astro-ph0106519/ms.tex
1: \documentclass[preprint]{aastex}
2: %\usepackage{emulateapj5}
3: %\usepackage[]{natbib}
4: %\usepackage{graphics}
5: %\usepackage{apjfonts}
6: \usepackage{psfig}
7: 
8: \newcommand{\et}{et~al.\ }
9: \newcommand{\Ha}{H$\alpha$}
10: \newcommand{\Hb}{H$\beta$}
11: \newcommand{\HeI}{\ion{He}{1}}
12: \newcommand{\OIII}{[\ion{O}{3}]}
13: \newcommand{\NII}{[\ion{N}{2}]}
14: \newcommand{\HII}{\ion{H}{2}}
15: \newcommand{\Msun}{\ensuremath{M_\odot}}
16: \newcommand\ltwid{\mathrel{\raise.5ex\hbox{$<$\kern-.75em\lower1.2ex\hbox{$\sim$}}}}
17: 
18: 
19: %\slugcomment{}
20: \shorttitle{COMPOSITE SEYFERT 2 X-RAY SPECTRUM}
21: \shortauthors{MORAN ET AL.}
22: 
23: \begin{document}
24: 
25: \title{A Composite Seyfert 2 X-ray Spectrum: Implications for the
26:        Origin of the Cosmic X-ray Background}
27: 
28: \author{Edward C.\ Moran\altaffilmark{1,2}, Laura E.\ Kay\altaffilmark{3},
29:         Marc Davis\altaffilmark{2}, Alexei V.\ Filippenko\altaffilmark{2},
30:         and Aaron J.\ Barth\altaffilmark{4}}
31: 
32: \altaffiltext{1}{Chandra Fellow.}
33: 
34: \altaffiltext{2}{Department of Astronomy, University of California,
35:                  Berkeley, CA 94720-3411.}
36: 
37: \altaffiltext{3}{Department of Physics and Astronomy, Barnard College,
38:                  Columbia University, New York, NY 10027.}
39: 
40: \altaffiltext{4}{Harvard-Smithsonian Center for Astrophysics, 60 Garden
41:                  Street, Cambridge, MA 02138.}
42: 
43: 
44: \begin{abstract}
45: We present a composite 1--10 keV Seyfert~2 X-ray spectrum, derived from
46: {\sl ASCA\/} observations of a distance-limited sample of nearby galaxies.
47: All 29 observed objects were detected.  Above $\sim$~3 keV, the composite
48: spectrum is inverted, confirming that Seyfert~2 galaxies as a class have
49: the spectral properties necessary to explain the flat shape of the cosmic
50: X-ray background spectrum.  Integrating the composite spectrum over redshift,
51: we find that the total emission from Seyfert~2 galaxies, combined with
52: the expected contribution from unabsorbed type~1 objects, provides an
53: excellent match to the spectrum and intensity of the hard X-ray background.
54: The principal uncertainty in this procedure is the cosmic evolution of
55: the Seyfert~2 X-ray luminosity function.  Separate composite spectra for
56: objects in our sample with and without polarized broad optical emission
57: lines are also presented.
58: \end{abstract}
59: \keywords{galaxies:~Seyfert --- X-rays:~diffuse background --- X-rays:~galaxies}
60: 
61: 
62: \newpage
63: \section{Introduction}
64: 
65: Extensive observational effort over the past two decades has provided
66: compelling evidence that various types of radio-quiet active galactic
67: nuclei (AGNs), despite the diversity of their apparent properties, are
68: essentially similar objects that {\it look\/} different because of the
69: presence or absence of dense circumnuclear material along our line of sight
70: (Antonucci 1993).  The primary effect of such material is to obscure the
71: innermost regions of an active nucleus; this gives rise to a type~2 optical
72: spectrum which consists of relatively narrow permitted and forbidden emission
73: lines and lacks the very broad Balmer lines that characterize the spectra
74: of unobscured (type~1) AGNs.  In addition, the dense material preferentially
75: attenuates the soft X-ray emission produced by an AGN, flattening its
76: X-ray spectrum at lower energies (Awaki \et 1991).  This unified picture of
77: AGNs has thus allowed significant progress to be made toward a resolution
78: of the most enduring problem in X-ray astronomy---the origin of the cosmic
79: X-ray background (XRB)---by providing a means by which the integrated
80: emission from intrinsically steep-spectrum AGNs might be able to match the
81: flat observed spectrum of the background.  Models based on AGN unification
82: have shown that absorbed active nuclei may indeed account for both the
83: spectrum and intensity of the XRB over~a broad range of energies
84: (Setti \& Woltjer 1989; %Madau \et 1994; Comastri \et 1995).
85: Madau, Ghisellini, \& Fabian 1994; Comastri et al.~1995).
86: 
87: These models, however, rely on several crucial assumptions: (1) that
88: {\it all\/} type~2 AGNs are obscured type~1 objects; (2) that the absorption
89: column densities of active galaxies have a specific distribution; and (3)
90: that the X-ray luminosity function of absorbed AGNs evolves with redshift
91: in a particular way.  None of these conditions has been fully verified
92: observationally.  To explore the issue of whether or not all type~2 AGNs
93: are hidden type~1 objects, we have conducted an optical spectropolarimetric
94: survey of a well-defined sample of 31 nearby Seyfert 2 galaxies (Moran
95: \et 2000).  Broad \Ha\ emission has been detected in 13 of the objects,
96: suggesting that at least 42\% of Seyfert~2 galaxies are indeed obscured
97: Seyfert~1s.\footnote{Re-observation of part of our sample has revealed
98: faint polarized broad emission lines in NGC~5347 and NGC~5929, raising the
99: number of hidden Seyfert~1s in our sample from 11 to 13.  Details will
100: be presented in a future paper.}  Recent progress has also been made
101: toward a determination of the column density distribution in Seyfert
102: galaxies (e.g., Bassani \et 1999; Risaliti, Maiolino, \& Salvati 1999).
103: An alternative approach for assessing the contribution of absorbed AGNs to
104: the XRB would be to measure their integrated broadband spectral properties
105: within a known volume.  A composite spectrum based on the summed emission
106: of such objects would simultaneously represent several characteristics of
107: the Seyfert~2 population that are free parameters in AGN models of the XRB.
108: In this {\it Letter}, we present for the first time a composite 1--10 keV
109: Seyfert~2 X-ray spectrum, derived from a large, unbiased sample of nearby
110: objects, and discuss its implications for the origin of the XRB.
111: 
112: 
113: \section{X-ray Observations and Analysis}
114: 
115: Our study is based on the distance-limited sample of Ulvestad \& Wilson
116: (1989, hereafter UW89; see also Moran \et 2000), which contains 31 type~2
117: Seyfert galaxies with recession velocities $cz < 4600$ km~s$^{-1}$ and
118: declinations $\delta > -45^{\circ}$.  As of mid-1998, 22 objects in the
119: sample had been (or were scheduled to be) observed with the {\sl ASCA\/}
120: and/or {\sl BeppoSAX\/} satellites.  We were awarded time for the nine
121: remaining targets in {\sl ASCA\/} Cycle~7 .  Unfortunately, two of the
122: observations scheduled prior to this were given low-priority status and
123: were never carried out.  The galaxies lacking X-ray data (NGC 5283 and
124: NGC 5728) have average radio and optical luminosities compared to the rest
125: of the sample, and their absence here is not expected to have a significant
126: impact on our results.  Although a number of objects were observed with
127: both {\sl ASCA\/} and {\sl BeppoSAX}, all 29 were observed with {\sl ASCA},
128: so we have limited our study to that dataset.  In addition, because of
129: the consistency of the response and field of view of {\sl ASCA}'s Gas
130: Imaging Spectrometers (GIS2 and GIS3), we have utilized data obtained with
131: those instruments only.
132: 
133: We have employed ``rev2''-processed and screened {\sl ASCA\/} data.\footnote
134: {See {\tt http://heasarc.gsfc.nasa.gov/docs/asca/ascarev2.html} for details
135: regarding rev2 data processing.}  Exposures for the majority of the sample
136: range between $\sim$~35 ks and 45 ks.  All of the galaxies were observed at
137: either the ``1-CCD'' or ``2-CCD'' positions in the field of view, so all
138: objects are
139: approximately the same distance off-axis in the GIS images.  Two galaxies,
140: NGC 1068 and Mrk 3, were observed twice each, but in both cases the additional
141: data were obtained at different off-axis locations and have therefore been
142: omitted.  Source counts were extracted within a $6'$ radius region centered
143: on the galaxy for most objects; $4'$--$5'$ radius regions were used for the
144: weakest sources.  The background was estimated within a $13'$ radius circle
145: centered on the optical axis, excluding an $8'$ region around the source
146: and small areas around resolved background sources, if present.  Table~1
147: lists the galaxies observed, along with their distances (assuming $H_0 =
148: 75$ km s$^{-1}$ Mpc$^{-1}$) and GIS exposure times.  Also listed are the
149: signal-to-noise ratios achieved with the combined GIS2 and GIS3 data in the
150: 1--10 keV band.  As the table indicates, all objects were detected above a
151: significance of $\sim 4\sigma$.  Note, however, that over a third of the
152: sample is detected at or below $10\sigma$; thus, it would be impossible
153: to determine an accurate distribution of absorption column densities with
154: these data.  Source and background spectra with 128 channels were created
155: for each object from the extracted events, and effective area files were
156: generated for each of the source spectra.
157: 
158: Our objective is to construct a composite spectrum that is representative
159: of the local Seyfert~2 population, i.e., one in which the contribution of
160: individual objects is governed by their apparent luminosities, not their
161: observed fluxes.  Thus, we have multiplied the exposure times in the FITS
162: headers of the background-subtracted spectra of each galaxy by the factor
163: $(60/d)^2$, where $d$ is the distance of the source in Mpc from Table~1.
164: The modified GIS count rates are listed in the final column of Table~1.
165: Summing the modified count rates of the objects (using the FTOOLS task
166: {\tt mathpha}), we obtain a composite spectrum that represents what we
167: would observe if the objects were all located at 60~Mpc.  The sums were
168: carried out for the GIS2 and GIS3 data separately, yielding nearly
169: identical results.  The GIS2 and GIS3 sums were then averaged to produce
170: a final composite 1--10 keV Seyfert~2 X-ray spectrum.  Poisson errors were
171: propagated at each step in the procedure.  The individual effective area
172: files were averaged with the {\tt addarf} task.  We note that if the
173: objects {\it were\/} all 60~Mpc away, the data obtained would correspond
174: to a total exposure time of $8.5 \times 10^6$~s, which illustrates why it
175: would be difficult to repeat this experiment for a more distant sample
176: of comparable size.
177: 
178: 
179: \section{The Composite Seyfert~2 X-ray Spectrum}
180: 
181: The composite 1--10 keV Seyfert 2 spectrum, derived from the combined
182: emission of 29 galaxies, is displayed in Figure~1.  There is a considerable
183: amount of steep-spectrum emission below 3 keV, which was not anticipated
184: in the original Seyfert~2 XRB models.  Above 3~keV, however, the composite
185: spectrum is inverted, with strong Fe~K$\alpha$ line emission at 6.4 keV
186: (rest).  Again, the objects in our sample are included on the basis of
187: their distances and optical classifications, not their X-ray properties.
188: The composite spectrum thus verifies that Seyfert~2 galaxies {\it as a
189: class\/} have the X-ray spectral properties necessary to explain the flat
190: shape of the XRB spectrum---something which cannot be claimed for any other
191: extragalactic population.  Examination of the modified count rates listed
192: in Table~1 indicates that a single source, NGC~4507, contributes $\sim 30$\%
193: of the flux of the composite spectrum and that eight other objects (Mrk 3,
194: NGC 262, 424, 788, 1068, 3081, 3281, and 5135) are responsible for an
195: additional $\sim 50$\%.  Among these, NGC 424, 1068, and 5135 contribute
196: mainly in the soft X-ray band.  The others, which are the strongest sources
197: above 4~keV, all have ``Compton-thin'' spectra with absorption column
198: densities (inferred with the model described below) of $N_{\rm H}$ =
199: 2--8 $\times 10^{23}$ cm$^{-2}$.
200: 
201: We have modeled the composite spectrum as the sum of a Gaussian line and
202: two power laws, one which is heavily absorbed and one which is not.  An
203: acceptable fit is obtained if we require the energy index $\alpha$ of the
204: power-law components to have the same value.  Such a model corresponds to
205: the ``partial scattering'' scenario, in which a single continuum component
206: is observed both in transmission through dense material and via scattering
207: along a less obscured line of sight.  However, for a handful of Seyfert~2
208: galaxies that have been studied with {\sl BeppoSAX}, this model underpredicts
209: the observed flux in the 20--30 keV range (e.g., Cappi \et 1999), probably
210: because it fails to account for the presence of Compton-reflected emission.
211: Limited to the 1--10 keV band, we are unable to constrain the strength of a
212: reflection component in the composite spectrum.  A double power-law
213: model in which the energy indices are independent also provides an acceptable
214: fit to the composite spectrum; the fit is shown in Figure~1.  This model may
215: not provide the correct physical description of the emission, but when it is
216: applied to the {\sl ASCA\/} spectra of individual Seyfert~2s observed with
217: {\sl BeppoSAX\/} (e.g., several of the objects studied by Cappi \et 2000)
218: it {\it does\/} accurately predict their measured fluxes in the 20--30 keV
219: range.  This is important because our estimate of the contribution of
220: Seyfert~2 galaxies to the intensity of the XRB (\S~4) necessarily involves
221: an extrapolation of the composite spectrum to higher energies.
222: 
223: We have compiled composite spectra separately for objects in our sample
224: which do and do not exhibit evidence for polarized broad optical emission
225: lines (see Table~1).  These are displayed in Figure~2.  Although there are
226: differences in detail between these spectra, they have similar qualitative
227: features: a declining continuum in the soft X-ray band which rises sharply
228: at harder energies, and a prominent Fe~K$\alpha$ line.  Thus, objects
229: lacking optical evidence for a hidden broad-line region---at least the
230: X-ray--brightest of them---display the characteristics of obscured type~1
231: Seyferts in the 1--10 keV range.  Note that the steep soft X-ray component
232: is relatively stronger in Seyfert~2s that do {\it not\/} have polarized
233: optical emission lines, which casts some doubt on the possibility that
234: the soft X-ray flux in these objects is mainly scattered emission.
235: 
236: 
237: \section{Seyfert 2 Galaxies and the X-ray Background}
238: 
239: As we have discussed previously (Moran \et 2000), the UW89 objects comprise
240: a reasonably complete, distance-limited sample of classical Seyfert~2 galaxies.
241: Since they occupy a known volume, the composite spectrum we have derived is
242: effectively an unbiased measure of the local Seyfert~2 volume emissivity
243: (luminosity per unit volume) as a function of X-ray energy.  This can be
244: integrated over redshift using the following equation to provide an estimate
245: of the sky brightness at energy $E$ due to such objects:
246: $$I_{\rm E} = {c \over {4 \pi H_0}}
247: \int\displaylimits_{0}^{z_{\rm max}} {{\rho(E')\; e(z)\; dz} \over 
248: {(1 + z)\, [\Omega_{\rm M}\, (1 + z)^3 + \Omega_{\Lambda}]^{1/2} }}\; ,$$
249: \noindent
250: where $\rho$ is the volume emissivity at $E' = E(1 + z)$, $e(z)$ is a term
251: that describes the evolution of the volume emissivity with redshift, and
252: $\Omega_{\rm M}$ and $\Omega_{\Lambda}$ are the matter and vacuum-energy
253: density parameters, respectively.
254: Since we have ``moved'' all of the galaxies to $d = 60$ Mpc,
255: the composite spectrum shown in Figure~1 (which is the sum of their fluxes)
256: is converted to a volume emissivity by multiplying by $4 \pi d^2$ and dividing
257: by 0.6 times the volume of a 60 Mpc sphere.\footnote{The sample does not
258: cover the Galactic plane ($|b| < 20^{\circ}$) or the south celestial cap
259: ($\delta < -45^{\circ}$), hence the factor of 0.6.}  Although the evolution
260: term $e(z)$ is unmeasured for type~2 AGNs, it has been characterized for
261: type~1 objects in the soft X-ray band by Miyaji, Hasinger, \& Schmidt
262: (2000).  For simplicity, we have adopted their pure-density evolution
263: model, which has the following form: $e(z) = (1 + z)^{4.4}$ for $z \le 1.6$,
264: and $e(z) = e(1.6)\, [(1 + z) / 2.6]$ for $z > 1.6$ (T.~Miyaji 2001, private
265: communication).  This evolution model can account for all of the soft XRB
266: with type~1 AGNs, so our results should be considered an upper limit for
267: the contribution of type~2 Seyferts to the hard XRB.  Finally, we have
268: adopted an $\Omega_{\rm M} = 0.3$, $\Omega_{\Lambda} = 0.7$ cosmology for
269: our calculations (Garnavich \et 1998; Balbi \et 2000; Hanany \et 2000)
270: and an upper integration limit of $z_{\rm max} = 6$ (AGNs have been detected
271: out to this redshift).
272: 
273: The results of the integration in the 1--10 keV band are displayed in Figure~3.
274: As anticipated (see Gilli \et 1999), the integration has smoothed out all of
275: the bumps and wiggles in the composite Seyfert~2 spectrum.  Assuming that
276: steep-spectrum ($\alpha = -0.8$) type~1 AGNs produce $\sim$~60\% of the XRB
277: at 1~keV---consistent with the findings of the {\sl ROSAT\/} deep survey
278: of the ``Lockman Hole'' (Schmidt \et 1998)---the combined contribution of
279: Seyfert~1 and Seyfert~2 galaxies agrees exceptionally well with the observed
280: spectrum and intensity of the XRB (from Gendreau \et 1995) above $\sim$~3 keV.
281: Our calculations are not particularly sensitive to the adopted cosmology.
282: For example, if we assume $\Omega_{\rm M} = 1.0$ and $\Omega_{\Lambda} =
283: 0.0$, the Seyfert~2 contribution we obtain has the same overall shape as
284: that shown in Figure~3 and about two-thirds of the integrated intensity.
285: 
286: In conclusion, the spectrum of the summed broadband X-ray emission of
287: nearby Seyfert~2 galaxies and its integration over redshift provide
288: strong support for the hypothesis that absorbed AGNs play a significant
289: (if not dominant) role in the production of the hard X-ray background.
290: Some important details need to be investigated further before we can
291: consider the XRB problem to be solved, namely, the average spectral
292: properties of Seyfert~2 galaxies above 10~keV and the cosmic evolution
293: of the Seyfert~2 X-ray luminosity function.  In addition, if optically
294: classified Seyfert~2 galaxies produce most of the hard XRB, as Figure~3
295: suggests, then the majority of the flat-spectrum sources being detected
296: in deep {\it Chandra\/} images should be Seyfert~2s.  Currently, this
297: does not appear to be the case---a significant fraction of faint, hard
298: {\it Chandra\/} sources have starlight-dominated optical spectra with
299: emission lines that are weak or absent (Barger \et 2001; Hornschemeier
300: \et 2001).  It is possible that the Seyfert~2 X-ray luminosity function
301: evolves less rapidly than we have assumed, which would leave room for
302: other types of AGNs---objects so obscured there is no optical evidence
303: of their activity---to emerge at faint flux levels.  Alternatively, the
304: limitations of ground-based observing have almost certainly affected our
305: ability to classify faint {\it Chandra\/} sources.  The majority of
306: these sources are quite distant, so in spectroscopic observations most of
307: the light from their host galaxies is collected along with the nuclear
308: emission.  At the low spectral resolution that is typically employed,
309: this starlight dilution can have a profound impact on the appearance of
310: the nuclear emission-line spectrum.  A powerful demonstration of this
311: effect has been presented by Kennicutt (1992).  Thus, the faint, hard
312: {\it Chandra\/} source population may {\it not\/} be significantly
313: different from the nearby Seyfert~2 galaxies we have investigated here.
314: 
315: 
316: \acknowledgments
317: 
318: We are grateful to T.~Miyaji for helpful discussions about the AGN
319: X-ray luminosity function, and to the {\sl ASCA\/} Cycle~7 peer review
320: panel for approving a substantial amount of satellite time for this
321: project. We also thank our referee for a number of insightful comments.
322: E.C.M.\ is supported by NASA through Chandra Fellowship grant PF8-10004
323: awarded by the Chandra X-ray Center, which is operated by the Smithsonian
324: Astrophysical Observatory for NASA under contract NAS 8-39073.  L.E.K.\
325: acknowledges the support of the NSF through CAREER grant AST-9501835.
326: A.J.B.\ is supported by a postdoctoral fellowship from the
327: Harvard-Smithsonian Center for Astrophysics.  We also acknowledge NASA
328: grant NAG 5-3556 and a Guggenheim Fellowship to A.V.F.
329: 
330: 
331: \begin{thebibliography}{}
332: 
333: \bibitem{}Antonucci, R.\ 1993, ARA\&A, 31, 473
334: \bibitem{}Awaki, H., Koyama, K., Inoue, H., \& Halpern, J.~P.\ 1991, PASJ,
335:           43, 195
336: \bibitem{}Balbi, A., et al.\ 2000, ApJ, 545, L1
337: \bibitem{}Barger, A.~J., Cowie, L.~L., Mushotzky, R.~F., \& Richards, E.~A.\
338:           2001, AJ, 121, 662
339: \bibitem{}Bassani, L., et al.\ 1999, ApJS, 121, 473
340: \bibitem{}Cappi, M., et al.\ 1999, A\&A, 344, 857
341: \bibitem{}Cappi, M., et al.\ 2000, in ``Broad Band X-ray Spectra of Cosmic
342:           Sources,'' eds.\ K.~Makishima, L.~Piro, \& T.~Takahashi
343:           (Pergamon Press), 815
344: \bibitem{}Comastri, A., Setti, G., Zamorani, G., \& Hasinger, G.\ 1995,
345:           A\&A, 296, 1
346: \bibitem{}Garnavich, P., et al.\ 1998, ApJ, 509, 74
347: \bibitem{}Gendreau, K.~C., et al.\ 1995, PASJ, 47, L5
348: \bibitem{}Gilli, R., Comastri, A., Brunetti, G., \& Setti, G.\ 1999, New
349:           Astronomy, 4, 45
350: \bibitem{}Hanany, S., et al.\ 2000, ApJ, 545, L5
351: \bibitem{}Hornschemeier, A.~E., et al.\ 2001, ApJ, in press
352: \bibitem{}Kennicutt, R.~C.\ 1992, ApJ, 388, 310
353: \bibitem{}Madau, P., Ghisellini, G., \& Fabian, A.~C.\ 1994, MNRAS, 270, L17
354: \bibitem{}Miyaji, T., Hasinger, G., \& Schmidt, M.\ 2000, A\&A, 353, 25
355: \bibitem{}Moran, E.~C., Barth, A.~J., Kay, L.~E., \& Filippenko, A.~V.\
356:           2000, ApJ, 540, L73
357: \bibitem{}Risaliti, G., Maiolino, R., \& Salvati, M.\ 1999, ApJ, 522, 157
358: \bibitem{}Setti, G., \& Woltjer, L.\ 1989, A\&A, 224, L21
359: \bibitem{}Schmidt, M., et al.\ 1998, A\&A, 329, 495
360: \bibitem{}Tully, R.~B.\ 1988, Nearby Galaxies Catalog (Cambridge:
361:           Cambridge Univ.\ Press)
362: \bibitem{}Ulvestad, J.~S., \& Wilson, A.~S.\ 1989, ApJ, 343, 659 (UW89)
363: 
364: \end{thebibliography}
365: 
366: 
367: \begin{center}
368: \begin{deluxetable}{lcccc}
369: \tablewidth{0pt}
370: \tablecaption{The Seyfert 2 Sample}
371: \tablehead{\colhead{Galaxy} &
372:            \colhead{$d$\tablenotemark{a}} &
373:            \colhead{Exp.\tablenotemark{b}} &
374:            \colhead{{~SNR}\tablenotemark{c}} &
375:            \colhead{{CR}\tablenotemark{d}}}
376: \startdata
377: IC  3639\tablenotemark{e}&  43.7  & 36.2  & ~10.2 &  $2.4 \times 10^{-3}$\\
378: MCG $-$05-18-002~~~~~    &  23.1  & 34.6  & ~12.9 &  $1.3 \times 10^{-3}$\\
379: MCG $+$01-27-020         &  46.8  & 41.6  & ~~8.0 &  $2.2 \times 10^{-3}$\\
380: Mrk 3\tablenotemark{e}   &  54.0  & 36.7  & ~44.3 &  $3.2 \times 10^{-2}$\\
381: Mrk 1066                 &  48.1  & 36.2  & ~12.6 &  $4.3 \times 10^{-3}$\\
382: NGC 262\tablenotemark{e} &  60.1  & 45.5  & ~41.4 &  $3.2 \times 10^{-2}$\\
383: NGC 424\tablenotemark{e} &  46.6  & 34.2  & ~20.6 &  $8.7 \times 10^{-3}$\\
384: NGC 591\tablenotemark{e} &  60.7  & 40.9  & ~~7.3 &  $2.8 \times 10^{-3}$\\
385: NGC 788\tablenotemark{e} &  54.4  & 40.2  & ~35.5 &  $2.3 \times 10^{-2}$\\
386: NGC 1068\tablenotemark{e}&  14.4  & 96.6  & 173.6 &  $1.1 \times 10^{-2}$\\
387: NGC 1358                 &  53.8  & 36.2  & ~~9.9 &  $3.6 \times 10^{-3}$\\
388: NGC 1386                 &  16.9  & 37.6  & ~18.2 &  $9.9 \times 10^{-4}$\\
389: NGC 1667                 &  60.7  & 42.3  & ~~4.2 &  $1.5 \times 10^{-3}$\\
390: NGC 1685                 &  60.4  & 34.0  & ~~3.8 &  $1.6 \times 10^{-3}$\\
391: NGC 2273\tablenotemark{e}&  28.4  & 36.9  & ~20.0 &  $2.3 \times 10^{-3}$\\
392: NGC 3081\tablenotemark{e}&  32.5  & 36.7  & ~39.8 &  $8.8 \times 10^{-3}$\\
393: NGC 3281                 &  42.7  & 18.4  & ~19.1 &  $1.1 \times 10^{-2}$\\
394: NGC 3982                 &  27.2  & 40.6  & ~10.3 &  $9.1 \times 10^{-4}$\\
395: NGC 4117                 &  17.0  & 38.0  & ~25.3 &  $1.2 \times 10^{-3}$\\
396: NGC 4388\tablenotemark{e}&  16.8  & 28.1  & ~37.1 &  $3.8 \times 10^{-3}$\\
397: NGC 4507\tablenotemark{e}&  47.2  & 40.4  & 102.9 &  $9.0 \times 10^{-2}$\\
398: NGC 4941                 &  ~6.4  & 36.0  & ~13.9 &  $1.0 \times 10^{-4}$\\
399: NGC 5135                 &  54.9  & 44.0  & ~18.6 &  $9.2 \times 10^{-3}$\\
400: NGC 5347\tablenotemark{e}&  36.7  & 36.7  & ~~6.6 &  $9.8 \times 10^{-4}$\\
401: NGC 5643                 &  16.9  & 41.5  & ~29.6 &  $1.4 \times 10^{-3}$\\
402: NGC 5695                 &  56.4  & 24.6  & ~~4.6 &  $1.9 \times 10^{-3}$\\
403: NGC 5929\tablenotemark{e}&  38.5  & 42.2  & ~22.1 &  $5.8 \times 10^{-3}$\\
404: NGC 6890                 &  31.8  & 34.7  & ~~7.0 &  $5.7 \times 10^{-4}$\\
405: NGC 7672                 &  53.5  & 39.3  & ~~4.7 &  $1.3 \times 10^{-3}$
406: \enddata
407: \tablenotetext{a}{Distances (in Mpc) from Tully (1988).  For the
408:                   more distant sources ($cz > 3000$ km~s$^{-1}$),
409:                   we assume $d = cz/H_0$.}
410: \tablenotetext{b}{Average exposure time (in ks) per detector.}
411: \tablenotetext{c}{Signal-to-noise ratio for the combined 1--10 keV GIS data.}
412: \tablenotetext{d}{Modified 1--10 keV GIS count rate (in ct s$^{-1}$), i.e.,
413:                   that which would be observed if the objects were at 60 Mpc.}
414: \tablenotetext{e}{Has polarized broad optical emission lines.}
415: \end{deluxetable}
416: \end{center}
417:  
418: 
419: \begin{figure}
420: \begin{center}
421: \centerline{\psfig{figure=fig1.ps,width=6truein,angle=270}}
422: \vskip 0.2truein
423: \caption{The composite Seyfert~2 X-ray spectrum, fitted with a double
424: power-law model.  An unabsorbed component (energy index $\alpha = -0.84$)
425: dominates below 3 keV; at higher energies, the spectrum is dominated by
426: a heavily absorbed component ($N_{\rm H} = 3.0 \times 10^{23}$ cm$^{-2}$,
427: $\alpha = -0.46$).  The equivalent width of the Fe K$\alpha$ line at
428: 6.4 keV is 420 eV.  In the 1--10 keV band, the spectrum of the XRB is modeled
429: as an $\alpha = -0.4$ power law.}
430: \end{center}
431: \end{figure}
432: 
433: \begin{figure}
434: \begin{center}
435: \centerline{\psfig{figure=fig2.ps,width=6truein,angle=270}}
436: \vskip 0.2truein
437: \caption{Composite X-ray spectra for the 13 Seyfert~2 galaxies in our
438: sample that have polarized broad optical emission lines, which indicate
439: the presence of a hidden broad-line region (HBLR), and for the 16 objects
440: that do not (non-HBLR).}
441: \end{center}
442: \end{figure}
443: 
444: \begin{figure}
445: \begin{center}
446: \centerline{\psfig{figure=fig3.ps,width=6truein,angle=270}}
447: \vskip 0.2truein
448: \caption{X-ray sky brightness due to Seyfert~2 galaxies ({\it dotted
449: line}), obtained via integration of the composite spectrum over redshift.
450: Note that much of the structure in the composite spectrum has been
451: smoothed out.  The contribution expected from steep-spectrum type~1
452: AGNs ($\sim$~60\% of the XRB at 1 keV; Schmidt \et 1998) is also shown
453: ({\it dashed line}).  The combined emission from type~1 and type~2
454: Seyfert galaxies ({\it solid line}) compares well with the spectrum of
455: the XRB measured by Gendreau \et (1995), especially at $E > 3$ keV.
456: The discrepancy between the XRB spectrum and the Seyfert sum below 2 keV
457: could represent the contributions of groups and clusters of galaxies
458: and star-forming galaxies, which are not included in this simple model.}
459: \end{center}
460: \end{figure}
461: 
462: \end{document}
463: