1: \documentclass[preprint,12pt]{aastex}
2: %\documentclass{emulateapj}
3: %\usepackage[dvips]{color}
4:
5: \newcommand{\kms}{\, {\rm km\, s}$^{-1}$}
6: \newcommand{\ha}{H\ensuremath{\alpha}}
7: \newcommand{\hb}{H\ensuremath{\beta}}
8: \newcommand{\hc}{H\ensuremath{\gamma}}
9: \newcommand{\hd}{H\ensuremath{\delta}}
10: \newcommand{\he}{H\ensuremath{\epsilon}}
11: \newcommand{\nha}{H\ensuremath{\alpha ^{n}}}
12: \newcommand{\bha}{H\ensuremath{\alpha ^{b}}}
13: \newcommand{\nhb}{H\ensuremath{\beta ^{n}}}
14: \newcommand{\bhb}{H\ensuremath{\beta ^{b}}}
15: \newcommand{\nhc}{H\ensuremath{\gamma ^{n}}}
16: \newcommand{\bhc}{H\ensuremath{\gamma ^{b}}}
17: \newcommand{\nhd}{H\ensuremath{\delta ^{n}}}
18: \newcommand{\bhd}{H\ensuremath{\delta ^{b}}}
19:
20: \newcommand{\nii}{[N\,II]}
21: \newcommand{\sii}{[S\,II]}
22: \newcommand{\oiii}{[O\,III]}
23: \newcommand{\feii}{Fe\,II}
24:
25: %\DeclareFixedFont{\rev}{OT1}{cmr}{bx}{10}
26: %\newfont{\rev}{cmbxsl10 scaled 1100}
27:
28:
29: \begin{document}
30: \title{Difference in Narrow Emission Line Spectra of Seyfert 1 and 2 galaxies}
31:
32: \author{
33: Kai Zhang, Tinggui Wang, Xiaobo Dong, Honglin Lu}
34: \affil{Center for Astrophysics, University of Science and
35: Technology of China (USTC), Hefei, Anhui, 230026, China; ~
36: zkdtc@mail.ustc.edu.cn, twang@ustc.edu.cn}
37:
38: \affil{Joint Institute of Galaxies and Cosmology, USTC and Shanghai
39: Observatory}
40:
41: \shorttitle{Narrow Line Spectra of Seyfert galaxies}
42: \shortauthors{Zhang et al.}
43:
44: \begin{abstract}
45: In the unification scheme of Seyfert galaxies, a dusty torus blocks
46: the continuum source and broad line region in Seyfert 2 galaxies.
47: However it is not clear whether or not and to what extent the torus
48: affects the narrow line spectra. In this paper, we show that Seyfert
49: 1 and Seyfert 2 galaxies have different distributions on the
50: [O\,III]/H$\,\beta $ vs [N\,II]/H$\,\alpha$ diagram (BPT diagram)
51: for narrow lines. Seyfert 2 galaxies display a clear left boundary
52: on the BPT diagram and only 7.3\% of them lie on the left. By
53: contrast, Seyfert 1 galaxies do not show such a cutoff and 33.0\%
54: of them stand on the left side of the boundary. Among Seyfert 1
55: galaxies, the distribution varies with the extinction to broad
56: lines. As the extinction increases, the distribution on BPT diagram
57: moves to larger [N\,II]/H$\,\alpha$ value. We interpret this as an
58: evidence for the obscuration of inner dense narrow line region by
59: the dusty torus. We also demonstrate that the [O\,III] and broad
60: line luminosity correlation depends on the extinction of broad lines
61: in the way that high extinction objects have lower uncorrected
62: [O\,III] luminosities, suggesting that [O\,III] is partially
63: obscured in these objects. Therefore, using [O\,III] as an indicator
64: for the nuclear luminosity will systematically under-estimate the
65: nuclear luminosity of Seyfert 2 galaxies.
66:
67: \end{abstract}
68:
69: \keywords{galaxies: active--galaxies:Seyfert--(galaxies:) quasars:
70: emission lines }
71:
72:
73: %\altaffiltext{3}{Max-Planck-Institute for Astrophysics,
74: %Karl-Schwarzschild-Str. 1, D-85741 Garching, Germany}
75:
76: %\altaffiltext{4}{MPA/SHAO Joint Center for Astrophysical Cosmology
77: %at Shanghai Astronomical Observatory, Nandan Road 80, Shanghai
78: %200030, China}
79:
80:
81: \section{Introduction}
82:
83: The unification scheme of Seyfert galaxies has been widely accepted
84: to explain the dichotomy of Seyfert 1 (hereafter Sy1) and Seyfert 2
85: galaxies (hereafter Sy2) (e.g. Antonucci 1993). The basic idea is
86: that the differences between Sy1s and Sy2s are caused by an
87: orientation effect. The broad line region (BLR) and continuum source
88: are blocked from our view in Sy2s by thick obscuring material,
89: presumably a dusty torus (Antonucci \& Miller 1985) at large
90: inclination angles, while they are observed directly in Sy1s. In
91: both types of Seyfert galaxies, narrow emission lines, produced on a
92: large scale, are observed. This scheme has been supported by many
93: observations (Antonucci \& Miller, 1985; Miller \& Goodrich, 1990;
94: Kl\"{o}ckner et al, 2003; Greenhill et al, 2003; Jaffe et al, 2004),
95: and even the temperature distribution of the dusty torus has been
96: measured (e.g., Tristram et al, 2007).
97:
98: While it is general agreed that dusty tori obscure BLR and
99: continuum source, and even the central narrow line region (NLR) of
100: Sy2s (e.g., Rhee \& Larkin 2005; Haas et al. 2005), it is still
101: unclear to what extent they obscure the NLR, and how the obscuration
102: depends on the orientation (i.e., the difference in NLR obscuration
103: among different types of Seyfert galaxies). This is an important
104: issue because the narrow line luminosity is often used as a
105: surrogate for the bolometric luminosity of Seyfert 2 nuclei. In
106: addition, it may yield important clues to the structure of NLR.
107: Jackson \& Browne (1990) compared [O\,III] emission from quasars
108: (i.e., radio-loud QSOs) with radio galaxies --- both of them are
109: believed to be the analogs of the two types of Seyfert galaxies
110: (Barthel 1989), and found that the [O\,III] emission of quasars is
111: much stronger than that of radio galaxies. It was interpreted as
112: part of the [O\,III] emission is obscured by the torus in radio
113: galaxies. A similar conclusion was reached by Haas et al. (2005) by
114: comparing [O\,III]$\lambda$5007 with [OIV]$\lambda$25.9$\mu$m. On
115: the other hand, Barthel \& Fosbury (1994) showed that, when
116: comparing [OII] emission of quasars and radio galaxies, both classes
117: of objects have very similar distributions, corroborating the
118: obscuration scenario. Unfortunately, [OII] emission may be enhanced
119: by shocks induced by radio plasma (Best et al. 1999) and thus may
120: not be a good tracer of AGN luminosity. Moreover, a fraction of FRII
121: radio galaxies may be intrinsically less luminous than quasars
122: (Ogle, Whysong \& Antonucci, 2006; also Meisenheimer et al. 2001).
123:
124: It is well known that NLR is stratified, at least for AGNs of some
125: sort (e.g.,Veilleux et al. 1991, Robinson et al. 1994). Emission
126: lines of higher critical density or ionization potential are
127: produced averagely closer to the active nucleus than those of lower
128: critical density or ionization. This fact can be used to check
129: whether the inner NLR region is blocked by torus in Sy2s; and if
130: yes, to what extended NLR is obscured. Murayama \& Taniguchi (1998)
131: found that Sy1s show excess [Fe VII]$\lambda$6087 with respect to
132: Sy2s, and proposed that [Fe VII] arises mainly from the inner wall
133: of dusty tori. Schmitt (1998) compared [OII]$\lambda$3727\AA,
134: [NeIII]$\lambda$3869\AA, [O\,III]$\lambda$5007\AA,
135: [NeV]$\lambda$3426\AA, as well as 60$\mu$m continuum flux from a
136: sample of 52 Sy1s to those of 68 Sy2s; they found that Sy1s have
137: higher excitation than Sy2s. However, the differences were
138: interpreted as that Sy1s have smaller number of ionization bounded
139: clouds than Sy2s by him.
140:
141:
142: In this paper, we study the location of Sy1s and Sy2s on the BPT
143: diagrams, short for Baldwin-Phillips-Terlevich diagrams (Baldwin et
144: al 1981), which are often used for classification of narrow emission
145: line galaxies. With the large spectroscopic sample of both Sy1s and
146: Sy2s available from the Sloan Digital Sky Survey (York et al, 2000),
147: and spectral decomposition, we are able to detect the different
148: distributions of Sy1s and Sy2s on BPT diagrams.
149:
150:
151:
152: \section{Sample and Data Analysis}
153: \subsection{Seyfert 1 and 2 galaxy samples}
154:
155:
156: The Seyfert 2 galaxies sample is selected from the galaxy catalog of
157: SDSS Data Release 4 (DR4). For our purpose, a redshift cut of
158: z$<$0.3 is used to ensure that the emission lines of interest,
159: H$\,\alpha$, H$\,\beta$ , [O\,III] and [N\,II], [S\,II], fall in the
160: spectra. We subtracted the stellar continuum to leave a clean
161: emission line spectrum following the recipe described in Lu et al.
162: (2006). In brief, we fit galaxy spectra with the templates derived
163: by applying Ensemble Learning for Independent Component Analysis
164: (EL-ICA) to the simple stellar population library (Bruzual \&
165: Charlot 2003). The templates were then broadened and shifted to
166: match the stellar velocity dispersion of the galaxy. In this way,
167: the stellar absorption lines are reasonably well subtracted to
168: ensure reliable measurement of weak emission lines. Next, we fitted
169: the emission line spectrum using gaussians (see next section) to
170: derive emission line parameters. In practice, stellar subtraction
171: and the emission line fitting are iterated because line parameters
172: are used to create correct mask regions for emission lines during
173: the stellar modeling. Only sources with H$\alpha$, H$\beta$ ,
174: [O\,III], [N\,II] and [S\,II] lines detected with S/N $> 5$ are
175: considered for further study. Broad emission line objects are
176: removed from the Seyfert 2 galaxies sample (described below). We
177: adopt the criteria of Kewley et al. (2006) to classify the emission
178: line galaxies as star forming galaxies, Seyfert galaxies, LINER,
179: AGN/HII composites. For this study, we only stick to the classical
180: Seyfert galaxies because only a very small fraction of Sy1s show a
181: composite/LINER spectrum while there are large portion of such Sy2s.
182: The final sample consists of 5544 sources.
183:
184:
185: Seyfert 1 galaxies\footnote{A small number of objects are in the
186: luminosity range for quasars.} were drawn from quasar and galaxy
187: catalogues of SDSS DR4 at redshift $z<0.3$. For nucleus dominated
188: sources where Fe\,II multiplets and other broad emission lines are
189: highly blended, we fit simultaneously the nuclear continuum, the
190: Fe\,II multiplets and emission lines (see Dong et al. 2008 for
191: details). The nuclear continuum is approximated by a broken
192: power-law, with free indices for \ha\ and \hb\ region respectively.
193: Fe II emission, both broad and narrow, is modeled using the spectral
194: data of the Fe II multiplets for I Zw 1 provided by Veron-Cetty et
195: al. (2004). Emission lines are modeled as multiple gaussians. For
196: those spectra with significant contribution of starlight as measured
197: by the equivalent widths (EWs) of the Ca\,II\,K $\lambda3934$
198: absorption line or high order Balmer lines or Na\,I
199: $\lambda\lambda5890,5896$, a starlight model is also included using
200: the 6 IC templates as described in the above. The decomposition of
201: host-galaxy starlight, nuclear continuum and FeII emission were
202: carried out following the procedure as described in detail in Zhou
203: et al. (2006). We fit and subtracted the above model from the SDSS
204: spectra, and then fit emission lines using multi-gaussian model (see
205: next section).
206:
207: Based on the fitted emission-line parameters, we construct
208: the sample of Sy1s according to the criteria as follows. (1) Adding
209: a broad gaussian component of H$\alpha$ to the model can improve the
210: fit significantly with a chance probability less than 0.05 according
211: to F-test; (2) The broad component is detected with S/N $\gtrsim 5$;
212: (3) The peak of the broad component is at least twice the
213: root-mean-square (rms) deviation in the regions surrounding \ha\
214: (Dong et al., in prep). To guarantee the reliability of the
215: measurements of the narrow lines of interest, we require that they
216: have S/N $> 5$, the same as in the cases of Sy2s. Here the flux
217: errors of narrow lines, especially \ha\ and \hb\ narrow lines, have
218: taken account for the uncertainty introduced by the emission-line
219: modeling (Dong et al., in preparation). The final sample has 2768
220: Sy1s.
221:
222:
223: \subsection{Emission Line Fitting}
224:
225: For Sy2s, after removing the starlight, we fit the emission-line
226: spectra using the code described in detail in Dong et al. (2005).
227: Briefly, every line is fitted with one or more Gaussians as is
228: statistically justified (mostly with 1--2 Gaussians); the line
229: parameters are achieved by minimizing $\chi^2$. The
230: \oiii$\lambda\lambda4959,5007$ doublet lines are assumed to have the
231: same profiles and redshifts; \nii$\lambda\lambda6548,6583$ and
232: \sii$\lambda\lambda6716,6731$ doublet are constrained in the same
233: way. Furthermore, the flux ratios of \oiii\ doublet and \nii\
234: doublet are fixed to the theoretical values. Usually, \ha\ line and
235: \nii\ doublet are highly blended and thus hard to isolate; in such
236: cases, we fit them assuming they have the same profile as \sii\
237: doublet, which is empirically justified (e.g., Filippenko \& Sargent
238: 1988; Ho et al. 1997; Zhou et al. 2006).
239:
240: For Sy1s, once the nuclear continuum and the Fe\,II emission are
241: subtracted, we perform a refined fit of the emission-line spectra.
242: The narrow lines are modeled as for Sy2s. The broad lines are
243: modeled with multi-gaussians, as many as is statistically justified.
244: One concern is the reliability of the decomposition of narrow and
245: broad lines. We found that when the broad line is significantly
246: broader than the narrow line and the narrow line is not weak, the
247: flux of narrow line is trustworthy (Dong et al., in preparation).
248: This is usually the case for narrow lines (S/N$>5$) of Sy1s in this
249: sample.
250:
251:
252: \section{Results}
253:
254: \subsection{A Comparison between Seyfert 1 and 2 galaxies}
255:
256: Dong et al.(2008) found that the distribution of broad-line Balmer
257: decrement for blue AGNs is a gaussian in the logarithmic space, with
258: a peak at 0.486 (H$\,\alpha$/H$\,\beta$=3.1) and an intrinsic
259: standard deviation only $\sim$0.03. The reddening interpretation of
260: large Balmer decrements is also supported by larger infrared to
261: broad line ratio (Dong et al. 2005) and excess X-ray absorption in
262: those objects (Wang et al. 2008). We estimate reddening to broad
263: lines for objects with $H\alpha/H\beta$ significantly larger than
264: the above value assuming a SMC extinction curve, by
265: $E_{B-V}^{b}=1.99\times [\log (H\alpha/H\beta)-0.486]$. Extinction
266: to narrow lines is also estimated using the Balmer decrement and
267: assuming an intrinsic narrow-line Balmer decrement of 3.1. Using the
268: $E_{B-V}$ and extinction curve, the intrinsic luminosities of
269: H$\,\alpha$, H$\,\beta$ broad lines, as well as H$\,\alpha$,
270: H$\,\beta$, [O\,III] [N\,II], [S\,II] narrow lines can be
271: calculated.
272:
273: We plot the 5544 Sy2 and 2768 Sy1 sources on the BPT diagram in
274: Figure 1. The intermediate broad line $E_{B-V}^{b}$ group are not
275: plotted for clarity. One can easily see that Sy1s and Sy2s
276: apparently occupy different regions on BPT diagram. We defined a
277: line (S12 line for short) such that Sy2s rarely appear on the left
278: side:
279: \begin{equation}
280: \log([O\,III]/H\beta)= 3.53\times \log([N\,II]/H\alpha)+1.65
281: \end{equation}
282: Only 409 (7.3\%) Sy2s lie on the left side of S12 line while 913
283: (33.0\%) Sy1s on the left side. The overall trend is very clear that
284: Sy1s lie to the left of Sy2s on the diagram. On average,
285: [N\,II]/H$\alpha$ ratio is 0.11 dex higher in Sy2s than in Sy1s.
286:
287: \subsection{Seyfert 1 galaxies of different extinction}
288: If we regard Sy2s as Seyfert galaxies with extremely high
289: $E_{B-V}^{b}$ in broad lines, the trend of different types of
290: Seyfert galaxies should also be traced for Sy1s of different broad
291: line extinction $E_{B-V}^b$. In the right panel of Fig 1, we split
292: the source according to their $E_{B-V}^{b}$ of broad emission lines.
293: Only Sy1s in $E_{B-V}^{b}\in[0, 0.2]$ and $E_{B-V}^{b}\in[0.6, 1]$
294: are plotted for clarity. It is evident that as the $E_{B-V}^{b}$
295: increases, the distribution moves to right. This is in accordance
296: with our hypothesis. One can see that a small fraction (23.8\%) of
297: objects in the $E_{B-V}^{b}\in[0.6, 1]$ group lie on the left side
298: of S12 line while the fraction for the low extinction group is
299: (43.3\%).
300:
301:
302: \subsection{On the narrow and broad line correlation}
303:
304: From the broad line sample, an empirical relationship between
305: uncorrected [O\,III] luminosity and broad H$\,\alpha$ line
306: luminosity can be obtained. The broad H$\,\alpha$ is corrected from
307: the extinction to the broad line region as in last section.
308: \begin{equation}
309: \log(L_{[O\,III]}^{uncorrected} )=(0.977\pm0.007)\times
310: log(L_{H\,\alpha^b}^{intri} )+(0.238\pm0.294)
311: \end{equation}
312: %The relation has a relative large [O\,III] scatter of 0.302 dex.
313: In Fig 2, we split the sample by their $E_{B-V}^{b}$ and plot them
314: on the [O\,III] vs H$\,\alpha$ diagram. The purple crosses are with
315: low $E_{B-V}^{b}\in[0,0.2]$ while green triangle symbols with high
316: $E_{B-V}^{b}\in[0.6,1]$. The average ratios (in logarithmic value)
317: of luminosities of [O\,III] and H$\,\alpha$ for these two groups are
318: $-0.678\pm0.278$ and $-1.012\pm0.158$. Thus the [O\,III]
319: luminosities of the low $E_{B-V}^{b}$ group are on average two times
320: (0.334 dex) larger than that of high $E_{B-V}^{b}$ group at a given
321: broad line luminosity. The mean values of the two groups are
322: significantly different at a chance probability less than $10^{-10}$
323: according to the Student's t-test.
324: %%The student T-test gives a result
325: %%of T=14.85 and a null hypothesis at a chance probability less than
326: %%$10^{-10}$.
327: % It is obvious that the they form two
328: %distinct lines with a [O\,III] dispersion of 0.25 dex.
329:
330: %The fitting to two groups lead to the result of
331: %\begin{eqnarray*}
332: % \log(L_{[O\,III]}^{obs} )&=&(0.925\pm0.009)\times
333: %\log(L_{H\,\alpha^b}^{intri}
334: %) \\
335: %& &+(2.521\pm0.406)
336: % \;\;\; (E_{B-V}^b\in [0,0.2])
337: %\end{eqnarray*} %($E_{B-V}\in$ [0,0.2])
338: %and
339: %\begin{eqnarray*}
340: %\log(L_{[O\,III]}^{uncorrected} ) & = &(1.026\pm0.017)\times
341: %\log(L_{H\,\alpha^b}^{intri}
342: %) \\
343: %& & -(2.190\pm0.736)
344: % \;\;\; (E_{B-V}^b\in [0.6,1])
345: %\end{eqnarray*}
346: %For a given broad line luminosity, AGNs in high $E_{B-V}^{b}$ bin
347: %has a factor three lower [O\,III] luminosity than in the low
348: %$E_{B-V}^{b}$ bin.
349:
350: One concern is that this may be introduced by the biases that the
351: measurement error in the $E_{B-V}^{b}$ will lead to a shift in the
352: relation for different $E_{B-V}^{b}$. However, Monte-Carlo
353: simulation shows that it causes a deviation of only 0.06 dex for a
354: typical H$\,\alpha$ luminosity, much smaller than the difference of
355: 0.334 dex. Thus this must be real.
356:
357:
358: \section{Discussion and conclusion}
359:
360: We find that Seyfert 1 galaxies have significantly smaller
361: [N\,II]/H$\alpha$ ratios than Seyfert 2 galaxies. A similar trend
362: has also been observed in Sy1s with different BLR reddening:
363: unreddened Seyfert 1 galaxies have smaller ratios than reddened
364: ones. We also show that reddened Seyfert 1 galaxies have average
365: lower uncorrected [O\,III] luminosity at a given nuclear luminosity
366: represented by broad H$\alpha$ luminosity.
367:
368: Before discussing the implication of these findings, we would like
369: to rule out the possibility that they are introduced by systematic
370: bias in our spectral modeling and sample selection. First, as a
371: sanity check, we fit all the above-mentioned narrow lines assuming
372: that they have the same profile; it does not change our results.
373: Thus our findings are not caused by the emission-line modeling
374: (different models for different lines). Second, stellar absorption
375: lines are properly subtracted in both Sy1s and Sy2s. We have
376: visually inspected the high-order Balmer absorption lines to ensure
377: they are well matched. Thus we are confident that the fluxes of
378: narrow lines are not affected seriously by stellar absorption
379: features. Third, we check Sy1s with either small difference in line
380: width between broad lines and narrow lines, e.g., narrow line Sy1s,
381: or with a relative low signal to noise ratio, where the
382: decomposition of narrow and broad components is most difficult. We
383: find that their location on BPT diagram is indistinguishable from
384: other Sy1s. Therefore, the line-profile decomposition should not
385: affect our results. Last, but not the least, we check the influence
386: of our sample selection by raising the criterion for the S/N of
387: narrow line to be 7. We find that the fraction of objects lying on
388: the left side of S12 line is 9\% for Sy2s, 26.4\% for Sy1s in the
389: $E_{B-V}^{b}\in[0.6, 1]$ group and 44.6\% for Sy1s in the
390: $E_{B-V}^{b}\in[0, 0.2]$ group. These values are almost the same as
391: those presented in \S3.
392:
393:
394: Let's consider the fact that there are substantial Sy1s with very
395: low [N\,II]/H$\alpha$ first. Low [N\,II]/H$\alpha$ can be produced
396: with very metal poor gas (Groves et al. 2006), or presence of gas
397: with a density above the [N\,II] critical density with high
398: ionization. We argue that metal-poor gas hypothesis is unlikely.
399: Groves et al. showed that gas metallicity is correlated with the
400: mass of galaxy bulge, which is in turn correlated with central black
401: hole mass. However, we find that Sy1s on the left of S12 line have a
402: similar black hole mass distribution as the whole sample.
403: Furthermore, if the metal-poor gas hypothesis is right, then the
404: fact that this group almost does not have type 2 counterparts is
405: directly in contrast to the well-supported unification scheme.
406:
407: The different distribution of Sy1s and Sy2s on BPT diagram does not
408: suggest the failure of AGN unification because reddened Sy1s show a
409: distribution in between. Rather it can be interpreted as partial
410: obscuration to the NLR in Sy2s. It is well established that the NLR
411: is stratified with high density and high ionization gas at close to
412: the continuum source whereas low density and low ionization gas is
413: in outer part of NLR(Nagao et al. 2003, Riffel et al. 2006). At high
414: densities, lines with low critical densities, such as [N\,II],
415: [S\,II], are suppressed by collisional de-excitation process, while
416: lines with high critical densities and recombination lines, are
417: unaffected. Considering [O\,III] is the dominant narrow line in
418: Sy1s, we hypothesis that even in the inner NLR, the gas density does
419: not exceed the critical density of [O\,III], i.e., [O\,III],
420: H$\alpha$ and H$\beta$ trace each other in the inner NLR. Once the
421: inner NLR is obscured by opaque dusty material in Sy2s, one would
422: expect that the H$\alpha$ emitting region to be shielded while
423: [N\,II] emitting region being affected to a less degree or not at
424: all, thus the [N\,II]/H$\alpha$ ratio will be shifted to a larger
425: value while [O\,III]/H$\beta$ remains the same. The uncorrected
426: [O\,III] luminosity of reddened Sy1s are underluminous in comparison
427: with unreddened one is consistent with the partial obscuration
428: interpretation \footnote{If the obscuring material is opaque and
429: covers only the inner part of NLR, narrow line flux will be smaller.
430: Meanwhile, due to its small size, the BLR is entirely covered by
431: dust and the broad lines can be efficiently corrected from Balmer
432: decrement.}.
433:
434: The obscuring material can be the extended part of the torus or a
435: dusty lane in the host galaxy. The inner edge of dusty torus is
436: known to be order of parsecs from the central engine (Rhee et al.
437: 2006), but the extent and height of the torus are not well
438: constrained. Schmitt et al. (2003) showed that all Sy1s have a
439: resolved bright [O\,III] emission knot at the nucleus on the scales
440: typically tens parsecs, and half [O\,III] emission radius is
441: typically tens parsecs. Thus, combined with their result, our
442: findings here suggest that, at least in a large fraction ($\gtrsim
443: 30$ per cent) of AGNs if not all, the obscuring material (likely
444: being the torus) must extend to several tens of parsecs.
445:
446: %\section*{Acknowledgements}
447: \acknowledgements We thank Guinevere Kauffmann and Xue-Guang Zhang,
448: as well as the anonymous referee for critical comments. We also
449: thank Ting Xiao and Shaohua Zhang for useful discussion. This work
450: is supported by Chinese NSF grants NSF-10533050 and NSF-10573015,
451: the Knowledge Innovation Program (Grant No. KJCX2-YW-T05). Funding
452: for the Sloan Digital Sky Survey (SDSS) has been provided by the
453: Alfred P. Sloan Foundation, the Participating Institutions, the
454: National Aeronautics and Space Administration, the National Science
455: Foundation, the U.S. Department of Energy, the Japanese
456: Monbukagakusho, and the Max Planck Society. The SDSS is managed by
457: the Astrophysical Research Consortium (ARC) for the Participating
458: Institutions. The SDSS Web site is http://www.sdss.org/.
459:
460:
461: \begin{thebibliography}{}
462: \bibitem[Antonucci(1993)]{1993ARA&A..31..473A} Antonucci, R.\ 1993, \araa,31, 473
463: \bibitem[Antonucci(1985)]{1985ApJ..297..621A} Antonucci, R. \&Miller ,\ 1985,
464: \apj,297, 621
465: \bibitem[Baldwin et al. 1981]{ba81} Baldwin, J. A., Phillips, M. M., \& Terlevich, R., 1981, PASP, 93, 5
466: \bibitem[Barthel(1994)]{1994ASPC...54..175B} Barthel, P.~D.\ 1994, The
467: Physics of Active Galaxies, 54, 175
468: \bibitem[Barthel (1989)] {1989ApJ...336..606B}
469: Barthel, Peter D.,\ 1989, \apj, 336, 606
470: \bibitem[Best et al.(1999)]{1999IAUS..194..241B} Best, P., R{\"o}ttgering,
471: H., \& Longair, M.\ 1999, Activity in Galaxies and Related
472: Phenomena, 194, 241
473: \bibitem[Bruzual et al. (2003)] {2003MNRAS.344.1000B}
474: Bruzual, G.; Charlot, S.,\ 2003 ,\mnras, 344, 1000
475:
476: \bibitem[Dong et al.(2005)]{2005ApJ...620..629D} Dong, X.-B., Zhou, H.-Y.,
477: Wang, T.-G., Wang, J.-X., Li, C., \& Zhou, Y.-Y.\ 2005, \apj, 620,
478: 629
479:
480: \bibitem[Dong et al.(2008)]{2008MNRAS...383...581D} Dong, X. B; Wang, T. G; Wang, J. G; Yuan, W. M; Zhou, H. Y; Dai, H. F; Zhang, K
481: \ 2008, \mnras, 383, 581
482: \bibitem[Filippenko\& Sargent(1988)]{1988ApJ...324..134F} Filippenko, A.~V., \&
483: Sargent, W.~L.~W.\ 1988, \apj, 324, 134
484:
485: \bibitem[Greenhill et al, (2003)]{2003ApJ...590..162G}Greenhill et al,\ 2003, \apj,
486: 590,162
487: \bibitem[Groves et al. (2006)] {2006MNRAS.371.1559G}
488: Groves, Brent A.; Heckman, Timothy M.; Kauffmann, Guinevere, \ 2006,
489: \mnras, 371, 1559
490: \bibitem[Haas et al.(2005)]{2005A&A...442L..39H} Haas, M., Siebenmorgen, R., Schulz,
491: B., Kr{\"u}gel, E., \& Chini, R.\ 2005, \aap, 442, L39
492: \bibitem[Ho et al.(1997)]{1997ApJS..112..391H} Ho, L.~C., Filippenko,
493: A.~V., Sargent, W.~L.~W., \& Peng, C.~Y.\ 1997, \apjs, 112, 391
494: \bibitem[Jackson et al.(1990)]{1990Nature...343...43J}
495: Jackson, N. \& Browne, I. W. A. 1990, Nature, 343, 43
496: \bibitem[Jaffe et al.(2004)]{2004ApJ...615...55J}
497: Jaffe et al, \ 2004 , \apj, 615, 55
498: \bibitem[Kewley et al. (2006)]{kew06} Kewley, L. J., Groves, B., Kauffmann, G., \& Heckman, T., 2006, \mnras, 372, 961
499: \bibitem[Kl{\"o}ckner et al.(2003)]{2003Natur.421..821K} Kl{\"o}ckner,
500: H.-R., Baan, W.~A., \& Garrett, M.~A.\ 2003, \nat, 421, 821
501: \bibitem[Lu et al.(2006)]{2006AJ....131..790L} Lu, H., Zhou, H., Wang, J.,
502: Wang, T., Dong, X., Zhuang, Z., \& Li, C.\ 2006, \aj, 131, 790
503: \bibitem[Meisenheimer et al. (2001)] {2001A&A...372..719M}
504: Meisenheimer, K.; Haas, M.; M¨¹ller, S. A. H.; Chini, R.; Klaas, U.;
505: Lemke, D., \ 2001, A\&A, 372, 719
506: \bibitem[Miller et al. (1990)] {1990ApJ...355..456M}
507: Miller, J. S.; Goodrich, R. W., \ 1990, \apj, 355, 456
508: \bibitem[Murayama et al. (1998)] {1998ApJ...503L.115M}
509: Murayama, Takashi; Taniguchi, Yoshiaki, \ 1998, \apj, 503L, 115
510: \bibitem[Nagao et al. (2003)] {2003AJ....126.1167N} Nagao, Tohru;
511: Murayama, Takashi; Shioya, Yasuhiro; Taniguchi, Yoshiaki, \ 2003,
512: \aj, 126, 1167
513: \bibitem[Ogle et al. (2006)]{ 2006ApJ...647..161O}
514: Ogle, Patrick; Whysong, David; Antonucci, Robert, \ 2006, \apj, 647,
515: 161
516: \bibitem[Rhee et al. (2006)] {2006ApJ...640..625R}
517: Rhee, Joseph H.; Larkin, James E., \ 2006, \apj, 640, 625
518: \bibitem[Riffel et al. (2006)]{2006A&A...457...61R}
519: Riffel, R.; Rodr¨ªguez-Ardila, A.; Pastoriza, M. G.,\ 2006, A\&A,
520: 457, 61
521: \bibitem[Robinson et
522: al, 1994]{1994A&A...291..351R} Robinson, A. et al. \ 1994, A\&A,
523: 291, 351
524: \bibitem[Schmitt, H. R. (1998)]{1998ApJ...506..647S} %
525: Schmitt, H R.\ 1998 ,\apj, 506, 647
526: \bibitem[Schmitt et al. 2003] {2003ApJ...597..768S}
527: Schmitt, H. R.; Donley, J. L.; Antonucci, R. R. J.; Hutchings, J.
528: B.; Kinney, A. L.; Pringle, J. E., \ 2003, \apj, 597, 768
529:
530: \bibitem[Tristram et al, 2007]{2007A&A...474..837T} Tristram, K.
531: R. W. et al. \ 2007, A\&A, 474, 837
532: \bibitem[1991]{veilleux} Veilleux, S., \& Sylvain, 1991, \apj, 369, 331
533: \bibitem[V{\'e}ron-Cetty et al.(2004)]{2004A&A...417..515V}
534: V{\'e}ron-Cetty, M.-P., Joly, M., \& V{\'e}ron, P.\ 2004, \aap, 417,
535: 515
536: \bibitem[Wang et al, (2008)]
537: Wang T. G. et al , \aj, submitted
538: \bibitem[York et al.(2000)]{2000AJ....120.1579Y} York, D.~G.~et al.\ 2000, %
539: \aj, 120, 1579
540:
541:
542: \bibitem[Zhou et al.(2006)]{2006ApJS..166..128Z} Zhou, H., Wang, T., Yuan,
543: W., Lu, H., Dong, X., Wang, J., \& Lu, Y.\ 2006, \apjs, 166, 128
544:
545:
546: \end{thebibliography}
547: \clearpage %
548: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% figures
549: \begin{figure*}
550: \begin{center}
551: \label{fig-1}
552: \includegraphics[width=16cm]{f1.eps}
553: \caption{BPT diagram for Seyfert 1 (right panel) and Seyfert 2 (left
554: panel) galaxies. The lower curve is the empirical line separating
555: AGN from star-forming galaxies (Kewley et al 2006). The straight
556: line is the S12 line described in the text. Most Seyfert 2 galaxies
557: locate on the right side of the line. In the right panel, purple
558: crosses represent objects with $E_{B-V}^{b}<0.2$ and green triangles
559: those with $E_{B-V}^{b}\in [0.6,1]$ while the intermediate
560: $E_{B-V}^{b}$ group are not plotted for clarity.}
561:
562: \end{center}
563: \end{figure*}
564:
565:
566:
567: %%\clearpage
568: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% figures
569: \begin{figure*}
570: \begin{center}
571: \label{fig-2}
572: \includegraphics[width=10cm]{f2.eps}
573: \caption{ The uncorrected luminosity of [O\,III] versus extinction
574: corrected luminosity of broad H$\,\alpha$ for Seyfert 1 galaxies.
575: The purple crosses are AGNs with $E_{B-V}^{b}<0.2$ and green
576: triangles with $E_{B-V}^{b}\in [0.6,1]$. The blue line shows the
577: best linear fit to the whole sample. The intermediate $E_{B-V}^{b}$
578: group is not plotted for clarity. }
579: \end{center}
580: \end{figure*}
581:
582: %\clearpage
583:
584:
585: \end{document}
586: