arXiv:astro-ph/0206018v1 3 Jun 2002 Abstract. ofimtesgeto htalrefato fLLAGNs of fraction large loud. a radio that are and suggestion AGNs, luminosity more-luminous the X-ray of confirm those to with ob- radio LLAGNs heavily of the be ratio compare to ex- We likely than most AGN. are scured less galaxies is three find these ratio we that results, this multi-wavelength with objects, Comparing three pected. For by powered LLAGN. that predominantly are a indicate lines and emission AGNs, characteristic optical value luminous their more the and with LLAGNs agreement of good in are H jects to X-ray The sdtce nalglxe xetoeadterXrylu- X-ray their 5 and range one the in except nucleus are galaxies X-ray minosities all An spectrum in (2000). inverted detected al. or et is Nagar low-luminosity flat by 48 of a (LLAGN) survey AGN VLA with the in objects core con- radio the compact sample all Our of presented. sists are object HII- transition one and LINER galaxies, Re- Seyfert Emission-line low-luminosity three Nuclear gions), (Low-Ionization LINERs 11 of ohs nosue G eg,Tre ta.20) This 2001). al. et Turner known (e.g., 2s AGN LINER obscured few to an a interesting host only is are to there It currently 2000b). that al. note et Terashima (e.g., sources l oee yselrpoess n h uioiyratio proba- luminosity most the are and others L processes, while stellar by AGN, powered signatures an bly of clear to presence show likely the 2s is of LINER 2s LINER Some On of heterogeneous. 1997b). source be al. energy the et hand, Ho other 2000a; the Ptak & Ho, Terashima, 2001; erclmnst esthan are bolo- less spectra) a luminosity with optical (LLAGN) metric AGN nuclear low-luminosity their a by in powered lines Balmer broad .Terashima Y. Compact a with Core LINERs Radio of Observations Snapshot Chandra .. hs aaishvn ra H broad having 1s, (LINER galaxies LINERs various those 1 Fil- in i.e., type Ho, studies that shown (e.g., Extensive have galaxies 1997a). wavelengths bright Sargent nearby & many ippenko, (LINERs) in regions found are emission-line nuclear Low-Ionization Introduction 1. 1 2 3 edi acig eebr36 01es h olr .Komo S. Boller, Th. eds. 2001 3-6, December Garching, in held Workshop: the of Proceedings X nttt fSaeadAtoatclSine -- Yoshin 3-1-1 Science, Astronautical and Space of Institute srnm eatet nvriyo ayad olg Par College Maryland, of University Department, Astronomy pc eecp cec nttt,30 a atnDie B Drive, Martin San 3700 Institute, Science Telescope Space / L H α a eue odsrmnt ewe hs power these between discriminate to used be can h eut of results The 1 α , 2 uioiyrto o 1oto 4ob- 14 of out 11 for ratios luminosity n ..Wilson A.S. and Xrysetocp fANwt hnr n XMM-Newton” and Chandra with AGN of spectroscopy “X-ray Chandra ∼ × 10 10 38 42 nphtobservations snapshot α o8 to rss ergs 2 n osbyother possibly and , 3 × 10 − 1 41 H tal. et (Ho rss ergs s,S an .Knea P eot pp. Report, MPE Kunieda, H. Kahn, S. ssa, − 1 di aaiaa aaaa2981,Japan 229-8510, Kanagawa Sagamihara, odai, . lioe D228 USA 21218, MD altimore, ,M 04,USA 20742, MD k, ihaclm est averaging density of obscuration heavy column extension show a low-luminosity often with a which that 2s, simply indicate Seyfert not may luminous are LINERs in 2s AGN LINER obscured of paucity aais(oe l 97) sapltsuyo h X-ray the a of here study pilot report bright we a of LLAGNs, As of 1997a). survey LLAGNs properties al. These spectroscopic et 96 Mpc. Palomar (Ho all galaxies 19 of the survey of from radio distance come cm a 2 within VLA al. LLAGNs a et Nagar reported absorption. have by (2002) affected keV. less much 2 absorbed are above quency, band less X-ray of the favor uses in one X- if biased even or probably ones, lines are emission fluxes optical exam- through ray For selected important. objects finding be ple, against may LLAGNs biases Alternatively, obscured heavily 1997). al. et Turner oe(rm ie04s omnmz ffcso iep 1/2 pileup. of effects minimize sub-frame to 1/8 s) 0.4 in time observed (frame were mode objects Eight back-illuminated chip. ACIS-S3 the CCD 2002). with al. observed et were Nagar objects 2001; Ho & Ulvestad 2000; al. et cke iin2ojc.1 u ftee1 bet aebe ob- temperature been brightness have high objects tran- and 14 one ( VLBA these and the of 2, out with Seyfert 12 served one object. three 1s, 1s, 2 LINER Seyfert sition seven two of 2s, consists summarized sample LINER are The radio targets 1. Table wavelength The in literature. longer in with published comparison data (2000) al.’s et rmrdocr ( core radio (1997a). trum al. et Ho of survey spectroscopic from optical consists selected and the Seyferts objects low-luminosity LINERs eight between 48 and objects nuclei, of HII transition sample 18 by LINERs, Their observations 22 VLA (2000). of GHz al. 15 et the on Nagar based is sample Our Observations and Sample The 2. AGN. of classes other with with it nuclei compare galactic and sample the of (2002) 13 al’s Chandra detected et Nagar have of We galaxies, sample. 14 comprising subset, a of T b ncnrs,rdoosrain,priual thg fre- high at particularly observations, radio contrast, In h xouetm a yial w scec.Althe All each. ksec two typically was time exposure The eslce 4ojcssoigafltt netdspec- inverted to flat a showing objects 14 selected We > 10 eas xmn h rdoluns”o our of loudness” “radio the examine also We . 7 )rdocrswr eetdi l fte (Fal- them of all in detected were cores radio K) α − ≥ 0 . 3, S ν ∝ ν N α H codn oNagar to according ) ∼ 10 Chandra 23 cm − 2 survey (e.g., 2 Y. Terashima & A.S. Wilson: Chandra Snapshot Observations of LINERs with a Compact Radio Core

Table 1. The Sample. Seyferts, and QSOs presented in Terashima et al. (2000a) and Ho et al. (2001). This indicates that their optical emis- Class Name sion lines are predominantly powered by a LLAGN. LINER 1 NGC 266, 2787, 3226, 4143, 4203, 4278, 4579 The three objects NGC 2787, NGC 5866, and NGC L L LINER 2 NGC 3169, 4548, 6500 6500, however, have much lower X/ Hα ratios (log < Seyfert 1 NGC 4565, 5033 LX/LHα ∼ 0) than expected from the correlation, and Seyfert 2 NGC 3147 their X-ray luminosities are not enough to power the Hα Transition 2 NGC 5866 luminosities. This X-ray faintness could indicate one or more of several possibilities such as (1) an AGN is the power source, but is heavily absorbed at energies above sub-frame modes were used for three objects. Detailed re- 2 keV, (2) an AGN is the power source, but is currently sults are given in Terashima & Wilson (2002c). switched-off or in a faint state, and (3) the optical narrow emission lines are powered by some other source(s) than 3. Results an AGN. If an AGN is present in these X-ray faint objects and An X-ray nucleus is seen in all the galaxies except for NGC absorbed in the hard energy band above 2 keV, only scat- 5866. The X-ray luminosities corrected for absorption are tered and/or highly absorbed X-rays can be observed, × 38 × 41 −1 in the range 5 10 to 8 10 ergs s . The positions and then the intrinsic luminosity would be much higher of the X-ray nuclei coincide with the radio core positions than that observed. This can account for the low LX/LHα to within the positional accuracy of Chandra. ratios and high radio to X-ray luminosity ratios (νLν (5 Spectral fits were performed for relatively bright ob- GHz)/LX). If the intrinsic X-ray luminosities are about jects. The pileup effect for the three objects with the one or two orders of magnitude higher than those ob- largest count rate per frame (NGC 4203, NGC 4579, and served, as is often inferred for Seyfert 2 galaxies, LX/LHα NGC 5033) is serious and we did not attempt detailed and νLν (5 GHz)/LX become typical of LLAGNs. spectral fits. Instead, we use the spectra and fluxes mea- Additional lines of evidence which support the pres- sured with ASCA for these three objects (Terashima et al. ence of an AGN include the fact that all three of these 2002b and references therein) in the following discussions. galaxies (NGC 2787, NGC 5866, and NGC 6500) have We confirmed that the nuclear X-ray source dominates the VLBI-detected, sub-pc scale, nuclear radio core sources hard X-ray emission within the beam size of ASCA. (Falcke et al. 2000), a broad Hα component (in NGC A power-law model modified by absorption was applied 2787, and an ambiguous detection in NGC 5866; Ho et al. and acceptable fits were obtained in all cases. The photon 1997b), a variable radio core in NGC 2787, and a jet-like indices of the nuclear sources are generally consistent with linear structure in a high-resolution radio map at 5 GHz the typical values observed in LLAGNs (photon index Γ = with the VLBA (NGC 6500; Falcke et al. 2000). Only an . − . 1 6 2 0, e.g., Terashima et al. 2002a, 2002b), although upper limit to the X-ray flux is obtained for NGC 5866. errors are quite large due to the limited photon statistics. If an X-ray nucleus is present in this galaxy and its lumi- The two objects (LINER 2s NGC 3169 and NGC 4548) 23 −2 nosity is only slightly below the upper limit, this source show large absorption column density NH=1.1×10 cm N 23 − could be an AGN obscured by a column density H∼ 10 . × 22 2 − and 1 6 10 cm , respectively, while NGC 3226 is cm 2 or larger. If the intrinsic luminosity of the nucleus is N . × 21 −2 less absorbed ( H=9 3 10 cm ). Others have small much lower than the observed upper limit, an AGN would column densities which are consistent with ‘type 1’ AGNs. have to be almost completely obscured and/or the optical NGC 2787 has only 8 detected photons in the 0.5–8 keV emission lines powered by some other source(s). The opti- band and is too faint to obtain spectral information. cal classification (transition object) suggests the presence of an ionizing source other than an AGN. 4. Discussion 4.1. Power Source of LINERs 4.2. Obscured LLAGNs We test whether the detected X-ray sources are the power In our sample, we found at least two highly absorbed source of their optical emission lines by examining the lu- LLAGNs (NGC 3169 and NGC 4548). In addition, if minosity ratio LX/LHα. The Hα luminosities (LHα) were the X-ray faint objects discussed in the previous subsec- taken from Ho et al. (1997a) and corrected for the redden- tion are indeed AGNs, they are most probably highly 23 −2 ing estimated from the Balmer decrement for the narrow absorbed with NH> 10 cm . Among these absorbed lines. The X-ray luminosities (LX) in the 2−10 keV band, objects, NGC 2787 is classified as a LINER 1.9, NGC and corrected for absorption, were used. The resulting 3169, NGC 4548, and NGC 6500 as LINER 2s, and NGC LX/LHα ratios of most objects are in the range of AGNs 5866 as a transition 2 object. Thus, heavily absorbed > (log LX/LHα ∼ 1) and in good agreement with the strong LINER 1.9s/2s, of which few are known, are found in the correlation between LX and LHα for LLAGNs, luminous present observations demonstrating that radio selection is Y. Terashima & A.S. Wilson: Chandra Snapshot Observations of LINERs with a Compact Radio Core 3

show high absorption columns in their X-ray spectra. In this subsection, we study radio loudness by comparing ra- dio and hard X-ray luminosities. Since the unabsorbed > 23 −2 luminosity for objects with NH ∼ 10 cm (equivalent > to AV ∼ 50 mag for a normal gas to dust ratio) can be reliably measured in the 2–10 keV band, it is clear that re- placement of optical by hard X-ray luminosity potentially yields considerable advantages. In the following analysis, radio data at 5 GHz taken from the literature are used since fluxes at this frequency are widely available for various classes of objects. We used the radio luminosities primarily obtained with the VLA at ′′ ∼< 1 resolution for the present sample. High resolution VLA data at 5 GHz are not available for several objects. For four objects among such cases, VLBA observations at 5 GHz with 150 mas resolution are published in the lit- erature (Falcke et al. 2000) and are used here. For two objects, we estimated 5 GHz fluxes from 15 GHz data by assuming a spectral slope of α = 0 (cf. Nagar et al. 2001). Since our sample is selected based on the presence of a compact radio core, the sample could be biased to more radio loud objects. Therefore, we constructed a larger sam- ple by adding objects taken from the literature for which Fig. 1. Examples of Chandra spectra. (a) NGC 3169 and 5 GHz radio, 2–10 keV X-ray, and RO measurements are (b) NGC 4548 available. First, we introduce the ratio RX = νLν (5 GHz)/LX as a measure of radio loudness and compare the ratio with a valuable technique for finding obscured AGNs. Along the conventional RO parameter. The X-ray luminosity LX with heavily obscured LLAGNs known in low-luminosity in the 2–10 keV band (source rest frame), corrected for Seyfert 2s (e.g., NGC 2273, NGC 2655, NGC 3079, NGC absorption, is used. We examine the behavior of RX using 4941, and NGC 5194; Terashima et al. 2002a), our obser- samples of AGN over a wide range of luminosity, including vations show that at least some type 2 LLAGNs are sim- LLAGN, the Seyfert sample of Ho & Peng (2001) and PG ply low-luminosity counterparts of luminous Seyferts in quasars which are also used in their analysis. The X-ray which heavy absorption is often observed. Some LINER luminosities (mostly measured with ASCA) are compiled 2s (e.g., NGC 4594, Terashima et al. 2002a; NGC 4374, from the literature. Finoguenov & Jones 2001; NGC 4486, Wilson & Yang Fig. 2 compares the parameters RO and RX for the 2002) and low-luminosity Seyfert 2s (NGC 3147) show no Seyferts and PG sample. These two parameters correlate strong absorption. Therefore, the orientation dependent well for most Seyferts. Some Seyferts have higher RO val- unification scheme does not always apply to AGNs in the ues than indicated by most Seyferts. This could be a result low-luminosity regime. of extinction. Seyferts showing X-ray spectra absorbed by a column greater than 1022 cm−2 (NGC 2639, 4151, 4258, 4388, 4395, 5252, and 5674) are shown as open circles in 4.3. Radio Loudness of LLAGNs Fig. 2. At least four of them have larger RO than indi- Earlier studies have suggested that LLAGNs tend to be cated by the correlation. The correlation between log RO radio loud compared to more luminous AGNs based on the and log RX for the less absorbed Seyferts can be described spectral energy distributions of seven LLAGNs (Ho 1999) as log RO = 0.88 log RX + 5.0. According to this relation, and, for a larger sample, on the conventional definition the boundary between radio loud and radio quiet object of radio loudness RO = Lν (5 GHz)/Lν(B) (the subscript (log RO = 1) corresponds to log RX = −4.5. “O” stands for optical), with RO > 10 being radio loud The PG quasars show systematically lower RO values (Ho & Peng 2001). Ho & Peng (2001) measured the lumi- than those of Seyferts at a given log RX. For the former nosities of the nuclei by spatial analysis of optical images objects, log RO = 1 corresponds to log RX = −3.5. This obtained with HST to reduce the contribution from stellar apparently reflects a luminosity dependence of the shape light. A caveat in the use of optical measurements for the of the SED: luminous objects have steeper optical-X-ray −α definition of radio loudness is extinction, which will lead slopes αox = 1.4 − 1.7 (S ∝ ν ), where αox is often to an overestimate of RO. Although Ho & Peng (2001) measured as the spectral index between 2200 A and 2 used only type 1–1.9 objects, some objects of these types keV, while less luminous AGNs have αox =1.0 − 1.2 (Ho 4 Y. Terashima & A.S. Wilson: Chandra Snapshot Observations of LINERs with a Compact Radio Core

Fig. 3. X-ray luminosity dependence of RX = νLν (5 GHz)/LX for the present LLAGN sample, Seyfert galax- ies, and PG quasars. The boundary between “radio loud” and “radio quiet” objects (log Rx = −4.5) is shown as a Fig. 2. Comparison between RO = Lν (5 GHz)/Lν (B) horizontal dashed line. and RX = νLν (5 GHz)/LX for Seyferts and PG quasars. The conventional boundary between “radio loud” and “ra- dio quiet” objects (log RO = 1) is shown as a horizontal References dashed line. Falcke, H., Nagar, N. M., Wilson, A. S., & Ulvestad, J., S. 2000, ApJ, 542, 197 Finoguenov, A. & Jones, C. 2001, ApJ, 547, L107 1999). This is related to the fact that luminous objects Ho, L. C. 1999, ApJ, 516, 672 show a more prominent “big blue bump” in their spectra. Ho, L. C., et al. 2001, ApJ, 549, L51 Figure 8 of Ho (1999) demonstrates that low-luminosity Ho, L. C., Filippenko, A. V., & Sargent, W. L. W. 1997a, ApJS, objects are typically 1–1.5 orders of magnitude fainter in 112, 315 the optical band than luminous quasars for an given X-ray Ho, L. C., Filippenko, A. V., Sargent, W. L. W., & Peng, C. Y. 1997b, ApJS, 112, 391 luminosity. Ho, L. C., & Peng, C. Y. 2001, ApJ, 555, 650 The definition of radio loudness using the hard X-ray Nagar, N. M., Falcke, H., Wilson, A. S., & Ho, L. C. 2000, ApJ, flux (RX) appears to be more robust because X-rays are 542, 186 less affected by both extinction at optical wavelengths and Nagar, N. M., Falcke, H., Wilson, A. S., & Ulvestad, J. S. 2002, the detailed shape of the blue bump. Further, measure- A&A, submitted ments of nuclear X-ray fluxes are much easier than mea- Nagar, N. M., Wilson, A. S., & Falcke, H. 2001, ApJ, 559, L87 surements of nuclear optical fluxes, since in the latter case Terashima, Y., Ho, L. C., Iyomoto, N., & Ptak, A. F. 2002a, the nuclear light must be separated from the surrounding ApJ, in preparation starlight. Terashima, Y., Ho, L .C., & Ptak, A. F. 2000a, ApJ, 539, 161 Terashima, Y., Ho, L .C., Ptak, A. F., et al. 2000b, ApJ, 533, Fig. 3 shows the X-ray luminosity dependence of R . X 729 In this plot, the LLAGN sample discussed in the present Terashima, Y., Iyomoto, N., Ho, L. C., & Ptak, A. F. 2002b, paper is shown in addition to the Seyfert and PG samples ApJS, 139, 1 R M nuc used above. This is an “X-ray version” of the log O- B Terashima, Y. & Wilson, A. S. 2002c, ApJ, submitted plot (Fig. 4 in Ho & Peng 2001). Our plot shows that a Turner, M. J., et al. 2001, A&A, 365, L110 42 −1 large fraction of LLAGNs (LX< 10 ergs s ) are radio Turner, T. J., George, I. M., Nandra, K., & Mushotzky, R. M. loud. This is a confirmation of Ho & Peng’s (2001) finding. 1997, ApJS, 113, 23 Since radio emission in LLAGNs is likely to be dominated Ulvestad, J. S., & Ho, L. C. 2001, ApJ, 562, L133 by emission from jets (Nagar et al. 2001; Ulvestad & Ho Wilson, A. S. & Yang, Y 2002, ApJ, 568, 133 2001), these results suggest that, in LLAGN, the fraction of the accretion energy that powers a jet, as opposed to electromagnetic radiation, is larger than in more luminous Seyfert galaxies and quasars.

Acknowledgements. Y.T. is supported by the Japan Society for a Promotion of Science Postdoctoral Fellowship for Young Sci- entists. This research was supported by NASA through grants NAG81027 and NAG81755 to the University of Maryland.