1: \documentclass[12pt,preprint]{aastex}
2: \newcommand{\fxfopt}{$F_{\rm X}/F_{\rm opt}$}
3: \newcommand{\fxfi}{$F_{\rm X}/F_{\rm I}$}
4:
5: \begin{document}
6:
7:
8: \title{``Hidden'' Seyfert 2 Galaxies in the Chandra Deep Field North}
9:
10: \author{Carolin N.\ Cardamone\altaffilmark{1} and Edward C.\ Moran}
11: \affil{Astronomy Department, Wesleyan University, Middletown, CT 06459
12: \\ \medskip{\rm and}\\}
13:
14: \author{Laura E.\ Kay}
15: \affil{Department of Physics and Astronomy, Barnard College, 3009 Broadway,
16: New York, NY 10027}
17:
18: \altaffiltext{1}{Present address: Department of Astronomy, Yale University,
19: P.O.\ Box 208101, New Haven, CT 06520.}
20:
21: \begin{abstract}
22: We have compared the X-ray--to--optical flux ratios (\fxfopt )
23: of absorbed active galactic nuclei (AGNs) in the {\it Chandra\/} Deep Field
24: North (CDF-N) with those of nearby, optically classified Seyfert~2 galaxies.
25: The comparison provides an opportunity to explore the extent to which the
26: local population of absorbed AGNs can account for the properties of the
27: distant, spectroscopically ambiguous sources that produce the hard X-ray
28: background. Our nearby sample consists of 38 objects that well represent
29: the local Seyfert~2 luminosity function. Integrated {\sl UBVRI\/} photometry
30: and broadband X-ray observations are presented. Using these data, we have
31: simulated the \fxfopt\ ratios that local Seyfert~2s would exhibit if they
32: were observed in the redshift range $0.2 \le z \le 1.3$ as part of the CDF-N.
33: In the simulations we account for the effects of redshift on flux measurements
34: in fixed observed-frame bands, and the way the luminosity function of a given
35: population is sampled in a flux-limited survey like the CDF-N. Overall, we
36: find excellent agreement between our simulations and the observed distribution
37: of \fxfopt\ ratios for absorbed AGNs in the CDF-N. Our analysis has thus
38: failed to reveal any physical differences between the local population of
39: Seyfert~2s and CDF-N sources with similar X-ray properties. These results
40: support the hypothesis that the nuclear emission lines of many distant hard
41: X-ray galaxies are hidden in ground-based spectra due to a combination of
42: observational effects: signal-to-noise ratio, wavelength coverage, and dilution
43: by host-galaxy light.
44:
45: \end{abstract}
46:
47: \keywords{galaxies:\ Seyfert --- X-rays:\ diffuse background ---
48: X-rays:\ galaxies}
49:
50:
51: \section{Introduction}
52:
53: Broadband X-ray observations have revealed that many active galactic nuclei
54: (AGNs) are heavily obscured by dense gas and dust located in their host
55: galaxies (e.g., Awaki et al.\ 1991). The selective absorption caused by the
56: obscuring medium flattens (or inverts) the intrinsically steep X-ray spectra
57: of these AGNs, making them attractive candidates for the origin of the hard
58: (2--10 keV) X-ray background (XRB; Setti \& Woltjer 1989). Detailed models
59: based on the observed properties of nearby AGNs have demonstrated that a
60: distant population of obscured objects is indeed capable of accounting for
61: the spectrum and intensity of the hard XRB (Comastri et al.\ 1995; Gilli,
62: Salvati, \& Hasinger 2001; Moran et al.\ 2001). Consistent with this
63: expectation, the X-ray
64: colors of sources detected in extremely deep exposures with the {\it Chandra
65: X-ray Observatory}, which have resolved the majority of the hard XRB, indicate
66: that obscured AGNs are the most prevalent sources at faint hard X-ray fluxes
67: (Alexander et al.\ 2003).
68:
69: Locally, the vast majority of obscured AGNs have the optical spectra of
70: Seyfert~2 galaxies, which are characterized by strong, narrow emission lines.
71: Spectroscopy of faint, hard {\it Chandra\/} sources should, therefore, provide
72: a straightforward means of confirming the Seyfert~2 model for the XRB. But
73: a different picture has emerged: Follow-up studies of distant {\it Chandra\/}
74: sources have instead revealed a significant population of apparently
75: {\it normal\/} galaxies whose starlight-dominated optical spectra have only
76: weak emission lines, if any (e.g., Mushotzky et al.\ 2000; Barger et al.\
77: 2001a, 2001b, 2002; Szokoly et al.\ 2004). Many such sources have the X-ray
78: properties of Seyfert~2 galaxies, but they seem to lack the corresponding
79: optical emission-line signatures.
80:
81: There are several viable explanations for the normal optical appearance of
82: distant absorbed AGNs. One possibility is that moderately luminous AGNs in
83: the past tend to be more heavily obscured than similar objects in the local
84: universe (Barger et al.\ 2001a, 2005). A higher covering factor of the nuclear
85: obscuration would reduce the illumination of the narrow emission-line region
86: by the ionizing continuum, resulting in weaker narrow lines. Alternatively,
87: extranuclear dust may play a greater role in obscuring our view of the
88: narrow emission-line regions of distant objects (Rigby et al.\ 2006). Yet
89: another possibility is that distant AGNs may accrete predominantly in a
90: radiatively inefficient mode, whereby they produce significant hard X-ray
91: emission but far less of the soft X-ray and ultraviolet flux that is chiefly
92: responsible for the ionization of the narrow-line gas (Yuan \& Narayan 2004).
93:
94: As an alternative to these scenarios,
95: Moran, Filippenko, \& Chornock (2002) have suggested that the limitations of
96: ground-based observing may be the culprit. The small angular sizes
97: of distant sources cause their ground-based spectra to be dominated by light
98: from stars and/or \ion{H}{2} regions in the host galaxy, which can mask the
99: emission lines associated with their nuclear activity. Integrated spectra of
100: local Seyfert~2s confirm that host-galaxy dilution would alter many of their
101: spectroscopic classifications if they were observed at modest redshifts with
102: ground-based facilities (Moran et al.\ 2002). Still, the extent to which this
103: dilution affects the demographics of the distant X-ray galaxy population has
104: yet to be demonstrated. Ultimately, a determination of whether the optically
105: normal appearance of distant absorbed AGNs is largely physical or observational
106: in origin has important implications for the nature of supermassive black
107: holes and their environments at earlier epochs.
108:
109: Unfortunately, distant X-ray galaxies tend to be faint at all wavelengths,
110: which limits the amount and quality of information we have about their
111: properties. For example, over half of the X-ray sources detected in the
112: 2~Ms {\it Chandra\/} Deep Field North (CDF-N; Alexander et al.\ 2003; Barger
113: et al.\ 2003) have optical counterparts that are fainter than $R = 23$.
114: Clearly, high-quality optical spectra can only be obtained for the small
115: fraction of relatively bright sources in that field. On the other hand,
116: broadband magnitudes and colors have been measured for nearly all of the
117: CDF-N sources. X-ray--to--optical flux ratios (\fxfopt), therefore, offer
118: one of the best handles we have on the nature of these objects. It has been
119: shown that the \fxfopt\ ratio broadly discriminates between various classes
120: of celestial X-ray sources (e.g., Stocke et al.\ 1991), in particular,
121: between luminous AGNs and truly normal (or quiescent) galaxies. Thus, a
122: comparison of the \fxfopt\ ratios of the optically normal, X-ray--bright
123: objects that have turned up in the deep {\it Chandra\/} surveys to those
124: of local active galaxies with similar high-energy properties could be very
125: informative. For instance, if host-galaxy dilution is generally not a
126: factor, we might expect the deficit of nuclear emission (line and continuum)
127: in the absorption or accretion-mode scenarios described above to lead to
128: systematically higher \fxfopt\ ratios in the distant population.
129:
130: A fair comparison of the \fxfopt\ ratios of nearby and distant objects
131: requires the consideration of several important factors. First, samples of
132: local and high-redshift AGNs are typically defined in very different ways.
133: Nearby samples contain objects recognized as AGNs for a variety of reasons
134: (e.g., X-ray brightness, strength of their emission in some other region of
135: the spectrum, optical emission-line properties, etc.) whereas distant X-ray
136: galaxies are usually identified on the basis of a sole property: detection
137: as an X-ray source. In addition, the volume surveyed in flux-limited studies
138: such as the CDF-N is a sharp function of luminosity, which leads to a deficit
139: of low-luminosity sources and an over-representation of (rare) high-luminosity
140: objects in the derived source catalogs. Thus, nearby and distant AGN samples
141: may contain inherently different types of objects and/or similar objects that
142: are drawn largely from different portions of the AGN luminosity function.
143: Another complication is that different portions of the rest-frame spectra
144: of nearby and distant galaxies fall within the fixed observed-frame bands
145: used to establish their \fxfopt\ ratios. Redshift effects can have a
146: significant
147: impact on the perceived \fxfopt\ ratios of AGNs (Moran 2004; Peterson et al.\
148: 2006) and must be accounted for. And finally, only the integrated fluxes of
149: distant sources can be measured, and the same must be obtained for local
150: objects.
151:
152: In this paper, we present a comparison of the \fxfopt\ ratios of absorbed AGNs
153: in the CDF-N with those of nearby galaxies classified optically as type~2
154: Seyferts. Our approach accounts for the observational factors described
155: above by (1) employing a nearby sample that well represents the local
156: Seyfert~2 luminosity function and (2) accurately simulating how the nearby
157: objects would appear if they were observed in the CDF-N, including the effects
158: of how pencil-beam surveys like the CDF-N sample the luminosity function of a
159: given population. This allows us to examine in detail the extent to which
160: nearby, well-characterized AGNs can explain the properties of distant,
161: spectroscopically ambiguous X-ray galaxies. The criteria used to define
162: the comparison sample of absorbed AGNs from the CDF-N are outlined in \S~2.
163: In \S~3, the local Seyfert~2 sample is described, along with the integrated
164: optical and X-ray data we have collected for the objects. Our simulations
165: are presented in \S~4, along with discussion of how the \fxfopt\ ratios of
166: Seyfert~2 galaxies are transformed by redshift and sampling effects. Our
167: findings are summarized in the final section.
168:
169:
170: \section{The CDF-N Sample of Absorbed AGNs}
171:
172: Our investigation of the \fxfopt\ ratios of absorbed AGNs requires an
173: appropriate sample of distant X-ray galaxies from a well-characterized survey,
174: and an unbiased sample of local objects with broadband X-ray and optical data.
175: For the distant X-ray galaxy sample, the 2~Ms CDF-N is an ideal resource. The
176: details of the {\it Chandra\/} observations and parameters of the $> 500$
177: sources detected in the survey have been thoroughly documented (Alexander et
178: al.\ 2003). In addition, deep optical imaging and spectroscopy of the sources
179: have been obtained with the Subaru 8~m and Keck 10~m telescopes (Barger et al.\
180: 2002, 2003), yielding optical fluxes and, for many objects, spectroscopic
181: redshifts.
182:
183: The sources we have selected from the CDF-N have X-ray properties similar to
184: those of nearby Seyfert~2s and are drawn from a well-defined portion of the
185: deep survey. First, we select only sources with total exposure times between
186: 1.5~Ms and 2.0~Ms. This exposure time range brackets the strong peak in the
187: CDF-N source exposure time distribution centered at 1.7~Ms (Alexander et al.\
188: 2003), and because it is narrow, it allows us to establish an effective X-ray
189: flux limit and solid angle for the deep survey, which are required for the
190: simulations described below. Next, since we are chiefly concerned with the
191: origin of the XRB, we select CDF-N sources with 2--8~keV hard-band detections
192: and absorbed X-ray spectra with effective photon indices $\Gamma<1.5$ (as
193: indicated by their ``hardness ratios''). These are the sources responsible for
194: the hard XRB, and based on observations of nearby objects, they are expected
195: to be Seyfert~2 galaxies. Finally, we require that the included sources have
196: a measured spectroscopic redshift.
197:
198: Over 80\% of the sources that satisfy these criteria have redshifts between
199: $z = 0.2$ and $z = 1.3$. We have further restricted our CDF-N sample to this
200: redshift range for two reasons. First, objects closer than $z \approx 0.2$
201: are probably extended in optical images, and published magnitudes for them
202: may not reflect their total optical emission. Second, our simulations (\S~4)
203: employ {\sl UBVRI\/} data for nearby Seyfert~2s to yield the observed-frame
204: $I$-band fluxes they would have at various redshifts. At $z = 1.3$, the
205: rest-frame $U$ band is roughly centered on the observed $I$ band. Adopting
206: this redshift limit thus eliminates the need for significant extrapolation
207: of our local galaxy spectra to wavelengths shortward of $U$.
208:
209: A total of 59 CDF-N sources meet all of our selection criteria. Using
210: published 2--8~keV fluxes and $I$-band magnitudes (Alexander et al.\ 2003;
211: Barger et al.\ 2002, 2003), we have computed their observed-frame \fxfi\
212: flux ratios. Optical spectra are published for only 38 of the objects
213: (Barger et al.\ 2002), but a visual inspection of these indicates only half
214: a dozen or so clearly have the spectral signatures of narrow-line AGNs.
215: Curiously, one other object is reported to have broad emission lines, though
216: they appear to be weak in the Barger et al.\ data. The spectra of the rest
217: of the objects are consistent with those of normal galaxies, or are ambiguous
218: because of the signal-to-noise ratio and/or wavelength coverage of the data.
219: Presuming the rest of the objects we have selected to be similar, it is safe
220: to conclude that the majority of the absorbed AGNs in our CDF-N sample are
221: not considered to be Seyfert~2s on the basis of their ground-based optical
222: spectra. The redshifts, {\it Chandra\/} exposure times, 2--8 keV fluxes,
223: and 2--8 keV luminosities of the CDF-N objects are shown in Figure~1.
224:
225: \section{The Local Sample of Seyfert 2 Galaxies}
226:
227: To ensure that our comparison of the \fxfopt\ ratios of nearby and distant
228: objects is fair, it is vital that we employ a local sample
229: that is as complete and unbiased as possible. However, because of the
230: variety of ways in which Seyfert~2s have been discovered and the fact that
231: their luminosity function is not firmly established, this is a
232: non-trivial matter. The biases that result when samples are flux-limited
233: and defined on the basis of a particular property (e.g., ultraviolet excess
234: or far-infrared color) are well documented (Ho \& Ulvestad 2001). In
235: addition, samples of Seyfert~2 galaxies can be tainted by spectroscopic
236: misclassifications.
237:
238: To minimize the effects of selection biases and contamination in our study,
239: we have chosen to use objects drawn from the distance-limited sample of
240: Ulvestad \& Wilson (1989; hereafter UW89), which consists of all Seyfert
241: galaxies known (at the time of its definition) with redshifts $cz \leq 4600$
242: km~s$^{-1}$ ($z \leq 0.0153$) and a declinations $\delta \geq -45^{\circ}$.
243: Because the objects were included on the basis of distance, and not some
244: observed property, and because their nuclear activity was noticed for a
245: variety of reasons, the sample is free of significant selection biases.
246: In addition, detailed optical investigations of this sample have verified
247: that all 31 of the Seyfert~2s it contains are bona fide narrow-line AGNs
248: (Moran et al.\ 2000). For this study, we also include the 7 objects listed
249: by UW89 as ``narrow-line X-ray galaxies'' (NLXGs), despite the fact that
250: several of them are technically intermediate type~1 Seyferts that display
251: weak, broad H$\alpha$ components in high-quality optical spectra. Our
252: analysis of {\sl ASCA\/} data for the NLXGs (\S~3.2) has confirmed that
253: all of the objects are absorbed X-ray sources, with column densities of
254: $\sim 10^{22}$ cm$^{-2}$ or more. Thus, over a range of redshifts they
255: would satisfy the spectral flatness criterion used above to select absorbed
256: AGNs in the CDF-N (\S~2). Including the NLXGs, our local sample of absorbed
257: AGNs (which we refer to as ``Seyfert~2s'' for convenience) stands at 38
258: objects.
259:
260: We note that not every galaxy within the UW89 distance and declination limits
261: has been searched for a Seyfert nucleus, so the sample must be incomplete
262: to some degree. Indeed, some Seyfert galaxies have been discovered within
263: the sample volume since 1989. The level of incompleteness is probably most
264: significant at low values of the nuclear luminosity, where, in many cases,
265: an accurate emission-line classification cannot be made without careful
266: starlight template subtraction (Ho, Filippenko, \& Sargent 1997). Still,
267: several lines of
268: evidence suggest that the UW89 sample, while falling short of perfection,
269: is nonetheless a very good one. First, the radio properties of the UW89
270: Seyferts are broadly consistent with those of objects in other samples, e.g.,
271: the CfA sample (Kukula et al.\ 1995). Second, as Figure~3 of Lumsden \&
272: Alexander (2001) illustrates, the UW89 sample extends to much lower
273: luminosities than other well-studied collections of Seyfert~2s, such as
274: the CfA/12~$\mu$m (Tran 2001) and {\sl IRAS\/} (Lumsden et al.\ 2001) samples.
275: Thus, it contains more typical Seyfert~2s and suffers less from an
276: over-representation of high-luminosity objects than these other samples.
277: Finally, the X-ray luminosity density of the Seyfert~2 population derived
278: from the UW89 sample is able to account for
279: both the intensity and spectral slope of the 2--10 keV X-ray background
280: (Moran et al.\ 2001). Taken in combination, these results suggest that the
281: UW89 sample must represent the luminosity function of type~2 Seyfert galaxies
282: reasonably well.
283:
284: \subsection{Broadband Optical Data}
285:
286: The fluxes measured for distant CDF-N objects reflect their integrated optical
287: and X-ray emission; comparable data are needed for local Seyfert~2 galaxies so
288: that we can simulate what their \fxfopt\ ratios would be if they were observed
289: at modest redshift in the CDF-N. Surprisingly, although the UW89 objects are
290: among the most well-studied Seyfert~2 galaxies, relatively little information
291: about their integrated optical fluxes has been published. Integrated magnitudes
292: in the blue and visual bands can be found for about 60\% of the sample, and
293: data at redder wavelengths are even more sparse. In this section we describe
294: our {\sl UBVRI\/} observations of over half of the UW89 sample, and our methods
295: of determining integrated magnitudes for the remainder of the objects.
296:
297: \subsubsection{UBVRI Observations}
298:
299: Our optical data were acquired with the 0.9-m WIYN telescope at Kitt Peak and
300: the 1.3-m McGraw Hill telescope at the MDM Observatory during six separate
301: runs between 2003 January and 2006 January. On the WIYN telescope, we used
302: the $2048 \times 2048$ S2KB CCD, which affords a $\sim$ $20' \times 20'$ field
303: of view and an image scale of $0\farcs6$ per pixel. At MDM we employed the
304: $1024 \times 1024$ ``Templeton'' CCD, which has an $8\farcm5 \times 8\farcm 5$
305: field of view and a scale of $0\farcs5$ per pixel. Images were obtained
306: with Harris {\sl UBVRI\/} filters on the 0.9-m, and on the 1.3-m, Harris
307: {\sl BVR\/} filters were used in conjunction with a Bessell $U$ filter and
308: an Arizona $I$ filter. During our 2003 October, 2004 March, and 2006 January
309: runs, we obtained photometric observations in all five bands for 21 UW89
310: galaxies.
311:
312: We processed our images using standard IRAF procedures. In each, an integrated
313: instrumental magnitude was measured within a circular aperture centered on the
314: nucleus of the UW89 galaxy. The size of the aperture was initially selected
315: to include all of the galaxy flux visible when the image was displayed with
316: extreme settings of the contrast and dynamic range. The aperture was then
317: resized to the point at which increasing its radius did not result in an
318: increase of the galaxy's flux. The flux from foreground stars was
319: measured and subtracted from the total flux in the galaxy aperture. The sky
320: background level was estimated within a concentric annulus placed well outside
321: the galaxy aperture. A few of the UW89 objects have nearby companions, which
322: have not been excluded in our measurements. Our objective is to compare the
323: UW89 sample and distant X-ray galaxies, and the flux from companions would
324: not be separable in observations of the latter. By including companion flux
325: in the nearby sample we preserve the true diversity of the morphologies and
326: integrated colors of its members, making our comparison as valid as possible.
327: In the end, this concerns only a handful of objects: NGC~5929 is interacting
328: with NGC~5930, a spiral galaxy of comparable brightness; NGC~262 has a minor
329: companion, LEDA 212600, and two fainter satellite galaxies; and NGC~1667 has
330: a single dwarf companion.
331:
332: We observed equatorial standard star fields (Landolt 1992) to calibrate our
333: instrumental magnitudes. Average extinction coefficients for each band were
334: obtained from Landolt (1983). Following Bessell (1995), we transformed the
335: magnitudes of the standard stars to the Johnson-Kron-Cousins system defined
336: by Bessell (1990). The formal uncertainties in our {\sl UBVRI\/} measurements,
337: which include the effects of photon statistics, flat-fielding accuracy,
338: aperture size, and transformation to the standard scale, are estimated to be
339: 0.02--0.04 mag. Table~1 lists the {\sl UBVRI\/} photometry results for the
340: 21 objects we observed.
341:
342: For 12 of these galaxies, integrated photoelectric photometry in $B$ and
343: $V$ (with typical uncertainties of 0.1--0.2 mag) is published in the Third
344: Reference Catalog of Bright Galaxies (RC3; de Vaucouleurs et al.\ 1991).
345: This provides a limited opportunity to check the accuracy of our measurements.
346: As the left panel of Figure~2 indicates, the differences between our $V$-band
347: magnitudes
348: and those listed in the RC3 are (for 11 objects) consistent with the expected
349: uncertainties in the two quantities (the median offset is 0.04~mag with a
350: standard deviation of 0.10~mag). One significant discrepancy was uncovered,
351: however:\ We find Mrk~3 to be brighter than the RC3 values by 0.58~mag in $V$
352: and 0.43~mag in $B$. The problem may be related to a very bright foreground
353: star located within the aperture we used to extract the galaxy's flux. The
354: star is {\it not\/} responsible for our flux being too high --- using an
355: aperture that completely excludes the star we obtain a $B$ value that
356: is fainter by only 0.2~mag. Thus, we are confident that we have successfully
357: removed the contribution of the star in our full-aperture data, but it is not
358: clear how the contamination was handled in the RC3 measurement. We note that
359: the RC3 also lists values of $m_B$ --- photographic magnitudes from the
360: Shapley-Ames catalog (Sandage \& Tammann 1981) that have been reduced to the
361: $B_T$ system. The $m_B$ value for Mrk~3 of $13.55\pm 0.17$ is
362: entirely consistent with our measurement of $B = 13.60$. Thus, given the
363: overall agreement between our measurements and those listed in the RC3, we
364: have confidence in the accuracy of our photometry.
365:
366: \subsubsection{Integrated Magnitudes of the Unobserved Galaxies}
367:
368: Information about the integrated magnitudes of the 17 (mostly southern)
369: galaxies we did not observe is also available from the RC3 ($U_T$, $B_T$,
370: and $V_T$) and the ESO-Uppsala survey ($B_T$ and $R_T$; Lauberts \& Valentijn
371: 1989). We have adopted photoelectric magnitudes from the RC3 whenever they
372: are available (12 galaxies), and have supplemented these with photographic
373: $R_T$ magnitudes from the ESO catalog when $B_T$(ESO) agrees with $B_T$(RC3).
374: Three other objects that have only $m_B$ photographic magnitudes in the RC3
375: have $B$ and $R$ measurements in the ESO catalog; the $B$ magnitudes agree
376: in each case, so we have adopted the ESO values for these galaxies. Only
377: $m_B$ data are available for the two remaining objects.
378:
379: To estimate integrated magnitudes in the bands lacking published data, we have
380: used the $B - V$ and/or $B - R$ colors of the objects to determine the most
381: appropriate Johnson-Cousins color template from the compilation of Fukugita
382: et al.\ (1995). The majority of the galaxies are best represented by an Sab
383: template, though for several an S0 (6 galaxies) or Sbc (2 galaxies) template
384: provides the closest match. For the two objects with $m_B$ magnitudes only
385: (thus no integrated colors), we have adopted the Sab template. In all cases,
386: the template we have selected is consistent with the galaxy's morphological
387: type listed in the NASA/IPAC Extragalactic Database (NED).
388:
389: Turning once more to the 11 objects we observed that have reliable
390: photoelectric data in the RC3, we have compared the $R$-band magnitudes
391: extrapolated from their $V_T$ values to the $R$ magnitudes that we derived
392: from our images. As the right panel of Figure~2 indicates, the median
393: difference in these magnitudes is 0.04~mag with a standard deviation of
394: 0.11~mag. The similarity between the left and right panels of Figure~2
395: suggests that the application of a Fukugita et al.\ (1995) template
396: does not add an appreciable systematic error to the extrapolated magnitudes,
397: giving us confidence that the extrapolated magnitudes for the unobserved
398: objects are reasonably accurate. The integrated magnitudes obtained from the
399: literature, together with those derived from application of the appropriate
400: color template, are listed in Table~2. The final column of Table~2 indicates
401: the bands for which published data are available and the Fukugita et al.\
402: (1995) galaxy template that was used.
403:
404: \subsubsection{Corrections for Galactic Extinction}
405:
406: Corrections for Galactic extinction are necessary for a determination of the
407: true optical fluxes of the UW89 Seyfert~2s and for comparisons to galaxies in
408: other samples. We corrected our magnitudes by computing $A_{\lambda}$ for
409: each object in each of the five bands. Values of the color excess $E(B-V)$
410: in the direction of each galaxy (originating from Schlegel et al.\ 1998)
411: were obtained from NED. The total absorption in each band was then calculated
412: from $A_{\lambda}/E(B-V)$ using Table~6 of Schlegel et al.\ (1998). The final
413: extinction-corrected magnitudes for all 38 UW89 objects are listed in Table~3.
414: For clarity, magnitudes obtained from our observations or the RC3 are given
415: in plain type and those derived with the use of a Fukugita template
416: are given in italics. The Galactic latitudes of the UW89 members span a wide
417: range, so the extinction corrections vary considerably from object to object.
418:
419: \subsubsection{Ultraviolet Data}
420:
421: In the simulations described in \S~4, information about the slope of
422: the near-UV spectra of the UW89 objects is needed to ensure that the
423: observed-frame $I$-band fluxes we predict for them are accurate for all
424: assumed redshifts up to our limit of $z = 1.3$. The atlas of galaxies
425: observed with {\sl GALEX\/} (Gil de Paz et al.\ 2006) provides integrated
426: fluxes for 8 UW89 Seyfert~2s (Mrk~3, NGC 262, 1068, 1386, 2992, 4117,
427: 4388, and 7582) at a near-UV wavelength of 2267~\AA . After correcting
428: for Galactic extinction [$A_{\rm NUV} = 8 \times E(B-V)$; Gil de Paz et
429: al.\ 2006], we find that the ${\rm NUV} - U$ colors of our 8 objects range
430: from 1.64 to 3.17. For the rest of the UW89 objects, we adopt the median
431: value of ${\rm NUV} - U = 2.12$.
432:
433: \subsection{X-ray Data}
434:
435: Broadband X-ray data in the $\sim 0.5$--10 keV energy range are available
436: for the entire UW89 Seyfert~2 sample. Nearly all (36/38) of the objects
437: were observed with the {\sl ASCA\/} satellite; the two remaining galaxies
438: (NGC~5283 and NGC~5728) have been observed with the {\it Chandra X-ray
439: Observatory}. Although several other UW89 galaxies have also been observed
440: with {\it Chandra}, the {\sl ASCA\/} data are preferred because of the
441: consistent depth of the exposures and the fact that, due to the large
442: {\sl ASCA\/} beam ($\sim 3$ arcmin half-power diameter), we can be sure
443: that they represent the total X-ray flux from the nucleus and host galaxy.
444: The nucleus is likely to dominate in the majority of cases, at least at
445: the higher X-ray energies, but many objects are weak and their extended
446: X-ray emission (from supernova remnants, X-ray binaries, etc.) could be
447: comparable to the nuclear flux.
448:
449: Details of the {\sl ASCA\/} observations and data reduction for the UW89
450: sample have been described by Moran et al.\ (2001); a brief summary is provided
451: here. The data were obtained from our own observations and from the HEASARC
452: data archive at NASA's Goddard Space Flight Center. The {\sl ASCA\/} exposure
453: times of the UW89 Seyfert~2s are uniformly long (most are in the 35--45 ks
454: range), and the targets were placed at the ``1-CCD'' off-axis position in most
455: of the images. For this work we focus on data collected with the Gas Imaging
456: Spectrometers (GIS) on board {\sl ASCA}; compared to the satellite's SIS
457: instruments, the GIS have better hard X-ray sensitivity and more consistent
458: response, and due to their larger field of view, background estimation is
459: more straightforward with them.
460:
461: The {\it Chandra\/} images of NGC~5283 and NGC~5728 were obtained from the
462: data archive at the Chandra X-ray Observatory Center (CXC). The objects were
463: observed with the ACIS-S instrument for 9.8 ks and 19.0 ks, respectively.
464: Both sources are relatively weak ($\sim 0.06$ count~s$^{-1}$), so spectral
465: distortions resulting from photon pile-up are not a concern.
466:
467: We extracted source and background events for all of the {\sl ASCA\/} and
468: {\it Chandra\/} data sets, and generated response and effective area files
469: specific to the individual observations. All 38 UW89 Seyfert~2s were detected
470: above a signal-to-noise ratio of 4 (full band). For 25 objects, the net counts
471: obtained were sufficient to allow spectral modeling with the XSPEC software
472: (Arnaud 1996). We have modeled the spectra as the sum of three components: a
473: weakly absorbed power law with a photon index $\Gamma_1$ and associated column
474: density of $(N_{\rm H})_1$, a heavily absorbed power law with slope $\Gamma_2$
475: and column density $(N_{\rm H})_2$, and a Gaussian Fe~K$\alpha$ line of width
476: $\sigma_{K\alpha}$ centered at energy $E_{K\alpha}$. In all instances but one,
477: an acceptable fit with reasonable best-fit spectral parameters was obtained.
478: The exception is NGC~1068, which has a far more complex broadband X-ray
479: spectrum (Iwasawa, Fabian, \& Matt 1997; Matt et al.\ 1997). Table~4 lists the
480: adopted distances to the galaxies (see Moran et al.\ 2001), the X-ray spectral
481: parameters derived from our fits, and the associated X-ray fluxes in the
482: 0.5--2 keV and 2--8 keV bands. We note that while our relatively simple
483: spectral models generally afford statistically acceptable fits, they may not
484: represent the best physical description of the X-ray emission in every case.
485: The main purpose of our spectral analysis is to provide accurate fluxes,
486: which it does. This is true even for NGC~1068; our approach yields soft-
487: and hard-band fluxes that are respectively within 10\% and 1\% of those
488: obtained using a more complex model that provides a good fit.
489:
490: For the 13 weakly detected objects, X-ray fluxes were estimated from ratios
491: of the counts detected in hard (4--10 keV) and soft (1--4 keV) bands. First,
492: we computed the median Seyfert~2 X-ray spectrum based on the spectral fits
493: obtained for the 25 ``strong'' sources above. The median spectrum is
494: characterized by the following parameters: $\Gamma_1 = 1.78$, $\Gamma_2 =
495: 1.70$, $(N_{\rm H})_1 = 0$, and $(N_{\rm H})_2 = 2.42 \times 10^{23}$
496: cm$^{-2}$. (An Fe~K$\alpha$ component is not included, for reasons that
497: will become clear below.) The median model is similar to the composite
498: Seyfert~2 X-ray spectrum derived from the summed emission of the UW89
499: objects (Moran et al.\ 2001), despite the fact that the latter is
500: dominated by the most luminous sources.
501:
502: To estimate the X-ray fluxes of the weak sources, we fixed the parameters
503: of the median model and varied the normalizations of the two power-law
504: components in XSPEC until the hard-to-soft counts ratio associated with the
505: model matched the observed counts ratio. We then fixed the ratio of the
506: normalizations and scaled them until the count rate implied by the model
507: was identical to the total observed count rate. The fluxes in 0.5--2.0 keV
508: and 2.0--8.0 keV ranges were then computed from the model. To validate
509: our approach, we applied the same procedure to the ``strong'' sources whose
510: spectra could be modeled directly. As Figure~3 indicates, the 2--8 keV
511: fluxes obtained directly from spectral fitting and those obtained using the
512: median model differ by only a few percent in most cases (even {\it without\/}
513: a contribution from an Fe~K$\alpha$ line). Based on this good agreement,
514: we are confident that the fluxes we have derived for the 13 weak objects
515: (also listed in Table~4) are reasonably accurate.
516:
517:
518: \section{Simulations}
519:
520: \subsection{Approach}
521:
522: A direct, fair comparison of the \fxfopt\ ratios of distant absorbed AGNs
523: with those of nearby Seyfert~2 galaxies cannot be made. First of all, the
524: \fxfopt\ ratio is measured in the observed frame, so its value for a given
525: object varies with redshift. Secondly, the ways in which samples of distant
526: and nearby sources are assembled naturally lead to different luminosity
527: distributions in the samples, which in turn affect the distributions of their
528: \fxfopt\ ratios. Our approach, therefore, is to take a sample of nearby
529: Seyfert~2s that well represents the local luminosity function and simulate the
530: distribution of flux ratios that would result if they were observed under
531: the same conditions (and with the same redshift distribution) as the distant
532: sources. This minimizes the effects of redshift and selection bias.
533:
534: We begin by applying the information listed in Table~4 to determine the fluxes
535: of each UW89 source in the observed 0.5--2 keV and 2--8 keV bands as a function
536: of redshift. The luminosity distances used in the calculations are based on an
537: $H_0$ = 70 km~s$^{-1}$~Mpc$^{-1}$, $\Omega_{\rm M} = 1/3$, $\Omega_{\Lambda}
538: = 2/3$ cosmology. The results establish the redshift range within which each
539: UW89 object would be detectable if observed as part of the $t \ge 1.5$~Ms
540: portion of the CDF-N survey. Specifically, we apply the same criteria used to
541: define our CDF-N sample of absorbed AGNs: a 2--8 keV flux limit corresponding
542: to the this exposure time ($1.8 \times 10^{-16}$ erg~cm$^{-2}$~s$^{-1}$) and
543: an effective spectral index $\Gamma \le 1.5$, which corresponds to a flux ratio
544: $F_{\rm 2-8}/F_{\rm 0.5-2} \ge 2$. It is interesting to note that, based on
545: these criteria, four UW89
546: galaxies would not be included in the CDF-N (as absorbed AGNs) in the $0.2 \le
547: z \le 1.3$ range. The spectra of NGC~1068, NGC~1386, and NGC~5135 are too
548: steep to meet the spectral flatness criterion while their 2--8 keV fluxes are
549: above the hard X-ray flux limit. The fourth object, NGC~4941, falls below the
550: flux limit before $z = 0.2$. Only 10 objects would be detectable in the CDF-N
551: all the way out to our redshift limit of $z = 1.3$.
552:
553: We use Monte Carlo methods to simulate the \fxfi\ distribution that nearby
554: Seyfert~2 galaxies would have if observed in the CDF-N, randomly selecting
555: a redshift (weighted by the CDF-N redshift distribution) and a UW89 galaxy
556: (unweighted, since to first order the UW89 sample is the local Seyfert~2
557: luminosity function). We first verify that the UW89 object would be included
558: in the CDF-N as an absorbed AGN at the chosen redshift. If not, another
559: galaxy is selected at the same redshift. Next, we determine the likelihood
560: that an object with the UW89 galaxy's X-ray luminosity would be included in
561: the CDF-N. For this test, we have combined the CDF-N flux limit and survey
562: solid angle (170 arcmin$^2$ for $t = 1.5$~Ms) to estimate the volume searched
563: in the CDF-N as a function of minimum detectable 2--8~keV luminosity. The
564: results are plotted in Figure~4, along with the fixed volume represented by
565: the UW89 sample (calculated by Moran et al.\ 2001). Below a luminosity of
566: $\sim 3 \times 10^{42}$ erg~s$^{-1}$, the volume searched in the CDF-N is
567: less than that of the UW89 sample. Therefore, in this $L_{\rm X}$ range, the
568: ratio of the CDF-N volume to the UW89 volume defines the probability that a
569: local object of a given luminosity would be included in the CDF-N.
570:
571: If a UW89 galaxy passes all the above tests, the {\sl UBVRI\/} photometry
572: reported in \S~3 is used to compute its integrated, observed-frame $I$-band
573: flux. The optical spectrum of the object is approximated by converting the
574: broadband magnitudes to flux densities at the band centers and assuming they
575: are joined by power laws. The spectrum is shifted and dimmed appropriately
576: for the selected redshift; the portion falling within the observed $I$ band
577: is then integrated over the width of the band to give us the optical flux.
578: As the redshift approaches $z = 1.3$, the rest-frame UV spectrum shortward of
579: the center of the $U$ band enters the observed-frame $I$ band. The NUV data
580: (\S~3.1.4) is used to extrapolate to shorter wavelengths, though the value
581: of the ${\rm NUV} - U$ color we adopt affects the flux by $< 1$\%. Using the
582: derived optical flux, the \fxfi\ ratio of the object is then calculated.
583: The process continues until an \fxfi\ distribution composed of $10^4$ UW89
584: objects is obtained.
585:
586: \subsection{Redshift and Sampling Effects}
587:
588: Before presenting the results of our simulations and a comparison to the
589: CDF-N, we explore the way source redshifts and the flux-limited nature of
590: deep surveys combine to influence the \fxfopt\ ratios of a population of
591: absorbed AGNs.
592:
593: As Table~3 indicates, the intrinsic integrated optical colors of the UW89
594: objects are quite red ($B - I \approx 2$). In the X-ray band, the heavy
595: absorption in Seyfert~2 galaxies usually hardens their observed X-ray spectra
596: considerably (see the composite UW89 X-ray spectrum in Fig.~1 of Moran et
597: al.\ 2001). Thus, as the redshift of a typical Seyfert~2 galaxy increases,
598: a brighter portion of its rest-frame X-ray spectrum is shifted into the
599: observed 2--8 keV band, and a fainter portion of its rest-frame optical
600: spectrum is shifted into the observed $I$ band. The observed-frame \fxfopt\
601: ratio should therefore increase significantly with redshift. This effect
602: is clearly demonstrated in Figure~5, where we have plotted $F_{\rm X}$,
603: $F_{\rm I}$, and \fxfi\ vs.\ redshift for four UW89 Seyfert~2s spanning a
604: wide range of intrinsic \fxfi\ ratios. Between $z = 0$ and $z = 1.5$, the
605: observed flux ratios of these objects increase by factors of 15 to 35.
606:
607: Of equal importance are the effects of sampling in a flux-limited survey
608: such as the CDF-N. As Figure~4 illustrates, the volume searched for X-ray
609: galaxies in the CDF-N is a strong function of the observed 2--8 keV
610: luminosity. This naturally leads to Malmquist bias in the CDF-N source
611: catalog, i.e., an under-representation of relatively abundant sources with
612: low X-ray luminosities, and an over-representation of rare, high-luminosity
613: sources. If \fxfi\ happens to depend on $L_{\rm X}$ (and it does; see \S~4.3),
614: these Malmquist effects will be imprinted on the \fxfi\ distribution for
615: absorbed AGNs in the CDF-N.
616:
617: In combination, the effects of redshift and sampling can dramatically alter
618: the observed \fxfi\ distribution for Seyfert~2 galaxies. In Figure~6 we have
619: plotted the rest-frame \fxfi\ distribution for the UW89 sample, along with
620: the distribution obtained by simulating CDF-N observations of the UW89 objects
621: (as described in the previous section). Clearly, the two distributions bear
622: no resemblance to each other, even though they are derived from the same set
623: of objects. This illustrates why a direct comparison of the \fxfopt\ ratios
624: of nearby and distant sources would yield misleading results. More generally,
625: Figure~6 indicates that \fxfopt , as an activity diagnostic, can be ambiguous.
626: X-ray survey results are often summarized with plots that compare the X-ray
627: and optical fluxes of the detected sources, with diagonal lines drawn for
628: constant values of \fxfopt\ (e.g., Alexander et al.\ 2003; Bauer et al.\ 2004).
629: Frequently, the region on these plots represented by
630: log \fxfopt\ $> -1$ are labeled ``AGNs,'' while that represented by log
631: \fxfopt\ $< -2$ are labeled ``galaxies.'' Our investigation of the UW89
632: sample reveals that Seyfert~2s at modest redshift can have \fxfopt\
633: ratios well outside the range expected for AGNs.
634:
635: \subsection{Comparison to the CDF-N}
636:
637: The \fxfi\ distribution for the absorbed AGNs in the CDF-N and the results
638: of our simulations are compared in Figure~7. As the Figure indicates, the
639: two \fxfi\ distributions are broadly consistent with each other: they peak
640: at the same place and have roughly the same width. The match is especially
641: good for values of log~\fxfi\ $\ge -1$. Note that the CDF-N distribution
642: comprises just 59 objects, so there is some statistical uncertainty associated
643: with the number of objects in each bin of that distribution. The only
644: possible discrepancy occurs at the lowest \fxfi\ ratios, where the simulated
645: UW89 distribution falls consistently below the CDF-N distribution. Given
646: the nature of the rest-frame \fxfi\ distribution of the UW89 sample (Fig.~6),
647: there is no chance that the good agreement between the CDF-N and simulated
648: UW89 flux-ratio distributions is coincidental. Instead, it must be a
649: reflection of the similarity between the nearby and distant populations of
650: absorbed AGNs.
651:
652: A more detailed comparison is provided in Figure~8, which plots the observed
653: \fxfi\ ratio as a function of observed 2--8 keV luminosity for the 59 CDF-N
654: sources and a UW89 simulation consisting of 75 successful trials. Two things
655: are immediately obvious in Figure~8: (1) the CDF-N and UW89 points occupy
656: similar locations in the \fxfi\ -- $L_{\rm X}$ plane, and (2) \fxfi\ scales
657: roughly linearly with $L_{\rm X}$ for both populations, albeit with a fair
658: amount of dispersion. The fact that UW89 and CDF-N galaxies of a certain
659: nuclear luminosity ($L_{\rm X}$) have about the same range of \fxfi\ ratios
660: indicates that they are fundamentally similar objects in terms of their
661: optical properties. Moreover, because the slope of the ``correlation''
662: between \fxfi\ and $L_{\rm X}$ is about unity, the median optical luminosity
663: of the objects must be roughly constant and independent of the luminosity of
664: the nucleus. In a statistical sense, therefore, it appears that the optical
665: luminosities of absorbed AGNs (with observed $L_{\rm X} \le 10^{43}$
666: ergs~s$^{-1}$) are dominated by emission from the host galaxy.
667:
668: It is also evident from Figure~8 that the match between the UW89 and CDF-N
669: samples is not perfect. In particular, there are no UW89 objects with
670: observed hard X-ray luminosities above $\sim 2 \times 10^{43}$ ergs~s$^{-1}$,
671: whereas the CDF-N sample extends to $L_{\rm X} = 10^{44}$ ergs~s$^{-1}$.
672: Also, there appear to be too few UW89 objects with $L_{\rm X} < 10^{42}$
673: ergs~s$^{-1}$ and log~\fxfi\ $< -1$, consistent with the discrepancy between
674: the flux-ratio distributions shown in Figure~7. Although several factors
675: can affect the exact location of an absorbed AGN in the \fxfi\ -- $L_{\rm X}$
676: plane, the differences indicated in Figure~8 are almost certainly related to
677: the completeness of the UW89 sample rather than physical differences between
678: the nearby and distant sources. At the high-$L_{\rm X}$ end, the volume
679: associated with the UW89 sample is too small to include rare, high-luminosity
680: objects, which are in fact over-represented in the CDF-N due to the large
681: volume it surveys for such sources (see Fig.~4). At the low-$L_{\rm X}$
682: end, the incompleteness of the UW89 sample noted in \S~3 is probably the
683: primary issue. Again, the nearby Seyfert~2 galaxies that are absent in
684: the UW89 sample are likely to be objects with low-luminosity nuclei, such
685: as those for which starlight template subtraction is required for an accurate
686: emission-line classification. Having more local galaxies with low X-ray
687: luminosities would increase the representation of low-\fxfi\ sources in our
688: simulations. The fact that our simulations reproduce the flux-ratio
689: distribution of distant absorbed AGNs as well as they do implies that the
690: shortcomings of the UW89 sample are not too severe.
691:
692: \subsection{Optically Normal X-ray Galaxies in the Deep Surveys}
693:
694: If nearby Seyfert~2 galaxies are able to account for the relative X-ray and
695: optical properties of distant absorbed AGNs, why do the objects detected in
696: the {\it Chandra\/} deep surveys often lack emission-line evidence for
697: nuclear activity? As Moran et al.\ (2002) have demonstrated, a combination
698: of observational factors --- host-galaxy dilution, signal-to-noise ratio, and
699: wavelength coverage --- are capable of making many UW89 Seyfert~2s appear
700: ``normal'' in their integrated spectra. The bulk of the UW89 and CDF-N
701: objects overlap in terms of their optical luminosities, which in both cases are
702: dominated by host-galaxy emission, so these observational factors should affect
703: the ground-based spectra of distant absorbed AGNs in a similar way. It would
704: seem, therefore, that additional hypotheses for the optically normal appearance
705: of the CDF-N population --- at least those objects that satisfy our selection
706: criteria --- are unnecessary at this time.
707:
708: However, in a recent study of objects from the {\it Chandra\/} Deep Field South
709: (CDF-S), Rigby et al.\ (2006) have argued that the absence of strong AGN-like
710: emission lines in the ground-based spectra of distant X-ray galaxies results
711: primarily from obscuration of the narrow-line region by extranuclear dust,
712: rather than host-galaxy dilution. Their conclusions are based on the finding
713: that the morphologies of optically active galaxies (with broad emission lines
714: or high-excitation narrow lines) and optically dull galaxies (with weak and/or
715: low-excitation emission lines) differ statistically. Optically active galaxies
716: in the CDF-S tend to have high ratios of their semiminor and
717: semimajor axes $b/a$, whereas optically dull objects have a relatively flat
718: distribution of $b/a$. Taking the measured axis ratio as a proxy for
719: inclination, Rigby et al.\ suggest that the optically dull sources are missing
720: AGN-like emission lines because extranuclear dust obscures the narrow-line
721: region in the more inclined galaxies.
722:
723: To examine the inclination hypothesis, we have compared axis ratio
724: distributions for appropriate subsets of the Rigby et al.\ and UW89 samples.
725: For the CDF-S galaxies, we have compiled the $b/a$ distribution for those
726: objects in the $0.5 \le z \le 0.8$ subsample of Rigby et al.\ that satisfy
727: our X-ray selection criteria, i.e., detection in the 2--8~keV band and a
728: (2--8 keV)/(0.5--2 keV) flux ratio in excess of 2. Although we have ignored
729: their emission-line classifications, this resticted CDF-S sample of 15 objects
730: includes just one optically active source --- a narrow-line object --- so any
731: possible confusion introduced by the presence of broad-line AGNs has been
732: eliminated. Likewise, we have limited the UW89 comparison sample to include
733: only the 18 objects that would be detected in the CDF-S ($F_{\rm 2-8} > 3
734: \times 10^{-16}$ ergs~cm$^{-2}$~s$^{-1}$) at $z \ge 0.5$. We have estimated
735: axis ratios for the UW89 galaxies using our images or images available from
736: NED. Our measurements, obtained both by hand and with the {\it ellipse\/} task
737: in IRAF, are based on the shape of the outer, low surface-brightness isophotes.
738: The two methods yield very similar results for all objects where both could
739: be successfully employed. In a handful of cases (e.g., interacting galaxies)
740: the output from {\it ellipse\/} is suspect and we favor the values measured
741: by hand. Our best estimates of $b/a$ for the UW89 subsample are listed in
742: Table~5. As Figure~9 indicates, local Seyfert~2s have a very broad
743: distribution of $b/a$, implying that they are at least as inclined as the
744: optically dull objects in the CDF-S. One caveat here is that $b/a$ has not
745: been measured in exactly the same way for the nearby and distant sources.
746: However, given the coarse binning used in Figure~9 it is unlikely that a
747: different measurement approach for the local sample would alter these results
748: significantly. The fact that the UW89 objects have strong nuclear emission
749: lines suggests that inclination, and the associated effects of extranuclear
750: dust, cannot be the primary origin of the optically normal appearance of the
751: distant, absorbed X-ray galaxies.
752:
753: In a recent complementary study, Peterson et al.\ (2006) have examined the
754: \fxfopt\ ratios that nearby AGNs would have if they were observed at a
755: redshift of $z = 0.3$. Their analysis revealed that many such objects would
756: have low \fxfopt\ ratios and modest X-ray luminosities, similar to the
757: optically bright, X-ray--faint sources (OBXFs; Hornschemeier et al.\ 2001,
758: 2003) that have been detected in the CDF-N. Spectroscopically, the OBXFs
759: appear to be quiescent, and Peterson et al.\ have reasoned that many could
760: harbor normal Seyfert~2 nuclei if host-galaxy dilution is significant in
761: their ground-based optical spectra. Our results support this conclusion.
762: In Figure~5, it is clear that redshift effects on \fxfi\ are slight at $z =
763: 0.3$. All but a few of the UW89 objects would be detectable at $z \approx
764: 0.3$ in the CDF-N, so the UW89 \fxfi\ distribution at that redshift would
765: look much like the $z = 0$ distribution shown in Figure~6, shifted by only
766: $\sim +0.3$ in log~\fxfi . A significant number of the UW89 galaxies would
767: therefore have log~\fxfopt\ $< -2$, similar to the OBXFs in the CDF-N. In
768: addition, the low-\fxfopt\ objects in the sample would have X-ray luminosities
769: in the range of normal galaxies ($\sim 10^{41}$ ergs~s$^{-1}$ or less), and
770: many would have quiescent optical spectra (Moran et al.\ 2002). Thus, as
771: Peterson et al.\ have suggested, a number of the OBXF objects could be
772: unrecognized Seyfert~2s.
773:
774:
775: \section{Summary and Conclusions}
776:
777: To investigate the nature of the ``normal'' X-ray--luminous galaxies in the
778: CDF-N, we have obtained {\sl UBVRI\/} photometry and broadband X-ray data for
779: a distance-limited Seyfert~2 galaxy sample that broadly represents the local
780: luminosity function for absorbed AGNs. From these data we have measured the
781: integrated fluxes of the galaxies, since this is what is normally derived
782: from multiwavelength observations of the distant objects detected in the
783: deep X-ray surveys.
784:
785: We have selected a sample of absorbed AGNs from a well-defined portion of the
786: CDF-N for comparison to the local objects. Using the redshift distribution
787: of the CDF-N sources, we have simulated the \fxfopt\ ratios that the UW89
788: objects would have if they were observed at modest redshift as part of the
789: CDF-N. By including (1) the effects of redshift on flux measurements in fixed
790: observed-frame bands, and (2) the way the luminosity function of a given
791: population is sampled in a flux-limited survey like the CDF-N, we have shown
792: that nearby Seyfert~2s with strong nuclear emission lines are able to account
793: for the X-ray and optical properties of distant absorbed AGNs, despite the
794: fact that the latter often lack optical evidence for nuclear activity in
795: ground-based data. The integrated spectra of UW89 galaxies indicate that
796: observational factors --- host-galaxy dilution, signal-to-noise ratio, and
797: wavelength coverage --- are capable of hiding the nuclear emission lines of
798: bona fide Seyfert~2s (Moran et al.\ 2002). We conclude, therefore, that the
799: same factors provide the simplest explanation for the ``normal'' appearance
800: of many absorbed AGNs in the {\it Chandra\/} deep surveys. Note that our
801: arguments are statistical --- it is certainly possible that some distant
802: absorbed AGNs appear to be normal because they are located in edge-on host
803: galaxies, or because they have unusually high amounts of nuclear obscuration.
804: In general, though, we have been unable to identify differences between the
805: nearby and distant populations of absorbed AGNs that cannot be attributed
806: to host-galaxy dilution. Until we do, it seems unnecessary to invoke the
807: existence of a significant new class of X-ray--bright, optically normal
808: galaxies (XBONGs; Comastri et al.\ 2002) that differ from nearby Seyfert~2s
809: in some fundamental way.
810:
811: The problem with the X-ray--luminous ``normal'' galaxies may be mainly a
812: matter of perception. In Figure~10, we have plotted the integrated spectra
813: of two galaxies from the UW89 sample, Mrk~3 and NGC~788 (Moran et al.\ 2002).
814: In most respects, these two AGNs are nearly identical: they have similar
815: X-ray luminosities and absorption column densities; optically, their
816: luminosities are comparable and both exhibit polarized broad emission lines;
817: and both reside in S0 host galaxies at a distance of $d \approx 54$ Mpc.
818: However, as Figure~10 illustrates, a wide range of line strengths exists
819: among ``real'' Seyfert~2s. Mrk~3 would be easily recognized as an AGN at
820: moderate redshifts, whereas NGC~788 would not. The main difficulty with
821: the deep X-ray survey results may lie with an expectation that the average
822: Seyfert 2 resembles Mrk~3, when in fact NGC~788 is the more typical object.
823:
824: \acknowledgments
825: We would like to thank John Salzer for helpful discussions regarding Malmquist
826: effects in flux-limited surveys, Seth Cohen for help with the axis-ratio
827: measurements, Eve Armstrong for obtaining the optical images of NGC~2110,
828: Kaitlin Kratter for extensive help with the observing at the MDM 1.3-m, and
829: Mary Hui for assistance with the WIYN 0.9-m observing. This work was supported
830: in part by NASA through a grant for {\sl HST\/} proposal \#09869 from the
831: Space Telescope Science Institute, which is operated the Association of
832: Universities for Research in Astronomy, Inc., under NASA contract NAS5-26555.
833:
834:
835: \begin{thebibliography}{}
836: \bibitem[]{}Alexander, D. M., et al.\ 2003, AJ, 126, 539
837: \bibitem[]{}Arnaud, K.\ 1996, in Astronomical Data Analysis Software and
838: Systems V, ed.\ G.~Jacoby \& J.~Barnes (San Francisco:\ ASP), 17
839: \bibitem[]{}Awaki, H., Koyama, K., Inoue, H., \& Halpern, J.~P.\ 1991, PASJ,
840: 43, 195
841: \bibitem[]{}Barger, A.~J., Cowie, L.~L., Bautz, M.~W., Brandt, W.~N.,
842: Garmire, G. P., Hornschemeier, A. E., Ivison, R. J., \&
843: Owen, F.~N.\ 2001b, AJ, 122, 2177
844: \bibitem[]{}Barger, A.~J., Cowie, L.~L., Mushotzky, R.~F., \& Richards, E.~A.\
845: 2001a, AJ, 121, 662
846: \bibitem[]{}Barger, A.~J., et al.\ 2002, AJ, 124, 1839
847: \bibitem[]{}Barger, A. J., et al.\ 2003, AJ, 126, 632
848: \bibitem[]{}Barger, A.~J., et al.\ 2005, AJ, 129, 578
849: \bibitem[]{}Bauer, F.~E., et al.\ 2004, AJ, 128, 2048
850: \bibitem[]{}Bessell, M.~S.\ 1990, PASP, 102, 118
851: \bibitem[]{}Bessell, M.~S.\ 1995, PASP, 107, 672
852: \bibitem[]{}Comastri, A., Setti, G., Zamorani, G. \& Hasinger, G.\ 1995,
853: A\&A, 296, 1
854: \bibitem[]{}Comastri, A., et al.\ 2002, ApJ, 571, 771
855: \bibitem[]{}de Vaucouleurs, G., de Vaucouleurs, A., Corwin, H.~G., Jr., Buta,
856: R.~J., Paturel, G., \& Fouqu\'e, P.\ 1991, Third Reference
857: Catalog of Bright Galaxies, (New York:\ Springer-Verglag)
858: \bibitem[]{}Fukugita, M., Shimasaku, K., \& Ichikawa, T.\ 1995, PASP, 107, 945
859: \bibitem[]{}Gil de Paz, A., et al.\ 2006, ApJS, in press (astro-ph/0606440)
860: \bibitem[]{}Gilli, R., Salvati, M., \& Hasinger, G.\ 2001, A\&A, 366, 407
861: \bibitem[]{}Ho, L.~C., Filippenko, A.~V., \& Sargent, W.~L.~W.\ 1997, ApJS,
862: 112, 315
863: \bibitem[]{}Ho, L.~C., \& Ulvestad, J.~S.\ 2001, ApJS, 133, 77
864: \bibitem[]{}Hornschemeier, A.~E., et al.\ 2001, ApJ, 554, 742
865: \bibitem[]{}Hornschemeier, A.~E., et al.\ 2003, AJ, 126, 575
866: \bibitem[]{}Iwasawa, K., Fabian, A.~C., \& Matt, G.\ 1997, MNRAS, 289, 443
867: \bibitem[]{}Kukula, M.~J., Pedlar, A., Baum, S.~A., \& O'Dea, C.~P.\ 1995,
868: MNRAS, 276, 67
869: \bibitem[]{}Landolt, A.~U.\ 1983, AJ, 88, 439
870: \bibitem[]{}Landolt, A.~U.\ 1992, AJ, 104, 340
871: \bibitem[]{}Lauberts, A., \& Valentijn, E.~A.\ 1989, The Surface Photometry
872: Catalogue of the ESO-Uppsala Galaxies (Garching:\ European
873: Southern Observatory)
874: \bibitem[]{}Lumsden, S.~L., \& Alexander, D.~M.\ 2001, MNRAS, 328, L32
875: \bibitem[]{}Lumsden, S.~L., Heisler, C.~A., Bailey, J.~A., Hough, J.~H., \&
876: Young, S.\ 2001, MNRAS, 327, 459
877: \bibitem[]{}Matt, G., et al.\ 1997, A\&A, 325, L13
878: \bibitem[]{}Moran, E.~C., 2004, in Supermassive Black Holes in the Distant
879: Universe, (Kluwer:\ Boston), ed.\ A.~J.\ Barger, p.~225
880: \bibitem[]{}Moran, E.~C., Barth, A.~J., Kay, L.~E., \& Filippenko, A.~V.\
881: 2000, ApJ, 540, L73
882: \bibitem[]{}Moran, E.~C., Filippenko, A.~V., \& Chornock, R.\ 2002, ApJ,
883: 579, L71
884: \bibitem[]{}Moran, E.~C., Kay, L.~E., Davis, M., Filippenko, A.~V., \&
885: Barth, A.~J.\ 2001, ApJ, 556, L75
886: \bibitem[]{}Mushotzky, R.~F., Cowie, L.~L., Barger, A.~J., \& Arnaud, K.~A.\
887: 2000, Nature, 404, 459
888: \bibitem[]{}Peterson, K.~C., Gallagher, S.~C., Hornschemeier, A.~E., Muno,
889: M.~P., \& Bullard, E.~C.\ 2006, AJ, 131, 133
890: \bibitem[]{}Rigby, J. E., Rieke, G.~H., Donley, J.~L., Alonso-Herrero, A.,
891: P\'erez-Gonz\'alez, P.~G.\ 2006, ApJ, 647, 922
892: \bibitem[]{}Sandage, A., \& Tammann, G.~A.\ 1981, Revised Shapley-Ames Catalog
893: of Bright Galaxies, Carnegie Institution of Washington Publ.\ 635
894: \bibitem[]{}Schlegel, D.~J., Finkbeiner, D.~P. \& Davis, M.\ 1998, ApJ,
895: 500, 525
896: \bibitem[]{}Setti, G., \& Woltjer, L.\ 1989, A\&A, 224, L21
897: \bibitem[]{}Stocke, J.~T., et al.\ 1991, ApJS, 76, 813
898: \bibitem[]{}Szokoly, G.~P., et al.\ 2004, ApJS, 155, 271
899: \bibitem[]{}Tran, H.~D.\ 2001, ApJ, 554, L19
900: \bibitem[]{}Ulvestad, J.~S., \& Wilson, A.~S.\ 1989, ApJ, 343, 659
901: \bibitem[]{}Yuan, F., \& Narayan, R.\ 2004, ApJ, 612, 724
902: \end{thebibliography}
903:
904:
905: \clearpage
906: \begin{center}
907: \begin{deluxetable}{lccccc}
908: \tablewidth{0pt}
909: \tablecaption{Photometry Results}
910: \tablehead{\colhead{Galaxy} &
911: \colhead{$U$} &
912: \colhead{$B$} &
913: \colhead{$V$} &
914: \colhead{$R$} &
915: \colhead{$I$}}
916: \startdata
917: MCG $-$05-18-002 & 14.12 & 13.52 & 11.92 & 11.28 & 10.49 \\
918: MCG $+$01-27-020 & 14.82 & 14.83 & 14.08 & 13.65 & 13.05 \\
919: Mrk 3 & 13.88 & 13.60 & 12.39 & 11.64 & 10.96 \\
920: Mrk 1066 & 14.50 & 14.17 & 13.19 & 12.56 & 11.83 \\
921: NCG 262 & 13.68 & 13.67 & 12.84 & 12.28 & 11.72 \\
922: NCG 591 & 14.15 & 14.00 & 13.18 & 12.62 & 11.96 \\
923: NGC 788 & 13.45 & 13.02 & 12.05 & 11.50 & 10.80 \\
924: NGC 1358 & 13.50 & 13.06 & 12.09 & 11.48 & 10.83 \\
925: NGC 1667 & 12.90 & 12.79 & 12.03 & 11.46 & 10.82 \\
926: NGC 1685 & 14.58 & 14.29 & 13.40 & 12.73 & 12.16 \\
927: NGC 2110 & 14.50 & 13.22 & 11.83 & 11.06 & 10.18 \\
928: NGC 2273 & 12.86 & 12.65 & 11.64 & 11.02 & 10.31 \\
929: NGC 3081 & 13.22 & 12.96 & 12.05 & 11.51 & 10.87 \\
930: NGC 3982 & 12.10 & 12.18 & 11.59 & 11.14 & 10.57 \\
931: NGC 4117 & 14.09 & 13.84 & 13.00 & 12.43 & 11.78 \\
932: NGC 4388 & 11.86 & 11.72 & 10.96 & 10.45 & ~9.76 \\
933: NGC 4941 & 12.03 & 12.06 & 11.15 & 10.60 & ~9.86 \\
934: NGC 5347 & 13.23 & 13.17 & 12.46 & 11.93 & 11.14 \\
935: NGC 5695 & 13.82 & 13.49 & 12.66 & 12.09 & 11.31 \\
936: NGC 5929 & 12.76 & 12.99 & 12.05 & 11.45 & 10.72 \\
937: NGC 7672 & 14.95 & 14.76 & 13.94 & 13.40 & 12.50
938: \enddata
939: \vskip -0.3truein
940: \tablecomments{Uncorrected for Galactic extinction.}
941: \end{deluxetable}
942: \end{center}
943:
944: \begin{center}
945: \begin{deluxetable}{lcccccc}
946: \tablewidth{0pt}
947: \tablecaption{Integrated Magnitudes from the Literature}
948: \tablehead{\colhead{Galaxy} &
949: \colhead{$U$} &
950: \colhead{$B$} &
951: \colhead{$V$} &
952: \colhead{$R$} &
953: \colhead{$I$} &
954: \colhead{Lit.\ data/template}}
955: \startdata
956: IC 3639 & 13.34 & 13.01 & 12.23 & 11.87 & 11.22 & $BR$/Sbc \\
957: MCG $-$05-23-016 & 14.49 & 14.07 & 13.29 & 12.44 & 11.83 & $BR$/S0 \\
958: NCG 424 & 14.18 & 13.76 & 12.91 & 12.38 & 11.77 & $BR$/S0 \\
959: NGC 1068 & ~9.70 & ~9.61 & ~8.87 & ~8.31 & ~7.66 & $UBV$/Sab \\
960: NGC 1386 & 12.42 & 12.09 & 11.23 & 10.76 & 10.15 & $UBVR$/S0 \\
961: NGC 2992 & 13.54 & 13.14 & 12.18 & 11.62 & 10.97 & $UBV$/S0 \\
962: NGC 3281 & 13.12 & 12.70 & 11.72 & 11.17 & 10.56 & $BV$/S0 \\
963: NGC 4507 & 13.05 & 12.92 & 12.07 & 11.70 & 11.05 & $UBVR$/Sab \\
964: NGC 5135 & 13.01 & 12.88 & 12.11 & 11.55 & 10.90 & $UBV$/Sab \\
965: NGC 5283 & 14.53 & 14.20 & 13.42 & 12.86 & 12.21 & $B$/Sab \\
966: NGC 5506 & 13.21 & 12.79 & 11.92 & 11.38 & 10.77 & $BV$/S0 \\
967: NGC 5643 & 10.89 & 10.74 & 10.00 & ~9.48 & ~8.87 & $UBV$/Sbc \\
968: NGC 5728 & 12.70 & 12.37 & 11.59 & 11.03 & 10.38 & $B$/Sab \\
969: NGC 6890 & 13.14 & 13.01 & 12.25 & 11.57 & 10.92 & $UBVR$/Sab \\
970: NGC 7172 & 13.24 & 12.85 & 11.91 & 11.15 & 10.54 & $UBVR$/S0 \\
971: NGC 7314 & 11.57 & 11.62 & 11.01 & 10.61 & ~9.99 & $UBVR$/Sbc \\
972: NGC 7582 & 11.62 & 11.37 & 10.62 & 10.06 & ~9.41 & $UBV$/Sab
973: \enddata
974: \vskip -0.3truein
975: \tablecomments{Uncorrected for Galactic extinction.}
976: \end{deluxetable}
977: \end{center}
978:
979: \begin{center}
980: \begin{deluxetable}{lccccc}
981: \tablewidth{0pt}
982: \tabletypesize \scriptsize
983: \tablecaption{Magnitudes Corrected for Galactic Extinction}
984: \tablehead{\colhead{Galaxy} &
985: \colhead{$U$} &
986: \colhead{$B$} &
987: \colhead{$V$} &
988: \colhead{$R$} &
989: \colhead{$I$}}
990: \startdata
991: IC 3639 & {\it 12.96} & 12.71 & {\it 12.00} & 11.69 & {\it 11.09}\\
992: MCG $-$05-18-002 & 13.05 & 12.67 & 11.27 & 10.75 & 10.10\\
993: MCG $-$05-23-016 & {\it 13.90} & 13.60 & {\it 12.93} & 12.15 & {\it 11.62}\\
994: MCG $+$01-27-020 & 14.66 & 14.70 & 13.98 & 13.57 & 12.99\\
995: Mrk 3 & 12.86 & 12.79 & 11.77 & 11.14 & 10.60\\
996: Mrk 1066 & 13.79 & 13.60 & 12.75 & 12.21 & 11.57 \\
997: NCG 262 & 13.32 & 13.38 & 12.62 & 12.10 & 11.59 \\
998: NCG 424 & {\it 14.10} & 13.69 & {\it 12.86} & 12.34 & {\it 11.74}\\
999: NCG 591 & 13.90 & 13.80 & 13.03 & 12.50 & 11.87\\
1000: NGC 788 & 13.30 & 12.91 & 11.96 & 11.43 & 10.75\\
1001: NGC 1068 & ~9.52 & ~9.47 & ~8.76 & {\it ~8.22} & {\it ~7.60}\\
1002: NGC 1358 & 13.16 & 12.79 & 11.88 & 11.31 & 10.71\\
1003: NGC 1386 & 12.35 & 12.04 & 11.19 & 10.73 & {\it 10.13}\\
1004: NGC 1667 & 12.46 & 12.45 & 11.76 & 11.24 & 10.67\\
1005: NGC 1685 & 14.27 & 14.05 & 13.21 & 12.58 & 12.05\\
1006: NGC 2110 & 12.46 & 11.60 & 10.56 & 10.06 & 9.45\\
1007: NGC 2273 & 12.41 & 12.30 & 11.37 & 10.81 & 10.15\\
1008: NGC 2992 & 13.21 & 12.88 & 11.98 & {\it 11.46} & {\it 10.85}\\
1009: NGC 3081 & 12.92 & 12.72 & 11.87 & 11.37 & 10.77\\
1010: NGC 3281 & {\it 12.60} & 12.29 & 11.40 & {\it 10.91} & {\it 10.37}\\
1011: NGC 3982 & 12.02 & 12.12 & 11.55 & 11.11 & 10.54\\
1012: NGC 4117 & 14.02 & 13.79 & 12.95 & 12.39 & 11.75\\
1013: NGC 4388 & 11.68 & 11.58 & 10.85 & 10.36 & ~9.69\\
1014: NGC 4507 & 12.52 & 12.50 & 11.75 & 11.44 & {\it 10.86}\\
1015: NGC 4941 & 11.83 & 11.90 & 11.03 & 10.51 & ~9.79\\
1016: NGC 5135 & 12.69 & 12.62 & 11.91 & {\it 11.39} & {\it 10.78}\\
1017: NGC 5283 & {\it 13.98} & 13.76 & {\it 13.08} & {\it 12.59} & {\it 12.01}\\
1018: NGC 5347 & 13.11 & 13.08 & 12.45 & 11.87 & 11.09\\
1019: NGC 5506 & {\it 12.89} & 12.53 & 11.72 & {\it 11.22} & {\it 10.65}\\
1020: NGC 5643 & ~9.97 & 10.01 & ~9.44 & {\it ~9.03} & {\it ~8.54}\\
1021: NGC 5695 & 13.73 & 13.41 & 12.60 & 12.05 & 11.28\\
1022: NGC 5728 & {\it 12.59} & 12.28 & {\it 11.52} & {\it 10.98} & {\it 10.34}\\
1023: NGC 5929 & 12.63 & 12.89 & 11.97 & 11.38 & 10.67\\
1024: NGC 6890 & 12.92 & 12.84 & 12.12 & 11.46 & {\it 10.84}\\
1025: NGC 7172 & 13.10 & 12.74 & 11.82 & 11.08 & {\it 10.49}\\
1026: NGC 7314 & 11.45 & 11.53 & 10.94 & 10.55 & {\it ~9.95}\\
1027: NGC 7582 & 11.54 & 11.31 & 10.57 & {\it 10.02} & {\it ~9.38}\\
1028: NGC 7672 & 14.56 & 14.45 & 13.70 & 13.21 & 12.36
1029: \enddata
1030: \end{deluxetable}
1031: \end{center}
1032:
1033: \begin{center}
1034: \begin{deluxetable}{lcccccccccc}
1035: \tablewidth{0pt}
1036: \rotate
1037: \tabletypesize \footnotesize
1038: \tablecaption{X-ray Spectral Parameters and Fluxes}
1039: \tablehead{\colhead{Galaxy} &
1040: \colhead{$d$(Mpc)} &
1041: \colhead{$(N_{\rm H})_1$} &
1042: \colhead{$\Gamma_1$} &
1043: \colhead{$(N_{\rm H})_2$} &
1044: \colhead{$\Gamma_2$} &
1045: \colhead{$E_{K\alpha}$} &
1046: \colhead{$\sigma_{K\alpha}$} &
1047: \colhead{EW$_{K\alpha}$} &
1048: \colhead{$F$(0.5--2 keV)} &
1049: \colhead{$F$(2--8 keV)}}
1050: \startdata
1051: IC 3639 & 43.7 & 0 & 1.78 & 2.42E23 & 1.70 & ... & ... & ... & 1.30E-13 & 2.43E-13\\
1052: MCG $-$05-18-002 & 23.1 & 0 & 1.78 & 2.42E23 & 1.70 & ... & ... & ... & 2.07E-13 & 5.28E-13\\
1053: MCG $-$05-23-016 & 33.3 & ... & ... & 1.50E22 & 1.80 & 6.35 & 0.39 & 0.29 & 8.25E-12 & 7.21E-11\\
1054: MCG $+$01-27-020 & 46.8 & 0 & 1.78 & 2.42E23 & 1.70 & ... & ... & ... & 3.40E-14 & 4.21E-13\\
1055: Mrk 3 & 54.0 & 0 & 2.05 & 4.74E23 & 1.83 & 6.31 & 0.10 & 0.73 & 8.26E-13 & 2.77E-12\\
1056: Mrk 1066 & 48.1 & 0 & 1.78 & 2.42E23 & 1.70 & ... & ... & ... & 1.62E-13 & 3.92E-13\\
1057: NCG 262 & 60.1 & 0 & 1.77 & 1.77E23 & 1.77 & 6.16 & 0.10 & 0.14 & 1.38E-13 & 3.56E-12\\
1058: NCG 424 & 46.6 & 0 & 1.61 & 2.42E23 & 1.61 & 6.41 & 0.10 & 1.49 & 2.14E-13 & 1.04E-12\\
1059: NCG 591 & 60.7 & 0 & 1.78 & 2.42E23 & 1.70 & ... & ... & ... & 3.76E-14 & 2.90E-13\\
1060: NGC 788 & 54.4 & 0 & 1.15 & 4.80E23 & 1.67 & 6.23 & 0.24 & 0.39 & 8.13E-14 & 3.48E-12\\
1061: NGC 1068 & 14.4 & 2.90E21 & 5.59 & 3.16E21 & 1.28 & 6.52 & 0.30 & 3.26 & 1.02E-11 & 5.19E-12\\
1062: NGC 1358 & 53.8 & 0 & 1.78 & 2.42E23 & 1.70 & ... & ... & ... & 5.14E-14 & 4.27E-13\\
1063: NGC 1386 & 16.9 & 2.04E21 & 3.47 & 1.94E23 & 1.96 & 6.40 & 0.10 & 0 & 4.90E-13 & 4.46E-13\\
1064: NGC 1667 & 60.7 & 0 & 1.78 & 2.42E23 & 1.70 & ... & ... & ... & 2.81E-14 & 1.27E-13\\
1065: NGC 1685 & 60.4 & 0 & 1.78 & 2.42E23 & 1.70 & ... & ... & ... & 3.38E-14 & 8.31E-14\\
1066: NGC 2110 & 29.1 & 5.80E20 & 1.61 & 3.61E22 & 1.61 & 6.26 & 0.51 & 0.41 & 1.20E-12 & 2.39E-11\\
1067: NGC 2273 & 28.4 & 0 & 1.95 & 4.41E23 & 1.95 & 6.36 & 0.11 & 1.38 & 2.25E-13 & 9.69E-13\\
1068: NGC 2992 & 30.5 & 0 & 4.58 & 9.74E21 & 1.61 & 6.46 & 0.47 & 1.59 & 1.30E-12 & 3.99E-12\\
1069: NGC 3081 & 32.5 & 0 & 2.01 & 5.36E23 & 1.58 & 6.22 & 0.13 & 0.41 & 2.12E-13 & 3.75E-12\\
1070: NGC 3281 & 42.7 & 0 & 1.85 & 6.02E23 & 1.64 & 6.30 & 0.10 & 0.99 & 2.72E-13 & 2.25E-12\\
1071: NGC 3982 & 27.2 & 0 & 1.78 & 2.42E23 & 1.70 & ... & ... & ... & 1.03E-13 & 3.22E-13\\
1072: NGC 4117 & 17.0 & 2.21E21 & 1.60 & 3.99E23 & 1.60 & 6.43 & 0.10 & 0.11 & 7.62E-14 & 1.31E-12\\
1073: NGC 4388 & 16.8 & 0 & 1.50 & 3.98E23 & 1.96 & 6.34 & 0.16 & 0.62 & 5.41E-13 & 4.86E-12\\
1074: NGC 4507 & 47.2 & 0 & 1.78 & 3.02E23 & 1.53 & 6.23 & 0.23 & 0.30 & 5.41E-13 & 1.73E-11\\
1075: NGC 4941 & ~6.4 & 1.18E21 & 1.61 & 9.13E23 & 1.61 & 6.35 & 0.10 & 0.64 & 1.08E-13 & 8.26E-13\\
1076: NGC 5135 & 54.9 & 0 & 2.88 & 3.84E23 & 1.80 & 6.34 & 0.10 & 1.44 & 5.36E-13 & 4.65E-13\\
1077: NGC 5283 & 41.4 & 1.10E21 & 3.28 & 1.06E23 & 1.60 & 6.40 & 0.10 & 0 & 3.91E-14 & 1.14E-12\\
1078: NGC 5347 & 36.7 & 0 & 1.78 & 2.42E23 & 1.70 & ... & ... & ... & 5.16E-14 & 1.94E-13\\
1079: NGC 5506 & 28.7 & 1.11E21 & 2.20 & 3.02E22 & 1.82 & 6.39 & 0.23 & 0.22 & 3.81E-12 & 6.03E-11\\
1080: NGC 5643 & 16.9 & 0 & 1.90 & 1.76E23 & 1.64 & 6.34 & 0.10 & 2.73 & 4.49E-13 & 1.12E-12\\
1081: NGC 5695 & 56.4 & 0 & 1.78 & 2.42E23 & 1.70 & ... & ... & ... & 5.05E-14 & 1.23E-13\\
1082: NGC 5728 & 42.2 & 3.63E20 & 2.07 & 7.17E23 & 1.67 & 6.33 & 0.10 & 1.12 & 9.22E-14 & 1.00E-12\\
1083: NGC 5929 & 38.5 & 5.16E21 & 1.70 & 2.77E23 & 1.70 & 6.19 & 0.10 & 0.35 & 8.06E-14 & 1.40E-12\\
1084: NGC 6890 & 31.8 & 0 & 1.78 & 2.42E23 & 1.70 & ... & ... & ... & 3.99E-14 & 1.69E-13\\
1085: NGC 7172 & 33.9 & 0 & 1.69 & 9.62E22 & 1.82 & 6.44 & 0.10 & 0.20 & 1.93E-13 & 1.01E-11\\
1086: NGC 7314 & 18.3 & ... & ... & 7.62E21 & 1.95 & 6.24 & 0.77 & 0.45 & 8.36E-12 & 3.42E-11\\
1087: NGC 7582 & 17.6 & 0 & 1.68 & 1.23E23 & 1.82 & 6.24 & 0.10 & 0.18 & 4.51E-13 & 1.23E-11\\
1088: NGC 7672 & 53.5 & 0 & 1.78 & 2.42E23 & 1.70 & ... & ... & ... & 4.80E-14 & 9.08E-14
1089: \enddata
1090: \vskip -0.1truein
1091: \tablecomments{Column densities ($N_{\rm H}$) are in units of atoms cm$^{-2}$.
1092: Iron line centroids ($E_{K\alpha}$), energy widths ($\sigma_{K\alpha}$),
1093: and equivalent widths (EW$_{K\alpha}$) are all in units of keV. Fluxes are
1094: in units of erg~cm$^{-2}$~s$^{-1}$.}
1095: \end{deluxetable}
1096: \end{center}
1097:
1098: \begin{center}
1099: \begin{deluxetable}{lc}
1100: \tablewidth{0pt}
1101: \tabletypesize \footnotesize
1102: \tablecaption{Host Galaxy Axis Ratios}
1103: \tablehead{\colhead{Galaxy} &
1104: \colhead{$b/a$}}
1105: \startdata
1106: MCG $-$05-23-016~~~~~~~~~~~~~~~~~~~~~ & 0.45\\
1107: Mrk 3 & 0.97\\
1108: NCG 262 & 0.79\\
1109: NCG 424 & 0.33\\
1110: NCG 788 & 0.76\\
1111: NCG 2110 & 0.80\\
1112: NCG 2992 & 0.39\\
1113: NCG 3081 & 0.57\\
1114: NCG 3281 & 0.43\\
1115: NCG 4388 & 0.28\\
1116: NCG 4507 & 0.86\\
1117: NCG 5283 & 0.91\\
1118: NCG 5506 & 0.32\\
1119: NCG 5728 & 0.74\\
1120: NCG 5929 & 0.78\\
1121: NCG 7172 & 0.52\\
1122: NCG 7314 & 0.39\\
1123: NCG 7582 & 0.46
1124: \enddata
1125: \vskip -0.1truein
1126: \tablecomments{Galaxies included here are those that would be detected as
1127: absorbed AGNs in the CDF-S at $z \ge 0.5$. NGC~5929 is interacting with
1128: NGC~5920; the axis ratio listed is for NGC~5929 alone.}
1129: \end{deluxetable}
1130: \end{center}
1131:
1132:
1133: \begin{figure}
1134: \begin{center}
1135: \epsscale{1.0}
1136: \centerline{\plotone{f1.eps}}
1137: \vskip -0.1truein
1138: \caption{Distributions of the redshifts, ACIS-I exposure times, 2--8 keV
1139: fluxes, and observed-frame 2--8 keV luminosities of the absorbed AGNs in
1140: the CDF-N that meet the selection criteria described in \S~2.}
1141: \end{center}
1142: \end{figure}
1143:
1144: \begin{figure}
1145: \begin{center}
1146: \epsscale{1.0}
1147: \centerline{\plottwo{f2a.eps}{f2b.eps}}
1148: \vskip -0.7truein
1149: \caption{({\it left panel}) Comparison of the integrated $V$ magnitudes of
1150: 11 objects from our CCD photometry with those from photoelectric measurements
1151: published in the RC3. ({\it right panel}) Integrated $R$ magnitudes of the
1152: same objects from our photometry, compared with those extrapolated from the
1153: RC3 $V$ magnitude using a Fukugita et al.\ (1995) galaxy color template.
1154: In both cases, the mean and dispersion of the magnitude differences are
1155: consistent with the uncertainties in the RC3 measurements.}
1156: \end{center}
1157: \end{figure}
1158:
1159: \begin{figure}
1160: \begin{center}
1161: \epsscale{1.0}
1162: \centerline{\plotone{f3.eps}}
1163: \vskip -0.5truein
1164: \caption{Comparison of the 2--8 keV fluxes of the well-detected UW89 objects
1165: obtained directly from spectral modeling with those obtained by scaling the
1166: median model to agree with the measured (4--10 keV)/(1--4 keV) counts ratio.
1167: For most sources, the two fluxes differ by only a few percent, suggesting
1168: that the application of the median model provides accurate flux estimates
1169: for the 13 weakly detected UW89 objects.}
1170: \end{center}
1171: \end{figure}
1172:
1173: \begin{figure}
1174: \begin{center}
1175: \epsscale{0.9}
1176: \centerline{\plotone{f4.eps}}
1177: %\vskip -1.5truein
1178: \caption{The volume searched in the CDF-N vs.\ minimum detectable luminosity in
1179: the 2--8 keV band. The volume is calculated assuming the flux limit and solid
1180: angle that correspond to an ACIS-S exposure time of 1500~ks. Also plotted is
1181: the (fixed) volume covered by the UW89 sample of nearby Seyfert~2 galaxies.}
1182: \end{center}
1183: \end{figure}
1184:
1185: \begin{figure}
1186: \begin{center}
1187: \epsscale{0.85}
1188: \centerline{\plotone{f5.eps}}
1189: \vskip -1.0truein
1190: \caption{({\it left}) Fluxes in the observed 2--8 keV and $I$ bands vs.\
1191: redshift, and ({\it right}) \fxfi\ vs.\ redshift for four UW89 Seyfert~2s
1192: that span a wide range of \fxfi\ at $z = 0$. The \fxfi\ ratios of these
1193: objects increase dramatically by factors of 15--35 as the redshift increases
1194: from $z = 0$ to $z = 1.5$.}
1195: \end{center}
1196: \end{figure}
1197:
1198: \begin{figure}
1199: \begin{center}
1200: \epsscale{0.9}
1201: \centerline{\plotone{f6.eps}}
1202: \vskip -0.0truein
1203: \caption{Intrinsic ($z = 0$) \fxfi\ distribution for the UW89 sample, and
1204: the distribution that would be obtained if the same sources were observed
1205: in the CDF-N. The dramatic transformation of the flux-ratio distribution
1206: results from a combination of the redshift effects displayed in Fig.~5 and
1207: the Malmquist effects (see \S~4.2) that arise because of the flux-limited
1208: nature of the CDF-N.}
1209: \end{center}
1210: \end{figure}
1211:
1212: \begin{figure}
1213: \begin{center}
1214: \epsscale{1.0}
1215: \centerline{\plotone{f7.eps}}
1216: \vskip -1.7truein
1217: \caption{Distribution of the observed-frame 2--8~keV/$I$-band flux ratio
1218: for absorbed AGNs in the CDF-N, compared to the \fxfi\ ratios that the UW89
1219: Seyfert~2 galaxies would have if they were observed in the CDF-N. The
1220: similarity of the distribuitions suggests that nearby Seyfert~2s and distant
1221: absorbed AGNs do not differ in some fundamental way, despite the fact that
1222: most of the latter lack spectroscopic evidence of activity in ground-based
1223: optical observations.}
1224: \end{center}
1225: \end{figure}
1226:
1227: \begin{figure}
1228: \begin{center}
1229: \epsscale{1.0}
1230: \centerline{\plotone{f8.eps}}
1231: \vskip -1.7truein
1232: \caption{\fxfi\ ratio for the CDF-N absorbed AGN sample, as a function of
1233: observed 2--8 keV luminosity. Also plotted are points associated with UW89
1234: galaxies from a simulation (see \S~4.1) consisting of 75 successful trials.
1235: Overall, the simulated UW89 Seyfert~2 sample provides a close match to the
1236: CDF-N objects over a broad range of luminosities and \fxfopt\ ratios.}
1237: \end{center}
1238: \end{figure}
1239:
1240: \begin{figure}
1241: \begin{center}
1242: \epsscale{0.9}
1243: \centerline{\plotone{f9.eps}}
1244: \vskip -0.0truein
1245: \caption{Distribution of host galaxy axis ratios for subsets of the CDF-S and
1246: UW89 samples. The CDF-S sources are limited to those in the $0.5 \le z \le
1247: 0.8$ subsample of Rigby et al.\ (2006) that meet our criteria for absorbed
1248: AGNs. The UW89 objects included here are those that would be detected as
1249: absorbed AGNs at $z \ge 0.5$ in the CDF-S.
1250: The similarity of the CDF-S and UW89 axis-ratio distributions, and the fact
1251: that the UW89 galaxies have strong nuclear emission lines, suggest that
1252: inclination effects are not the main reason the CDF-S objects appear
1253: optically inactive in ground-based spectra.}
1254: \end{center}
1255: \end{figure}
1256:
1257: \begin{figure}
1258: \begin{center}
1259: \epsscale{1.0}
1260: \centerline{\plotone{f10.eps}}
1261: \vskip 0.0truein
1262: \caption{Integrated spectra of two UW89 Seyfert~2 galaxies from the study
1263: of Moran et al.\ (2002), plotted at the same scale. Although the objects
1264: are nearly identical in most respects, the strengths of their nuclear
1265: emission lines are very different. Objects like NGC~788, which is more
1266: typical of nearby Seyfert~2s than Mrk~3 is, would not be recognized as
1267: AGNs at $z > 0.5$ in the {\it Chandra\/} deep surveys.}
1268: \end{center}
1269: \end{figure}
1270:
1271: \end{document}
1272: