X-ray and infrared diagnostics of nearby active galactic nuclei with MAXI and AKARI
Naoki Isobe, Taiki Kawamuro, Shinki Oyabu, Takao Nakagawa, Shunsuke Baba, Kenichi Yano, Yoshihiro Ueda, Yoshiki Toba
aa r X i v : . [ a s t r o - ph . GA ] A ug Publ. Astron. Soc. Japan (2014) 00(0), 1–14doi: 10.1093/pasj/xxx000 X-ray and infrared diagnostics of nearby activegalactic nuclei with MAXI and AKARI.
Naoki I
SOBE , Taiki K
AWAMURO , Shinki O YABU , Takao N AKAGAWA ,Shunsuke B ABA , Kenichi Y
ANO , Yoshihiro U
EDA , & Yoshiki T OBA School of Science, Tokyo Institute of Technology, 2-12-1 Ookayama, Meguro, Tokyo152-8551, Japan Institute of Space and Astronautical Science (ISAS), Japan Aerospace Exploration Agency(JAXA), 3-1-1 Yoshinodai, Chuo-ku, Sagamihara, Kanagawa 252-5210, Japan Department of Astronomy, Kyoto University, Oiwake-cho, Sakyo-ku, Kyoto 606-8502, Japan Graduate School of Science, Nagoya University, Furo-cho, Chikusa-ku, Nagoya, Aichi464-8602, Japan Department of Physics, The University of Tokyo, 7-3-1, Hongo, Bunkyo-ku, Tokyo, 133-0033,Japan Research Center for Space and Cosmic Evolution, Ehime University, 2-5 Bunkyo-cho,Matsuyama, Ehime 790-8577, Japan Academia Sinica Institute of Astronomy and Astrophysics, PO Box 23-141, Taipei 10617,Taiwan ∗ E-mail: [email protected]
Received 2016 June 20; Accepted 2016 August 23
Abstract
Nearby active galactic nuclei were diagnosed in the X-ray and mid-to-far infrared wavelengths,with Monitor of All-sky X-ray Image (MAXI) and the Japanese infrared observatory AKARI,respectively. Among the X-ray sources listed in the second release of the MAXI all-sky X-raysource catalog, 100 ones are currently identified as a non-blazar-type active galactic nucleus.These include 95 Seyfert galaxies and 5 quasars, and they are composed of 73 type-1 and27 type-2 objects. The AKARI all-sky survey point source catalog was searched for their mid-and far-infrared counterparts at , , and µ m. As a result, 69 Seyfert galaxies in theMAXI catalog (48 type-1 and 21 type-2 ones) were found to be detected with AKARI. TheX-ray ( – keV and – keV) and infrared luminosities of these objects were investigated,together with their color information. Adopting the canonical photon index, Γ = 1 . , of theintrinsic X-ray spectrum of the Seyfert galaxies, the X-ray hardness ratio between the – and – keV ranges derived with MAXI was roughly converted into the absorption column density.After the X-ray luminosity was corrected for absorption from the estimated column density, thewell-known X-ray-to-infrared luminosity correlation was confirmed at least in the Compton-thinregime. In contrast, NGC 1365, only one Compton-thick object in the MAXI catalog, was foundto deviate from the correlation toward a significantly lower X-ray luminosity by nearly an orderof magnitude. It was verified that the relation between the X-ray hardness below 10 keV and X-ray-to-infrared color acts as an effective tool to pick up Compton-thick objects. The difference inthe infrared colors between the type-1 and type-2 Seyfert galaxies and its physical implicationon the classification and unification of active galactic nuclei were briefly discussed. c (cid:13) Publications of the Astronomical Society of Japan , (2014), Vol. 00, No. 0
Key words: galaxies: active — galaxies: Seyfert — X-rays: galaxies — infrared: galaxies
The unified picture of active galactic nuclei (e.g., Antonucci1993; Urry & Padovani 1995) invokes a supermassive blackhole accompanied by a mass accreting disk as their central en-gine. It is widely believed that they are surrounded by a parsec-scale torus filled with dust and gas clouds. However, the de-tailed physical condition and spatial configuration of these es-sential ingredients in the active galactic nuclei still remain oneof the important issues yet to be solved in the modern astro-physics.The active galactic nuclei are widely known as an X-rayemitter. Except for jet-dominated sources including blazars andBL Lacertae objects, their X-ray emission is thought to origi-nate in a hot corona above the accretion disk, where disk pho-tons are Comptonized (e.g., Sunyaev & Titarchuk 1980). Afterthe nuclear radiation is absorbed by the dust torus, it is re-emitted into the mid-to-far infrared (IR) wavelength. As a re-sult, the active galactic nuclei, without any significant jet con-tamination, typically exhibit a strong IR bump in their spec-tral energy distribution around – µ m (e.g., Sanders et al.1989; Antonucci 1993; Elvis et al. 1994; Elitzur 2008). Theseproperties make a combination of X-ray and IR observations anideal probe for the central region of the active galactic nuclei.Actually, a number of recent studies indicate a linear correlationbetween the absorption-corrected (or -unaffected) X-ray lumi-nosities and the observed IR ones among nearby active galac-tic nuclei with a moderate absorbing Hydrogen column densityof N H < ∼ cm − , regardless of their optical classification(Gandhi et al. 2009; Matsuta et al. 2012; Ichikawa et al. 2012).Such a correlation is supposed to prefer a so-called clumpy torusgeometry (e.g., Krolik & Begelman 1988), instead of a simpletorus model with a smooth and homogeneous dust distribution(e.g., Pier & Krolik 1993), because the latter model requests adeficit in the observed IR luminosity from obscured objects dueto self-extinction within the dust torus.Owing to tremendous progress in multi-wavelength obser-vations, active galactic nuclei that are deeply enshrouded by thedust (e.g., Ueda et al. 2007) have gradually been uncovered.Among such heavily obscured active galactic nuclei, those with N H > . × cm − are widely called as Compton-thickobjects. In spite of their astrophysical importance for variousreasons, such as the origin of the X-ray background radiation(Gilli et al. 2007; Ueda et al. 2014), the Compton-thick sourcesare rather elusive in optical observations because of severe dustextinction. Even hard X-ray surveys performed with the BurstAlert Telescope (BAT) onboard the Swift observatory is inferred to have eventually undercounted the Compton-thick objects bya factor of ∼ (Burlon et al. 2011), although hard X-ray pho-tons above 10 keV are expected to exhibit a high penetratingpower. In contrast, it is recently proposed that a combinationof X-ray and IR color information is very useful to select theheavily obscured population (Severgnini et al. 2015; Terashimaet al. 2015).Previous unbiased X-ray and IR diagnostics on nearby activegalactic nuclei are usually based on the hard X-ray survey above10 keV (Matsuta et al. 2012; Ichikawa et al. 2012). In con-trast, the – keV X-ray spectral information, obtained withrecent X-ray telescopes, such as Suzaku (Mitsuda et al. 2007),Chandra, and XMM-Newton, is one of the most standard andpowerful tools to investigate the properties of active galaxies. Itis important to compare results from the detailed spectral anal-ysis with those from all-sky X-ray surveys in the same energyrange, namely below 10 keV. However, the – keV samples ofactive galactic nuclei adopted for the X-ray-to-IR studies werefrequently constructed from a restricted sky field (e.g., Gandhiet al. 2009). Therefore, unbiased all-sky survey data below 10keV with a high sensitivity has been strongly requested.In the present study, we overcome such situation, by mak-ing use of the X-ray source catalog with the Monitor of All-sky X-ray Image (MAXI; Matsuoka et al. 2009) and the IR onewith the Japanese space IR observatory AKARI (Murakami etal. 2007). MAXI has continuously monitored all the X-ray skybelow 10 keV with the sensitivity higher than any other pre-vious all-sky X-ray survey missions. In the second release ofthe MAXI all-sky X-ray source catalog (hereafter the 2MAXIcatalog; Hiroi et al. 2013), more than 100 active galactic nu-clei, including Seyfert galaxies, quasars and blazars, are listed.Considering the size of the dust torus (typically a parsec scale,corresponding to the light-crossing time of a few years), theX-ray flux averaged over ∼ years, which is provide by the2MAXI catalog, is valuable rather than snap-shot data obtainedwith the pointing X-ray satellites. The hardness ratio within theMAXI energy range is regarded as a good indicator of the ab-sorption column density of the dust torus. The ability of MAXIfor studies of active galactic nuclei was demonstrated by severalauthors (e.g., Isobe et al. 2010; Isobe et al. 2015; Tachibana etal. 2016).The IR characteristics of the active galactic nuclei, pickedup from the 2MAXI catalog, were examined with the AKARIall-sky survey Point Source Catalog (AKARI/PSC). With thetwo IR instruments, the InfraRed Camera (IRC; Onaka et al.2007) and the Far-Infrared Surveyor (FIS; Kawada et al. 2007), ublications of the Astronomical Society of Japan , (2014), Vol. 00, No. 0 Table 1.
X-ray and IR properties of the active galactic nuclei, detected both with MAXI and AKARI ∗ . (1) (2) (3) (4) (5) (6) (7) (8) (9) (10)2MAXI Object Name F S F H HR f f f z typeJ0044 −
238 NGC 235A . ± . . ± . . ± . — ± — . Sy1J0048 +
320 Mrk 348 . ± . . ± . . ± . — ±
43 736 ±
60 0 . Sy2J0124 −
349 NGC 526A . ± . . ± . . ± .
04 141 ±
14 292 ± — . Sy1.5J0124 −
588 Fairall 0009 . ± . . ± . ± .
04 229 ±
20 440 ± — . Sy1J0228 +
314 NGC 931 . ± . ± . . ± .
06 349 ±
12 763 ±
48 2430 ±
64 0 . Sy1.5J0233 +
324 NGC 973 . ± . . ± . . ± . — ±
77 1704 ±
69 0 . Sy2J0334 −
360 NGC 1365 . ± . . ± . . ± .
07 2234 ±
38 5364 ±
42 80384 ± . Sy1.8J0342 −
214 ESO 548 − G081 . ± . . ± . . ± . — — ±
61 0 . Sy1J0424 −
571 1H 0419 − . ± . . ± . . ± . — ± — . Sy1.5J0433 +
054 3C 120 . ± . . ± . . ± .
04 203 ±
20 497 ±
68 1468 ±
85 0 . Sy1J0437 −
106 MCG − . ± . . ± . . ± . — — ±
111 0 . Sy1.2J0443 +
288 UGC 03142 . ± . . ± . . ± . — — ±
79 0 . Sy1J0453 −
752 ESO 033-G 002 . ± . . ± . . ± .
04 163 ± ±
15 646 ±
120 0 . Sy2J0516 −
001 Ark120 . ± . . ± . . ± .
04 252 ±
18 253 ± — . Sy1J0516 −
102 MCG-02-14-009 . ± . . ± . ± .
09 90 ± — ±
56 0 . Sy1J0518 −
325 ESO 362-18 . ± . . ± . . ± .
09 166 ±
31 366 ±
36 1277 ±
88 0 . Sy1.5J0523 −
363 PKS 0521 − . ± . . ± . . ± .
07 96 . ± . ± — . BL LacJ0552 −
073 NGC 2110 . ± . . ± . . ± .
02 300 ±
19 566 ±
30 4594 ±
47 0 . Sy2J0554 +
464 MCG + . ± . . ± . . ± .
03 340 ±
18 1283 ±
35 2377 ±
62 0 . Sy1.5J0615 +
712 Mrk 3 . ± . . ± . . ± .
17 322 ± ±
49 2939 ±
282 0 . Sy2J0640 −
257 ESO 490-IG026 . ± . . ± − . ± . — — ±
41 0 . Sy1.2J0740 +
497 Mrk 79 . ± . . ± . . ± . ± ±
38 1358 ±
64 0 . Sy1.2J0804 +
052 Mrk 1210 . ± . . ± . . ± .
14 274 ±
19 1310 ± ±
61 0 . Sy2J0808 +
757 PG 0804 + ± . . ± . − . ± .
07 118 ±
10 186 ± — . Sy1J0920 −
079 MCG − . ± . ± . . ± . — ± — . Sy2J0923 +
228 MCG + . ± . . ± . . ± .
09 78 ±
15 178 ± — . Sy1.2J0945 −
140 NGC 2992 . ± . . ± . ± .
11 299 ±
49 826 ±
54 9220 ±
194 0 . Sy2J0947 −
308 MCG − . ± . . ± . . ± .
01 384 ±
14 1391 ±
21 1277 ±
84 0 . Sy2J1000 −
313 2MASX J09594263 − . ± . . ± . . ± .
07 85 ± ± — . Sy1J1023 +
199 NGC 3227 . ± . . ± . . ± .
04 444 ±
71 1128 ±
44 10596 ±
535 0 . Sy1.5J1031 −
143 2MASSi J1031543 − . ± . . ± . . ± .
08 94 ± — — . Sy1J1105 +
724 NGC 3516 . ± . . ± . ± .
04 262 ±
20 651 ±
16 1317 ±
83 0 . Sy1.5J1139 −
376 NGC 3783 . ± . . ± . . ± .
03 502 ±
10 1530 ±
41 2716 ±
108 0 . Sy1J1203 +
445 NGC 4051 . ± . . ± . . ± .
05 346 ±
30 885 ±
42 4557 ±
254 0 . Sy1.5J1210 +
394 NGC 4151 . ± . . ± . . ± .
01 1032 ±
19 3629 ±
72 4594 ±
126 0 . Sy1.5J1217 +
073 NGC 4235 . ± . . ± . . ± . — — ±
50 0 . Sy1J1325 −
429 Centaurus A . ± . . ± . . ± .
01 10191 ± ± ± . Sy2J1335 −
341 MCG − . ± . . ± . . ± .
03 280 ±
32 591 ±
11 1035 ±
119 0 . Sy1.2J1338 +
046 NGC 5252 . ± . . ± . . ± . — — ±
120 0 . Sy1.9J1349 −
302 IC 4329A . ± . . ± . . ± .
01 769 ±
12 1790 ±
34 1785 ±
210 0 . Sy1.2J1413 −
030 NGC 5506 . ± . . ± . . ± .
02 823 ±
27 2240 ±
69 8413 ±
302 0 . Sy1.9J1417 +
253 NGC 5548 . ± . . ± . . ± .
03 157 ± ±
40 1073 ±
237 0 . Sy1.5J1419 −
265 ESO 511-G030 . ± . . ± . . ± . — — ±
150 0 . Sy1J1423 +
250 NGC 5610 . ± . . ± . . ± . ±
19 371 ±
41 5795 ±
205 0 . Sy2J1437 +
588 Mrk 817 . ± . . ± . ± .
08 188 ±
10 669 ±
27 1575 ±
60 0 . Sy1.5J1512 −
213 2MASX J15115979 − . ± . . ± . . ± .
08 95 ± — ±
89 0 . Sy1J1513 +
423 NGC 5899 ± . . ± . . ± . — — ±
184 0 . Sy2J1535 +
581 Mrk 290 . ± . ± . ± . — ± — . Sy1J1548 −
136 NGC 5995 . ± . . ± . ± .
07 325 ±
19 671 ± ±
329 0 . Sy2J1555 −
793 PKS 1549 − . ± . ± . . ± .
08 82 ± ±
45 925 ±
95 0 . Sy1J1614 +
660 Mrk 876 . ± . . ± . . ± .
08 82 ± ±
19 541 ±
39 0 . Sy1J1717 −
628 NGC 6300 . ± . . ± . ± . ±
24 1336 ±
97 14928 ± . Sy2J1834 +
327 3C 382 . ± . . ± . . ± .
04 120 ± — — . Sy1J1836 −
594 Fairall 0049 ± . . ± . . ± .
05 411 ±
20 920 ±
59 2619 ±
180 0 . Sy2J1838 −
654 ESO 103-035 . ± . . ± . . ± .
04 300 ±
25 1446 ±
12 1227 ±
77 0 . Sy2J1842 +
797 3C 390.3 . ± . . ± . . ± .
03 90 ±
10 242 ± — . Sy1J1845 −
626 Fairall 0051 . ± . . ± . . ± .
11 301 ± ±
62 1705 ±
180 0 . Sy1J1921 −
587 ESO 141-G055 ± . . ± . ± .
03 150 ± ± — . Sy1J1937 −
060 2MASX J19373299 − . ± . . ± . . ± . — ±
28 3709 ±
127 0 . Sy1.5J2009 −
611 NGC 6860 . ± . . ± . . ± .
06 155 ±
13 357 ±
61 1369 ±
73 0 . Sy1J2042 +
751 4C + . ± . . ± . . ± .
04 147 ± ± — . Sy1J2044 −
106 Mrk 509 . ± . . ± . . ± .
02 247 ±
18 499 ± — . Sy1.2J2115 +
821 2MASX J21140128 + . ± . . ± . . ± .
07 71 ± ± — . Sy1J2136 −
624 1RXS J213623.1 − . ± . . ± . . ± . — ± — . Sy1J2200 +
105 Mrk 520 . ± . . ± . . ± .
12 146 ±
12 320 ±
11 5040 ±
100 0 . Sy1.9J2201 −
317 NGC 7172 . ± . . ± . . ± .
05 316 ±
17 424 ±
44 8087 ±
218 0 . Sy2J2238 −
126 Mrk 915 ± . . ± . ± . — ± — . Sy1J2303 +
086 NGC 7469 . ± . . ± . . ± .
04 767 ±
17 2692 ±
60 27694 ± . Sy1.2J2304 −
085 Mrk 926 . ± . ± . . ± .
02 60 ± ±
36 647 ±
119 0 . Sy1.5J2318 +
001 NGC 7603 . ± . . ± . . ± .
07 295 ±
11 321 ±
12 1340 ±
104 0 . Sy1.5 ∗ Remarks:(1) Source name in the 2MAXI catalog.(2) Name of the optical counterpart.(3), (4) Soft ( – keV) and hard ( – keV) band X-ray flux in the unit of − ergs cm − s − (Hiroi et al. 2013)(5) X-ray Hardness ratio, defined as HR = ( F ′ H − F ′ S) / ( F ′ H + F ′ S) , where F ′ H and F ′ S is the hard and soft band fluxes in the Crab unit;i.e., F ′ H = F H / . × − − − and F ′ S = F S / . × − − − (see Hiroi et al. 2013) .(6) – (8) IR flux density at µ m, µ m and µ m in mJy of the AKARI counterpart (Ishihara et al. 2010; Yamamura et al. 2012)(9) Object redshift.(10) Optical Seyfert type (Hiroi et al. 2013; and reference therein). Publications of the Astronomical Society of Japan , (2014), Vol. 00, No. 0
Fig. 1.
Distribution of the separation angle between the optical position ofthe Seyfert galaxies in the 2MAXI catalog and their AKARI counterparts.The hatched and thick histograms indicate the distributions of the IRC andFIS counterparts, respectively. the AKARI/PSC widely covers the mid-to-far IR sources in the – µ m range, where the dust torus is expected to radiate asignificant fraction of its energy. Therefore, a combined use ofthe 2MAXI catalog and AKARI/PSC helps us to investigate thevicinity of the active galactic nuclei, and to reveal the nature ofthe dust torus. The 2MAXI catalog (Hiroi et al. 2013) was constructed fromall-sky X-ray survey data accumulated with the gas slit camera(Mihara et al. 2011; Sugizaki et al. 2011) onboard MAXI inthe first 37 months from 2009 September to 2012 October. Thecatalog lists 500 X-ray sources detected at high Galactic latitudeof | b | ≥ ◦ with a significance of > σ in the – keV range.Its σ detection sensitivity, . × − erg cm − s − in – keV for about half of the sky, is highest among those of previousall-sky X-ray surveys in the similar energy range. The X-rayflux of the faintest source in the catalog is measured as . × − erg cm − s − .Hiroi et al. (2013) cross-matched all the 2MAXI X-raysources to previous X-ray source catalogs, including the firstMAXI catalog (Hiroi et al. 2011), the meta-catalog of X-raydetected clusters of galaxies (Piffaretti et al. 2011), and theSwift/BAT 70-month catalog (Baumgartner et al. 2013). Asa result, they reported that sources,mostly due to the moderate position uncertainty of the 2MAXI Electrically available at http://vizier.cfa.harvard.edu/viz-bin/VizieR?-source=J/ApJS/207/36 . Table 2.
Statistics of Source Identification
Type Sy1 Sy2 Sy1+Sy2 N X ∗
68 27 95 N IR †
48 21 69 N ‡
36 16 52 N ‡
38 19 57 N ‡
30 20 50 ∗ Number of the Seyfert galaxies listed in the2MAXI catalog. † Number of the 2MAXI Seyfert galaxies,that have an AKARI counterpart in at leastone of the three photometric bands. ‡ Number of the 2MAXI Seyfert galaxies,detected at µ m ( N ), µ m ( N ) or µ m ( N ). catalog (90% error radius of ∼ . ◦ for σ sources; Hiroi et al.2013). Thus, the current completeness of source identificationto the 2MAXI catalog is about ∼ %. For the remaining 204sources, identification studies are ongoing (private communica-tion with the MAXI team).In the present study, all the active galactic nuclei tabulated inthe 2MAXI catalog were picked up; these consist of 95 Seyfertgalaxies, 5 quasars, and 15 blazars including BL Lacertae ob-jects. We noticed that all of these sources are listed in theSwift/BAT 70-month catalog. We refer to the 2MAXI cata-log (and references therein) for their source name, the 3-yearaveraged X-ray fluxes in the – keV and – keV ranges ( F S and F H , respectively), X-ray hardness ratio between these bands HR (its definition is discussed in § z , and object type (optical classification). The AKARI/PSC is electrically available at the AKARICatalogue Archive Server (Yamauchi et al. 2011) . The cata-log contains , mid-IR sources (Ishihara et al. 2010) and , far-IR ones (Yamamura et al. 2012), detected with theIRC and the FIS, respectively. The IRC is equipped with pho-tometric bands at the effective wavelength of λ = 9 µ m and µ m. The IRC sensitivity for an % detection completenessis 0.12 Jy and 0.22 Jy at µ m and µ m, respectively. TheFIS has photometric bands centered at λ = 65 µ m, µ m, µ m, and µ m. In the present study, we refer only to µ m FIS sources, since the sensitivity at this band (the % com-pleteness limit of 0.43 Jy) is nearly an order of magnitude betterthan those at the other 3 bands. The typical position accuracyof the IRC and FIS sources is ∼ ′′ and ∼ ′′ , respectively.A quality flag, indicating a reliability of the source detec-tion and flux determination, is assigned to each AKARI source http://darts.isas.jaxa.jp/astro/akari/cas.html ublications of the Astronomical Society of Japan , (2014), Vol. 00, No. 0 Fig. 2.
Distributions of the redshift z (panel a) and – keV X-ray flux F H (panel b). The thick histograms indicate the distributions for all the Seyfert galaxiesin the 2MAXI catalog, while the hatched histograms show the distributions for those with an AKARI counterpart. for the individual photometric bands. According to the recom-mendation from the AKARI team, we adopted only the AKARIsources with a quality flag of 3, which assures a high flux accu-racy. The same criterion was commonly adopted in similar stud-ies (e.g., Matsuta et al. 2012; Ichikawa et al. 2012; Terashima etal. 2015). We searched the AKARI/PSC for IR counterparts to the ac-tive galaxies selected from the 2MAXI catalog. For the sourceidentification, we referred to the optical position of the individ-ual 2MAXI sources, instead of their position determined withMAXI. A search radius of ′′ and ′′ was adopted for theIRC and FIS catalog, respectively, since these values nearly cor-respond to their σ position accuracy. The same angular thresh-old was widely imposed in previous studies (e.g., Matsuta et al.2012).The result of the source identification is summarized in ta-ble 1, where the X-ray properties ( F H , F S , and HR ) and IRflux densities at µ m, µ m, and µ m ( f , f , and f re-spectively) are tabulated for those listed in both the 2MAXI cat-alog and AKARI/PSC, together with their optical information.Figure 1 plots the distribution of the angular separation betweenthe 2MAXI sources and their AKARI counterparts. As shownwith the hatched histogram, the IRC counterparts are concen-trated at a relatively narrow range with a separation of < ∼ ′ .Due to the slightly worse angular resolution at the longer wave-length, the distribution of the FIS sources (up to ∼ ′ as in-dicated by the thick histogram) is wider than that of the IRCsources. These trends are qualitatively consistent with the pre-vious study by Ichikawa et al. (2012).No AKARI counterpart was found to the 5 quasars in the2MAXI catalog. This is reasonable because these quasars are a relatively high-redshift and faint source ( z > . and F H < . × − erg cm − s − ; see figure 2). Among 15 blazars,only one BL Lacertae object PKS 0521 −
36, corresponding tothe MAXI X-ray source 2MAXI J0523 −
363 with a hard andsoft X-ray flux of F H = (12 . ± . × − erg cm − s − and F S = (3 . ± . × − erg cm − s − respectively, was iden-tified with an AKARI source with µ m and µ m IR fluxdensities of f = 96 . ± . mJy and f = 216 ± mJy, re-spectively. Therefore, in the statistical argument below, we dealonly with Seyfert galaxies.Statistics of the source matching for the Seyfert galaxies be-tween the 2MAXI catalog and AKARI/PSC are summarizedin Table 2. Among the 95 2MAXI Seyfert galaxies, 69 ones( ∼ %) were successfully identified with an AKARI source atleast at one of the three photometric bands. The number of the2MAXI Seyfert galaxies detected at µ m, µ m, and µ mis N = 52 , N = 57 , and N = 50 , respectively. Out of the68 type-1 Seyfert (Sy1) galaxies, including those optically cate-gorized into Sy1.2 and Sy1.5 sources, ( ∼ %) are found tohave an AKARI counterpart ( N = 36 , N = 38 , and N = 30 ),while 21 out of 27 type-2 Seyfert (Sy2) galaxies ( ∼ %), in-cluding Sy1.8 and Sy1.9 objects, were detected with AKARI( N = 16 , N = 19 , and N = 20 ).The redshift distribution of the 2MAXI Seyfert galaxies isdisplayed in the panel (a) of figure 2. Their average redshiftwas evaluated as z ≃ . . When we focus on the objects withan AKARI counterpart, the average redshift reduces to z ≃ . .All the 2MAXI Seyfert galaxies located at z < . (41 ones)are detected with AKARI. In contrast, only 5 Seyfert galaxies(out of 12 ones) at z > . are found to have an AKARI coun-terpart, with PKS 1549 −
79 (Sy1) being the most distant one( z = 0 . ). Thus, our final sample is limited to local sourcesmainly at z < ∼ . .The – X-ray flux, F H , of the 2MAXI Seyfert galaxies Publications of the Astronomical Society of Japan , (2014), Vol. 00, No. 0
Fig. 3.
Relation between the observed X-ray and IR luminosities; (a) L H – L , (b) L H – L , (c) L H – L , (d) L S – L , (e) L S – L , and (f) L S – L . TheSy1 and Sy2 galaxies are plotted with the filled blue circles and open red diamonds, respectively. The logarithmic average of the X-ray to IR luminosity ratiofor the Sy1 galaxies is shown with the solid line in each panel, while the σ r and σ r ranges are indicated by the dashed and dash-dotted lines, respectively. is distributed as shown in the panel (b) of figure 2. The X-raybrightest source in the sample is Centaurus A, with an X-rayflux of F H = 3 . × − erg cm − s − averaged over the 3years. It is reasonable that the objects, that are not detected withAKARI, are mainly the fainter ones with an X-ray flux of F H < ∼ × − erg cm − s − . Above this flux threshold, out of objects ( ∼ %) were found to coincide with an AKARI source.Among the objects without any AKARI counterpart, the X-raybrightest one is the Sy1.9 galaxy 2MASX J09235371 − F H = 6 . × − erg cm − s − . In figure 3, the soft- and hard-band X-ray luminosities, L S and L H respectively, of the 69 Seyfert galaxies tabulated in both the2MAXI and AKARI catalogs, are plotted against their IR onesat µ m, µ m and µ m, L , L and L . The solid line oneach panel corresponds to the average of the X-ray-to-IR lumi-nosity ratio among the Sy1 galaxies evaluated in the logarithmicspace, while the dashed and dash-dotted lines show its σ r and σ r ranges, respectively, where σ r indicates the standard devi-ation of the logarithmic luminosity ratio. Here, the X-ray flux observed with MAXI was simply converted into the luminos-ity as L i = 4 πD F i ( i = H and S for the hard and soft bandrespectively). The source redshift z was transformed to the lu-minosity distance D L , by assuming the cosmological constantsof H = 71 km s − Mpc − , Ω M = 0 . , and Ω Λ = 0 . . Themonochromatic IR luminosity was calculated from the flux den-sity at each photometric band as L λ = 4 πD ν λ f λ , where ν λ isthe representative frequency of the AKARI photometric bands(i.e., λ = 9 µ m, µ m, and µ m). We neglected the so-called K -correction because the sample is limited to the low-redshiftsources, as shown in figure 2. Even for the most distant sourcein the sample, PKS 1549 −
79 located at z = 0 . , the effectof the K-correction on its luminosity is evaluated as less than afew %.Figure 3 suggests that the X-ray luminosities of the Sy1galaxies in the 2MAXI-AKARI sample linearly correlates withthe IR ones in the logarithmic space, while the correlation ap-pears to be rather vague for the Sy2 sources. In order to quantifysuch a correlation, we calculated the Spearman’s rank correla-tion coefficient ( ρ L ), between the X-ray and IR luminosities inthe logarithmic space. This coefficient was commonly adoptedfor similar studies (e.g., Matsuta et al. 2012; Ichikawa et al.2012). The result is summarized in table 3. A tight correlationwas confirmed for the Sy1 galaxies with ρ L ∼ . . In contrast, ublications of the Astronomical Society of Japan , (2014), Vol. 00, No. 0 Fig. 4.
Distributions of the absorption-inclusive X-ray to IR luminosity ratio in the logarithmic space; (a) log( L H /L ) , (b) log( L H /L ) , (c) log( L H /L ) ,(d) log( L S /L ) , (e) log( L S /L ) , and (f) log( L S /L ) . The distributions of the Sy1 and Sy2 galaxies are indicated with the thick blue and hatched redhistograms respectively. the X-ray luminosities of the Sy2 objects are found to exhibita moderate correlation to the mid-IR luminosities ( ρ L ≃ . – . to log( L ) and log( L ) ), and no meaningful one to thefar-IR luminosity ( ρ L ∼ to log( L ) ). After the Sy2 galaxiesare added to the Sy1 ones (indicated as Sy1+Sy2 in table 3),the coefficients to log( L ) and log( L ) still indicate a strongcorrelation ( ρ L > ∼ . ), while those to log( L ) decreased to amoderate value ( ρ L ∼ . ).We also evaluated the Spearman’s rank correlation coeffi-cients between the logarithms of the observed X-ray fluxes andIR flux densities ( ρ F ), and tabulate them in table 4. In the caseof flux-limited samples, the flux correlation is useful, since it isfree from artifacts due to the source redshift. A comparison be-tween the results in tables 3 and 4 pointed out that ρ F exhibits asimilar trend to ρ L , except for the fact that the tight luminositycorrelation suggested for the Sy1 and Sy1+Sy2 categories hasreduced to a moderate one in the flux space ( ρ F = 0 . – . ). Wethink that this is probably because the observed X-ray flux range(typically log( F H ) = − . – − . as shown in figure 2) is only comparable to the dispersion of the X-ray-to-IR flux/luminosityratio (see below).Figure 4 plots the distribution of the X-ray-to-IR luminosityratio for the 2MAXI-AKARI sample in the logarithmic space.The Sy1 galaxies are found to be distributed in a relativelynarrow range with σ r ≃ . – . (corresponding to a factor of – in the linear space). This range is found to be similarto those reported in the previous studies (e.g., Matsuta et al.2012; Ichikawa et al. 2012). In contrast, the observed X-ray lu-minosity of the Sy2 galaxies tends to be lower than that of theSy1 galaxies with a similar IR luminosity. This is clearly vi-sualized in figure 3 where the majority of the Sy2 galaxies arelocated below the solid line, indicating the logarithmic averageof the X-ray-to-IR luminosity ratio for the Sy1 galaxies. Thistrend is more prominent in the soft X-ray band, as is clearlyindicated in figure 4. As we discuss in § Publications of the Astronomical Society of Japan , (2014), Vol. 00, No. 0
Fig. 5.
The relation between the X-ray hardness HR and IR colors. In panels (a) and (b), HR is plotted against log( L /L ) and log( L /L ) , respectively.The filled blue circles indicate the Sy1 galaxies, while the open red diamonds point the Sy2 galaxies. Fig. 6.
Distributions of the IR colors among the 2MAXI-AKARI Seyfert galaxies. The panels (a) and (b) display the distributions of log( L /L ) and log( L /L ) , respectively. The thick blue and hatched red histograms indicate the distributions of the Sy1 and Sy2 galaxies, respectively. The openblue and filled red arrows on the top of the figure represent the mean IR color of the Sy1 and Sy2 galaxies, respectively, in the 9-month Swift/BAT sample,estimated from Ichikawa et al. (2012). In figure 5, the X-ray hardness ratio HR , representing theX-ray spectral color, is plotted against the IR colors L /L and L /L for the 2MAXI-AKARI Seyfert sample. TheX-ray hardness, taken from Hiroi et al. (2013), is defined as HR = ( F ′ H − F ′ S ) / ( F ′ H + F ′ S ) , where F ′ H = F H /F H , C and F ′ S = F S /F S , C represent the X-ray fluxes normalized to the Crab onesin the hard and soft bands, respectively ( F H , C = 1 . × − ergcm − s − and F S , C = 3 . × − erg cm − s − ). This meansthat a Crab-like X-ray spectrum with a photon index of Γ = 2 . and N H = 3 . × cm − (Kirsch et al. 2005) corresponds to HR = 0 . Figure 5 helps us to discriminate the Sy1 galaxies from theSy2 ones. On the color-color plots, the Sy1 objects tend to con-centrate in a relatively narrow area represented by < ∼ HR < ∼ . and | log( L /L ) | < ∼ . , or | log( L /L ) | < ∼ . . Comparedto the Sy1 sources, the Sy2 ones exhibit a rather hard X-rayspectrum, with a typical X-ray hardness of HR > ∼ . .We accumulated the IR-color distributions of the sample, asshown in figure 6. The arrows on the top of each panel in-dicate the IR colors, which were estimated from the averagespectral energy distribution of the Sy1 and Sy2 galaxies in the9-month Swift/BAT sample taken from Ichikawa et al. (2012).Thus, the IR color of the 2MAXI sample is consistent to that of ublications of the Astronomical Society of Japan , (2014), Vol. 00, No. 0 Fig. 7. (a) The line-of-sight hydrogen column density, N H , plotted as a func-tion of the X-ray hardness, HR , derived with MAXI. The model predictions,corresponding to the intrinsic photon index of Γ = 1 . , . , and . , are in-dicated with the dashed, solid and dash-dotted lines, respectively. (b) Theabsorbed-to-intrinsic flux ratio α abs in the soft (thin lines) and hard (thicklines) bands, plotted against HR . the 9-month Swift/BAT sample. We found no significant differ-ence in the mid-IR color, log( L /L ) , between the two Seyfertclasses, although there are a few relatively red Sy2 galaxies withMrk 3 being the reddest one with log( L /L ) = − . . In con-trast, the distribution of the mid-to-far IR color, log( L /L ) ,seems to rather differ between the Sy1 and Sy2 galaxies in thesense that the Sy2 galaxies tend to show a redder mid-to-far IRcolor with log( L /L ) < ∼ . From the K-S test, the proba-bility of the difference between the two source categories wasestimated as ∼ % HR The X-ray continuum from Seyfert galaxies in the – keVrange is thought to be basically dominated by the direct nuclearX-ray emission, which is absorbed by their dust torus, exceptfor Compton-thick sources (e.g., Gilli et al. 2007; Ueda et al.2014). Because X-ray photons in the soft band are more easilysubjected to absorption than those in the hard band, an objectwith higher absorption column density is predicted to exhibit aharder X-ray spectrum (at least in the Compton-thin regime).Therefore, the higher X-ray hardness observed from the Sy2 objects, which is clearly visualized in figure 5, is naturally as-cribed to the X-ray absorption.From the X-ray hardness ratio derived with MAXI, we canroughly estimate the absorbing column density of the dust torus(Ueda et al. 2011). Here, we simply assume that the dominantX-ray spectral component from the Seyfert nuclei in the MAXIenergy range is described with an absorbed power-law model.Panel (a) of figure 7 displays the relation of the line-of-sight hy-drogen column density, N H , to the MAXI hardness, HR , forsome representative values of the photon index ( Γ = 1 . , . ,and . ). This figure indicates that the HR value is sensitive tothe X-ray absorption in the range of N H = 10 – × cm − (i.e., in the Compton-thin regime). In panel (b) of figure 7, theratio of the absorbed X-ray flux to the intrinsic one, α abs , isplotted against HR . It is possible to estimate the absorption-corrected intrinsic flux/luminosity of the objects in our sam-ple, by dividing α abs into their flux/luminosity observed withMAXI.In order to validate the method for absorption correction,we here focus on Centaurus A, the brightest Seyfert (Sy2)galaxy in the sample. Based on the result presented in figure7 (a), the hardness ratio of this object measured with MAXI, HR = 0 . ± . , is converted into the column density of N H ≃ . × cm − . From a close examination on the 0.5– 300 keV wide-band X-ray spectrum of Centaurus A obtainedwith Suzaku in 2005, Markowitz et al. (2007) revealed that itsdominant power-law component in the 2 – 10 keV range, witha photon index of Γ = 1 . – . , is subjected to X-ray absorp-tion with a column density of N H = 1 . × cm − . A simi-lar N H value was recently derived in a coordinated observationwith NuStar and XMM-Newton performed in 2013 (F¨urst et al.2016). The hydrogen column density inferred from the X-rayhardness with MAXI agrees with these results. From the panel(b) of figure 7, the systematic uncertainty on α abs (and hencethe intrinsic luminosity estimate) due to the difference betweenthe MAXI and Suzaku results on N H is evaluated as at most ∼ %. By the method demonstrated in § L i , cor = L i /α abs ( i = H and S ). For the absorption correction, we adopted the typical photonindex of the Seyfert galaxies, Γ = 1 . (e.g., Ueda et al. 2014).No correction was performed to the sources with the hardnesssmaller than HR = 0 . (corresponding to α abs = 0 . and . in the soft and hard bands respectively), since we think that theabsorption put only a negligible impact on these sources. Therelation between the X-ray and the IR luminosities after the ab- Publications of the Astronomical Society of Japan , (2014), Vol. 00, No. 0
Fig. 8.
The absorption-corrected X-ray luminosities plotted against the observed IR luminosities; (a) L H , cor – L , (b) L H , cor – L , (c) L H , cor – L , (d) L S , cor – L , (e) L S , cor – L , and (f) L S , cor – L . The same symbol and color notations as for figure 3 are adopted, except for the open triangles indicatingNGC 1365. sorption correction is displayed in figure 8.After the absorption correction, we re-evaluated the X-ray-to-IR correlation coefficients, ρ L and ρ F , and summarize themin tables 3 and 4, respectively. It is found that the correctionput only a minor impact on both ρ L and ρ F , in comparison tothose before the correction. When we take all the Sy1 and Sy2objects into account, the absorption-removed X-ray luminosi-ties are found to highly correlate to the µ m and µ m mid-IRluminosities ( ρ L > ∼ . ), while their correlation to the µ mfar-IR luminosity is moderate ( ρ L ∼ . ). The flux correlationcoefficients of the X-ray to mid-IR wavelengths indicate a mod-erate correlation ( ρ F = 0 . – . ), while those to the far-IR bandcorrespond to a weak correlation ( ρ F ∼ . ).Figure 9 shows the histograms of the X-ray-to-IR luminos-ity ratio in the logarithmic space, after the X-ray absorption wascorrected. Thanks to the correction, the Sy2 galaxies have typ-ically moved into the σ r range of the Sy1 galaxies (i.e., theregions between the dash-dotted lines on figure 8). This resultis thought to reinforce the X-ray-to-IR luminosity/flux correla-tion throughout the Sy1 and Sy2 galaxies. For comparison, weestimated the X-ray-to-IR color of the 9-month and 22-monthSwift/BAT samples derived in the previous studies by Ichikawaet al. (2012) and Matsuta et al. (2012), respectively. The ratiosof the – keV X-ray luminosity to the IR ones and their er- rors, both of which are presented in Ichikawa et al. (2012) andMatsuta et al. (2012), were converted to match the MAXI en-ergy range, by assuming a power-law like X-ray spectrum witha photon index of Γ = 1 . . The horizontal arrows in figure 9indicate the σ r range of the X-ray-to-IR luminosity ratio, thusobtained, for the 9-month and 22-month Swift/BAT samples.Our result from the 2MAXI sample is found to be consistent tothe Swift/BAT results.A number of recent researches indicate that the intrinsic X-ray luminosity of Seyfert galaxies tightly correlates with theirobserved IR luminosities, irrespective of their optical classifi-cation (e.g., Gandhi et al. 2009; Matsuta et al. 2012; Ichikawaet al. 2012). The correlation is widely interpreted under theframework of the so-called clumpy torus geometry (e.g., Krolik& Begelman 1988), since the simple torus model with a smoothdust distribution infers a significant dust extinction to the IR lu-minosity for obscured (i.e., type-2) sources.Figures 8 and 9, together with the correlation coefficientssummarized in tables 3 and 4, support the X-ray-to-IR correla-tion, especially that to the mid-IR band. Basically, these meanthat our result strengthens the clumpy torus scenario. Most ofthe previous results were usually based on the hard X-ray sur-vey above 10 keV (Matsuta et al. 2012; Ichikawa et al. 2012),or a small sample below 10 keV derived in a restricted sky ublications of the Astronomical Society of Japan , (2014), Vol. 00, No. 0 Table 3.
Summary of the Spearman’s rank correlation coeffi-cients, ρ L , between the X-ray and IR luminosities. IR luminosityType X-ray luminosity log( L ) log( L ) log( L ) Sy1 log( L H ) 0 .
92 0 .
89 0 . L H , cor ) 0 .
91 0 .
89 0 . L S ) 0 .
92 0 .
89 0 . L S , cor ) 0 .
92 0 .
88 0 . Sy2 log( L H ) 0 .
53 0 . − . L H , cor ) 0 .
44 0 . − . L S ) 0 .
45 0 . − . L S , cor ) 0 .
44 0 . − . Sy1+Sy2 log( L H ) 0 .
89 0 .
82 0 . L H , cor ) 0 .
89 0 .
82 0 . L S ) 0 .
88 0 .
80 0 . L S , cor ) 0 .
88 0 .
82 0 . Table 4.
Summary of the Spearman’s rank correlation coeffi-cients, ρ F , between the X-ray flux and IR flux density. IR flux densityType X-ray flux log( f ) log( f ) log( f ) Sy1 log( F H ) 0 .
42 0 .
39 0 . F H , cor ) 0 .
42 0 .
39 0 . F S ) 0 .
47 0 .
43 0 . F S , cor ) 0 .
41 0 .
37 0 . Sy2 log( F H ) 0 .
53 0 . − . F H , cor ) 0 .
46 0 . − . F S ) 0 .
57 0 . − . F S , cor ) 0 .
44 0 . − . Sy1+Sy2 log( F H ) 0 .
42 0 .
33 0 . F H , cor ) 0 .
47 0 .
39 0 . F S ) 0 .
34 0 .
23 0 . F S , cor ) 0 .
46 0 .
38 0 . field (Gandhi et al. 2009). Our study successfully complementsthe previous ones, because the sample originates in the rela-tively unbiased all-sky X-ray survey below 10 keV, conductedby MAXI with the highest sensitivity.Figures 8 and 9 suggest that the dispersion of the ratio be-tween the X-ray luminosity to the far-IR one L ( σ r ≃ . forthe Sy1 galaxies) is slightly larger than those to the mid-IR ones, L and L ( σ r ≃ . ). In relation, the X-ray correlation coeffi-cients to log( L ) tends to be smaller than those to log( L ) and log( L ) . It is pointed out that within the AKARI angular res-olution non-negligible contribution to the far-IR emission fromwarm dust related to star formation activity in the host galaxyresulted in this larger scatter and lower correlation coefficientsat µ m (Matsuta et al. 2012). Interestingly, we found an outlier on figures 8 and 9, which islocated significantly outside the σ r region of the Sy1 galaxies.This object, indicated with the red open triangles in figure 8, is NGC 1365. Although this source is optically classified asSy1.8, it is known to exhibit occasionally a Compton-thick X-ray spectrum with N H > cm − (e.g., Risaliti et al. 2005;Risaliti et al. 2007; Risaliti et al. 2009a; Risaliti et al. 2009b).A standard population synthesis model by Ueda et al. (2014)which was constructed from the X-ray luminosity function ofactive galactic nuclei to reproduce the X-ray background spec-trum predicts the fraction of the Compton-thick objects as < %and ∼ % in the – keV and – keV bands, respectively,at the flux limit of our sample. The observed Compton-thickfraction (i.e., 1 out of 100) is in a reasonable agreement withthese predictions, given the fact that our sample is essentiallyselected by MAXI and the Swift/BAT (see § N H = 10 – cm − ). The X-ray lu-minosity below 10 keV of the reflected/scattered component isestimated to be by typically an order of magnitude less than theunabsorbed luminosity of the direct component. Therefore, forthe Compton-thick sources, the absorption correction based onthe HR – N H relation in figure 7 should yield a column densitylower than the real value, and thus, significantly underestimatestheir intrinsic luminosity. Using the hard X-ray catalog withSwift/BAT of nearby Seyfert galaxies, Matsuta et al. (2012)reported a sign of a deficit in the hard X-ray luminosity fromCompton-thick sources.The hardness ratio, HR = 0 . ± . , of NGC 1365 iswithin the range of those of the Sy1 galaxies, and thus indicatesa soft X-ray spectrum in the MAXI energy range. This HR value corresponds to a low column density of N H ∼ × cm − (in the case of Γ = 1 . ), in spite of its Sy2 nature. TheX-ray-to-IR luminosity ratio of NGC 1365 after the absorptioncorrection is by a factor of > ∼ lower than the average value ofthe Sy1 galaxies, as shown in figure 8. These two observationalfacts seem suggestive of the Compton-thick picture for NGC1365.From recent studies (Severgnini et al. 2015; Terashima et al.2015), the relation between the observed X-ray hardness andX-ray-to-IR color is proposed as an effective tool to pick upCompton-thick active galactic nuclei. In panels (a) and (b) offigure 10, we plot HR against the absorption-inclusive values of log( L H /L ) and log( L S /L ) , respectively, for the 2MAXI-AKARI sample. On these diagrams, the model prediction isplotted as a function of N H (from to × cm − ) with Publications of the Astronomical Society of Japan , (2014), Vol. 00, No. 0
Fig. 9.
Distributions of the logarithmic ratio between the X-ray and IR luminosities, after the X-ray absorption correction; (a) log( L H , cor /L ) , (b) log( L H , cor /L ) , (c) log( L H , cor /L ) , (d) log( L S , cor /L ) , (e) log( L S , cor /L ) , and (f) log( L S , cor /L ) The distributions among the Sy1 and Sy2galaxies are indicated with the thick blue and hatched red histograms respectively. The σ r range of the luminosity ratio for the Sy1 galaxies estimated fromMatsuta et al. (2012) and Ichikawa et al. (2012) are shown with the horizontal arrows indicated as ”M” and ”I”, respectively. the dotted lines. The dashed lines indicate the absorption col-umn of N H = 4 × , × , × , and × cm − .By referring to Terashima et al. (2015), we introduced to theabsorbed power-law model, an additional unabsorbed power-law one which represents the reflected/scattered component.The photon index of the reflected/scattered component was pre-sumed to be identical to the direct one (i.e., Γ = 1 . ; see § %, %, %, and % to theintrinsic luminosity of the direct component. The log( L H /L ) and log( L S /L ) values of a source unaffected by absorption(i.e., N H = 0 ) were assumed to be their mean among Sy1 galax-ies (i.e., represented by the solid lines in figure 3).In the Compton-thin regime ( N H < cm − ), a sourcewith a higher N H value is expected to exhibit a harder X-rayspectrum and a lower ratio of the soft X-ray to IR luminosities.In contrast, the hard-X-ray to IR color is relatively insensitive tothe absorption. For heavily obscured objects, an N H increase is predicted to result in a significant X-ray spectral softening and adecrease in log( L H /L ) . The log( L S /L ) value is inferred toremain unchanged because the direct X-ray emission in the softband has been fully blocked by the Compton-thick dust torus,and hence, is totally dominated by the scattered/reflected com-ponent. We clearly recognize in figure 10 that the behaviorof the 2MAXI-AKARI Seyfert galaxies, qualitatively followsthese trends. Especially, we can explain the colors of NGC 1365by tuning the fraction of the scattered/reflected component andthe X-ray-to-IR luminosity ratio for N H = 0 , both of which arethought to reflect the geometry of the dust torus around the ac-cretion disk. We have, thus, re-confirmed that a combination ofthe X-ray spectral hardness below 10 keV and the X-ray-to-IRcolor is useful to distinguish the Compton-thick sources fromthe Compton-thin ones, without any detailed spectral analysis,or hard X-ray data above 10 keV.Finally, it is important to mention that figure 10 implies afew additional Compton-thick candidates, which are located in ublications of the Astronomical Society of Japan , (2014), Vol. 00, No. 0 Fig. 10.
X-ray hardness ratio HR is plotted against the logarithm of the absorption-inclusive X-ray to IR luminosity ratios, log( L H /L ) and log( L S /L ) ,in panels (a) and (b) respectively. The blue filled circles and red open diamonds shows the Sy1 and Sy2 galaxies, respectively, while the red open triangleindicates NGC 1365. The model prediction estimated for some representative values of the fraction of the scattered component ( %, %, % and %) aredrawn with the dotted lines, where the log( L H /L ) and log( L S /L ) values for a source with no absorption ( N H = 0 ) are simply assumed to be theiraverage over the Sy1 galaxies in the sample. The dashed lines indicate the absorption column density of N H = 4 × , × , × , and × cm − . the transition area between the Compton-thin and -thick objects,represented by log( L H /L ) < ∼ − or log( L S /L ) < ∼ − (cor-responding to N H > × cm − ). Detailed X-ray spectralanalyses are considered to be required to find out the nature ofthese objects (e.g., Terashima et al. 2015), although we regardthem as out of the scope of the present paper. The mid-IR spectrum of Seyfert galaxies is utilized as a goodprobe into physical or geometrical properties of their dust torus,because it is known to be dominated by emission from the dusttorus especially in the – µ m range (Mor et al. 2009). It ispointed out that the – µ m IR spectra of nearby Sy2 galaxiesare possibly redder than those of the Sy1 objects (e.g., Mateoset al. 2016). Based on the numerical simulation on the clumpytorus model (e.g., Nenkova et al. 2008a; Nenkova et al. 2008b),these spectral characteristics are proposed to be explained by apossible idea that the covering factor of the dust torus is largerin the Sy2 sources than in the Sy1 ones (Mateos et al. 2016).However, our result shown in the left panel of figure 6 doesnot seem to give a support to such a scenario, since no cleardiscrepancies are found between the Sy1 and Sy2 sources.As briefly mentioned in § L /L color,in comparison to the Sy1 objects. We reconfirmed this trend(with a probability of ∼ % by the K-S test), by utilizing the2MAXI-AKARI sample, as shown in the right panel of figure6. In addition, no meaningful X-ray correlation to log( L ) in-dicated to the Sy2 galaxies (see tables 3 and 4) is possible tobe also related to a significant contamination from a higher star-formation activity in such objects. These results suggest thatthe optical Seyfert classification is not only determined by theorientation of the dust torus to our line of sight, but also by thewarm-dust distribution connected to the star-forming activity inthe host galaxy. Acknowledgments
We thank the anonymous referee for her/his supportive comments tomodify the present paper. This research is financially supported by theMinistry of Education, Culture, Sports, Science and Technology (MEXT)of Japan, through the Grant-in-Aid 24103002 (N.I.), 26247030 (T.N.) and26400228 (Y.U.). T.K. is supported by the Grant-in-Aid for JSPS Fellows.We made use of the MAXI catalog (Hiroi et al. 2013), which was con-structed from the MAXI data provided by RIKEN, JAXA and the MAXIteam. This work is based on the data taken with AKARI, a JAXA project,with the participation of ESA. In advance of publication, Dr. Ichikawakindly provided us with the IR data of the Swift/BAT sample. Publications of the Astronomical Society of Japan , (2014), Vol. 00, No. 0