Nobeyama Millimeter Interferometric HCN(1-0) and HCO+(1-0) Observations of Further Luminous Infrared Galaxies

0.88 MB·English

by  Masatoshi Imanishi

AstronomicalJournal
PreprinttypesetusingLATEXstyleemulateapjv.11/26/04
NOBEYAMA MILLIMETER INTERFEROMETRIC HCN(1–0) AND HCO+(1–0) OBSERVATIONS OF
FURTHER LUMINOUS INFRARED GALAXIES
Masatoshi Imanishi 1
NationalAstronomicalObservatory,2-21-1,Osawa,Mitaka,Tokyo181-8588, Japan
Kouichiro Nakanishi 1
NobeyamaRadioObservatory,Minamimaki,Minamisaku,Nagano, 384-1305,Japan
9
0
0 Yoichi Tamura 2
2 DepartmentofAstronomy,TheUniversityofTokyo,7-3-1Hongo,Bunkyo-ku, Tokyo113-0033, Japan

n and
a Chih-Han Peng
J
formerlyDepartmentofAstronomy,SchoolofScience, GraduateUniversityforAdvancedStudies,Mitaka,Tokyo181-8588

6 Astronomical Journal

] ABSTRACT
A
We report the results of interferometric HCN(1–0) and HCO+(1–0) observations of four luminous
G
infrared galaxies (LIRGs), NGC 2623, Mrk 266, Arp 193, and NGC 1377, as a final sample of our
. systematic survey using the Nobeyama Millimeter Array. Our survey contains the most systematic
h
interferometric, spatially-resolved, simultaneous HCN(1–0) and HCO+(1–0) observations of LIRGs.
p
Ground-basedinfraredspectraoftheseLIRGsarealsopresentedtoelucidatethenatureoftheenergy
-
o sources at the nuclei. We derive the HCN(1–0)/HCO+(1–0) brightness-temperature ratios of these
r LIRGs and confirm the previously discovered trend that LIRG nuclei with luminous buried AGN
t
s signaturesininfraredspectratendtoshowhighHCN(1–0)/HCO+(1–0)brightness-temperatureratios,
a as seen in AGNs, while starburst-classified LIRG nuclei in infrared spectra display small ratios, as
[
observed in starburst-dominated galaxies. Our new results further support the argument that the
1 HCN(1–0)/HCO+(1–0) brightness-temperature ratio can be used to observationally separate AGN-
v important and starburst-dominantgalaxy nuclei.
4 Subject headings: galaxies: active — galaxies: nuclei — galaxies: ISM — radio lines: galaxies —
0 galaxies: individual (NGC 2623, Mrk 266, Arp 193, and NGC 1377)
7
0
1. 1. INTRODUCTION optically. However, it is crucial to understand the ener-
0 Luminous infrared galaxies (LIRGs) radiate the bulk getic role of such optically elusive buried AGNs in LIRG
9 oftheir largeluminosities (L>1011L ) as infrareddust nuclei.
⊙
0 emission. Large infrared luminosities mean that: (1) lu- A starburst (nuclear fusion) and buried AGN (mass
: accretion onto a SMBH) have very different energy gen-
v minous energy sources are present, but are hidden be-
eration mechanisms. Specifically, while UV emission is
i hinddust;(2)energeticradiationfromthehiddenenergy
X predominant in a starburst, an AGN emits strong X-
sourcesisabsorbedbythe surroundingdust; and(3)the
ray emission in addition to UV. Additionally, in a nor-
r heated dust grains re-emit this energy as infrared ther-
a mal starburst, the stellar energy sources and dust are
mal radiation. To understand the nature of LIRGs, it is
spatially well mixed, so energy sources are spatially ex-
essential to unveil the hidden energy sources; namely, to
tended. Since the energy generationefficiency of the nu-
distinguish whether starbursts (= nuclear fusion inside
clearfusionreactionisonly∼0.5%ofMc2 (Misthemass
stars)aredominant,oractivegalacticnuclei(AGNs; ac-
ofmaterialusedinthenuclearfusionreaction),theemis-
tivemassaccretionontoacentralcompactsupermassive
black hole [SMBH] with >106M ) are also energetically sion surface brightness of a starburst is modest and has
⊙
both observational (Soifer et al. 2000) and theoretical
important.
(Thompson et al. 2005) upper limits (∼1013L kpc−2).
Unlike AGNs surrounded by toroidally shaped dust, ⊙
Ontheotherhand,inanAGN,themassaccretingSMBH
which are classified optically as Seyferts (Veilleux & Os-
is spatially very compact, and thus more centrally con-
terbrock 1987), concentrated molecular gas and dust in
centrated than the surrounding gas and dust. The high
LIRGnuclei(Sanders&Mirabel1996)caneasilybury(=
energy generation efficiency of an AGN (6–42% of Mc2;
obscure in virtually all directions) the putative compact
M is the mass of accreting material; Thorne 1974) can
AGNs, making AGN signatures very difficult to detect
generate large luminosities from a very compact region,
Electronicaddress: [email protected] producingaveryhighemissionsurfacebrightness(Soifer
1Department of Astronomy, School of Science, Graduate Uni- et al. 2000). These differences between an AGN and a
versityforAdvancedStudies,Mitaka,Tokyo181-8588 starburstcouldcreatedifferencesinthepropertiesofthe
2National Astronomical Observatory of Japan, 2-21-1 Osawa,
surrounding molecular gas and dust that may be distin-
Mitaka,Tokyo181-8588, Japan
2 Imanishi et al.
guishable based on observations at wavelengths of low temperatures ratios, without invoking luminous AGNs,
dust extinction. have been present. Examples include the high turbu-
Molecular gas emission lines in the millimeter wave- lence (Papadopoulos 2007) or high density (Imanishi
length range can be used effectively to investigate the et al. 2007b) of molecular gas. In fact, ULIRGs have
hidden energy sources of LIRG nuclei. First, dust ex- very dense, highly disturbed molecular gas compared to
tinction is very small. Second, theoretical calculations normal LIRGs with L < 1012L (i.e., non-ULIRGs)
IR ⊙
predictthatX-ray-andUV-emittingenergysourcesshow (Sanders & Mirabel 1996; Soifer et al. 2000, 2001).
different behaviors of molecular line emission due to dif- Observations of normal LIRGs (non-ULIRGs), with or
ferent chemical reactions (Meijerink et al. 2006). Fi- without buried AGN signatures, would help elucidate
nally, mid-infrared 10–20 µm emission is strong in an whether the high HCN(1–0)/HCO+(1–0) brightness-
AGN because of hot dust thermal emission heated by temperature ratios in observed ULIRGs are indeed
the high surface brightness energy source of an AGN. caused by luminous buried AGNs, or are simply due to
The strong mid-infrared emission could selectively en- different molecular gas properties in ULIRGs.
hance particular molecular line emission through an in- Inthispaper,wepresenttheresultsofinterferometric,
frared radiative pumping mechanism (Aalto et al. 1995; simultaneousHCN(1–0)andHCO+(1–0)observationsof
Garcia-Burillo et al. 2006; Guelin et al. 2007; Weiss et such normal LIRGs (non-ULIRGs), to investigate the
al. 2007; Aalto et al. 2007). In this paper, we focus HCN(1–0)/HCO+(1–0)brightness-temperatureratiosin
on HCN(1–0) and HCO+(1–0) lines, because an AGN the nuclear regions. This is the final paper of our series
could enhance HCN(1–0) emission through an increased of interferometric HCN(1–0) and HCO+(1–0) surveys of
HCN abundance (Lintott & Viti 2006) and/or infrared LIRGs using NMA (Imanishi et al. 2004, 2006b; Iman-
radiative pumping (Aalto et al. 1995; Garcia-Burillo et ishi & Nakanishi 2006; Imanishi et al. 2007b). Ancillary
al.2006;Weiss et al.2007). Infact, Kohno (2005)found ground-based infrared K- (1.9–2.5 µm) and L- (2.8–4.1
observationally that HCN(1–0) emission is stronger, rel- µm) band spectra are also utilized to better understand
ative to HCO+(1–0), in AGN-dominated nuclei than in the hidden energy sources of these millimeter-observed
starburst galaxies, demonstrating the potential of this LIRG nuclei. Throughoutthis paper, we adopt H = 75
0
HCN(1–0) and HCO+(1–0) based method for the pur- kms−1Mpc−1,Ω =0.3,andΩ =0.7,tobeconsistent
M Λ
pose of distinguishing between an AGN and a starburst. with our previously published papers.
This method may be effective even for a Compton thick
(N > 1024 cm−2) buried AGN, the detection of which 2. TARGETS
H
is very difficult with direct X-ray observations. WeobservedthefourLIRGs,NGC2623,Mrk266,Arp
HCN(1–0) and HCO+(1–0) observations of a large 193, and NGC 1377, primarily because they were non-
number of LIRGs have been made, using single-dish ra- ULIRGs with or without buried AGN signatures and so
diotelescopes(Gao&Solomon2004;Gracia-Carpioetal. wereappropriatesourcestoaddresstheabove-mentioned
2008;Kripsetal.2008). However,thebeamsizesofthese issue (§1), and because we estimated that we could de-
observations are so large (>25 arcsec) that severe con- tect HCN(1–0) and HCO+(1–0) emission with NMA in
tamination from spatially-extended star-forming emis- a reasonable amount of telescope time. Table 1 sum-
sion is unavoidable, possibly hampering the detection marizesthe infraredemissionproperties of these LIRGs.
of the AGN signatures of LIRG nuclei, where putative An angular scale of 1′′ corresponds to a physical size of
AGNs are expected to be present. Furthermore, LIRGs 0.12–0.53kpcattheredshiftsoftheseLIRGs(z =0.006–
oftenshowmultiple-nucleusmorphologieswithsmallsep- 0.028).
arations(<10 arcsec),for which spatially-resolvedstudy NGC2623(z=0.018)isaLIRG(L =1011.5L )with
IR ⊙
is not possible with single-dish radio telescopes. Fi- twoprominentmergingtailsalongthenorth-easternand
nally, the HCN(1–0) and HCO+(1–0) data were col- south-western directions (Rothberg & Joseph 2004). It
lected at different times under different weather condi- is optically unclassified(Keel 1984;Veilleux et al.1995);
tions,possiblyincreasinguncertaintyregardingtheirrel- noobviousAGNsignatureexistsintheopticalspectrum.
ative strengths due to the ambiguity of inter-calibration Itsinfrared5–35µmspectrumobtainedwithSpitzerIRS
of data. isdominatedbypolycyclicaromatichydrocarbon(PAH)
To overcomethese issues, we performed millimeter in- emission, as is usually observed in starburst galaxies
terferometric, spatially-resolved, simultaneous HCN(1– (Brandl et al. 2006). However, high spatial resolution
0) and HCO+(1–0) observations of LIRGs, using the ground-based infrared 10-µm imaging observations re-
Nobeyama Millimeter Array (NMA) (Imanishi et al. veal that the bulk of the 10-µm emission in NGC 2623
2004, 2006b; Imanishi & Nakanishi 2006; Imanishi et comes from compact nuclear cores with <400 pc, the
al. 2007b). In our study, only nuclear HCN(1–0) and surface brightness of which can be as high as ∼1013L
⊙
HCO+(1–0) emission was extracted to investigate their (Soifer et al. 2001). This value is close to the maximum
ratios in brightness temperature (∝ flux × λ2). The found in starburst phenomena (Soifer et al. 2000). The
sample contained mostly ultraluminous infrared galax- 2–10 keV X-ray emission from NGC 2623 is character-
ies (ULIRGs; LIR > 1012L⊙; Sanders et al. 1988a) ized by a hard spectrum, suggesting that the scattered
showingluminousburiedAGNsignaturesatotherwave- componentofX-rayemissionfromaComptonthick(N
H
lengths, and we found that their HCN(1–0)/HCO+(1– > 1024 cm−2) AGN dominates the observed 2–10 keV
0) brightness-temperature ratios were generally higher X-rayflux (Maiolino et al. 2003). The intrinsic AGN lu-
than starburst galaxies with LIR < 1012L⊙. This re- minosity is highly dependent on the unknown value of
sult supports the presence of luminous buried AGNs the scatteringefficiency,andcanbe highif the efficiency
in these observed ULIRGs. However, scenarios that islow. Theopticalnon-detectionofSeyfertsignaturesin
try to explain large HCN(1–0)/HCO+(1–0) brightness- NGC 2623 suggests that the putative AGN is a buried
HCN/HCO+ in four LIRGs 3
one, whose scattering efficiency is expected to be low extended flux ratio for the HCN(1–0) line is similar to
(Fabianetal.2002). Evansetal.(2008)performedmulti- or higher than that of the CO(1–0) line. Papadopoulos
wavelength observations of NGC 2623, and detected the (2007)found that the HCN(4–3) to HCN(1–0)flux ratio
high-excitation forbidden-emission line [NeV]14.3µm in ofArp193is muchlowerthanthatofotherLIRGs. The
the infrared,indicative of narrowline regionclouds pho- weak high-excitation HCN(4–3) line may suggest that
toionized by AGN radiation. The intrinsic AGN lumi- the dominant energy source is a spatially-extended dif-
nosity can be high because the narrow line regions are fuse one, rather than a compact high-emission surface-
expectedtobe under-developedinaburiedAGN(Iman- brightness one. So far, no obvious AGN signatures have
ishi et al. 2006a, 2007a). The high CO(1–0) flux at the been seen in the Arp 193observationaldata (Vega et al.
coreofNGC2623(Bryant&Scoville1999)suggeststhat 2008).
detection of nuclear HCN(1–0) emission is feasible. NGC1377(z=0.006)hasaninfraredluminosityofL
IR
Mrk 266 (NGC 5256; z = 0.028) is a LIRG (L = = 1010.1L , and so is not a LIRG in a strict sense. This
IR ⊙
1011.5L ), consisting of two main merging nuclei sepa- galaxy shows a deficit of 21-cm emission relative to in-
⊙
rated by ∼10 arcsec, a south-western (SW) and north- frared emission (Roussel et al. 2003). Although it is op-
eastern(NE)nucleus (Mazzarella&Boroson1993). Op- tically unclassified (Veilleux et al. 1995), and thus no
tical spectroscopy by various groups suggested that the AGN signatures are presentin the opticalspectrum, the
SW nucleus shows clear Seyfert 2 signatures but the NE infrared 3–25 µm spectrum is dominated by a PAH-free
one does not (Osterbrock & Dahari 1983; Kollatschny continuum (Imanishi 2006; Roussel et al. 2006), typi-
& Fricke 1984; Mazzarella & Boroson 1993; Wu et al. cal of AGN-dominated galaxies. The 9.7-µm and 18-
1998; Goncalves et al. 1999; Ishigaki et al. 2000). How- µmsilicatedustabsorptionfeaturesareextremelystrong
ever, the most systematic optical spectral classification (Roussel et al. 2006), suggesting that the putative AGN
ofgalaxiesclassifiedtheSWandNEnucleiofMrk266as is deeply buried in dust. NGC 1377 is included in our
LINERandSeyfert2,respectively(Veilleux etal.1995). sample to investigate whether even a buried AGN can-
High spatial resolution X-ray data obtained with Chan- didate with low absolute infrared luminosity can have a
dra found that the X-ray spectrum of the NE nucleus high HCN(1–0)/HCO+(1–0) brightness-temperature ra-
is characterized by a heavily-obscured AGN, while that tio. This inclusion is important to test whether the
of the SW nucleus consists of an obscured AGN and a generaltrendofhighHCN(1–0)/HCO+(1–0)brightness-
strong starburst component (Brassington et al. 2007). temperature ratios found in ULIRGs with buried AGN
The detection of a strong 3.3-µm PAH emission feature signatures (Imanishi et al. 2006b, 2007b) is indeed due
(Imanishi 2002), a good starburst indicator (Moorwood to AGN-related phenomena, or simply the result of ex-
1986;Imanishi&Dudley2000),supportsthepresenceof treme moleculargaspropertiesin ULIRGs (see §1). The
a strong starburst in the SW nucleus. The radio 8.44- CO(1–0)emissionisdetected withsingle-dishradiotele-
GHz (3.55 cm) to infrared (1–1000µm) luminosity ratio scope observations (Roussel et al. 2003). If a significant
of Mrk 266 is a factor of 2–3 higher than other LIRGs, fraction of the CO(1–0) emission comes from the nu-
suggestingthepresenceofaradio-intermediate(orradio- clearregions,andunlessthenuclearHCN(1–0)/CO(1–0)
loud) AGN (Condon et al. 1991). The radio emission brightness-temperature ratio is extremely low, then we
of the NE nucleus is dominated by a spatially compact estimate that detection of nuclear HCN(1–0) emission is
(<0′.′3)componentandismuchbrighterthanthatofthe feasible.
SWnucleus (Condonetal.1991). The infrared5–35µm
3. OBSERVATIONSANDDATAREDUCTION
spectrum, takenwith Spitzer IRS, is dominated by large
equivalent-width PAH emission, which is typical of star- 3.1. Millimeter Interferometry
burst galaxies (Brandl et al. 2006). However, we note We performed interferometric observations of LIRGs,
that the SW and NE nuclei are not sufficiently covered at HCN(1–0) (λ = 3.3848 mm and ν = 88.632
rest rest
with SL (5.2–14.5 µm) and LL (14–38 µm) spectra, re- GHz) and HCO+(1–0) (λ = 3.3637 mm or ν =
rest rest
spectively (Brandl et al. 2006), possibly missing AGN 89.188 GHz) lines, using the Nobeyama Millimeter Ar-
signatures, given the lack of full Spitzer spectral cover- ray(NMA) atthe NobeyamaRadio Observatory(NRO)
age for both nuclei. between 2006 December and 2008 March. For Mrk 266
Arp 193 (z=0.023) is a LIRG (LIR = 1011.6L⊙) with and NGC 1377, since no interferometric CO(1–0) maps
a long narrow tail along the south-eastern direction hadbeen published, we conducted NMA observationsat
(Rothberg & Joseph 2004). It is classified optically CO(1–0) (λ = 2.6026 mm or ν = 115.271 GHz).
rest rest
as a LINER (Veilleux et al. 1995), showing no obvi- Table 2 summarizes the detailed observing log. The
ous optical AGN signatures. The mid-infrared 8–20 NMA consists of six 10-m antennas and observations
µm dust emission is dominated by a spatially-extended were undertaken using the AB (the longest baseline was
component with no prominent compact core (Soifer et 351 m), C (163 m), and D (82 m) configurations.
al. 2001). The emission surface brightness is ∼2 × The backend was the Ultra-Wide-Band Correlator
1012L⊙ (Soifer et al. 2001), which is within the range (UWBC) (Okumura et al. 2000) which can cover 1024
of starburst phenomena. The HCN(1–0) and HCO+(1– MHz with 128 channels at 8-MHz resolution. For
0) fluxes of Arp 193, measured with single-dish radio HCN(1–0)andHCO+(1–0)observations,the centralfre-
telescopes, are high (Solomon et al. 1992;Gracia-Carpio quencyforeachsource(Table2)wassettocoverboththe
et al. 2008). Millimeter interferometric observations re- redshifted HCN(1–0) and HCO+(1–0) lines simultane-
vealedthattheCO(1–0)emissionshowsastrongnuclear ously. A bandwidth of 1024 MHz corresponds to ∼3500
compact component (Downes & Solomon 1998; Bryant km s−1 for the redshifted HCN(1–0) and HCO+(1–0)
& Scoville 1999). Hence, we expect that detection of lines at ν ∼ 86–89 GHz and ∼2700 km s−1 for the red-
nuclear HCN(1–0) emission is feasible if the nuclear to shifted CO(1–0) lines at ν ∼ 112–115 GHz. The fields
4 Imanishi et al.
of view at 86–89 GHz and 112–115 HGz are ∼77′′ and bands.
∼62′′ at full-width at half-maximum (FWHM), respec- Photometric conditions persisted throughout the ob-
tively. The Hanning window function was applied to re- servations. The seeing at K was measured in the range
duce side lobes in the spectra, so that the effective reso- 0′.′6–0′.′8 in FWHM. A standard telescope nodding tech-
lutionwaswidenedto16MHzor54kms−1 (42kms−1) nique(ABBApattern)withathrowof7′.′5wasemployed
at ν ∼ 86–89 GHz (112–115GHz). alongtheslit. Eachexposurewas15sec,andtwocoadds
The UVPROC-II package developed at NRO (Tsut- were made at each position. The telescope tracking was
sumi et al. 1997) and the AIPS package of the National monitored with the SpeX infrared slit-viewer. Table 4
Radio Astronomy Observatory were used for standard summarizes the detailed observing log.
data reduction. Corrections for the antenna baselines, Appropriatestandardstars(Table 4), with anairmass
band-passproperties,andtime variationin the visibility difference of <0.1 for individual LIRGs, were observed
amplitude andphasewereappliedto allofthe data (Ta- to correct for the wavelength-dependent transmission of
ble 2). After discarding a fractionof the data with large Earth’s atmosphere. The K- and L-band magnitudes of
phase scatter due to poor millimeter seeing and clipping thestandardstarswereestimatedbasedontheirV-band
a small fraction of data with unusually high amplitude, (0.6 µm) magnitudes, and V −K and V −L colors, of
the data were Fourier-transformed using a natural UV the corresponding stellar types (Tokunaga 2000). The
weighting. The flux calibrationwasmade using observa- estimated K-band magnitudes agree very well with the
tions of Uranus or appropriate quasars whose flux levels 2MASS measurements (Skrutskie et al. 2006).
had been measured at least every month in the NMA Standarddatareductionprocedureswereemployedus-
observing seasons (Table 2). A conventional CLEAN ingIRAF3.Initially,framestakenwithanA(orB)beam
method was applied to deconvolvethe synthesizedbeam were subtracted from frames subsequently taken with a
pattern. Table 3 summarizes the total net on-source in- B(orA)beam,andtheresultingsubtractedframeswere
tegration times and synthesized beam patterns. added and divided by a spectroscopic flat image. Then,
bad pixels and pixels hit by cosmic rays were replaced
3.2. Ground-based Infrared K- (1.9–2.5 µm) and L- with interpolated values from surrounding pixels. Fi-
(2.8–4.1 µm) band Spectroscopy
nallythespectraofLIRGnucleiandstandardstarswere
To put stronger constraints on the hidden energy extracted by integrating signals over 1′.′8–2′.′7, depend-
sources at these millimeter-observed LIRG nuclei, we ingonactualsignalprofiles. Wavelengthcalibrationwas
performedground-basedinfraredK-(1.9–2.5µm)andL- performed based on standard star spectra which reflect
(2.8–4.1 µm) band spectroscopy of NGC 2623,Mrk 266, thewavelength-dependentnatureofEarth’satmospheric
and Arp 193, as their L-band spectra were not available transmission curve with high signal-to-noise ratios. The
in the literature. NGC 1377 was not observed, because LIRGnucleispectraweredividedbytheobservedspectra
its L-band spectrum was given by Imanishi (2006). of standard stars, and then multiplied by the spectra of
Inshort,throughinfraredL-bandspectroscopy,wecan blackbodieswithtemperaturesappropriatetoindividual
investigate the energy sources based on the equivalent standard stars (Table 4).
widthofthe3.3µmPAHemissionandtheopticaldepths Flux calibration was done based on signals of LIRGs
of absorption features at λ ∼ 3.05 µm by ice-covered andstandardstarsdetectedinsideourslitspectra. Since
rest
dust grains and at λ ∼ 3.4 µm by bare carbonaceous we employed a narrow 0′.′8-wide slit, our spectra are
rest
dust grains (Imanishi & Dudley 2000; Imanishi & Mal- sensitive only to spatially-compact emission, and miss
oney 2003; Imanishi et al. 2006a). A normal starburst a spatially-extended component. To reduce data point
galaxy should always show large equivalent-width 3.3 scatter,appropriatebinningofspectralelementswasper-
µm PAH emission, while a pure AGN produces a PAH- formed.
free continuum (Imanishi & Dudley 2000; Imanishi et
al. 2006a). For absorption features, optical depths have 4. RESULTS
upper limits in a normal starburst in which stellar en-
4.1. Millimeter Interferometric Data
ergysourcesanddustarespatiallywellmixed, while the
depths can be arbitrarily large in a buried AGN with For NGC 2623, Mrk 266, and Arp 193, our spectra at
a more centrally concentrated energy source geometry the HCN(1–0) or HCO+(1–0) emission peaks show that
than the surrounding dust (Imanishi & Maloney 2003; the flux levels between these lines are above zero, indi-
Imanishi et al. 2006a). In the infrared K-band spectra, catingthatcontinuumemissionispresent. Wecombined
emission from stars and AGNs is distinguishable using data points that are unaffected by these lines and made
CO absorption features at λ = 2.3–2.4 µm because interferometric maps of the continuum emission. Figure
rest
the CO absorption features are produced by stars older 1 presents the contours of the continuum emission for
than 106 yr, but not by AGN-heated hot dust emission. NGC 2623, Mrk 266, and Arp 193, in the vicinity of the
Infrared K- (1.9–2.5 µm) and L- (2.8–4.1 µm) band redshiftedHCN(1–0)and HCO+(1–0)emissionat86–89
spectra were taken using SpeX (Rayner et al. 2003) at- GHz. For Arp 193, the continuum emission (∼15 mJy)
tached to the IRTF 3-m telescope atop Mauna Kea, is clearly (>8σ) detected. For NGC 2623, and Mrk 266
Hawaii, on 2008 Apr 19 and 20 (UT). The 1.9–4.2 µm SWandNE,signs ofcontinuumemissionaremarginally
cross-dispersed mode was employed, so that K- and L- seen (∼3σ). Continuum emission was not detected in
band spectra were obtained simultaneously. We chose the HCN(1–0)/HCO+(1–0) data of NGC 1377, or the
a narrow 0′.′8-wide slit to pinpoint the infrared contin-
uumemissionpeaksofindividualLIRGnucleiwherethe 3IRAFisdistributedbytheNationalOpticalAstronomyObser-
vatories, operated by the Association of Universities for Research
putative AGN may be located. The resulting spectral
in Astronomy, Inc. (AURA), under cooperative agreement with
resolution using this slit is R ∼ 1000 in the K- and L- theNationalScienceFoundation.
HCN/HCO+ in four LIRGs 5
CO(1–0) data of Mrk 266 and NGC 1377. maps in Fig. 2. The values estimated in this way agree
Figure 2 displays integrated intensity maps of the withthosebasedontheGaussianfitstowithin30%. Ta-
HCN(1–0)andHCO+(1–0)emissionoftheobservedfour ble 5 also includes CO(1–0) fluxes at the nuclear peaks,
LIRGs. Figure 3 presents the interferometric maps of derived from our NMA data (Mrk 266 and NGC 1377)
CO(1–0) emission for Mrk 266 and NGC 1377. Since or taken from the literature (NGC 2623 and Arp 193).
LIRGscontainhighlyconcentratednuclearmoleculargas TheCO(1–0)emissionatMrk266SWclearlydisplays
(Sanders & Mirabel 1996), the fraction of high-density spatially-extended structures (Fig. 3). The total CO(1–
molecular gas (n > 104 cm−3) is expected to increase 0) flux, including this spatially-extended component, is
H
(Gao & Solomon 2004). Such molecular gas is better estimated to be 162 Jy km s−1.
probedwith HCN(1–0)and HCO+(1–0)rather than the CO(1–0) fluxes of Mrk 266 and NGC 1377, and
widely used CO(1–0) because the dipole moments of HCN(1–0)andHCO+(1–0)fluxes ofArp 193,havebeen
HCN and HCO+ (µ > 3 debye) are much larger than measured using single-dish radio telescopes (Sanders et
CO (µ ∼ 0.1 debye; Botschwina et al. 1993; Millar et al. 1986;Roussel et al. 2003; Gracia-Carpioet al. 2008).
al. 1997). Hence, the maps in Fig. 2 reflect the spatial Table 6 compares these single-dish measurements with
distribution of the dense molecular gas phase. our NMA interferometric data. For Mrk 266 CO(1–0)
Mrk266hasadouble-nucleusmorphology. TheCO(1– emission, our interferometric flux measurements provide
0) emission is sufficiently bright to investigate its veloc- good agreement, within ∼15%, to old single-dish mea-
ity information for each individual nucleus, as shown in surements. ForArp193,ourHCN(1–0)fluxismorecom-
Figure 4. In this CO(1–0) channel map (Fig. 4), the parable to the latest IRAM 30m measurement (Gracia-
NE nucleus shows a CO(1–0) emission peak at lower Carpio et al. 2008), but is a factor of two smaller than
frequency (112.08–112.11 GHz) than the SW nucleus the old IRAM one (Solomon et al. 1992), whose high
(112.17–112.19GHz), suggesting that the NE nucleus is flux measurement was noted by Gracia-Carpio et al.
more redshifted than the SW nucleus. Previously ob- (2008). For Arp193 HCO+(1–0) emission, our measure-
tained optical spectra also show that the NE nucleus is ment agrees with the latest IRAM 30m data (Gracia-
more redshifted than the SW nucleus (20–130 km s−1) Carpioetal.2008). Fortheselines,thebulkoftheemis-
(Mazzarella&Boroson1993;Kollatschny&Fricke1984; sion should be covered in our interferometric data. For
Kim et al. 1995). the nearest source, NGC 1377, however, our CO(1–0)
Figure 5 shows spectra around the HCN(1–0) and flux measurement is more than a factor of four smaller
HCO+(1–0) lines at the nuclear positions of the ob- than the SEST telescope measurement, suggesting that
served four LIRGs. Spectra around the CO(1–0) line our interferometric data miss a spatially very extended
for Mrk 266 SW and NE, and NGC 1377 are presented component, and detect only nuclear compact molecular
in Figure 6. For NGC 1377, HCN(1–0) emission might gas.
be marginally detected in the integrated intensity map
4.2. Ground-based Infrared Spectra
close to the CO(1–0) emission peak (Figs. 2 and 3). In
the spectrum at this peak position, five successive spec- Figure 8 presents ground-based infrared K- (1.9–2.5
tral elements at the expected HCN(1–0) frequency show µm) and L- (2.8–4.1 µm) band spectra of NGC 2623,
flux excess, compared to the surrounding pixels (Fig. 5, Mrk 266 SW and NW, and Arp 193. K-band spectra
lower-left). For reference, a spectrum with less binning, withnarrowerwavelengthcoverageareavailableforNGC
despite being slightly noisier, also displays signatures of 2623, Arp 193, and Mrk 266 SW (Ridgway et al. 1994;
HCN(1–0)emission(Fig. 5,lower-right). Wethusregard Goldaderetal.1997). ForMrk266SW,anL-bandspec-
the HCN(1–0) emission signature as real in NGC 1377. trum is also available (Imanishi 2002). A K-band spec-
Figure7presentsGaussianfitstothedetectedHCN(1– trum of Mrk 266 NE, and L-band spectra of NGC 2623,
0), HCO+(1–0), and CO(1–0) lines. In the HCN(1– Mrk 266 NE, and Arp 193 are first presented in this pa-
0)/HCO+(1–0) spectra, the central velocity and line per. For Mrk 266, this is the first infrared spectrum
width of the Gaussian components are determined in- that clearly resolves emission from the Mrk 266 NE and
dependently for both lines. For NGC 1377, although SW nuclei, allowing separate investigation of the energy
HCN(1–0) and CO(1–0)emission are seen in similar po- sources of individual nuclei for the first time.
sitions (Figs. 2 and 3), their velocity profiles are signifi-
4.2.1. Continuum Flux Level
cantly different (Fig. 7). It may be that the HCN(1–0)
line reflectshighdensitymoleculargasinthe centralnu- ForNGC 2623,andMrk266NE andSW,theK-band
clearregionofNGC1377,whileCO(1–0)emissionprobes flux levels of our SpeX spectra (0′.′8 slit width) are a
thesurroundingdiffusemoleculargastowardthenuclear factor of 1.7–2.7 smaller than the K-band photometric
direction. We should take care interpreting the HCN(1– measurements with a 5” aperture (Carico et al. 1990).
0)/CO(1–0) brightness-temperature ratio of NGC 1377 For NGC 2623, Mrk 266 SW, and Arp 193, the K-band
because HCN(1–0) and CO(1–0) lines may probe physi- flux levels of our SpeX spectra (0′.′8 slit width) are 30–
cally unrelated molecular gas components. 50% smaller than those measured in 3” × 9” aperture
Table 5 summarizes the derived Gaussian fitting pa- spectra (Goldader et al. 1997). The difference is great-
rameters. ThederivedfluxesofHCN(1–0)andHCO+(1– est for Mrk 266 SW, whose spatially-extended K-band
0),andtheirbrightness-temperatureratiosatthenuclear emission is strongest in our spectra along the slit direc-
peak positions (i.e., spatiallyunresolvedcomponent)are tion. The smaller flux levels in our spectra are reason-
also summarized in Table 5. For NGC 2623 and Arp able, given that our smaller aperture misses a spatially
193, where continuum emission is subtracted, the inte- extended emission component.
grated HCN(1–0) and HCO+(1–0) fluxes are also esti-
4.2.2. Hydrogen Recombination Lines
matedfromthepeakcontoursoftheintegratedintensity
6 Imanishi et al.
Strong Paα (1.87 µm) and Brα (4.05 µm) emission estimated CO = 0.23, 0.16, and 0.18 for NGC 2623,
ph
lines are visible in some spectra. Enlarged spectra Mrk 266SW, and Arp 193, respectively. Using the for-
aroundthesestrongemissionlinesareshowninFigure9. mula CO = 1.46 × CO − 0.02 (Goldader et al.
spec ph
In the K-bandspectra, Brγ (2.17 µm) and H (1–0)S(1) 1997), these CO values are converted to CO =
2 phot spec
(2.12 µm) emission lines are detectable. Table 7 sum- 0.32,0.21,and0.24forNGC2623,Mrk266SW,andArp
marizes the strengths and line widths of these emission 193, respectively, which are slightly (0.02–0.05) larger
lines. Ifemissionfromthe broadlineregionsclosetothe than our measurements (Table 8).
central SMBH in an AGN is strong and only modestly
obscured, then we may be able to detect broad (FWHM 5. DISCUSSION
>1500kms−1)emissionlinecomponentsintheinfrared, 5.1. Comparison of HCN(1–0)/HCO+(1–0)
where dust extinction is much smaller than the optical, Brightness-Temperature Ratios with Other Galaxies
aswasseeninopticalSeyfert2galaxies(Hilletal.1996;
Figure 11 plots the HCN(1–0)/HCO+(1–0) and
Veilleux et al. 1997; Lutz et al. 2002). However, we see
HCN(1–0)/CO(1–0) brightness-temperature ratios for
no obvious signatures of broad line components. Since
the four LIRGs. Previously obtained data points of
buriedAGNs tend to containa largeramountofobscur-
nearby LIRGs (Imanishi et al. 2004, 2006b; Imanishi &
ing dust than Seyfert 2 AGNs (Imanishi et al. 2006a,
Nakanishi 2006; Imanishi et al. 2007b), starbursts, and
2007a, 2008), detection of the broad emission line com-
Seyfertgalaxies(Kohno2005)arealsoplotted. Asstated
ponentswouldbemoredifficult,eveniftheburiedAGNs
by Imanishi et al. (2006b), the HCN(1–0)/HCO+(1–0)
emit intrinsically luminous broad emission lines.
brightness-temperatureratiosinthe ordinatearemainly
In Table 7, since our small aperture spectra miss
used in our discussions for the following reasons. First,
spatially-extended emission, the absolute flux is not
since HCN(1–0) and HCO+(1–0) have similarly high
meaningful. Only the flux ratios between different emis-
dipole moments (µ > 3 debye; Botschwina et al. 1993;
sionlinesmeasuredwiththe sameapertureswillbeused
Millar et al. 1997), it is very likely that similar high-
in our discussion.
density molecular gas is probed with both lines. Next,
asboththeHCN(1–0)andHCO+(1–0)linesareobserved
4.2.3. PAH Emission
simultaneouslywith the same NMA arrayconfiguration,
The 3.3-µm PAH emission is clearly seen in the L- theirbeampatternsarevirtuallyidentical. Finally,both
band spectra of all the observed LIRGs. To estimate HCN(1–0) and HCO+(1–0) fluxes are measured at the
the 3.3-µm PAH emission strength, we make the rea- same time with the same receiver and same correlator
sonable assumption that the profiles of the 3.3-µm PAH unit,underthesameweatherconditions,sothatpossible
emission in these LIRGs are similar to those of Galactic absolute flux calibration uncertainties of interferometric
star-formingregionsandnearbystarburstgalaxies(type- data do not propagate to the ratio, which is dominated
1sourcesinTokunagaetal. 1991),followingImanishiet by statistical noise and fitting errors (see Fig. 7).
al.(2006a). Theadoptedprofilereproducestheobserved In Fig. 11, we note the following points: first, NGC
3.3-µm PAH emission features of the LIRGs reasonably 1377 exhibits a high HCN(1–0)/HCO+(1–0) brightness-
well. Table 8 summarizes the fluxes, luminosities, and temperatureratio,asseeninAGN-dominatedgalaxynu-
rest-frameequivalentwidthsofthe3.3µmPAHemission clei. Second, the Mrk 266 NE nucleus shows a signif-
feature. Onlytheequivalentwidthisnotsignificantlyaf- icantly higher ratio than the SW nucleus. Finally, the
fected by flux loss in our small-aperture spectra, and so HCN(1–0)/HCO+(1–0)brightness-temperatureratios of
will be used in our discussion. Mrk 266 SW and Arp 193 are low, as observed in star-
burstgalaxies,andthatofNGC2623ishigherthanMrk
4.2.4. CO Absorption 266 SW and Arp 193.
Figure 10 shows enlarged spectra in the longer wave-
5.2. Buried AGN Signatures in Individual Objects
length portion of the K-band spectra. All spectra show
spectral gaps in the continuum at λ > 2.35 µm in In this subsection, we look for buried AGN signatures
obs
the observed frame. We attribute the gaps to CO ab- in the infrared spectra of individual LIRG nuclei and
sorption features at λ = 2.31–2.4 µm produced by theninvestigatetheirHCN(1–0)/HCO+(1–0)brightness-
rest
stars older than 106 yr (Oliva et al. 1999; Ivanov et temperature ratios. In doing so, we have to account
al. 2000; Imanishi & Alonso-Herrero 2004; Imanishi & for the fact that starburst activity surrounds the cen-
Wada 2004). To estimate the CO absorption strengths, tral compact buried AGN, so starburst emission is less
we adopt the spectroscopic CO index (CO ) defined obscuredbygasanddustthantheburiedAGNemission.
spec
by Doyon et al. (1994) and follow the procedures previ- The contribution from the buried AGN to the observed
ously applied to other LIRGs by ourselves (Imanishi et infrared flux can be small, even if the buried AGN is in-
al.2004;Imanishi&Nakanishi2006). Power-lawcontin- trinsically luminous and is energetically important, due
uum levels (F = α×λβ), shown as solid lines in Fig. to dust extinction. Thus, we need to examine the in-
λ
frared spectra carefully.
10, are determined using data points at λ = 2.05–
rest
2.29 µm, excluding obvious emission lines. The derived
5.2.1. Mrk 266 NE
CO values are summarized in column 5 of Table 8.
spec
Ridgway et al. (1994) also estimated CO = Although optical spectroscopy by various groups has
spec
0.24±0.03 for NGC 2623, with a 2′.′7 aperture, which failedtofindAGNsignaturesintheMrk266NEnucleus
agreeswithourmeasurementofCO =0.27(Table8) (Osterbrock & Dahari 1983; Kollatschny & Fricke 1984;
spec
to within uncertainty. Goldader et al. (1997) measured Mazzarella&Boroson1993;Wuetal.1998;Goncalveset
the photometric CO index inside 3” × 9” apertures and al.1999;Ishigakietal.2000),buriedAGNsignaturesare
HCN/HCO+ in four LIRGs 7
discernible in our infrared spectra. First, Mrk 266 NE Shields et al. (2006), the measured CO(1–0) line width
shows a high H (1–0) S(1) to Brγ flux ratio, > 1 (Ta- providesasimilarSMBHmassof∼2×105M . TheEd-
2 ⊙
ble 7). Observationally,starburstgalaxies show H (1–0) dington luminosity of this SMBH mass is ∼7 × 109L .
2 ⊙
S(1) to Brγ flux ratios of <1 (Mouri & Taniguchi 1992), This is only slightly below the observed infrared lumi-
whilehighH (1–0)S(1)toBrγfluxratios(>1)arefound nosity (∼1 × 1010L ). In a buried AGN, the infrared
2 ⊙
inopticalAGNs(=Seyfertgalaxies)(Mouri&Taniguchi luminosity should be comparable to the bolometric lu-
1992),aswellasburiedAGNcandidates(Imanishietal. minosity, because almost all of the energetic radiation
2004). Next, the rest-frame equivalent width of the 3.3- from the central AGN is absorbed by the surrounding
µm PAH emission of Mrk 266 NE (EW ∼ 30 nm; dust and re-emitted as infrared thermal dust emission.
3.3PAH
Table 8) is significantly smaller than that observed in Thus, even if current constraints on the supermassive
starburst galaxies (∼100 nm; Moorwood 1986; Imanishi black hole mass are accurate, it is still possible that the
&Dudley2000),suggestingthataPAH-freeAGN-heated bulkoftheobservedluminosityinNGC1377comesfrom
hot dust continuum contributes significantly to the ob- buried AGN activity.
served 3–4 µm flux of Mrk 266 NE. However, uncertainties in the estimated SMBH mass
The HCN(1–0)/HCO+(1–0) brightness-temperature in NGC 1377could exist. First, the observed20-cmflux
ratio of Mrk 266 NE is also higher than that of the Mrk from a buried AGN can be severely attenuated by free-
266SWnucleus. AputativeburiedAGN,suggestedfrom free absorption, possibly leading to an underestimate of
X-ray data (Brassington et al. 2007), may enhance the the SMBH mass. Second, the line width ofH (1–0)S(1)
2
HCN(1–0) emission in Mrk 266 NE. emission is not well calibrated for an SMBH mass esti-
mate. Finally,inthesmalllinewidthrangeofmillimeter
5.2.2. NGC 1377 CO(1–0)(<100kms−1),theSMBHmassestimatedfrom
The HCN(1–0)/HCO+(1–0) brightness-temperature CO(1–0) line width yields systematically smaller results
ratio of NGC 1377 is high, as found in AGNs. The in- than other measurements (Shields et al. 2006). Hence,
frared L-band spectrum is also dominated by PAH-free the actual SMBH mass in NGC 1377 could be higher
continuum emission (Imanishi 2006), as usually seen in than above estimates.
AGNs. The strong 9.7 µm silicate dust absorption fea- Regarding the second point, the non-detection of high
ture detected in the Spitzer IRS infrared 5–35 µm spec- excitation emission lines is a natural consequence of a
trum (Roussel et al. 2006) is also incompatible with a buried AGN because the central AGN is obscured by
normal starburst, where stellar energy sources and dust dust and gas along virtually all directions at the inner
are spatially well mixed. It requires a buried AGN-type part(<1 pc), producing virtually no narrowline regions
centrallyconcentratedenergysourcegeometry(Imanishi (= the main emitting sources of high excitation lines).
etal.2007a). Theseoverallobservationalresultsarenat- Therefore,noneofthecurrentlyavailableobservational
urally explained by the presence of a luminous buried results preclude the presence of a luminous buried AGN
AGN in NGC 1377. in NGC 1377. Strong H2 emission is observed in NGC
APAH-freecontinuumandstrong9.7-µmsilicatedust 1377 (Roussel et al. 2006). In general, H2 emission
absorption could be explained by an exceptionally cen- is stronger in AGNs than starburst galaxies (Mouri &
trally concentrated extreme starburst whose emitting Taniguchi1992). Althougha majorgalaxymergercould
volumeispredominantlyoccupiedwithHII-regions,with produce strong H2 emission (Van der Werf et al. 1993),
virtuallynomoleculargasandphoto-dissociationregions NGC 1377 shows no sign of a major merger (Roussel et
(Fig. 1e of Imanishi et al. 2007a). Unlike ULIRGs with al.2006). EnergysourceobscurationforNGC1377isex-
L > 1012L , the absolute infrared luminosity of NGC tremely high (Roussel et al.2006),and the observedlow
IR ⊙
1377isonly∼1010.1L ,whichcouldbe accountedforby radio to infrared luminosity ratio (Roussel et al. 2003)
⊙
a smallnumber of super star cluster whose emitting size could be explained by severe flux attenuation of the ra-
is very small (<<100 pc) and emission surface bright- dio 20-cm (1.5 GHz) emission by free-free absorption.
ness is high (Gorjian et al. 2001). In fact, Roussel et
5.2.3. NGC 2623
al.(2003)preferredsuchacompactstarburst(superstar
cluster)scenario,butruledoutthepossibilityofanener- The EW3.3PAH and COspec values, and the H2(1–0)
getically important buried AGN, based on the following S(1) to Brγ flux ratio of NGC 2623 are all within star-
arguments: (1) estimated SMBH mass is too small to burst range. The HCN(1–0)/HCO+(1–0) brightness-
account for the luminosity of NGC 1377 with AGN ac- temperature ratios are slightly higher than Arp 193 and
tivity, and (2) high excitation forbidden emission lines, Mrk 266 SW, which may be due to an HCN(1–0) emis-
usually seen in Seyfert galaxies, are not detected. sionenhancementbyanX-raydetectedburiedAGN(see
Regarding the first point, the SMBH mass was esti- §2).
mated to be <2 × 105M , from the observed radio 20-
⊙ 5.2.4. Mrk 266 SW and Arp 193
cm (1.5-GHz) continuum flux and the small line width
of the infraredH (1–0)S(1) emissionline (FWHM < 25 The large EW and CO values, the small
2 3.3PAH spec
km s−1) (Roussel et al. 2003). The measured line width H (1–0) S(1) to Brγ flux ratios, and the small HCN(1–
2
of millimeter CO(1–0) lines in our NMA spectrum (Fig- 0)/HCO+(1–0)brightness-temperatureratiosofMrk266
ure 6) is ∼100 km s−1 in FWHM. After correction for SW and Arp 193 are all explained by starburst activity
the velocityresolutionofNMAdata(∼42kms−1; §3.1), only, with no significant AGN contribution required.
we obtain the intrinsic CO(1–0) line width of ∼90 km
s−1 in FWHM, which is similar to the SEST 15-m tele- 5.3. Interpretation of High HCN(1–0)/HCO+(1–0)
Brightness-Temperature Ratios in Buried AGN
scope measurement (Roussel et al. 2003). Assuming the
Candidates
CO(1–0) line width and SMBH mass relation given by
8 Imanishi et al.
One natural explanation for the high HCN(1– 0) brightness-temperature ratio, even for a low absolute
0)/HCO+(1–0) brightness-temperature ratios found in infrared luminosity galaxy (e.g. NGC 1377), if infrared
AGNs is an HCN abundance enhancement. If molec- emission surface brightness is high.
ular gas consists of small dense gas clumps with low InFig. 12,wemayseeaweakcorrelationbutthescat-
volume filling factor, as widely supported from obser- ter is large and the total sample size is small. Based on
vations (Solomon et al. 1987), an enhanced HCN abun- our current dataset, we cannot determine whether the
dance can result in higher HCN(1–0) flux regardless of infrared radiative pumping scenario is indeed at work in
whethertheHCN(1–0)emissionisopticallythinorthick buriedAGNsatLIRG’snuclei. Furtherdetailedtheoret-
(Imanishi et al. 2007b). Several chemical calculations of ical calculations that realistically incorporate the actual
HCNabundanceinmoleculargasaroundUV-andX-ray energy level populations of HCN and HCO+, as well as
emitting energy sources have been published (Meijerink the clumpy structure of molecular gas around an AGN
et al. 2006; Lintott & Viti 2006). Although an HCN (Yamada et al. 2007), are needed to properly interpret
abundance enhancement around an X-ray emitting en- the origin of strong HCN(1–0) emission in AGNs. Com-
ergy source (i.e., AGN) is predicted in some parameter binationwithhighertransitionlinesinthesubmillimeter
rangesthat are realistic for molecular gas aroundburied wavelengthrange(Wilson et al.2008)will also help bet-
AGNs in LIRGs (high-F range of Table 3 in Meijerink ter understand the physical properties of molecular gas
X
et al. 2006), an HCN abundance decrease is suggested in LIRG’s nuclei.
in other reasonable parameter ranges (Meijerink et al.
2006). Thechemicalcalculationresultsarehighlydepen- 6. SUMMARY
dent on parameters, and it is currently unclear whether We presented the results of millimeter interferometric
an HCN abundance is indeed enhanced in molecular gas simultaneousHCN(1–0)andHCO+(1–0)observationsof
in the close vicinity of buried AGNs in LIRG nuclei. fourLIRGsusingNMA.Whencombinedwithpreviously
ThesecondpossibleexplanationforthestrongHCN(1– observedLIRGs,oursisthelargestLIRGinterferometric
0)emissioninAGNsisaninfraredradiativepumpingsce- HCN(1–0) and HCO+(1–0) survey. From our interfero-
nario(Aaltoetal.1995;Garcia-Burilloetal.2006;Weiss metricdata,weextractedtheHCN(1–0)andHCO+(1–0)
et al. 2007). It is usually assumed that molecular gas is
fluxes at the nuclei, where putative AGNs are expected
excited by collision. However, the HCN molecule has a to be present. We derived the HCN(1–0)/HCO+(1–0)
line at infrared 14 µm that can be excited by absorbing
brightness-temperature ratios of the observed LIRG nu-
photons at λ ∼ 14 µm. HCN(1–0) emission in the mil-
clei and compared them with the ratios found in AGNs
limeter wavelength range could be enhanced through a and starburst galaxies. The main result of this paper is
cascadeprocess. Since the emissionsurfacebrightnessof
the confirmation of the previously discovered trend that
an AGN is high, much of the surrounding dust is heated LIRGs with luminous buried AGN signatures at other
to several 100 K, producing strong mid-infrared 10–20 wavelengths tend to show higher HCN(1–0)/HCO+(1–
µm emission (§1). Hence, this infrared radiative pump- 0) brightness-temperature ratios than those without.
ing scenario for HCN could work effectively in AGNs. This reinforces the utility of the HCN(1–0)/HCO+(1–0)
HCO+ also has a line at infrared 12 µm, so that the
brightness-temperatureratioasapowerfulobservational
infrared radiative pumping mechanism could work in a tool for discovering elusive AGNs deeply buried in dust
similarway. Observationally,the14µmHCNabsorption
and molecular gas.
features are detected in highly obscured LIRGs (Lahuis
et al. 2007), but the 12 µm HCO+ absorption features
arenot(Farrahetal.2007). Itmaybe thatthis infrared We are grateful to the NRO staff for their continu-
radiative pumping scenario is working more effectively ous efforts to make our NMA observations possible. We
for HCN than HCO+, possibly enhancing the HCN(1– also thank D. Griep for his support during our IRTF
0)/HCO+(1–0) brightness-temperature ratios in AGNs. observing runs, and the anonymous referee for his/her
To test this infrared radiative pumping scenario, in usefulcomments. M.I.issupportedbyGrants-in-Aidfor
Figure12wecomparetheobservedHCN(1–0)/HCO+(1– Scientific Research (19740109). Y.T. is financially sup-
0) brightness-temperature ratio with the infrared emis- portedbytheJapanSocietyforthePromotionofScience
sion surface brightness estimated by Soifer et al. (2000, (JSPS)forYoungScientists. NROisabranchoftheNa-
2001) and Evans et al. (2003). We use the infrared tional Astronomical Observatory, National Institutes of
emission surface brightness, rather than infrared lumi- Natural Sciences, Japan. This study utilized the SIM-
nosity (Gracia-Carpio et al. 2006), because the bulk of BAD database, operated at CDS, Strasbourg, France,
the observed normal LIRGs with <1012L (i.e., non- andtheNASA/IPACExtragalacticDatabase(NED)op-
⊙
ULIRGs) are chosen because they display luminous erated by the Jet Propulsion Laboratory, California In-
buried AGN signatures. Although Gracia-Carpio et al. stitute of Technology, under contract with the National
(2008)foundanincreasingtrendofHCN(1–0)/HCO+(1– Aeronautics and Space Administration. This publica-
0) brightness-temperature ratios with increasing galaxy tion makes use of data products from the Two Micron
infraredluminosities,ourheterogeneoussampleselection AllSkySurvey,whichis ajointprojectofthe University
ofnormalLIRGs(non-ULIRGs)couldbebiasedtoAGN ofMassachusettsandtheInfraredProcessingandAnaly-
candidates(=strongHCNemitters)andartificiallyelim- sis Center/California Institute of Technology, funded by
inate this trend. The infrared radiative pumping sce- the NationalAeronautics andSpace Administrationand
nariocan,inprinciple,enhancetheHCN(1–0)/HCO+(1– the National Science Foundation.
REFERENCES
Aalto,S.,Booth, R.S.,Black,J. H.,&Johansson, L.E.B.1995, Aalto, S., Spaans, M., Wiedner, M. C., & Huttemeister, S. 2007,
A&A,300,369 A&A,464,193
HCN/HCO+ in four LIRGs 9
Botschwina, P., Horn, M., Flugge, J., Seeger, S. 1993, J. Chem. Lahuis,F.etal.2007, ApJ,659,296
Soc.FaradayTrans.,89,2219 Lintott,C.,&Viti,S.2006, ApJ,646,L37
Brassington,N.J.,Ponman,T.J.,&Read,M.2007,MNRAS,377, Lutz, D.,Maiolino,R.,Moorwood,A.F.M.,Netzer, H.,Wagner,
1439 S.J.,Sturm,E.,&Genzel,R.2002, A&A,396,439
Brandl,B.etal.2006,ApJ,653,1129 Maiolino,R.etal.2003,MNRAS,344,L59
Bryant,P.M.,&Scoville,N.Z.1999, AJ,117,2632 Mazzarella,J.M.,&Boroson,T.A.1993,ApJS,85,27
Carico, D. P., Sanders, D. B., Soifer, B. T., Matthews, K., & Meijerink,R.Spaans,M.,&Israel,F.P.2006, ApJ,650,L103
Neugebauer, G.1990,AJ,100,70 Millar,T.J.,Farquhar,P.R.A.,&Willacy,K.1997,A&AS,121,
Condon, J. J., Huang, Z. -P., Yin, Q. F., & Thuan, T. X. 1991, 139
ApJ,378,65 Moorwood,A.F.M.1986,A&A,166,4
Downes,D.,&Solomon,P.M.1998, ApJ,507,615 Mouri,H.,&Taniguchi,Y.1992,ApJ,386,68
Doyon,R.,Joseph,R.D.,&Wright,G.S.1994, ApJ,421,101 Okumura,S.etal.2000,PASJ,52,393
Evans,A.S.etal.2003, AJ,125,2341 Oliva, E., Origlia, L., Maiolino, R., & Moorwood, A. F. M. 1999,
Evans,A.S.etal.2008, ApJ,675,L69 A&A,350,9
Fabian, A. C., Wilman, R. J., & Crawford,C. S., 2002, MNRAS, Osterbrock,D.E.,&Dahari,O.1983, ApJ,273,478
329,L18 Papadopoulos,P.P.2007,ApJ,656,792
Farrah,D.etal.2007,ApJ,667,149 Rayner, J. T., Toomey, D. W., Onaka, P. M., Denault, A. J.,
Gao,Y.,&Solomon,P.M.2004,ApJS,152,63 Stahlberger, W. E., Vacca, W. D., Cushing, M. C., & Wang,
Garcia-Burillo,S.,etal.2006,ApJ,645,L17 S.2003, PASP,115,362
Goldader,J.D.,Joseph,R.D.,Doyon,R.,&Sanders,D.B.1997, Ridgway,S.E.,Wynn-Williams,C.G.,&Becklin,E.E.1994,ApJ,
ApJS,108,449 428,609
Goncalves, A. C., Veron-Cetty, M. -P., & Veron, P. 1999, A&AS, Rothberg,B.,&Joseph,R.D.2004, AJ,128,2098
135,437 Roussel, H., Helou, G., Beck, R., Condon, J. J., Bosma, A.,
Gorjian,V.,Turner,J.L.,&Beck,S.C.2001, ApJ,554,L29 Matthews,K.,&Jarrett,T.H.2003, ApJ,593,733
Gracia-Carpio, J., Garcia-Burillo, S, Planesas, P., & Colina, L. Roussel,H.,etal.2006,ApJ,646,841
2006, ApJ,640,L135 Sanders,D.B.,&Mirabel,I.F.1996, ARA&A,34,749
Gracia-Carpio, J., Garcia-Burillo, S, Planesas, P., Fuente, A., & Sanders,D.B.,Scoville,N.Z.,Young,J.S.,Soifer,B.T.,Schloerb,
Usero,A.2008, A&A,479,703 F.P.,Rice,W.L.,&Danielson,G.E.1986, ApJ,305,L45
Guelin,M.,etal.2007,A&A,462,L45 Sanders,D.B.,Soifer,B.T.,Elias,J.H.,Madore,B.F.,Matthews,
Hill,G.J.,Goodrich,R.W.,&Depoy,D.L.1996,ApJ,462,163 K.,Neugebauer,G.,&Scoville,N.Z.1988a,ApJ,325,74
Imanishi,M.2002,ApJ,569,44 Sanders, D. B., Soifer, B. T., Elias, J. H., Neugebauer, G., &
Imanishi,M.2006,AJ,131,2406 Matthews,K.1988b, ApJ,328,L35
Imanishi,M.,&Alonso-Herrero,A.2004,ApJ,614,122 Shields,G.A.,Menezes,K.L.,Massart,C.A.,&VandenBout,P.
Imanishi,M.,&Dudley,C.C.2000,ApJ,545,701 2006,ApJ,641,683
Imanishi,M.,&Maloney,P.R.2003,ApJ,588,165 Skrutskie,M.F.,etal.2006,AJ,131,1163
Imanishi,M.,&Nakanishi,K.2006,PASJ,58,813 Soifer,B.T.etal.2000,AJ,119,509
Imanishi,M.,&Wada,K.,2004,ApJ,617,214 Soifer,B.T.etal.2001,AJ,122,1213
Imanishi, M., Dudley, C. C., & Maloney, P. R. 2006a, ApJ, 637, Solomon, P.M.,Downes, D.,& Radford,S. J. E.1992, ApJ,387,
114 L55
Imanishi, M., Dudley, C. C., Maiolino, R., Maloney, P. R., Solomon,P.M.,Rivolo,A.R.,Barrett,J.,&Yahil,A.1987,ApJ,
Nakagawa, T.,&Risaliti,G.2007a,ApJS,171,72 319,730
Imanishi, M., Nakagawa, T., Ohyama, Y., Shirahata, M., Wada, Thompson,T.A.,Quataert,E.,&Murray,N.2005,ApJ,630,167
T.,Onaka,T.,&Oi,N.2008,PASJ,inpress(ArXiv:0808:0363) Thorne,K.S.1974,ApJ,191,507
Imanishi,M.,Nakanishi,K.,Kuno,N.,&Kohno,K.2004,AJ,128, Tokunaga, A.T.2000, inAllen’sAstrophysical Quantities, ed.A.
2037 N.Cox(4thed;Berlin:Springer),143
Imanishi,M.,Nakanishi,K.,&Kohno,K.2006b, AJ,131,2888 Tokunaga A. T.,SellgrenK.,SmithR. G.,Nagata T.,Sakata A.,
Imanishi, M., Nakanishi, K., Tamura, Y., Oi, N., & Kohno, K. NakadaY.,1991,ApJ,380,452
2007b, AJ,134,2366 Tsutsumi,T.Morita,K.-I.&Umeyama,S.1997,inAstronomical
Ishigaki,T.etal.2000,PASJ,52,185 Data Analysis Software and Systems VI, ed. G. Hunt, & H. E.
Ivanov, V. D., Rieke, G. H., Groppi, C. E., Alonso-Herrero, A., Payne(SanFrancisco:ASP),50
Rieke,M.J.,&Engelbracht, C.W.2000,ApJ,545,190 VanderWerf,P.P.,Genzel, R.,Krabbe,A.,Blietz,M.,Lutz, D.,
Keel,W.C.1984,ApJ,282,75 Drapatz,B.,Ward,M.J.,&Forbes,D.A.,1993,ApJ,405,522
Kenney,J.D.,&Young,J.S.1988, ApJS,66,162 Vega, O., Clemens, M. S., Bressan, A., Granato, G. L., Silva, L.,
Kim, D. -C., Sanders, D. B., Veilleux, S., Mazzarella, J. M., & &Panuzzo,P.2008,A&A,484,631
Soifer,B.T.1995,ApJS,98,129 Veilleux,S.,Goodrich,R.W.,&Hill,G.J.1997,ApJ,477,631
Kohno, K. 2005, in AIP Conf. Ser. 783, The Evolution of Veilleux, S., Kim, D. -C., Sanders, D. B., Mazzarella, J. M., &
Starbursts, ed. S. Hu¨ttemeister, E. Manthey, D. Bomans, & K. Soifer,B.T.1995,ApJS,98,171
Weis(NewYork:AIP),203(astro-ph/0508420) Veilleux,S.,&Osterbrock,D.E.1987,ApJS,63,295
Kohno, K., Matsushita, S., Vila-Vilaro, B., Okumura, S. K., Weiss,A.,Downes,D.,Neri,R.,Walter,F.,Henkel,C.,Wilner,D.
Shibatsuka, T., Okiura, M., Ishizuki, S., Kawabe, R. 2001, in J.,Wagg, J.,&Wiklind,T.2007,A&A,467,955
ASP Conf. Ser. 249, The Central kpc of Starbursts and AGN, Wilson,C.D.,etal.2008,ApJ,178,189
ed.J.H.Knapen,J.E.,Beckman,I.Shlosman,&T.J.Mahoney Wu,H.,Zou,Z.L.,Xia,X.Y.,&Deng,Z.G.1998,A&A,132,181
(SanFrancisco:ASP),672(astro-ph/0206398) Yamada,M.,Wada,K.,&Tomisaka,K.2007,ApJ,671,73
Kollatschny,W.,&Fricke,K.J.1984, A&A,135,171
Krips, M., Neri, R., Garcia-Burillo, S., Martin, S., Combes, F.,
Gracia-Carpio,J.,&Eckart,A.2008,ApJ,677,262
10 Imanishi et al.
TABLE 1
Detailedinformationon theobservedLIRGs
Object Redshift f12 f25 f60 f100 logLIR (logLIR/L⊙) Far-infrared
(Jy) (Jy) (Jy) (Jy) (ergss−1) Color
(1) (2) (3) (4) (5) (6) (7) (8)
NGC2623(Arp243) 0.018 0.21 1.74 23.13 27.88 45.1(11.5) 0.08(cool)
Mrk266(NGC5256) 0.028 0.23 0.98 7.34 11.07 45.0(11.5) 0.13(cool)
Arp193(IC 883,UGC8387) 0.023 0.26 1.36 15.44 25.18 45.2(11.6) 0.09(cool)
NGC1377 0.006 0.44 1.81 7.25 5.74 43.7(10.1) 0.25(warm)
Note. — Col.(1): Object name. Col.(2): Redshift. Cols.(3)–(6): f12, f25, f60, and f100 are IRAS FSC fluxes at 12, 25, 60, and 100 µm,
respectively. Col.(7): Decimallogarithmoftheinfrared(8−1000µm)luminosityinergss−1 calculatedasfollows: LIR=2.1×1039×D(Mpc)2 ×
(13.48×f12 +5.16×f25 +2.58×f60+f100)ergss−1 (Sanders&Mirabel1996). Thevaluesinparenthesesarethedecimallogarithmsofthe
infraredluminositiesinunitsofsolarluminosities. Col.(8): IRAS25µmto60µmfluxratio(f25/f60). LIRGswithf25/f60 <(>)0.2areclassified
ascool(warm)(Sandersetal.1988b).
TABLE 2
Observing log of NMAmillimeterinterferometricobservations
Object Array ObservingDate Central Phase Band-pass Flux
configuration (UT) frequency calibrator calibrator calibrator
(1) (2) (3) (4) (5) (6) (7)
NGC2623 C 2007Dec13,14,15,16 87.30 0851+202 3C84 3C345
Mrk266 C 2008Jan8,9,10 86.52 1418+546 3C84 1749+096
D 2008Feb20,21,22,Mar4,5
C(CO) 2007Dec19,23 112.43 1418+546 3C84 1749+096
Arp193 C 2007Dec20 86.89 1308+326 3C84 1749+096
D 2008Feb18,19
NGC1377 D 2006Dec12,13 88.47 0420−014 3C454.3 Uranus
AB 2007Jan27,28
C 2007Dec13,14,15,16,19,20
C(CO) 2007Dec10,11 114.95 0420−014 3C454.3 Uranus
Note. — Col.(1): Objectname. Col.(2): NMAarrayconfiguration. Themark“(CO)”meansCO(1–0)observations,whichwereexecutedonly
forMrk266andNGC1377. Col.(3): ObservingdateinUT.Col.(4): Centralfrequencyusedfortheobservations. Col.(5): Objectnameusedasa
phasecalibrator. Col.(6): Objectnameusedasaband-passcalibrator. Col.(7): Objectnameusedasafluxcalibrator.
TABLE 3
Parametersof finalNMAmaps
Object Line Onsource Beamsize Positionangleof
integration(hr) (arcsec×arcsec) thebeam(◦)
(1) (2) (3) (4) (5)
NGC2623 HCN/HCO+ 10 5.4×4.2 −48.4
Mrk266 HCN/HCO+ 34 4.3×3.4 −31.5
CO 5 4.3×3.4 −31.5
Arp193 HCN/HCO+ 8 7.1×6.6 −52.5
NGC1377 HCN/HCO+ 23 7.1×4.0 −13.6
CO 4 6.7×3.6 −18.8
Note. — Col.(1): Object name. Col.(2): Observed line. HCN(1–0)/HCO+(1–0) or CO(1–0). Col.(3): Net on-source integration time in
hours. Col.(4): Beam size in arcsec × arcsec. Col.(5): Position angle of the beampattern. It is 0◦ for the north-south direction, and increases
counterclockwiseontheskyplane.