CON-quest: Searching for the most obscured galaxy nuclei
N. Falstad, S. Aalto, S. König, K. Onishi, S. Muller, M. Gorski, M. Sato, F. Stanley, F. Combes, E. González-Alfonso, J. G. Mangum, A. S. Evans, L. Barcos-Muñoz, G. C. Privon, S. T. Linden, T. Díaz-Santos, S. Martín, K. Sakamoto, N. Harada, G. A. Fuller, J. S. Gallagher, P. P. van der Werf, S. Viti, T. R. Greve, S. García-Burillo, C. Henkel, M. Imanishi, T. Izumi, Y. Nishimura, C. Ricci, S. Mühle
AAstronomy & Astrophysics manuscript no. conquest_acc © ESO 2021March 1, 2021
CON-quest
Searching for the most obscured galaxy nuclei
N. Falstad , S. Aalto , S. König , K. Onishi , S. Muller , M. Gorski , M. Sato , F. Stanley , F. Combes ,E. González-Alfonso , J. G. Mangum , A. S. Evans , , L. Barcos-Muñoz , G. C. Privon , S. T. Linden ,T. Díaz-Santos , , , S. Martín , , K. Sakamoto , N. Harada , , G. A. Fuller , , J. S. Gallagher ,P. P. van der Werf , S. Viti , , T. R. Greve , , S. García-Burillo , C. Henkel , , M. Imanishi , T. Izumi ,Y. Nishimura , , C. Ricci , , , and S. Mühle (A ffi liations can be found after the references) ABSTRACT
Context.
Some luminous and ultraluminous infrared galaxies (LIRGs and ULIRGs) host extremely compact ( r <
100 pc) and dusty nuclei. Thehigh extinction associated with large column densities of gas and dust toward these objects render them hard to detect at many wavelengths. Theintense infrared radiation arising from warm dust in these sources can provide a significant fraction of the bolometric luminosity of the galaxyand is prone to excite vibrational levels of molecules such as HCN. This results in emission from the rotational transitions of vibrationally excitedHCN (HCN-vib), with the brightest emission found in compact obscured nuclei (CONs; Σ HCN − vib > (cid:12) pc − in the J = Aims.
We aim to establish how common CONs are in the local Universe ( z < . Methods.
We have conducted an Atacama Large Millimeter / submillimeter Array (ALMA) survey of the rotational J = Results.
Compact obscured nuclei are identified in 38 + − % of the ULIRGs, 21 + − % of the LIRGs, and 0 + − % of the lower luminosity galaxies. Wefind no dependence on the inclination of the host galaxy, but strong evidence of lower IRAS 25 to 60 µ m flux density ratios ( f / f ) in CONs(with the exception of one galaxy, NGC 4418) compared to the rest of the sample. Furthermore, we find that CONs have stronger silicate features( s . µ m ) but similar polycyclic aromatic hydrocarbon equivalent widths (EQW . µ m ) compared to other galaxies. Besides signatures of molecularinflows seen in the far infrared in most CONs, submillimeter observations also reveal compact, often collimated, outflows. Conclusions.
In the local Universe, CONs are primarily found in (U)LIRGs, in which they are remarkably common. As such systems are oftenhighly disturbed, inclinations are di ffi cult to estimate, and high resolution continuum observations of the individual nuclei are required to determineif the CON phenomenon is related to the inclinations of the nuclear disks. Further studies of the in- and outflow properties of CONs should also beconducted in order to investigate how these are connected to each other and to the CON phenomenon. The lower f / f ratios in CONs as well asthe results for the mid-infrared diagnostics investigated (EQW . µ m and s . µ m ) are consistent with the notion that large dust columns gradually shiftthe radiation from the hot nucleus to longer wavelengths, making the mid- and far-infrared “photospheres” significantly cooler than the interiorregions. Finally, to assess the importance of CONs in the context of galaxy evolution, it is necessary to extend this study to higher redshifts where(U)LIRGs are more common. Key words. galaxies: evolution – galaxies: nuclei – galaxies: ISM – ISM: molecules – ISM: jets and outflows
1. Introduction
Over the last decade, it has been found that some luminous(10 L (cid:12) < L IR (8 − µ m) < L (cid:12) ) and ultraluminous( L IR > L (cid:12) ) infrared galaxies (LIRGs and ULIRGs) in thelocal Universe exhibit emission from rotational lines of HCNin its vibrationally excited v = Σ HCN − vib > (cid:12) pc − in the J = r > ff ect that mayincrease the central dust temperatures to several hundreds ofKelvin (González-Alfonso & Sakamoto 2019). These high tem-peratures give rise to an intense mid-infrared radiation field that is able to populate the vibrational states of molecules such asHCN. With sizes on the order of 10 to 100 pc and dust temper-atures around 100 K at the far-infrared photosphere, CONs maybe able to supply a significant fraction of the total infrared lumi-nosity of their host galaxies (e.g., González-Alfonso et al. 2012;Falstad et al. 2015). Due to the obscured nature of their nuclei, itis still unclear what the ultimate embedded source of the high lu-minosity in CONs is; it may be an accreting supermassive blackhole (SMBH) in an active galactic nucleus (AGN), a nuclear star-burst, or a combination of both. If CONs are mainly powered byhidden AGN activity, they may represent a phase of rapid ac-cretion onto the SMBH, almost completely surrounded by highcolumn densities of obscuring material, following a merger orinteraction event (Kocevski et al. 2015; Ricci et al. 2017; Blechaet al. 2018; Boettcher et al. 2020). Regardless of the exact na-ture of the hidden power source, studies of these objects couldhelp us understand growth processes in galaxy nuclei, as well asthe relations between black hole mass and bulge properties (e.g., Article number, page 1 of 16 a r X i v : . [ a s t r o - ph . GA ] F e b & A proofs: manuscript no. conquest_acc
Magorrian et al. 1998; Kormendy & Ho 2013). An interestingcomparison may be made with the compact star-forming galax-ies found at redshifts 2 − ffi cult. The radiative trans-fer of traditional tracers of dense molecular gas such as HCNand HCO + is complicated by absorption due to large columns ofcooler foreground gas along the line of sight to the nucleus (e.g.,Aalto et al. 2015b; Martín et al. 2016; Imanishi et al. 2016b).In these situations, HCN-vib provides a tool to probe deeperinto the nucleus (e.g., Martín et al. 2016; Aalto et al. 2019), but itcan also be used as a survey tool to search for obscured activity(e.g., Aalto et al. 2019). The rotational transitions of HCN-viboccur within the vibrationally excited v = E / k = J = v = f transition which we use in this paperoccurs at a frequency 267 .
199 GHz. The critical density of the v = ∼ × cm − , high enough for colli-sional excitation to be unlikely. Instead, the vibrational state islikely populated through radiative excitation directly from theground state by 14 µ m photons (Ziurys & Turner 1986). Forthe state to be e ffi ciently populated in this way, radiation witha brightness temperature in excess of 100 K at 14 µ m is required(Aalto et al. 2015b). Since the first extragalactic detection bySakamoto et al. (2010), di ff erent HCN-vib transitions have beenobserved in many other galaxies (e.g., Imanishi & Nakanishi2013; Imanishi et al. 2016b,c, 2018; Aalto et al. 2015a,b; Martínet al. 2016). E ff orts have been made to make inventories of theseexisting extragalactic detections. For example, in a sample ofnine galaxies, mainly (U)LIRGs, by Aalto et al. (2015b), HCN-vib lines were detected in eight. Five of these could be classifiedas CONs with the definition used at the time. A follow-up studyby Falstad et al. (2019) found 19 (U)LIRGs with existing obser- vations covering the HCN-vib frequencies. Of these, 13 were de-tected and 7, including the 5 found by Aalto et al. (2015b), couldbe classified as CONs. However, there have been no systematicsearches for CONs, and an understanding of the prevalence ofCONs will help assess their importance to the galaxy evolutionprocess.In this paper, we present the first results of a volume limitedsurvey of infrared luminous galaxies, named CON-quest, withthe aim to establish how common CONs are in the local ( z < .
08) Universe. We introduce our selection criterion in Sect. 2and the sample in Sect. 3, describe the new observations and datareduction in Sect. 4, and present the results in Sect. 5. In Sect.6 we discuss the results before summarizing our conclusions inSect. 7.
2. How to find compact obscured nuclei
Compact obscured nuclei are characterized by high column den-sities of warm dust and gas within relatively small distancesof their centers, typically much smaller than the size of centralmolecular zones which have radii of several 100 pc (e.g., Israel2020). In principle, this means that properties such as the columndensity and temperature of obscuring material, and the physicalsize of the nucleus must be determined to identify CONs. Eachsource would then require detailed radiative transfer modelingand excitation analysis (e.g., González-Alfonso et al. 2012),or high-resolution multifrequency continuum observations (e.g.,Aalto et al. 2019). However, with the discovery of bright HCN-vib emission from known CONs, a new tool to be used in surveyssuch as this one was provided.Previous studies of CONs have often focused on the ratio be-tween the HCN-vib luminosity and the total infrared luminosityof the galaxy, with CONs generally having L HCN − vib / L IR > − (e.g., Aalto et al. 2015b; Falstad et al. 2019). While this criterionis easy to use, it has a significant drawback in that it includesboth emission originating mainly in the nucleus ( L HCN − vib ) andemission that is likely to have a substantial contribution from theentire host galaxy ( L IR ). This means that it may miss CONs insystems where the infrared emitting region is spatially extendedor where multiple nuclei are present. Instead of using this crite-rion, we identify as CONs galaxies with Σ HCN − vib > (cid:12) pc − in the J = . Σ HCN − vib can be esti-mated by comparing the extents of the continuum and HCN-vibemission in sources where the latter is resolved and not a ff ectedby line blending. On doing so, we find that the surface brightnessof HCN-vib is generally underestimated by less than a factor oftwo (1 . . . N H (cid:38) × cm − toward the central ∼
50 pc (González-Alfonso et al.2012; Falstad et al. 2015), have Σ HCN − vib > (cid:12) pc − (Aaltoet al. 2015b; Martín et al. 2016, ; this work) while Mrk 231,which is generally not considered a CON due to its lower columndensities ( N H ∼ cm − ) on similar scales (González-Alfonsoet al. 2014), has Σ HCN − vib ∼ . (cid:12) pc − (Aalto et al. 2015a). For Article number, page 2 of 16. Falstad et al.: CON-quest further comparison, our limit is approximately two orders ofmagnitude larger than the HCN-vib surface brightness seen inthe super star clusters of NGC 253 (Krieger et al. 2020).Due to the complex dependence of the HCN-vib emission onother properties of the source (González-Alfonso & Sakamoto2019), a measured HCN-vib surface brightness can not be di-rectly translated into physical properties such as the column den-sity and size of an observed source. However, the examples in theprevious paragraph may give some indication of the compact-ness and amount of obscuration in sources selected by our crite-rion. Furthermore, some properties can be estimated from the re-cent work by González-Alfonso & Sakamoto (2019) who modelthe continuum and HCN-vib emission from buried nuclear re-gions of galaxies. The models employ a spherically symmetricapproach, simulating either an AGN or a nuclear starburst as theheating source and exploring the ranges N H = –10 cm − for the H column density and Σ IR = . × –2 . × L (cid:12) pc − for the surface brightness. For the typical luminosity of a LIRGor ULIRG in our sample, this corresponds to radii of ∼ ∼ Σ HCN − vib > (cid:12) pc − in the J = N H > cm − , corresponding to an opticaldepth of ∼ . µ m. We note that this criterion, and in factany criterion based on the HCN-vib luminosity, assumes that,apart from large columns of obscuring material, a CON also hasa power source that is strong enough to produce the mid-infraredradiation field required to excite HCN-vib.
3. Sample
As mentioned in Sect. 1, no systematic searches for CONs havebeen carried out and only seven have been found so far: two inULIRGs, five in LIRGs and none in sub-LIRGs (10 L (cid:12) ≤ L IR < L (cid:12) ; Aalto et al. 2015b; Falstad et al. 2019). Due to the lim-ited statistical data, it is still uncertain how common the CONsare, and how their prevalence depends on luminosity. As a firststep to remedy this, we have compiled a sample of far-infraredluminous galaxies drawn from the IRAS revised bright galaxysample (RBGS; Sanders et al. 2003). The RBGS is a completesample of extragalactic objects with IRAS 60 µ m flux densitiesgreater than 5 .
24 Jy, covering the entire sky surveyed by IRAS atGalactic latitudes | b | > ◦ . Our CON-quest sample was selectedfrom the RBGS based on the far-infrared luminosities, declina-tions, and distances listed in Table 1 of Sanders et al. (2003)using the following far infrared luminosity, L FIR (40 − µ m),criteria: – L (cid:12) ≤ L FIR , δ < ◦ , D <
330 Mpc – L (cid:12) ≤ L FIR < L (cid:12) , δ < ◦ , D <
76 Mpc – L (cid:12) ≤ L FIR < L (cid:12) , δ < ◦ , 10 < D < . t int (cid:46) ≤
20 in each lu-minosity bin) to make the survey feasible. Calculating the far-infrared luminosity using the prescription in Table 1 of Sanders& Mirabel (1996) and the relatively conservative assumptions f = × f and C = .
8, we note that the full RBGS is vol-ume limited down to the relevant far-infrared luminosities forthe low-, and mid-luminosity bins. For the distance limit in thehigh-luminosity bin, however, the RBGS is only volume limitedfor L FIR (cid:38) . × L (cid:12) . On comparing with the more sensitive IRAS 1 . µ m), the high-, mid-, andlow-luminosity bins contain ULIRGs, LIRGs, and sub-LIRGs,which are defined using total infrared luminosity (between 8and 1000 µ m), respectively, and we will refer to them by thesenames. In total, our criteria select 48 systems (some of whichcontain multiple galaxies), eight in the ULIRG-sample and 20each in the LIRG and sub-LIRG samples. However, observationshave not been obtained of the galaxy pairs IC 4687 / / The selection based on far-infrared luminosity (between 40 and400 µ m) instead of total infrared luminosity (between 8 and1000 µ m) may introduce a bias against warm sources that emit alarger fraction of their radiation at shorter wavelengths. One wayto investigate this potential infrared emission bias is to comparethe distributions of the IRAS 25 to 60 µ m flux density ratios( f / f ) in our sample and the IRAS RBGS from which it wasselected. In Fig. 1 we make this comparison: the top panel showsthe distribution in the RBGS, and the middle panel shows thedistribution for the CON-quest sample using a dark shade. The f / f distribution of the CON-quest sample is clearly skewedtoward lower values when compared to the RBGS. In the mid-dle panel we have also included the sources required, togetherwith the CON-quest sample, to complete a sample selected basedon the total (instead of far-infrared) infrared luminosities. Thedistribution of this combined sample is more similar to that ofthe underlying RBGS, as can also be seen in the bottom panelwhere we show the empirical distribution functions of the di ff er-ent samples.
4. Observations and data reduction
For this survey, we targeted the HCN-vib l = f J = .
199 GHz. In the literature andthe Atacama Large Millimeter / submillimeter Array (ALMA)archives, we found pre-existing observations of this transition infour ULIRGs, six LIRGs, and one sub-LIRG. Three sub-LIRGshad ALMA observations of the corresponding J = J = New observations of the HCN-vib l = f J = − The four sources observed as part of thisproject were the ULIRGs IRAS 09022-3615, IRAS 13120-5453,
Article number, page 3 of 16 & A proofs: manuscript no. conquest_acc
Table 1.
Sample galaxies
Name RA Dec z D La L FIR b L IR c f / f d Incl. e (J2000) (J2000) (Mpc) (10 L (cid:12) ) (10 L (cid:12) ) ( ◦ )ULIRGsIRAS 17208-0014 17:23:21.95 -00:17:00.9 0.0428 183 ±
12 22 25 ± ±
24 18 20 ± + + ±
21 17 19 ± ± ± ±
17 15 18 ± + . ± . ± ±
20 12 14 ± . . . IRAS F22491-1808 22:51:49.31 -17:52:24.0 0.0778 328 ±
22 12 13 ± . . . LIRGsNGC 1614 04:34:00.03 -08:34:44.6 0.0159 64 . ± . . . ± . + . ± . . . ± . . ± . . . ± . . ± . . . ± . . ± . . . ± . . ± . . . ± . . ± . . . ± . . ± . . . ± . . ± . . . ± . + . ± . . . ± . . ± . . . ± . . ± . . . ± . . ± . . . ± . + . ± . . . ± . . ± . . . ± . . ± . . . ± . . ± . . . ± . . ± . . . ± . . ± . . . ± . f . ± . .
78 2 . ± . . ± . .
35 0 . ± .
08 0.16 83NGC 4254 12:18:49.63 + . ± . .
26 0 . ± .
10 0.12 20NGC 4303 12:21:54.95 + . ± . .
23 0 . ± .
10 0.13 18NGC 660 01:43:02.35 + . ± . .
24 0 . ± .
04 0.11 79NGC 4527 12:34:08.50 + . ± . .
19 0 . ± .
10 0.11 81NGC 3627 11:20:15.03 + . ± . .
17 0 . ± .
05 0.13 68NGC 613 01:34:18.24 -29:25:06.6 0.0049 15 . ± . .
17 0 . ± .
09 0.16 36NGC 4666 12:45:08.68 -00:27:42.9 0.0051 12 . ± . .
17 0 . ± .
09 0.10 70NGC 1792 05:05:14.45 -37:58:50.7 0.0040 12 . ± . .
17 0 . ± .
09 0.12 63NGC 4501 12:31:59.22 + . ± . .
17 0 . ± .
08 0.15 63NGC 4536 12:34:27.13 + . ± . .
15 0 . ± .
08 0.13 73NGC 5643 14:32:40.78 -44:10:28.6 0.0040 14 . ± . .
12 0 . ± .
03 0.20 30NGC 3628 11:20:17.02 + . ± . .
14 0 . ± .
04 0.09 79NGC 1559 04:17:35.78 -62:47:01.3 0.0043 12 . ± . .
13 0 . ± .
07 0.10 60NGC 5248 13:37:32.07 + . ± . .
12 0 . ± .
06 0.14 56NGC 3810 11:40:58.74 + . ± . .
10 0 . ± .
05 0.12 48NGC 4654 12:43:56.64 + . ± . .
10 0 . ± .
05 0.13 60NGC 1055 02:41:45.18 + . ± . .
10 0 . ± .
05 0.12 63
Notes. ( a ) Luminosity distances calculated from the redshifts following the procedure outlined by Sanders et al. (2003). For consistency, in galaxieswhere Sanders et al. (2003) instead use direct primary or secondary distance measurements, we use the same distances as they do. For the samereason, a H =
75 km s − Mpc − , Ω m = .
3, and Ω Λ = . ( b ) Far-infrared luminosities used for selection, taken fromSanders et al. (2003). ( c ) Infrared luminosities calculated using the prescription in Table 1 of Sanders & Mirabel (1996, originally from Pérault1987) using IRAS fluxes taken from Sanders et al. (2003). ( d ) Ratio of IRAS fluxes at 25 and 60 µ m taken from Sanders et al. (2003). ( e ) Inclinationsof the host galaxies taken from the HyperLEDA database a (Makarov et al. 2014). ( f ) NGC 1068 is a LIRG, but as we selected our samples basedon L FIR , which is lower, it is included in the sub-LIRG sample. This is the only border case. a http: // leda.univ-lyon1.fr / Article number, page 4 of 16. Falstad et al.: CON-quest . . . . . . f / f N u m b e r o f s ou r ce s IRAS RBGS0 . . . . . . f / f N u m b e r o f s ou r ce s CON-questnon-CON-quest0 . . . . . . f / f . . . . . . C u m u l a ti v e p r ob a b ilit y IRAS RBGSCON-questnon-CON-questL IR sample Fig. 1.
Distributions of IRAS 25 / µ m ratios in various samples. Top: histogram showing the distribution in the IRAS revised bright galaxysample (RBGS).
Middle: stacked histogram showing the distribution inthe sample that would have resulted if we had selected sources basedon the total infrared luminosity (between 8 and 1000 µ m) instead ofthe far-infrared luminosity (between 40 and 400 µ m), with the CON-quest sample marked using a darker shade. Bottom: empirical distribu-tion functions of the RBGS (solid black), the CON-quest sample (dash-dotted gray), the additional sources required to complete an infraredselected sample (dotted gray), and the total infrared selected sample(dashed red).
IRAS F14378-3651 and IRAS 17208-0014. The observationstook place in September 2018 with the intermediate array con-figuration C43-5 (baseline lengths between 15 m and 2 km). Themaximum recoverable scale (MRS, defined as 0 . λ/ B min , where B min is the minimum baseline length) is ∼ (cid:48)(cid:48) (6 −
13 kpc).
Thirteen LIRGs (ESO 173-G015, ESO 221-IG10, ESO 286-G035, ESO 320-G030, IC 4734, IC 5179,IRAS F17138-1017, IRAS 17578-0400, NGC 2369, NGC 3110, NGC 5135, NGC 5734, UGC 2982) and 15 sub-LIRGs(NGC 660, NGC 1055, NGC 1559, NGC 1792, NGC 3627,NGC 3628, NGC 3810, NGC 4254, NGC 4303, NGC 4501,NGC 4527, NGC 4536, NGC 4654, NGC 4666, NGC 5248)were observed for this project between October 2018 and Oc-tober 2019. The observations were carried out with the inter-mediate array configurations C43-4 and C43-5 with baselinelengths between 15 m and 2 . ∼ (cid:48)(cid:48) (0 . − . + ( ν rest = − ν rest = .
875 GHz, and a channelwidth of 3 . / gain calibrators.The imaging of the visibility sets for all 32 sources was per-formed using the ”tclean” task in the Common Astronomy Soft-ware Applications (CASA ; McMullin et al. 2007) package ver-sion 5.4.0. Applying a natural weighting resulted in data cubeswith beam sizes of roughly 0.3 (cid:48)(cid:48) (spatial scales between ∼
45 and ∼
95 pc) for the ULIRGs in 2017.1.00759.S, and beam sizes be-tween ∼ (cid:48)(cid:48) and ∼ (cid:48)(cid:48) (spatial scales between ∼ ∼
35 pc)for the LIRGs and sub-LIRGs in 2018.1.01344.S (see Table 2 formore information). For analysis purposes we Hanning-smoothedthe data cubes to a velocity resolution of 20 km s − . The result-ing 1- σ noise levels per channel for each source are listed inTable 2. After calibration and imaging within CASA, all imagecubes were converted into FITS format for further analysis.To obtain the HCN-vib fluxes and upper limits in sourceswith relatively narrow well separated spectral lines, we inte-grated the data cubes over the velocity range ±
150 km s − aroundthe systemic velocity of the transition, calculated from the red-shifts given in Table 1 in the heliocentric frame. This velocityrange was set to avoid blending issues with the HCO + tran-sition that is blueward of the HCN-vib line by approximately400 km s − . We then extracted the fluxes from a region with adiameter of twice the Gaussian full width at half maximum ofthe velocity integrated HCO + emission. To compensate for lineblending in sources with strong and broad lines, we instead ob-tained the flux through a fit to the spatially integrated spectrumextracted from the same region. In the case of a nondetection ofthe HCN-vib line, we set a 3- σ upper limit to the flux based onthe rms in the velocity integrated map and the line width of theground state transitions.We obtained estimates for the continuum properties fromtwo-dimensional Gaussian fits to the continuum maps using theCASA task IMFIT which deconvolves the synthesized beamfrom the fitted component size, see Condon (1997) for a discus-sion of the error estimates in this process. In sources with multi-ple continuum components, we use the results from the strongestone, as this is where a potential CON would be most likely to befound. Details about the observation setups and data reduction forsources with pre-existing data can be found in the referenceslisted in the last column of Table 3. For previous ALMA obser-vations, we have not used the published values for the HCN-vib http: // casa.nrao.edu / Article number, page 5 of 16 & A proofs: manuscript no. conquest_acc
Table 2.
Properties of the new ALMA observations.
Name Beam size a Sensitivity a,b ( (cid:48)(cid:48) ) (mJy beam − )ULIRGsIRAS 17208-0014 0 . × .
33 0 . . × .
33 0 . . × .
29 0 . . × .
27 0 . . × .
31 0 . . × .
30 0 . . × .
30 0 . . × .
28 0 . . × .
30 0 . . × .
28 0 . . × .
32 0 . . × .
24 0 . . × .
31 0 . . × .
27 0 . . × .
26 0 . . × .
29 0 . . × .
30 0 . . × .
52 0 . . × .
55 0 . . × .
52 0 . . × .
64 0 . . × .
52 0 . . × .
53 0 . . × .
54 0 . . × .
52 0 . . × .
57 0 . . × .
52 0 . . × .
44 0 . . × .
54 0 . . × .
55 0 . . × .
54 0 . . × .
49 0 . Notes. ( a ) With natural weighting of the interferometric visibilities. ( b ) Given as 1- σ rms for channel widths of 20 km s − . line fluxes in cases where these exist. We have instead extractedthese in the same way as for newly observed sources, using thecalibrated data sets obtained from the principal investigators ofthe projects in question. Some di ff erences in the instrumentalsetup compared to our new observations exist, but none that arelarge enough to impact the appropriate extraction procedure.
5. Results
Luminosities of HCN-vib, continuum properties, and HCN-vibsurface brightnesses are presented in Table 3. For comparisonwith earlier work, the L HCN − vib / L IR ratio is also listed. Line lu-minosities are calculated using Eq. 1 of Solomon & Vanden Bout(2005), applied to HCN-vib: L HCN − vib = . × − S HCN − vib ∆ v ν rest (1 + z ) − D , (1)where L HCN − vib is the HCN-vib luminosity measured in L (cid:12) , S HCN − vib ∆ v is the velocity integrated flux in Jy km s − , ν rest is the rest frequency in GHz, and D L is the luminosity distance inMpc.In total, emission from vibrationally excited HCN is detectedin five of the eight ULIRGs, five of the 19 LIRGs, and in none ofthe 19 sub-LIRGs. The corresponding numbers when consider-ing individual nuclei resolved by our observations are six of 12in ULIRGs and five of 20 in the LIRGs. For three of the LIRGs,ESO 320-G030, ESO 173-G015, and IRAS 17578-0400, and oneof the ULIRGs, IRAS F14378-3651, this is the first detection ofvibrationally excited HCN. Another ULIRG, IRAS 17208-0014,has previously been detected in the HCN-vib J = J = J = . (cid:12) , has recently been detected in NGC 1068 by Iman-ishi et al. (2020). In Fig. 2, we present the spectra of sourceswith new detections of HCN-vib J = + transition. Self-and continuum absorption by the latter (see Aalto et al. 2015b,for a more thorough discussion) complicates the situation fur-ther, warranting a short explanation of the features in this spec-trum. The peaks on each side of the central frequency of theHCO + transition are both part of the HCO + feature. The peculiarshape is caused by an emission component, peaking close to thecentral frequency of the transition, combined with an absorptioncomponent that has its maximum at a slightly higher frequency.Finally, the prominent wing on the low-frequency side of theHCO + transition is HCN-vib emission. Using the criterion Σ HCN − vib > (cid:12) pc − in the J = + − % in the ULIRG sample, 21 + − % in the LIRGsample, and 0 + − % in the sub-LIRG sample. If we instead con-sider individual resolved nuclei, the detection rates are 25 + − %in the ULIRG sample and 20 + − % in the LIRG sample. The 1- σ confidence intervals were estimated using the beta distribu-tion quantile technique (Cameron 2011). For reference, with theold criterion, L HCN − vib / L IR > − , the same seven CONs wouldhave been identified, together with an additional one in IRASF12112 + Armed with the CON detection rates, we can explore links withthe properties of their host galaxies. As there were no detectionsin the sub-LIRG sample we only include the (U)LIRGs in thiscomparison.Falstad et al. (2019) suggested that galaxies with high incli-nation may be more likely to also have high L HCN − vib / L IR ratios,and thus be classified as CONs. To assess whether CONs arepreferentially found in high inclination systems, we can comparethe distribution of inclinations for CONs with that of the restof the sample galaxies. To do this, we use optical estimates ofthe inclinations taken from the HyperLEDA database (Makarovet al. 2014). An important caveat is that inclinations are di ffi cult http: // leda.univ-lyon1.fr / Article number, page 6 of 16. Falstad et al.: CON-quest
Table 3.
Results of HCN-vib J = Name L HCN − vib S cont . Continuum size Σ HCN − vib L HCN − vib / L IR Ref.(10 L (cid:12) ) (mJy) (mas × mas) (L (cid:12) pc − ) (10 − )ULIRGsIRAS 17208-0014 62 . ± . . ± . ± × ± . ± .
38 2 . ± . < .
776 1 . ± .
21 145 ± × ± < . < .
283 2IRAS F14348-1447 SW < .
17 2 . ± .
18 113 ± × ± < . < .
451 2IRAS F12112 + . ± . . ± .
24 201 ± × ±
41 0 . ± .
11 1 . ± . + < .
071 1 . ± .
22 445 ± × ± < . < .
481 3IRAS 13120-5453 < .
99 32 . ± .
88 869 ± × ± < . < .
108 1IRAS 09022-3615 < .
552 5 . ± .
16 601 ± × ± < . < .
308 1Arp 220 W 61 . ± . . ± . ± × ±
17 11 . ± .
31 3 . ± . . ± . . ± . ± × ±
30 0 . ± .
22 0 . ± .
06 4IRAS F14378-3651 E 3 . ± . . ± .
08 350 ± × ±
17 0 . ± .
003 0 . ± .
05 1IRAS F14378-3651 W < . . ± .
04 345 ± × ± < . < .
082 1IRAS F22491-1808 a . ± . . ± .
07 66 ± × ± . ± .
31 2 . ± . < .
425 16 . ± . ± × ± < . < .
105 3NGC 7469 < .
121 3 . ± .
38 342 ± × ± < . < .
031 3NGC 3256 S < .
065 9 . ± .
79 1247 ± × ± < . < .
017 5NGC 3256 N < .
065 17 . ± . ± × ± < . < .
017 5IRAS F17138-1017 < .
247 3 . ± . ± × ± < . < .
095 1IRAS 17578-0400 15 . ± . . ± .
71 227 ± × ±
10 4 . ± .
72 6 . ± .
15 1NGC 7130 < .
203 13 . ± .
77 745 ± × ± < . < .
089 6ESO 173-G015 1 . ± . . ± . ± × ±
13 0 . ± .
03 0 . ± .
06 1NGC 3110 < .
157 1 . ± .
14 987 ± × ± < . < .
077 1IC 4734 < .
18 10 . ± .
84 767 ± × ± < . < .
095 1Zw 049.057 11 . ± . . ± . ± × ±
61 3 . ± .
68 6 . ± . < .
126 3 . ± .
42 669 ± × ± < . < .
085 1ESO 221-IG10 < .
143 1 . ± .
25 1130 ± × ± < . < .
098 1IC 5179 < .
055 1 . ± .
10 408 ± × ± < . < .
038 1UGC 2982 < .
124 0 . ± .
07 800 ± × ± < . < .
088 1ESO 286-G035 < .
15 2 . ± .
22 1213 ± × ± < . < .
112 1NGC 4418 a . ± . . ± . ± × ± . ± .
23 3 . ± . < .
075 11 . ± . ± × ± < . < .
06 1NGC 5734 < .
115 0 . ± .
09 638 ± × ± < . < .
101 1ESO 320-G030 2 . ± . . ± . ± × ± . ± . . ± . < .
01 13 . ± . ± × ± < . < .
004 9NGC 1808 b < .
003 . . . . . . . . . < .
005 10,11NGC 4254 < .
005 0 . ± .
08 685 ± × ± < . < .
016 1NGC 4303 < .
005 1 . ± .
08 855 ± × ± < . < .
016 1NGC 660 < .
004 28 . ± . ± × ± < . < .
014 1NGC 4527 < .
009 2 . ± .
23 1388 ± × ± < . < .
033 1NGC 3627 < .
003 7 . ± .
24 984 ± × ± < . < .
014 1NGC 613 b < .
004 . . . . . . . . . < .
018 10,12NGC 4666 < .
006 1 . ± .
11 1383 ± × ± < . < .
025 1NGC 1792 < .
004 1 . ± .
09 704 ± × ± < . < .
021 1NGC 4501 < .
006 1 . ± .
11 438 ± × ± < . < .
026 1NGC 4536 < .
008 4 . ± .
31 2166 ± × ± < . < .
039 1NGC 5643 b < .
006 . . . . . . . . . < .
034 13NGC 3628 < .
003 38 . ± . ± × ± < . < .
017 1NGC 1559 < .
006 0 . ± .
02 793 ± × ± < . < .
038 1NGC 5248 < .
007 4 . ± .
29 1077 ± × ± < . < .
045 1NGC 3810 < .
007 0 . ± .
08 1380 ± × ± < . < .
050 1NGC 4654 < .
006 0 . ± .
10 1230 ± × ± < . < .
044 1NGC 1055 < .
004 3 . ± .
28 994 ± × ± < . < .
034 1
Notes. ( a ) Continuum properties from observations by Imanishi et al. (2018) and Sakamoto et al. (in prep.) ( b ) Observed in the J = References. (1) This work; (2) Imanishi et al. (2019); (3) Imanishi et al. (2016b); (4) Martín et al. (2016); (5) Harada et al. (2018); (6) Privon, priv.comm.; (7) Aalto et al. (2015b); (8) Sakamoto et al. (2010); (9) Imanishi et al. (2016a); (10) Combes et al. (2019); (11) Audibert et al. (2020); (12)Audibert et al. (2019); (13) García-Burillo et al. in prep. Article number, page 7 of 16 & A proofs: manuscript no. conquest_acc . . . . . . I n t e n s it y ( m J y / b ea m ) H C N H C O + H C N - v i b HO C + H C N - v i b C H NH C H OH IRAS 17208-0014 . . . . . . . I n t e n s it y ( m J y / b ea m ) H C N H C O + H C N - v i b HO C + H C N - v i b C H NH C H OH ESO 320-G030 . . . . . . I n t e n s it y ( m J y / b ea m ) H C N H C O + H C N - v i b HO C + H C N - v i b C H NH C H OH IRAS 17578-0400 . . . . . . . I n t e n s it y ( m J y / b ea m ) H C N H C O + H C N - v i b HO C + H C N - v i b C H NH C H OH ESO 173-G015 . . . . . . I n t e n s it y ( m J y / b ea m ) H C N H C O + H C N - v i b HO C + H C N - v i b C H NH C H OH IRAS F14378-3651
Fig. 2.
Continuum subtracted central spectra of galaxies with new HCN-vib J = + transitions are self-absorbed. to determine in the disturbed interacting galaxies that are com-mon in the most luminous systems. The histogram and the em-pirical distribution functions, which are estimates of the true cu-mulative distribution functions, in Fig. 3 suggest that, while highinclinations might be preferred, the distribution of inclinations isstatistically not significantly di ff erent. An Anderson-Darling test(e.g., Scholz & Stephens 1987) applied to the data confirms this,giving a probability of more than 25% that the populations aredrawn from the same underlying distribution. However, we cau-tion that the nuclear orientation may not be well connected tothat of the larger scale structure (e.g., Pjanka et al. 2017).Another interesting property in the context of CONs, whichare thought to contain warm cores with intense mid-infrared ra-diation fields, is the relative strength of the mid- and far-infraredemission as traced by the IRAS 25 to 60 µ m flux density ratio( f / f ). Visual inspection of both the histogram and the empir-ical distribution functions in Fig. 4 suggests that the CONs and non-CONs have di ff erent distributions, with most of the CONshaving lower f / f ratios than the non-CONs. An Anderson-Darling test between the f / f distributions of the CONs andthe rest of the sample confirms this notion; the probability thatthe populations are drawn from the same underlying distributionis less than 1%. This result is interesting in the light of the samplebeing biased toward objects with lower f / f ratios, see Sect.6.2.1 for further discussion.As the CONs are characterized by their high column den-sities of obscuring material, it may be of interest to investi-gate the strength of the 9 . µ m silicate feature ( s . µ m ) and theequivalent width of the 6 . µ m polycyclic aromatic hydrocar-bon (PAH) feature (EQW . µ m ), two properties used in the mid-infrared classification plot devised by Spoon et al. (2007) toseparate AGN-dominated, starburst-dominated, and deeply ob-scured nuclei. The strength of the silicate feature is defined as s . µ m = ln ( f . µ m / C . µ m ) where f . µ m is the measured flux den- Article number, page 8 of 16. Falstad et al.: CON-quest ◦ )0123456 N u m b e r o f s ou r ce s CONsnon-CONs
20 40 60 80Inclination ( ◦ )0 . . . . . . C u m u l a ti v e p r ob a b ilit y Non-CONsCONs
Fig. 3.
Distribution of inclinations in the (U)LIRG sample of CON-quest.
Top: stacked histogram of the CON-quest sample with CONsmarked using a darker shade.
Bottom: empirical distribution functionsof the CON and non-CON parts of the sample. The 1- σ confidence in-tervals are marked with a gray shade. sity at 9 . µ m and C . µ m is the continuum flux density at thesame wavelength in the absence of an absorption feature. In thediagnostic plot of Spoon et al. (2007, their Fig. 1), deeply ob-scured nuclei are characterized by strong silicate absorptions andlow PAH equivalent widths. In Figs. 5 and 6 we present the his-tograms and empirical distribution functions of these two prop-erties for the (U)LIRG samples, using data collected by Stier-walt et al. (2013). While the distributions of s . µ m seem to di ff erbetween the CONs and the rest of the sample, the distributionsof EQW . µ m are more similar. Anderson-Darling tests indicatethat the probabilities that the CON and non-CON populationsare drawn from the same underlying distributions are lower than5% and higher than 25%, respectively, for s . µ m and EQW . µ m . An important way in which the CON and the extended hostgalaxy may a ff ect each other is by in- or outflows of gas ontoor from the nucleus. The presence of such noncircular motionscould also provide clues to how the CONs may evolve withtime. Falstad et al. (2019) noted that galaxies with bright HCN-vib emission tend to show redshifted OH ground state absorp-tion lines at 119 µ m, indicating possible molecular inflows. Oneof the CONs in our survey, IRAS 17578-0400, does not haveany observations of this doublet. However, it has observations ofthe other ground state doublet at 79 µ m in the Herschel science . . . . . . f / f N u m b e r o f s ou r ce s CONsnon-CONs0 .
05 0 .
10 0 .
15 0 . f / f . . . . . . C u m u l a ti v e p r ob a b ilit y Non-CONsCONs
Fig. 4.
Distribution of IRAS 25 / µ m ratios in the (U)LIRG sam-ple of CON-quest. Top: stacked histogram of the CON-quest samplewith CONs marked using a darker shade.
Bottom: empirical distribu-tion functions of the CON and non-CON parts of the sample. The 1- σ confidence intervals are marked with a gray shade. archive (see appendix A), showing a small redshift with respectto the systemic velocity of the galaxy.Another result from Falstad et al. (2019) is that at longerwavelengths CONs instead tend to show evidence of molecu-lar outflows, often compact and collimated. The sensitivity ofour new observations allows for the detection of faint spectralfeatures, probing possible low mass outflows. For IRAS 17208-0014, this is hindered by the large line widths which cause severeblending with nearby lines of CH NH and vibrationally excitedHC N, but a possible outflow signature in the form of blueshifted( ∼ −
70 km s − ) self-absorption features can be seen in the HCNand HCO + ground-state lines in Fig. 2. In ESO 320-G030 (Fig.7)we see that the HCN emission at projected velocities between 0and ±
100 km s − and between ±
200 and ±
300 km s − from thesystemic velocity has a component that is elongated along anaxis with a position angle (PA) of ∼ ◦ east of north. This isalmost perpendicular to the major kinematic axis which has aPA of ∼ ◦ as determined from the velocity field seen outsideof the most nuclear region, consistent with the value found byPereira-Santaella et al. (2016) from observations of CO J = ◦ and 40 ◦ . Interestingly, the absorption featuresin the HCN and HCO + ground-state lines in ESO 320-G030 areredshifted with respect to the emission, possibly indicating thatthe absorbing gas is moving inward. We do not see any signa-tures of misaligned high-velocity emission in IRAS 17578-0400but there is a tentative signature at lower velocities. In Fig. 8 Article number, page 9 of 16 & A proofs: manuscript no. conquest_acc − − − − . µ m N u m b e r o f s ou r ce s CONsnon-CONs − − − − . µ m . . . . . . C u m u l a ti v e p r ob a b ilit y Non-CONsCONs
Fig. 5.
Distribution of the apparent depth of the 9 . µ m silicate absorp-tion feature in the (U)LIRG sample of CON-quest. Top: stacked his-togram of the CON-quest sample with CONs marked using a darkershade.
Bottom: empirical distribution functions of the CON and non-CON parts of the sample. The 1- σ confidence intervals are marked witha gray shade. we see that the emission within ±
80 km s − (projected) of thesystemic velocity is elongated along the kinematic minor axis(PA ∼ ◦ , again determined from the velocity field outside themost nuclear region), with a possible o ff set between the red- andblueshifted emission components along this axis. For this galaxy,we have not found any published observations of CO to com-pare the orientations of the kinematic axes in the nucleus andon larger scales. The absorption features in the HCN and HCO + ground-state lines of IRAS 17578-0400 do not show any strongvelocity shifts, although the one in the HCN line appears slightlyredshifted compared to the emission.
6. Discussion
With the currently used definition (see Sect. 2), a remarkablyhigh fraction of local LIRG and ULIRG systems, ∼
20% and ∼ ∼ × − Mpc − in the local Universe (Sanders et al.2003), we should thus expect a CON density of ∼ − Mpc − ,taking into account the lower detection rate in less luminous sys-tems. As discussed in Sect. 3.1, there is a potential bias due to . . . . . . . . . µ m ( µ m)02468 N u m b e r o f s ou r ce s CONsnon-CONs0 . . . . . . . . . µ m ( µ m)0 . . . . . . C u m u l a ti v e p r ob a b ilit y Non-CONsCONs
Fig. 6.
Distribution of the equivalent width of the 6 . µ m PAH featurein the (U)LIRG sample of CON-quest. Top: stacked histogram of theCON-quest sample with CONs marked using a darker shade.
Bottom: empirical distribution functions of the CON and non-CON parts of thesample. The 1- σ confidence intervals are marked with a gray shade. our selection criteria which misses “warm” galaxies. If CONsare predominantly found in cold sources, as suggested in Sect.6.2, the inclusion of the warm sources will result in an over-all decrease of ∼
50% in the CON detection rate. On the otherhand, it is possible that the only “warm” CON so far, NGC 4418,is not an exception but rather an example of a group of sourcesbelonging to the second peak of a possible bimodal distribution.If this is the case, the local number density of CONs may evenbe slightly higher than estimated here.Local CONs are important in order to conduct detailed stud-ies, but in the context of galaxy evolution it is also importantto determine how common they were at earlier times when thestar formation rate and rate of black hole growth in the Universewas higher (e.g., Chary & Elbaz 2001; Marconi et al. 2004). At z ∼
1, the number density of (U)LIRGs is up to two orders ofmagnitude larger than at z < . ff erent from those oftheir local counterparts (e.g., Rujopakarn et al. 2011). Our sample contains one LIRG: ESO 173-G015, and twoULIRG nuclei: Arp 220 E, and IRAS F12112 + Σ HCN − vib that Article number, page 10 of 16. Falstad et al.: CON-quest
Right Ascension (J2000)100 pc − −
80 0 80 160Velocity (km / s) 0 – 100 km / s100 pc11h53m11.8s 11.6s-39 ◦ D ec li n a ti on ( J )
100 – 200 km / s100 pc 200 – 300 km / s100 pc Fig. 7.
Intensity weighted velocity (moment 1) map and integrated in-tensity contours of HCN taken over the velocity ranges 0–100, 100–200,and 200–300 km s − from the systemic velocity ( ∼ − ) on thered- (solid) and blueshifted (dashed) sides in ESO 320-G030. Velocitycontours are given every 30 km s − . Integrated intensity contours startat 0 .
24, 0 .
16, and 0 .
11 Jy beam − km s − (5- σ ) for the three velocity in-tervals, respectively, and increase by factors of two. Right Ascension (J2000)100 pc − −
80 0 80 160Velocity (km / s) 0 – 80 km / s100 pc18h00m31.9s 31.8s-4 ◦ D ec li n a ti on ( J )
80 – 160 km / s100 pc 160 – 240 km / s100 pc Fig. 8.
Intensity weighted velocity (moment 1) map and integrated in-tensity contours of HCN taken over the velocity ranges 0–80, 80–160,and 160–240 km s − from the systemic velocity ( ∼ − ) on thered- (solid) and blueshifted (dashed) sides in IRAS 17578-0400. Ve-locity contours are given every 30 km s − . Integrated intensity contoursstart at 0 .
16, 0 .
20, and 0 .
16 Jy beam − km s − (5- σ ) for the three velocityintervals, respectively, and increase by factors of two. are up to a factor of five too low to be classified as CONs. Aquestion that should be asked is how these objects should beclassified and what their relation to the CONs is. If we com-pare the surface brightnesses to those of the CONs, we see thatthe weakest CON has a Σ HCN − vib value that is less than a factorof two larger than that of the strongest non-CON. Interestingly,when considering the L HCN − vib / L IR ratio using the infrared emis-sion from the individual nuclei, Arp 220 E has a higher ratiothan the neighboring CON Arp 220 W (Martín et al. 2016). Inaddition to the HCN-vib emission, the spectra of these galax-ies also exhibit some other features commonly seen in CONs.These include possible CH NH lines, double peaked profiles inthe HCN and HCO + ground-state lines (see Fig. 2), and, at leastin ESO 173-G015, a clearly detected HOC + line. With these sim-ilarities in mind, it is possible that these objects are related to theCONs, either as less extreme versions, as former CONs, or as ob-jects in a transition phase to become CONs. With an empiricallybased, but still somewhat arbitrary, limit separating CONs fromnon-CONs, it may also be fruitful to think about the CONs as apopulation of sources at one of the extreme ends of a continuousdistribution of source properties. The sources that fall just belowthe limit to be considered a CON would then not be qualitativelydi ff erent from those that are above the same limit. For example,one possibility is that these objects are as obscured as the regu-lar CONs, but without a central engine powerful enough to cre-ate the environment required for the HCN-vib emission to riseabove the limit in our definition. In any case, further studies ofthese, and similar, objects may be important for our understand-ing of the timescales and processes of formation and destructionof CONs. There is no statistically significant evidence that CONs are foundprimarily in highly inclined systems, which was one of the expla-nations for the lack of wide angle outflows in CONs suggestedby Falstad et al. (2019). However, it is important to rememberthat it is di ffi cult to determine the inclination of the disturbedinteracting galaxies that are common in the (U)LIRG samples.These systems may also contain two nuclei with distinct orien-tations that will not necessarily be well-resolved. Furthermore,the obscuration in CONs typically occurs on scales of ∼
100 pcor less (e.g., Aalto et al. 2019), and the orientation of this struc-ture may di ff er from that of the host galaxy (e.g., Pjanka et al.2017). This will be further investigated in a follow-up projectwith an angular resolution of ∼ . (cid:48)(cid:48) ( ∼ J = . ff erent distribution of f / f ratios compared to the restof the sample galaxies. The majority of the detected CONs, aswell as the other sources with detected HCN-vib emission, havemid to far-infrared continuum ratios smaller than 0 .
1, indicatingrelatively cool dust emission. This result may seem somewhatsurprising at first, given that e ffi cient excitation of HCN-vib re-quires intense mid-infrared radiation. However, suppression ofthe mid-infrared continuum is a natural consequence of the highcolumn densities that are found in CONs. Similar argumentswere invoked already by Bryant & Scoville (1999) to explainthe low f / f ratios in other LIRGs with high central gas sur-face densities. The low f / f ratios are also consistent with thegreenhouse scenario of González-Alfonso & Sakamoto (2019)where the mid- and far-infrared “photospheres” are significantlycooler than the interior regions. As they point out, however, their Article number, page 11 of 16 & A proofs: manuscript no. conquest_acc study assumes spherical symmetry and no clumpiness. In a morerealistic situation, for example with a clumpy medium or a disk-like structure, some amount of mid-infrared radiation leaking outthrough sightlines with lower column density is expected. An in-dication that this is occuring comes from the results of Lahuiset al. (2007), who find excitation temperatures of 200 −
300 Kin their analysis of the 14 µ m HCN absorption band, compara-ble to the central temperature found in IC 860 by Aalto et al.(2019) from millimeter observations. This leaking would reducethe greenhouse e ff ect, but the radiative transfer in disk-outflowsystems as the one suggested in IC 860 (Aalto et al. 2019) iscomplex, and requires further study. We note also that one ofthe CONs, NGC 4418, stands out with a substantially higher f / f ratio of 0 .
22. This could be due to a larger fraction ofmid-infrared radiation leaking out in this source, either due to alower column density or a more clumpy medium, or to radiationfrom surrounding structures as, for example, the super star clus-ters suggested by Varenius et al. (2014) based on radio VLBIobservations. It could also be that the di ff erence is due to di ff er-ent fractions of the total L IR coming from the CON and the hostgalaxy. For example, we see in Fig. 2b of González-Alfonso &Sakamoto (2019) that, although lower than in less obscured nu-clei, the intrinsic f / f ratio of the model with high columndensity is still close to the 0 .
22 observed in NGC 4418. It is thuspossible that the CON in NGC 4418 is dominating the infraredluminosity of the galaxy, while the CONs in other galaxies are“diluted” by (colder) emission from the host. In any case, furtherstudies are required to investigate if, and in that case how, coldand warm CONs di ff er.If we turn to the mid-infrared diagnostics used by Spoonet al. (2007), it seems that CONs generally have similarEQW . µ m but stronger s . µ m when compared to the rest of thesample. This may be an indication that the CONs are distributedalong most of the diagonal branch, going from strong s . µ m andlow EQW . µ m to weak s . µ m and high EQW . µ m , in the diag-nostic plot by Spoon et al. (2007, their Fig. 1). This notion isconfirmed if we look at the combination of the two values foreach individual CON (Fig. 9). Interestingly, the galaxies alongthis diagonal branch are described by Spoon et al. (2007) as “in-termediate stages between a fully obscured galactic nucleus andan unobscured nuclear starburst”, a statement that is seeminglyinconsistent with the fact that the highly obscured CONs arefound along most of the range of this branch. This could how-ever, again, be explained by the greenhouse scenario in whichmost of the mid-infrared emission from CONs is hidden by theouter, cooler, parts of the nucleus. We note that NGC 4418 hasthe deepest silicate absorption of the CONs, possibly favoringthe leakage explanation for its high f / f ratio as the 9 . µ mcontinuum against which the absorption takes place would bevery low in a pure greenhouse scenario (Fig. 2b of González-Alfonso & Sakamoto 2019).There are also striking di ff erences between the three lumi-nosity bins: none of the sub-LIRGs, ∼
20% of the LIRGs, and ∼
40% of the ULIRGs in our sample host CONs. This may berelated to the conditions necessary for CONs to form, and it istherefore worth examining what sets the three subsamples apart.Besides the trivial di ff erences in luminosity, they mainly dif-fer in the morphologies of the constituent galaxies. While mostsub-LIRGs are single gas-rich spirals, the fraction of interact-ing galaxies, as well as the severity of the interaction, increaseswith luminosity in the LIRG luminosity range, to dominate thepopulation of ULIRGs (Sanders & Mirabel 1996). If the CONphenomenon is primarily linked to rapid gas inflows we expectto find them primarily in actively interacting systems, and this is − − PAH EQW . µ m ( µ m) − − − − s . µ m CONsnon-CONs
Fig. 9.
Diagnostic plot of the strength of the 9 . µ m silicate featureagainst the equivalent width of the 6 . µ m PAH feature suggested bySpoon et al. (2007). The CON-quest LIRGs and ULIRGs are plottedusing data presented by Stierwalt et al. (2013), with CONs plotted asred triangles and non-CONs plotted as white circles. Shaded and whiterectangles indicate the approximate regions of the plot used by Spoonet al. (2007) to divide sources into di ff erent classes. consistent with the higher fraction seen in ULIRGs, which are allmajor mergers. It is also possible that the higher detection ratein ULIRGs is due to the larger fraction of systems with multiplenuclei, as the rate decreases to ∼
25% if we consider individualresolved nuclei. The LIRG CONs, on the other hand, are oftenfound in seemingly isolated galaxies with relatively settled mor-phologies, although some of them, for example IRAS 17578-0400 (Stierwalt et al. 2013) and NGC 4418 (Boettcher et al.2020), do have nearby companions. In the latter case, the cen-tral gas concentration may be explained by a minor interactionwith the companion (Boettcher et al. 2020), but the presence ofCONs in isolated and undisturbed galaxies raises questions onthe origin of the massive amounts of gas and dust in their nuclei,which in turn should a ff ect the growth of the SMBH. It is pos-sible, and maybe even likely, that there are multiple formationprocesses where the ULIRG CONs are formed in major mergers,while others form predominantly through minor interactions orsecular processes internal to the galaxies. This would be similarto, for example, how pseudobulges are thought to form throughsecular evolution while classical bulges are formed through ma-jor mergers (e.g., Kormendy & Kennicutt 2004). Detailed studiesof CONs, their host galaxies, and the environments of the hostswill be required to determine whether CONs are always formedas a result of interactions or if secular evolution is capable ofdepositing enough gas in the nuclei for CONs to form. The di ff erence in the distributions of the f / f ratios in theCON and non-CON parts of the sample calls for a discussion ofhow the selection bias presented in Sect. 3.1 a ff ects our conclu-sions. The main e ff ect of the bias is to prevent the real fractionof warmer sources that host CONs to be determined. Depend-ing on the value of this fraction there are two main scenarios:either NGC 4418 is an outlier, and in general warm CONs arerare; or there is a second population of warm CONs. In order todiscern between the two scenarios, a sample selected on the totalinfrared luminosity could be observed. This would include an ad-ditional six ULIRGs, 24 LIRGs, and four sub-LIRGs comparedto the current CON-quest sample. In the first scenario mentioned Article number, page 12 of 16. Falstad et al.: CON-quest above, no additional CONs would be found in this extended sam-ple and the fraction of (U)LIRGs that host CONs would have tobe revised down by ∼
50% as the ULIRG and LIRG samples ap-proximately double. In the second scenario, the detection rate ofnew CONs would depend on the actual fraction of warm galaxiesthat host CONs.
The four previously known CONs in this survey were all in-cluded in a study of OH outflows in obscured galaxies (Falstadet al. 2019). Contrary to most other (U)LIRGs in that study, theCON hosts all had positive median velocities, indicating inflow-ing motion, in the OH 119 µ m absorption lines. One other CON,IRAS F22491-1808, was included in the same study, but wasjust below the limit to be considered a CON with the defini-tion based on the L HCN − vib / L IR ratio. Its OH 119 µ m lines areredshifted as in other CONs. One of the other newly identifiedCONs in this survey, ESO 320-G030, has observations of the OH119 µ m doublet reported by González-Alfonso et al. (2017) un-der the name IRAS 11506-3851. Similarly to those in the alreadyknown CONs, the 119 µ m absorption lines of OH in ESO 320-G030 show evidence of inflowing gas in the form of a significantredshift relative to the systemic velocity of the galaxy. The othernewly identified CON, IRAS 17578-0400, had an observation ofthe other ground state doublet at 79 µ m, with the absorption linespeaking at a slightly redshifted velocity (see Appendix A). In amultitransition OH analysis of molecular outflows by González-Alfonso et al. (2017) it is seen (their Fig. 10) that all sources inwhich the 79 µ m doublet peaks at velocities (cid:38) − also haveOH 119 µ m lines which peak at positive velocities.At (sub)millimeter wavelengths, all seven CONs in our sur-vey show possible signatures of molecular outflows. At leastthree of these, those in Arp 220 W (Barcos-Muñoz et al. 2018),ESO 320-G030 (Pereira-Santaella et al. 2016), and Zw 049.057(Falstad et al. 2018), appear collimated. This may be true also forthe outflow in IRAS 17208-0014 (García-Burillo et al. 2015), al-though its morphology still requires further study. IRAS F22491-1808 has outflow signatures on kpc scales in the J = J tran-sitions of CO has been reported by Fluetsch et al. (2019) andLutz et al. (2020) but its geometry is not known. The final CON,IRAS 17578-0400, does not have any published outflow signa-tures, but our HCN observations reveal a low-velocity elongationalong its kinematic minor axis (Fig. 8). This may be a signatureof a minor axis outflow which is directed almost perpendicular toour line of sight toward the galaxy. Possible signatures of noncir-cular motions are also seen in the ground state HCN and HCO + lines in the form of asymmetric line profiles in some CONs. Inmost of them, the absorptions are stronger on the blueshiftedside: Zw 049.057 (Aalto et al. 2015b), IRAS 17208-0014 (Fig.2), and Arp 220 (Sakamoto et al. 2009; Aalto et al. 2015b; Martínet al. 2016) or close to the line center: IRAS 17578-0400 (Fig. 2),and NGC 4418 (Sakamoto et al. 2013). In ESO 320-G030 (Fig.2), however, it is clearly stronger on the redshifted side, possi-bly indicating that the absorbing gas is moving toward the nu-cleus, consistent with what is seen in CO J = + ◦ ) but has been studied at high spatial resolution by Aalto et al. (2019).In this galaxy, the J = J = ff ect mayalso be present in other CONs, for example ESO 320-G030 andIRAS 17578-0400 where the nuclear HCN emission is more ex-tended along the kinematic minor axis than along the kinematicmajor axis (see Figs. 7 and 8). Determining the HCN-vib surface brightness requires obser-vations with both high sensitivity and spatial resolution. It isnot always possible to achieve both of these, so alternativeways to find CONs are sometimes desirable. With high enoughspatial resolution, one possible way to find sources with veryhigh column density is to examine the continuum brightness at(sub)millimeter wavelengths, where the optical depth of the dustapproaches unity for column densities of N H2 ∼ cm − . Thismethod has been used in, for example, Arp 220 (Sakamoto et al.2008), NGC 4418 (Sakamoto et al. 2013), and IC 860 (Aaltoet al. 2019). In Fig. 10 we plot the continuum surface brightnessat 1 . ffi cient of > .
99. We note that the surface brightnesses onboth axes are calculated using the same area, that of the 1 . ff ectively meaning that brighter 1 . ff ectedby synchrotron or free-free emission (e.g., Aalto et al. 2019),a problem that gets worse at longer wavelengths. This methodmay thus render some false positives when searching for CONs.For example, NGC 1068 has a continuum surface brightness at1 . ∼ − , similar to many CONs, but aHCN-vib surface brightness of less than 0 . L (cid:12) pc − .In some cases, for example at higher redshift ( z (cid:38) . L HCN − vib / L IR > − selects mostlythe same sources as the currently used definition and that it there-fore may be useful in such situations. In Fig. 11 we plot the L HCN − vib / L IR ratio as a function of the HCN-vib surface bright-ness. The correlation between these quantities is weaker with aPearson correlation coe ffi cient of ∼ .
3. This can be considereda moderate correlation and the criterion based on L HCN − vib / L IR ratio may be used when the surface brightness is hard to derive. Article number, page 13 of 16 & A proofs: manuscript no. conquest_acc − − − Σ HCN − vib (L (cid:12) / pc ) S mm / ∆ Ω ( m J y / a r c s ec ) Fig. 10.
Continuum surface brightness at 1 . − − − Σ HCN − vib (L (cid:12) / pc )10 − − L H C N − v i b / L I R ( − ) Fig. 11.
Luminosity of HCN-vib relative to the total (between 8 and1000 µ m) infrared luminosity as a function of the HCN-vib surfacebrightness for the CON-quest sample.
7. Conclusions
We present the first results of a systematic survey of infraredluminous galaxies, called CON-quest, with the aim to searchfor compact obscured nuclei, traced by strong HCN-vib emis-sion. Our sample consists of literature, archival, and new ALMAdata toward 46 far-infrared selected ULIRGs (10 L (cid:12) ≤ L FIR ),LIRGs (10 L (cid:12) ≤ L FIR < L (cid:12) ), and sub-LIRGs (10 L (cid:12) ≤ L FIR < L (cid:12) ).Using our definition of a CON as a galaxy where Σ HCN − vib > (cid:12) pc − over nuclear regions with radii of ∼ −
100 pc, we findthat ∼
40% of the ULIRGs, ∼
20% of the LIRGs, and <
9% ofthe sub-LIRGs, host CONs. It has been suggested that the CONphenomenon is related to inclination, but there is no evidencethat the CON hosts have a di ff erent distribution of inclinationsthan the rest of the sample. However, the disk of the host galaxymay not be aligned with that of the nucleus, and high-resolutionobservations should be conducted to determine the inclinationsof the nuclear disks in CONs. One property that does di ff er be-tween CONs and other galaxies is the IRAS f / f ratio, withCONs generally having lower ratios indicating colder dust spec-tral energy distributions (SEDs). This is consistent with the no-tion that the large dust columns in CONs work to gradually shiftthe radiation to longer wavelengths, making the mid- and far-infrared “photospheres” significantly cooler than the interior re-gions. There is however one outlier, NGC 4418, with a signifi-cantly higher f / f ratio which leaves open the possibility of a broad or bimodal f / f distribution among CONs. The obscu-ration of the warm interiors by cooler foreground gas may alsoexplain why the mid-infrared diagnostics EQW . µ m and s . µ m indicate that some of the deeply obscured CONs are insteadrelatively unobscured starbursts. So far, all CONs have possi-ble signatures of inflowing molecular gas in the far-infrared. At(sub)millimeter wavelengths, however, all CONs show possiblesignatures of compact and collimated molecular outflows. Fur-ther high-resolution observations should be conducted to inves-tigate if and how these properties are connected to each other andto the CON phenomenon. Detailed studies of individual sources,including for example multifrequency continuum observations(e.g., Aalto et al. 2019) or excitation analysis of HCN-vib atfrequencies less a ff ected by dust obscuration (e.g., Salter et al.2008), would also make it possible to better constrain physicalproperties such as column densities, sizes, and temperatures ofthe nuclei.The fact that our sample is based on far-infrared luminosi-ties means that it is biased toward sources with cooler SEDs.The bias mainly a ff ects the LIRG and ULIRG parts of the sam-ple and can be remedied by obtaining ALMA measurements ofthe HCN-emission in a complementary sample of sources se-lected based on their total infrared luminosity. This would ap-proximately double the number of LIRGs and ULIRGs in oursample and increase the sub-LIRG sample by ∼ ∼ ff er. Finally, to make an assessment of the importance of CONsin the context of galaxy evolution, it is necessary to conduct stud-ies at higher redshifts where (U)LIRGs are more common. Acknowledgements.
This paper makes use of the following ALMA data:ADS / JAO.ALMA / JAO.ALMA / JAO.ALMA / JAO.ALMA / JAO.ALMA / JAO.ALMA / JAO.ALMA / NRAO and NAOJ. We thank the anonymous referee for usefulcomments and suggestions. We acknowledge support from the Nordic ALMARegional Centre (ARC) node based at Onsala Space Observatory. The NordicARC node is funded through Swedish Research Council grant No 2017-00648.S.A. gratefully acknowledges support from an ERC AdvancedGrant 789410 andfrom the Swedish Research Council. K.S. S.G-B. acknowledges support fromthe research projects PGC2018-094671-B-I00 (MCIU / AEI / FEDER, UE) andPID2019-106027GA-C44 from the Spanish Ministerio de Ciencia e InnovaciónT.D-S. acknowledges support from the CASSACA and CONICYT fund CAS-CONICYT Call 2018. G.A.F acknowledges financial support from the StateAgency for Research of the Spanish MCIU through the AYA2017-84390-C2-1-R grant (co-funded by FEDER) and through the “Center of Excellence SeveroOchoa” award for the Instituto de Astrofísica de Andalucia (SEV-2017-0709).T.R.G. acknowledges the Cosmic Dawn Center of Excellence funded bythe Danish National Research Foundation under grant No. 140. E.G-A. is aResearch Associate at the Harvard-Smithsonian Center for Astrophysics, andthanks the Spanish Ministerio de Economía y Competitividad for support underprojects ESP2017-86582-C4-1-R and PID2019-105552RB-C41. This researchhas made use of NASA’s Astrophysics Data System. This research has made useof the NASA / IPAC Extragalactic Database (NED) which is operated by the JetPropulsion Laboratory, California Institute of Technology, under contract withthe National Aeronautics and Space Administration.
References
Aalto, S., Garcia-Burillo, S., Muller, S., et al. 2015a, A&A, 574, A85Aalto, S., Martín, S., Costagliola, F., et al. 2015b, A&A, 584, A42Aalto, S., Muller, S., König, S., et al. 2019, A&A, 627, A147
Article number, page 14 of 16. Falstad et al.: CON-quest
Audibert, A., Combes, F., García-Burillo, S., et al. 2019, A&A, 632, A33Audibert, A., Combes, F., García-Burillo, S., et al. 2020, arXiv e-prints,arXiv:2011.09133Barcos-Muñoz, L., Aalto, S., Thompson, T. A., et al. 2018, ApJ, 853, L28Barcos-Muñoz, L., Leroy, A. K., Evans, A. S., et al. 2015, ApJ, 799, 10Barro, G., Faber, S. M., Pérez-González, P. G., et al. 2013, ApJ, 765, 104Barro, G., Faber, S. M., Pérez-González, P. G., et al. 2014, ApJ, 791, 52Blecha, L., Snyder, G. F., Satyapal, S., & Ellison, S. L. 2018, MNRAS, 478,3056Boettcher, E., Gallagher, John S., I., Ohyama, Y., et al. 2020, A&A, 637, A17Bryant, P. M. & Scoville, N. Z. 1999, AJ, 117, 2632Cameron, E. 2011, PASA, 28, 128Chary, R. & Elbaz, D. 2001, ApJ, 556, 562Combes, F., García-Burillo, S., Audibert, A., et al. 2019, A&A, 623, A79Condon, J. J. 1997, PASP, 109, 166Falstad, N., Aalto, S., Mangum, J. G., et al. 2018, A&A, 609, A75Falstad, N., González-Alfonso, E., Aalto, S., et al. 2015, A&A, 580, A52Falstad, N., Hallqvist, F., Aalto, S., et al. 2019, A&A, 623, A29Fisher, K. B., Huchra, J. P., Strauss, M. A., et al. 1995, ApJS, 100, 69Fluetsch, A., Maiolino, R., Carniani, S., et al. 2019, MNRAS, 483, 4586García-Burillo, S., Combes, F., Usero, A., et al. 2015, A&A, 580, A35Ginsburg, A. & Mirocha, J. 2011, PySpecKit: Python Spectroscopic Toolkit, As-trophysics Source Code LibraryGonzález-Alfonso, E., Fischer, J., Graciá-Carpio, J., et al. 2014, A&A, 561, A27González-Alfonso, E., Fischer, J., Graciá-Carpio, J., et al. 2012, A&A, 541, A4González-Alfonso, E., Fischer, J., Spoon, H. W. W., et al. 2017, ApJ, 836, 11González-Alfonso, E., Pereira-Santaella, M., Fischer, J., et al. 2021, A&A, 645,A49González-Alfonso, E. & Sakamoto, K. 2019, ApJ, 882, 153Harada, N., Sakamoto, K., Martín, S., et al. 2018, ApJ, 855, 49Imanishi, M. & Nakanishi, K. 2013, AJ, 146, 91Imanishi, M., Nakanishi, K., & Izumi, T. 2016a, ApJ, 822, L10Imanishi, M., Nakanishi, K., & Izumi, T. 2016b, AJ, 152, 218Imanishi, M., Nakanishi, K., & Izumi, T. 2016c, ApJ, 825, 44Imanishi, M., Nakanishi, K., & Izumi, T. 2018, ApJ, 856, 143Imanishi, M., Nakanishi, K., & Izumi, T. 2019, ApJS, 241, 19Imanishi, M., Nguyen, D. D., Wada, K., et al. 2020, ApJ, 902, 99Israel, F. P. 2020, A&A, 635, A131Kocevski, D. D., Barro, G., Faber, S. M., et al. 2017, ApJ, 846, 112Kocevski, D. D., Brightman, M., Nandra, K., et al. 2015, ApJ, 814, 104Kormendy, J. & Ho, L. C. 2013, ARA&A, 51, 511Kormendy, J. & Kennicutt, Robert C., J. 2004, ARA&A, 42, 603Krieger, N., Bolatto, A. D., Leroy, A. K., et al. 2020, ApJ, 897, 176Lahuis, F., Spoon, H. W. W., Tielens, A. G. G. M., et al. 2007, ApJ, 659, 296Le Floc’h, E., Papovich, C., Dole, H., et al. 2005, ApJ, 632, 169Lutz, D., Sturm, E., Janssen, A., et al. 2020, A&A, 633, A134Magnelli, B., Elbaz, D., Chary, R. R., et al. 2009, A&A, 496, 57Magnelli, B., Popesso, P., Berta, S., et al. 2013, A&A, 553, A132Magorrian, J., Tremaine, S., Richstone, D., et al. 1998, AJ, 115, 2285Makarov, D., Prugniel, P., Terekhova, N., Courtois, H., & Vauglin, I. 2014, A&A,570, A13Marconi, A., Risaliti, G., Gilli, R., et al. 2004, MNRAS, 351, 169Martín, S., Aalto, S., Sakamoto, K., et al. 2016, A&A, 590, A25McMullin, J. P., Waters, B., Schiebel, D., Young, W., & Golap, K. 2007, in As-tronomical Society of the Pacific Conference Series, Vol. 376, AstronomicalData Analysis Software and Systems XVI, ed. R. A. Shaw, F. Hill, & D. J.Bell, 127Ott, S. 2010, in Astronomical Society of the Pacific Conference Series, Vol. 434,Astronomical Data Analysis Software and Systems XIX, ed. Y. Mizumoto,K.-I. Morita, & M. Ohishi, 139Pérault, M. 1987, PhD thesis, PhD dissertation, Université Paris VII, (1987)Pereira-Santaella, M., Colina, L., García-Burillo, S., et al. 2016, A&A, 594, A81Pereira-Santaella, M., Colina, L., García-Burillo, S., et al. 2018, A&A, 616,A171Pilbratt, G. L., Riedinger, J. R., Passvogel, T., et al. 2010, A&A, 518, L1Pjanka, P., Greene, J. E., Seth, A. C., et al. 2017, ApJ, 844, 165Poglitsch, A., Waelkens, C., Geis, N., et al. 2010, A&A, 518, L2Ricci, C., Bauer, F. E., Treister, E., et al. 2017, MNRAS, 468, 1273Rujopakarn, W., Rieke, G. H., Eisenstein, D. J., & Juneau, S. 2011, ApJ, 726, 93Sakamoto, K., Aalto, S., Barcos-Muñoz, L., et al. 2017, ApJ, 849, 14Sakamoto, K., Aalto, S., Costagliola, F., et al. 2013, ApJ, 764, 42Sakamoto, K., Aalto, S., Evans, A. S., Wiedner, M. C., & Wilner, D. J. 2010,ApJ, 725, L228Sakamoto, K., Aalto, S., Wilner, D. J., et al. 2009, ApJ, 700, L104Sakamoto, K., Wang, J., Wiedner, M. C., et al. 2008, ApJ, 684, 957Salter, C. J., Ghosh, T., Catinella, B., et al. 2008, AJ, 136, 389Sanders, D. B., Mazzarella, J. M., Kim, D.-C., Surace, J. A., & Soifer, B. T.2003, AJ, 126, 1607Sanders, D. B. & Mirabel, I. F. 1996, ARA&A, 34, 749 Scholz, F. W. & Stephens, M. A. 1987, Journal of the American Statistical As-sociation, 82, 918Scoville, N., Murchikova, L., Walter, F., et al. 2017, ApJ, 836, 66Solomon, P. M. & Vanden Bout, P. A. 2005, ARA&A, 43, 677Spoon, H. W. W., Marshall, J. A., Houck, J. R., et al. 2007, ApJ, 654, L49Stierwalt, S., Armus, L., Surace, J. A., et al. 2013, ApJS, 206, 1Strauss, M. A., Huchra, J. P., Davis, M., et al. 1992, ApJS, 83, 29Tacchella, S., Carollo, C. M., Förster Schreiber, N. M., et al. 2018, ApJ, 859, 56Varenius, E., Conway, J. E., Martí-Vidal, I., et al. 2014, A&A, 566, A15Veilleux, S., Meléndez, M., Sturm, E., et al. 2013, ApJ, 776, 27Ziurys, L. M. & Turner, B. E. 1986, ApJ, 300, L19Zolotov, A., Dekel, A., Mandelker, N., et al. 2015, MNRAS, 450, 2327 Department of Space, Earth and Environment, Chalmers Universityof Technology, Onsala Space Observatory, 439 92 Onsala, Swedene-mail: [email protected] Observatoire de Paris, LERMA, College de France, CNRS, PSLUniv., Sorbonne University, UPMC, Paris, France Universidad de Alcalá, Departamento de Física y Matemáticas,Campus Universitario, E-28871 Alcalá de Henares, Madrid, Spain National Radio Astronomy Observatory, 520 Edgemont Road, Char-lottesville, VA 22903, USA Department of Astronomy, 530 McCormick Road, University of Vir-ginia, Charlottesville, VA 22904, USA Núcleo de Astronomía de la Facultad de Ingeniería, UniversidadDiego Portales, Av. Ejército Libertador 441, Santiago, Chile European Southern Observatory, Alonso de Córdova 3107, Vitacura763 0355, Santiago, Chile Joint ALMA Observatory, Alonso de Córdova 3107, Vitacura 7630355, Santiago, Chile Institute of Astronomy and Astrophysics, Academia Sinica, 11Fof AS / NTU Astronomy-Mathematics Building, No.1, Sec. 4, Roo-sevelt Rd, Taipei 10617, Taiwan, R.O.C. Jodrell Bank Centre for Astrophysics, Department of Physics & As-tronomy, School of Natural Sciences, The University of Manchester,M13 9PL, UK Department of Astronomy, University of Wisconsin-Madison, 5534Sterling, 475 North Charter Street, Madison WI 53706, USA Leiden Observatory, Leiden University, P.O. Box 9513, NL-2300RA Leiden, The Netherlands Department of Physics and Astronomy, University College London,Gower Street, London WC1E 6BT, UK Cosmic Dawn Center (DAWN), DTU-Space, Technical Universityof Denmark, Elektrovej 327, DK-2800 Kgs. Lyngby, Denmark Observatorio de Madrid, OAN-IGN, Alfonso XII, 3, E-28014-Madrid, Spain Max-Planck-Institut für Radioastronomie, Auf dem Hügel 69,53121, Bonn, Germany Astron. Dept., King Abdulaziz University, P.O. Box 80203, 21589Jeddah, Saudi Arabia National Astronomical Observatory of Japan, National Institutes ofNatural Sciences (NINS), 2-21-1 Osawa, Mitaka, Tokyo 181–8588,Japan Institute of Astronomy, The University of Tokyo, 2-21-1, Osawa,Mitaka, Tokyo 181-0015, Japan Chile Observatory, National Astronomical Observatory of Japan, 2-21-1, Osawa, Mitaka, Tokyo 181-8588, Japan Kavli Institute for Astronomy and Astrophysics, Peking University,Beijing 100871, China George Mason University, Department of Physics & Astronomy, MS3F3, 4400 University Drive, Fairfax, VA 22030, USA Argelander Institut für Astronomie, Universität Bonn, Auf demHügel 71, 53121 Bonn, Germany Chinese Academy of Sciences South America Center for Astronomy(CASSACA), National Astronomical Observatories, CAS, Beijing100101, China Institute of Astrophysics, Foundation for Research and Technology–Hellas (FORTH), Heraklion, GR-70013, Greece Instituto de Astrofísica de Andalucia (CSIC), Glorieta de al As-tronomia s / n E-18008, Granada, SpainArticle number, page 15 of 16 & A proofs: manuscript no. conquest_acc
Appendix A: Herschel OH detection in IRAS17578-0400
One of the newly identified CONs, IRAS 17578-0400, also hadobservations of the OH ground state doublet at 79 µ m takenwith the Photodetector Array Camera and Spectrometer (PACS;Poglitsch et al. 2010) on the Herschel
Space Observatory (Pil-bratt et al. 2010). The observations (obsid: 1342239716, PI:L. Armus) were conducted on 2012 February 25 with the highspectral sampling, range spectroscopy mode for a duration of1176 s and the data were processed with version 14.2 of the stan-dard pipeline. The nuclear far-infrared emission of IRAS 17578-0400 is spatially unresolved in the central 9 . (cid:48)(cid:48) ( ∼ Herschel interactive processing environment (HIPE; Ott2010) version 14.0.1. Before analyzing the absorption lines, apolynomial of order two was fitted to the continuum and thensubtracted from the spectrum. To ease comparison with the OH119 µ m doublets used in previous studies, the profiles of the OH79 µ m doublets were modeled using the same procedure as inVeilleux et al. (2013). Each line was fitted with two Gaussiancomponents characterized by their amplitude, velocity centroid,and width. The separation between the two lines of the dou-blet was fixed at 0 . µ m in the rest frame and the amplitudeand width of the two lines were the same for each component.The median velocity of the absorptions was then found to be33 km s − . The fitting procedure was carried out using the spec-troscopic analysis toolkit PySpecKit (Ginsburg & Mirocha 2011)and the continuum subtracted spectrum with the fits overplottedare presented in Fig. A.1. − −
500 0 500 1000Velocity (km s − )-4.0-2.00.0 F l uxd e n s it y ( J y ) IRAS 17578
Fig. A.1.
Spectral fits to the OH 79 µ m absorption lines in IRAS 17578-0400. The solid black histogram represents the data, the solid magentaline is the best multicomponent fit to the data, and the dashed blue linesare the individual components. The velocity scale is set relative to thefrequency of the blue component of the doublet. Dashed vertical linesindicate the expected positions of the two absorption components giventhe adopted redshift of 0 ..