The Different Nature in Seyfert 2 Galaxies With and Without Hidden Broad-Line Regions
Yu-Zhong Wu, En-Peng Zhang, Yan-Chun Liang, Cheng-Min Zhang, Yong-Heng Zhao
aa r X i v : . [ a s t r o - ph . H E ] J a n Preprint typeset using L A TEX style emulateapj v. 11/10/09
THE DIFFERENT NATURE IN SEYFERT 2 GALAXIES WITH AND WITHOUT HIDDEN BROAD-LINEREGIONS
Yu-Zhong Wu , En-Peng Zhang , Yan-Chun Liang , Cheng-Min Zhang , and Yong-Heng Zhao ABSTRACTWe compile a large sample of 120 Seyfert 2 galaxies (Sy2s) which contains 49 hidden broad-lineregion (HBLR) Sy2s and 71 non-HBLR Sy2s. From the difference in the power sources between twogroups, we test if HBLR Sy2s are dominated by active galactic nuclei (AGNs), and if non-HBLR Sy2sare dominated by starbursts. We show that: (1) HBLR Sy2s have larger accretion rates than non-HBLR Sy2s; (2) HBLR Sy2s have larger [Ne v ] λ . ii ] λ .
81 and [O iv ] λ . ii ] λ . IRAS f /f flux ratio which showsthe relative strength of the host galaxy and nuclear emission than non-HBLR Sy2s. So we suggestthat HBLR Sy2s and non-HBLR Sy2s are AGN-dominated and starburst-dominated, respectively.In addition, non-HBLR Sy2s can be classified into the luminous ( L [O III] > ergs s − ) and lessluminous ( L [O III] < ergs s − ) samples, when considering only their obscuration. We suggestthat: (1) the invisibility of polarized broad lines (PBLs) in the luminous non-HBLR Sy2s depends onthe obscuration; (2) the invisibility of PBLs in the less luminous non-HBLR Sy2s depends on the verylow Eddington ratio rather than the obscuration. Subject headings: galaxies: active — galaxies: Seyfert — galaxies: statistics INTRODUCTION
According to the unified model of active galactic nuclei(AGNs; Antonucci 1993), type 1 AGNs are seen face-onand have both narrow and broad emission lines; type2 AGNs are seen edge-on and have only narrow emis-sion lines, which are commonly believed to be intrinsi-cally the same as type 1 AGNs. With the discoveriesof both polarized broad lines (PBLs) in NGC 1068 (An-tonucci & Miller 1985) and some hidden broad-line re-gions (HBLRs) in other Seyfert 2 galaxies (Sy2s; Miller& Goodrich 1990; Tran et al. 1992; Young et al. 1996;Heisler et al. 1997; Kay & Moran 1998), Seyfert 2 galax-ies are classified into HBLR and non-HBLR Sy2s. About50% of the total currently known Seyfert 2 galaxies showthe presence of HBLRs in their polarized optical spec-tra, while the remaining half do not (Tran 2001, 2003;Nicastro et al. 2003; Haas et al. 2007).It seems to have an indication that the activities ofthe two kinds of objects may be powered by differentmechanisms. Based on the results of a spectropolarimet-ric survey of the CfA and 12 µ m samples of Sy2s, Tran(2001) proposed the existence of a population of galac-tic nuclei whose activity is powered by starburst ratherthan accretion onto a supermassive black hole (SMBH)and in which, therefore, the BLRs simply do not exist(Nicastro et al. 2003). With respect to the radio, far-infrared, and near-infrared emissions of the two groups,Yu & Hwang (2005) found that an HBLR Sy2 is similarto a Sy1, suggesting that this type of object does harbora central AGN; on the other hand, the non-HBLR Sy2 is [email protected] National Astronomical Observatories, Chinese Academy ofSciences, 20A Datun Road, Beijing 100012, China. Graduate University of Chinese Academy of Sciences, 19AYuquan Road, Beijing 100049, China. Key Laboratory of Optical Astronomy, National Astronomi-cal Observatories, Chinese Academy of Sciences, Beijing 100012,China. more like a starburst galaxy.Considerable efforts have been devoted in the pastdecade to understanding the HBLR and non-HBLR Sy2s.The absence of PBLs could be attributed to the edge-on line of sight and hidden of electron scattering region(Heisler et al. 1997; Wang & Zhang 2007). Many evi-dences showed that the presence or absence of HBLRs inSeyfert 2 galaxies depends on the AGN luminosity, withthe HBLR sources having, on average, larger luminosities(Lumsden & Alexander 2001; Gu & Huang 2002; Martoc-chia & Matt 2002; Tran 2001, 2003; Nicastro et al. 2003).Examining the sample extracted from the spectropolari-metric survey of Tran (2001, 2003), Nicastro et al. (2003)found that all HBLR sources have accretion rates largerthan the threshold value of ˙ m ≃ − (in Eddingtonunits), while non-HBLR sources lie at ˙ m ≤ ˙ m thres . Col-lecting a sample of 90 Sy2s with radio, infrared, optical,and X-ray (2-10 keV) data, Gu & Huang (2002) indicatedthat the majority of non-HBLR Sy2s have less powerfulAGN activity, which is likely caused by a low accretionrate. Based on the observed upper limit of emission linewidth of 25,000 km s − , Laor (2003) also proposed amodel to describe the existence of BLRs in AGNs.Seyfert 2 galaxies have large columns of circumnuclearobscuring material that prevents the direct view of thenucleus. X-ray observations are useful for providing anindication of the level of obscuration by the torus. Oneusually uses the column density of neutral hydrogen ( N H )to show the obscuration. In the local universe, abouthalf of Sy2s are found to be Compton-thick sources with N H > cm − (Maiolino et al. 1998; Bassani etal. 1999; Risaliti et al. 1999). However, some Sy2sdo not show HBLRs in spectropolarimetric observationsand have column densities lower than 10 cm − in the X-ray observations (Panessa & Bassini 2002), which indeedchallenge the unified model (Bian & Gu 2007).About the nature of the power sources in HBLRand non-HBLR Sy2s, there are still some controversies. Wu et alMoreover, the reason that Sy2s with column densitieslower than 10 cm − do not show HBLRs is still un-clear. In this paper, therefore we firstly devote to distin-guish between HBLR and non-HBLR Sy2s in dominantmechanisms (AGNs or starbursts); then we investigateand discuss physical reasons of the absence of PBLs innon-HBLR Sy2s. We assume H =75 km s − Mpc − ,Ω M = 0 .
3, and Ω Λ = 0 . THE SAMPLE AND DATA
We collect multi-wavelength data for the large sampleof 120 Sy2s which consists of 71 non-HBLR Sy2s and 49HBLR Sy2s listed in Tables 1 and 2, respectively, includ-ing radio, far-infrared, infrared, optical, and X-ray (2-10keV) bands. The sample selection is mainly from Gu &Huang (2002), Tran (2003), Wang & Zhang (2007), andShu et al. (2007). According to the Sy2 classification ofTran (2003), we classify the two objects, NGC 5347 andNGC 5929, as the non-HBLR Sy2 sample. Except the 18objects in Table 5 of Wang & Zhang (2007), all other ob-jects of our sample have their spectropolarimetric obser-vations which are described in Appendix in detail. Withregard to the 18 objects, Wang & Zhang (2007) took thetwo criteria to classify unabsorbed Seyfert 2 galaxies intonon-HBLR Sy2s and HBLR Sy2s (see section 2 of Wang& Zhang 2007; 14 non-HBLR Sy2s and 4 HBLR Sy2s).To present the properties and the dominant mecha-nisms between the two groups, we calculate some pa-rameters, for example, far-infrared, infrared, 1.49 GHz,[O iii ] luminosities, and high excitation lines ratios ([Ne v ] /[Ne ii ] and [O iv ] /[Ne ii ] ) and introduce some ofthem as follows.In order to get more luminosities of different bands,we need to calculate the luminosity distances of someobjects. The luminosity distance can be shown as D L = (1 + z ) cH Z z [(1 + x ) Ω M + Ω Λ ] − . dx (Darling & Giovanelli 2002; Ballantyne et al. 2006).We calculate the luminosities of far-infrared andinfrared bands of most of sources in our sampleby the fluxes from either the published papers orthe NASA/IPAC Extragalactic Database (NED). Thefluxes and luminosities of far-infrared and infraredcan be shown to be: F fir = 1 . × − (2 . f + f )[Wm − ], L fir = 4 π C D L2 F fir [ L ⊙ ], F ir = 1 . × − (13 . f + 5 . f + 2 . f + f )[Wm − ], and L ir = 4 πD L2 F ir [ L ⊙ ], where the constant C is the correc-tion factor required to account principally for the extrap-olated flux longer than the Infrared Astronomical Satel-lite ( IRAS ) 100 µm filter (Sanders & Mirable 1996), andhere C=1.6; the fluxes for 25, 60, and 100 µ m are listedin Tables 1 and 2, while the 12 µ m fluxes (not appearingin Tables 1 and 2 for simplicity) are also selected fromthe same literatures or NED as the 25, 60, 100 µ m fluxes.Besides the luminosities of radio band (1.49 GHz) ofsome sources from literatures, we also obtain the lumi-nosities of other sources using L = 4 πD F , where L and F are the luminosity and flux of the radio band (1.49GHz) from NED. The 2-10 keV fluxes come directly fromsome published literatures.The [O iii ] luminosity could be taken as an indicator of the nuclear activity only after correction for extinction(Maiolino et al. 1998; Bassani et al. 1999; Gu & Huang2002). The luminosity of the extinction-corrected [O iii ] λ L [O III] = 4 πD F cor[O III] , where F cor[O III] is the extinction-corrected flux of [O iii ] λ F cor[O III] = F obs[O III] [ (H α/ H β ) obs (H α/ H β ) ] . where an intrinsic Balmer decrement (H α/ H β ) = 3 . v ] and [O iv ] because they arenot affected by photoionization of stars and because theyare generally among the brightest highly-ionized lines(Sturm et al. 2002). For examining starburst and AGNactivities, the [Ne v ] λ .
32 and [O iv ] λ .
89 lines arethe most useful single line diagnostics. Both lines arestrong in spectra of AGNs (Farrah et al. 2007), whilethey are weak in spectra of star-forming regions (Lutzet al. 1998). We consider diagnostics based on the fine-structure line ratios. In Tables 1 and 2, we list variousparameters both types of objects. RESULTS AND DISCUSSION
In sections 3.1 and 3.2, we will show the propertiesthat differ between the two groups and test if HBLRSy2s are dominated by AGNs, and if non-HBLR Sy2sare dominated by starbursts. In section 3.3, we employthe separation of Shu et al. (2007) who noted that non-HBLR Sy2s are divided into the luminous ( L [O III] > ergs s − ) and less luminous ( L [O III] < ergs s − )classes. We will investigate their differences in the ob-scuration between the two groups. We also discuss theirproperties and compare their obscuration with that ofHBLR Sy2s. Distributions of Main properties for Non-HBLRand HBLR Sy2s
In this section, we report the distributions ofseveral parameters for HBLR and non-HBLRSy2s, for example, the SMBH mass, redshift, N H , F − /F [O III] ( F HX /F [O III] ) ratio, K α iron-lineequivalent width (EW), and mid-infrared line ratios,both F [Ne V] /F [Ne II] and F [O IV] /F [Ne II] .In Figure 1.a, we show the distribution of SMBHmasses for HBLR and non-HBLR Sy2s. Since there arecensored data points (upper limits; densely shaded ar-eas) among non-HBLR Sy2s, we use the astronomicalsurvival analysis package (ASURV; Feigelson & Nelson1985) for statistical analysis. The distributions of theSMBH masses are different between HBLR and non-HBLR Sy2s. For the whole sample, the mean SMBHmass of HBLR Sy2s is larger than that of non-HBLRSy2s by the amount of 0.31, with a confidence level of96 .
41% (see Table 3).Figure 1.b shows the distribution of redshifts for HBLRand non-HBLR Sy2s. The distributions of redshifts aredifferent between HBLR and non-HBLR Sy2s. The meanvalue of log z of HBLR Sy2s is larger than that of non-he different dominant mechanisms in Sy2s 3
TABLE 1
The non-HBLR Sy2 sample
Name z M BH f f f [Ne ii ][Ne v ] [O iv ] L [O III] L . F HX ˙ M N H EW Reference(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16)ESO 428-G014 0.006 7.34 1.77 4.40 6.05 ... 82.9 ... 42.23 28.542 3.80 1.04 > < > < a ... ... 21.51 .... 7,1,1,7F03362-1642 0.037 ... 0.50 1.06 2.01 ... ... ... 41.62 29.370 ... 0.26 ... .... 3,3,5F04103-2838 0.117 ... 0.54 1.82 1.71 ... ... ... ... 30.539 0.38 ... ... .... 1,5,15F04210+0401 0.045 7.34 0.25 0.60 < < < >
24 2000 2,18,4,4,4,4F20210+1121 0.056 ... 1.40 3.39 2.68 ... ... ... 43.31 30.485 3.0 12.51 > <
130 1.34 20.643 .... 9,1,17,4,5,4,4Mrk 573 0.017 6.04 0.85 1.24 1.43 ... ... ... 42.39 29.133 1.2 1.50 > > <
321 3,3,19,3,5,4,4,4Mrk 1066 0.012 7.5 2.26 11.0 12.2 ... 17.8 ... 42.27 29.467 2.3 1.14 > < < ≤ a ≤ < < < <
120 0.40 22.00 .... 3,3,3,5,4,3NGC1241 0.014 < < < > < < < a ≤ a
150 0.0004 ≤ d b ,3,5,11,4,4NGC3147 0.009 8.64 1.08 8.40 29.96 ... ... ... 40.19 29.350 a ≤ < a < a ≤ ≤ > < < < < a < a a c ,7,1,11,7,11NGC4594 0.004 8.52 0.50 4.26 22.86 ... 0.3 ... 39.26 ... 19.0 0.001 21.23 .... 7,8,1,7,11,7NGC4698 0.003 7.43 < <
425 7,1,7,16,7,12NGC4941 0.004 6.34 0.46 1.87 4.79 ... 9.0 19.0 41.18 27.629 7.0 0.09 23.65 1600 3,3,17,3,5,11,3,11NGC5033 0.003 7.48 1.15 13.8 43.9 13.3 0.4 5.1 39.47 28.992 a > > < > b ,4,5,4,4,4NGC5283 0.01 7.14 0.089 0.132 0.751 ... ... ... 40.88 28.355 14.6 0.05 23.18 <
220 3,1,3,5,4,3,4NGC5347 0.008 7.3 0.96 1.42 2.64 3.0 ... 4.0 41.22 27.852 2.2 0.10 > < a d > < < < < a <
200 7,14,7,1,1,7,12UGC6100 0.03 8.26 0.202 0.574 1.50 ... ... ... 42.30 29.265 < Wu et al
TABLE 2
The HBLR Sy2 sample
Name z M BH f f f [Ne ii ] [Ne v ] [O iv ] L [O III] L . F HX ˙ M N H EW Reference(1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16)Circinus 0.001 ... 68.44 248.7 315.85 900.0 239.0 ... 40.92 ... 100.0 0.05 24.633 2250 1,1,4,11,4,4ESO273-IG04 0.039 ... 1.72 4.76 4.92 ... ... ... 42.48 ... ... 1.85 ... .... 1,4F00317-2142 0.027 8.08 0.56 3.85 8.42 ... ... ... 41.13 30.005 a ... 0.08 20.28 <
900 7,1,7,1,7,12F00521-7054 0.069 ... 0.80 1.02 < < > > < < a < <
70 7,1,7,1,11,7,11IC 3639 0.011 6.83 2.54 8.90 13.79 ... ... ... 41.89 29.265 0.80 0.48 > > > a a ... 0.004 ≤ > > < Notes : Col. (1): Source name; Col. (2): Redshift; Col. (3): Log of SMBHs masses in units of M ⊙ ; Col. (4), (5), and (6):Infrared flux (in Janskys) for 25, 60, and 100 µm ; Col. (7), (8), and (9): Flux (10 − W cm − ) for [Ne ii ] λ . v ] λ . iv ] λ .
89; Col. (10): Log of extinction-corrected [O iii ] λ − ; Col. (11): Log ofluminosity of radio for 1.49 GHz in ergs s − Hz − ; Col. (12): Fluxes of observed X-ray (2-10 keV) in units of 10 − ergs s − cm − for Sy2s; Col. (13): Accretion rates ( M ⊙ yr − ); Col. (14): Log of gaseous absorbing column density ( N H ) in units of cm − ;Col. (15): EW is the Fe k α equivalent width in eV. Col. (16): References (for cols. [3], [4]-[6], [7]-[9], [10], [11], [12], [14], and[15], respectively). a L . are the luminosities at 1.4 GHz. b
19 are references for 19, 1, and 19, respectively. c
19 is reference for 19 and 1, respectively. d EW are the equivalent width of Fe K α line at 6.7 keV. References : (1) NED; (2) Tran 2003; (3) Zhang & Wang 2006; (4) Shu et al. 2007; (5) Gu & Huang 2002; (6) Bian & Gu2007; (7) Wang & Zhang 2007; (8) Ho et al. 1997; (9) Wang et al. 2007; (10) Imanishi, M 2002; (11) Bassani et al. 1999; (12)Panessa & Bassani 2002; (13) Surace et al. 2004; (14) Sanders et al. 2003; (15) Teng et al. 2005; (16) Akylas & Georgantopolos2009; (17) Deo et al. 2007; (18) Farrah et al. 2007; (19) Tommasin et al. 2008; (20) Goulding & Alexander 2009; (21) Marconiet al. 2001. he different dominant mechanisms in Sy2s 5 N (a) log (M BH /M sun ) non-HBLR Sy2s N HBLR Sy2s -3.0 -2.5 -2.0 -1.5 -1.0 -0.502468101214161820 N (b) log z non-HBLR Sy2s -3.0 -2.5 -2.0 -1.5 -1.0 -0.50246810121416 N HBLR Sy2s (c) log (F HX /F [O III] ) -2 -1 0 1 2 30246810 HBLR Sy2s N -2 -1 0 1 2 3024681012 non-HBLR Sy2s N
20 21 22 23 24 25 26024681012 N (d) log (N H /cm -2 ) non-HBLR Sy2s
20 21 22 23 24 25 2602468 N HBLR Sy2s
20 21 22 23 24 25 26
20 21 22 23 24 25 26024681012
Fig. 1.—
Distributions of the mass of the black hole, redshift, F − /F [OIII] ratio, and column density of neutral hydrogen. Denselyand sparsely shaded areas denote the upper and lower limits, respectively. Wu et al
TABLE 3Summary of HBLR and non-HBLR Sy2s.
Parameters non-HBLR Sy2s p null HBLR Sy2sMean N (%) Mean N(1) (2) (3) (4) (5) (6)log( N H ) 22.96 ± ± a ± α d
32 1.83 544 ±
99 38 F HX /F [O III] ± ± M BH ± ± ± ± L . ± ± L FIR (10 L ⊙ ) 6.21 ± ± L IR (10 L ⊙ ) b ± ± f /f ± − ± M ± ± F [Ne V] /F [Ne II] ) 0.40 ± ± F [O IV] /F [Ne II] ) 1.03 ± ± L [O III] ± ± L bol /L Edd ) e -0.98 ± ± Note : Col.(1): Parameters; Cols.(2)-(3) and (5)-(6): For each sample of non-HBLR Sy2 and HBLR Sy2 galaxies,“Mean” is the mean value of the various parameters and N is the number of data points. Col. (4): the probability p null (in percent) for the null hypothesis that the two distributions are drawn at random from the same parent population.When there are censored data, we use Gehan’s generalized Wilcoxon test (hypergeometric variance) in ASURV. a EW(Fe) is the Fe K α equivalent width in eV. b We have removed 3 sources (NGC 1241, NGC 3362, and NGC 7682) because they have no detections in their 12 µm band. c Detections only. d An ASURV test does not give the value of α . e Here the Eddington ratios, L bol /L Edd are given by L bol = 3500 L [O III] and L Edd = 1 . × ( M BH /M ⊙ )ergs s − ,respectively. -5 -4 -3 -2 -1 0 1 2-0.20.00.20.40.60.81.01.21.4 HBLR Sy2s log . (M/M sun yr -1 ) non-HBLR Sy2s l og ( f /f ) Fig. 2.—
IR color f /f ratio vs. accretion rate (defined as ˙ M = L bol /ηc ) for HBLR and non-HBLR Sy2s, where η =0.1 is the accretionefficiency (Wang et al. 2007); the filled circles denote HBLR Sy2s and the open squares denote non-HBLR Sy2s. he different dominant mechanisms in Sy2s 7 N e V ( . ) / N e II ( . ) L (IR) non-HBLR Sy2s HBLR Sy2s non-HBLR Sy2s HBLR Sy2s O I V ( . ) / N e II ( . ) L (IR)
Fig. 3.—
IR luminosity L(IR) vs. [Ne v ] λ . ii ] λ .
81 ratio (left) and [O iv ] λ . ii ] λ .
81 ratio (right). The solid lineprobably shows the starburst or AGN dominated region. The filled circles denote HBLR Sy2s and the open squares denote non-HBLRSy2s.
HBLR Sy2s by the amount of 0 .
27. A Kolmogorov-Smirnov (K-S) test shows that the probability for thetwo samples to be extracted from the same parent pop-ulation is 2 . F HX /F [O III] , which is the ratio “ T ”, is agood indicator of obscuration (Bassani et al. 1999; Tran2003; here, [O iii ] fluxes have been corrected for extinc-tion, and X-ray fluxes have not been corrected for ab-sorption). Table 3 shows little difference in F HX /F [O III] between the two groups. An ASURV test shows thatthe probability for the two samples to be extracted fromthe same parent population is about 28 . T is 24 . ± .
49 for non-HBLR Sy2s and6 . ± .
55 for HBLR Sy2s.In Figure 1.d, the N H distributions between HBLR andnon-HBLR Sy2s are not significantly different (with aconfidence level of 53 . N H upper limits as themeasured values). Our N H distribution is consistent withthe results of Gu & Huang (2002), Tran (2003) and Shuet al. (2007). This may be explained by the followingreason: since the mean value of N H is 10 . ± . cm − for the less luminous non-HBLR Sy2s (see Table 4), theyweaken greatly the difference in N H between non-HBLRSy2s and HBLR S2ys (Shu et al. 2007). In section 3.3,we will find that N H has the significant differences amongthe luminous, less luminous non-HBLR Sy2s, and HBLRSy2s (see Table 4 and Figure 4).In Table 3, non-HBLR Sy2s are obviously larger interms of the mean value of EW(Fe) than HBLR Sy2s.An ASURV test shows that the difference between thetwo samples is present (at a level of 98 . . F [Ne V] /F [Ne II] and F [O IV] /F [Ne II] , can better distinguish HBLR from non-HBLR Sy2s (see Table 3). Table 3 shows the significant differences in the two ratios between the two groups. AK-S test displays that the probabilities for the two sam-ple to be extracted from the same parent population are0 .
17% and 0 . Starburst or AGN Domination in HBLR andNon-HBLR Sy2s
In this section, we test if HBLR Sy2s are dominatedby AGNs, and if non-HBLR Sy2s are dominated by star-bursts. Next we use two methods to demonstrate dif-ferent dominant mechanisms between non-HBLR andHBLR Sy2s.As mentioned in section 1, the AGN activity is the keyto understanding the differences between the two kindsof Sy2s. AGN luminosity comes from the disk accre-tion onto central SMBHs. [O iii ] λ L [O III] isa good indicator of the AGN activity for Type II AGNs(Kauffmann et al. 2003).The IRAS f /f flux ratio is not a good indica-tor of the inclination, but of the relative strength ofthe host galaxy and nuclear emission (Alexander 2001;Shu et al. 2007). It has been shown that HBLR Sy2shave smaller values of f /f ratio, compared to non-HBLR Sy2s (Heisler et al. 1997). In Table 3, the meanvalue of the f /f ratio is 5 . ± .
45 for non-HBLRSy2s and 2 . ± .
30 for HBLR Sy2s. The difference(at the 99 . f /f ra-tio denotes the relative strength of starburst and AGN Wu et al luminous non-HBLR Sy2s (a) log (F HX /F [O III] ) HBLR Sy2s N N -2 -1 0 1 2 3 N less luminous non-HBLR Sy2s luminous non-HBLR Sy2s N HBLR Sy2s N less luminous non-HBLR Sy2s (b) log (N H /cm -2 )
20 21 22 23 24 25 26 N Fig. 4.—
Distributions of the F − /F [OIII] ratio and column density of neutral hydrogen for the HBLR Sy2s, luminous ( L [O III] > ergs s − ) non-HBLR Sy2s, and less luminous ( L [O III] < ergs s − ) non-HBLR Sy2s. Densely and sparsely shaded areas denotethe upper and lower limits, respectively. emissions.In Figure 2, we show the correlation of the f /f ratioversus the accretion rate (defined as ˙ M = L bol /ηc ), withPearson’s correlation coefficient of − .
54 and a probabil-ity of < . µm or 60 µm .).These results show that they have a significant anticorre-lation. As the f /f ratio drops, the ˙ M value increases.HBLR Sy2s show the smaller f /f ratios and larger˙ M values, which may indicate a higher ratio of AGN-to-starburst activity in the SED; non-HBLR Sy2s showthe larger f /f ratios and smaller ˙ M values, whichmay indicate a lower ratio of AGN-to-starburst activityin the SED. Therefore, we suggest that the non-HBLRSy2s are dominated by starbursts, while the HBLR Sy2sare dominated by AGNs.We also use another diagnostic for examining starburstand AGN activities. Due to the intense star formation inthe nuclear region of many active galaxies, some fractionof the measured fluxes of low lying fine structure lines(excitation potential ≤
50 eV) will be produced by pho-toionization from stars rather than AGNs, while the highexcitation lines ([O iv ] , [Ne v ] ) show little or no contam-ination from possible starburst components (Sturm et al.2002). Genzel et al. (1998) found that [O iv ] /[Ne ii ] and[Ne v ] /[Ne ii ] are much higher in AGNs than in star-bursts, which can now be confirmed on a broader statis-tical basis. [O iv ] originates purely from the narrow lineregion (NLR) in AGNs. In a unified scheme, the NLR line luminosity should be independently orientated andbe a good tracer of AGN power, in particular when usingan extinction insensitive and modest excitation line like[O iv ] (Sturm et al. 2002). Since the ionization potentialof [Ne v ] λ .
32 is E ion = 97 . v ] is unlikely tobe strong in galaxies without an AGN (Voit 1992; Farrahet al. 2007). While [Ne ii ] is a fairly good tracer of hotstar emission in starburst activity. In AGNs, the [Ne ii ]from the NLR is more easily contaminated by starburstemission than the higher excitation [O iv ] line (Sturm etal. 2002).In Figure 3, the relations of infrared luminosity versusflux ratios of [Ne v ] λ . ii ] λ .
81 and [O iv ] λ . ii ] λ .
81 are shown. We can see in the tworegions of each plot that the upper region is primarilyHBLR Sy2s and the lower one is primarily non-HBLRSy2s. The [Ne v ] λ .
32 and [O iv ] λ .
89 lines are themost useful single line diagnostics (Farrah et al. 2007).As a result, we suggest that the non-HBLR Sy2s arestarburst-dominated, while the HBLR Sy2s are AGN-dominated.In Figures 2 and 3, we find that HBLR and non-HBLRSy2s clearly show the differences in the power sources.In Table 3, the differences in the accretion rate ( ˙ M = L bol /ηc ), f /f ratio, and two mid-infrared line ratio, F [Ne V] /F [Ne II] and F [O IV] /F [Ne II] , are significant. Sowe hold that non-HBLR Sy2s are starburst-dominated,while HBLR Sy2s are AGN-dominated. Physical Nature of the Various Obscuration
In this section, we will investigate differences in theobscuration and reasons of the absence of PBLs in lu-he different dominant mechanisms in Sy2s 9
TABLE 4Summary of luminous, less luminous non-HBLR Sy2s and HBLR Sy2s.
Parameters non-HBLR Sy2sA(S1) a HBLR Sy2s(S2) non-HBLR Sy2sB(S3) a p null (%)Mean N Mean N Mean N S1-S2 S2-S3 S1-S3(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)log N H ± ± ± b ± α c
21 544 ±
99 38 955 ± α c
11 0.05 97.97 2.91 F HX /F [O III] ± ± ± L [O III] ± ± ± − − log( L bol /L Edd ) d ± ± ± Note : Col.(1): Parameters. Cols.(2)-(3), (4)-(5), and (6)-(7): For each sample of the non-HBLR Sy2sA, HBLRSy2s, and non-HBLR Sy2sB, “Mean” is the mean value of various parameters and “N” is the number of data points.Col.(8): From the K-S or ASURV test of luminous non-HBLR Sy2s (S1) vs HBLR Sy2s (S2), the probability p null forthe null hypothesis that the two distributions are drawn at random from the same parent population. Col.(9): As incol (8), but for HBLR Sy2s (S2) vs less luminous non-HBLR Sy2s (S3). Col.(10): As in col.(8), but for luminous non-HBLR Sy2s (S1) vs less luminous non-HBLR Sy2s (S3). When there are censored data, we use Gehan’s generalizedWilcoxon test (hypergeometric variance) in ASURV. a Non-HBLR Sy2sA and Sy2sB indicate the luminous ( L [O III] > ergs s − ) and less luminous ( L [O III] < ergs s − ) non-HBLR Sy2s, respectively. b EW is the Fe K α equivalent width in eV. c An ASURV test does not give the value of α . d Here the Eddington ratio is the same as Table 3. minous and less luminous non-HBLR Sy2s, and comparethem with those of HBLR Sy2s.With regard to the obscuration between HBLR andnon-HBLR Sy2s, our result (see Table 3) and previousresults (Tran 2003; Gu & Huang 2002; Shu et al. 2007)all show little difference. This reason is that adding theless luminous Sy2s to the Sy2 sample weakens the dif-ference in obscuration found in the luminous Sy2 sample(Shu et al. 2007). Next we discuss the possible physicalexplanations of the absence of PBLs in the luminous andless luminous non-HBLR Sy2s, respectively.Table 4 shows that all differences in N H , EW(Fe), and F HX /F [O III] between the luminous non-HBLR Sy2s andHBLR Sy2s are significant. An ASURV test shows thatthe probabilities for the two samples to be extractedfrom the same parent population are 3 . . . . L [O III] between luminous non-HBLR Sy2s and HBLR Sy2s, andthe mean values of log( L [O III] / ergs s − ) are 41 . ± . . ± .
11, respectively. So we suggest that theobscuration is the key cause that makes PBLs weakeror nondetectable for the luminous non-HBLR Sy2s. Ourexplanation supports Shu et al. (2007)’s suggestion thatthe absence of PBLs in the luminous Sy2s arises from theobscuration. However, PBLs are not detected in most(24/28) of the less luminous Sy2s in our sample. Thereason is still unclear.To explore the natural reason of the absence of PBLsfor less luminous non-HBLR Sy2s, we analyse their ob-scuration: N H = 10 . ± . cm − and F HX /F [O III] =55 . ± .
94 (see Table 4), suggesting that the less lumi-nous non-HBLR Sy2s have the smaller obscuration thanthe luminous non-HBLR Sy2s or HBLR Sy2s . Their ob- Because the sample size of less luminous non-HBLR Sy2s with scuration seems to be close to that of face-on Sy1s. If thescaleheight of the scattering zone varies with the centralsource luminosity (Lumsden & Alexander 2001), the ab-sence of their PBLs may be due to either the small scale-height in the scattering region or the inexistence of BLRs.Since the less luminous non-HBLR Sy2s have very small L [O III] , their scattering screens may have smaller scalesthan those of the luminous non-HBLR Sy2s or HBLRSy2s. However, the obscuration of this type of objects isvery small and seems to be the same as that of the hostgalaxy. So we suggest that the invisibility of PBLs forless luminous non-HBLR Sy2s does not arise from theobscuration.The Eddington ratios of the less luminous non-HBLRSy2s are generally very small and their mean value is10 − . ± . (see Table 4). This is consistent with whatNicastro et al. (2003) argued, that at very low accretionrates, the clouds of BLRs would cease to exist. Sincethe obscuration of this type of objects is very small, akey factor in the absence of PBLs is the very low Ed-dington ratio rather than the obscuration. When theaccretion rate drops to extremely sub-Eddington val-ues, their central engines undergo fundamental changesand the BLR disappears (Ho 2008). Recently, Tran etal. (2010) suggested that the low-luminosity AGNs areprobably powered by radiatively inefficient, or advectiondominated accretion flow (ADAF), that intrinsically lackBLRs, as suggested observationally by e.g., Tran (2001,2003); Bianchi et al. (2008); Panessa et al. (2009); Shi etal. (2010), and inspired theoretically by Nicastro (2000);Laor (2003); Elitzur & Shlosman (2006); Elitzur & Ho(2009); Cao (2010).In Table 4, we find that non-HBLR Sy2s can be classi- EW(Fe) measurements is only 11 and NGC 3982 has an EW(Fe) of6310 eV, Table 4 shows almost no difference in EW(Fe) between lessluminous non-HBLR and HBLR Sy2s. However, the differences in N H and F HX /F [O III] are significant. So we can accept the result. L [O III] > ergs s − ) and lessluminous samples, when considering only their obscura-tion. In light of the above discussion, we hold that theinvisibility of polarized broad lines (PBLs) in the lumi-nous non-HBLR Sy2s depends on the obscuration; theinvisibility of PBLs in the less luminous non-HBLR Sy2sdepends on the very low Eddington ratio rather than theobscuration. CONCLUSION
We conclude that HBLR Sy2s are dominated by AGNs,and non-HBLR Sy2s are dominated by starbursts. Thisidea is supported by the evidences listed below: (1) com-pared with non-HBLR Sy2s, HBLR Sy2s have larger ac-cretion rates and smaller f /f ratio which may denotesthe relative strength of starbursts and AGN emissions;(2) HBLR Sy2s are intrinsically more powerful than non-HBLR Sy2s from the analysis of [Ne v ] λ .
32, [O iv ] λ .
89, and [Ne ii ] λ .
81, which are the useful singleline diagnostics for distinguishing AGN from starburstactivity.In addition, we find that the obscuration of less lumi-nous non-HBLR Sy2s is much smaller than that of lumi-nous non-HBLR Sy2s or HBLR Sy2s. We conclude thatin luminous non-HBLR Sy2s, the invisibility of PBLs is due to the obscuration (Shu et al. 2007); in less luminousnon-HBLR Sy2s, the invisibility of PBLs may not be dueto the scattering screen obscured by the obscuring ma-terial, but is very likely due to the very low Eddingtonratio and the BLRs are not exist.Although these results are from our large sample, weshould further consider sample completeness and have aslarge a sample size as possible. In the future, both morecomplete and unbiased sample of HBLR and non-HBLRSy2s and fine measurements in various bands will presentthe physical nature of non-HBLR and HBLR Sy2s.
ACKNOWLEDGMENTS referees for the careful reading of the manuscript andvery helpful comments. We thank Chen Hu, and Xin-Lin Zhou for helpful suggestions and discussions. We alsothank James Wicker, Ali Tanni, Ping-Yan Zhou, and WeiDu for polishing the English. This work was supportedby the Natural Science Foundation of China (NSFC)Foundation under grants 10933001 and 10778726, theNational Basic Research Program of China (973 Pro-gram) No.2007CB815404, and the Young ResearcherGrant of National Astronomical Observatories, ChineseAcademy of Sciences.
REFERENCESAkylas, A., & Georgantopoulos, I. 2009, A&A, 500, 999Alexander, D. M., et al. 1999, MNRAS, 303, L17Alexander, D. M., et al. 2000, MNRAS, 313, 815Alexander, D. M. 2001, MNRAS, 320, L15Antonucci, R. 1993, ARA&A, 31, 473Antonucci, R., & Miller, J. S. 1985, ApJ, 297, 621Ballantyne, D. R., Everett, J. E., & Murray, N. 2006, ApJ, 639,740Bassani, L., et al. 1999, ApJS, 121, 473Baum, S. A., et al. 2010, ApJ, 710, 289Bian, W., & Gu, Q. 2007, ApJ, 657, 159Bianchi, S., et al. 2008, ApJ, 385, 195Cao, X. 2010, ApJ, 724, 855Darling, J., & Giovanelli, R. 2002, ApJ, 572, 810Deo, R. P., et al. 2007, ApJ, 671, 124Elitzur, M., & Shlosman, I. 2006, ApJ, 648, L101Elitzur, M., & Ho, L. C. 2009, ApJ, 701, L91Farrah, D., et al. 2007, ApJ, 667, 149Feigelson, E. D., & Nelson, P. I. 1985, ApJ, 293, 192Genzel, R., et al. 1998, ApJ, 498, 579Goulding, A. D., & Alexander, D. M. 2009, MNRAS, 398, 1365Gu, Q., & Huang, J. 2002, ApJ, 579, 205Gu, Q., Maiolino, R., & Dultzin-Hacyan, D. 2001, A&A, 366, 765Haas, M., et al. 2007, A&A, 473, 369Heisler, C. A., Lumsden, S. L., & Bailey, J. A. 1997, Nature, 385,700Ho, L. C., et al. 1997, ApJS, 112, 315Ho, L. C. 2008, ARA&A, 46, 475Imanishi, M. 2002, ApJ, 569, 44Inglis, M. D., et al. 1993, MNRAS, 263, 895Kauffmann, G., et al. 2003, MNRAS, 346, 1055Kay, L. E., & Moran, E. C. 1998, PASP, 110, 1003Laor, A. 2003, ApJ, 590, 86Lumsden, S. L., Heisler, C. A., Bailey, J. A., Hough, J. H., &Young, S. 2001, MNRAS, 327, 459Lumsden, S. L., & Alexander, D. M. 2001, MNRAS, 328, L32Lumsden, S. L., Alexander, D. M., & Hough, J. H. 2004,MNRAS, 348, 1451 Lutz, D., et al. 1998, A&A, 333, L75Maiolino, R., et al. 1998, A&A, 338, 781Marconi, A., et al. 2001, ApJ, 549, 915Martocchia, A., & Matt, G. 2002, preprint (astro-ph/0210332)Miller, J. S., & Goodrich, R. W. 1990, ApJ, 355, 456Moran, E. C., et al. 2000, ApJ, 540, L73Moran, E. C., et al. 2001, ApJ, 556, L75Nagao, T., Taniguchi, Y., & Murayama, T. 2000, ApJ, 119, 2605Nicastro, F. 2000, ApJ, 530, L65Nicastro, F., Martocchia, A, & Matt, G. 2003, ApJ, 589, L13Oliva, E., et al. 1998, A&A, 329, L21Panessa, F., & Bassani, L. 2002, A&A, 394, 435Panessa, F., et al. 2009, MNRAS, 398, 1951Pernechele, C., et al. 2003, MNRAS, 338, L13Risaliti, G., Maiolino, R., & Salvati, M. 1999, ApJ, 522, 157Ruiz, M., Rieke, G. H., & Schmidt, G. D. 1994, ApJ, 423, 608Sanders, D. B., et al. 2003, AJ, 126, 1607Sanders, D. B., & Mirable, I. F. 1996, ARA&A, 34, 749Shi. Y., et al. 2010, ApJ, 714, 115Shu, X.-W., et al. 2007, ApJ, 657, 167Sturm, D., et al. 2002, A&A, 393, 821Surace, J. E., et al. 2004, AJ, 127, 3235Teng, S. H., et al. 2005, ApJ, 633, 664Tommasin, S., et al. 2008, ApJ, 676, 836Tran, H. D., Miller, J. S., & Kay, L. E. 1992, ApJ, 397, 452Tran, H. D. 1995, ApJ, 440, 565Tran, H. D. 2001, ApJ, 554, L19Tran, H. D. 2003, ApJ, 583, 632Tran, H. D.,, Lyke, J. E., & Mader, J. A. 2010, ApJ, 726, L21Tran, H. D. 2010, ApJ, 711, 1174Voit, G. M. 1992, ApJ, 399, 495Wang, J.-M., & Zhang, E.-P. 2007, ApJ, 660, 1072Wang, J.-M., et al. 2007, ApJ, 661, L143Young, S., et al. 1996, MNRAS, 281, 1206Yu, P.-C., & Hwang, C.-Y. 2005, ApJ, 631, 720Zhang, E.-P., & Wang, J.-M. 2006, ApJ, 653, 137 he different dominant mechanisms in Sy2s 11
APPENDIX
TABLE 5The non-HBLR Sy2 sample
Name Reference a Name Reference a Name Reference a Name Reference a Name Reference a (1) (2) (3) (4) (5) (6) (7) (8) (9) (10)ESO 428-G014 3 Mrk 334 24 NGC1685 3 NGC4501 5L NGC5695 3,5LF00198-7926 12 Mrk 573 5L NGC2685 10 NGC4565 10 NGC5728 3,4AF01428-0404 10 Mrk 938 5P NGC3031 10 NGC4579 10 NGC5929 5P,6K,12F03362-1642 5L Mrk 1066 2L NGC3079 5L NGC4594 10 NGC6251 10F04103-2838 4A Mrk 1361 12 NGC3147 16KT,26K NGC4698 16KT,26K NGC6300 13AF04210+0401 4A NGC676 10 NGC3281 3 NGC4941 3 NGC6890 3F04229-2528 4A NGC1058 10 NGC3362 5L NGC5033 10 NGC7130 12F04259-0440 12 NGC1143 12 NGC3393 17,11 NGC5128 18A NGC7172 19A,12F08277-0242 4A NGC1144 5P NGC3486 10 NGC5135 12,19A NGC7496 4AF10340+0609 3,8 NGC1241 5P NGC3660 5L NGC5194 12 NGC7582 19A,12F13452-4155 4A NGC1320 5L NGC3941 10 NGC5256 12 NGC7590 19AF19254-7245 14E NGC1358 3 NGC3982 5L NGC5283 5L NGC7672 2LF20210+1121 4A NGC1386 3 NGC4117 3 NGC5347 5L NGC7679 10F23128-5919 4A NGC1667 3,5L NGC4472 10 NGC5643 3 UGC6100 5LIC 5298 12 ... ... ... ... ... ... ... ...The HBLR Sy2 sampleName Reference a Name Reference a Name Reference a Name Reference a Name Reference a (1) (2) (3) (4) (5) (6) (7) (8) (9) (10)ESO273-IG04 4A F18325-5926 13L MCG-3-58-7 5P NGC 591 2L,3K NGC 5252 4A,15KF00317-2142 10 F20050-1117 10 MCG-5-23-16 13A NGC 788 25L NGC 5506 5,13AF00521-7054 4A F20460+1925 4A Mrk 3 2L NGC 1068 20L NGC 5995 12F01475-0740 5P F22017+0319 4A,5P Mrk 78 2L NGC 2110 15K NGC 6552 5PF02581-1136 5L F23060+0505 7 Mrk 348 2L NGC 2273 3K NGC 7212 1LF04385-0828 5LP IC 1631 10 Mrk 463E 2L,4A NGC 2992 13A NGC 7314 13AF05189-2524 4A IC 3639 12 Mrk 477 1L NGC 3081 3K NGC 7674 2L,4AF11057-1131 4A IC 5063 13A,23A Mrk 1210 1L NGC 3185 10 NGC 7682 5PF15480-0344 4A Circinus 9E,21A NGC 424 3C NGC 4388 4A Was 49b 1LF17345+1124 7 MCG-3-34-64 4A NGC 513 22L NGC 4507 3K ... ... Notes : Column 1, 3, 5, 7, and 9: source name; Column 2, 4, 6, 8, and 10: the corresponding reference of the spectropolarimetricobservations.. a Letters denote references that used the following telescope: C=CTIO (4 m), P=Palomar (5 m), K=Keck (10 m), L=Lick (3 m), S=Subaru(8.2 m), E=ESO (3.6), KT=Kitt (2.3 m), and A=AAT (3.9 m).