Molecules as tracers of galaxy evolution: an EMIR survey. I. Presentation of the data and first results

0.84 MB·English

by  F. Costagliola

Astronomy&Astrophysicsmanuscriptno.costagliola˙evol (cid:13)c ESO2011
January12,2011
Molecules as tracers of galaxy evolution: an EMIR survey
I. Presentation of the data and first results
F.Costagliola1,⋆,S.Aalto1,⋆⋆,M.I.Rodriguez2,S.Muller1,H.W.W.Spoon3,S.Mart´ın4,M.A.Pere´z-Torres2,
A.Alberdi2,J.E.Lindberg1,5,F.Batejat1,E.Ju¨tte6,P.vanderWerf7,8,andF.Lahuis7,9
1 Department of Earth and Space Sciences, Chalmers University of Technology, Onsala Space Observatory, SE-439 92 Onsala,
Sweden,e-mail:[email protected]
2 InstitutodeAstrof´ısicadeAndaluc´ıa(IAA-CSIC),POBox3004,E-18080Granada,Spain
1 3 CornellUniversity,AstronomyDepartment,Ithaca,NY14853,USA
1 4 EuropeanSouthernObservarory,AlonsodeCo´rdova3107,Vitacura,Casilla19001,Santiago19,Chile
0 5 CentreforStarandPlanetFormation,NaturalHistoryMuseumofDenmark,UniversityofCopenhagen,ØsterVoldgade5-7,1350
2 KøbenhavnK,Denmark
n 6 AstronomischesInstitutRuhr-UniversitaetBochum,Universitaetsstr.150,44780Bochum,Germany
a 7 LeidenObservatory,LeidenUniversity,NL-2300RA,Leiden,TheNetherlands
J 8 InstituteforAstronomy,UniversityofEdinburgh,RoyalObservatory,BlackfordHill,EdinburghEH93HJ,UnitedKingdom
9 SRONNetherlandsInstituteforSpaceResearch,P.O.Box800,NL-9700AV,Groningen,TheNetherlands
1
1

ABSTRACT
]
O
Aims.Weinvestigatethemoleculargaspropertiesofasampleof23galaxiesinordertofindandtestchemicalsignaturesofgalaxy
C evolutionandtocomparethemtoIRevolutionarytracers.
. Methods.Observationat3mmwavelengthswereobtainedwiththeEMIRbroadbandreceiver,mountedontheIRAM30mtelescope
h
onPicoVeleta,Spain.Wecomparetheemissionofthemainmolecularspecieswithexistingmodelsofchemicalevolutionbymeans
p
oflineintensityratiosdiagramsandprincipalcomponentanalysis.
-
o Results.We detect molecular emission in 19 galaxies in two 8 GHz-wide bands centred at 88 and 112 GHz. The main detected
r transitions are the J=1–0 lines of CO, 13CO, HCN, HNC, HCO+, CN, and C2H. We also detect HC3N J=10–9 in the galaxies
t IRAS17208,IC860,NGC4418,NGC7771,andNGC1068.TheonlyHC NdetectionsareinobjectswithHCO+/HCN<1andwarm
s 3
a IRAScolours.GalaxieswiththehighestHC3N/HCNratioshavewarmIRAScolours(60/100µm>0.8).ThebrightestHC3Nemission
[ isfoundinIC860,wherewealsodetectthemoleculeinitsvibrationallyexcitedstate.WefindlowHNC/HCNlineratios(<0.5),that
cannotbeexplainedbyexistingPDRorXDRchemicalmodels.TheintensitiesofHCO+andHNCappearanti-correlated,because

1 galaxieswithlowHCO+/HCNintensityratioshavehighHNC/HCN.NocorrelationisfoundbetweentheHNC/HCNlineratioand
v dusttemperature.AllHNC-brightobjectsareeitherluminousIRgalaxies(LIRG)orSeyferts.Galaxieswithbrightpolycyclicaromatic
2 hydrocarbons(PAH)emissionshowlowHNC/HCO+ratios.TheCO/13COratioispositivelycorrelatedwiththedusttemperatureand
2 isgenerallyhigherthaninourgalaxy.TheemissionofCNandC18Oiscorrelated.
1 Conclusions.BrightHC NemissioninHCO+-faintobjectsmayimplythatthesearenotdominatedbyX-raychemistry.Thusthe
3
2 HCN/HCO+lineratioisnot,byitself,areliabletracerofXDRs.BrightHC NandfaintHCO+couldbesignaturesofembeddedstar-
3
. formation,insteadofAGNactivity.MechanicalheatingcausedbysupernovaexplosionsmayberesponsibleforthelowHNC/HCN
1
andhighHCO+/HCNratiosinsomestarbursts.Wecannotexclude,however,thatthediscussedtrendsarelargelycausedbyoptical
0
deptheffectsorexcitation.Chemicalmodelsalonecannotexplainallpropertiesoftheobservedmolecularemission.Betterconstraints
1
tothegasspacialdistributionandexcitationareneededtodistinguishabundanceandexcitationeffects.
1
: Keywords.galaxies:evolution—galaxies:starburst—galaxies:active—radiolines:ISM—ISM:molecules
v
i
X
r 1. Introduction AGN and starburst are still not well understood and must be
a
further explored. At high redshift, LIRGs dominate the cos-
Luminous infrared galaxies (LIRGs) radiate most of their lu-
mic infrared background and, by assuming that they are pow-
minosity (L >1011 L ) as dust thermal emission in the
IR ⊙ ered by starburts, we can use these galaxies to trace the dust-
infrared and have been studied at almost all wavelenghts
obscured star-formation rate, the dust content and the metal-
(Sanders&Mirabel 1996). However, the nature of the power
licity in the early Universe (Bargeretal. 1999). The most well
source is still unclear when the inner region of the LIRG is
knowntechniquestodistinguishbetweenAGNandstar-powered
obscured by dust. The high central IR surface brightness im-
galaxiesrelyontheobservationofemissionlinesintheoptical
plies that this power source can be either an embedded com-
(e.g., Veilleux&Osterbrock1987).Overthelastdecadeseveral
pact starburst or an enshrouded AGN - or a combination of
otherdiagnosticdiagrams,basedonIRspectra,havebeenpro-
both. The evolutionof the activity and the connectionbetween
posed to quantify the contribution of star-formation and AGN
activity to the infrared luminosities of LIRGs (Genzeletal.
⋆ F.C.wishestothanktheEUESTRELAprogrammeforsupport. 1998; Lutzetal. 1998; Spoonetal. 2007). In the mm and sub-
⋆⋆ S.A.wishestothanktheSwedishResearchCouncilforgrantsup- mm there have been attempts as well to classify the activ-
port.
2 F.Costagliolaetal.:EMIRMolecularTracers
ity of galaxies via diagnostic diagrams, of which the most 4.5
well known is the HCN/HCO+ line ratio plot of Kohnoetal.
(2001) and Imanishietal. (2004). A multi-transition study of 4 8 24
the HCN/HCO+ ratio in Seyfertand starburstgalaxieshas also
3.5
been reportedby Kripsetal. (2008). These authorsfind an un- 21
derluminosityof HCO+ in some AGN-dominatedcores, which
3
they suggested to be owing to an underabundance of HCO+ ht 20
causedbytheX-ray-dominatedchemistryinducedbytheAGN ng2.5 3A 3B 3C
e
(Maloneyetal. 1996). This latter interpretation has been dis- Str
puted since it does not agree with recent models of X-ray- e 2
dinosmteiandatsehdowregainonesnh(XanDcRems)en(MtoefijHerCinOk+&aSbupnadanansc2e0s0c5a)u,swedhibcyh Silicat1.5 18 10 2 513
anincreaseinionization.Otherinterpretationsintermsofstar- 1 2A 7 2B 2C
burst evolution have been put forward as an alternative (e.g.,
3
B1U–aL0aInRlueGmtsailn(.Lo2IsRi0ti0>e8s1)0.w1G2ithrLa⊙crei)a´gi-anCrdagrsepntieooraeHltCaselN.em(-20cto0o6mh)apvaaelrseoldowfwoeiurtnhHdmCtoOhra+et 0.05 1A 15 1 1 9111 B22 6 14C 9
moderateLIRGs.TheHCN/CN1–0andHCN/HNC1–0linera- 10−2 10−1 100
tioshavebeenusedaswelltohelpinterpretagalaxy’spositionin PAH equivalent width (6.2 µm) [µm]
anevolutionaryscheme(Aaltoetal.2002;Baanetal.2008).For
Fig.1. Position of the observed galaxies on the mid-infrared
thebrightestnearbygalaxies,aschemebasedontheHNCO/CS
diagnostic diagram from Spoonetal. (2007). Galaxies are di-
ratio hasbeen proposed(Mart´ınetal. 2009). Moreover,promi-
vided into nine classes (1A–3C), depending on their spectral
nent sources such as NGC 253 and IC 342 allow the detec-
propertiesand position on the graph.These classes range from
tionofrarerspeciesthatcanbeusedinturntohelpidentifying
PAH-dominated spectra (1C) to continuum-dominated spec-
andevenresolvingthedominantactivity(e.g., Meier&Turner
tra with faint PAH emission (1A) and obscured galaxies with
2005;Mart´ınetal.2006).
deepsilicateabsorption(3A).Galaxytypesaredistinguishedby
Lineratiosneedtobeveryaccuratelymeasuredifwewantthem
their plotting symbol: Squares: LIRGs. Stars: Starburst galax-
to be sensitive tracers of molecular properties. The new EMIR
ies. Triangles: Seyfert galaxies. The numbers indicate differ-
receiver,mountedin2009ontheIRAM30mtelescopeinSpain,
ent galaxies, as explained in Table 2. The shadowed area cor-
offerstheopportunityofachievingthishighaccuracy.Theavail-
responds to the two main sequences discussed by Spoonetal.
ablebandwidthofnearly8GHz at3mmallowsustofitmany
(2007).AnincreaseofthePAHequivalentwidthindicateanin-
keymolecularlinesinthesameband,thereforeeliminatingthe
creaseofthestarburstcontributiontotheemission,whileAGN
uncertainties owing to relative calibration and pointing errors
spectraarecontinuum-dominated.SeeSection2forfurtherdis-
that affect most single-dish observations. We present a survey
cussion.
of molecular lines for a sample of 23 galaxies observed with
EMIR in the period June-November 2009. In section 3 we re-
port the details about the observationsand source selection. In branches, reflecting the diverse nature of the LIRG family. In
Sect.4wepresenttheresults.InSect.5thelineratiodiagrams this work, the term LIRG broadly refers to objects which emit
forHCN,HNC,HCO+,andHC Narepresentedanddiscussed. mostoftheirenergyintheIR.ThusagalaxyclassifiedasLIRG
3
Themolecularemissionisalsocomparedwithfar-infrared(FIR) may be a starburst, an AGN, or both, but the dust obscuration
coloursandpolycyclicaromatichydrocarbons(PAH)emission. hinders a clear classification. We do not explicitly distinguish
InSect.6wepresentourconclusionsandanoutlook.Thespec- between LIRGs (LIR >1011 L⊙) and ULIRGs (LIR >1012 L⊙).
tra of all observedsourcesand the tablessummarizingthe line However,IRluminositiesare reportedin Table2 forreference.
parametersofthedetectedspeciescanbefoundintheAppendix. ObjectswithbothLIRGandAGNsignaturesarelabelledasA,L
inTable2andappearasAGNs(triangles)inthegraphs.
Spoonetal.(2007)interprettheobserveddistributionasapos-
2. Sourceselection
sibleevolutionaryeffect,withsourcesmovingfromthediagonal
The first diagnostic diagram to take into account the effects of to the horizontal branch as the dust distribution evolves from
strong obscuration of the nuclear power source was presented a uniformto a clumpygeometry.The underpopulated2A class
bySpoonetal.(2007),usingtheequivalentwidthofthe6.2µm implieseitherthatthistransitionforLIRGsisveryrapid,orthat
PAHemissionfeatureandthestrengthofthe9.7µmsilicateab- LIRGsmostlyevolveintounobscuredstarbursts.
sorption(seeFig.1). We aim to compare this mid-IR evolution scheme with mm
Basedonthepositioninthediagram,galaxiesareputintonine molecularobservations.Our targetswere mainlyselected from
classes,rangingfromcontinuum-dominatedAGNhotdustspec- thesampleofSpoonetal.(2007),plusafewinterestingobjects
tra(1A)toPAH-dominatedstarburstspectra(1C)toabsorption- forwhichwehadPAHandsilicatemid-IRdata.Thesourcese-
dominatedspectraofdeeplyobscuredgalacticnuclei(3A). lectioncriteriawerethefollowing:
Spoonetal. (2007) find that galaxies are systematically dis-
– Uniformcoverageofthemostsignificantclassesinthediag-
tributedalongtwodistinctbranches:onehorizontalsequenceat
nosticdiagramofSpoonetal.(2007),
low silicate depths, ranging from AGN to starburst-dominated
– Uniform representation of different galaxy types (Seyferts,
spectra, and one diagonal sequence at higher silicate strength,
starbursts,LIRGs),
rangingfromobscurednucleitopure-starburstobjects.Thesep-
– SourcevisibilityatIRAMsite.
arationintotwobrancheslikelyreflectsfundamentaldifferences
in the dust geometry in the two sets of sources. Spectra of lu- The resulting sample is composed of 23 galaxies, whose
minousinfraredgalaxiesarefoundalongthefulllengthofboth main properties are listed in Tab 2. All targets were observed
F.Costagliolaetal.:EMIRMolecularTracers 3
Galaxy RA Dec VHelio HHCCON+((11−−00)) HHNCNC((11−−00)) HHCC3NN((11−0−0)9) HCC2HN((11−−00)) 13CCOO(1(1−−00)) Log(LIR/L⊙) Class Type Num.
[h:m:s] [◦:′:′′] [km/s]
IRAS17208 17:23:21.9 -00:17:01 12834 0.78(0.12) 0.77(0.14) 0.31(0.07) 0.42(0.10) - 12.35 1B L 1
IC860 13:15:03.5 +24:37:08 3347 0.62(0.16) 0.62(0.16) 0.42(0.13) 1.04(0.29) 18.96(3.83) 11.14 2B L 2
Mrk231 12:56:14.2 +56:52:25 12642 0.56(0.08) 0.38(0.06) <0.07 0.29(0.12) - 12.37 1A A,L 3
NGC1614 04:33:59.8 -08:34:44 4778 1.83(0.37) 0.33(0.15) <0.35 1.00(0.33) 28.81(2.10) 11.43 1C S 4
NGC3079 10:01:57.8 +55:40:47 1116 1.12(0.11) 0.27(0.05) <0.06 0.54(0.09) 17.10(1.16) 10.65 2C A 5
NGC4194 12:14:09.5 +54:31:37 2501 1.32(0.35) 0.53(0.23) <0.20 0.98(0.36) 19.10(1.90) 10.93 1C S 6
NGC4388 12:25:46.7 +12:39:44 2524 1.38(0.40) 0.62(0.24) <0.35 0.81(0.52) - 9.66 2A A 7
NGC4418 12:26:54.6 -00:52:39 2110 0.59(0.10) 0.47(0.09) 0.37(0.09) 0.64(0.18) - 11.00 3A L 8
NGC6090 16:11:40.7 +52:27:24 8785 1.67(0.44) 0.25(0.17) <0.18 0.46(0.20) - 11.34 1C L 9
NGC6240 16:52:58.9 +02:24:03 7339 1.63(0.14) 0.20(0.06) <0.09 0.36(0.08) 28.79(2.99) 11.69 2B L 10
NGC7469 23:03:15.6 +08:52:26 4892 1.12(0.11) 0.55(0.08) <0.07 0.75(0.24) 20.80(0.47) 11.41 1B A 11
NGC7771 23:51:24.9 +20:06:43 4277 0.94(0.09) 0.44(0.07) 0.06(0.04) 0.32(0.06) 13.61(0.52) 11.24 - L 12
NGC660 01:43:02.4 +13:38:42 850 1.04(0.09) 0.52(0.06) <0.07 0.45(0.08) 16.45(0.49) 10.40 2C S 13
NGC3556 11:11:31.0 +55:40:27 699 1.57(0.37) 0.37(0.24) <0.27 1.56(0.37) 12.50(0.31) 10.00 - S 14
NGC1068 02:42:40.7 -00:00:48 1137 0.67(0.03) 0.40(0.03) 0.04(0.02) 0.37(0.09) - 10.89 1A A 15
NGC7674 23:27:56.7 +08:46:45 8671 - - - - 14.46(1.18) 11.50 - A 16
UGC2866 03:50:14.9 +70:05:41 1232 1.46(0.20) 0.51(0.13) <0.14 0.88(0.17) 20.68(0.80) 10.69 - S 17
UGC5101 09:35:51.6 +61:21:11 11802 0.36(0.20) 0.82(0.28) <0.28 0.89(0.37) >8.63 11.87 2B A,L 18
NGC2273 06:50:08.6 +60:50:45 1840 1.05(0.37) 1.09(0.38) <0.89 <0.89 - 10.11 1B A 19
Arp220 15:34:57.2 +23:30:09 5382 0.47(0.07) 0.49(0.12) 0.19(0.06) - - 12.15 3B L 20
IRAS15250 15:26:59.4 +35:58:38 16535 - - - - - 12.02 3A L 21
NGC1140 02:54:33.6 -10:01:40 1501 - - - - - 9.50 1C S 22
NGC1056 02:42:48.3 +28:34:27 1545 - - - - - 9.50 - S 23
NGC1377 03:36:39.1 -20:54:07 1792 - - - - - 9.63 3A O 24
Table 2. Line ratios and general properties of the observed galaxies. The mid-IR Class column refers to the classification by
Spoonetal.(2007),basedonthePAHequivalentwidthandsilicateabsorptioninthemid-infrared.ThecolumnTypesummarizes
thepropertiesoftheobject,distinguishingbetweenstarbursts(S),Seyferts(A),andLIRGs(L).ThegalaxyNGC1377isclassified
asobscured(O),becauseitemitsmostlyintheIR,butitsluminosityisnothighenoughtobeclassifiedasLIRG.Foradiscussion
aboutsourcetypesseeSection2.InFigs.1,4and5,thesourcesarelabelledwiththenumbersreportedincolumnNum.
in the 88 GHz band, but because of time constraints, only 12 Calibration scans on the standard two load system were taken
wereobservedinthe112GHzband. every 5 minutes. The observing conditions were optimal, with
characteristicsystemtemperaturesof100and150Kforthe88
GHzand112GHzobservations,respectively.Thisresultedina
3. Observations rms noise per channelof roughly0.6 mK at 60 km s−1 resolu-
tion,for2hoursofon–sourceobservingtime.
3.1.Lineratiosat88GHz
Data were reduced with the CLASS1 software. A first order
The observations were obtained in June-November 2009 with baseline was removed from all spectra, which are shown in
the IRAM 30 m telescope on Pico Veleta, Spain. The 8 GHz Appendix D. The intensity scale in all spectra is in T∗, which
A
bandoftheEMIRreceiverwascentredontwodifferenttunings, isrelatedtothesourcebrightnesstemperatureas
at88.675and112.15GHz.Thesefrequencieswerechoseninor-
dertoaccommodateasmanypotentiallystronglinesaspossible T∗ θ2+θ2
T = A × s b, (1)
inthesameband.Alistofthetransitionsofthemostimportant b η θ2
mb s
moleculargastracersisgiveninTable1,togetherwithbeamef-
ficiencies (ηmb) and beam sizes (HPBW) for the two observed with θs and θb the angular sizes of source and beam, respec-
bands. Observationswere performedin dual polarization,each tively.We assume a constantmain beamefficiencyacross both
covering8GHz,whichnicelyfittedthe4×4GHzbackendbot- observedbands.Theerroronthemainbeambrightnesstemper-
tleneck.TheE090frontendwasconnectedtothelow-resolution ature estimate introduced by this assumption is of the order of
WILMAautocorrelator,whichiscapableofprocessingthelarge onepercent.
input bandwidth with a resolution of about 2 MHz (≃7 and 5
km s−1 at 88 GHz and 112 GHz, respectively). Observations
4. Results
wereperformedinwobblerswitchingmode,withathrowof60-
120′′(dependingonsourcesize), inorderto maximizebaseline Rest frequencies were taken from the NIST database
quality. Recommended Rest Frequencies for Observed Interstellar
Thepointingmodelwascheckedagainstbright,nearbycalibra-
torsforeverysource,andeverytwohoursforlongintegrations. 1 http://iram.fr/IRAMFR/GILDAS/
4 F.Costagliolaetal.:EMIRMolecularTracers
Fig.2.SpectraobservedwithEMIRinIC860at88and112GHz.TheintensityscaleisinT⋆,notcorrectedformainbeamefficiency.
A
Theregionmarkedwith(1/10),aroundtheCO1–0line,hasbeenscaleddownbyafactor10.Themainmoleculartransitionsare
labelled regardlessof line detection. The C H 3/2 and 1/2 labels mark the limits of the C H multipletat 87 GHz. Transitionsof
2 2
vibrationallyexcitedHC Narelabelledasv 1eandv 1f.Thefrequencyscaleistheobservedfrequency,notcorrectedforredshift.
3 7 7
MolecularMicrowaveTransitions2.Ourlineidentificationtakes galactic nuclei do reside at low HCO+/HCN values compared
into account the distortion of the velocity scale caused by the withstarbursts.
large observed bandwidth. For a discussion of this effect, see TheHC N/HCNlineratioisalsoreportedinthegraphs.Circles
3
Gordonetal.(1992). are drawn around sources where HC N J=10–9 has been de-
3
AnexampleofanEMIRspectrumisshowninFig.2,whileall tected, the diameter of the circle being proportional to the
theobservedspectraareshowninAppendixD.Lineintensities HC N/HCN ratio. Evidently that all the HC N detections have
3 3
were extracted by means of Gaussian fitting and are reported HCO+/HCN<1andHNC/HCN>0.4.
alongwithotherlineparametersinAppendixE. Plots a and c show strong correlations. As HNC/HCO+ in-
Integratedintensitiesofthemostrelevantspecieswerecombined creases,galaxiesmoveonthegraphstowardshigherHNC/HCN
to form line intensity ratios, which are listed in Table 2. These and HCO+/HCN ratios. An inverse correlation between HNC
wereusedtoproducethe diagramsshowninFigs. 4and5. We and HCO+ line intensities also emerges from plot b. Galaxies
choosetocompareintensitiesoftransitionsinsidethesamefre- at HCO+/HCN < 1 present on average a 30 % increase in
quencybandtomaximizetheaccuracyofthederivedlineratios. HNC/HCN, comparedwith those at higher HCO+/HCN ratios.
Thediagramsderivedfromthetwoobservedbandsaredescribed IfweexcludetheSeyfertNGC2273,thehigherendoftheHNC
insections3.1and4.1.Aprincipalcomponentanalysisofthe88 luminosity distribution (see graphs b and c in Fig. 4) is com-
GHzdatasetisdiscussedinsection4.2. posedmainlyofluminousinfraredgalaxies.
In plots a, b, and c of Fig. 4, we reportthe line ratios between Thegeneralpictureemergingfromthe graphsisthatstarbursts
thefirsttransitions(J=1–0)ofHCN,HNCandHCO+forallde- are characterized by faint HNC and bright HCO+ emission
tected galaxies. The different symbols refer to different galaxy (compared with HCN), while LIRGs mostly occupy the oppo-
types. We note that all non-compact starburst galaxies (stars) siteendoftheHNC-HCO+ correlation,withhighHNCandlow
are HCO+-luminous,with line ratiosHCO+/HCN ≥ 1 (plotsa, HCO+ intensities.AllHNC-brightgalaxiesareeitherLIRGsor
b).Luminousinfraredgalaxies(squares)andSeyferts(triangles) Seyferts.MostoftheHC NdetectionsareLIRGs,theonlyex-
3
do not show any strong trend. However, the majority of active ceptionbeingtheSeyfertNGC1068.
2 http://physics.nist.gov/PhysRefData/Micro/Html/contents.html
F.Costagliolaetal.:EMIRMolecularTracers 5
Line HPBW ηmb ν0 Eup nc(20K) PC1 PC2 PC3 PC4 PC5 PC6
[′′] [GHz] [K] [cm−3]
Variance% 45 29 16 6 3 1
88GHzBand 29 0.81
SiO(2-1) ′′ ′′ 86.847 6.25 1×105 HNC/HCN 0.03 -0.19 0.59 -0.32 0.02 -0.71
C2H(1-0) ′′ ′′ 87.316 4.19 1×105 HCO+/HCN -0.24 0.17 -0.26 -0.18 -0.9 -0.21
HCN(1-0) ′′ ′′ 88.633 4.25 2×105 C2H/HCN -0.02 0.10 0.26 -0.79 -0.01 0.54
HCO+(1-0) ′′ ′′ 89.188 4.28 3×104 Silicate 0.83 0.52 -0.06 -0.06 -0.07 -0.12
HNC(1-0) ′′ ′′ 90.663 4.35 1×105 PAH -0.46 0.68 -0.20 -0.17 0.42 -0.28
HC3N(10-9) ′′ ′′ 90.978 24.0 9×104 25/100µm 0.17 -0.43 -0.69 -0.45 0.22 -0.24
112GHzBand 22 0.78
Table 4. Projectionof the principalcomponentsontothe base
HC3N(12-11) ′′ ′′ 109.173 34.0 2×105 ofobservables.ForeachPC,itscontributiontothetotalvariance
C18O(1-0) ′′ ′′ 109.757 5.27 4×102 inthedataisalsoshown.
13CO(1-0) ′′ ′′ 110.201 5.29 4×102
CN(1-0)J=1/2-1/2 ′′ ′′ 113.191 5.43 2×106
CO(1-0) ′′ ′′ 115.271 5.53 4×102
eral, however, the observed quantities are not independentand
Table1.Propertiesofthetwoobservedbands.Restfrequencies different environments are best described by a combination of
ofthebrightestlinesineachbandareshownwiththebeamsizes observables.Asthenumberofobservablesincreases,itbecomes
(HPBW) and main beam efficiencies (η ). Energies of upper moreandmoredifficulttointerpretmultidimensionaldatasetsby
mb
level(E )andcriticaldensitiesat20K(n )weretakenfromthe meansof2Dsections(i.e.,line-ratiodiagrams).
up c
Leidenatomicandmoleculardatabase. To address this complexity, we applied a principal component
analysis(PCA)toourdata.Thisiscommonlyusedtoreducethe
dimensionalityofadatasetandisveryeffectiveinfindinghidden
trendsthatcouldotherwisebeburiedinthenoise.Foranappli-
HNC/HCN HCO+/HCN C2H/HCN Silicate PAH 25/100µm cation of a PCA to molecularmaps, see, e.g.,Ungerechtsetal.
(1997)andMeier&Turner(2005).
HNC/HCN 1.00 - - - - - Each galaxy in our sample is described by a set of n observed
HCO+/HCN -0.43 1.00 - - - - quantities (e.g. line ratios, IR properties) that represent an ini-
C2H/HCN 0.46 0.09 1.00 - - - tial base of vectors. The PCA algorithm first computes the co-
Silicate -0.11 -0.37 0.03 1.00 - - variance matrix of the data along the n directions and finds its
PAH -0.41 0.66 0.19 -0.22 1.00 - eigenvaluesandeigenvectors.Theeigenvaluesarethensortedin
25/100µm -0.24 -0.10 -0.2 0.03 -0.46 1.00 descending order, and the corresponding eigenvectors labelled
asprincipalcomponent(PC)1,2,etc...
Table3.Correlationmatrix Thefirstprincipalcomponent(PC1)isthusthelinearcombina-
tion of the initial galaxy propertiesalong which the dispersion
ofthedataismaximum.PC2isthevector,perpendiculartoPC
1, which has the second highestdispersion, and so on for all n
PCs. The valuesof the observedpropertiesfor eachgalaxyare
4.1.Lineratiosat112GHz
thenprojectedonthenewbaseofPCs.
Themaindetectionsinthe112GHzbandaretheJ=1-0emission In our analysis we choose as a base of observablesthe line in-
lines of CO, 13CO and C18O, and the spin doublet of CN 1–0 tensity ratios HNC/HCN, HCO+/HCN, and C H/HCN and the
2
J=3/2-1/2 and J=1/2-1/2. These lines are detected in most of IR propertiessilicate absorption, PAH EW and the ratio of the
the galaxies in our sample, and their integrated intensities are IRAS fluxes at 25 and 100 µm. Because our algorithm cannot
comparedintheplotsinFig.5. dealwith upperlimits, we do notincludefaintlines, as HC N.
3
AparticularlyrichchemistryisfoundinIC860,whosespectrum Inordertomaximizethenumberofgalaxies,welimitouranal-
shows bright emission lines of methanol and HC N J=12–11. ysistothe88GHzband,sinceonlyafractionofoursamplewas
3
ThisdetectionofHC Nistheonlyoneinthe112GHzsample observedat112GHz.
3
andwillbefurtherdiscussedinsection5.8.
4.2.1. ResultsofthePCanalysis
4.2.Principalcomponentanalysis
TheresultsofthePCAalgorithmareshowninTable4.Herethe
Line-ratio diagrams as the ones shown in sections 3.1 and 4.1 projectionof the PCs onto the originalbase are reportedalong
represent the standard framework for interpreting molecular withtheircontributiontothetotaldispersioninthedataset.Most
data and can be easily compared with previous studies (e.g., oftheinformationiscontainedinthefirstthreePCs,whichac-
Baanetal. 2008; Loenenetal. 2008; Baanetal. 2010), as dis- countforabout90%ofthetotalvariance.Inthefollowinganal-
cussedinsection4.3.Inthisapproach,thepropertiesofdifferent ysiswewillthusfocusonPC1,2,3.
galaxytypesarecomparedtwoatatimetofindtheobservables InFig.3 rightweshowtheprojectionofthePCsontotheorigi-
that best characterize different physical environments. In gen- nalbaseofobservables,whileinFig.3 leftweplottheposition
6 F.Costagliolaetal.:EMIRMolecularTracers
PC1 vs PC2 PC Components
2 01 9 4 6 1 53 10 2 18 8 2 000...246 PAH C2H/HCN silicate
C C
P P 0 HCO+/HCN
−1 111 19 7 −0.2 HNC/HCN
3 −0.4
15 25/100
−1 0 1 2 3 0 0.5 1
PC1 PC1
PC2 vs PC3 PC Components
1 1 19 18 0.5 HNC/HCN
2 C2H/HCN
0.5
3 3 0 silicate
PC 0 11 7 13 5 PC
6 HCO+/HCN PAH
−0.5 3 4 8 10 9 −0.5
−1 15 25/100
−1.5 −1 −0.5 0 0.5 1 −0.4 −0.2 0 0.2 0.4 0.6 0.8
PC2 PC2
Fig.3. Results from the principalcomponentanalysis. Left: Position of the observedgalaxiesin the PC base [PC1,PC2,PC3].
Differentsymbolsrefertostarbursts(stars),Seyferts(triangles)andLIRGs(squares).GalaxieswithbothAGNandLIRGproperties
areplottedastriangles.Numbersindicatedifferentgalaxies,asexplainedinTable2.ValuesforthePAHequivalentwidthsaretaken
fromSpoonetal.(2007).InfraredfluxeswereobtainedfromtheNASA/IPACExtragalacticDatabase.Right:ProjectionofthePC
componentsontotheoriginalbaseofobservables.WeconsideronlythefirstthreePCs,sincetheyaccountfor90%ofthevariance
andthuscontainmostoftheinformationinthedataset.
ofthegalaxiesinthePCbase. component.
Themaincontributiontothedispersionofthedatacomesfrom Most of the molecular information is contained in PC 3,
PAHEWandsilicateabsorption,whichhavethehighestprojec- where a substantial fraction of the total dispersion is along
tionsonPC1.Theoppositesignsmeanthatthedispersionalong the HNC/HCN–HCO+/HCN direction, while the contribution
PC1 is mainly caused by galaxies that have low PAH EW and by PAH and silicates is almost negligible. The IRAS colour
highsilicate absorption,orvice versa.ThefirstPC is therefore (25/100µmratio)stillplaysanimportantrole,withaprojection
dominatedby the diagonalsequenceon the mid-IRdiagramof alongPC3(≃-0.7)slightlyhigherinabsolutevaluethantheone
Fig.1.GalaxieswithhighvaluesofPC1willhavehighsilicate fortheHNC/HCNlineintensityratio(≃0.6).Theoppositesign
absorption and low PAH EW. This is evident when comparing ofthesetwoprojectionsiscausedbyaslightanti-correlationof
Figs.1and3 left,wherethegalaxiesNGC4418andNGC6090 the HNC/HCN ratio with dust temperature, which also results
(8and9onthegraphs)lieattheoppositeendsofbothPC1and from the correlation matrix in Table 3. This trend, however, is
themid-IRsequence. not evident in Fig. 4e. The distribution of galaxies along PC 3
Among the molecular line intensity ratios the largest contri- does not show any obvious correlation with galaxy properties
bution to PC 1 is given by HCO+/HCN. From Fig. 3 left, anditsinterpretationisnotstraightforward.Wewillfurtherdis-
HCO+/HCN appears to be positively correlated with PAH EW cusstheresultsofthePCanalysisinPaperII,whenwewillhave
(same sign of the projection along PC 1). This agrees with additionalinformationfromVLAradioobservationsabout,e.g.,
Figs. 4d and B.1 and generally with the trend that will be dis- gassurfacedensityandstar-formationrate.
cussedinSect.5.10.
Thesecondprincipalcomponent(PC2)isalsodominatedbyIR
observables, but with different contributions than the ones de- 4.3.Comparisonwithpreviousobservations
rivedforPC1.SilicatesandPAHprojectionsarenowpositively
correlated,andtheratioIRAS25/100µmgivesasignificantcon- A previousstudyofthe molecularemission ofdensegasin lu-
tributiontothedispersion.AscanbeseeninFig.3 left,PC2ef- minousinfraredgalaxieswasreportedbyBaanetal.(2008).The
ficientlyseparatesAGNsfrom obscuredorstarburst-dominated authorsanalysedatafor117galaxies,withinfraredluminosities
sources. All AGNs, with the exception of NGC 3079 (number rangingoveraboutthree ordersof magnitude.Ourdiagramsin
5 on the graph), have negative PC 2 values, which correspond Fig.4canbedirectlycomparedwiththoseofFig.9inBaanetal.
roughlytoclasses1A-1Bonthemid-IRdiagramofFig. 1.An (2008).
interestinganti-correlationbetweenPAHEWandIRAS25/100 The correlations between HNC/HCO+ and the two ratios
µmfluxratioisalsoevidentinFig.3 left. Thismaybecaused HNC/HCN and HCO+/HCN, shown by Baanetal. (2008) in
by higher dust temperatures in systems with a dominant AGN plotsaandc,areconfirmedbyourobservations.Ourmeasure-
F.Costagliolaetal.:EMIRMolecularTracers 7
menthaveamuchsmallerscatter,thankstotheexcellentrelative excitationacrossourgalaxysamplewillbefurtherdiscussedin
calibrationofthelineintensitiesprovidedbytheEMIRreceiver. afollowingpaper,wherewewillinclude1mmcounterpartsof
However,wecannotexcludethatthesmallerscatterispartially theobservedlines.
causedbysourceselectioneffects.
Plot b in Baanetal. (2008), showing HCO+/HCN versus
HNC/HCN,however,cannotbereproducedbyourdata.Inpar-
ticular,wedidnotobserveanygalaxywithlogHNC/HCN<-0.4 5.2.IsHCO+/HCNdrivenbyXDRchemistry?
andlogHCO+/HCN<0,whileabout25%ofthesourcesplotted
byBaanetal.(2008)areinthisrange.
Asaresult,thecorrelationbetweenHCO+/HCNandHNC/HCN The HCO+/HCN J=1-0 line ratio has been proposed as a reli-
lineratiosinourgraphbgoesintheoppositedirection,withlow ablediagnostictooltodistinguishbetweenAGN-andstarburst-
HNC/HCN corresponding to high HCO+/HCN. It is not clear poweredgalaxies(e.g., Kohnoetal.2001;Imanishietal.2004,
whetherthisiscausedbyselectioneffectsthatmayaffectourob- 2007). Interferometric observations by Kohnoetal. (2001)
servations,ortothelargescatterinthedatainBaanetal.(2008). reveal that Seyfert galaxies have lower HCO+/HCN ratios
InTableC.1wecomparelineintensitiesfromBaanetal.(2008) compared with starbursts. Seyfert galaxies with starburst-like
with our EMIR values. In most cases the two datasets agree at HCO+/HCN ratios are generally interpreted as mixed AGN-
the20%level,withsomeexceptions.Themoststrikingdiffer- starburstobjects(Imanishietal.2007).Inourobservations,star-
ences are seen for HNC/HCN in NGC 7469 and HCO+/HCN bursts and LIRGs have, on average, higher HCO+/HCN ra-
inNGC6240,whichvarymorethan60%,passingfromvalues tios than AGN, which is consistent with the trend observed
higherthanunityinonedatasettomuchlowerratiosintheother. by Kohnoetal. (2001) andImanishietal. (2007). This is com-
These discrepanciescouldbe causedby calibrationor pointing monly attributed to an enhancementof HCN abundance in the
inaccuracies,whichaffectmostnarrowbandsingledishobserva- X-ray-dominatedregion(XDR, Maloneyetal.1996)surround-
tions.ThedeterminationoflineratioswithEMIRismorerobust, ing an AGN (Lepp&Dalgarno 1996). Chemical models by
becausetheseinaccuraciesdonotaffecttherelativeintensityof Meijerink&Spaans(2005)showthatanHCO+/HCNlineratio
linesdetectedinthesameband. lower than one can be observed on the surface of low-density
(n <105 cm−3) XDRs, for H column densities <1022 cm−2.
2
However, at higher depths inside the molecular cloud, HCO+
5. Discussion abundanceincreasesandeventuallythecumulativeHCO+/HCN
column density ratio becomeshigher than one. Meijerinketal.
5.1.Interpretingline-intensityratios
(2007) report that for N(H )>1023 cm−2, the intensity of the
2
Line-intensity ratios are often interpreted as estimates of the J=1–0and J=4–3transitionsemergingfromanXDRishigher
abundanceratios of two molecular species. This is true only if forHCO+ thanforHCNbyafactorofatleastthree.Thisisat-
theemissionisopticallythinandthetwomoleculeshavesimi- tributedtothefactthatHCNabundanceinanXDRisenhanced
larexcitationproperties.Moreover,singledishobservationsonly (compared with HCO+) only in a narrow range of ionization
revealglobalratios,averagedonthewholegalaxy,andtheinter- rate/density(Fig.3in Lepp&Dalgarno1996).Meijerinketal.
pretationoftheresultshastotakeintoaccountthelackofspacial (2007) conclude that a low HCO+/HCN line ratio is a good
information. tracer of UV-dominated regions (PDR) for gas densities >105
Inourstudywedetecttransitionswithcriticaldensities(n )that cm−3, rather than X-ray-irradiated gas. Thus the observed low
c
varybyalmostfourordersofmagnitude(seeTable1),anditis HCO+/HCN line ratio could be either caused by low-density
thereforelikelythatexcitationplaysanimportantroleindeter- (n<105cm−3)XDRsordense(n>105cm−3)PDRs.
miningtheobservedratios. ObservationsbyBlakeetal.(1987)intheOrionmolecularcloud
Transitions with large n are efficiently excited in high density reveal extremely low HCO+/HCN abundance ratios (<10−3) in
c
gas,whichismostlyconcentratedintheinnerpartsofgalaxies, hotcores,i.e.warmanddensegasaroundmassiveyoungstars.
whiletheemissionoflineswithlowern ,whichcanbeexcitedin This is supported by calculations by Bayetetal. (2008), who
c
morediffusegas,isgenerallymoreextended.Low-n lineswill find abundances of HCO+ as low as 10−12 (relative to H ) in
c 2
thushaveahigherbeamfillingfactor,resultinginanincreased awiderangeofhotcoremodels.ThefaintHCO+ emissionmay
T⋆.Inthiscontext,itisadvisabletocomparelineintensitiesof thus be caused by deeply embedded star-formation, instead of
A
transitionswithsimilarn ,sincetheresultingratioscanbemore XDRorPDRchemistry.
c
directlyrelatedtomolecularabundances. Thisinterpretationisbasedontheassumptionthatlineratiosare
Also, the excitation temperatureof the transitions is an impor- directlylinkedto relativeabundances.Thismaynotbetruefor
tantfactorwheninterpretingsingledishintensities.Iftheemis- HCNandHCO+,whichhavedifferentexcitationproperties.The
sioniscomingfromregionswithdifferentgastemperatures,this criticaldensityofthe J=1–0transitionofHCO+ isoneorderof
translates into a differentpopulationof the low-J levels, which magnitudelowerthanforHCN.Asaconsequence,HCO+emis-
directlyaffectsthelineintensityratios. sioncanoriginatefromlowerdensitygas,whichingeneralhas
Another complication arises when we take into account radia- ahigherfillingfactorcomparedwiththedense(n >105 cm−3)
H
tive excitation. Many of the observed molecules, as e.g. HNC, component.ThisimpliesthatthehighHCO+/HCNlineratioob-
HCN, and HC N, can in fact be affected by radiative pumping served in starburst galaxies may be caused by different filling
3
viaIRvibrationalmodes.RadiativepumpingofHNChasbeen factorsforthetwomoleculesinsteadofbytheirabundance.
proposed by Aaltoetal. (2007) as an explanation of the high Although Kripsetal. (2008) suggest that the HCN/HCO+ line
HNC/HCNratioobservedinLIRGs,whilethestrongconnection ratiocanoftenserveasameasureofabundance,weadvisecau-
ofHC NexcitationwiththeIRcontinuumhasbeendiscussedin tion in doing so. Recent work by Sakamotoetal. (2009) and
3
Costagliola&Aalto(2010). Aaltoetal. (2009) show that in luminous, dusty nucleiabsorp-
Extreme caution should be used therefore when interpreting tion and radiative excitation affect the line emission of these
global line intensity ratios as abundance indicators. Molecular moleculesonglobalscales.
8 F.Costagliolaetal.:EMIRMolecularTracers
+og HCO/HCN J=1−010−−−−0000000.......01234321 a 1)0 9 4 154 1 1627 17 51113 3 8 121290 18 +og HCO/HCN J=1−010−−000...0242 b) 10 9 5 31 4152 14 8 2 1 0 176 311 72 1 18 19
L L
−0.5
−0.6
−0.6
−0.7
−0.8
−1.2 −1 −0.8 −0.6 −0.4 −0.2 0 0.2 0.4 −1 −0.8 −0.6 −0.4 −0.2 0
Log HNC/HCO+ J=1−0 Log HNC/HCN J=1−0
10 10
0.2
c) 19 0.4 d) 18
0
1 18 0.2
NC/HCN J=1−0 −−00..42 4 1 417 612 7 1131 15 3 8 220 +NC/HCO J=1−0 −−00..042 8 3 1 57 1 19 2 011 2 613
Log H10−−00..86 10 9 5 Log H10−−00..86 10 5 4 9
−1
−1
−1.2
−1.2 −1 −0.8 −0.6 −0.4 −0.2 0 0.2 0.4 −2 −1.5 −1 −0.5
Log HNC/HCO+ J=1−0 Log PAH equivalent width (6.2 µm) [µm]
10 10
0.2
e) 19 f) 8
0 0.1
18 1
−0.2 7 2 0 2 3
HNC/HCN J=1−0−−00..64 14 5 12 13 195 11 17 1 06 420 3 8 µIRAS 60/100 m−−00..21 1 18 2103 91 10 179 161 7 4 16 15
−0.8 −0.3 5 12
−1 −0.4 14
−0.5
−0.5 −0.4 −0.3 −0.2 −0.1 0 0.1 −1.8 −1.6 −1.4 −1.2 −1 −0.8 −0.6 −0.4
IRAS 60/100 µm IRAS 25/100 µm
Fig.4.Diagramsderivedfromobservedlineratios.SymbolsandgalaxylabellingarethesameasinFig.3.GalaxieswhereHC N
3
wasdetectedaremarkedwithacircle.ThediameterofthecircleisproportionaltotheHC N10–9/HCN1–0lineratio.Errorbars
3
show1-σuncertainties.
5.3.TheHNC/HCNratio HCN. Interestingly, Greaves&Nyman (1996) find HCN/HNC
abundance ratios of up to 6 for Galactic spiral arm clouds
In the Galaxy, the HNC/HCN observed line ratio ranges from (GMCs). The highest ratios are observed in regions of low-
1/100 in hot cores (Schilkeetal. 1992) to values as high est N(H ), consistent with models by Schilkeetal. (1992),
2
as 4 in dark clouds (Hirotaetal. 1998). The abundance of which show and increase of HCN/HNC with increasingC/CO.
the HNC molecule decreases with increasing gas temperature. BrightHNCemissioniscommonlyobservedinextragalacticob-
Hirotaetal.(1998)suggestthatthismaybeowingtothetemper- jects(Aaltoetal.2002;Wangetal.2004;Meier&Turner2005;
ature dependence of neutral-neutral reactions, which, for tem- Pe´rez-Beaupuitsetal. 2007). In particular, overluminous HNC
peraturesexceeding24K,selectivelydestroyHNCinfavourof
F.Costagliolaetal.:EMIRMolecularTracers 9
J=3–2 is found in LIRGs (e.g., Aaltoetal. 2007), where gas intermsofanascentstarburst.
temperatures, derived by the IR dust continuum (Evansetal. The HC N-bright galaxies reside in the HCO+/HCN interval
3
2003) and mid-IR molecular absorption (Lahuisetal. 2007), thatinthediagnosticdiagramsdiscussedbyKohnoetal.(2001)
are usually > 50 K and can reach values as high as a few is associated with Seyfert galaxies. Kohnoetal. (2001) and
100 K. Here, ion-moleculechemistry in PDRs may be respon- Imanishietal. (2007) interpret the observed ratios as the sig-
sible for the observed ratios. Meijerinketal. (2007) find that nature of XDR chemistry driven by the hard radiation from
the HNC/HCN J=1–0 line ratio is enhanced in PDRs and can the AGN. Our data seem to challenge this interpretation, since
reach a maximum value of one for H column densities ex- HC N cannotsurvivein significantabundancein highlyirradi-
2 3
ceeding 1022 cm−2. Even higher HNC/HCN line ratios can re- atedgas.BrightHC NemissioncanemergefromanAGNonly
3
sult from XDR emission. Models by Meijerinketal. (2007) ifthecentralsourceisshieldedbyaconsiderableamountofgas
show that for the J=4–3 transition, this ratio can be as high and dust. In these objects, the bulk of the molecular emission
as1.6indense(n(H)>106cm−3)X-ray-dominatedregions.The would be coming from an hot core-like region instead of be-
average HNC/HCN J=1–0 line ratio for AGN and LIRGs in ing dominated by XDR chemistry. Plots e and f in Fig. 4 also
our sample is ≃0.5 (see plot c in Fig. 4), which is consis- showthatthegalaxieswiththehighestHC N/HCNratioshave
3
tent with the results by Meijerinketal. (2007) for XDRs with highratiosofIRAS60to100µmfluxes.Thismayindicatethat
densities <105 cm−3 and high X-ray fluxes, F >10 erg cm−2 HC Nismoreabundantinwarmenvironments.
X 3
s−1. For the same sources, low-density XDR chemistry could An enhanced HC N emission at high gas temperatures could
3
also explain the observed HCO+/HCN ratios. The HNC/HCN also be owing to excitation effects. The upper state energy of
ratio of order unity observed in UGC 5101 and NGC 2273 theobservedHC Ntransitionsisindeedoftheorderof20-30K,
3
(see Table 2) is consistent with the emission coming from ei- higherbyafactorof≃5thantypicalenergiesforotherhighden-
ther a PDR or a low-density XDR (n <105 cm−3). Models sitytracers,meaningthatthemoleculeismoreefficientlyexcited
H
of PDRs with different density and radiation field strength can athighertemperatures.ThebrightHC Nemissionisthuslikely
3
explain HNC/HCN J=1-0 line ratios ranging from 0.5 to 1 tobeacombinationofenhancedabundanceandfavourableex-
(Meijerink&Spaans2005; Loenenetal. 2008). In oursample, citationconditionsinwarm,densegas.
theobservedHNC/HCNratioreachesvaluesaslowas0.2(plots Radiative pumping of the molecule’s rotational levels by
b, c in Fig. 4), which cannot be explained by existing PDR or means of IR vibrational modes was also suggested by
XDRmodels.However,lineratiosHNC/HCN<0.5areobserved Costagliola&Aalto(2010).
inGalacticPDRs.Abundanceratiosofabout0.2arefound,e.g.
,byFuenteetal.(1993)inthePDRregionofthereflectionneb-
5.5.CO/13COvsdusttemperature
ulaNGC7023.LowvaluesofHNC/HCNarealsofoundinother
galaxies (see , e.g., Baanetal. (2008)) as discussed in Section Plot g in Fig. 5 shows how the ratio ℜ ≡CO/13CO J=1–0
1−0
4.3. varies with dust temperature. The dust temperature was calcu-
latedfromIRASfluxesat100and60µm,viatheformula
5.4.LuminousHCN:anascentstarbursttracer?
3  82 
cBorrigesht(RHoCd3rNigueezm-Fisrsainocnoeistaol.fte1n998o;bsdeervVediceinnteGetaalal.cti2c00h0o)t. Tdust =−(1+z)ln(cid:16)0.3f60µm/f100µm(cid:17) −0.5, (2)
ThemoleculeiseasilydestroyedbyUVradiationandreactions see,e.g.,Solomonetal.(1997).TheIRASfluxeswereobtained
withC+ andHe+ (e.g., Turneretal.1998)andcanonlysurvive fromtheNASA/IPACExtragalacticDatabase.
in regions shielded by large gas and dust columns. In a high- A clear trend is observed, with ℜ increasing as T
1−0 dust
resolution study of the chemistry of IC 342, Meier&Turner increases. This is a well established result, see, e.g.,
(2005) find that HC N emission follows the 3 mm contin- Young&Sanders (1986) and Aaltoetal. (1995) for a detailed
3
uum dust emission and anti-correlates with regions of in- discussion. The trend is generally attributed to opacity effects
tense UV radiation,such as PDRs. Observationsin the Galaxy arising in a collection of non-uniform clouds, averaged in the
(Wyrowskietal.1999)andintheLIRGNGC4418(Aaltoetal. beam. In a molecular cloud, the emission of CO 1-0 arises
2007; Costagliola&Aalto 2010) show that HC N is strongly mostly from the outer layers, up to optical depths τ ≃1, while
3
connected with the IR field by means of its vibrational bend- the13COemissionstaysingeneralopticallythindeeperintothe
ing modes. In our sample, only five galaxies have detectable volumeofthecloud.Thelineratioℜ isthusmostlyaffected
1−0
HC N emission:NGC 4418,IRAS17208,IC 860,NGC 1068, bytherelativeopacityofthetwospecies.
3
andNGC7771.Forcompleteness,wealsoincludeinouranaly- TheopacityoftheCO 1-0line stronglydependsonthe excita-
sis the detection of HC N J=10–9 in the ULIRG Arp 220 by tiontemperature(see,e.g., Aaltoetal.1995, Eq.1).Withrising
3
Aaltoetal. (2002). Since these objects span a wide range of T ,the J=1levelisinfactde-populated,causingadecreasein
ex
LSR velocities, in most cases the detectability of HC N does optical depth of the CO 1-0 transition. Where τ ≃1, we can
3 CO
notdependonthesensitivityofourmeasurements,butonintrin- assumeτ <<1andtheintensityofthe13COJ=1–0lineisan
13CO
sicpropertiesoftheobservedgalaxies.Thisisfurtherdiscussed even steeper function of T than the 12CO intensity (see, e.g.,
ex
in Appendix A, where we compare HCN and HC N detection Mauersberger&Henkel 1993). As a consequence, ℜ must
3 1−0
thresholds. increasewithtemperature.
All galaxies with detected HC N have HCO+/HCN <1 and InourGalaxyYoung&Sanders(1986)findthatℜ hasanav-
3 1−0
higher-than-average HNC/HCN ratios. This is consistent with eragevalueofroughly8,while,inextragalacticsurveys,values
the hot core models described in Bayetetal. (2008), as dis- ℜ >20areobserved(e.g., Aaltoetal.1995, thiswork),the
1−0
cussedinSect.5.2.Itisthuspossiblethatthebulkofthemolec- highest ratios being found in interacting systems and mergers.
ular emission in these objects comes from regions of young In oursample,the highestℜ ≃ 28is foundforthe galaxies
1−0
star-formation. This interpretation was already proposed by NGC6240andNGC1614.Inthese objects,highgastempera-
Aaltoetal.(2007)toexplaintheHC NemissionofNGC4418 turesand a diffuse ISM may resultin an extremelylow optical
3
10 F.Costagliolaetal.:EMIRMolecularTracers
1.6 1.8
1.5 g) 10 1.6 h) 14
4 5
1.4
13CO/CO J=1−0 11..23 5 12 13 16 11 17 6 2 CO J=1−0/CNtot11..24 2 1 612 17 11 4 6
og 101.1 14 og 10 1 13 10
L L
1
18 0.8
0.9 18
0.8 0.6
1.55 1.6 1.65 1.7 1.75 1.8 1.85 1.2 1.4 1.6 1.8 2 2.2 2.4 2.6
Log Dust Temperature [K] Log CO/C18O J=1−0
10 10
Fig.5. SeecaptionofFig.4anddiscussioninSections5.5and5.6.
depthoftheCO1–0line,whichwouldexplaintheobservedra- sample. The observedcorrelationwould then be a result of
tios. opticaldeptheffects.
Additional processes have been proposed in order to explain – Itisalsopossibletospeculateonotherexplanations.Itisnot
these extreme ratios. In the central regions of merging and in- immediatelyself-evidentthattheC18OandCNcolumnden-
teracting galaxies, large amounts of gas and dust are trapped sitiesshouldfolloweachother.InPDRs,theCNmoleculeis
in the deep potential well, giving rise to high gas pressures. usuallyabundantupto a visualextinctionA =2. Athigher
V
Largeturbulentmotions,poweredbydifferentialrotationorstel- depths in the clouds the radical CN may react with other
lar windsfroman enhancedstar-formationare requiredto sus- moleculesto form– forexample– HCN. Thus,CN can,to
tainthecloudsagainstcollapse.Thiswouldcauseabroadening somedegree,tracetheamountofPDRsurfaces.
oftheCO emissionline andadecreaseinopticaldepth,witha In PDRs, it is likely that C18O could become photo-
consequentincreaseinℜ (Aaltoetal.1995). dissociated. However, if the CN is tracing PDRs, it is also
1−0
Chemistry could also provide an explanation of the observed tracing the impact of the star-formation on host clouds.
high ratios. In star-forming regions the surfaces of molecular ThesearetheregionswheretheISMenrichmentcausedby
cloudsare indeeddominatedby UV-photonsfromyoungstars. the starburst hits first., i.e., where C18O would become en-
Intheseregions,13COisdissociatedmoreeasilythanCO,which hancedearlyon.Hence,inthisscenario,aC18O–CNcorre-
is less affected by the UV radiation thanks to self-shielding lationwouldbeexpectedasaresultofISMenrichmentfrom
(Aaltoetal. 1995). This process could be relevant in starburst thestarburst.
galaxies,asNGC6240andNGC1614inoursample.
The age of the starburstcould also have an influence on ℜ1−0. In this context, it is interesting to note the distinct deviations
Nuclear processingin stars favours13CO overCO, and the ev- fromthiscorrelation.WehavethethreeoutlierswithfaintC18O
idence for an enhanced ℜ1−0 at low metallicities has been re- compared with their CN luminosities: NGC 4194, NGC 1614
ported by, e.g., Young&Sanders (1986). A large ℜ1−0 would andNGC7469.Apossiblewaytoobtainthiscomparableover-
bethuscompatiblewithayoungstar-formationage. luminosityinCNistohaveCNemergingfromanAGNcompo-
nent,withnoassociatedenrichmentfromastarburstcomponent.
TheC18Oenrichmentshouldcomequicklyinastarburst.If18O
5.6.CO/CNvsCO/C18O isaproductofα-captureon14NduringHeburning,itwillcome
frommasslossinHe-burningsystems,suchasWolf-Rayetstars
Plot h in Fig. 5 showsthe ratio CO/CN vsCO/C18O. The total
andSNetypeII.Thismeansthatitshouldcomeupwithprompt
CN intensity was obtained by adding the integrated intensities
nucleosynthesis–e.g.,followOabundance.
of the two lines of the spin doublet J=3/2-1/2 and J=1/2-1/2.
Because the molecular emission from NGC 7469is dominated
The ratio CO/CN varies in our sample over roughly one order
bythestarburstring,thiscomparativelyC18Oweaknessissome-
of magnitude and shows a general trend of increasing CO/CN
whatpuzzlingandshouldbefurtherinvestigated.Theobserved
withincreasingCO/C18O.ThisimpliesthatCNandC18Oemis-
highCO/C18Oratiocouldalsobecausedbyalowopticaldepth
sionarepositivelycorrelated.Thereareseveralpotentialexpla-
oftheCO1–0line.
nationsofthis:
The other two outliers are different since they are both merg-
erswheretheCO emissionisdominatedbyaprominentcross-
– WecanassumethatbothC18OandCNareopticallythin,for ing dust lane and not the starburst region (Olssonetal. 2010;
CN weknowthisfromtheCN 1–0spingroupratio,which Aalto&Hu¨ttemeister 2000). It is therefore possible that the
is at the optically thin level for all detected galaxies (apart moleculargasintheIRAMbeamisdominatedbyin-fallinggas
from,possibly,UGC5101).Inthiscase,theintensityofCN inthedustlane,whichhasnotyetbeenenrichedbythestarburst.
andC18Owouldscalewiththetotalmolecularcolumnden- Indeed,Mart´ınetal.(2010)findhigh13C/CratiosfromCOand
sity. This linear scaling with total column would not hap- CCHobservationsinthestarburstsM82andNGC253,suggest-
penfortheCO1–0intensity,whichcanbeconsideredtobe ingthatthebulkofthemolecularmassinthesegalaxiesmostly
at least moderatelyoptically thick in all the galaxiesin our consistsofunprocessedmaterial.