1: %\documentclass[12pt,preprint]{aastex}
2: \documentclass{emulateapj}
3:
4: \usepackage{apjfonts}
5: %\usepackage{amsmath}
6:
7: \newcommand{\ts}{\textstyle}
8: \newcommand{\tnm}{\tablenotemark}
9: \newcommand{\tnt}{\tablenotetext}
10:
11: \newcommand{\ergs}{ergs s$^{-1}$}
12: \newcommand{\flux}{ergs cm$^{-2}$ s$^{-1}$}
13: \newcommand{\intens}{ergs cm$^{-2}$ s$^{-1}$ deg$^{-2}$}
14: \newcommand{\ctsintens}{counts cm$^{-2}$ s$^{-1}$ deg$^{-2}$}
15: \newcommand{\pers}{s$^{-1}$}
16: \newcommand{\cdens}{cm$^{-2}$}
17: \newcommand{\vdens}{cm$^{-3}$}
18:
19: \newcommand{\hst}{{\it HST}}
20: \newcommand{\xmm}{{\it XMM}}
21: \newcommand{\xmmnewton}{{\it XMM-Newton}}
22: \newcommand{\chandra}{{\it Chandra}}
23: \newcommand{\spitzer}{{\it Spitzer}}
24: \newcommand{\asca}{{\it ASCA}}
25: \newcommand{\rxte}{{\it RXTE}}
26: \newcommand{\rosat}{{\it ROSAT}}
27: \newcommand{\bepposax}{{\it BeppoSAX}}
28: \newcommand{\ginga}{{\it Ginga}}
29: \newcommand{\exosat}{{\it EXOSAT}}
30: \newcommand{\einstein}{{\it Einstein}}
31:
32: \newcommand{\mekal}{{\it MEKAL}}
33:
34: \newcommand{\rxj}{RX J0059.2--7138}
35: \newcommand{\xtej}{XTE J0111.2--7317}
36: \newcommand{\fouru}{4U 1626--67}
37: \newcommand{\aofive}{A 0538--66}
38: \newcommand{\exo}{EXO 053109--6609.2}
39:
40: \newcommand{\phip}{$\phi_{35}$}
41: \newcommand{\tout}{\theta_{\rm out}}
42: \newcommand{\tin}{\theta_{\rm in}}
43: \newcommand{\rout}{r_{\rm out}}
44: \newcommand{\rin}{r_{\rm in}}
45: \newcommand{\logn}{$\log{N}/\log{S}$}
46:
47: \newcommand{\im}{\item}
48:
49: \newcommand{\al}{\alpha}
50: \newcommand{\be}{\beta}
51: \newcommand{\ta}{\theta}
52:
53: \newcommand{\alfven}{$\rm{Alfv\grave{e}n}$}
54:
55: \newcommand{\alm}{$\alpha$}
56: \newcommand{\bem}{$\beta$}
57: \newcommand{\tam}{$\theta$}
58: \newcommand{\dgr}{$^{\circ}$}
59: \newcommand{\average}[1]{\ensuremath{\langle#1\rangle} }
60:
61: \newcommand{\rstar}{R_{\star}}
62: \newcommand{\msun}{M_{\sun}}
63:
64: \newcommand{\W}{\hphantom{0}}
65: \newcommand{\Wp}{\hphantom{.}}
66: \newcommand{\Wf}{\hphantom{\footnotemark{a}}}
67:
68: \newcommand{\pso}{$\phi_{\rm SO}$}
69:
70: \newcommand{\z}{$z_{850}$}
71: \newcommand{\B}{$B_{435}$}
72: \newcommand{\V}{$V_{606}$}
73: \newcommand{\iband}{$i_{775}$}
74: \newcommand{\miz}{$i_{775}-z_{850}$}
75: \newcommand{\mBV}{$B_{435}-V_{606}$}
76:
77: \newcommand{\hm}{\citetalias{hick06a}}
78: \newcommand{\hmp}{\citepalias{hick06a}}
79:
80:
81: \begin{document}
82:
83: \title{Resolving the unresolved cosmic X-ray background in the {\it
84: Chandra} Deep Fields}
85: \shorttitle{RESOLVING THE COSMIC X-RAY BACKGROUND}
86: \shortauthors{HICKOX \& MARKEVITCH}
87: \author{Ryan C. Hickox\altaffilmark{1}}
88: \author{Maxim Markevitch\altaffilmark{1}}
89: \altaffiltext{1}{Harvard-Smithsonian Center for Astrophysics, 60 Garden Street,
90: Cambridge, MA 02138; rhickox@cfa.harvard.edu, maxim@head.cfa.harvard.edu.}
91:
92: \slugcomment{Accepted for publication in {\it Astrophysical Journal Letters}}
93:
94: \begin{abstract}
95: We present a measurement of the surface brightness of the cosmic X-ray
96: background (CXB) in the \chandra\ Deep Fields, after excluding all
97: detected X-ray, optical and infrared sources. The work is motivated
98: by a recent X-ray stacking analysis by Worsley and collaborators,
99: which showed that galaxies detected by \hst\ but not by \chandra\ may
100: account for most of the unresolved CXB at $E>1$ keV. We find that
101: after excluding \hst\ and \spitzer\ IRAC sources, some CXB still
102: remains, but it is marginally significant: $(3.4\pm1.4)\times10^{-13}$
103: \intens\ in the 1--2 keV band and $(4\pm9)\times10^{-13}$ \intens\ in
104: the 2--5 keV band, or $7\%\pm3\%$ and $4\%\pm9\%$ of the total CXB,
105: respectively. Of the 1--2 keV signal resolved by the \hst\ sources,
106: $34\%\pm 2\%$ comes from objects with optical colors typical of
107: ``normal'' galaxies (which make up 25\% of the \hst\ sources), while
108: the remaining flux comes from
109: objects with colors of starburst and irregular galaxies. In the 0.65--1 keV band (just above the bright Galactic
110: \ion{O}{7} line) the remaining diffuse intensity is
111: $(1.0\pm0.2)\times10^{-12}$ \intens. This flux includes emission
112: from the Galaxy as well as from the hypothetical warm-hot
113: intergalactic medium (WHIM), and provides a conservative upper limit
114: on the WHIM signal that comes interestingly close to theoretical
115: predictions.
116:
117:
118: \end{abstract}
119: \keywords{methods: data analysis --- X-rays: diffuse background --- X-rays: galaxies}
120:
121:
122: \defcitealias{hick06a}{HM06}
123:
124: \section{Introduction}
125: The cosmic X-ray background (CXB) is known to be primarily
126: the integrated emission from X-ray point sources,
127: mostly active galactic nuclei \citep[e.g.,][]{hasi05,baue04}. Recent studies
128: have shown that the \chandra\ Deep Fields (CDFs) North and South, the
129: most sensitive observations of the X-ray sky to date, resolve $\sim$80\%
130: of the extragalactic CXB at 1--2 keV \citep[][hereafter
131: HM06]{more03, wors05,hick06a}, and
132: $\sim$50\% at 7 keV \citep{wors05}. While the majority of the CXB at
133: $E<5$ keV has been accounted for, the nature of the still-unresolved
134: component (the absolute intensity of which we derived in \hm) remains unclear.
135:
136: Part of the unresolved CXB is due to faint point
137: sources that have yet to be detected in X-rays. However, \hm\ showed
138: that an extrapolation of the observed \logn\ curve to zero fluxes
139: falls far short of the unresolved CXB. To account for it, the slope of the
140: relation must steepen below the CDF-N X-ray source detection limit of
141: $S_{\rm 0.5-2\; \rm keV}= 2.4\times10^{-17}$ \flux. This is not
142: implausible; \citet{baue04} showed that the CDFs contain a population
143: of normal and starburst galaxies with $S_{\rm 0.5-2\; \rm keV}
144: <10^{-16}$ \flux\ whose \logn\ is indeed steeper, and which should
145: dominate the source counts at $S_{\rm 0.5-2\; \rm keV}<10^{-17}$
146: \flux. These faint sources, while not detected in X-rays, can be
147: associated with optical or infrared sources found in the
148: CDFs. A recent X-ray stacking
149: analysis of {\it Hubble Space Telescope} (\hst) sources in the CDFs
150: has found that their X-ray emission can account for most of the
151: unresolved $E=1-6$ keV CXB \citep{wors06}.
152:
153: Still, some of the unresolved flux should be truly diffuse. At $E<1$
154: keV, there is flux from the Local
155: Bubble \citep{snow04} as well as charge exchange local to the Sun
156: \citep[e.g.,][]{crav00, warg04}. In addition, some signal may come
157: from the warm-hot intergalactic medium (WHIM) with temperatures
158: $10^5$--$10^7$ K, that is thought to comprise the majority of the
159: baryons in the local Universe
160: \citep[e.g.,][]{cen99,dave01}. Models of the WHIM predict a signal
161: dominated by line emission from oxygen and iron with prominent lines
162: at 0.5--0.8 keV, but some flux up to 1.5 keV \citep{phil01,fang05a,
163: ursi06}. To date the WHIM has proved challenging to observe, although
164: there have been reports of detections of WHIM X-ray absorption lines
165: \citep[e.g,][however see Kaastra et~al.\ 2006]{nica05}, as well as
166: excess soft X-ray emission around galaxies that has been attributed to the
167: WHIM \citep[e.g.,][]{solt06}. Diffuse emission could
168: also come from more exotic sources such as decay of dark matter particles
169: \citep[e.g.,][]{abaz07}.
170:
171: \nocite{kaas06}
172:
173: In this Letter, we constrain the diffuse CXB by directly measuring
174: its absolute intensity in deep \chandra\ pointings after the removal
175: of all sources detected in deep X-ray, optical, and IR observations.
176: Our analysis is complementary to that of \citet{wors06}, and is a
177: follow-up to \citetalias{hick06a}, in which we excluded only the X-ray
178: sources.
179:
180: \section{Data}
181: \label{data}
182: {\it Chandra} X-ray observations of the CDFs North and South (CDF-N
183: and CDF-S) have respective total exposure times of roughly 2 Ms and 1
184: Ms. The CDF-N data consist of two subsets, observed in in Very Faint
185: (VF) or Faint (F) telemetry mode. Each of the three subsets (CDF-N
186: VF, CDF-N F, and CDF-S) contains 6 to 9 individual observations with
187: exposure times 30 to 170 ks each. The datasets are identical to those
188: used in \hm, except for the ACIS calibration updates. For subtraction
189: of the instrumental background, we add the latest ACIS stowed
190: observations, for a total background exposure of 325 ks, compared to
191: 256 ks used in \hm. These updates caused no significant change to the
192: CXB fluxes.
193:
194: In \hm, we measured the CXB intensity after excluding X-ray
195: detected sources from the catalog of \citet{alex03}. To further
196: exclude the contributions of X-ray faint objects, we use sources detected in the optical and
197: IR observations of the Great Observatories Origins Deep Survey
198: \citep[GOODS,][]{dick03}. Optical and near-IR data were taken using
199: the Advanced Camera for Surveys (ACS) on \hst\ in the \B, \V,
200: \iband, and \z\ bands \citep{giav04}. We use the public
201: catalog\footnotemark\ based on detections in the \z\ band ($\lambda \sim 8300-9500$ \AA), with a limiting magnitude
202: of $z_{850}\simeq 27$. The \hst\ catalogs for CDF-N and CDF-S
203: each contain $\sim$4500 objects within the area for which we extract CXB
204: spectra (inside a 3.2\arcmin\ circle, and away from detected X-ray
205: sources, see \S\ \ref{analysis}).
206:
207: \footnotetext{http://archive.stsci.edu/pub/hlsp/goods/catalog\_r1/h\_r1.1z\_readme.html}
208:
209: We also exclude sources detected with the Infrared Array Camera (IRAC)
210: on {\it Spitzer}, in four bands centered on 3.6, 4.5, 5.8, and 8
211: $\mu$m. There is no published catalog currently available for these
212: sources, so we have performed our own source detection using the
213: calibrated IRAC images\footnotemark. Sources were detected using the
214: {\tt SExtractor} code \citep{bert96} in all four bands. In the most
215: sensitive band (3.6 $\mu$m), we detect $\sim$1000 sources in the X-ray
216: extraction region to a 5$\sigma$ limiting flux of $\sim$0.2 $\mu$Jy.
217: Most ($>$95\%) of the IRAC sources have counterparts within 1\arcsec\
218: in the \hst\ \z\ catalog.
219:
220:
221:
222: \footnotetext{http://data.spitzer.caltech.edu/popular/goods/Documents\\/goods\_dataproducts.html}
223: \setcounter{footnote}{1}
224:
225: We also considered \spitzer\ Multiband Imaging
226: Photometer (MIPS) 24 $\mu$m sources\footnotemark\ and for CDF-N, Very
227: Large Array (VLA) 21 cm sources \citep{rich00}, with limiting fluxes
228: of 80 and 40 $\mu$Jy, respectively. All the sources within our
229: X-ray extraction regions have \hst\ or IRAC counterparts, except
230: for 2--4 of the 80--90 MIPS sources and 1 of the 6 VLA sources, so we
231: do not include these data in the analysis.
232:
233:
234: \begin{figure}
235: \plotone{f1.eps}
236: \caption{Smoothed 0.5--8 keV image of CDF-N, showing the source exclusion
237: regions. The large circle is the 3.2\arcmin\ extraction region,
238: medium-size circles and ellipses show X-ray point and extended
239: sources, and small circles show \hst\ sources. \label{figim}}
240: \end{figure}
241:
242:
243: \section{Analysis}
244: \label{analysis}
245: We calculate the CXB surface brightness as discussed in detail in \S\
246: 5 of \hm, with only minor differences. For the \hst\ and IRAC sources
247: not detected in X-rays, we use smaller exclusion regions with radii
248: equal to $r_{90}$, the X-ray point-spread function (PSF) 90\% enclosed flux
249: radius at 1.5 keV (see Eqn.\ 1 of \hm), compared to 4.5--$9r_{90}$ for
250: X-ray detected sources. Due to the large number of optical and IR
251: sources, at large off-axis angles there is little area remaining after
252: source exclusion. Therefore we only extract spectra from a circle
253: around the optical axis with radius 3.2\arcmin\ (where
254: $r_{90}=2.2$\arcsec), compared to 5\arcmin\ in \hm. The solid angle
255: we use here is 0.45 of that in \hm\ when we exclude only X-ray
256: sources, and 0.32 when we exclude all \hst\ and IRAC sources.
257:
258:
259: %\input{jnk1}
260:
261:
262:
263: \begin{figure}
264: \epsscale{1.15}
265: \plotone{f2.eps}
266: \caption{CXB spectrum for the CDF-N VF subset, with
267: all sources (\chandra, \hst\, and IRAC) excluded. The spectrum is
268: fit with an APEC model with $kT=0.20$ keV plus a power law with
269: $\Gamma=1.5$. Black points show the data included in the fit (see \hm), and
270: the red dashed line shows the best-fit model for the CDF-N VF
271: spectrum excluding only X-ray sources \hmp. \label{figspecvf}}
272: \end{figure}
273:
274: For each of the datasets, we fit the spectrum of the remaining area
275: with a model consisting of an optically-thin thermal component (APEC)
276: with solar abundances and $kT\simeq 0.15$--0.2 keV, to account for
277: emission from the Galaxy as well as any diffuse thermal CXB. We also
278: include a power law component that dominates at $E>1$ keV, and fix the
279: photon index at $\Gamma=1.5$, the best-fit value from \hm\ (we do not
280: vary $\Gamma$ as in \hm, due to poorer statistics). These models do
281: not necessarily provide a physical description of the emission, but
282: they are sufficient for our purpose of calculating fluxes. A CDF-N VF
283: spectrum with the model fit is shown in Fig.\ \ref{figspecvf}.
284:
285:
286:
287:
288: \begin{deluxetable}{lcccc}
289: \tablecaption{CXB surface brightness after excluding all sources}
290: \tablewidth{3in}
291: \tablehead{
292: \colhead{Energy} &
293: \multicolumn{3}{c}{CXB intensity\tnm{a}} &
294: \colhead{CXB frac.\tnm{b}} \\
295: \colhead{(keV)} &
296: \colhead{CDF-S} &
297: \colhead{CDF-N (VF+F)} &
298: \colhead{Average} &
299: \colhead{(\%)}}
300: \startdata
301: 0.5--2\tnm{c} & $45\pm 8$\W & \Wf$41\pm 11$\tnm{d}\W\Wp & \Wf$43\pm 7$\tnm{d}\W\Wp & \nodata \\
302: 0.65--1\tnm{c} & $11\pm 2$\W & \Wf$9\pm 3$\tnm{d}\Wp & \Wf$10\pm 2$\tnm{d}\W\Wp & \nodata \\
303: 1--2 & $ 5.0\pm 1.8$ & $ 2.4\pm 1.6$ & $ 3.4\pm 1.4$ & $ 7.3\pm 3.0$ \\
304: 1--1.3 & $ 1.7\pm 0.7$ & $ 1.0\pm 0.7$ & $ 1.3\pm 0.6$ & $ 8.3\pm 3.8$ \\
305: 2--5 & \W$ 7\pm12$ & \W$ 2\pm10$ & $ 4\pm 9$ & $ 4\pm 9$ \\
306: 2--8 & $14\pm39$ & \W$ 4\pm30$ & \W$ 7\pm26$ & \W$ 4\pm16$
307: \enddata
308: \tablenotetext{a}{Units are $10^{-13}$ \intens, errors are 68\%.}
309: \tablenotetext{b}{Percent of total CXB, assuming a
310: power-law CXB spectrum with $\Gamma=1.4$ and normalization $10.9\pm0.5$ photons cm$^{-2}$
311: s$^{-1}$ sr$^{-1}$ keV$^{-1}$ at 1 keV (\hm).}
312: \tablenotetext{c}{Intensities for $E<1$ keV are dominated by Galactic
313: and local line emission that varies with time and position. These are
314: not extragalactic values.}
315: \tablenotetext{d}{Does not include CDF-N VF, to exclude the local temporary excess.}
316: \end{deluxetable}
317:
318: \begin{figure*}
319: \plottwo{f3a.eps}{f3b.eps}
320: \vskip-0.3cm
321: \plottwo{f3c.eps}{f3d.eps}
322: \caption{(a) Solid angles of the X-ray extraction regions, and (b--d)
323: unresolved CXB surface brightness, for subsets of excluded \chandra,
324: \hst, and IRAC sources. In all panels, points on the far left show
325: the value when only X-ray sources are excluded (as in \hm, but
326: using the smaller 3.2 \arcmin\ region). Points with \z\ labels on
327: the $x$-axis
328: correspond to the additional exclusion of \hst\ sources brighter
329: than \z\ (note that $z_{850}=27$ is only an approximate detection
330: limit for all HST sources). Points on the right show intensities
331: excluding X-ray sources plus only \hst\ ``normal'' galaxies (\S\
332: \ref{whichsource}), only IRAC 3.6 $\mu$m sources, and finally all \hst\ and
333: IRAC sources. For the 0.65--1 keV band we show the CDF-N VF and
334: CDF-N F subsets separately, as time variability in the local
335: emission \citep[apparently charge exchange in the Solar wind,
336: e.g.,][]{warg04} caused these intensities to differ. For clarity,
337: errors are shown only on the far left and right points. For (c) and
338: (d), on the right axis we show the corresponding fraction of the
339: total CXB, for which we assume a power law
340: spectrum with $\Gamma=1.4$ and normalization 10.9 photons cm$^{-2}$
341: s$^{-1}$ sr$^{-1}$ at 1 keV \hmp. \label{figres}}
342: \end{figure*}
343:
344:
345:
346: For each spectrum, we calculate fluxes in the 0.5--2 keV, 1--2
347: keV, 2--8 keV, and 2--5 keV bands, as well as the 0.65--1 keV and 1--1.3 keV
348: bands, where the hypothetical WHIM signal or galaxy group emission are
349: best distinguished from the diffuse Galactic foreground (see \S\
350: \ref{whim}). Errors on the observed flux include statistical errors
351: in the sky and background count rates, and a 2\% systematic
352: uncertainty in the background normalization (\S\ 4 of \hm). We
353: carefully account for background uncertainties. Because of
354: differences in pointing of up to 2\arcmin\ and the small X-ray
355: extraction area, the composite stowed background spectrum includes 3
356: times (for CDF-N) and 1.5 times (for CDF-S) as many statistically
357: independent photons as do the spectra for each separate observation.
358: We accordingly increase the effective background exposure times by
359: these factors. We conservatively neglected this effect in
360: HM06, as it was much smaller due to the larger extraction area and
361: fewer excluded sources.
362:
363:
364:
365: A further small correction is necessary to subtract the point source
366: flux that is scattered outside our exclusion regions. For each \hst\
367: or IRAC source, 10\% of the X-ray flux should lie outside $r_{90}$.
368: However, because many regions overlap, the
369: contribution of this scattered flux is smaller. Using a model of the
370: \chandra\ PSF, we estimate that when all \hst\ and IRAC sources are
371: excluded with radius $r_{90}$, 5\% (0.5--2 keV) and 8\% (2--8 keV) of
372: their flux should remain in our unresolved spectrum. Likewise, 0.3\%
373: and 0.6\% of the flux from detected
374: X-ray sources (for which we use larger exclusion regions) will
375: be scattered into the spectrum. We account for both these small
376: contributions in the final unresolved intensities. As a check, we
377: tried exclusion regions of radius $1.5r_{90}$ and obtained
378: identical results.
379:
380:
381:
382:
383: \section{Results}
384: We first calculate the CXB intensity after excluding \hst\ subsamples with various limiting magnitudes from
385: $z_{850}<21-26$, and finally excluding all \hst\ sources
386: ($z\lesssim27$). We also try excluding only ``normal'' galaxies, selected
387: based on their \hst\ colors (see \S\ \ref{whichsource}), only IRAC
388: 3.6 $\mu$m sources, and finally all \hst\ and IRAC sources. In every
389: case we also exclude all detected X-ray sources. The results are
390: shown in Fig.\ \ref{figres}, and the residual CXB intensities excluding
391: all sources are given in Table
392: 1. We also calculate the mean CXB intensity for all three subsets, including a detailed error
393: analysis as described in \S\ 8.2.3 of \hm.
394:
395:
396: Fig.\ \ref{figres} shows that excluding faint \hst\ sources has little
397: effect on the CXB at $E<1$ keV, where it is dominated by the truly
398: diffuse Galactic and local emission. At $E>1$ keV, as fainter objects
399: are excluded, the surface brightness of the CXB decreases, indicating
400: that the \hst\ sources do indeed contribute to the X-ray unresolved
401: CXB. In the 1--2 keV band, the remaining intensity after excluding
402: all \hst\ and IRAC sources is above zero at only 1.5$\sigma$ for CDF-N
403: and 2.8$\sigma$ for CDF-S. The CXB intensity decreases even as we
404: excluded the faintest \hst\ sources, suggesting that if we were able
405: to go slightly deeper, we would resolve the 1--2 keV CXB to within our
406: absolute uncertainties. The X-ray undetected \hst\ sources contribute
407: 1--2 keV intensities of $(5.7\pm0.3)\times10^{-13}$ and
408: $(4.3\pm0.3)\times10^{-13}$ \intens\ in CDF-N and CDF-S, respectively,
409: or $59\%\pm9\%$ and
410: $45\%\pm7\%$ of the unresolved CXB from \hm.
411: The fluxes resolved by \hst\ sources (given here and in
412: \S~\ref{whichsource}) represent {\em differences} between the points
413: in Fig.~\ref{figres}, and do not depend strongly on the instrumental
414: background. Thus their errors are much smaller than the uncertainties
415: on the absolute surface brightness shown in Fig.~\ref{figres}.
416: Stacking the X-rays from the same \hst\ sources, \citet{wors06} found larger
417: intensities of $(8.4\pm0.5)\times10^{-13}$ and
418: $(9.3\pm0.7)\times10^{-13}$ \intens\ for CDF-N and CDF-S,
419: respectively. The cause of this difference is not immediately clear,
420: but may be related to different accounting for overlap between the
421: \hst\ source regions.
422:
423:
424: \begin{figure}
425: \epsscale{1.1}
426: \plotone{f4.eps}
427: \caption{Color selection of ``normal'' galaxies from \hst\ data. All
428: points show $B_{435}-V_{606}$ vs. $i_{775}-z_{850}$ colors for
429: sources in the UDF with $z_{850}<27$. Sources are classified as
430: elliptical and spiral galaxies (red circles) or starburst and Im
431: galaxies (black dots) from photo-$z$ analysis \citep{coe06}. Lines
432: show our criteria to select ``normal'' galaxies from the GOODS data.
433: \label{figcol}}
434: \end{figure}
435:
436:
437:
438: In the 2--5 keV band, the \hst\ and IRAC sources account for the
439: entire unresolved CXB within the large absolute uncertainties.
440: Similarly, \citet{wors06} found that the stacked flux from
441: \hst\ sources is consistent with completely resolving the 2--6
442: keV CXB.
443:
444: \section{Discussion}
445: \subsection{Which sources resolve the CXB?}
446: \label{whichsource}
447:
448:
449:
450:
451: \citet{baue04} and \hm\ show that AGN likely contribute a relatively
452: small fraction of the CXB at X-ray fluxes below the CDF limit, where
453: the X-ray source population is dominated by starburst and normal
454: galaxies. To determine what class of galaxies gives the largest
455: contribution to the unresolved CXB, we use \hst\ colors to
456: approximately divide the \z\ sources into two subsets by galaxy type.
457: While we do not have optical classifications for the \hst\ GOODS
458: galaxies, such a classification exists for the \hst\ Ultra Deep Field
459: (UDF), which covers a small region roughly 3\arcmin\
460: from the CDF-S aimpoint with a flux limit of $z_{850}\simeq30$. The
461: UDF was observed in the same four ACS bands as the wider GOODS fields,
462: and \citet{coe06} performed a detailed photometric redshift analysis
463: of the UDF sources, which classified each galaxy with one of eight
464: templates (elliptical, Sab, Sbc, Im, and four types of starburst).
465:
466:
467:
468: %\input{jnk2}
469:
470:
471: In Fig.\ \ref{figcol}, we show \mBV\ vs. \miz\ colors for
472: \citet{coe06} sources with $z_{850}<27$, corresponding to the flux
473: limits of the larger GOODS fields. In this diagram, normal
474: (elliptical and spiral) galaxies occupy a different region from
475: starburst and Im galaxies, with approximate boundaries
476: $i_{775}-z_{850}=0.2$ and $B_{435}-V_{606}=0.9-(i_{775}-z_{850})$, as
477: shown in Fig. \ref{figcol}. Of the 255 ``normal'' galaxies in the UDF
478: sample, 201 (73\%) are in this region, while only 52 (21\%) of the 253
479: galaxies in this region have starburst or Im classifications and would
480: be mis-classified by this criterion. In the GOODS catalog,
481: these criteria classify 25\% of $z_{850}$ sources as ``normal'' galaxies and
482: 70\% as ``starburst/Im'' objects (5\% of the
483: sources do not have detections in all four ACS bands). The ``normal''
484: galaxies contribute $34\%\pm 2\%$ of the 1--2 keV intensity
485: resolved by all the \hst\ sources, and so have higher average
486: X-ray fluxes than other X-ray undetected \hst\ objects.
487:
488: We also examine the contribution from the 3.6
489: $\mu$m sources, which are on average only 0.25 magnitudes brighter in \z\
490: than the \hst\ sources, but account for a disproportionately large
491: fraction of the unresolved CXB. The $\sim$1000 3.6 $\mu$m sources
492: contribute $63\%\pm2\%$ of the total 1--2 keV CXB resolved by
493: all \hst\ and IRAC sources. In comparison, the 1000 brightest \hst\
494: sources (which have $z_{850}<24.5$, and of which only 20\% are
495: IRAC sources) account for $53\%\pm2\%$. Therefore,
496: the X-ray flux from undetected X-ray sources is apparently more highly
497: correlated with 3.6 $\mu$m flux than with \z\ flux. This may be an
498: important clue to the nature of these faint X-ray sources, although
499: a detailed discussion is beyond the scope of this Letter. In a
500: forthcoming paper, we will determine how bright these undetected sources can be
501: in the X-rays and whether \chandra\ could detect them in a reasonable
502: longer exposure.
503:
504:
505: \subsection{Comparison to WHIM predictions}
506:
507:
508: \label{whim}
509: While most of the remaining CXB signal should be due to Galactic or
510: local diffuse emission and unresolved X-ray point sources, our results
511: provide conservative upper limits on emission from the WHIM (Cen et
512: al.\ 2000 and later works). The brightest WHIM emission feature
513: should be a redshifted \ion{O}{7} 0.57 keV line, but it cannot be
514: distinguished from the bright Galactic \ion{O}{7} line without much
515: better ($\sim$1 eV) spectral resolution. However, the hotter WHIM
516: phases with $kT\sim 0.2-0.5$ keV would also emit significant
517: \ion{Fe}{17} lines at $E\approx 0.72-0.83$ keV (rest-frame). Thus, the ratio of
518: WHIM emission to Galactic foreground (which is well-approximated by a
519: $kT\approx 0.15-0.2$ keV thermal spectrum, Fig.\ 1) may be higher at
520: $E\gtrsim 0.6$ keV.
521:
522: We can compare our 0.65--1 keV signal to predictions from cosmological
523: simulations. \citet{fang05a} find WHIM intensities of $\sim3\times
524: 10^{-8}$, $3\times 10^{-10}$, and $3\times 10^{-14}$ \intens\ toward a
525: galaxy group, filament, and void, respectively. \citet{ronc06a}
526: obtained a sky-averaged 0.65--1 keV signal of $5\times10^{-13}$
527: \intens\ after excluding galaxy groups and clusters. (For these two
528: works, we have approximately converted fluxes from 0.1--1 keV and
529: 0.5--2 keV to 0.65--1 keV). The CDFs were selected to lie away from
530: bright groups, but if they do not lie in voids, then our observed
531: signal, $(1.0\pm0.2)\times10^{-12}$ \intens, comes interestingly close to the
532: above predictions. Galactic superwinds may increase the WHIM
533: metallicity \citep{cen06}, so the line emission may be even higher. At
534: higher energies, our 1--2 keV signal is 15 times higher than the
535: predicted sky-averaged WHIM intensity of $0.19\times10^{-13}$ \intens\
536: \citep{phil01}, so does not give an interesting limit. For a more
537: quantitative comparison, it will be necessary to excise the regions
538: around individual galaxies in the WHIM simulations similarly to our
539: analysis. With such a simulation, one may be able to place
540: interesting constraints on the distribution and metallicity of the
541: WHIM in the CDFs.
542:
543:
544:
545:
546:
547:
548: \acknowledgements
549:
550: We thank J.P. Ostriker and L.A. Phillips for fruitful discussions, and
551: the referee, Franz Bauer, for helpful comments. RCH was supported by
552: NASA GSRP and Harvard Merit Fellowships, and MM by NASA contract
553: NAS8-39073 and \chandra\ grant G04-5152X.
554:
555: \begin{thebibliography}{}
556:
557: \bibitem[{{Abazajian} {et~al.}(2007){Abazajian}, {Markevitch}, {Koushiappas},
558: \& {Hickox}}]{abaz07}
559: {Abazajian}, K.~N., {Markevitch}, M., {Koushiappas}, S.~M., \& {Hickox}, R.~C.
560: 2007, \prd, in press, astro-ph/0611144
561:
562: \bibitem[{{Alexander} {et~al.}(2003){Alexander}, {Bauer}, {Brandt},
563: {Schneider}, {Hornschemeier}, {Vignali}, {Barger}, {Broos}, {Cowie},
564: {Garmire}, {Townsley}, {Bautz}, {Chartas}, \& {Sargent}}]{alex03}
565: {Alexander}, D.~M., et~al.\ 2003, \aj, 126, 539
566:
567: \bibitem[{{Bauer} {et~al.}(2004){Bauer}, {Alexander}, {Brandt}, {Schneider},
568: {Treister}, {Hornschemeier}, \& {Garmire}}]{baue04}
569: {Bauer}, F.~E., {Alexander}, D.~M., {Brandt}, W.~N., {Schneider}, D.~P.,
570: {Treister}, E., {Hornschemeier}, A.~E., \& {Garmire}, G.~P. 2004, \aj, 128,
571: 2048
572:
573: \bibitem[{{Bertin} \& {Arnouts}(1996)}]{bert96}
574: {Bertin}, E. \& {Arnouts}, S. 1996, \aaps, 117, 393
575:
576: \bibitem[{{Cen} \& {Ostriker}(1999)}]{cen99}
577: {Cen}, R. \& {Ostriker}, J.~P. 1999, \apj, 514, 1
578:
579: \bibitem[{{Cen} \& {Ostriker}(2006)}]{cen06}
580: ---. 2006, \apj, 650, 560
581:
582: \bibitem[{{Coe} {et~al.}(2006){Coe}, {Ben{\'{\i}}tez}, {S{\'a}nchez}, {Jee},
583: {Bouwens}, \& {Ford}}]{coe06}
584: {Coe}, D., {Ben{\'{\i}}tez}, N., {S{\'a}nchez}, S.~F., {Jee}, M., {Bouwens},
585: R., \& {Ford}, H. 2006, \aj, 132, 926
586:
587: \bibitem[{{Cravens}(2000)}]{crav00}
588: {Cravens}, T.~E. 2000, \apjl, 532, L153
589:
590: \bibitem[{{Dav{\'e}} {et~al.}(2001){Dav{\'e}}, {Cen}, {Ostriker}, {Bryan},
591: {Hernquist}, {Katz}, {Weinberg}, {Norman}, \& {O'Shea}}]{dave01}
592: {Dav{\'e}}, R., et~al.\ 2001, \apj, 552, 473
593:
594: \bibitem[{{Dickinson} {et~al.}(2003){Dickinson}, {Giavalisco}, \& {The GOODS
595: Team}}]{dick03}
596: {Dickinson}, M., {Giavalisco}, M., \& {The GOODS Team}. 2003, in The Mass of
597: Galaxies at Low and High Redshift, ed. R.~{Bender} \& A.~{Renzini}
598: (Berlin; New York: Springer), 324
599:
600: \bibitem[{{Fang} {et~al.}(2005){Fang}, {Croft}, {Sanders}, {Houck}, {Dav{\'e}},
601: {Katz}, {Weinberg}, \& {Hernquist}}]{fang05a}
602: {Fang}, T., {Croft}, R.~A.~C., {Sanders}, W.~T., {Houck}, J., {Dav{\'e}}, R.,
603: {Katz}, N., {Weinberg}, D.~H., \& {Hernquist}, L. 2005, \apj, 623, 612
604:
605: \bibitem[{{Giavalisco} {et~al.}(2004){Giavalisco}, {Ferguson}, {Koekemoer},
606: {Dickinson}, {Alexander}, {Bauer}, {Bergeron}, {Biagetti}, {Brandt},
607: {Casertano}, {Cesarsky}, {Chatzichristou}, {Conselice}, {Cristiani}, {Da
608: Costa}, {Dahlen}, {de Mello}, {Eisenhardt}, {Erben}, {Fall}, {Fassnacht},
609: {Fosbury}, {Fruchter}, {Gardner}, {Grogin}, {Hook}, {Hornschemeier}, {Idzi},
610: {Jogee}, {Kretchmer}, {Laidler}, {Lee}, {Livio}, {Lucas}, {Madau},
611: {Mobasher}, {Moustakas}, {Nonino}, {Padovani}, {Papovich}, {Park},
612: {Ravindranath}, {Renzini}, {Richardson}, {Riess}, {Rosati}, {Schirmer},
613: {Schreier}, {Somerville}, {Spinrad}, {Stern}, {Stiavelli}, {Strolger},
614: {Urry}, {Vandame}, {Williams}, \& {Wolf}}]{giav04}
615: {Giavalisco}, M., et~al.\ 2004, \apjl, 600, L93
616:
617: \bibitem[{{Hasinger} {et~al.}(2005){Hasinger}, {Miyaji}, \& {Schmidt}}]{hasi05}
618: {Hasinger}, G., {Miyaji}, T., \& {Schmidt}, M. 2005, \aap, 441, 417
619:
620: \bibitem[{{Hickox} \& {Markevitch}(2006)}]{hick06a}
621: {Hickox}, R.~C. \& {Markevitch}, M. 2006, \apj, 645, 95
622:
623: \bibitem[{{Kaastra} {et~al.}(2006){Kaastra}, {Werner}, {Herder}, {Paerels}, {de
624: Plaa}, {Rasmussen}, \& {de Vries}}]{kaas06}
625: {Kaastra}, J.~S., {Werner}, N., {Herder}, J.~W.~A.~d., {Paerels}, F.~B.~S., {de
626: Plaa}, J., {Rasmussen}, A.~P., \& {de Vries}, C.~P. 2006, \apj, 652, 189
627:
628: \bibitem[{{Moretti} {et~al.}(2003){Moretti}, {Campana}, {Lazzati}, \&
629: {Tagliaferri}}]{more03}
630: {Moretti}, A., {Campana}, S., {Lazzati}, D., \& {Tagliaferri}, G. 2003, \apj,
631: 588, 696
632:
633: \bibitem[{{Nicastro} {et~al.}(2005){Nicastro}, {Mathur}, {Elvis}, {Drake},
634: {Fang}, {Fruscione}, {Krongold}, {Marshall}, {Williams}, \& {Zezas}}]{nica05}
635: {Nicastro}, F., et~al.\ 2005, \nat, 433, 495
636:
637: \bibitem[{{Phillips} {et~al.}(2001){Phillips}, {Ostriker}, \& {Cen}}]{phil01}
638: {Phillips}, L.~A., {Ostriker}, J.~P., \& {Cen}, R. 2001, \apjl, 554, L9
639:
640: \bibitem[{{Richards}(2000)}]{rich00}
641: {Richards}, E.~A. 2000, \apj, 533, 611
642:
643: \bibitem[{{Roncarelli} {et~al.}(2006){Roncarelli}, {Moscardini}, {Tozzi},
644: {Borgani}, {Cheng}, {Diaferio}, {Dolag}, \& {Murante}}]{ronc06a}
645: {Roncarelli}, M., {Moscardini}, L., {Tozzi}, P., {Borgani}, S., {Cheng}, L.~M.,
646: {Diaferio}, A., {Dolag}, K., \& {Murante}, G. 2006, \mnras, 368, 74
647:
648: \bibitem[{{Snowden}(2004)}]{snow04}
649: {Snowden}, S.~L. 2004, in Soft X-ray Emission from Clusters of Galaxies and
650: Related Phenomena (Dordrecht; Norwell, MA: Kluwer Academic Publishers), 103
651:
652: \bibitem[{{So{\l}tan}(2006)}]{solt06}
653: {So{\l}tan}, A.~M. 2006, \aap, 460, 59
654:
655: \bibitem[{{Ursino} \& {Galeazzi}(2006)}]{ursi06}
656: {Ursino}, E. \& {Galeazzi}, M. 2006, \apj, 652, 1085
657:
658: \bibitem[{{Wargelin} {et~al.}(2004){Wargelin}, {Markevitch}, {Juda},
659: {Kharchenko}, {Edgar}, \& {Dalgarno}}]{warg04}
660: {Wargelin}, B.~J., {Markevitch}, M., {Juda}, M., {Kharchenko}, V., {Edgar}, R.,
661: \& {Dalgarno}, A. 2004, \apj, 607, 596
662:
663: \bibitem[{{Worsley} {et~al.}(2006){Worsley}, {Fabian}, {Bauer}, {Alexander},
664: {Brandt}, \& {Lehmer}}]{wors06}
665: {Worsley}, M.~A., {Fabian}, A.~C., {Bauer}, F.~E., {Alexander}, D.~M.,
666: {Brandt}, W.~N., \& {Lehmer}, B.~D. 2006, \mnras, 368, 1735
667:
668: \bibitem[{{Worsley} {et~al.}(2005){Worsley}, {Fabian}, {Bauer}, {Alexander},
669: {Hasinger}, {Mateos}, {Brunner}, {Brandt}, \& {Schneider}}]{wors05}
670: {Worsley}, M.~A., et~al.\ 2005, \mnras, 357, 1281
671:
672:
673: \end{thebibliography}
674:
675:
676:
677: \end{document}
678: