAOE)
58: Apartado Postal 51 y 216, 72000, Puebla, Pue., Mexico}
59: \altaffiltext{4}{currently at: Max-Planck-Institut für extraterrestrische
60: Physik, Giessenbachstaße, Postfach 1312, 85748 Garching, Germany}
61: \altaffiltext{4}{Academy of Athens, Research Centre for Astronomy \& Applied
62: Mathematics, Soranou Efessiou 4, 11-527 Athens, Greece}
63: \altaffiltext{5}{ Columbia Astrophysics Laboratory, Columbia University,
64: Pupin Physics Laboratories, 550 West 120th St., New York, USA}
65:
66:
67: \begin{abstract}
68: We have analysed the angular clustering of X-ray selected active
69: galactic nuclei (AGN) in different
70: flux-limited sub-samples of the {\em Chandra} Deep Field North (CDF-N)
71: and South (CDF-S) surveys. We find a strong dependence of the
72: clustering strength on the sub-sample flux-limit, a fact which explains most of
73: the disparate clustering results of different XMM and {\em Chandra}
74: surveys.
75: Using Limber's equation, we find that the inverted CDF-N and
76: CDF-S spatial clustering lengths are consistent with direct spatial clustering
77: measures found in the literature,
78: while at higher flux-limits the clustering length increases
79: considerably; for example, at
80: $f_{x,{\rm limit}}\sim 10^{-15}$ erg s$^{-1}$ cm$^{-2}$
81: we obtain $r_0\simeq 17 \pm 5$ and $18\pm 3 \; h^{-1}$ Mpc, for the
82: CDF-N and CDF-S, respectively.
83: We show that the observed flux-limit clustering trend hints towards
84: an X-ray luminosity dependent clustering of X-ray selected, $z\sim 1$, AGNs.
85:
86: \end{abstract}
87:
88: \keywords{galaxies: active --- quasars: general --- surveys --- cosmology:
89: observations --- large-scale structure of the universe}
90:
91: \section{Introduction}
92: X-ray selected AGNs provide a relatively unbiased census of the
93: AGN phenomenon, since obscured AGNs, largely missed in optical surveys,
94: are included in such surveys.
95: Furthermore, they can be detected out to high redshifts and thus trace
96: the distant density fluctuations providing important
97: constraints on supermassive black hole formation,
98: the relation between AGN activity and Dark Matter (DM) halo hosts,
99: the cosmic evolution of the AGN phenomenon (eg. Mo \& White 1996,
100: Sheth et al. 2001),
101: and on cosmological parameters and the dark-energy
102: equation of state (eg. Basilakos \& Plionis 2005; 2006, Plionis \& Basilakos
103: 2007).
104:
105: Until quite recently our knowledge of X-ray AGN clustering came
106: exclusively from analyses of ROSAT data ($\le 3$keV)
107: (eg. Boyle \& Mo 1993; Vikhlinin \& Forman 1995; Carrera et al. 1998;
108: Akylas, Georgantopoulos, Plionis, 2000; Mullis et al. 2004).
109: These analyses
110: provided conflicting results on the nature of high-$z$ AGN
111: clustering. Vikhlinin \& Forman (1995), using the angular correlation approach
112: and inverting to infer the spatial correlation length, found a strong amplitude
113: of $\bar{z}\sim 1$ sources ($r_{0}\simeq 9 \; h^{-1}$ Mpc), which translates
114: into $r_{0}\simeq 12 \; h^{-1}$ Mpc for a
115: $\Lambda$CDM cosmology and a luminosity
116: driven density evolution (LDDE) luminosity function (eg. Hasinger et al. 2005).
117: Carrera et al. (1998), however,
118: %investigated the spatial clustering directly
119: using spectroscopic data, could not confirm such a
120: large correlation amplitude.% for the high-$z$ X-ray AGNs.
121:
122: With the advent of the XMM and {\em Chandra} X-ray observatories,
123: many groups have attempted to settle this
124: issue. Recent determinations of the high-$z$ X-ray selected AGN
125: clustering, in the soft and hard bands, have provided
126: again a multitude of conflicting results, intensifying the debate
127: (eg. Yang et al. 2003; Manners et al. 2003;
128: Basilakos et al. 2004; Gilli et al. 2005; Basilakos et al 2005;
129: Yang et al. 2006; Puccetti et al. 2006; Miyaji et al. 2007; Gandhi et al.
130: 2006; Carrera et al. 2007).
131:
132: In this letter we investigate these clustering
133: differences by re-analysing the CDF-N and CDF-S surveys,
134: using the Bauer et al. (2004) classification to select
135: only AGNs in the 0.5-2 and 2-8 keV bands.
136: To use all the available sources, and not only those having
137: spectroscopic redshifts, we work in
138: angular space and then invert the angular correlation
139: function using Limber's equation.
140: Hereafter, we will be using $h\equiv H_{\circ}/100$ km $s^{-1}$ Mpc$^{-1}$.
141:
142: \section{ The X-ray source catalogues}
143: The 2Ms CDF-N and 1Ms CDF-S {\it Chandra} data represent the deepest
144: observations currently available at X-ray wavelengths (Alexander et al. 2003,
145: Giaconni et al. 2001).
146: The CDF-N and CDF-S cover an area of 448 and 391 arcmin$^{2}$,
147: respectively. We use the source catalogues of Alexander et al. (2003)
148: for both CDF-N and CDF-S.
149: The flux limits that we use for the CDF-N
150: are $3\times 10^{-17}$ and $2\times 10^{-16}$
151: $\rm erg~cm^{-2}~s^{-1}$ in the soft and hard band,
152: while for the CDF-S the respective values are
153: $6\times 10^{-17}$ and $5\times 10^{-16}$ erg cm$^{-2}$ s$^{-1}$.
154: Note that sensitivity maps were produced following the prescription of
155: Lehmer et al. (2005) and in order to produce random catalogues in a
156: consistent manner to the source selection, we discard sources which
157: lie below our newly determined
158: sensitivity map threshold, at their given position.
159: Our final CDF-N catalogues contain 383 and 263 sources
160: in the soft (0.5-2 keV) and hard band (2-8 keV),
161: respectively, out of which 304 and 255 are AGNs, according to
162: the ``pessimistic'' Bauer et al. (2004) classification.
163: The corresponding CDF-S catalogues
164: contain 257 and 168 sources in the same bands, out of which 227 and
165: 165 are AGNs.
166: A number of sources (roughly half) have spectroscopic redshift determinations
167: (mostly taken from Barger et al. 2003;
168: Szokoly et al. 2004; Vanzella et al. 2005; Vanzella et al. 2006;
169: Le F\'evre et al. 2004; Mignoli et al. 2005).
170:
171: \section{Correlation function analysis}
172: \subsection{The angular correlation}
173: The clustering properties of the X-ray AGNs are
174: estimated using the two-point angular correlation function, $w(\theta)$,
175: estimated using $w(\theta)=f(N_{DD}/N_{DR})-1$,
176: where $N_{DD}$ and $N_{DR}$ is the number of data-data and data-random
177: pairs, respectively, within separations $\theta$ and $\theta+d\theta$.
178: The normalization factor is given by
179: $f = 2 N_R /(N_D-1)$, where $N_D$ and $N_R$ are the total number of
180: data and random points respectively. The Poisson uncertainty in
181: $w(\theta)$ is estimated as $\sigma_{w}=\sqrt{(1+w(\theta))/N_{DR}}$
182: (Peebles 1973).
183:
184: The random catalogues are produced to account for the different
185: positional sensitivity and edge effects of the surveys. To this end
186: we generated 1000 Monte Carlo random realizations of the
187: source distribution, within the CDF-N and CDF-S survey areas,
188: by taking into account the local variations in sensitivity.
189: We also reproduce the desired $\log N - \log S$ distribution,
190: either the Kim et al. (2007) or the one recovered
191: directly from the CDF data (however our results remain mostly unchanged
192: using either of the two).
193: Random positioned sources with fluxes lower
194: than that corresponding to the particular position of the sensitivity
195: map are removed from our final random catalogue.
196:
197: We apply the correlation analysis evaluating $w(\theta)$ in the range
198: $[5^{''}, 900^{''}]$ in 10 logarithmic intervals with $\delta \log
199: \theta= 0.226$.
200: We find statistically significant signals for all bands and for both
201: CDF-N and CDF-S. As an example we provide in Table 1 the integrated signal
202: to noise ratios, given for two different flux-limits and two different
203: angular ranges.
204: The significance appears to be
205: low only for the CDF-S hard-band, but for the lowest flux-limit.
206:
207: The angular correlation function for two different flux-limits
208: are shown in Figure 1, with the lines corresponding to the best-fit
209: power law model: $w(\theta)=(\theta_{0}/\theta)^{\gamma-1}$,
210: using $\gamma=1.8$ and the standard $\chi^{2}$
211: minimization procedure. Note that for the CDF-N, we get
212: at some $\theta$'s very low or negative
213: $w(\theta)$ values. These, however, are taken into account
214: in deriving the integrated signal, presented in Table 1.
215:
216: Applying our analysis for different flux-limited sub-samples, we find
217: that the clustering strength increases with increasing flux-limit, in
218: agreement with the CDF-S results of Giacconi et al (2001).
219: In Figure 2 we plot the angular clustering scale, $\theta_0$, derived from
220: the power-law fit of $w(\theta)$, as a function of different sample
221: flux-limits. The trend is true for both energy bands and
222: for both CDF-N and CDF-S, although for the latter is apparently
223: stronger.
224:
225: We also find that at their lowest respective flux-limits the
226: clustering of CDF-S
227: sources is stronger than that of CDF-N (more so for the soft-band), in
228: agreement with the spatial clustering analysis of Gilli et al. (2005). This
229: difference has been attributed to cosmic variance, in the sense that
230: there are a few
231: large superclusters present in the CDF-S (Gilli et al. 2003).
232: However, selecting CDF-N and CDF-S sources at the same flux-limit reduces
233: this difference, which remains strong only for the highest flux-limited
234: sub-samples (see Fig.2).
235:
236: It is worth mentioning that our results could in principle
237: suffer from the so-called {\em amplification bias}, which can enhance
238: artificially the clustering signal due to the detector's PSF smoothing
239: of source pairs with intrinsically small angular separations
240: (see Vikhlinin \& Forman 1995; Basilakos et al. 2005).
241: However, we doubt whether this bias can significantly affect our results
242: because at the median redshift of the sources ($z\sim 1$) the {\em Chandra}
243: PSF angular size of $\sim 1^{''}$ corresponds to a rest-frame
244: spatial scale of only $\sim 5$ $h^{-1}$ kpc (even at large
245: off-axis angles, where the PSF size increases to $\sim 4^{''}$, the
246: corresponding spatial scale is only $\sim 20$ $h^{-1}$ kpc).
247: In any case, and ignoring for the moment the additional effect of
248: the variable PSF size, the above imply that only the $w(\theta)$
249: of the lowest flux-limited samples could in principle be affected,
250: but in the direction of reducing (and not inducing) the observed
251: $\theta_0-f_{x, {\rm limit}}$ trend (since the
252: uncorrected $\theta_0$ values are, if anything,
253: artificially larger than the true underlying one).
254:
255: However, the variability of the PSF size through-out the {\em Chandra} field
256: can have an additional effect, and possibly enhance or even produce
257: the observed $\theta_0-f_{x, {\rm limit}}$ trend.
258: %since in principle lower-fluxes can be observed
259: %more efficiently near the center of the field while brighter ones can
260: %be observed equally through-out the field.
261: To test for this we have repeated our analysis,
262: restricting the data to a circular
263: area of radius 6$^{'}$ around the center of the Chandra
264: fields, where we expect to have a relatively small variation of the PSF size.
265: This choice of radius was dictated as a compromise between excluding as much
266: external area as possible but keeping enough sources ($\mincir$50\% of
267: original) to perform the clustering analysis. The results show that indeed the
268: $\theta_0-f_{x, {\rm limit}}$ trend is present and qualitatively the same
269: as when using all the sources, implying that the
270: previously mentioned biases do not create the observed trend.
271:
272: \subsection{Comparison with other $w(\theta)$ results}
273: We investigate here whether the large span of published X-ray AGN
274: clustering results can be explained by the derived $\theta_0-f_{x,{\rm
275: limit}}$ trend. To this end we attempt to take into account
276: the different survey area-curves, by estimating
277: a characteristic flux for each survey, $f_{x}(\frac{1}{2} AC)$,
278: corresponding to half its area-curve
279: (easy to estimate from the different survey published
280: area-curves).
281:
282: In Figure 3 we plot the corresponding values of
283: $\theta_0$ (for fixed $\gamma=1.8$) as a function of $f_{x}(\frac{1}{2} AC)$
284: (for both hard and soft bands) for
285: the {\em Chandra} Large AREA Synoptic X-ray survey (CLASXS)
286: (Yang et al. 2003), the XMM/2dF (Basilakos et al. 2004; 2005),
287: XMM-COSMOS (Miyaji et al. 2007), XMM-ELAIS-S1 (Puccetti et al. 2006),
288: XMM-LSS (Gandhi et al. 2006) and AXIS (Carrera et al. 2007) surveys.
289: With the exception of the Yang et al. (2003) and the
290: Carrera et al. (2007) hard-band results, the rest are
291: consistent with the general flux-dependent trend. Note also that
292: Gandhi et al. (2006) do not find any significant clustering of their
293: hard-band sources. Of course,
294: cosmic variance is also at work (as evidenced also by the clustering
295: differences between the CDF-N and CDF-S; see Fig. 2 and Gilli et al. 2005)
296: which should be responsible for the observed scatter around the main trend
297: (see also Stewart et al. 2007).
298:
299: We would like to stress that the CDF surveys have a large flux
300: dynamical range which is necessary in order to investigate the
301: $f_{x,{\rm limit}}-\theta_o$ correlation.
302: This is probably why this effect has not been clearly detected in other surveys,
303: although recently, a weak such effect was found also in the CLASXS survey
304: (Yang et al. 2006).
305:
306: \subsection{The spatial correlation length using $w(\theta)$}
307: We can use Limber's equation to invert the angular clustering and
308: derive the corresponding spatial clustering length, $r_0$ (eg. Peebles 1993).
309: To do so it is necessary to model the spatial
310: correlation function as a power law and to assume a clustering evolution
311: model, which we take to be that of constant clustering in comoving
312: coordinates (eg. de Zotti et al. 1990; Kundi\'c 1997).
313: %($\epsilon=-1.2$).
314: %\begin{equation}
315: %\xi(r,z)=(r/r_{\circ})^{-\gamma} (1+z)^{-(3+\epsilon)+\gamma} \;,
316: %\end{equation}
317: %where the parameter $\epsilon$ defines the clustering evolution model.
318: %Then the amplitude, $\theta_0$, of the angular correlation function
319: %is related to the spatial
320: %correlation length $r_0$ (eg. Efstathiou et al. 1991) through the equation:
321: %\begin{equation}
322: %\theta_{\circ}^{\gamma-1}=H_{\gamma}r_{\circ}^{\gamma}
323: %\left(\frac{H_\circ}{c}\right)^\gamma
324: %\int_{0}^{\infty} \left( \frac{1}{N}\frac{{\rm d}N}{{\rm d}z}
325: %\right)^{2} \frac{E(z)}{x^{\gamma-1}(z)} %(1+z)^{-3-\epsilon+\gamma$}
326: %{\rm d}z \;,
327: %\end{equation}
328: %where we have used the constant clustering in comoving coordinates model
329: %($\epsilon=\gamma-3$). Note that $x(z)$ is the proper distance,
330: %$E(z)=\sqrt{\Omega_{\rm m}(1+z)^{3}+\Omega_{\Lambda}}$ and
331: %$H_{\gamma}=\Gamma(\frac{1}{2}) \Gamma(\frac{\gamma-1}{2})/\Gamma
332: %(\frac{\gamma}{2})$.
333: For the inversion to be possible it is necessary to know
334: the X-ray source redshift distribution,
335: %, ${\rm d}N/{\rm d}z$, and the total number, $N$, of the X-ray
336: %sources. Both
337: which can be determined by integrating the
338: corresponding X-ray source luminosity function above the minimum
339: luminosity that corresponds to the particular flux-limit used.
340: To this end we use the
341: Hasinger et al. (2005) and La Franca et al. (2005) LDDE luminosity functions
342: for the soft and hard bands, respectively.
343:
344: We perform the above inversion in the framework of the {\em
345: concordance} $\Lambda$CDM cosmological model
346: ($\Omega_{\rm m}=1-\Omega_{\Lambda}=0.3$) %,$H_{\circ}=70$km s$^{-1}$Mpc$^{-1}$
347: and the comoving clustering paradigm. The resulting values of the
348: spatial clustering lengths, $r_0$, show the same dependence on
349: flux-limits, as in Fig.2.
350:
351: We can compare our results with direct determinations of the spatial-correlation function
352: from Gilli et al. (2005), who used
353: a smaller ($\sim 50\%$) spectroscopic sample from the CDF-N and CDF-S.
354: They found a significant difference between the CDF-S and CDF-N
355: clustering, with $r_0=10.3 \pm 1.7 \; h^{-1}$ Mpc and
356: $r_0=5.5 \pm 0.6 \; h^{-1}$ Mpc, respectively (note also that the corresponding
357: slopes were quite shallow, roughly $\gamma\simeq 1.4-1.5$). Since, Gilli et al.
358: used sources from the full (0.5-8 keV) band,
359: we compare their results with our soft-band results which,
360: dominate the total-band sources. This comparison is possible, because
361: as we have verified using a Kolmogorov-Smirnov test, the flux distributions of
362: the sub-samples that have spectroscopic data are statistically equivalent with
363: those of the whole samples.
364: Our inverted clustering lengths, for the lowest flux-limit used, are:
365: $r_0=10.3 \pm 2 \; h^{-1}$ Mpc and $r_0=6.4 \pm 2.5 \; h^{-1}$ Mpc
366: (fixing $\gamma=1.8$) for the CDF-S and CDF-N respectively,
367: in good agreement with the Gilli et al. (2005) direct 3D determination\footnote{Leaving both
368: $r_{0}$ and $\gamma$ as free parameters in the fit,
369: %, to be fitted by the data,
370: we obtain $\gamma$'s quite near their nominal value ($\gamma\sim 1.6 - 1.8$)}.
371:
372: Our values can be also compared with the re-calculation of the CDF-N
373: spatial clustering by Yang et al. (2006), who find
374: $r_0\simeq 4.1 \pm 1.1 \; h^{-1}$ Mpc.
375:
376: We return now to the strong trend between $\theta_0$ (or the
377: corresponding $r_0$) and the sample flux-limit (see Fig.2 and 3),
378: which could be due to two possible effects
379: (or the combination of both). Either the different flux-limits correspond
380: to different intrinsic luminosities, ie., a luminosity-clustering dependence
381: (see also hints in the CLASXS and CDF-N based Yang et al. 2006 results; while for
382: optical data see Porciani \& Norberg 2006) or a redshift-dependent effect (ie., different
383: flux-limits correspond to different redshifts traced).
384: Using the sources
385: which have spectroscopic redshift determinations we have derived their
386: intrinsic luminosities, in each respective band,
387: from their count rates using a spectral index $\Gamma=1.9$ and the
388: {\em concordance} $\Lambda$CDM cosmology.
389: We have also applied an absorption correction by assuming a power-law
390: X-ray spectrum with an intrinsic $\Gamma=1.9$, obscured by an optimum
391: column density to reproduce the observed hardness ratio.
392: We then derive, for each flux-limit used, the median
393: redshift and median luminosity of the corresponding sub-sample. We find
394: relatively small variations and no monotonic change of the median redshift
395: with subsample flux-limit. For example, the median spectroscopic redshift for the
396: soft and hard bands, at the lowest flux-limit used, is ${\bar z}\sim 0.8$ and $\sim 0.95$,
397: respectively, while its mean variation between the different flux-limits used
398: is $\langle \delta z/z \rangle\simeq -0.11$ and -0.03 for the CDF-N and
399: $\langle \delta z/z \rangle \simeq -0.27$ and 0.05 for the CDF-S soft and
400: hard-bands, respectively. The large redshift variation of the soft-band CDF-S data
401: should be attributed to the presence of a few superclusters at $z\sim 0.7$;
402: see Gilli et al. 2003).
403:
404: In Figure 4 we present the correlation between the subsample
405: median X-ray luminosity and the corresponding subsample clustering length, as provided
406: by Limber's inversion. Although the CDF luminosity dynamical range is
407: limited, it is evident that the median X-ray luminosity systematically increases with
408: increasing sample flux-limit and it is correlated to $r_0$ (as expected from Fig.2).
409: It should be noted that the correlation length of the highest-flux limited CDF-S
410: soft-band subsample
411: is by far the largest ever found ($\sim 30 h^{-1}$ Mpc), but one has to keep in mind
412: that the CDF-S appears not to be a typical field, as discussed earlier (see Gilli et
413: al. 2003). The CDF-N high-flux results appear to converge to a value of
414: $r_0 \sim 18h^{-1}$ Mpc, similar to that of some other surveys
415: (eg. Basilakos et al. 2004; 2005 and Puccetti et al. 2007).
416: %Furthermore, note that the Yang et al (2003) hard-band angular
417: %clustering result ($\theta_0 \sim 40$ arcsec) corresponds to an even larger
418: %spatial clustering length.
419: %Also note that the soft-band clustering amplitude of Miyaji et
420: %al.(2007) is $\sim 11 \;h^{-1}$ Mpc (a factor of two higher than the clustering
421: %amplitude of optical QSOs).Interestingly, the luminosities of the X-COSMOS sources
422: %are in the range $\log L = 42.5-44$. At these luminosities the high X-COSMOS
423: %clustering is not very different from what we have found (see CDFN
424: %results in our Figure 4).
425:
426: We therefore conclude that not only are there indications for a luminosity dependent
427: clustering of X-ray selected high-$z$ AGNs, but also that they are significantly more
428: clustered than their lower-$z$ counterparts, which have $r_0 \sim 7-8 \;
429: h^{-1}$ Mpc (eg. Akylas et al. 2000; Mullis et al. 2004). This is a clear indication of
430: a strong bias evolution (eg. Basilakos, Plionis \& Ragone-Figueroa 2007).
431:
432: \section{Conclusions}
433: We have analysed the angular clustering of the CDF-N and CDF-S X-ray AGNs and find:
434:
435: \noindent
436: (1) A dependence of the angular clustering strength
437: on the sample flux-limit. Most XMM and
438: {\em Chandra} clustering analyses provide
439: results that are consistent with the observed
440: trend; a fact which appears to lift the confusion that arose from the
441: apparent differences in their respective clustering lengths.
442:
443: \noindent
444: (2) Within the concordance cosmological model, the comoving clustering
445: evolution model and the LDDE luminosity function, our angular clustering
446: results are in good agreement with direct estimations of the
447: CDF-N and CDF-S spatial clustering, which are based however on
448: roughly half the
449: total number of sources, for which spectroscopic data were available.
450:
451: \noindent
452: (3) The apparent correlation between clustering strength and sample flux-limit
453: transforms into a correlation between clustering strength and intrinsic
454: X-ray luminosity, since no significant redshift-dependent trend was found.
455:
456: %\acknowledgments
457:
458: \begin{references}
459: \reference{}Akylas, A., Georgantopoulos, I., Plionis, M., 2000, \mnras,
460: 318, 1036
461: \reference{} Alexander, D.M., et al., 2003, AJ, 126, 539
462: %\reference{BP01} Basilakos, S. \& Plionis, M., 2001, ApJ, 550, 522
463: %\reference{BP03} Basilakos, S. \& Plionis, M., 2003, ApJ, 593, L61
464: \reference{BP05} Basilakos, S. \& Plionis, M., 2005, MNRAS, 360, L35
465: \reference{BP06} Basilakos, S. \& Plionis, M., 2006, ApJ, 650, L1
466: \reference{B04} Basilakos, S., Georgakakis, A., Plionis, M.,
467: Georgantopoulos, I., 2004, ApJL, 607, L79
468: \reference{B05} Basilakos, S., Plionis, M., Georgantopoulos, I.,
469: Georgakakis, A., 2005, MNRAS, 356, 183
470: \reference{BP07} Basilakos, S., Plionis, M. \& Ragone-Figueroa, 2007, ApJ, {\em submitted}
471: \reference{}Barger, A.J., et al., 2003, AJ, 126, 632
472: \reference{} Bauer, F.E., Alexander, D.M., Brandt, W.N., Schneider, D.P.,
473: Treister, E., Hornschemeier, A.E., Garmire, G.P., 2004, AJ, 128, 2048
474: \reference{} Boyle B.J., Mo H.J., 1993, \mnras, 260, 925
475: %\reference{} Boyle B.J., et al. 1998, \mnras, 296, 1
476: \reference{Carr} Carrera, F.J., Barcons, X., Fabian, A. C.,
477: Hasinger, G., Mason, K.O., McMahon, R.G., Mittaz, J.P.D., Page, M.J.,
478: 1998, MNRAS, 299, 229
479: \reference{} Carrera, F.J., et al. 2007, A\&A, 469, 27
480: %\reference{} Capak, P. et al., 2007, {\tt astro-ph/0704.2430}
481: %\reference{Cr96}Croom, S. M., \& Shanks, T., 1996, MNRAS, 281, 893
482: %\reference{} Croom, S.M., Smith, R.J., Boyle, B.J., Shanks, T., Loaring,
483: % N.S., Miller, L., Lewis, I.J., 2001, MNRAS, 322, L29
484: %\reference{Cr02}Croom, S.M., Boyle, B.J., Loaring, N.S., Miller, L.,
485: %Outram, P.J., Shanks, T., Smith,R.J., 2002, MNRAS, 335, 459
486: %\reference{Cr05} Croom, S.M., et al., 2005, MNRAS, 365, 415
487: \reference{deZ}de Zotti, G., Persic, M., Franceschini, A., Danese, L.,
488: Palumbo, G.G.C., Boldt, E.A., Marshall, F.E., 1990, ApJ, 351, 22
489: %\reference{} Efstathiou, G., 2002, MNRAS, 330, L29
490: %\reference{} Efstathiou, G., Bernstein, G., Katz, N., Tyson, J. A.,
491: %Guhathakurta, P., 1991, \apj, 380, L47
492: \reference{Gan} Gandhi, P., et al., 2006, A\&A, 457, 393
493: \reference{Gil1} Gilli, R., et al. 2003, ApJ, 592, 721
494: \reference{Gil2} Gilli, R., et al. 2005, A\&A, 430, 811
495: %\reference{} Giacconi, R., et al. 2002, \apjs, 139, 369
496: \reference{} Giacconi, R., et al. 2001, \apj, 551, 624
497: %\reference{Gra} Grazian, A.,
498: %Negrello, M., Moscardini, L., Cristiani, S., Haehnelt, M.G., Mataresse, S., Omizzolo, A., Vanella, E., 2004, AJ, 127, 592
499: %\reference{} Hasinger, G., et al., 2001, A\&A, 365, L51
500: \reference{Has05} Hasinger, G., Miyaji, T., Schmidt, M., 2005, A\&A, 441, 417
501: %\reference{} Infante, L., 1994, A\&A, 282, 353
502: %\reference{LaF98} La Franca F., Andreani, P., Cristiani, S., 1998, ApJ, 497, 529
503: \reference{} Kim, M., Wilkes, B.J., Kim, D-W., Green, P.J., Barkhouse, W.A.,
504: Lee, M.G., Silverman, J.D., Tananbaum, H.D., 2007, ApJ, 659, 29
505: \reference{} Kundi\'c, T., 1997, ApJ, 482, 631
506: \reference{LaF05} La Franca, F. et al., 2005, ApJ, 635, 864
507: \reference{}Le F\'evre, O., et al., 2004, A\&A, 428, 1043
508: \reference{} Lehmer, B.D., et al., 2005, ApJS, 161, 21
509: \reference{} Manners, J.C., et al., 2003, MNRAS, 343, 293
510: %\reference{Mi00}Miyaji, T., Hasinger, G., Schmidt, M., 2000, A\&A, 353, 25
511: \reference{}Mignoli, M. et al., 1005, A\&A, 437, 883
512: \reference{Mi06}Miyaji, T., et al., 2007, ApJS, 172, 396
513: \reference{Mu}Mullis C.R., Henry, J. P., Gioia I. M.,
514: B\"{o}hringer H., Briel, U. G., Voges, W., Huchra, J. P., 2004, ApJ, 617, 192
515: %\reference{} Mullis C.R., 2002, PASP, 114, 668
516: \reference{}Mo, H.J, \& White, S.D.M 1996, MNRAS, 282, 347
517: %\reference{} Peebles P.J.E., 1980, ``The Large Scale Structure of the Univerce'', Princeton Univ. Press, Princeton, NJ
518: \reference{} Peebles P.J.E., 1993, ``Principles of Physical Cosmology'',
519: Princeton Univ. Press, Princeton, NJ
520: \reference{} Peebles, P.J.E., 1973, \apj, 185, 413
521: \reference{}Plionis, M. \& Basilakos, S., 2007, proceedings of the 6$^{th}$
522: International Workshop on the Identification of Dark Matter, held in Rhodes,
523: Greece, {\tt astro-ph/0701696}
524: \reference{} Porciani, C. \& Norberg, P., 2006, MNRAS, 371, 1824
525: \reference{Pu} Puccetti, S., et al., 2006, A\&A, 457, 501
526: \reference{} Sheth, R.K., Mo, H.J., Tormen, G., 2001, MNRAS, 323, 1
527: \reference{} Stewart, G.C., 2007, in ``X-ray Surveys: Evolution of
528: Accretion, Star-Formation and the Large-Scale Structure", Rodos island,
529: Greece \\
530: {\tt http://www.astro.noa.gr/xray07/rodos-talks/}
531: \reference{}Szokoly, G.P., et al., 2004, ApJS,
532: 155, 271
533: \reference{} Vanzella, E., et al., 2005, A\&A, 434, 53
534: \reference{} Vanzella, E., et al., 2006, A\&A, 454, 423
535: \reference{VF}Vikhlinin, A. \& Forman, W., 1995, ApJ, 455, 109
536: \reference{Y03} Yang, Y., Mushotzky, R.F., Barger, A.J.,
537: Cowie, L.L., Sanders, D.B., Steffen, A.T., 2003, ApJ, 585, L85
538: \reference{Y06}Yang, Y., Mushotzky, R.F., Barger, A.J., Cowie, L.L., 2006,
539: ApJ, 645, 68
540: %\reference{} Peebles, P.J.E., 1973, \apj, 185, 413
541: \end{references}
542:
543:
544: \begin{table}[h]
545: \caption[]{The integrated angular clustering signal.}
546: \vspace{1cm}
547:
548: \tabcolsep 15pt
549: \begin{tabular}{lcccc}
550: Sample & \multicolumn{2}{c}{$f_{x,{\rm limit}}$$^a$} &
551: \multicolumn{2}{c}{$f_{x,{\rm limit}}$$^b$} \\ \hline
552: & $<400^{''}$ & $<900^{''}$ & $<400^{''}$ & $<900^{''}$ \\ \hline
553: CDF-N soft & $2.1\sigma$ & $0.0\sigma$ & $2.1\sigma$ & $1.5\sigma$ \\
554: CDF-N hard & $3.3\sigma$ & $2.6\sigma$ & $6.7\sigma$ & $5.5\sigma$ \\
555: CDF-S soft & $4.2\sigma$ & $2.6\sigma$ & $3.4\sigma$ & $3.7\sigma$ \\
556: CDF-S hard & $0.3\sigma$ & $1.3\sigma$ & $4.1\sigma$ & $2.7\sigma$ \\ \hline
557:
558: \multicolumn{5}{l}{$^a$ the lowest flux-limit.} \\
559: \multicolumn{5}{l}{$^b$ $5 \times 10^{-16}$ erg s$^{-1} $cm$^{-2}$
560: (soft band) and $10^{-15}$ erg s$^{-1} $cm$^{-2}$ (hard band).} \\
561:
562: \end{tabular}
563: \end{table}
564:
565: \clearpage
566:
567: \begin{figure}
568: \epsscale{0.6}
569: \plotone{f1.eps}
570: %\plotone{wtheta.ps}
571: \figcaption{The CDF-S and CDF-N angular correlation function. The open points
572: correspond to the overall sample, while the filled points to the highest
573: flux-limit used ($f_x= 3\times 10^{-15}$ ${\rm erg/s/cm^2}$ for the soft and
574: $f_x= 5\times 10^{-15}$ ${\rm erg/s/cm^2}$
575: for the hard bands, respectively). The straight
576: lines correspond to the best power-law fit to the clustering
577: data. Errorbars correspond to $1\sigma$ Poisson uncertainties.}
578: \end{figure}
579:
580: \begin{figure}
581: \epsscale{0.6}
582: \plotone{f2.eps}
583: %\plotone{theta_o1.ps}
584: \figcaption{The angular clustering scale as a function of the flux-limit of
585: the different samples. The dependence of clustering strength to the
586: flux-limit is evident. The {\em left} and {\em right} panels correspond to
587: the hard and soft bands respectively. Filled symbols correspond to the CDF-N
588: while open ones to the CDF-S.}
589: \end{figure}
590:
591: \begin{figure}
592: \epsscale{0.7}
593: \plotone{f3.eps}
594: %\plotone{theta_eff1.ps}
595: \figcaption{The angular correlation scale, $\theta_0$, as a function of
596: different survey characteristic flux, defined as that corresponding to half the
597: respective survey area-curves. Most results appear to be consistent with
598: the clustering flux-limit dependence, found from the CDF-N and
599: CDF-S. Note that in the left panel we plot the 4.5-10 keV results of
600: Miyaji et al. (2007). Errorbars correspond to $1\sigma$ Poisson uncertainties.}
601: \end{figure}
602:
603: \begin{figure}
604: \epsscale{0.7}
605: \plotone{f4.eps}
606: \figcaption{The correlation between the clustering length, $r_0$, and the
607: median intrinsic X-ray luminosity of each subsample.
608: The {\em left} and {\em right} panels correspond to
609: the hard and soft bands respectively. Filled symbols correspond to the CDF-N
610: while open ones to the CDF-S.}
611: \end{figure}
612:
613: \end{document}
614: