1: \documentclass[12pt,preprint]{aastex}
2:
3: %\usepackage{color}
4:
5: \shorttitle{Mid--IR emission of Seyfert galaxies.}
6: \shortauthors{Ramos Almeida et al.}
7:
8:
9: \begin{document}
10:
11:
12: \title{The mid--infrared emission of Seyfert galaxies.
13: A new analysis of ISOCAM data\footnote{Based on observations with ISO, an
14: ESA project with instruments funded by ESA Member States
15: (especially the PI countries: France, Germany, the Netherlands and
16: the United Kingdom) and with the participation of ISAS and NASA.}}
17:
18:
19:
20: \author{C. Ramos Almeida\altaffilmark{1}, A.M. P\'{e}rez Garc\'\i a\altaffilmark{1},
21: J.A. Acosta-Pulido\altaffilmark{1}, and J.M. Rodr\'\i guez Espinosa\altaffilmark{1}}
22:
23:
24: \altaffiltext{1}{Instituto de Astrof\'\i sica de Canarias (IAC),
25: C/V\'\i a L\'{a}ctea, s/n, E-38200, La Laguna, Tenerife, Spain.
26: cra@iac.es, apg@iac.es, jap@iac.es, and jre@iac.es}
27:
28:
29:
30:
31: \begin{abstract}
32:
33: We present mid--infrared data of a sample of 57 AGNs obtained with the instrument ISOCAM on board
34: the satellite ISO. The images were obtained through the LW2 (6.75~\micron) and LW7 (9.62~\micron)
35: filters.
36: This is a new analysis of \citet{Clavel00} galaxy sample, which is divided into 26 type 1 ($\le$ 1.5)
37: and 28 type 2 ($>$ 1.5) Seyfert galaxies, plus three QSOs.
38: The spatial resolution of the images allow us to separate the nuclear and the extended
39: contributions to the total emission after decomposing the brightness profiles into different
40: morphological components. The most common components are a central point source (identified
41: as the active nucleus) and an exponential disk. In some cases a bulge, a bar or a ring are needed.
42: The relative contribution of the nucleus to the total emission appears larger in Seyfert 1
43: than in Seyfert 2. This result confirms that both types of Seyfert galaxies are
44: different in the mid-infrared wavelength range and supports
45: the existence of an structure which produces anisotropic emission in this wavelength range.
46: We have also explored correlations between the mid-infrared and the radio and X--ray wavelength ranges.
47: The well established radio/infrared correlation is mantained in our sample for the global
48: emission of the galaxies. If only the nuclear infrared emission is considered
49: then a non--linear correlation is apparent in the luminosity--luminosity scatter diagram.
50: The ratio between the intrinsic hard X--ray and the
51: nuclear mid-infrared emission presents large scatter and slightly larger values for type 2
52: Seyfert galaxies. These results seem to be consistent with the presence of a clumpy
53: dusty torus surrounding the active nucleus.
54:
55: \end{abstract}
56:
57:
58: \keywords{galaxies: active --- galaxies: Seyfert --- galaxies: nuclei --- infrared: galaxies}
59:
60: \section{Introduction}
61:
62: Despite the broad variety of type of objects included under the denomination of Active Galactic
63: Nuclei (AGN), it seems to be possible to explain all of them under a common scenario, the so called
64: Unification Models. The most successful models for the unification of Seyfert galaxies predict the
65: existence of a blocking structure surrounding the nucleus \citep{Antonucci93}.
66: This structure is likely a dusty torus, whose size
67: and structure is still a matter of debate \citep{Fritz06,Honig06}.
68: According to this, type 1 and 2 Seyfert galaxies (hereafter Sy1 and Sy2,
69: respectively) are proposed as the same kind of objects but viewed at different angles:
70: Sy1 are observed close to
71: face-on such that we have a direct view of the nucleus and the Broad emission Line Region (BLR),
72: whereas the Sy2 are seen at an inclination such that our view is blocked by the
73: optically thick dusty torus.
74: The dust grains in the torus will absorb the UV photons from the central engine
75: and after reprocessing the radiation will appear as strong emission in the infrared range.
76: In particular, the mid-infrared emission is produced by a mixture of stochastic heating
77: from Polycyclic Aromatic Hydrocarbon (PAHs) and thermal emission from very small dust grains at
78: high temperatures.
79: Active galaxies are significantly stronger radiators in the mid-infrared than non
80: AGN-dominated galaxies \citep{Spinoglio89,Rowan89, Fadda98, Perez01}.
81: Hence, this spectral range seems a natural window in which to study the properties
82: of such a structure.
83:
84: The mid-infrared spectra of Sy1 and Sy2 galaxies present important differences: whereas Sy1
85: spectra are characterized by a strong continuum with only weak emission features from PAH bands, most
86: Sy2 display a weak continuum but very strong PAH emission bands \citep{Clavel00}.
87: Despite the strong dilution by the nuclear continuum in the case of Sy1, both types 1 and 2 share
88: similar PAH luminosities.
89: This molecular emission results unrelated to the
90: nuclear activity, and arises in the interstellar medium of the underlying galactic bulge.
91: The majority of dusty torus models predict different mid--infrared spectra for type 1 and 2
92: objects. In case of Sy1 the silicate feature at 10~\micron\ is expected in emission, as recently
93: confirmed with instruments on board of the Spitzer satellite \citep{Siebenmorgen05}. In contrast,
94: for Sy2 this feature is observed in absorption \citep{Clavel00,Jaffe04}.
95:
96:
97: In this paper, we analyze mid-infrared images of a sample of Seyfert galaxies.
98: The data sample practically coincides with the previosly studied by
99: \citet{Clavel00}, in which the images are complemented by spectra
100: obtained with the instrument ISOPHOT--S \citep{Lemke96}.
101: Our main goal is to isolate the nuclear emission of the galaxies in the sample,
102: and provide a better estimate of the active component.
103: This new analysis of the data can be used for testing the unification models.
104: In this line we have investigated correlations with photometric data obtained in different
105: spectral ranges, namely radio and hard X--rays.
106:
107:
108: \section{ISOCAM Data}
109:
110: We analyzed the mid-infrared morphology of a sample of AGNs, mostly Seyfert galaxies.
111: These data were taken as part of the ISO Guaranteed Program {\it Seyfert 1 and Seyfert 2} lead by
112: J. Clavel.
113: We have selected this dataset because it was taken in a homogeneous way, with all
114: the exposures obtained in the same mode and keeping the same instrumental configuration
115: (pixel scale and filters).
116: Morphological classification, Seyfert type, spectroscopic redshifts
117: and observational data (exposure and observation time) are reported in Table \ref{info}.
118: The sample was originally drawn from the CfA hard X-ray flux limited complete sample
119: \citep{Piccinotti82}, but lacks the most well known objects (e.g. NGC~4151 or NGC~1068)
120: which were obtained within the frame of other ISO guaranteed time programs.
121: The sample is about equally divided into
122: 26 Sy1 ($\le$ 1.5) and 28 Sy2 ($>$ 1.5), plus a starburst galaxy and three QSOs.
123: Taking into account only the Seyfert galaxies, the mean and {\it rms} of the redshift
124: distributions are 0.023 $\pm$ 0.014 and 0.016 $\pm$ 0.011 for Sy1 and Sy2, respectively.
125: The mean values are not significatively different, considering the width
126: of the distributions, clarifying that the results presented in this paper
127: are not a consequence of differences between Sy1 and Sy2 redshift distributions.
128: In addition to this, we have compiled the total H magnitudes for all Seyferts in the sample from
129: The Two Micron All Sky Survey (2MASS), in order to derive their luminosity distributions.
130: The mean log(L$_{TOT}$[H]) are 44.81 $\pm$ 0.39 and 44.72 $\pm$ 0.42, indicating that
131: Sy1 and Sy2 luminosity, and consequently, mass distributions are very similar.
132: The images were obtained using ISOCAM, the mid-infrared camera \citep{Cesarsky96}
133: on board of the ISO spacecraft \citep{Kessler96}.
134: The ISOCAM images are formed by an array of 32x32 pixels, with a pixel size of 3 arcsec.
135: The employed filters were LW2 (5-8.5 \micron,
136: covering PAHs emission) and LW7 (8.5-10.7 \micron, including the silicate band),
137: which provide an effective resolution on the difraction limit
138: of FWHM=$3\farcs8$ and $4\farcs5$, respectively.
139: Reduced data were retrieved from the ISO data archive\footnote{http://www.iso.vilspa.esa.es/ida/}.
140:
141: \section{Data Analysis}
142:
143: \subsection{Brightness profile decomposition}
144:
145: We performed an isophotal analysis of the images in both filters to obtain the surface
146: brightness profiles, by azimuthally averaging over elliptical annuli. We fitted the isocontours
147: of surface brightness with the ELLIPSE task in IRAF\footnote{IRAF is
148: distributed by the National Optical Astronomy Observatories, which
149: are operated by the Association of Universities for the Research in
150: Astronomy, Inc., under cooperative agreement with the National science
151: Foundation. http://iraf.noao.edu/}, which employes the algorithm described in \citet{Jedrzejewski87}.
152: The spatial interval between two consecutive isophotes chosen was equal to the pixel size in all cases.
153: Given the limited spatial resolution and the relatively small field of view of the camera,
154: the brightness profiles are limited to a maximum of 15 data points,
155: which means a relatively poor sampling of the profile.
156: We subtracted the sky background emission, estimated as the median background value for each image.
157: As a first approach, we assume that the profiles are the sum of two contributions:
158: a nuclear component modeled with a Gaussian PSF of FWHM equals to 2~pixels\footnote{Here we assumed
159: that the nominal spatial resolution cannot be achieved due to insufficient sampling of the
160: PSF. The width of the profile for a point source should correspond to the minimum sampling criterium,
161: {\it i.e.} 2 pixels or equivalently 6\arcsec.},
162: and central amplitude as free parameter plus an exponential component with
163: two free parameters (central amplitude and length scale). Using this simple model we
164: were able to reproduce correctly the brightness profiles
165: for about half of the galaxies in our sample. In the rest of cases we could not
166: succesfully reproduce the brightness profiles,and extra components had to be introduced
167: in order to improve the fit. The extra components are a bulge described as a Sersic's
168: law; a ring described by a Gaussian profile \citep{Buta96}:
169: $$ I_{ring}(r) = I_{ring}^0 exp[-{1\over2}({r-r_{ring} \over l_{ring}})^{2}] ,$$
170: where I$_{ring}^0$ is the amplitude, r$_{ring}$ is the center, and l$_{ring}$ is the width; or a flat
171: bar profile \citep{Prieto97}:
172: $$ I_{bar}(r) = {I_{bar}^0 \over 1+exp[(r-r_{bar})/l_{bar}]} , $$
173: where I$_{bar}^0$ is the amplitude, r$_{bar}$ is the lenght, and l$_{bar}$ is the downward gradient.
174:
175: In all cases, the fits were performed with the criterium of minimum number of components and
176: consistency between both filters.
177: %We also took special care with the unresolved component, in order
178: %to characterize as well as possible the nuclear emission of each galaxy.
179: Whenever bars or rings were employed in the fitting, we also looked at the ellipticity
180: and orientation of the isophotes searching for abrupt changes coincident with the
181: position of these components. However, the limited resolution
182: of our imaging prevents a fully reliable component identification.
183: In addition to this, we used as a guideline the morphology of the galaxies in the visible (WFPC2/HST)
184: from \citet{Malkan98}
185: and in the near-infrared (NICMOS/HST) from \citet{Hunt04}, although in
186: some cases we fit nuclear bars that do not agree with the optical classification, or rings that
187: reproduce spiral arms, as in the case of NGC~1241. The maximum number of fitted parameters in our fits
188: was seven.
189: The fitting procedure looks for minimization of the Chi-Square function formed by the squared
190: difference between the observed and model profile, both in logarithmic scale.
191: The multidimensional minimization of the merit function is reached using the downhill simplex method.
192: This is the same fitting technique employed in \citet{Melo02} in the study of NGC~253.
193: In Table \ref{params} we report the fit parameters and their corresponding errors for 54 objects of the sample,
194: in both LW2 and LW7 filters. Errors were estimated using a bootstrapping
195: technique. This technique is based on a Monte Carlo simulation: new brightness profiles are obtained
196: by perturbation of the measured profile, using a normal distribution of the same width as the
197: error of the measured point. New fit parameters are determined for each one of the simulated profiles
198: (30 simulations) and the uncertainty of each parameter is computed as the standard deviation
199: of the resulting values.
200: We calculated the fluxes integrating
201: all the emission contained in each fitted component and in the total fit, for both LW2 and LW7 filters.
202: As an additional check point, we verified that the
203: resultant flux from the total fit was at least $\sim 95$\% of the total measured flux.
204: All fluxes and its errors are reported in Table \ref{fluxes}.
205: We found that bulges are only needed in the case of three galaxies: Mrk~3, MGC6-30-15,
206: and Mrk~841, which are morphologically classified as S0 or
207: elliptical (see Table \ref{info}).
208: 27 galaxies were fitted using a PSF plus an exponential component, 14 with PSF plus exponential component
209: plus bar, 4 with exponential component plus bar, 3 with PSF plus exponential component plus ring, 2
210: with PSF plus bar plus ring, and one with exponential component
211: only. The galaxies NGC~7592 and ESO137-G34
212: can not be fitted because of their double nucleus, NGC~5929 because of the proximity of its companion
213: (the starburst galaxy NGC~5930), and the QSO HS1700+6416 because of their high $z$=2.736.
214: For these four galaxies without profile decomposition, total fluxes were
215: obtained by means of aperture photometry, using the PHOT task, within the IRAF environment.
216:
217: We show in Fig.~\ref{ic4329a} examples of
218: brightness profile decomposition, fitting different morphological components, in both LW2 (left panels) and LW7
219: (right panels) filters. The origin of X-axis corresponds to the center of fitted isophotes,
220: and distance from this origin is given in pixels, being the pixel size of 3\arcsec.
221:
222: \clearpage
223:
224: \begin{figure*}
225: \centering
226: \includegraphics[width=18cm,height=20cm]{f1_color.eps}
227: \figcaption{\scriptsize{Brightness profiles of the Sy1 galaxy IC~4329 in both
228: LW2 and LW7 filters, fitted by an exponential
229: component (dashed line) plus a Gaussian PSF (dot-dashed line). Filled line is the final fit.
230: The same for the Sy2 galaxy NGC~4507, fitted by an
231: exponential component (dashed line) plus a Gaussian PSF (dot-dashed line) and a flat bar (dotted line);
232: for the Sy1 galaxy NGC~1097, fitted by an
233: exponential component (dashed line) plus a Gaussian PSF (dot-dashed line) and a nuclear ring
234: (double-dot-dashed line); and finally for the Sy2 galaxy NGC~5033, fitted by a
235: Gaussian PSF (dot-dashed line) plus a flat bar (dotted line) and a ring (double-dot-dashed line).
236: Figures 1.1 - 1.50 are available in the electronic edition of the journal.}
237: \label{ic4329a}}
238: \end{figure*}
239:
240: %\clearpage
241:
242: %\begin{figure*}
243: %\centering
244: %\includegraphics[width=9cm,angle=90]{f2_color.eps}
245: %\figcaption{\scriptsize{Same as in Fig.~\ref{ic4329a}, but for the Sy2 galaxy NGC~4507, fitted by an
246: %exponential component (dashed line) plus a Gaussian PSF (dot-dashed line) and a flat bar (dotted line). }
247: %\label{ngc4507}}
248: %\end{figure*}
249:
250:
251: %\begin{figure*}
252: %\centering
253: %\includegraphics[width=9cm,angle=90]{f3_color.eps}
254: %\figcaption{\scriptsize{Same as in Fig.~\ref{ic4329a}, but for the Sy1 galaxy NGC~1097, fitted by an
255: %exponential component (dashed line) plus a Gaussian PSF (dot-dashed line) and a nuclear ring
256: %(double-dot-dashed line).}
257: %\label{ngc1097}}
258: %\end{figure*}
259:
260:
261: %\begin{figure*}
262: %\centering
263: %\includegraphics[width=9cm,angle=90]{f4_color.eps}
264: %\figcaption{\scriptsize{Same as in Fig.~\ref{ic4329a}, but for the Sy2 galaxy NGC~5033, fitted by a
265: %Gaussian PSF (dot-dashed line) plus a flat bar (dotted line) and a ring (double-dot-dashed line).}
266: %\label{ngc5033}}
267: %\end{figure*}
268:
269:
270: \subsection{Comparison with previous studies}
271:
272: The main interest of our work is focused into a reliable
273: measurement of the nuclear flux, as well as its relative value to the total flux
274: for each galaxy.
275: Once we have obtained the fluxes for each morphological component,
276: reported in Table \ref{fluxes}, the first task we accomplished was
277: the comparison of our results with the previous ones obtained from the same
278: data sample and reported by \citet{Clavel00} (hereafter C00).
279: Here we recall that in C00, nuclear fluxes were obtained by integrating within a
280: circular aperture of 3 pixel radius (9\arcsec) for point--like sources,
281: and within an aperture of 4.5 pixels (13.\arcsec5) for extended ones.
282: In order to account for the
283: emission in the wings of the PSF they also corrected by a factor 1.23.
284: We note that this photometric correction makes sense for point sources,
285: although their effect is not clear on extended sources like most of our
286: targets are. We present in Fig.~\ref{clavel1} the
287: difference between our total values and C00 measurements, normalized to our total
288: fluxes. It can be seen that
289: for most of the galaxies, our total fluxes are underestimated about 10\% compared
290: with C00 ones. This might be due to the introduction of the correction for
291: the PSF wings in C00 fluxes and not in our values, which is compensated by the fact that
292: our measurements extended all over the detector, as opposite to C00, which uses a limited aperture.
293: %Fluxes underestimated over 10\% correspond to brightness profiles with few points,
294: %as is the case of QSOs.
295: On the contrary, for large galaxies, as for example NGC~5033, NGC~5674, NGC~1241,
296: and NGC~3982, our flux measurements are much higher than those reported by C00.
297: This discrepancy is due to the use of a 4.5 pixels aperture by C00 in clearly
298: extended sources, resulting in underestimations of $\sim 60$\%.
299: We also present in Fig.~\ref{clavel1} the comparison between our nuclear fluxes
300: and the C00 values, normalized to C00 values.
301: In this case all our nuclear fluxes are lower than C00 ones,
302: since we have subtracted the contribution of the galaxy to the nuclear flux,
303: and moreover the
304: radius of our Gaussian PSF is smaller than 3 pixels (the aperture employed by COO for point-like sources).
305: Note that for Sy1 galaxies our values are in better agreement with C00,
306: due to the fact that the compact nuclear component is dominant relative to the
307: total galaxy emission.
308: Summarizing, we believe that the differences found between our flux measurements
309: and those reported by C00 justify the need of
310: a brightness profile decomposition in order to get a more accurate
311: determination of the nuclear flux.
312:
313: %\clearpage
314:
315: \begin{figure}[!h]
316: \centering
317: {\par
318: \includegraphics[width=5.5cm,angle=90]{f2a_color.eps}
319: \includegraphics[width=5.5cm,angle=90]{f2b_color.eps}\par}
320: \figcaption{\footnotesize{Left panel: comparison of our total fluxes (see Table \ref{fluxes})
321: with the reported ones in \citet{Clavel00}. Filled circles represent Sy1
322: (25 in total), open diamonds Sy2 (24 in total), open
323: triangles QSOs (2 in total), and the starburst galaxy, NGC~701, is represented by
324: a cross. NGC~1144 and Mrk~789 fluxes were not reported by C00. Right panel: the same comparison, but for
325: our nuclear fluxes.}
326: \label{clavel1}}
327: \end{figure}
328:
329: %\clearpage
330:
331: %\begin{figure}[!h]
332: %\centering
333: %\includegraphics[width=8cm,angle=90]{f6_color.eps}
334: %\figcaption{\footnotesize{Same as Fig.~\ref{clavel1},
335: % but for our nuclear fluxes (see Table \ref{fluxes}).}
336: %\label{clavel2}}
337: %\end{figure}
338:
339: \section{Results}
340:
341:
342: \subsection{Nuclear vs total emission}
343: \label{nuclear}
344:
345:
346: Once we have performed the profile decomposition we can measure
347: the relative contribution of the nuclear source to the total emission for 54 objects of our sample.
348: Note that for 7 galaxies our profile decomposition does not include a nuclear
349: component (see Table \ref{params}), which includes 6 Sy2, namely NGC~1144, Mrk~3,
350: NGC~1667, NGC~5728, and NGC~5953, plus
351: the Sy1 Mrk~841 (it is an elliptical galaxy and the central component is better fitted
352: with a bulge solely).
353: We show in Fig.~\ref{hist}
354: the ratio of nuclear to total emission at 6.75 and 9.62 $\mu$m, respectively.
355: It can be seen in both cases that there is a clear difference between the ratios
356: for the Seyfert types 1 and 2. In fact, the
357: median of the ratio between nuclear and total fluxes for type 1 galaxies are 0.61 and 0.49,
358: in the 6.75 and 9.62 $\mu$m filters, respectively. These values are both 0.14 for type
359: 2 galaxies.
360: %For the QSOs in our sample the nuclear emission becomes dominant, in contrast with the
361: %starburst galaxy, NGC~701, whose nuclear emission is quite small compared with the total flux.
362: %%\clearpage
363: %We have also represented the ratio of nuclear to total flux according to the Seyfert type and
364: %filter used but in the form of histograms in Fig.~\ref{hist}.
365: In case of type 2 objects, most of them are grouped around the lowest values
366: of the nuclear vs total flux ratio, whereas for type 1, the maximum number of
367: objects are located around or above the value 0.5. We have applied the Kolmogorov--Smirnov
368: test to check the significance level probability of the apparent difference between Sy1 and Sy2
369: distributions. We found that in both wavelenght ranges, the nuclear to total flux distributions
370: for Sy1 and Sy2 are different in a 99.9\%.
371: %In the case of the LW2 filter, we obtain a normalized probability value of 3.66x10$^{-6}$, and
372: %2.3x10$^{-4}$ for the distributions in the LW7 filter. These low values indicate that nuclear to total flux
373: %distributions for Sy1 and Sy2 in both filters are statistically different.
374:
375:
376: We conclude that nuclear emission in the mid-infrared is a significative contribution of the total
377: flux in Sy1 galaxies, whereas for Sy2 other components overcome the
378: nuclear emission. This result is consistent with the unification model predictions, since
379: the orientation of the molecular torus for type 1 Seyfert would be face-on with respect
380: to our line of sight whereas for type 2 would be edge--on.
381: Similar results have been found previously at different spectral ranges \citep{Yee83,Alonso96}.
382:
383: %\begin{figure}[!h]
384: %\centering
385: %\includegraphics[width=8cm,angle=90]{f7_color.eps}
386: %\figcaption{\footnotesize{Ratio between nuclear and total emission versus total emission at
387: %6.75 $\mu$m. Filled circles represent Sy1 (27 in total), open diamonds Sy2 (24
388: %in total), the two QSOs are represented by open triangles, and the starburst galaxy, NGC~701, by a cross.}
389: %\label{psf1}}
390: %\end{figure}
391: %\begin{figure}[!h]
392: %\centering
393: %\includegraphics[width=8cm,angle=90]{f8_color.eps}
394: %\figcaption{\footnotesize{Same as Fig.~\ref{psf1}, but for fluxes measured at 9.62 $\mu$m.}
395: %\label{psf2}}
396: %\end{figure}
397:
398: %\clearpage
399:
400: \begin{figure}[!h]
401: \centering
402: \includegraphics[width=9cm,angle=90]{f3_color.eps}
403: \figcaption{\footnotesize{Histograms of nuclear vs total emission for Sy1 (left panels) and Sy2
404: galaxies (right panels) at 6.75 $\mu$m (top plots) and 9.62 $\mu$m (bottom plots). Note that
405: for Sy2 galaxies, most of them are concentrated around the lowest values of this ratio,
406: what does not happen in the case of Sy1.}
407: \label{hist}}
408: \end{figure}
409:
410: %\clearpage
411:
412: \subsection{Nuclear and host galaxy mid-infrared colors}
413:
414: We have studied the mid-infrared color distributions for both the nuclear and
415: the host galaxy emission. The colors of host galaxies are computed after subtracting the
416: nuclear to the global emission.
417: We represent in Fig.~\ref{color1} the 9.62/6.75~\micron\ color histograms
418: for both the nuclear and the host galaxy emission, for each Seyfert type.
419: The nuclear 9.62/6.75~\micron\ color distributions look very similar for both types,
420: being the median values 1.14 and 1.02 for Sy1 and Sy2, respectively.
421: We have performed a Kolmogorov-Smirnov test, obtaining a probability of 18\%, which
422: indicates that both color distributions are not statistically different.
423: Previous works \citep{Spinoglio89,Fadda98,Kuraszkiewicz03,
424: Alonso03,Lutz04,Rigby04} have found similar
425: observed spectral energy distributions (SEDs) in the mid--infrared ($\lambda$ $>$ 5 \micron) for Sy1
426: and Sy2 galaxies,
427: These results are in conflict with the predictions of many compact torus models \citep{Pier92,Granato94,
428: Efstathiou95,Granato97}, that has promoted the search of a more distributed or complex geometry
429: of the absorbing material around the AGN \citep{Nenkova02,Fritz06,Elitzur06,Ballantyne06}.
430:
431:
432: %\citet{Fritz06} predict no differences for type 1 and 2 in the region around 10~\micron\ when models
433: %with low optical depths and small outer-to-inner radii are considered.
434: Despite the small number of galaxies in our sample and the proximity of
435: the bands analised, we can conclude that the nuclear mid-infrared emission
436: seems to be very similar and relatively flat in both types of Seyfert galaxies.
437: %In addition, we have also checked that the colors of the QSOs are in the middle of the
438: %range of nuclear colors for Seyfert galaxies, whereas the color of
439: %the starburst NGC701 is among the bluest of the sample.
440:
441:
442:
443: On the other hand, the color distributions of host galaxies look different, being those
444: of type 1 redder than those of type 2 (median values of 2.00 and 1.41, respectively).
445: The Kolmogorov-Smirnov test confirms that both distributions are different with a probability larger than 99\%.
446: This result seems to contradict the core idea of Seyfert unification scheme, which
447: predicts no much influence of the host galaxy to determine the type of active nucleus.
448: In the same line of our result, \citet{Hunt99} pointed out that Sy1 could be older, more evolved
449: than Sy2, since they are found more commonly in earlier morphological types.
450:
451: %\clearpage
452:
453: \begin{figure}[!h]
454: \centering
455: \includegraphics[width=9cm,angle=90]{f4_color.eps}
456: \figcaption{\footnotesize{9.62/6.75~\micron\ colors for nuclear fluxes (top panels) and for
457: host galaxy fluxes (bottom). Left histograms represent Sy1 and right ones, Sy2.}
458: \label{color1}}
459: \end{figure}
460:
461: %\clearpage
462:
463: %\begin{figure}[!h]
464: %\centering
465: %\includegraphics[width=8cm,angle=90]{f11_color.eps}
466: %\figcaption{\footnotesize{Same as Fig.~\ref{color1}, but for nuclear fluxes.}
467: %\label{color2}}
468: %\end{figure}
469:
470:
471: %according with
472: %the results obtained by \citet{Fritz06} for a modelling of the emission by dust in a torus geometry,
473: %a flared disk, and a dust grain distribution function including a full range of grain sizes. This model
474: %predicts redder 9.62/6.75~\micron\ colors for Sy1 than for Sy2, specially for the lowest equatorial optical
475: %depths and outer-to-inner radii ratios.
476:
477:
478:
479: \subsection{The radio/mid--infrared correlation in our sample}
480: \label{radio}
481:
482: A well-known correlation between global far--infrared and radio emission from galaxies
483: apply to a wide range of Hubble types \citep{Fitt88,Hummel88,Wunderlich87}.
484: %This correlation is displayed over 3 orders
485: %of magnitude in luminosity, from dust--rich dwarfs to ultraluminous infrared galaxies (ULIRGs).
486: The most natural explanation for such correlation is related to
487: star--formation activity.
488: In addition, \citet{Elbaz02,Gruppioni03} found that it
489: can be also extended to the mid--infrared range although with larger dispersion.
490: %than in the case of the radio/far-infrared.
491: %A priori, there is not obvious reason why such correlation should exist. The mid-infrared emission
492: %is mainly produced by
493: %a mixture of complex PAH plus thermal emission from small dust grains at
494: %high temperatures, whereas the far--infrared emission is dominated by large dust grains at lower
495: %temperatures.
496: %More recently, \citet{Appleton04} presented evidences that the radio/far- and mid--infrared
497: %correlation is valid to cosmologically significant distances (up to $z\simeq1$).
498: Thus, the study of the radio/far--infrared correlation in active galactic nuclei turns out to be
499: an useful tool for studying the starburst--AGN relationship.
500: A priori, the presence of nuclear radio activity not related to supernova remnants
501: should introduce departures from the mentioned correlation.
502: \citet{Roy98} reported that Seyfert galaxies
503: also display the radio/far-infrared correlation, although with a larger scatter than
504: non-active galaxies. They also found that Seyfert with compact radio cores tend to
505: deviate from the correlation, contrary to those without compact cores. However, they
506: noted that the correlation does not improve after subtraction of the compact radio emission.
507:
508:
509:
510: We have investigated the presence of the radio/mid-infrared correlation in our
511: sample of Seyfert galaxies. The radio data were taken from those available in
512: the literature. Most of them (data of 32 galaxies) come
513: from the sample observed by \citet{Rush96} at 20 cm using the VLA\footnote{For another 6 objects
514: (Mrk~789, NGC~4579, ESO141-G55, NGC~4593, Mrk~509, and NGC~701)
515: we have used VLA measurements at 20~cm, reported by different authors
516: \citep{Ulvestad84,Condon98,Ho01,Wadadekar04}.}.
517: The resolution ($\sim$ 1.\arcmin5) of these radio data samples the emission of large scales
518: in the galaxy, namely $\sim$ 35~kpc at the median distance of the sample.
519: %These structures should be similar to those reported by \citet{Baum93}.
520: From here and in all later sections, we will present results only using the LW2 filter,
521: but we have always checked that the results obtained
522: with both filters are practically the same.
523: % We present in Fig.~\ref{flujo_total} the radio and 6.75~\micron\ total flux--flux scatter diagram.
524: We have excluded two galaxies from subsequent analysis because their behaviour deviates from
525: the rest of the sample. They are identified as NGC~4593 (its radio flux is a lower limit)
526: and Mrk~335 (a S0 galaxy with a low radio emission likely due to be an early type galaxy).
527: %As can be seen there is an apparent correlation between the 20~cm and 6.75~\micron\ data sets.
528: The luminosity-luminosity scatter diagram is presented in Fig. \ref{rush1}. A linear correlation
529: appears in log-log scale, with a slope of 1.07 and a coefficient $r= 0.88$\footnote{In order
530: to verify that the correlation is not a distance effect, we have performed a test
531: by upsetting in a random way the redshifts ($z$) of galaxies. If the linear correlation
532: is merely a distance effect, it should be mantained with any $z$ distribution. In our case,
533: the correlation disappears when the distance distribution is changed, confirming that
534: the observed correlation is not a distance effect.}.
535: Similar correlations are found when the data are fitted for types 1 and 2, separately
536: (see Fig. \ref{rush1}). The highest correlation coefficient ($r=0.95$) appears when only
537: Sy2 are considered.
538: In addition, we have applied a non--parametric correlation
539: test, such as the Spearman's rank correlation test. The results indicate that
540: the correlation is significant ($p <0.01$) in all cases (all galaxies, and Sy1 and Sy2,
541: separately).
542:
543: %\clearpage
544:
545: \begin{figure}[!h]
546: \centering
547: \includegraphics[width=8cm,angle=90]{f5_color.eps}
548: \figcaption{\footnotesize{Radio luminosity (L$_{20cm}$) versus mid-infrared global luminosity
549: (L$_{TOT}$~6.75~$\mu$m). Filled circles represent Seyfert 1 (19 in total),
550: open diamonds Seyfert 2 (18 in total) and the cross the starburst galaxy, NGC~701.
551: The continuous black line represents a linear fit to all data,
552: the dashed lines correspond to the limits of the correlation at one $\sigma$.
553: The values of $\alpha$ and k are derived from the expression
554: $L_{20cm}/10^{28}=k\cdot(L_{TOT}/10^{29})^{\alpha}$ and are given explicitely in the
555: legend. Separate fits to types 1 and 2 data are also presented.}
556: \label{rush1}}
557: \end{figure}
558:
559: %\clearpage
560:
561: We have also explored the existence of such a correlation between the same radio data and
562: our estimations of the nuclear mid-infrared emission. The luminosity scatter diagram
563: is presented in Fig. \ref{rush2}.
564: In this case, the correlation appears worse ($r \simeq 0.68$) than when global
565: mid-infrared luminosity is considered. Moreover, the correlation becomes distant
566: from linear in log-log space, being the slopes different from unity (see values
567: in Fig. \ref{rush2}). We also noticed that the correlation slightly improves when
568: only Sy2 galaxies are considered.
569: The less luminous galaxies (mostly Sy2) display an excess of radio luminosity compared
570: to their nuclear mid-IR luminosity, indicating that most of the radio emission
571: in these objects is not related to nuclear processes.
572: Summarizing, the global radio emission seems to be
573: related closely to AGN activity in the most nuclear-infrared luminous galaxies (mostly Sy1),
574: whereas for the less luminous the radio emission would be more related to non-nuclear
575: stellar processes. In this respect it is worth to mention the result of
576: \citet{Baum93}, who found that extra-nuclear (several kpc) radio emission, similar
577: to the lobes of powerful radio galaxies, appears frequently in galaxies whose properties
578: are dominated either by an AGN or a starburst.
579:
580: %\clearpage
581:
582: \begin{figure}[!ht]
583: \centering
584: \includegraphics[width=8cm,angle=90]{f6_color.eps}
585: \figcaption{\footnotesize{Radio luminosity (L$_{20cm}$) versus mid-infrared nuclear
586: luminosity (L$_{NUC}$~6.75~$\mu$m). The meaning of symbols is the same as in
587: Fig.~\ref{rush1}. The continuous line represents the fit to all galaxies,
588: the dashed line represents the correlation using global mid-IR luminosity,
589: from Fig.~\ref{rush1}.
590: The corresponding fits to Sy1 and Sy2 data separately, are also represented.}
591: \label{rush2}}
592: \end{figure}
593:
594: %\clearpage
595:
596: According to our results, we claim that for the less radio luminous galaxies (mostly Sy2),
597: radio emission is more related to stellar processes, given the
598: fact that most of the mid-infrared emission does not come from the nuclear region.
599: However, for the more luminous galaxies (mostly Sy1),the correlation could be
600: attributed to extended radio emission which is somewhat related to
601: the presence of the AGN as claimed by \citet{Baum93} and indirectly by \citet{Roy98}.
602:
603:
604:
605: %The data corresponding to all considered galaxies displays a linear correlation of
606: %slope 0.8 and a correlation coefficient $r=0.7$. If we perform the analysis separating by Seyfert types,
607: %the fit slopes are 0.81 and 0.83 for Sy1 and Sy2, respectively, and
608: %the correlation coefficients are 0.76 and 0.70 for those types.
609: %Apparently, the correlation is marginally better for type 2 Seyfert galaxies in the flux--flux
610: %Then, the correlation is marginally better for type 2 Seyfert galaxies in the flux--flux
611: %space.
612: %when all galaxies and only type 2 are considered, whereas it does not appear as
613: %significant when only type 1 galaxies are taken into account ($p=0.08$).
614: %In the luminosity--luminosity scatter diagram (see Fig.~\ref{rush1}), founding a linear
615: %radio/mid-infrared correlation,
616: %, showing an improvement with respect to the flux--flux case
617: %\footnote{In order to verify that the
618: %previous correlation exists for luminosities, and it is not a distance effect
619: %(Seyferts in our sample are distributed in a redshift range of [0, 0.06], what can stretch the points
620: % %distribution of Fig.~\ref{rush1}), we have perform a test by upsetting in a random way the redshifts ($z$) of galaxies. If
621: % the lineal correlation is only a distance effect, points should mantain it when we modify their $z$. Fortunately,
622: % correlation disapears when we perform the test, so we can conclude that luminosities fit is not an effect of $z$
623: % distribution.}.
624: %, whereas for type 1 the
625: %value of $r=0.81$ gives also confidence about the presence of such correlation.
626: %We have also done the rank correlation test using the luminosity data, finding
627: %that in all cases the probability for correlation is very significant.
628: % although
629: %type 1 galaxies present the less significant case.
630:
631: % \begin{figure}[!h]
632: % \centering
633: % \includegraphics[width=8cm,angle=90]{flujo_total.ps}
634: % \figcaption{\footnotesize{Radio flux (F$_{20cm}$) versus total mid-infrared flux (F$_{TOT}$~6.75~$\mu$m),
635: % displaying a quasi-linear
636: % correlation (continuous black line) of $\alpha$=0.8 (excluding Mrk~335 and NGC~4593 from the fit). $\alpha$ and k
637: % are derived from the expression $F_{20cm}/10^{28}=k\cdot(F_{TOT}/10^{29})^{\alpha}$.
638: % Dashed lines represent the limits of the correlation at one $\sigma$.
639: % Filled circles represent Sy1 (18 in total), open diamonds Sy2 (19 in total) and the cross corresponds to the
640: % starburst galaxy, NGC~701. Fits to Sy1 and Sy2 data are represented separately too.}
641: % \label{flujo_total}}
642: % \end{figure}
643:
644: % We represent in Figs.~\ref{radionucflux} and
645: % \ref{rush2} the radio versus mid-infrared flux--flux and luminosity--luminosity scatter diagrams.
646: % As can be seen in Fig.~\ref{radionucflux}, the correlation between fluxes does not exist, being
647: %In fact, the correlation between fluxes does not exist.
648: %When we look at the luminosities scatter diagram (see Fig \ref{rush2}) the
649: %correlation
650: %coefficient rise to a value near 0.7. However, the correlation becomes distant from linear in
651: %log-log space, being the slopes different from unity (see values in Fig. \ref{rush2}).
652: %$\alpha=0.53$ for all galaxies,
653: %and $\alpha=0.4, 0.7$ for types 1 and 2, respectively.
654: %The previous results have been also confirmed
655: %using the non-parametric Spearman's rank correlation test, being all correlations in luminosities
656: %significative.
657:
658:
659: %We believe that the non-existence of a correlation for nuclear fluxes, and that a
660: %loose correlation appears when luminosities are considered, is not caused by a distance effect.
661: %As explained by \citet{Feigelson83}, if the intrinsic correlation
662: %between $L_{20cm}$ and $L_{NUC}$ 6.75~\micron\ is not linear ($\alpha \neq 1$),
663: %then a random redshift distribution will smear out any flux-flux correlation.
664: %In Fig.~\ref{rush2} we have also overplotted the observed correlation using the total mid-infrared emission.
665: %Sy1 galaxies are located within the dashed lines indicating the dispersion
666: %around the correlation (one $\sigma$). This fact is expected since the
667: %nuclear to total flux/luminosity ratio is higher for Sy1 galaxies. However, the less
668: %luminous Sy2 objects departure considerably from this region, indicating an excess of radio
669: %luminosity compared to the mid-infrared one. This fact suggests that in those objects most of the radio
670: %emission is not related to nuclear processes.
671: %On the contrary, those galaxies showing the largest
672: %radio luminosity tend to return to the correlation observed in total emission (see Fig.~\ref{rush2}).
673:
674: %A similar result was found by \citet{Roy98}, who noted that Seyferts with compact radio cores depart from
675: %the normal correlation between far--infrared and radio emission, being more radio--loud than non-active
676: %spirals. Moreover, they also found that the correlation does not improve when the compact emission is
677: %subtracted, which seems to indicate that the origin of the kpc scale radio emission is
678: %related directly or indirectly to AGN activity.
679: %This can also be related with the extra-nuclear radio emission (aligned with the galaxy minor axis) appearing
680: %in objects whose properties are dominated either by a starburst or by an AGN \citep{Baum93}.
681:
682:
683: %In principle, non-stellar AGNs should not display such correlation
684: %mainly due to the presence of nuclear radio activity, likely not related to
685: %supernova remnants. Based on a small sample of Seyfert galaxies,
686: %\citet{Baum93} found that these objects obey the normal correlation after
687: %subtracting the radio emission from only the central kpc.
688: %The most noticeable result from that study is the presence of radio emission
689: %on scales up to $\sim 5~\mathrm{kpc}$ similar to the lobes of powerful radio galaxies,
690: %aligned with the minor axis.
691: %On the other hand, \citet{Roy98} report that Seyfert galaxies display the radio/far-infrared
692: %correlation, although with a larger scatter than non-active galaxies. In particular,
693: %they found that Seyferts without compact radio-core follow the correlation, whereas those who
694: %shelter it, they move apart.
695:
696: %When we consider only the nuclear mid-infrared emission for the Seyfert sample, in the most luminous
697: %objects the radio emission seems to be related closely to the nuclear mid-infrared source, i.e. the AGN
698: %activity, whereas for the less luminous the radio emission is more related to stellar processes.
699:
700: % \begin{figure}[!h]
701: % \centering
702: % \includegraphics[width=8cm,angle=90]{flujo_psf.ps}
703: % \figcaption{\footnotesize{F$_{20cm}$ versus nuclear F$_{NUC}$~6.75~$\mu$m.
704: % Continuous line in black is the total fit and dashed line is the quasi-linear correlation from Fig.~\ref{flujo_total}.
705: % Filled circles represent Sy1 (17 in total), open diamonds Sy2 (15 in total) and the cross corresponds to the
706: % starburst galaxy, NGC~701. Fits to Sy1 and Sy2 data are represented separately too.}
707: % \label{radionucflux}}
708: % \end{figure}
709:
710:
711:
712: \subsection{Comparison among X-rays and mid-infrared emission}
713:
714:
715: The hard X--ray (2-10 keV) spectral region is of particular interest for
716: the study of AGN. It provides a direct view to the central engine and it is believed to
717: be a reasonable isotropic indicator of the bolometric luminosity of the
718: nuclei. On contrast, in the simple unified models, the mid-infrared emission is expected to vary
719: as a function, not only of the AGN luminosity, but also of the distribution of
720: the obscuring matter along the viewing direction of the observer.
721: As a result of this, the comparison of the mid-infrared versus
722: the hard X--ray measurements constitutes an important test for unification models.
723: % Previously, a tight correlation between the two quantities had been reported by \citet{Krabbe01},
724: % using a very reduced sample of only 8 Seyfert galaxies.
725: % Later, \citet{Lutz04} found a similar correlation for a larger sample of 71 Seyferts.
726: % However, they found considerable scatter in the ratio between intrinsic (absorption corrected)
727: % X--ray and mid-infrared emissions, being higher for type 2 Seyferts. They ascribe part of this dispersion
728: % to variations in the AGN spectral energy distribution and geometry of obscuring dust, as well as
729: % time variability.
730: % More surprisingly, \citet{Lutz04} reported no significant difference between type 1 and 2 Seyferts
731: % in the average ratio of X-ray and mid-infrared continuum. They explained this fact as a consequence of
732: % a large isotropic contribution from extended dust components which largely dilutes
733: % the anisotropic emission from the torus.
734: % In a recent work, \citet{Horst06} claimed a tight correlation between
735: % nuclear mid-infrared and intrinsic hard X--ray luminosities, using
736: % very high spatial resolution mid-infrared images from VISIR on the VLT. They found the best fit power law index
737: % different of unity ($log L_{MIR} \propto 1.6~log L_X$).
738:
739: %Smooth torus models predict a flux difference between type 1 and 2 nuclei of
740: %an order of magnitude. However, clumpy torus models \citep{Nenkova02, Honig06}
741: %are compatible with {\bf no flux difference in the mid-infrared}. Clumpy torus can appear
742: %as optically thin in the mid-infrared, although individual clouds appear optically thick.
743:
744: Here, we compare our mid-infrared measurements with the hard X-ray fluxes compiled by
745: \citet{Lutz04}, obtained from various literature sources, taken by different satellites
746: (e.g. ASCA, BeppoSAX, Chandra, XMM-Newton).
747: They reported measurements of 35 objects which are also included in our sample,
748: from which only 27 galaxies have intrinsic or absorption--corrected values.
749: %X-ray emission is significantly more absorbed in Sy2 than in Sy1, then, for Sy2 for which
750: %there are absorbing column density N$_{H}$ measurements,
751: %\citet{Lutz04} report the intrinsec absorption-corrected hard X-ray fluxes (unfortunately,
752: %there are only 27 galaxies in our sample with intrinsec fluxes measurements).
753: %We have compared these data with our nuclear and total MIR fluxes.
754: %, in order to verify the conclusions found by \citet{Lutz04}, who calculated their 6\micron fluxes
755: %by integration of continuum from ISOPHOT spectra.
756: We represent in Fig. \ref{lutz_hist} the histograms corresponding to
757: observed hard X--ray fluxes for Sy1 and Sy2, plus absorption corrected hard X--ray fluxes for Sy2.
758: %We have then performed the analysis of this section based only in the filter centered at 6.75~\micron,
759: %after having verified that there are not significative differences between both ranges.
760: % There is a clear segregation between the two types
761: % of Seyfert galaxies, being type 2 nuclei less luminous in X-rays (and, as we have seen before, in
762: % mid-infrared).
763: Given the fact that the hard X--ray emission in Sy2 galaxies is commonly strongly absorbed,
764: it makes sense to consider the intrinsic, instead than the observed, hard X--ray fluxes.
765: For those cases in which absorption corrected fluxes are not provided we adopt the observed
766: flux as a lower limit.
767: %Sy2 nuclei show similar hard X--ray intrinsic fluxes,
768: % and generally lower than those of type 1.
769: %The mid-infrared fluxes of Sy2 nuclei are
770: %much lower than those of type 1.
771: In this section we do not refer to the total infrared flux since hard X--ray emission is
772: expected to be uniquely related to nuclear emission.
773: %Indeed, there is not
774: %an apparent correlation between X--ray and total mid-infrared luminosities
775: %\textcolor{red}{Hemos comprobado esto?}.
776:
777: %It is worth to mention that the value of the X-ray to mid-infrared ratio for the QSO
778: %H 1821+643 is equal to -1.42, which is among the type 1 nuclei values.
779:
780: %\clearpage
781:
782: \begin{figure}[!h]
783: \centering
784: \includegraphics[width=8cm,angle=90]{f7_color.eps}
785: \figcaption{\footnotesize{Histograms of observed hard X--ray fluxes for Sy1 (left), Sy2 (center), and
786: absorption corrected hard X--ray fluxes for Sy2 (right).}
787: \label{lutz_hist}}
788: \end{figure}
789:
790: %\clearpage
791:
792: % In addition to this, we have analyzed the variation of the intrinsic hard X-ray to nuclear mid-infrared luminosities
793: % ratio with the X-ray absorption column, N$_{H}$. We have asigned a lower limit of
794: % N$_{H}$ = 10$^{25}$ cm$^{-2}$ for the Sy2 Compton thick objects.
795: %
796: % The median value of this ratio for type 2 nuclei is
797: % higher than that for the Sy1, due to the lower nuclear emission in the mid-infrared of the Sy2.
798: % The value of this ratio for both Sy1 and Sy2 does not change in a consistent way with N$_{H}$,
799: % as found by \citet{Lutz04}. Nevertheless, at high values of N$_{H}$ ($\geq 10^{25} \mathrm{cm}^{-2}$),
800: % there are hints that Compton thick Sy2 exhibit higher values of the ratio. Basic unification models
801: % predict an upturn in this ratio at high column densities. Unfortunately we have not got
802: % enough data points with high N$_{H}$ to confirm this expected trend.
803:
804:
805: %\begin{figure}[!h]
806: %\centering
807: %\includegraphics[width=8cm,angle=90]{f17_color.eps}
808: %\figcaption{\footnotesize{Absorption corrected hard X--ray fluxes versus nuclear 6.75~\micron~fluxes. Linear variation
809: %is overplotted for eye guidance. Filled circles represent Sy1 (16 in total), open diamonds Sy2 (14 in total), and
810: %the QSO H~1821+643 is represented by an open triangle. Observed that hard X--ray fluxes are
811: %represented as lower limits when no N$_{H}$ data is reported.}
812: %\label{lutz_nuc_corr}}
813: %\end{figure}
814:
815:
816: %\begin{figure}[!h]
817: %\centering
818: %\includegraphics[width=8cm,angle=90]{f18_color.eps}
819: %\figcaption{\footnotesize{Intrinsic hard X-ray to nuclear mid-infrared (6.75~\micron) luminosities ratio versus X-ray
820: %absorption column, N$_{H}$. Filled circles represent Sy1
821: %(14 in total) and open diamonds Sy2 (14 in total). The ratio shows a great scatter for type 1 nuclei,
822: %and appears larger in the case of type 2 objects. Median values for type 1 and 2 are represented (Sy1 = -1.51 and
823: %Sy2 = -1.03). The ratio for the QSO H 1821+643 is equal to -1.42, what is among the type 1 values.}
824: %\label{nh}}
825: %\end{figure}
826:
827: We have also looked for a correlation between X--ray and mid-infrared luminosities
828: (see Fig.~\ref{lutz_nuc_lumin}).
829: The data corresponding to all galaxies display a linear correlation
830: of slope 0.8 with correlation coefficient r = 0.83.
831: We noticed that a good correlation appears between X--ray and nuclear
832: mid-infrared luminosities for type 1 nuclei ($\alpha$ = 0.97 and r = 0.95). However,
833: the data for type 2 nuclei present higher dispersion
834: ($\alpha$ = 0.56 and r = 0.58), showing lower mid-infrared luminosities relative to their X--ray luminosities.
835: We have also applied the Spearman's rank correlation test,
836: finding that the correlation is significant when all galaxies and only type 1 nuclei
837: are considered.
838: %However, in the flux-flux space diagram, we found not correlation, so we are in the
839: %same situation as we see in Section \ref{radio} for the nuclear emission: if the intrinsic correlation between X-ray and
840: %mid-infrared luminosities is not linear ($\alpha \neq 1$), then a random redshift distribution will smear out any
841: %flux-flux correlation \citep{Feigelson83}.
842: We have checked that this correlation is not a distance effect by upsetting in a random way the
843: redshift distribution of the galaxies.
844:
845: The ratio of hard--X ray to nuclear mid-infrared emission
846: appears larger in the case of type 1 nuclei
847: ($<log (L^X_{intr}/L^{MIR})>=-1.62\pm0.35$), than in the case of type 2
848: nuclei ($<log (L^X_{intr}/L^{MIR})> -1.19\pm0.67$). Both distributions appear
849: significatively different according to the Kolmogorov-Smirnov test.
850: We have excluded the galaxies NGC1386 and NGC7674 due to their extremely low ratios,
851: which in addition are lower limits.
852:
853: The difference in the ratios for Sy1 and Sy2 appears contradictory to that has been claimed by \citet{Lutz04}
854: and \citet{Horst06}. Nevertheless, we expect to find different results when we compare with the
855: work of \citet{Lutz04}, since they have extracted the nuclear emission by spectral
856: decomposition of very large aperture data. We have checked that their estimations
857: of the AGN contribution are largely overestimated compared to our measurements
858: of the nuclear component. In the case of \citet{Horst06}, they have obtained
859: very high resolution data in the mid-infrared range but their sample is very reduced.
860: The small dispersion claimed by \citet{Horst06} could be due to the limited
861: number of galaxies included.
862:
863:
864: Our results are in qualitative agreement with the predictions of unification models.
865: In case of type 1 nuclei we would see direct emission from the nucleus at any wavelength
866: range. In type 2 nuclei the intrinsic hard X--ray
867: emission would be similar, {\it i.e.}, after absorption correction, to type 1.
868: The mid-infrared emission coming from a dusty torus would depend on the viewing angle as a
869: function of the optical depth of the obscuring structure. The predictions vary
870: drastically for different models. The initially proposed torus models by \citet{Pier92} predict
871: large variation for both Seyfert type in the mid-infrared range. However, more recent clumpy
872: torus models do not predict large offset between both Seyfert types \citep{Nenkova02, Honig06}.
873: This variation will be larger for shorter wavelengths where the innermost part of the
874: torus dominates, and will be attenuated for longer wavelengths where the outermost
875: parts dominate. This may be one of the reasons to explain why we detect large difference
876: in the X--ray to mid-infrared ratio for types 1 and 2, contrary to that has been found by previous
877: authors using a longer wavelength filter \citep{Horst06}.
878:
879: %\clearpage
880:
881: \begin{figure}[!h]
882: \centering
883: \includegraphics[width=8cm,angle=90]{f8_color.eps}
884: \figcaption{\footnotesize{Absorption corrected hard X-ray luminosities versus nuclear
885: 6.75~\micron\ luminosities.
886: Continuous line in black is the total fit and dotted lines correspond to one $\sigma$ limits.
887: Filled circles represent Sy1 (17 in total), open diamonds Sy2
888: (13 in total), and the QSO H~1821+643 is represented by an open triangle.
889: Fits to Sy1 and Sy2 data are represented too.
890: Observed hard X--ray luminosities are represented as lower limits when no N$_{H}$ data is reported.}
891: \label{lutz_nuc_lumin}}
892: \end{figure}
893:
894: %\clearpage
895:
896: % Summarizing, we have found a general trend for Sy1 to show larger observed X--ray and mid-infrared nuclear fluxes
897: % as compared to Sy2. The intrinsic (absorption corrected) X--ray fluxes are very similar for both types of nuclei, which
898: % almost washed out the correlation between X--ray and mid-infrared fluxes. The luminosity--luminosity scatter diagram
899: % seems to indicate a good correlation between the two quantities, at least in the case of type 1 nuclei.
900:
901: \subsection{The mid-infrared properties of hidden broad-line region Seyfert 2 and narrow-line Seyfert 1 galaxies}
902:
903: According to the unification scheme of Seyfert galaxies each Sy2 galaxy should have a hidden
904: broad--line region (HBLR), as found in the prototypical NGC~1068 \citep{Antonucci85}.
905: However, spectropolarimetric surveys of complete samples of Sy2 galaxies show that hidden type 1
906: nuclei have been detected in less than $50$~\% of the galaxies from the CfA and 12~\micron~samples
907: \citep{Lumsden01, Moran00, Tran01,Tran03}.
908: These objects with failed detection of a type 1 nuclei are known as non--hidden broad-line
909: region (NHBLR) Sy2 galaxies.
910: The non--detection of broad lines could be explained in the case of an edge-on line of sight
911: or in the absence of an electron scattering region \citep{Miller90,Taniguchi99}.
912: However, some large--scale characteristics of the HBLR galaxy population are not shared by
913: the non--HBLR population.
914: The HBLR galaxies display distinctly higher radio power relative to their far-infrared output and
915: hotter dust temperature (F$_{25\micron}/F_{60\micron}$ color),
916: compared to the NHBLR Sy2 galaxies \citep{Tran03}.
917: The NHBLR galaxies also appear sistematicly as weaker radio sources
918: than their HBLR counterparts \citep{Thean01}.
919: The level of obscuration, as measured by the Balmer decrement, is indistinguishable between both types
920: of Sy2 as well as the high level of starlight domination \citep{Moran00}.
921: Thus, the relative number of HBLR and non--HBLR galaxies
922: cannot be explained by different orientations,
923: challenging the unification scheme \citep{Tran01,Tran03,Lumsden01}.
924: These results strongly support the existence of two intrinsically different populations of Sy2
925: galaxies: one harbouring an energetic, hidden Sy1 nucleus with a broad-line region
926: and the other, "true" Sy2 galaxies, with a weak or absent type 1 nucleus and a strong,
927: likely dominating starburst component \citep{Tran03}.
928:
929: Three galaxies from \citet{Tran03}
930: with spectropolarimetric confirmed HBLR are included in our initial sample of 57 AGNs,
931: namely NGC~4388, NGC~7674, and IR~05189-2524, and also
932: seven NHBLR, namely Mrk~266, NGC~1144, NGC~1241, NGC~1386, NGC~1667, NGC~3982, and NGC~5929.
933: We have investigated possible differences in the mid--infrared properties segregating both types of
934: Sy2 galaxies in our sample. We add two Sy2 galaxies to the \citet{Tran03} ones, namely IC~4397 and NGC~7592, and other
935: Seyfert types for which we have found far-infrared data in the literature \citep{Perez01}.
936: We present in Fig. \ref{tran1} the variation of the ratio between our mid-infrared nuclear and total
937: fluxes versus the far--infrared color $F_{25\micron}/F_{60\micron}$.
938: The nuclear versus total emission ratios of the HBLR are among the highest values for Sy2.
939: On the contrary, for NHBLR these ratios are among the lowest values.
940: The ratio of the nuclear versus total flux is below 0.02 in NGC~1241 and NGC~3982, and does not exceed
941: 0.5 in Mrk~266, IC~4397, and NGC~1386. For the galaxies NGC~1144 and NGC~1667 we could not identify
942: a central unresolved source, which could be associated with the presence of an AGN.
943: NGC~5929 and NGC~7592 are excluded from this analysis because their nuclear emission cannot be isolated
944: because of its close--by companion, and because of its double nucleus, respectively,
945: due to our limited spatial resolution. Both galaxies appear in Fig. \ref{tran1} in order to show their
946: far--infrared colors.
947:
948: %\clearpage
949:
950: \begin{figure*}
951: \centering
952: \includegraphics[width=8cm,angle=90]{f9_color.eps}
953: \figcaption{\footnotesize{Nuclear/total mid-infrared flux ratio versus the far-infrared
954: color $F_{25\micron}/F_{60\micron}$
955: for galaxies in our sample. NHBLR are shown as filled diamonds and confirmed HBLR
956: as open diamonds. Circles represent Sy1, squares Sy1.9,
957: asterisks Sy1.8, and the starburst galaxy is marked with a cross. IC~4397 and NGC~7592 are represented
958: as NHBLR type 2 nuclei based only in their position in this diagram.}
959: \label{tran1}}
960: \end{figure*}
961:
962: %\clearpage
963:
964: \citet{Tran03} found that the far--infrared colors $F_{25\micron}/F_{60\micron}$ compared to the radio
965: flux can be used as a good discriminant between HBLR and NHBLR (Fig. 4 in Tran 2003).
966: We present a similar diagram in Fig. \ref{fir} for the galaxies in our sample.
967: Radio fluxes (20 cm) are the same used in Section \ref{radio}.
968: Despite the low number of objects, it appears that recognized HBLR Sy2 galaxies are located
969: in the upper--right corner, whereas NHBLR galaxies tend to occupy the bottom--left corner of the diagram.
970: As a result of the use of Figs. \ref{tran1} and \ref{fir} as diagnostic diagrams we can propose new NHBLR
971: candidates. For instance, IC~4397 and NGC~7592 are located in the left--bottom quadrant
972: of Fig. \ref{fir}, and the former also shows a low nuclear to total flux ratio.
973: These facts would support the classification of IC~4397 and NGC~7592 as possible NHBLR Sy2 nucleus.
974:
975: %\clearpage
976:
977: \begin{figure*}
978: \centering
979: \includegraphics[width=8cm,angle=90]{f10_color.eps}
980: \figcaption{\footnotesize{Ratio of radio flux density S$_{20cm}$ and far-infrared flux
981: F$_{60\micron}$ as a function of
982: F$_{25\micron}/F_{60\micron}$ for galaxies of our sample. IC~4397 and NGC~7592 are represented
983: as NHBLR type 2 nuclei based only in their position in this diagram. Symbols are the same as in Fig.
984: \ref{tran1}}
985: \label{fir}}
986: \end{figure*}
987:
988: %\clearpage
989:
990: It is interesting to look at the position of intermediate Seyfert type galaxies in Fig. \ref{fir}.
991: \citet{Tran01} only reports pure Sy2 data, but in Fig. \ref{fir} we are including
992: Sy1.8 and Sy1.9 too. All the three Sy1.9 have low S$_{20cm}$/F$_{60\micron}$ and
993: F$_{25\micron}/F_{60\micron}$ ratios, corresponding to
994: the NHBLR region. Sy1.8 galaxies ocuppy random positions in the diagram.
995:
996:
997: \citet{Zhang06} suggest that NHBLR Sy2 are the counterparts of Narrow Line Seyfert 1 (NL Sy1)
998: viewed at larger angles.
999: The NL Sy1 class is characterized for very narrow Balmer lines
1000: [H$_{\beta}$ FWHM $\le$~2000 km~s$^{-1}$],
1001: strong [Fe II] lines \citep{Osterbrock85}, and violent variability in soft X-rays \citep{Boller96}.
1002: %They likely contain less massive black holes at
1003: %the Eddington limit rates \citep{Boller96,Laor97} and can be described by a slim disk \citep{Wang99,Chen04}.
1004: We have looked for galaxies in our sample classified by them as NL Sy1 and we found three:
1005: Mrk~335, Mrk~766, and NGC~4051, whose fluxes are reported in Table \ref{fluxes}. In our diagrams
1006: (Figs. \ref{tran1} and \ref{fir}) these galaxies occupy the same
1007: region as Sy1 and HBLR Sy2. According to our view this result contradicts the hypothesis
1008: of \citet{Zhang06}.
1009:
1010: \section{Conclusions}
1011:
1012: We have presented and analyzed mid-infrared data of a sample of Seyfert galaxies obtained with the
1013: instrument ISOCAM, being able to separate the nuclear and extended contributions to the total emission.
1014: The following results were found:
1015:
1016: \begin{itemize}
1017: \item The nuclear emission in the mid-infrared is a significant contribution to the total flux in Sy1 galaxies
1018: whereas for Sy2 other components overcome the nuclear emission.
1019: This result is consistent with the unification model predictions.
1020: \item The mid-infrared color distribution of the host galaxies of Sy1 appears redder than that of Sy2,
1021: whereas the nuclear mid-infrared emission seems to be more similar in both types of Seyfert galaxies, being
1022: also relatively flat.
1023: % The colors of the QSOs are in the middle of the range of nuclear colors for Seyfert galaxies,
1024: % whereas the color of the starburst is among the bluest of the sample.
1025:
1026: \item The global radio emission of Seyfert galaxies seems to be
1027: related closely to AGN activity in the most nuclear-infrared luminous galaxies (mostly Sy1),
1028: whereas for the less luminous (mostly Sy2) the radio emission would be more related to non-nuclear
1029: stellar processes.
1030:
1031: %\item Seyfert galaxies in our sample displayed the radio/mid-infrared correlation, when global
1032: %mid-infrared emission is considered, although appearing less tight that those
1033: %displayed by non-active spirals and starburst galaxies. This is naturally attributed to a relevant
1034: %contribution of non-stellar process from the AGN.
1035: %When only the nuclear mid-infrared emission is considered for the Seyfert sample,
1036: %the radio emission seems to be related closely to the nuclear mid-infrared source in the most
1037: %luminous objects, whereas for the less luminous ones there seems to be an excess of radio emission.
1038:
1039: \item The luminosity--luminosity scatter diagram between hard X-ray and mid-infrared emission seems to indicate
1040: a good correlation between the two quantities, at least in the case of type 1 nuclei.
1041: The ratio between the intrinsic hard X--ray and the
1042: nuclear mid-infrared emission presents large scatter and slightly larger values for type 2
1043: Seyfert galaxies. These results seem to be consistent with the presence of a clumpy
1044: dusty torus surrounding the active nucleus.% scatter diagram seems to indicate a good correlation between the two quantities, at least in the case
1045: % of type 1 nuclei.
1046: % A general trend is found for Sy1 to show larger observed X--ray and mid-infrared nuclear fluxes
1047: % as compared to Sy2. The intrinsic (absorption corrected) X--ray fluxes are very similar for both types of nuclei, which
1048: % almost washed out the correlation between X--ray and mid-infrared fluxes. The luminosity--luminosity
1049: % scatter diagram seems to indicate a good correlation between the two quantities, at least in the case
1050: % of type 1 nuclei.
1051: \item The mid-infrared properties of HBLR and NHBLR nuclei appear different. The nuclear to total flux
1052: ratios of the HBLR objects are among the highest values for Sy2. On the contrary, for NHBLR these ratios
1053: are among the lowest values. A diagram representing the mid-infrared
1054: nuclear to total emission ratio versus the far-infrared colors seems to be a useful tool to segregate both
1055: types of Sy2 nuclei.
1056: \end{itemize}
1057:
1058: \acknowledgments
1059:
1060: We thank the anonymous referee for his very useful comments that led to the improvement of our work.
1061: This research has made use of the NASA/IPAC Extragalactic Database (NED) which is operated by the
1062: Jet Propulsion Laboratory, California Institute of Technology, under contract with the National
1063: Aeronautics and Space Administration.
1064: This publication makes use of data products from the Two Micron All Sky Survey, which is a joint
1065: project of the University of Massachusetts and the Infrared Processing and Analysis
1066: Center/California Institute of Technology, funded by the National Aeronautics and Space
1067: Administration and the National Science Foundation. We thank the anonymous referee for his very useful comments that led to the improvement of our work.
1068:
1069: \begin{thebibliography}{}
1070:
1071: \bibitem[Alonso-Herrero et al.(2003)]{Alonso03} Alonso-Herrero, A., Quillen, A. C., Rieke, G. H., Ivanov, V. D.,
1072: \& Efstathiou, A. 2003, \aj, 126, 81
1073:
1074: \bibitem[Alonso-Herrero et al.(1996)]{Alonso96} Alonso-Herrero, A., Ward, M.J., \& Kotilainen, J.K. 1996,
1075: \mnras, 278, 902
1076:
1077: \bibitem[Antonucci \& Miller (1985)]{Antonucci85} Antonucci, R. R. J. \& Miller, J. S. 1985, \apj, 297, 621
1078:
1079: \bibitem[Antonucci (1993)]{Antonucci93} Antonucci, R. 1993, \araa, 31,473
1080:
1081: %\bibitem[Appleton et al.(2004)]{Appleton04} Appleton, P. N., et al. 2004, \apjs, 154, 147
1082:
1083: \bibitem[Ballantyne et al.(2006)]{Ballantyne06} Ballantyne, D. R., Shi, Y., Rieke, G. H., Donley, J. L.,
1084: Papovich, C., \& Rigby, J. R. 2006, \apj, 653, 1070
1085:
1086: \bibitem[Baum et al.(1993)]{Baum93} Baum S. A., O'Dea C. P, Dallacassa, D., De Bruyn A. G., \& Pedlar, A. 1993, \apj, 419, 553
1087:
1088: \bibitem[Boller et al.(1996)]{Boller96} Boller, Th., Brandt, W. N., \& Fink, H. 1996, \aap,305,53
1089:
1090: \bibitem[Buta (1996)]{Buta96} Buta, R. 1996, in IAU Colloq., 157, Barred Galaxies, ed. R. Buta, D.
1091: A. Crocker, \& B. G. Elmegreen (ASP Conf. Ser. 91; San Francisco: ASP), 11
1092:
1093: \bibitem[Cesarsky et al.(1996)]{Cesarsky96} Cesarsky, C., et al. 1996, \aap,315,32
1094:
1095: \bibitem[Chen \& Wang (2004)]{Chen04} Chen, L.-H. \& Wang, J.-M. 2004, \apj,614,101
1096:
1097: \bibitem[Clavel et al.(2000)]{Clavel00} Clavel, J., et al. 2000, \aap,357,839
1098:
1099: \bibitem[Condon et al.(1998)]{Condon98} Condon, J. J., Cotton, W. D., Greisen, E. W., Yin, Q. F., Perley, R. A.,
1100: Taylor, G. B., \& Broderick, J. J. 1998, \aj,115,1693
1101:
1102: \bibitem[Efstathiou \& Rowan-Robinson (1995)]{Efstathiou95} Efstathiou, A., \& Rowan-Robinson, M. 1995, \mnras,
1103: 273, 649
1104:
1105: \bibitem[Elbaz et al.(2002)]{Elbaz02} Elbaz, D., Cesarsky, C. J., Chanial, P., Aussel, H., Franceschini, A., Fadda, D.,
1106: \& Chary, R. R. 2002, \aap, 384, 848
1107:
1108: \bibitem[Elitzur \& Shlosman (2006)]{Elitzur06} Elitzur, M., \& Shlosman, I. 2006, \apj, 648, L101
1109:
1110: \bibitem[Fadda et al.(1998)]{Fadda98} Fadda, D., Giuricin, G., Granato, G. L., \& Vecchies, D. 1998, \apj, 496, 117
1111:
1112: \bibitem[Feigelson \& Berg (1983)]{Feigelson83} Feigelson, E. D. \& Berg, C. J. 1983, \apj, 269, 400
1113:
1114: \bibitem[Fitt et al.(1988)]{Fitt88} Fitt, A. J., Alexander, P., \& Cox, M. J. 1988, \mnras, 233, 907
1115:
1116: \bibitem[Fritz et al.(2006)]{Fritz06} Fritz, J., Franceschini, A., \& Hatziminaoglou, E. 2006, \mnras, 366, 767
1117:
1118: \bibitem[Granato \& Danese (1994)]{Granato94} Granato, G. L., \& Danese, L. 1994, \mnras, 268, 235
1119:
1120: \bibitem[Granato et al.(1997)]{Granato97} Granato, G. L., Danese, L., \& Franceschini, A. 1997, \apj, 486, 147
1121:
1122: \bibitem[Gruppioni et al.(1999)]{Gruppioni99} Gruppioni, C., Mignoli, M., \& Zamorani, G. 1999,\mnras, 304, 199
1123:
1124: \bibitem[Gruppioni et al.(2003)]{Gruppioni03} Gruppioni, C., Pozzi, F., Zamorani, G., Ciliegi, P., Lari,
1125: C., Calabrese, E., La Franca, F., \& Matute, I. 2003, \mnras, 341, L1
1126:
1127: \bibitem[H\"onig et al.(2006)]{Honig06} H\"onig, S. F., Beckert, T., Ohnaka, K., \& Weigelt, G. 2006, \aap, 452, 459
1128:
1129: \bibitem[Horst et al.(2006)]{Horst06} Horst, H., Smette, A., Gandhi, P., \& Duschl W.J. 2006, \aap, 457, L17
1130:
1131: \bibitem[Hunt \& Malkan (1999)]{Hunt99} Hunt, L.K., \& Malkan, M.A. 1999, \apj, 516, 660
1132:
1133: \bibitem[Hunt \& Malkan (2004)]{Hunt04} Hunt, L.K., \& Malkan, M.A. 2004, \apj, 616, 707
1134:
1135: \bibitem[Jaffe et al.(2004)]{Jaffe04} Jaffe, W., Meisenheimer, K., R\"ottgering, H. J. A., et al.
1136: 2004, \nat, 429, 47
1137:
1138: \bibitem[Jedrzejewski (1987)]{Jedrzejewski87} Jedrzejewski, R. I. \mnras, 226, 747
1139:
1140: \bibitem[Kessler et al.(1996)]{Kessler96} Kessler, M. F., et al. 1996, \aap, 315, 27
1141:
1142: \bibitem[Krabbe et al.(2001)]{Krabbe01} Krabbe, A., B\"oker, T., \& Maiolino, R. 2001, \apj, 557, 626
1143:
1144: \bibitem[Kuraszkiewicz et al.(2003)]{Kuraszkiewicz03} Kuraszkiewicz, J. K. 2003, \apj, 590, 128
1145:
1146: \bibitem[Ho \& Ulvestad (2001)]{Ho01} Ho, L. C., \& Ulvestad, J. S. 2001, \apjs, 133, 77
1147:
1148: \bibitem[Hummel et al.(1988)]{Hummel88} Hummel, E., Davies, R. D., Pedlar, A., Wolstencroft, R. D., \& van der Hulst, J. M. 1988,
1149: \aap, 199, 91
1150:
1151: \bibitem[Laor et al.(1997)]{Laor97} Laor, A., et al. 1997, \apj, 477, 93
1152:
1153: \bibitem[Lemke et al.(1996)]{Lemke96} Lemke, D., et al. 1996, \aap, 315, 64
1154:
1155: \bibitem[Lumsden \& Alexander (2001)]{Lumsden01} Lumsden, S. L. \& Alexander, D. M. 2001, \mnras, 328, 32
1156:
1157: \bibitem[Lutz et al.(2004)]{Lutz04} Lutz, D., Maiolino, R., Spoon, H. W. W., \& Moorwood, A. F. M.
1158: 2004, \aap, 418, 465
1159:
1160: \bibitem[Malkan et al.(1998)]{Malkan98} Malkan, M. A., Gorjian, V., \& Tam, R. 1998, \apjs, 117, 25
1161:
1162: \bibitem[Melo et al.(2002)]{Melo02} Melo, V. P., P\'{e}rez Garc\'{i}a, A. M., Acosta-Pulido, J. A., Muņoz-Tuņ\'{o}n,
1163: C., \& Rodr\'{i}guez Espinosa, J. M. 2002, \apj, 574, 709
1164:
1165: \bibitem[Miller \& Goodrich (1990)]{Miller90} Miller, J. S. \& Goodrich, R. W. 1990,\apj, 355, 456
1166:
1167: \bibitem[Moran et al.(2000)]{Moran00} Moran, E. C., Barth, A. J., Kay, L. E., \& Filippenko, A. V. 2000,
1168: \apj, 540, L73
1169:
1170: \bibitem[Nenkova et al.(2002)]{Nenkova02} Nenkova, M., Ivezi\'{c}, $\check{Z}$., Elitzur, M. 2002, \apj, 570, L9
1171:
1172: \bibitem[Oliver et al.(2000)]{Oliver00} Oliver, S., et al. 2000, \mnras, 316, 749
1173:
1174: \bibitem[Osterbrock \& Pogge (1985)]{Osterbrock85} Osterbrock, D. E. \& Pogge, R. W. 1985, \apj, 297, 166
1175:
1176: \bibitem[P\'{e}rez Garc\'\i a \& Rodr\'\i guez Espinosa (2001)]{Perez01} P\'{e}rez Garc\'\i a, A. M. \& Rodr\'\i guez Espinosa, J. M. 2001,
1177: \apj,557,39
1178:
1179: \bibitem[Piccinotti et al.(1982)]{Piccinotti82} Piccinotti, G., Mushotzky, R. F., Boldt, E. A.,
1180: Holt, S. S., Marshall, F. E., Serlemitsos, P. J., \& Shafer, R. A. 1982,\apj, 253, 485
1181:
1182: \bibitem[Pier \& Krolik (1992)]{Pier92} Pier, E. A., \& Krolik, J. H. 1992, \apj, 401, 99
1183:
1184: \bibitem[Prieto et al.(1997)]{Prieto97} Prieto, M., Gottesman, S. T., Aguerri, J.-A.L., \& Varela, A.-M. 1997,
1185: \aj, 114, 1413
1186:
1187: \bibitem[Rigby et al.(2004)]{Rigby04} Rigby, J. R., et al. 2004, \apjs, 154, 160
1188:
1189: \bibitem[Rowan-Robinson \& Crawford (1989)]{Rowan89} Rowan-Robinson, M., \& Crawford, J. 1989, \mnras, 38, 523
1190:
1191: \bibitem[Roy et al.(1998)]{Roy98} Roy, A. L., Norris, R. P., Kesteven, M. J., Troup, E. R., \& Reynolds, J. E.
1192: 1998, \mnras, 301, 1019
1193:
1194: \bibitem[Rush et al.(1996)]{Rush96} Rush, B., Malkan, M. A., \& Edelson, R. A. 1996, \apj, 473, 130
1195:
1196: \bibitem[Siebenmorgen et al.(2005)]{Siebenmorgen05} Siebenmorgen, R., Haas, M., Kr\"ugel, E.,
1197: \& Schulz, B. 2005, \aap, 436, L5
1198:
1199:
1200: \bibitem[Spinoglio \& Malkan (1989)]{Spinoglio89} Spinoglio, L., \& Malkan, M. A. 1989, \apj, 342, 83
1201:
1202: \bibitem[Taniguchi \& Anabuki (1999)]{Taniguchi99} Taniguchi, Y. \& Anabuki, N. 1999, \apj,521,103
1203:
1204: \bibitem[Thean et al.(2001)]{Thean01} Thean, A., Pedlar, A., Kukula, M. J., Baum, S. A., \& O'Dea, C. P. 2001,
1205: \mnras, 325, 737
1206:
1207: \bibitem[Tran (2001)]{Tran01} Tran, H. D. 2001, \apj, 554, 19
1208:
1209: \bibitem[Tran (2003)]{Tran03} Tran, H. D. 2003, \apj, 583, 632
1210:
1211: \bibitem[Ulvestad \& Wilson (1984)]{Ulvestad84} Ulvestad, J. S., \& Wilson, A. S. 1984, \apj, 285, 439
1212:
1213: \bibitem[Ulvestad \& Wilson (1989)]{Ulvestad89} Ulvestad, J. S., \& Wilson, A. S. 1989, \apj, 343, 659
1214:
1215: \bibitem[Ulvestad \& Ho (2001)]{Ulvestad01} Ulvestad, J. S., \& Ho, L. C. 2001, \apj, 558, 561
1216:
1217: \bibitem[Wadadekar (2004)]{Wadadekar04} Wadadekar, Y. 2004, \aap, 416, 35
1218:
1219: \bibitem[Wang et al.(1999)]{Wang99} Wang, J.-M., Szuszkiewicz, E., Lu, F.-J., \& Zhou, Y.-Y. 1999, \apj, 522, 839
1220:
1221: \bibitem[Wunderlich et al.(1987)]{Wunderlich87} Wunderlich E.,Wielebinski R., Klein U. 1987, \aaps, 69, 487
1222:
1223: \bibitem[Yee (1983)]{Yee83} Yee, H. K. C. 1983, \apj, 473, 473
1224:
1225: \bibitem[Zhang \& Wang (2006)]{Zhang06} Zhang, E.-P. \& Wang, J.-M. 2006, accepted 2006 June 5 by \apj
1226:
1227: \end{thebibliography}
1228:
1229:
1230: \clearpage
1231:
1232:
1233: \input{tab1.tex}
1234:
1235: \clearpage
1236:
1237: \input{tab2.tex.orig}
1238:
1239: \clearpage
1240:
1241: \input{tab3.tex}
1242:
1243: \end{document}
1244: