1: \documentclass[12pt,preprint]{aastex}
2:
3:
4: \begin{document}
5:
6: \title{Identifying high redshift AGNs using X-ray hardness}
7:
8: \author{J. X. Wang\altaffilmark{1,3}, S. Malhotra\altaffilmark{2},
9: J. E. Rhoads\altaffilmark{2}, C. A. Norman\altaffilmark{1,2}}
10: \begin{abstract}
11: The X-ray color (hardness ratio) of optically undetected X-ray sources
12: can be used to distinguish obscured active galactic nuclei (AGNs) at
13: low and intermediate redshift from viable high-redshift (i.e., $z > 5$) AGN
14: candidates. This will help determine the space density, ionizing
15: photon production, and X-ray background contribution of the earliest
16: detectable AGNs.
17: High redshift AGNs should appear soft in X-rays, with hardness
18: ratio HR $\sim$ -0.5, even if there is strong absorption by a
19: hydrogen column density $N_H$ up to
20: $10^{23} cm^{-2}$, simply because the absorption redshifts out of the soft
21: X-ray band in the observed frame. Here the X-ray hardness ratio is defined
22: as HR= (H-S)/(H+S), where S and H are the soft and hard band net counts
23: detected by $Chandra$. High redshift AGNs that are Compton thick
24: ($N_H \gtrsim 10^{24} cm^{-2}$) could have HR $\sim$ 0.0 at $z >$ 5.
25: However, these should be rare in deep Chandra images, since they have
26: to be $\gtrsim$ 10 times brighter intrinsically, which implies
27: $\gtrsim$ 100 times drop in their space density.
28: Applying the hardness criterion (HR $<$ 0.0) can filter out
29: about 50\% of the candidate high redshift AGNs selected from deep
30: Chandra images.
31: \end{abstract}
32:
33: \keywords{galaxies: active --- galaxies: high-redshift --- X-rays: galaxies}
34:
35: \altaffiltext{1}{Department of Physics and Astronomy, Johns Hopkins University,
36: 3400 N. Charles Street, Baltimore, MD 21218; jxw@pha.jhu.edu, norman@stsci.edu.}
37: \altaffiltext{2}{Space Telescope Science Institute, 3700 San Martin Drive,
38: Baltimore, MD 21218; san@stsci.edu, rhoads@stsci.edu.}
39: \altaffiltext{3}{Center for Astrophysics, University of Science and Technology of China, Hefei, Anhui 230026, P. R. China; jxw@ustc.edu.cn.}
40: \section {Introduction}
41:
42: In the past few years, many deep $Chandra$ images of the extragalactic sky
43: have been obtained, with the 2 Ms $Chandra$ Deep Field North (CDF-N, e.g.,
44: Alexander et all. 2003) and 1Ms $Chandra$ Deep Field South (CDF-S, Giacconi
45: et al. 2002; Rosati et al. 2002) being the two deepest.
46: Combining such exposures with deep optical images allows easy selection of
47: candidate high redshift (i.e., $z > 5$, hereafter high-z) AGNs.
48: Such high-z AGNs are extremely faint in optical bands blueward of
49: the Ly$\alpha$ wavelength, because of the
50: heavy absorption of UV light by the high redshift IGM (e.g., Fan et al.
51: 2001).
52: Optically undetected X-ray sources are thus good candidates for high-redshift
53: AGNs. Their space density provides an upper limit on the density of high-z
54: AGNs, and can help determine their cosmological evolution and
55: contribution to reionization (e.g., see Alexander et al. 2001;
56: Barger et al. 2003a; Koekemoer et al. 2004; Wang et al. 2004).
57: However, the absence of these candidates in optical bands makes
58: them difficult to identify spectroscopically.
59: This motivates different approaches to studying them.
60: Currently, the efforts mainly focus on their infrared
61: colors (e.g., Yan et al. 2003; Koekemoer et al. 2004).
62:
63: In this letter, we point out for the first time that the X-ray hardness
64: ratio can be used to filter out low-z sources from these X-ray
65: selected, optically undetected high-z candidates. High-z AGNs cannot be hard
66: in $Chandra$ images, because the absorption that makes the X-ray spectra harder
67: is redshifted out of the soft X-ray band in the observed frame.
68: In section 2, we present detailed simulations to quantify this effect.
69: Here the hardness ratio (HR) is defined as (H-S)/(H+S), where S and H are
70: the soft (0.5 -- 2.0 keV) and hard (2.0 -- 8.0 keV)\footnote{In some papers, only the 2.0 -- 7.0 keV band
71: net counts were given (e.g., Giacconi et al. 2002; Stern et al. 2002; Wang
72: et al. 2004). The difference of the hardness ratios (HR) using
73: different hard bands (2.0 -- 8.0 keV or 2.0 -- 7.0 keV) is negligible
74: ($\Delta$HR $<$ 0.008 from our simulations). This is actually expected because $Chandra$ has
75: much lower effective area above 7 keV and 7.0 -- 8.0 keV X-ray net count makes
76: a very small contribution to the whole hard band.
77: } X-ray band net counts detected by $Chandra$.
78:
79: \section{Simulations}
80:
81: The X-ray spectra of low and intermediate redshift AGNs have been well
82: studied using the observations from several generations of X-ray satellites,
83: including EINSTEIN, ROSAT, ASCA, BeppoSAX, $Chandra$, and XMM.
84: For type 1 AGNs (i.e., Seyfert 1 galaxies, and QSOs), the basic
85: component of their X-ray spectra is a power law
86: with photon index $\Gamma \sim 1.9$ (e.g., Nandra et al. 1997; George
87: et al. 2000, Malizia et al. 2003) and an exponential cut off at high
88: energies ($\sim$ 200 keV, see Malizia et al. 2003).
89: For type 2 AGNs (i.e., Seyfert 2 galaxies, and type 2 QSOs),
90: the power law is cut off at low energies
91: by photo-electric absorption, and the cutoff energy increases
92: with the column density of the intercepted torus (e.g.,
93: Turner et al. 1997a; Norman et al. 2002).
94: Recent $Chandra$ observations show that the X-ray spectra of QSOs
95: (i.e., luminous AGNs) at $z$ = 4.0 -- 6.3 are also
96: well fitted by a power law with photon index $\Gamma$ = 1.9 (Vignali et al. 2004).
97: This indicates that although the space density of AGNs varies significantly
98: from $z \sim$ 0 -- 6, the shape of their intrinsic X-ray spectra evolves
99: relatively little.
100: Evidence for warm absorbers and/or
101: soft excess emission (e.g., Krolik \& Kriss 2001; Piro, Matt \& Ricci 1997) have
102: also been found in significant numbers of AGNs.
103: However, at high-z, these features shift out
104: of $Chandra$'s soft band.
105:
106: In this section we present simulations to predict the X-ray colors
107: in $Chandra$ images for high-z AGNs by assuming a power law spectrum
108: ($\Gamma$ = 1.9) with different absorption column densities
109: (N$_H$ = 10$^{21}$, 10$^{22}$, 10$^{23}$, 10$^{24}$ cm$^{-2}$ respectively;
110: see Fig. 1 for the model spectra).
111: We used
112: XSPEC 11.0.1%\footnote {See http://legacy.gsfc.nasa.gov/docs/xanadu/xspec/index.html for the most recent version of XSPEC.}
113: to do the simulations, and
114: model $wabs$ in XSPEC, a photoelectric absorption using Wisconsin
115: cross-sections
116: (Morrison and McCammon, 1983), to simulate
117: the neutral absorption in the rest frame. The
118: $Chandra$ ACIS on-axis instrument response for CDF-S (Giacconi et al. 2002)
119: was used,
120: and the Galactic HI column density (N$_H$ = 0.8 $\times 10^{20} cm^{-2}$)
121: in CDF-S was taken account during the simulations\footnote{
122: Using slightly different Galactic HI column density, or $Chandra$ on-axis instrument response
123: calculated for other ACIS-I fields, does not affect any results
124: presented in this paper.}.
125: The output X-ray hardness ratios are plotted in Fig. \ref{hr}.
126: We can see that the predicted HR is a constant (-0.58 for
127: $\Gamma$ = 1.9) for AGNs without intrinsic absorption at any
128: redshift, because of the power law shape of the
129: X-ray spectrum. The corresponding HR for
130: different photon indices are also shown in Fig. \ref{hr}.
131: While the photon index for QSOs varies
132: from 1.5 to 3.0 (e.g., George et al. 2000), the dominant source of variation
133: in hardness ratio HR is absorption. We show in Fig. \ref{hr}
134: that an extreme power-law
135: with $\Gamma = 1.5$ would still give a soft color
136: (HR = -0.41).
137: For AGNs with intrinsic absorption, the predicted hardness ratio
138: varies with redshift: at lower redshift, the X-ray spectra
139: are much harder because the soft X-ray emission are significantly attenuated
140: by the absorber; but at $z > 5$,
141: we barely see differences between the X-ray hardness ratios of X-ray
142: spectra with absorption up to N$_H$ = 10$^{23}$ cm$^{-2}$, because
143: the absorption has largely redshifted out
144: of the soft X-ray band.
145: If the absorber is Compton thick (N$_H \ge 10^{24} cm^{-2}$), even the hard
146: X-ray emission would be significantly attenuated.
147: At $z > 5$, the predicted hardness ratio is $\sim$ 0.0.
148: In the compton thick regime (N$_H$ = $10^{24} cm^{-2}$), further correction to
149: photoelectric absorption is needed, due to
150: Compton scattering (see Matt, Pompilio \& La Franca 1999; Yaqoob 1997).
151: Based on figure 3 of Matt et al. (1999)
152: which includes compton scattering,
153: we conclude that the {\it shape} of the transmitted curve is unchanged
154: by compton scattering, while the amplitude decreases
155: by a factor of 1.7.
156:
157: For higher column density (N$_H > 10^{24} cm^{-2}$), the direct X-ray
158: emission is strongly attenuated,
159: and the X-ray spectra are dominated by a reflection component
160: from cold and neutral gas (e.g., Turner et al. 1997b).
161: We used the XSPEC model $pexrav$ (Magdziarz \& Zdziarski
162: 1995) to simulate such pure reflection spectra.\footnote{The model of
163: Magdziarz \& Zdziarski is angle-dependent. The results we present in this
164: paper were derived by setting cos$\theta$ at 0.45, which is closest
165: in overall shape to the reflected spectrum averaged over all viewing angles.
166: Reflection from smaller viewing angles tends to be slightly harder ($\Delta HR < 0.1$), but
167: obviously, this is not the case for type II AGNs which
168: are supposed to edge-on.}
169: The Fe K emission line at 6.4 keV has a higher equivalent width (EW) in
170: the reflection dominated X-ray spectra of AGNs (e.g., Ghisellini, Haardt \&
171: Matt 1994; Levenson et al. 2002) since the direct component is absent.
172: Therefore we add
173: an Fe K emission line at 6.4 keV with EW of 1 keV in the rest frame,
174: which is normal in the
175: reflection dominated X-ray spectra of AGNs (e.g., Levenson et al. 2002).
176:
177: \section{Discussion}
178: The X-ray spectra of AGNs at low to intermediate redshifts can be extremely
179: hard due to heavy absorption ( with hardness ratios up to HR $\sim$ 1.0). However, they are
180: much softer at higher redshift because the absorbed energy shifts out of the observed bands (see Fig. 2 for
181: simulations, and Fig. 12 of Szokoly et al. 2004 for the HR distribution
182: of a large sample of AGNs from $z$ = 0-4).
183: AGNs at $z \gtrsim 5$ with intrinsic absorption up to N$_H = 10^{23} cm^{-2}$
184: should have HR $\sim$ - 0.5, and the Compton-thick ones (N$_H \gtrsim 10^{24}
185: cm^{-2}$) should have HR $\lesssim$ 0.1.
186: Due to heavy absorption, and compton scattering, the X-ray flux of
187: Compton-thick AGNs (N$_H \sim 10^{24}$ $cm^{-2}$) is attenuated by a factor of 9.5
188: (See Fig. 1).
189:
190:
191: For pure reflection spectra, the attenuation is even larger:
192: assuming a reflection efficiency of 3\% in rest frame 2.0 -- 10.0 keV band
193: (e.g., see Norman et al. 2002) yields a factor of 21.
194: Thus any high redshift sources detected with a large hardness ratio would have intrinsic
195: X-ray luminosity of $\sim 10^{45}$ erg s$^{-1}$.
196: We know that brighter QSOs are much rarer; according to the X-ray luminosity
197: function of AGNs (e.g., see Miyaji, Hasinger, \& Schmidt 2001; Ueda et al. 2003),
198: a tenfold increase in luminosity
199: implies a hundredfold drop in the space density.
200: Furthermore, there is evidence that the fraction of type 2 AGNs
201: decreases at higher intrinsic
202: luminosity (e.g., Steffen et al. 2003; Ueda et al. 2003).
203: We conclude that those candidates with $HR \gtrsim 0.0$ are statistically unlikely to
204: be at $z \gtrsim 5$. They are either obscured AGNs or QSOs at low to
205: intermediate redshift.
206:
207: $Chandra$ has detected a number of AGNs at high redshift.
208: Presently, there are 66 AGNs\footnote{See the excellent Web site
209: http://wwww.astro.psu.edu/users/niel/papers/highz-xray-detected.dat
210: maintained by Niel Brandt and Christian Vignali for the list of the
211: high-redshift AGN detected in X-rays so far} at $z > 4$ detected
212: by $Chandra$ ACIS, and 41 of them have published
213: soft (0.5 -- 2.0 keV) and hard (2.0 -- 8.0 keV) band net counts (or 0.5 --
214: 2.0 keV band and 0.5 -- 8.0 keV band net counts)
215: (Alexander et al. 2003; Barger et al. 2002; Brandt et al. 2001, 2002;
216: Castander et al. 2003; Vignali et al. 2001, 2003a, 2003b; Bassett et al. 2004).
217: All of the 41 sources have HR $\le$ 0.0, with an average value of -0.60$\pm0.21$,
218: in excellent agreement with our estimates above.
219: All of these sources are type 1
220: AGNs, and most of them are optically selected.
221: Since strong X-ray emission is expected from both type 1 and type 2 AGNs,
222: we expect the X-ray selected high-z AGN sample to include both
223: types.
224: However we argue that the
225: hardness ratio distribution of the X-ray selected high-z AGNs should be similar
226: to that of the known $z>4$ AGNs based on following reasons: I) the 3 X-ray selected AGNs with $z > 4$ have
227: consistent soft X-ray colors with the rest; II) high-z type 2 AGNs with N$_H$
228: upto 10$^{23} cm^{-2}$ are also expected to be X-ray soft with HR $\sim$ -0.5;
229: III) type 2 AGNs with heavier absorption are much fainter in observed X-ray
230: fluxes, thus are much rarer.
231:
232: Wang et al. (2004) presented 168 X-ray sources detected by an 172 ks
233: $Chandra$ ACIS exposure in the Large Area Lyman Alpha (LALA, e.g,
234: Rhoads et al. 2003) Bo\"{o}tes field. 19 of them are not detected
235: in deep R and bluer band images ($R > 25.7$, Vega mag)
236: and are possible $z \gtrsim 5$ objects.
237: The sources and their hardness ratios are listed in Table 1 of Wang et al.
238: In Fig. 3, we plot their hardness
239: ratios comparing with the 41 $Chandra$ detected $z > 4$ AGNs.
240: The two distributions are distinguishable at 99.99\% level
241: according to the Kolmogorov-Smirnov (K-S) test.
242: Down to a flux limit of 1.7 $\times$ 10$^{-15}$ ergs cm$^{-2}$ s$^{-1}$,
243: the 19 sources
244: contribute $\sim$ 8\% to the total 2 -- 10 keV band X-ray background
245: (Wang et al. 2004).
246: After removing sources with HR $>$ 0, we lose 50\% of the candidates
247: and the contribution to the total X-ray background drops to $\sim$ 4\%.
248:
249:
250: Yan et al. (2003) studied the infrared colors of 6 $R$ band nondetected X-ray
251: sources in CDF-S.
252: These sources are listed in table 1. Two of them have HR $\gtrsim$
253: 0.0, and cannot be at $z > 5$. This agrees with
254: Yan's statement, that all these sources are unlikely to be at $z > 5$ based
255: on their infrared colors.
256: Koekemoer et al. (2004) presented 7 X-ray sources in CDF-S which are
257: undetected in deep multi-band GOODS HST ACS images, with extremely
258: high X-ray-to-optical ratios and red colors ($z_{850}$ -- K).
259: These sources might
260: be located at $z >$ 6 such that even their Ly$\alpha$ emission is redshifted
261: out of the bandpass of ACS $z_{850}$ filter.
262: We find that 3 of the 7 sources have HR $\gtrsim$ 0.0,
263: indicating that their X-ray colors are too hard to be at $z > 6$.
264: These 3 sources could instead be type 2 AGNs at low to intermediate redshift.
265: Their nuclear optical emission should be heavily obscured,
266: and their host galaxies need to be substantially underluminous, or
267: dust-obscured, compared to other known sources (Koekemoer et al. 2004).
268: This confirms that there is a population of AGNs at low to
269: intermediate redshift which are extremely red, with high X-ray-to-optical
270: ratios, and undetected at the depth of GOODS.
271: The analogous sample among non-active galaxies is the population of
272: extremely red objects (EROs), which have surface density comparable to
273: Lyman break galaxies but a much lower typical redshift (e.g., Vaisanen
274: \& Johansson 2004).
275:
276: Using deep multicolor optical data, Barger et al. (2003a) searched
277: candidate $z > 5$ AGNs in the 2 Ms X-ray exposure of the CDF-N, and
278: found that besides the one X-ray source spectroscopically confirmed
279: at $z = 5.19$, only 31 X-ray sources with $z' >$ 25.2 and no
280: B or V band detection could lie at $z > 5$. Barger et al. (2003b) provided
281: multiband photometry for the CDF-N X-ray sources, which allows us
282: to identify the 31 candidate high-z AGNs\footnote
283: {Because the magnitudes discussed in Barger et al. 2003a and 2003b were measured
284: with different aperture diameters, it is difficult to pick up exactly the same
285: 31 sources discussed in Barger et al. 2003a based on the magnitudes provided
286: in Barger et al. 2003b.
287: We identify 31 candidate high-z AGNs in CDF-N following the same manner
288: discussed in Barger et al. 2003a (i.e., z' band faint and $B,V$ band
289: undetected), and tune the threshold magnitudes to match the numbers
290: of sources discussed in Barger et al. 2003a. Thus the sample we picked up
291: could statistically represent that of Barger et al. 2003a}. The hardness ratio distribution
292: of the 31 sources is plotted in Fig. 3. 15 of the 31 sources have hardness ratios
293: HR $>$ 0.0, and thus cannot have $z > 5$.
294: This directly supports the deduction that the majority of the optically
295: undetected X-ray sources are extreme examples of
296: the optically faint X-ray source population, most of which are obscured AGNs
297: at z $\le$ 3 (Alexander et al. 2001).
298: Barger et al. (2003a) pointed out that Haiman \& Loeb (1999) overestimated
299: the surface density of $z > 5$ AGNs by at least an order of magnitude, and
300: similar conclusions can be seen in Alexander et al. (2001) and
301: Szokoly et al. (2004). Our analyses indicate that applying the X-ray hardness
302: ratio cutoff (HR $>$ 0.0) could further reduce the surface density of candidate
303: $z > 5$ AGNs, and strengthen the above conclusions by a factor of 2.
304: This also supports the statement that AGNs made little contribution to
305: the reionization at $z \sim 6$ (Barger et al. 2003a; also see
306: Dijkstra, Haiman \& Loeb 2004; Moustakas \& Immler 2004).
307:
308: \section {Conclusions}
309:
310: In this letter we present detailed simulations showing that high-z AGNs
311: cannot be hard in $Chandra$ images since X-ray absorption will shift out
312: of soft band at high redshift.
313: High redshift AGNs should appear soft in X-rays, with hardness
314: ratio HR $\sim$ -0.5 at $z \gtrsim 5$, even if there is strong absorption
315: with $N_H$ up to $10^{23} cm^{-2}$.
316: High-z AGNs that are Compton thick ($N_H \gtrsim 1
317: 0^{24} cm^{-2}$) could have HR $\sim$ 0.0. However, these should be rare
318: in deep Chandra images, since they have to be $\gtrsim$ 10 times brighter
319: intrinsically, which implies $\gtrsim$ 100 times drop in their space density.
320: Most optically undetected X-ray sources with HR $\gtrsim$ 0.0 should be
321: obscured AGNs at low to intermediate redshift.
322: Applying the hardness criterion (HR $<$ 0.0) can filter out about 50\% of the
323: candidate high redshift AGNs selected from deep Chandra images.
324: This criterion can thereby help us to understand the
325: nature of these $Chandra$ X-ray sources, put additional robust constraints
326: to the space density of high-z AGNs, and significantly reduce the
327: expensive telescope time needed to spectroscopically confirm
328: high-z AGN samples based on deep $Chandra$ images.
329:
330: \acknowledgements
331: We would like to thank Dr. A. Hornschemeier and T. Yaqoob for helpful
332: discussions. The work of JW was supported by the CXC grant GO2-3152x and GO3-4148X.
333: We also would like to thank the referee for a prompt and helpful report.
334:
335: \begin{references}
336: \reference{} Alexander, D. M., et al. 2001, AJ, 122, 2156
337: \reference{} Alexander, D. M., et al. 2003, AJ, 126, 539
338: \reference{} Barger, A. J., et al. 2002, AJ, 124, 1839
339: \reference{} Barger, A. J., et al. 2003a, ApJ, 584, L61
340: \reference{} Barger, A. J., et al. 2003b, AJ, 126, 632
341: \reference{} Basset, L. C., Brandt, W. N., Schneider, D. P., Vignali, C., Chartas, G., \&
342: Garmire, G. P. 2004, AJ in press, astro-ph/0404543
343: \reference{} Brandt, W. N., et al. 2001, ApJ, 122, 1
344: \reference{} Brandt, W. N., et al. 2002, ApJ, 569, L5
345: \reference{} Castander, F. J., et al. 2003, AJ, 125, 1689
346: \reference{} Dijkstra, M., Haiman, Z., \& Loeb, A. 2004, ApJ submitted, astro-ph0403078
347: \reference{} Fan, X. et al. 2001, AJ, 121, 54
348: \reference{} Ghisellini, G., Haardt, F., \& Matt, G., 1994, MNRAS, 267, 743
349: \reference{} Giacconi, R., et al. 2002, ApJS, 139, 369
350: \reference{} George, I. M., et al. 2000, ApJ, 531, 52
351: \reference{} Haiman, Z, \& Loeb, A. 1999, ApJ, 512, L9
352: \reference{} Koekemoer, A. M., et al. 2004, ApJ, 600, L123
353: \reference{} Krolik, J. H., \& Kriss, G. A. 2001, ApJ, 561, 684
354: \reference{} Levenson, N. A., Krolik, J. H., \.{Zycki}, P. T., Heckman, T, M., Weaver, K. A., \& Awaki, H. 2002, ApJ, 573, L81
355: \reference{} Lyons, L. 1991, Data Analysis for Physical Science Students, Cambridge University Press, Cambridge
356: \reference{} Magdziarz, P, \& Zdziarski, A. 1995, MNRAS, 273, 837
357: \reference{} Malizia, A., Bassani, L., Stephen, J. B., \& Di Cocco, G. 2003, ApJ, 589, L17
358: \reference{} Matt, G., Pompilio, F., La Franca, F. 1999, NewA, 4, 191
359: \reference{} Miyaji, T., Hasinger, G, \& Schmidt, M. 2001, A\&A, 369, 49
360: \reference{} Morrison, R., \& McCammon, D. 1983, ApJ, 270, 119
361: \reference{} Moustakas, L. A. \& Immler, S. 2004, ApJL submitted, astro-ph0405270
362: \reference{} Nandra, K., George, I. M., Mushotzky, R. F., Turner, T. J., \& Yaqoob, T. 1997, ApJ, 477, 602
363: \reference{} Norman, C. et al. 2002, ApJ, 571, 218
364: \reference{} Piro, L., Matt, G., \& Ricci, R. 1997, A\&AS, 126, 525
365: \reference{} Rhoads, J. E. et al. 2003, AJ, 125, 1006
366: \reference{} Rosati, P., et al. 2002, ApJ, 566, 667
367: \reference{} Steffen, A. T., Barger, A. J., Cowie, L. L., Mushotzky, R. F., \& Yang, Y. 2003, ApJ, 596, L23
368: \reference{} Stern, D., et al. 2002, AJ, 123, 2223
369: \reference{} Szokoly, G. P. et al. 2004, ApJS submitted, astro-ph0312324
370: \reference{} Turner, T. J., George, I. M., Nandra, K., Mushotzky, R. F. 1997a, ApJS, 113, 23
371: \reference{} Turner, T. J., George, I. M., Nandra, K., Mushotzky, R. F. 1997b, ApJ, 488, 164
372: \reference{} Ueda, Y., Akiyama, M., Ohta, K., \& Miyaji, T. 2003, ApJ, 598, 886
373: \reference{} Vaisanen, P., \& Johansson, P. H. 2004, A\&A, 421, 821
374: \reference{} Vignali, C., et al. 2001, AJ. 122, 2143
375: \reference{} Vignali, C., Brandt, W. N., \& Schneider, D. P. 2004, in ASP Conf.
376: Series, AGN Physics with the SDSS, eds. G.T. Richards and P.B. Hall, astro-ph0310659
377: \reference{} Vignali, C., Brandt, W. N., Schneider, D. P., Garmire, G. P., Kaspi, S. 2003a, AJ, 125, 418
378: \reference{} Vignali, C., et al. 2003b, AJ, 125, 2876
379: \reference{} Wang, J. X., et al. 2004, AJ, 127, 213
380: \reference{} Yan, H, et al. 2003, ApJ, 585, 67
381: \reference{} Yaqoob, T. 1997, ApJ, 479, 184
382: \end{references}
383:
384: \begin{deluxetable}{lccc}
385: \tablecaption{CDF-S X-ray sources in Koekemoer et al. 2004 and Yan et al. 2003}
386: \tablecolumns{9}
387: \tablewidth{0pt}
388: \tablehead
389: {
390: \colhead {ID$^a$} & \colhead {R.A(J2000)} & \colhead {Dec.(J2000)} & \colhead{HR$^b$}
391: }
392: \startdata
393: Koekemoer et al. 2004\\
394: 66& 3:32:08.39&-27:40:47.0& $ 0.30^{+0.17}_{-0.17}$\\
395: 69$^c$& 3:32:08.89&-27:44:24.3& $-0.71^{+0.52}_{-0.29}$\\
396: 93& 3:32:13.92&-27:50:00.7& $-0.32^{+0.11}_{-0.68}$\\
397: 133& 3:32:20.36&-27:42:28.5& $-0.01^{+0.28}_{-0.30}$\\
398: 161& 3:32:25.83&-27:51:20.3& $-0.29^{+0.21}_{-0.71}$\\
399: 191& 3:32:33.14&-27:52:05.9& $-0.35^{+0.11}_{-0.65}$\\
400: 216& 3:32:51.64&-27:52:12.8& $0.15^{+0.22}_{-0.23}$\\
401: Yan et al. 2003\\
402: 98$^d$& 3:32:14.67&-27:44:03.4& 0.09\\
403: 140& 3:32:22.44&-27:45:43.9& $-0.39^{+0.25}_{-0.61}$\\
404: 188& 3:32:32.17&-27:46:51.4& $0.39^{+0.15}_{-0.16}$\\
405: 214& 3:32:38.03&-27:46:26.2& $-0.45^{+0.08}_{-0.08}$\\
406: 222& 3:32:39.06&-27:44:39.1& $-0.11^{+0.10}_{-0.10}$
407: \enddata
408: \tablenotetext{a}{X-ray Source ID in Giacconi et al. (2002).}
409: \tablenotetext{b}{
410: The X-ray net counts in different bands used to calculate HR are from
411: Alexander et al. (2003). We subtract the soft band net counts from the
412: total band (0.5 -- 8.0 keV) net counts to calculate the hard band net
413: counts for
414: sources with only upper limits of the hard band net counts in Alexander et al.
415: Errors for this quantity are calculated following the "numerical method"
416: described in \S1.7.3 of Lyons 1991.
417: }
418: \tablenotetext{c}{Also included by Yan et al. 2003.}
419: \tablenotetext{d}{Only 0.5 -- 8.0 keV band net count is available in Alexander
420: et al., and the
421: hardness ratio is derived using the upper limit of the soft and hard band
422: counts, the sum of which is very close to the total count (32.2 vs 27.9).
423: }
424: \end{deluxetable}
425:
426: \begin{figure}
427: \plotone{f1.eps}
428: \caption{
429: The input model spectra for N$_H$ = 10$^{21}$, 10$^{22}$, 10$^{23}$, 10$^{24}$ cm$^{-2}$,
430: and pure reflection spectrum.
431: }
432: \label{spectra}
433: \end{figure}
434:
435: \begin{figure}
436: \plotone{f2.eps}
437: \caption{
438: The predicted X-ray hardness ratio HR at different redshift.
439: The input spectrum is a power law with photon index
440: $\Gamma$ = 1.9, and absorbed by different
441: column densities in the rest frame (N$_H$ = 10$^{21}$, 10$^{22}$, 10$^{23}$,
442: and 10$^{24}$ cm$^{-2}$ respectively) and pure reflection spectrum.
443: The output is the
444: hardness ratio, if observed by $Chandra$. The hardness ratio
445: HR is defined as (H-S)/(H+S), where H and S are $Chandra$ net
446: counts in the soft (0.5 -- 2.0 keV) and hard (2.0 -- 8.0 keV)
447: X-ray band.
448: Along the right ordinate, we mark the photon indices $\Gamma$ of absorption-free
449: power-law spectra which could reproduce the corresponding
450: hardness ratios.
451: The known AGNs at $z > 4$ detected by $Chandra$ are also marked:
452: open circles are optically selected, solid
453: circles are X-ray selected, and stars are radio selected.
454: See Fig. 12 of Szokoly et al. 2004 for a sample of Chandra detected
455: AGN at $z < 4$.
456: }
457: \label{hr}
458: \end{figure}
459:
460: \begin{figure}
461: \plotone{f3.eps}
462: \caption{
463: The X-ray hardness ratio distributions of 41 $Chandra$ detected
464: $z > 4$ AGNs ($z>4$), 19 high-z candidates in LALA Bo\"{o}tes field
465: (LALA), 12 CDF-S sources in Koekemoer et al. (2004) and Yan et al (2003),
466: and 31 CDF-N sources discussed in Barger et al. (2003a). The later three
467: distributions are significantly different from the first one at the
468: level $>$ 99.99\% based on the K-S test.
469: See the text for details.}
470: \end{figure}
471:
472:
473: \end{document}
474:
475: