SMA/PdBI multiple line observations of the nearby Seyfert2 galaxy NGC 1068: Shock related gas kinematics and heating in the central 100pc?
M. Krips, S. Martin, A. Eckart, R. Neri, S. Garcia-Burillo, S. Matsushita, A. Peck, I. Stoklasova, G. Petitpas, A. Usero, F. Combes, E. Schinnerer, L. Humphreys, A.J. Baker
aa r X i v : . [ a s t r o - ph . C O ] M a y Draft version November 13, 2018
Preprint typeset using L A TEX style emulateapj v. 03/07/07
SMA/PDBI MULTIPLE LINE OBSERVATIONS OF THE NEARBY SEYFERT 2 GALAXY NGC 1068:SHOCK RELATED GAS KINEMATICS AND HEATING IN THE CENTRAL 100 PC? ⋆ M. Krips , S. Mart´ın , A. Eckart , R. Neri , S. Garc´ıa-Burillo , S. Matsushita , A. Peck , I. Stoklasov´a , G.Petitpas , A. Usero , F. Combes , E. Schinnerer , L. Humphreys , A.J. Baker Draft version November 13, 2018
ABSTRACTWe present high angular resolution (0 . ′′ − . ′′
0) observations of the mm continuum andthe CO(J=3–2), CO(J=3–2), CO(J=2–1), C O(J=2–1), HCN(J=3–2), HCO + (J=4–3) andHCO + (J=3–2) line emission in the circumnuclear disk (r .
100 pc) of the proto-typical Seyfert type-2 galaxy NGC 1068, carried out with the Submillimeter Array. We further include in our analysisnew CO(J=1–0) and improved CO(J=2–1) observations of NGC 1068 at high angular resolution(1 . ′′ − . ′′
0) and sensitivity, conducted with the IRAM Plateau de Bure Interferometer. Based on thecomplex dynamics of the molecular gas emission indicating non-circular motions in the central ∼
100 pc,we propose a scenario in which part of the molecular gas in the circumnuclear disk of NGC 1068 isradially blown outwards as a result of shocks. This shock scenario is further supported by quite warm(T kin ≥
200 K) and dense (n(H ) ≃ cm − ) gas constrained from the observed molecular line ratios.The HCN abundance in the circumnuclear disk is found to be [HCN]/[ CO] ≈ − . . This is slightlyhigher than the abundances derived for galactic and extragalactic starforming/starbursting regions.This results lends further support to X-ray enhanced HCN formation in the circumnuclear disk ofNGC 1068, as suggested by earlier studies. The HCO + abundance ([HCO + ]/[ CO] ≈ − ) appearsto be somewhat lower than that of galactic and extragalactic starforming/starbursting regions. Whentrying to fit the cm to mm continuum emission by different thermal and non-thermal processes, itappears that electron-scattered synchrotron emission yields the best results while thermal free-freeemission seems to over-predict the mm continuum emission. Subject headings:
Galaxies: individual: NGC 1068 – Galaxies: ISM – Galaxies: active – Galaxies:kinematics and dynamics – Galaxies: nuclei – Galaxies: Seyfert – Radio continuum:galaxies – Radio lines: galaxies – Submillimeter: galaxies INTRODUCTION
Little is known about the effects of active processesin galaxies on the chemical and kinematic properties of ⋆ Based on observations carried out with the IRAM Plateaude Bure Interferometer. IRAM is supported by INSU/CNRS(France), MPG (Germany) and IGN (Spain). Institut de Radio Astronomie Millim´etrique, Saint Martind’H`eres, F-38406, France; email: [krips,neri]@iram.fr European Southern Observatory, Alonso de C´ordova 3107, Vi-tacura, Casilla 19001, Santiago 19 Chile; e-mail: [email protected] Universit¨at zu K¨oln, I.Physikalisches Institut, Z¨ulpicher Str.77, 50937 K¨oln, Germany; email: [email protected] Observatorio Astron´omico Nacional (OAN) - Observatoriode Madrid, C/ Alfonso XII 3, 28014 Madrid, Spain; email:[s.gburillo,a.usero]@oan.es Institute of Astronomy and Astrophysics, Academia Sinica,PO Box 23-141, Taipei 10617, Taiwan, R.O.C.; email:[email protected] Joint ALMA Observatory, Alonso de C´ordova 3107, Santiago,Chile; email: [email protected] NRAO, 520 Edgemont Rd, Charlottesville, VA 22903 Astronomical Institute of the Academy of Sciences of the CzechRepublic, v.v.i., Boˇcn´ı II 1401, 14131 Prague, Czech Republic Harvard-Smithsonian Center for Astrophysics, SMA project,60 Garden Street MS 78, Cambridge, MA 02138; email:[email protected] Observatoire de Paris, LERMA, 61 Av. de l’Observatoire,75014 Paris, France; email: [email protected] Max-Planck-Institut f¨ur Astronomie, K¨onigstuhl 17, 69117Heidelberg, Germany; email: [email protected] ESO, Karl-Schwarzschild-Str. 2, D-85748 Garching, Ger-many; email: [email protected] Department of Physics and Astronomy, Rutgers, the StateUniversity of New Jersey, 136 Frelinghuysen Road, Piscataway, NJ08854-8019, USA; email: [email protected] the surrounding molecular gas and vice versa, whetherthe activity is in form of an active galactic nucleus(AGN) or a starburst (SB) or both. Information on thecharacteristics of the molecular gas in the vicinity of theactivity is essential to reveal the underlying physicalprocesses because molecular gas constitutes a largefraction of the fuel for the central activity and thus helpsto keep it alive over cosmologically relevant time scales.Also, the feedback of activity onto the surroundingmolecular gas represents an important factor for theevolution of the activity (with respect to outflows orshocks for instance). The large diversity and oftentimesalso simultaneity of the physical processes accompanyingthe different activity types certainly complicate anyinterpretation of the interaction between the activityand the molecular gas. These processes include largescale shocks, gas out- and inflow, other dynamical per-turbations, and strong radiation fields, such as throughUV- or X-ray radiation, cosmic rays, or supernovaeexplosions (e.g., Mart´ın et al. 2011; Garc´ıa-Burillo et al.2010; Sakamoto et al. 2010; Papadopoulos 2010;P´erez-Beaupuits et al. 2009; Krips et al. 2008; Aalto2008; Matsushita et al. 2007; Garc´ıa-Burillo et al. 2007;Mart´ın et al. 2006; Usero et al. 2006; Sakamoto et al.2006; Matsushita et al. 2005; Fuente et al. 2005;Meier & Turner 2005; Usero et al. 2004). Thus, a thor-ough study of the kinematics, excitation conditions and chemistry of the molecular gas close to AGN and SBs isessential for understanding the nature and evolution ofthese active environments. Krips et al.High angular resolution observations of the CO emis-sion, a reliable tracer of the global molecular gas reser-voir, are an important step to study the dynamics in ac-tive galaxies. CO alone, however, can certainly not de-scribe the complexity of the molecular gas close to the ac-tivity processes (especially with respect to the chemistryand excitation conditions of the molecular gas). More-over, CO has been found to be an unreliable tracer ofdense molecular gas environments (at least its lower ro-tational transitions), in which most AGN or star forma-tion activity is supposed to take place (e.g., Krips et al.2007; Gao & Solomon 2004). Observations at high angu-lar resolution of different molecular tracers is thus a nextlogical step. Given recent upgrades of current (sub)mminterferometers, the detection and spatial resolution ofweaker molecular lines becomes a feasible task.In this paper we present high angular resolution ob-servations of the (sub-)mm-continuum and CO, CO,C O, HCN and HCO + line emission in the nearbySeyfert type-2 galaxy NGC 1068, conducted with theSubmillimeter Array (SMA) and the IRAM Plateau deBure Interferometer (PdBI). NGC 1068
The nearby Seyfert type-2 galaxy NGC 1068 (seeTable 1 for a general overview) has become not onlythe figurehead for the viewing angle unification the-ory for Seyfert galaxies (e.g. Krolik & Kallman 1987),as its centre most impressively exhibits the character-istics for harboring an obscured type-1 Seyfert AGN(Antonucci & Miller 1985). It also advanced to a figure-head for the significantly different effects that AGN canhave on the excitation conditions and chemistry of thesurrounding molecular gas when compared to SB or qui-escent galaxies (e.g. Krips et al. 2008; Usero et al. 2004;Kohno et al. 2001; Sternberg et al. 1994). The prototyp-ical nature of NGC 1068 is certainly in part due to itsrelatively small distance (12.6 Mpc) from us and strongcontinuum as well as line emission from X-ray to radiofrequencies, making it hence an ideal target to study theaccretion and feedback processes of its active nucleus atunprecedented detail. As a (fortunate) consequence, awealth of information is already available for this source,the most revelant of which will be summarised in thissection.The molecular gas in NGC 1068 is distributed ina starburst ring/spiral of ∼ ∼ ′′ ) in diame-ter, a stellar bar of ∼ ∼ ′′ ) in length and acirumnuclear disk/ring (CND) of ∼
200 pc ( ∼ ′′ ) indiameter (e.g., Schinnerer et al. 2000, and referencestherein). At the very center of the CND, pronouncedH CO(2–1) line emission within theCND were interpreted as a consequence of a warpeddisk (Schinnerer et al. 2000). However, more recentobservations of the MIR rovibrational H emissionstart to raise doubts about this interpretation (e.g.,M¨uller-S`anchez et al. 2009) and alternatively suggestthat the complex gas kinematics are due to a funnelingof the gas toward the AGN along the jet (inner 60 pc)plus an expanding ring (on scales of r=100-150 pc), thelatter having been already proposed a few years beforeby Galliano & Alloin (2002).Additional fascinating characteristics of the CND ofNGC 1068, besides the complex kinematic behaviourof its molecular gas, are its chemistry and excita-tion conditions which appear to significantly differ fromstarburst/star-forming environments. Traced by the“abnormal” line ratios of different molecules and tran-sitions, mostly by HCN, HCO + and CO, it has beensuggested that the CND in NGC 1068 harbors a giantX-ray-dominated-region (XDR; e.g., Kohno et al. 2008;Usero et al. 2004; Tacconi et al. 1994; Sternberg et al.1994). XDRs are defined in a similar way to the Photon-Dominated-Regions (PDRs) in starburst galaxies (suchas M82; e.g., Fuente et al. 2005) but are driven by X-ray rather than UV-radiation. High HCN-to-CO(J=1–0)( ≥
1) and HCN-to-HCO + (J=1–0) ( ≥
1) ratios are foundin the CND of NGC 1068, indicating enhanced HCNabundances there. The X-ray radiation of the AGN isthereby supposed to be the main driver for the enhance-ment of HCN. It can penetrate much deeper into the sur-rounding molecular gas than the UV-radiation in PDRsleading to a stimulated “hyper”-production of HCN.A multi-transition, multi-molecular line study of HCNand HCO + conducted with the IRAM 30m-telescope(Krips et al. 2008) supports an increased abundance ofHCN and/or increased kinetic temperatures. Both canequally explain the elevated HCN-to- CO(1–0) line ra-tios either by the aforementioned enhancement of theHCN abundance and/or a hypo-excitation of the low-J CO transitions (see also M51 and NGC 6951 as exam-ples of increased HCN/ CO(J=1–0) ratios; Krips et al.2009, 2007; Matsushita et al. 2007, 2004; Kohno et al.1996).Recent SiO interferometric observations of the CNDin NGC 1068 carried out by Garc´ıa-Burillo et al. (2010)testify further to the complexity of the gas chemistry inthis galaxy. The bright SiO emission in its CND suggestsan enhanced abundance of this molecule which is inter-preted by the authors as closely related to (high-velocity)shocks. The shocks are believed to be a consequence ofa jet-gas interaction. OBSERVATIONS illimeter observations of NGC 1068 3A summary of all (sub-)millimeter interferometric ob-servations is given in Table 2, where observing parame-ters as well as achieved rms noise and angular resolutionsare listed.
SMA
For all SMA observations, unless otherwisestated, the phase reference centre has been set to α J2000 =02h42m40.70s and δ J2000 = − ◦ ′ . ′′ LSR =1137 km s − . Both the uppersideband (USB) and lower sideband (LSB) were usedfor the observations, yielding a bandwidth of 2 GHzeach separated by 10 GHz. A spectral resolution of0.81 MHz was used for all observations, correspondingto 0.8 km s − at 1 mm. The 225 GHz zenith opacity τ was measured regularly throughout all observations atthe nearby Caltech Submillimeter Observatory (CSO).The accuracy of the flux calibration for all tracks isestimated to be at a conservative level of ∼ HCN(J=3–2) emission
We observed the HCN(J=3–2) line emission inNGC 1068 using the extended and very extended con-figurations with up to eight 6 m dishes in January, Oc-tober and November 2006. The 345 GHz receivers weretuned to the HCN(J=3–2) line (265.886 GHz at rest)in the LSB; the USB was used for continuum measure-ments. The weather conditions were good with opacitiesof τ =0.05-0.1 in the January track (extended config-uration) and τ =0.13-0.22 in the October/Novembertracks (very extended configuration). We used 3C273,3C111 and/or 3C454.3 as bandpass and Uranus, Titan,and/or Neptune as flux calibrators . We observed twoquasars (0235+164, 0238-084, 0339-017, and/or 0423-013) every ∼
15 minutes to calibrate the gains (amplitudeand phase versus time). The data from all three trackshave been combined into one single data file, resulting inan rms noise of 24 mJy in 17 km s − wide velocity chan-nels. The synthesized beam is determined to be 1 . ′′ × . ′′ ◦ (natural weighting) and0 . ′′ × . ′′
46 at PA=30 ◦ (robust weighting). HCO + (J=3–2) emission We carried out observations of the HCO + (J=3–2)emission in NGC 1068 using seven antennas in extendedconfiguration during November 2006. The 345 GHz re-ceivers were tuned to the HCO + (3–2) line (267.558 GHzat rest) in the LSB; the USB was used for continuummeasurements. The weather conditions were good withopacities of τ =0.06-0.15. Bandpass calibration wasperformed on 3C273, Titan and Uranus, while absolutefluxes were determined using Titan. The gains have been We used the line-free USB respectively for Titan and Neptuneto determine the absolute flux level as they are known to have broadHCN lines which could contaminate a flux calibration in the LSB. calibrated on 0423-013 and 0339-017. For this data set,we reach an rms noise of 33 mJy in 17 km s − wide ve-locity channels. The synthesized beam is determined tobe 1 . ′′ × . ′′ ◦ (natural weighting). CO(J=3–2) and HCO + (J=4–3) emission The CO(J=3–2) emission of NGC 1068 was ob-served in extended configuration using all eight anten-nas during September 2007. HCO + (J=4–3) was addi-tionally observed in a separate track in August 2007.The 345 GHz receivers were tuned to the CO(J=3–2) line (345.796 GHz at rest) in the LSB such that theHCO + (J=4–3) (356.734 GHz at rest) still falls withinthe USB. The opacities ranged between τ =0.06-0.13.Bandpass calibration was performed on 3C454.3 andUranus. Uranus was also used for flux calibration. Gainswere determined using 0238+166 and 0423-013. An rmsnoise of 80 mJy is reached in 7 km s − wide velocitychannels. The synthesized beam is determined to be1 . ′′ × . ′′ ◦ for natural weighting when also us-ing a uv-taper to better match the angular resolutions ofthe other SMA observations. The original (un-tapered)angular resolution amounts to 0.6 × ◦ . CO(J=2–1) and C O(J=2–1) emission
The CO(J=2–1) emission in NGC 1068 has been ob-served in extended configuration using all eight anten-nas during January and February 2008. The 230 GHzreceivers have been tuned to the CO(J=2–1) line(220.399 GHz at rest) in the USB such that theC O(J=2–1) (219.560 GHz at rest), the HC N(J=23-22) (209.230 GHz at rest) and H32 α (210.502 GHzat rest) still fall within the LSB. However, only the CO(J=2–1) and C O(J=2–1) line emission was de-tected. The opacities ranged between τ =0.1-0.2.Bandpass calibration was performed on 0423-013, 3C111and Titan, while gains were determined using 0339-017and 0423-013. Titan was further used as a flux calibra-tor. An rms noise of 12 mJy was reached in 17 km s − wide velocity channels. The synthesized beam is deter-mined to be 1 . ′′ × . ′′ ◦ for natural weightingwhen also using a uv-taper to better match the angularresolutions of the other SMA observations. CO(J=3–2) emission
The CO(J=3–2) emission in NGC 1068 was observedin compact configuration using seven antennas in Oc-tober 2005. These observations were part of the ob-serving campain presented by Humphreys et al. (2005),which aimed to detect extragalactic H O maser emis-sion at (sub-)millimeter wavelengths. The 345 GHz re-ceivers have been tuned to the H O(10(2,9)-9(3,6)) maserline (321.226 GHz at rest) in the LSB such that the CO(J=3–2) line (356.734 GHz at rest) was still lo-cated within the USB. The 225 GHz zenith opacity hasremained stable around 0.05-0.06. Bandpass calibra-tion has been performed on 3C454.3, 3C111 and Uranus.Uranus has been also used as a flux calibrator. The gainswere determined using 0234+285 and verified against0215+015 and 0420-014. An rms noise of 61 mJy isreached in 7 km s − wide velocity channels. The syn-thesized beam is determined to be 2 . ′′ × . ′′ ◦ (natural weighting). Krips et al. IRAM PdBI CO(2–1) and CO(J=1–0) emission
Observations of the CO(2–1) emission in NGC 1068were carried out with the IRAM PdBI in February2003 using all six antennas in A configuration. Si-multaneously, we observed the CO(J=1–0) using the3 mm PdBI receivers. The bandpass was calibrated onNRAO150 and 0420-014 while phase and amplitude cal-ibration were performed on 0235+164 and 0238-084. Atotal bandwidth of 580 MHz with a spectral resolutionof 1.25 MHz was used. We reach an RMS of ∼ − wide channels (natural weighting) at 1 mmand of ∼ − wide channels (naturalweighting) at 3 mm. Applying natural weighting in themapping process, beam sizes are derived to be 1 . ′′ × . ′′ ◦ at 1 mm and 2 . ′′ × . ′′ ◦ at 3 mm.However, to better match the SMA observations, wemapped the CO(2–1) data with a uv-taper giving aneffective angular resolution of 1.0 ′′ × ′′ at PA=30 ◦ . Asthe uv-coverages between the SMA and PdBI data (forthe high angular resolution data) are very similar, theusage of a simple uv-taper already provides the neces-sary accuracy to match the restoring beam of the PdBIobservations with that of the SMA. RESULTS
Continuum emission: from 850 µ m to 1.4 mm The sub(mm) continuum emission at the wavelengthspresented in this paper was derived from the line-freechannels of the respective line observations (Table 2),averaging emission from USB and LSB where possible(for the SMA data; the PdBI data were obtained fromsingle-sideband observations). The continuum emissionhas further been merged between data sets with very sim-ilar observed frequencies (i.e., within ∼
10 GHz). Beforeaveraging data from different sidebands and/or observa-tions, it has been carefully verified that the absolute flux,position and structure of the emission in the individualdata sets are consistent with each other within the cali-brational uncertainties of 20% (flux) and 0.1 ′′ (position)in order to reduce systematic errors.The continuum emission is clearly detected at all wave-lengths at a ≥ σ level. To obtain accurate fluxes, posi-tions and sizes, elliptical Gaussians were fitted to thedata in the uv-plane, except for the uniformly weightedmap of the 1.0 mm continuum emission for which a cir-cular Gaussian was fitted given its apparently unresolvednature. The results of these fits are listed in Table 3.The continuum emission at 1.0 mm (NA), 1.3 mm and1.4 mm appear to be consistent with each other in termsof their flux, position and structure (see Table 3 andFig. 1; see also Krips et al. 2006). All show peak fluxesof around 15-19 mJy/beam and spatially integrated fluxdensities of 22-28 mJy, indicating extended emission.Their positions, although self-consistent, are slightly tothe North ( ∼ ′′ ) of the radio position of the AGN (com-ponent S1 from Gallimore et al. 2004, marked with awhite cross in Fig. 1) and that of the uniformly weighted1.0 mm continuum emission (Fig. 1b; white contours).The shift between the mm and cm data is larger thanthe positional uncertainty of 0.1 ′′ and thus assumed tobe real. NGC 1068 is known to have a pronounced radio(and mm-) jet in a North-East-to-South-West direction, of which the North-Eastern part exhibits the strongeremission (e.g., Gallimore et al. 2004; Krips et al. 2006).Despite the steep spectral index of the synchrotron emis-sion of the jet, the extended (i.e., > ′′ ) emission fromboth the jet and counter-jet are still visible at 3 mm (e.g.,Krips et al. 2006; Schinnerer et al. 2000) but are signifi-cantly fainter or undetected at shorter wavelengths (i.e., . ∼ ′′ . Dueto the higher angular resolution, the 1.0 mm contin-uum emission of the jet in the uniformly weighted map(Fig. 1b) is almost entirely resolved out leaving behindonly the more compact emission of the AGN. Thus, thecentroid of the emission at 1.0 mm (NA), 1.3 mm and1.4 mm will naturally be shifted towards the North, whilethe 1.0 mm (UN) continuum emission should reveal theactual position of the AGN (or at least the base of thejet).Going to even shorter wavelengths of 850 µ m, it ap-pears that not only the continuum flux increases again,but also its position seems to be now consistent withthe AGN, independent of the weighting (i.e., synthesizedbeam) used for mapping/cleaning and unlike the 1.0 mm(NA), 1.3mm, and 1.4 mm continuum emission. Thelatter may indicate that the emission from the radio jetis negligeable at 850 µ m (see also Fig. 3 in Krips et al.2006) and the AGN (i.e., the S1 component) dominates.The increased flux at 850 µ m, which appears to be largerby almost a factor of 2 compared to the 1.0 mm-1.4 mmemission, shows that thermal dust emission already playsa significant role at 850 µ m (see Section 5.1). Also, thesize and shape of the continuum emission appear to havechanged compared to that at longer wavelengths. ThePA of the 850 µ m emission of ∼ ◦ is significantly differ-ent from that ( ∼ ◦ ) of the 1.0 mm-1.4 mm emission.Moreover, the 850 µ m emission appears to be extended(Fig. 1c), in contrast to the 1.0 mm-1.4 mm emission,which seems to be extended only in the jet componentbut not in the ’left-over’ AGN component in the uni-formly weighted 1.0 mm map (Fig. 1b). The uniformlyweighted 850 µ m continuum emission appears to be alsoresolved (white contours in Fig. 1c, compare also peakflux with total flux density in Table 3). Line emission
General Characteristics and Distribution of theMolecular Gas
The continuum emission has been subtracted from allline data in the uv plane to avoid any contamination ofthe line by continuum emission even if in some cases thecontinuum emission does not exceed the noise level inthe individual channel maps (see Table 2 & 3). In or-der to reduce systematic effects due to spatial filtering,we used a slight uv-taper (giving some more weight tothe shorter baselines) to map and clean all line emissiondata with the same synthesized beam, except for the highangular resolution ( ∼ ′′ ) of the uniformly weightedHCN(J=3–2) map (shown additionally in Fig. 2a usingwhite contours) and the low angular resolution ( ∼ ′′ )of the CO(J=1–0), CO(J=1–0) and CO(J=3–2)maps.illimeter observations of NGC 1068 5Fig. 2 shows the velocity integrated intensity mapsof the molecular line emission from HCN(J=3–2), HCO + (J=3–2), HCO + (J=4–3), CO(J=2–1), CO(J=2–1), CO(J=3–2), C O(J=2–1), CO(J=1–0) and CO(J=1–0); CO(J=3–2) is plotted in greyscale in all images to facilitate a comparison. All thesemolecules have been clearly detected above the 5 σ level(except CO(J=1–0)). The emission in all lines revealsa pronounced peak on the stronger eastern knot and inthe stronger lines also a weaker peak on the western knot,both of which are already known from previous COobservations (e.g., Schinnerer et al. 2000). An ellipticalGaussian has been fitted to the uv-data for all lines inorder to obtain the position, peak- and spatially inte-grated flux of the emission in the two knots. The re-sults of the fits are given in Table 4. The position of theemission in the eastern and western knot is very simi-lar in all observed lines, excluding the CO(J=2–1) andC O(J=2–1) emission which seem to peak closer to theAGN.The spectrum of the spatially integrated emission (overthe central ∼ ′′ ) of each line is plotted in Fig. 3. Wealso show the CO(J=1–0) line emission taken fromSchinnerer et al. (2000) for consistency. While the veloc-ity integrated line emission seems to be very similar in itsshape and position for most lines, the line profiles varysignificantly from each other. While for the dense gastracers (HCN, HCO + ) a single Gaussian fit is sufficientto reproduce the line profiles, the CO lines need a dual,triple or quadrupole Gaussian fit. However, to simplify acomparison, the results given in Table 5 represent a singleGaussian fit to all lines. The line centers are roughly con-sistent with each other, differing by less than 20 kms s − .Excluding the CO(J=2–1) and C O(J=2–1) emission,the line widths also agree with each other within the un-certainties. Except for the CO(J=2–1) and CO(J=2–1) line emission for which roughly half of the emissionseems to be resolved out, the interferometric observationshave captured most of the emission measured with sin-gle dish observations (Table 5). Please note that for the CO(J=1–0) (Schinnerer et al. 2000) and CO(J=1–0)emission, the single dish fluxes are much higher thanthe interferometric ones because they contain significantemission from the star-forming ring/spiral-arms and thebar.
Dynamical Characteristics of the Molecular Gas
The kinematic behaviour of the different molecules ispresented in detail in Fig. 4 to 11. To better under-stand the puzzling complexity of the different profiles ofthe various molecular lines and test whether it is dueto dynamical effects, we spatially split the spectra byderiving the spectrum of the western and eastern knotseparately (Fig. 4). Please note that the CO(J=1–0)and C O(J=2–1) line emission were discarded becauseof their insufficent sensitivity and/or lack of emission inthe western knot while CO(J=1–0) and CO(J=3–2)are not included because of their insufficient angular reso-lution. The iso-velocity maps (Fig. 5) of the CO, CO,HCN and HCO + line emission clearly show a dynamicalstructure that seems to be dominated by standard diskrotation with a blueshifted eastern knot and a redshiftedwestern knot. If disk rotation were the only underlyingkinematics, one would expect to find a simple blueshifted peak at the eastern knot and a redshifted peak at thewestern knot. Although disk rotation is observed, Fig. 4shows kinematic features significantly differing from sim-ple rotation. Instead, the blueshifted eastern knot alsoexhibits redshifted emission and the redshifted westernknot blueshifted emission. These ’wings’ appear to bepresent in all three CO emission at a high significancelevel as well as in the CO, HCN and HCO + emissionbut, given the lower signal-to-noise ratio (SNR) for theselines, not as pronounced as for CO. At this point, itshould be emphasized that in such a case the moment onemap can be very misleading as it derives only the domi-nant kinematic structure and might overlook more com-plex underlying kinematics. Integrating (in velocity) thered- and blue-shifted parts of the line spectrum (Fig.6) aswell as analysing the channel maps (Fig. 7 & 8) might bethe more appropriate approach. Fig. 6 indicates a morecomplex distribution than expected from simple disk ro-tation. We find blueshifted emission spatially coincidentwith redshifted emission and vice versa; this seems to bemost pronounced for the CO, CO and HCN emis-sion. By looking at the channel maps of the CO(J=2–1) and CO(J=3–2) emission (which have the highestSNR), the red- on blueshifted and blue- on redshiftedemission is not only at low-velocities but also at highervelocities (which is especially visible in the CO(J=2–1) emission; see channels < − > +100 km/s inFig. 7). The same is true for the CO(J=3–2) emission(Fig. 8, although less pronounced, especially for veloc-ities >
80 km s − for which no emission can be foundanymore as opposed to CO(J=2–1)). We find a be-haviour of the HCN(J=3–2) emission similar to that ofthe CO(J=2–1) and CO(J=3–2) emission though ona much lower significance level. The CO(J=2–1) emis-sion seems to indicate, however, a different behaviour(see Fig. 9). Instead of being distributed in a ’ring’-like manner, the emission appears to be elongated morein a South-West to North-East direction (see especiallychannels maps between +50 km s − to −
20 km s − ).However, given the low sensitivity level, this structurehas to be treated with caution and needs confirmationby either higher sensitivity observations or other molec-ular lines such as SiO. Indeed, the SiO emission seemsto indicate a similar behaviour as discussed in separatepapers (see Garc´ıa-Burillo et al. 2008, 2010).The position-velocity diagrams of the CO and HCNemission, taken at different Position Angles (PA) in stepsof 30 ◦ across the CND, are shown in Fig. 10. The greyscale denotes the CO(J=2–1) emission for better com-parison. Overall, the kinematic structures in the differentlines strongly resemble each other. Also, the overlap ofthe red- on blueshifted emission can be seen quite wellin the position-velocity diagrams (see especially PA=60-120 ◦ in Fig. 10), strongly indicating pronounced non-circular motions in the CND of NGC 1068.In order to quantify and parametrise the observedcomplex kinematics, we follow the approach used byHeckman et al. (1989) and Baum et al. (1992). We de-termine three kinematic parameters from the slits takenat the different position angles used in Fig. 10 for the CO emission. These parameters are: 1.) the av-erage line-of-sight velocity dispersion σ , determined as0.426 × FWHM along each slit, 2.) the “rotational” veloc-ities ∆, determined from the difference between the aver- Krips et al.age velocities on either side of the nucleus along each slit,and 3.) the rms variation of the velocity ǫ for each pointalong the slit, defined as ǫ = q N Σ Ni =1 ( v i − v av ) , where N is the number of points along the slit, v i the intensityweighted velocity for point i and the v av the intensityweighted average velocity along the slit. A comparisonof the different parameters with each others, especiallythe ratios ∆ σ and ∆ ǫ allows to classify the dynamics intothree different groups (for a more detailed explanationwe refer to Baum et al. (1992)):ROTATORS: ∆ /σ &
1, ∆ /ǫ & /σ <
1, ∆ /ǫ ∼ /σ <
1, ∆ /ǫ < CO(J=2–1) are shown in Fig. 11, highlighting our pre-vious findings (we find very similar values when takingthe CO(J=3–2) emission). For most position angles,we find that the ratios are most consistent with calm non-rotation with one exception (PA=60 ◦ ) being located inthe area of the violent non-rotators. This strongly em-phasizes the fact that although there is an underlyingdominating disk rotation, the dynamics of the molecu-lar gas in the CND of NGC 1068 is significantly dis-turbed by a non-rotational process, most significantly forthe North, North-Eastern part of the CND (i.e., alongPA=0-90 ◦ ). We also attempted to determine the rota-tional velocities directly from azimuthally averaging thevelocities within the CND and then subsequently fittingrotational curves to the data using rotcurv in gipsy. Wethereby assumed different scenarios ranging from puredisk rotation to adding a radial dependency. However,no physically meaningful result could be obtained. Wemostly find positive velocities that decrease with radiuswhich is inconsistent with simple disk rotation (see alsothe dynamical analysis done in Schinnerer et al. 2000)and necessitates the inclusion of a bulge or central masscomponent whether in form of nuclear star cluster or amassive black hole in addition to the disk. This certainlyfurther emphasizes the complexity of the molecular gasdynamics in the CND of NGC 1068. Molecular Line Ratios
In order to constrain the excitation conditions andchemistry of the molecular gas, we derive the line ra-tios for the different molecules and transitions in sev-eral ways, by accounting for the different angular res-olutions (especially with respect to the CO(J=1–0), CO(J=1–0) and CO(J=3–2) emission). Before de-termining any line ratio, the line emission was brought tothe same (lower) angular resolution by using a uv-taper;this seems appropriate for most of the lines as we recov-ered most of the emission with the interferometric obser-vations. Also, we compute line ratios only for emissioncoming from the samwe spatial regions (see Figs.12-14).Fig. 12 shows the velocity integrated line ratios betweenvarious combinations of the molecular lines. Separatingspatially the eastern and western knot we derive spatiallyaveraged line ratios from Fig. 12 which are listed in Ta-ble 6. Please note that the values in Table 6 might varyfrom those derived from Tables 4 and 5. However, the differences can be easily explained by the different se-quence of averaging (i.e., first in space, then in velocityversus first in velocity, then in space) for Table 5, the us-age of an elliptical Gaussian fit for Table 4 as opposedto spatially averaging without a fit as done for Table 6,and the different spatial resolutions of the line emissionin Tables 4 & 5.We identify some spatial variance of the different lineratios (mostly a factor ∼ CO and CO line ratios (Fig. 12a,d,o).The HCN and HCO + line ratios seem to be more con-stant over the CND than CO and CO. The highervalues are found closer to the position of the AGN formost maps.In order to investigate whether there might be a veloc-ity (and spatial) dependence on the line ratios, we deter-mined the line ratio channel maps for the two strongest CO transitions (J=2–1 and J=3–2) as function of ve-locity in Fig. 13. We find somewhat higher (i.e., fac-tor of & CO(J=3–2)-to- CO(J=2–1) ratios closeto the AGN at velocities around the systemic velocitybut also on the eastern knot at high negative velocities( . −
130 km s − ). Both knots seem to show (more orless) the same velocity and spatial behaviour as can beseen in Fig. 14. This plot shows the spatially averaged CO(J=3–2)-to- CO(J=2–1) line ratios for the east-ern and western knot as function of velocity. The twocurves follow each other nicely except for velocities be-tween −
120 to −
140 km s − for which the eastern knotexposes higher values (by a factor of 2). The error barsdenote the variance of each averaged value which in mostcases indicates a variation by a factor of 1.5. DISCUSSION
Spectral Energy Distribution of the ContinuumEmission
The nature of the continuum emission (from IR oversub-mm to cm wavelengths) represents a highly debatedand complicated matter for NGC 1068, recently gain-ing a revival by newly published VLTI/MIDI (i.e., IR)and radio data (e.g., H¨onig et al. 2008; Cotton et al.2008). As mentioned in the previous section, the ra-dio continuum emission splits up into several compo-nents, a jet plus counter-jet and a core component (S1)associated with the AGN itself. While the emissionfrom the jet is certainly pure non-thermal synchrotronemission, as supported by its steep continuum spec-trum (e.g., Gallimore et al. 2004; Cotton et al. 2008),the nature of the emission from S1 is highly contro-versial. Gallimore et al. (2004) already rule out syn-chrotron emission as origin for the continuum spectrumof S1 and discuss electron-scattered synchrotron emis-sion as well as thermal free-free absorption as alter-natives. While Krips et al. (2006) present argumentsfor electron-scattered synchrotron emission based on aturnover seen between cm and mm-data, Cotton et al.(2008) rather support the thermal free-free absorptionmodel. A highly complicating factor in this discussion iscertainly the mismatch in angular resolution between the Although this a reasonable first order fit, the line emission iscertainly not exactly of an elliptical Gaussian shape so that some ofthe emission is not well reproduced by fitting an elliptical Gaussianto the velocity integrated maps. illimeter observations of NGC 1068 7cm, mm and IR data. As discussed in Krips et al. (2006)and this paper, the mm-continuum emission is contami-nated by emission from the jet at angular resolutions of & ′′ , introducing large uncertainties in the estimate ofS1’s flux (compare H¨onig et al. 2008; Cotton et al. 2008;Krips et al. 2006) due to the lack of angular resolution.However, the previous estimate of the 1.3 mm continuumflux of (10 ±
4) mJy in Krips et al. (2006), translating to(9 ±
4) mJy at 1.0 mm, is very similar to the measured ±
2) mJy from our high-angular resolution SMA observations. Given the unre-solved nature of the latter, jet emission does not seemto be significant anymore at this angular resolution (seealso Fig. 1 in Cotton et al. 2008) although the obtainedangular resolution is still an order of magnitude largerthan that at cm wavelengths.Taking also the new 850 µ m (UN) continuum flux mea-surements into account, we recomputed the spectral en-ergy distribution (SED) and replotted the models used byH¨onig et al. (2008) and Krips et al. (2006). We therebybase our graphs on the formula and parameters specifiedin Equations (2)-(5) and Table 2&3 in H¨onig et al. (2008)and Equations (1)-(2) and Fig. 3 in Krips et al. (2006).The results are shown in Fig. 15a-c. We marked all datapoints with a circle for which the obtained angular reso-lution of the observations did not exceed 1 ′′ . We alsofitted a two-temperature grey body to the IR data, in or-der to estimate the contribution of thermal dust emissionto the sub-mm continuum emission. Although this grey-body fit is certainly not as sophisticated as the clumpytorus model used in H¨onig et al. (2008), it represents areasonable approximation as demonstrated by the goodmatch to the IR data points.As can be seen in Fig. 15b and c, the models usedby H¨onig et al. (2008) significantly overestimate the ob-served 1 mm (UN) flux by a factor of 2-3, althoughthey reproduce correctly the 850 µ m one. It seemsthat the electron-scattered synchrotron emission modelin Fig. 15a and the thermal free-free emission model inFig. 15c both need the extra contribution from the ther-mal dust emission to correctly reproduce the 850 µ m(UN) flux, while the synchrotron model in Fig. 15bdoes not require it. Thus, it appears to be indeed verylikely that the continuum emission is dominated by ther-mal dust emission starting at wavelengths . µ m, asposited above.Based on the 1 mm (UN) flux, it seems that the modelbest reproducing the SED at cm and mm wavelengthsis the electron-scattered synchrotron emission; the ther-mal free-free absorption seems to overpredict the 1 mmflux. However, observations of the continuum emission inNGC 1068 have to be conducted at similarly high angu-lar resolutions ( ≪ ′′ ), in order to dispel all remainingdoubts, although the new mm observations presented inthis paper are already a step in the right direction. Dynamics of the Molecular Gas
In previous studies, the complex kinematic behaviourof the molecular gas has been thought to be a conse-quence of a warped disk. The warped disk has been Please note, that this is true for all data points except that at3 mm. The encircled S1 data point has been estimated at 3 mm,not observed. We added this data point for consistency reasonsonly. modelled with a tilted ring model (e.g., Schinnerer et al.2000). However, the spatial overlap between the red- andblueshifted emission (i.e., the existence of highly non-circular motions) cannot be reproduced by these tiltedring models because they are based on circular motionsand thus cannot account for non-circular motions of thegas (within the plane).Even though we cannot rule out a warped disk sce-nario in which part of the gas could be trapped in el-liptical orbits producing the non-circular motions, wewant to propose an alternative approach, following re-cent findings on the H (1–0) S(1) emission at scales of100-150pc by M¨uller-S`anchez et al. (2009) and the modelproposed by Galliano & Alloin (2002). The nature of thedynamics displayed in Fig. 4, 5, 6, 7 and 8 could alsobe explained by the following scenario: a rotating diskplus an outflow of the disk gas due to shocks and/or aCND-jet interaction. This hypothesis seems to furthergain support when considering besides the H µ m map (Bock et al. 2000, see their Fig. 4) andthe 5cm radio-continuum map (Gallimore et al. 2004, seetheir Fig. 1). The H µ m emission follownicely that of the radio jet in the inner 1 ′′ (North/North-East direction) which seems to interact with emissionfrom the molecuar gas in the CND at ± ′′ in the North-ern part (see next Sub-Section); both the blue and red-shifted components associated with the non-circular mo-tions to the East and West of where the jet goes throughor lies in front of the CND (see also the case of M51;Matsushita et al. 2007, 2004). As the jet shows a biconi-cal structure with a change in direction close to the CND,it is not unreasonable to believe that part of it indeed hitsthe CND (see also Kraemer et al. 1998). Such an inter-action could easily produce a shock through/along theCND (at least in the northern part, i.e. the bridge be-tween the eastern and western knot) feeding the assump-tion that some part of the molecular gas in the ring/diskmight be blown outwards.Other causes for expanding/shocked gas include hy-pernovae explosions, stellar winds from a super stellarcluster as suspected in some nearby starburst galax-ies (such as NGC 253 or M82, Sakamoto et al. 2006;Matsushita et al. 2000) or cloud-cloud collision withinthe CND (i.e., within the inner Lindblad resonance; seeGarc´ıa-Burillo et al. 2010). However, they seem ratherunlikely since the CND of NGC 1068 does not show anysigns of starburst activity and the expansion seems tobe too “ordered”, i.e., too symmetric, to be caused byhighly chaotic cloud-cloud collisions. Also, we cannotentirely rule out that the dynamics we see in the molec-ular gas is due to inflowing rather than outflowing gas,especially since indications of inflowing gas along the jetwithin the central ∼
70 pc (i.e., on scales smaller than theCND) have been already presented in previous studies(see M¨uller-S`anchez et al. 2009). However, based on theappearance of the molecular gas within the disk (ring-like structure with an apparent void of gas in the innerpart) lets us favor the jet-gas interaction rather than aninflow scenario (on scales of 100-150 pc). An outflowon scales of 100-150 pc is not necessarily in contradic-tion with an inflow scenario on scales .
70 pc proposedby M¨uller-S`anchez et al. (2009). The jet-gas interactioncould equally drag gas outwards on scales larger than Krips et al.100 pc but trigger an inflow at smaller scales, dependingon the type of interaction between the jet and the molec-ular clouds. However, higher-angular resolution obser-vations at sub-arcsecond angular resolution (as possiblewith ALMA) will certainly help to distinguish betweenthe different scenarios.Based on this alternative approach, we tried to repro-duce the molecular gas dynamics with a very simplisticmodel that includes a dominant (Keplerian) rotating diskplus an outflow of some of the disk gas. We thereby as-sume a velocity gradient of ∆v =200 km s − , a radius of ∼ ′′ and an inclination of ∼ ◦ for the rotating disk.We additionally add a slight ellipticity of 5% and assym-metry ratio between the eastern and western part of 1:0.7to the model. The outflow is approximated by a slightlyexpanding elliptical ring. We assume the same elliptic-ity of 5% as before and that the expansion starts at theinner radius of the disk. The expansion rate (H) is cho-sen to be 200 km s − per 1 ′′ or 67 pc equivalently (i.e.,H=3 km/s/pc); this value is similar to what has been es-timated for M51 (2.2 km/s/pc; Matsushita et al. 2007).We also introduced a slight asymmetry ratio of 1:0.9 be-tween the blue- and redshifted emission. We further as-sume that up to ∼
30% of the disk gas is expanding. Mostof these values that were chosen for the model parameterswere almost instantly obvious from the observations, es-pecially the velocities and radii. We hence used them asstarting values and scanned then through a reasonableparameter space for the optimal combination of inputvalues that matches best the observed maps. However,given the larger number of parameters used in this modeland hence the many degrees of freedom, we did not ac-tually conduct a true fit to the data but rather a “fit”by eye. The results of this so found “best-fit” model aredisplayed and compared to the CO(J=2–1) emissionin Fig. 16-20. Indeed, the major dynamical character-stics of the molecular gas emission can be reproducedby this simplistic model supporting the hypothesis of anadditional gas outflow. Also, using the same kinematicparametrisation for the model as used for the CO emis-sion (see Fig. 11), we find very similar ratios between theobserved and simulated emission. We have to stress thatthis suggestion does not exclude a warped CND but it isnot needed for this model.It is interesting to note as further support for our ap-proach that the distribution of the HCN(J=3–2) emissionmatches almost exactly that of H CO(J=2–1) emission (see Fig. 1 inM¨uller-S`anchez et al. 2009). Similar to the H map thatindicates the brightest emission toward the North, alsomost of the HCN emission is found toward the northernpart of the CND at which most of the potentially shockedgas would lie. Thus, one would expect the densest (andhottest) part of the gas in that area (see next section). Excitation conditions of the gas
The line ratios derived from the interferometric maps,especially for the HCN, HCO + and CO emission(Fig. 12 and Table 6), are consistent with previousfindings from single-dish observations (e.g., Krips et al.2008). They support a picture in which the molecu-lar gas in the CND is relatively dense (n(H ) ≤ . )and warm (T k >
40 K) with potentially higher thannormal (i.e., in galactic giant molecular clouds) HCN abundances (Z(HCN)=[HCN]/[H ]) of Z(HCN)=1-50 × Z galactic (HCN) (Z galactic (HCN)=2 × − ; e.g., Irvine1987). Krips et al. (2008) argue that the high HCN-to- CO(J=1–0) and HCN-to- CO(J=1–0) in NGC 1068can be explained by either higher than normal HCNabundances due to an XDR (see also Usero et al. 2004;Sternberg et al. 1994) and/or higher gas temperaturesleading to hypo-excited CO(J=1–0) emission. The lat-ter is supported by decreasing HCN-to-CO line ratioswith increasing rotational number J. However, strongconstraints on the kinetic temperatures could not be setbased on the single-dish observations alone. Also, mostsingle-dish observations are unable to unambigiously dis-tinguish between the molecular gas emission in the centerand that in the star-forming spiral arms, complicatingany interpretation of the data. Furthermore, most of theinterferometric data previously published mostly focuson CO at moderate angular resolution and only two ofits transitions.The new interferometric maps, obtained for severaltransitions and molecules at sufficiently high angular res-olution, overcome some of the short-comings of previousobservations/analyses.In the following we will discuss results from simulationsof the excitation conditions of the molecular gas carriedout with the radiative transfer code RADEX developedby Van der Tak et al. (2007). Please note that we didnot find significant difference when using the LVG codein MIRIAD or the RADEX code. Given simplified simu-lations with RADEX, we decided to use RADEX in thispaper as opposed to Krips et al. (2008) in which an LVGcode was used.RADEX offers three different possibilities for the es-cape probability method: a) a uniform sphere, b) an ex-panding sphere (Large-Velocity-Gradient, LVG), and c)a plane parallel slab (shock). We used all three methodsbut did not find significant differences for our data be-tween them. Thus, in order to keep the interpretation assimple as possible, we will discuss the results with respectto a uniform sphere in the following.We carried out simulations with RADEX for eachmolecule using the following grid of parameters (dimen-sion: 51 × × • Kinetic Temperature: T kin = 1 - 500 K • Gas density: n(H )=10 - 10 cm − • Column density: N( CO) = 10 - 10 cm − We define the abundance ratio be-tween one molecule (MOL1) and another(MOL2) as X
MOL1MOL2 ≡ Z(MOL1)/Z(MOL2) withZ(MOL1) ≡ [MOL1]/[H ]. For the different molecu-lar abundance ratios we assume the following ranges: • Abundance ratios:X CO CO ≃ HCNHCO + ≃ C/ C abundance ratio are ∼
20 whichillimeter observations of NGC 1068 9increase to values of 80-100 in the outer parts of thegalaxy (e.g. Wilson & Rood 1994; Wilson & Matteucci1992). Nearby starburst galaxies and ULIRGs showsomewhat higher C/ C abundance ratios of >
30 thanfound in the Galactic Center (e.g., NGC 253, M82,IC 342, NGC 4945, NGC 6240 see Greve et al. 2009;Henkel et al. 1998, 1994, 1993a; Henkel & Mauersberger1993b) with Arp 220 being the exception. Greve et al.(2009) find a very low abundance ratio of only 8 forthis galaxy. Furthermore, some high redshift galax-ies seem to exhibit also rather high C/ C values of >
30 as determined for the ISM in the graviational lensof PKS 1830 −
211 (Muller et al. 2006) or the Clover-leaf (Henkel et al. 2010). Values of X
HCNHCO + show anequally large scatter ranging from the order of unity instarforming regions in the Milky Way (such as Orionand SgrB2; see Blake et al. 1987) and nearby starburstgalaxies/ULIRGs (such as M82, NGC 2146, NGC 253,NGC 4945, NGC 6240 and Arp 220; see Naylor et al.2010; Greve et al. 2009; Krips et al. 2008; Wang et al.2004) to ≥
10 in nearby Seyfert galaxies (e.g., Krips et al.2008).For the simulations, we will concentrate on the regionof the CND that contains all molecules. This region cor-responds to the bridge between the eastern and westernknot, i.e., the northern part of the CND, and has a sizeof roughly ∼ ′′ ( ≃ CO & CO emission
We conducted a χ -test on the RADEX grid by usingthe observed line ratios for the CO and CO line emis-sion. Fig. 21 shows the parameters for the best χ -testfor four exemplary abundance ratios (10,26,52,110) fromthe aforementioned range; abundance ratios at the lowerand higher end of the range show somewhat higher χ values.The middle panel shows the lowest χ found in eachrange of column densities and abundance ratios. Thelower and upper panels show the respective lower andupper limits for the column density for which still a rea-sonably low χ was found to indicate the spread in col-umn densities.The availability of the three lowest transitions for COand CO allows us to set tight constraints on T kin ,n(H ) and N( CO). The observed CO line ratios mostimpressively restrict the kinetic temperatures to valueswell above 200 K. This strengthens previous indicationsof warm/hot ( T kin >
50 K) molecular gas in the CND ofNGC 1068 (Kamenetzky et al. 2011; Krips et al. 2008;Matsushita et al. 1998; Sternberg et al. 1994) but be-ing much higher than the value found by Tacconi et al.(1994). However, Tacconi et al. (1994) base their sim-ulations on single-dish data which cannot distinguishwell between emission from the starforming ring/spiralarms and the CND. Especially the CO(J=1–0) emis-sion might be overestimated for the CND which clearlyis hardly detected in the interferometric map despite thehigh sensitivity of the observations. The lowest χ isactually found for kinetic temperatures around 450 K,which seems to be fairly high. Although we used aone-gas component model due to the lack of sufficient observational constraints (i.e., higher-J CO transitionswith J upper ≥ CO or NH ;see Ao et al. 2011). High kinetic temperatures wouldbe expected in several scenarios, among them shocks aswell as heating through X-ray radiation from the AGN(e.g., Meijering et al. 2007; Meijering & Spaans 2005).Nevertheless, a significant fraction ( >> ≤ . -10 . cm − , consistent withprevious findings (Krips et al. 2008; Matsushita et al.1998). Considering the range of assumed abundanceratios, the column density of CO approximatelyspans a range between ∼ . cm − and 10 . cm − .Even though CO might not be the best tracer ofthe C/ C isotopic ratio (e.g., Mart´ın et al. 2010)we assume it to be a first approximation to this ra-tio. The carbon ratio of 26 found for the abso-lute lowest χ is similar to the value of 20 measuredin the Galactic Center region (e.g., Wilson & Rood1994) and lower than that derived towards nearby star-bursts (e.g., Henkel et al. 1993a; Henkel & Mauersberger1993b; Henkel et al. 1994). Such C enrichment wouldpoint towards a highly nuclear processing of the ISM inthe central region of NGC 1068.
HCN & HCO + emission In Krips et al. (2008) we carried out an LVG analy-sis for NGC 1068 based on the HCN and HCO + single-dish lines ratios with kinetic temperatures not exceed-ing 200 K. As our new simulations indicate kinetic tem-peratures lying significantly above 200 K, we repeatedthe simulations with RADEX allowing for a larger rangein kinetic temperatures. The line fluxes for the HCNand HCO + (J=1–0) emission are taken from PdBI obser-vations at ∼ ′′ angular resolution (Garc´ıa-Burillo et al.2008) which will be analysed in more detail in a later pa-per by Usero et al. (in prep.). The results of the RADEXsimulations for HCN and HCO + are shown in Fig. 22.Considering the restrictions for the kinetic tempera-tures ( >
200 K), we find solutions (i.e., with low χ ) withRADEX for which the HCN-to-HCO + abundance ratiolies between X HCNHCO + ≃ HCN , gHCO + ≃ + are not as sensitive to changes in the kinetic tem-peratures as CO and CO. They are better indicatorsof changes in the volume density, as can be nicely seenin Fig. 22; the volume density is restricted by the χ -testto a very small area and independently yields values ofn(H ) ≃ . -10 . cm − similar to the CO and CO0 Krips et al.results.The simulations indicate column densities for HCNin the range of N(HCN) ≃ . -10 . cm − which issmaller (a factor of ∼
10) than what was found byKrips et al. (2008) with the LVG code. However, resultsare quite similar to Krips et al. (2008) when assumingsimilar kinetic temperatures.Comparing the HCN column densities to those of CO(N( CO ) ≃ . -10 . cm − ), we obtain abundance ra-tios between HCN and CO of X COHCN & . which seemsto be still compatible with a slightly increased abundanceof HCN. Comparing the column densities of HCO + and CO, we find a somewhat decreased HCO + abundance(by a factor of at least 10 lower than found in galac-tic starforming regions). This agrees well with previ-ous results (e.g., Garc´ıa-Burillo et al. 2010; Krips et al.2008; Usero et al. 2004) suggesting an increased forma-tion (and hence increased abundance) of HCN due to anXDR in the center of NGC 1068. SUMMARY & CONCLUSIONS
The SMA and PdBI observations of the (sub-)mmemission in NGC 1068 presented in this paper show acomplex distribution, kinematics and excitation condi-tions of the molecular gas. The (sub)mm continuum andmolecular line emission is interpreted as follows:1.) The cm/mm-continuum emission seems to be bestreproduced by electron-scattered synchrotron emis-sion. Thermal free-free emission as proposed byH¨onig et al. (2008) overpredicts the high angularresolution 1 mm continuum emission.2.) The molecular gas is found to display a very com-plex kinematic behaviour in the CO, HCN andHCO + lines which is not reproducable by a tilted-ring model approximating a warped disk with cir-cular motions. Instead, a dominant rotating diskplus a radial outflow of some of the gas in the CND is proposed as an alternative explanation to ac-count for the non-circular motions.3.) The different line ratios from the CO, CO,HCN, and HCO + emission seems to be consis-tent with moderately dense and warm gas, bothbeing further support for a gas scenario in whichheated and compressed by a shock (at least in theNorth/North-Eastern part of the ring). The high-est line ratios are found close to the AGN and/orjet-CND ’contact’-point. In this picture, the in-creased kinetic temperatures seem to be one of theculprits for the unusually high HCN-to-CO(J=1–0)line ratios due to a hypo-excitation of the CO(J=1–0) line emission.5.) Consistent with previous papers, we find furtherindications of an increased HCN abundance inNGC 1068 (by a factor of ∼ + abundance (by a factor of ∼ + abundance ratio in theCND of this source.The Submillimeter Array is a joint project betweenthe Smithsonian Astrophysical Observatory and theAcademia Sinica Institute of Astronomy and Astro-physics and is funded by the Smithsonian Institution andthe Academia Sinica. We thank the anonymous refereefor a very careful and constructive report. We also wouldlike to thank Gaelle Dumas for very useful discussionon and help with the kinematic analysis of the COdata. I.S. acknowledges grant LC06014 of Ministry ofEducation of the Czech Republic. ABP thanks the Na-tional Radio Astronomy Observatory (NRAO) for sup-port on this project. NRAO is a facility of the NationalScience Foundation operated under cooperative agree-ment by Associated Universities, Inc. SM is supportedby the National Science Council (NSC) of Taiwan, NSC97-2112-M-001-021-MY3. Please note that we consider the two highest contours inFig. 22 as being acceptable solutions REFERENCESAalto, S., 2008, Ap&SS, 313, 273Antonucci, R. R. J., & Miller, J. S. 1985, ApJ, 297, 621Ao, Y., Henkel, C., Braatz, J. A., Wei, A., Menten, K. M., Mhle,S., 2011, A&A, 529, 154Baum, S.A., Heckman, T.M., van Breugel W., 1992, ApJ, 389, 208Blake, G. A., Sutton, E. C., Masson, C. R., & Phillips, T. G., 1987,ApJ, 315, 621Bland-Hawthorn, J., Gallimore, J. F., Tacconi, L. J., Brinks, E.,Baum, S. A., Antonucci, R. R. J., & Cecil, G. N. 1997, Ap&SS,248, 9Bock, J. J., et al., 2000, AJ, 120, 2904Cotton, W. D., Jaffe, W., Perrin, G., Woillez, J., 2008, A&A, 477,517Dale, D.A., Sheth, K., Helou, G., Regan, M.W., H¨uttemeister, S.,2005, AJ, 129, 2197Fuente, A., Garc´ıa-Burillo, S., Gerin, M., Teyssier, D., Usero, A.,Rizzo, J. R., de Vicente, P., 2005, ApJ, 619, L155Galliano, E., Pantin, E., Alloin, D., Lagage, P. O., 2005, MNRAS,363, L1Galliano, E., Alloin, D., 2002, A&A, 393, 43Gallimore, J.F., Baum, S.A., O’Dea, C.P., 2004, ApJ, 613, 794Gallimore, J.F., Henkel, C., Baum, S.A., Glass, I.S., Claussen,M.J., Prieto, M.A., & von Kapherr, A., 2001, ApJ, 556, 694Gallimore, J. F., Baum, S. A., & O’Dea, C. P. 1996, ApJ, 464, 198 Gao, Y., Solomon, Philip M., 2004, ApJ, 606, 271Garc´ıa-Burillo, S., et al., 2010, A&A, 519, 2Garc´ıa-Burillo, S., Combes, F., Usero, A., Graci´a-Carpio, J., 2008,J. Phys.: Conf. Ser. 131, 2031Garc´ıa-Burillo, S., Combes, F., Usero, A., Graci´a-Carpio, J., 2007,NewAR, 51, 160Greenhill, L. J., Gwinn, C. R., Antonucci, R., Barvainis, R., 1996,ApJ, 472, L21Greve, T. R., Papadopoulos, P. P., Gao, Y., Radford, S. J. E., 2009,ApJ, 692, 1432Guilloteau, S. & Lucas, R. 2000, in Imaging at Radio ThroughSubmillimeter Wavelengths, ed. J. G. Mangum & S. J. E.Radford (San Francisco: ASP), 299Heckmann, T.M., Baum, S.A., van Breugel, W.J.M., McCarthy, P.,1989, ApJ, 338, 48Henkel, C., Downes, D., Weiß, A., Riechers, D., Walter, F., 2010,A&A, 516, 111Henkel, C., Chin, Y.-N., Mauersberger, R., Whiteoak, J. B., 1998,A&A, 329, 443Henkel, C., Whiteoak, J. B., Mauersberger, R., 1994, A&A, 284,17Henkel, C., Mauersberger, R., Wiklind, T., Huettemeister, S.,Lemme, C., Millar, T. J., 1993, A&A, 268, L17Henkel, C. & Mauersberger, R., 1993, A&A, 274, 730 illimeter observations of NGC 1068 11
TABLE 1Properties of NGC 1068
Characteristic NGC 1068 ReferenceHubble Class . . . . . . . (R)SA(rs)b NED a AGN type . . . . . . . . . . Seyfert 2 Khachikian & Weedman (1974)Dynamical Center . . α J2000 =02h42m40.70s Gallimore et al. (2004) δ J2000 = − ◦ ′ . ′′ ± ± − Huchra et al. (1999)Luminosity Distance 12.6 Mpc b NED a Scale . . . . . . . . . . . . . . . 61 pc/arcsec NED a,b
Inclination Angle . . . 40 ◦ Bland-Hawthorn et al. (1997)Position Angle . . . . . 278 ◦ Bland-Hawthorn et al. (1997) a NED: NASA/IPAC Extragalactic Database b Used cosmology: Hubble constant H = 73 km s − /Mpc, Omega(matter) = 0.27, Omega(vacuum) = 0.73 Hoenig, S.F., Almudena Prieto, M., Beckert, Th., 2008, A&A, 485,33Huchra, J. P., Vogeley, M. S., & Geller, M. J. 1999, ApJS, 121, 287Humphreys, E. M. L., Greenhill, L. J., Reid, M. J., Beuther, H.,Moran, J. M., Gurwell, M., Wilner, D. J., & Kondratko, P. T.2005, ApJ, 634, L133Irvine, W. M., Goldsmith, P. F., Hjalmarson, A., in Interstellarprocesses, Proceedings of the Symposium, Grand Teton NationalPark, WY, July 1-7, 1986 (A88-14501 03-90). Dordrecht, D.Reidel Publishing Co., 1987, ASSL 134, p. 561-609Israel, F.P., 2009, A&A, 493, 525Jaffe, W., et al., 2004, Natur, 429, 47Kamenetzky, J., et al., 2011, ApJ, 731, 83Khachikian, E. Y., & Weedman, D. W. 1974, ApJ, 192, 581Kohno, K., Nakanishi, K., Tosaki, T., Muraoka, K., Miura, R.,Ezawa, H., Kawabe, R., 2008, Ap&SS, 313, 279Kohno, K., Matsushita, S., Vila-Vilar´o, B., Okumura, S. K.,Shibatsuka, T., Okiura, M., Ishizuki, S., & Kawabe, R. 2001,in ASP Conf. Ser. 249, The Central kpc of Starbursts and AGN,ed. J. H. Knapen et al. (San Francisco: ASP), 672Kohno, K., Kawabe, R., Tosaki, T., Okumura, S., 1996, ApJ, 461,L29Kraemer, S.B., Ruiz, J.R., Crenshaw, D. M., 1998, ApJ, 508, 232Krips, M., Mart´ın, S., Neri, R., Garc´ıa-Burillo, S., Combes,F., Petitpas, G., Fuente, A., Usero, A., “The Starburst-AGNConnection”, ASP Conference Series, Vol. 408, proc. of theconference held 27-31 October 2008, at Shanghai NormalUniversity, Shanghai, China. Edited by Weimin Wang, ZhaoqingYang, Zhijian Luo, and Zhu Chen. San Francisco: AstronomicalSociety of the Pacific, 2009, p.142Krips, M., Neri, R., Garc´ıa-Burillo, S., Mart´ın, S., Combes, F.,Graci´a-Carpio, J., Eckart, A., 2008, ApJ, 677, 262Krips, M., et al., 2007, A&A, L468Krips, M., et al., 2006, A&A, 446, 113Krolik, J. H., & Kallman, T. R. 1987, ApJ, 320, L5Mart´ın, S., et al., 2011, A&A, 527, 36Mart´ın, S., Aladro, R., Mart´ın-Pintado, J., Mauersberger, R., 2010,A&A, 522, 62Mart´ın, S., Mauersberger, R., Mart´ın-Pintado, J., Henkel, C.,Garc´ıa-Burillo, S., 2006, ApJS, 164, 450Meier, David S., Turner, Jean L., 2005, ApJ, 618, 259Muller, S., Guelin, M., Dumke, M., Lucas, R., Combes, F., 2006,A&A, 458, 417M¨uller-S`anchez, F., Davies, R. I., Genzel, R., Tacconi, L. Jz.,Eisenhauer, F., Hicks, E. K. S., Friedrich, S., Sternberg, A., 2009,ApJ, 691, 749Matsushita, S., Muller, S., Lim, J., 2007, A&A, 468, L49 Matsushita, S., et al., 2004, ApJ, 616, L55Matsushita, S., Kawabe, R., Kohno, K., Matsumoto, H., Tsuru,T.G., Vila-Vilar´o, B., 2005, ApJ, 618, 712Matsushita, S., Kawabe, R., Matsumoto, H., Tsuru, T.G., Kohno,K., Morita, K.-I., Okumura, S.K., Vila-Vilar´o, B., 2000, ApJ,545, L107Matsushita, S., Kohno, K., Vila-Vilar´o, B., Tosaki, T., Kawabe,R., 1998, ApJ, 495, 267Meijerink, R., Spaans, M., Israel, F. P., 2007, A&A, 461, 793Meijerink, R., & Spaans, M., 2005, A&A, 436, 397Naylor, B. J., et al., 2010, ApJ, 722, 668Paglione, T. A. D., Jackson, J. M., Bolatto, A. D., Heyer, M. H.,1998, ApJ, 493, 680Papadopoulos, P. P., 2010, ApJ, 720, 226P´erez-Beaupuits, J. P., Spaans, M., van der Tak, F.F.S., Aalto, S.,Garc´ıa-Burillo, S., Fuente, A., Usero, A., 2009, A&A, 503, 459Poncelet, A., Perrin, ., Sol, H., 2008, A&A, 481, 305Rigopoulou, D., Kunze, D., Lutz, D., Genzel, R., Moorwood, A. F.M., 2002, A&A, 389, 374Roussel, H., et al., 2007, ApJ, 669, 959Roy, A.L., Colbert,E.J.M., Wilson, A.S., & Ulvestad, J.S., 1998,ApJ, 504, 147Sakamoto, K., Aalto, S., Evans, A. S., Wiedner, M.C., Wilner, D.J., 2010, ApJ, 725, L228Sakamoto, K., Ho, P. T. P., Peck, A. B., 2006, ApJ, 644, 862Schinnerer, E., Eckart, A., Tacconi, L.J., Genzel, R. & Downes, D.,2000, ApJ, 533, 850Scoville, N.Z., Carlstrom, J.E., Chandler, C.J., Phillips, J.A., Scott,S.L., Tilanus, R. P.J., Wang, Z., 1993, PASP, 105, 1482Siemiginowska, A., & Elvis, M., 1997, ApJ, 482, L9Sternberg, A., Genzel, R., & Tacconi, L. 1994, ApJ, 436, L131Tacconi, L. J., Genzel, R., Blietz, M., Cameron, M., Harris, A. I.,Madden, S., 1994, ApJ, 426, L77Tomono, D., Terada, H., Kobayashi, N., 2006, ApJ, 646, 774Usero, A., Garc´ıa-Burillo, S., Mart´ın-Pintado, J., Fuente, A., Neri,R., 2006, A&A, 448, 457Usero, A., Garc´ıa-Burillo, S., Fuente, A., Mart´ın-Pintado, J., &Rodr´ıguez-Fern`andez, N. J. 2004, A&A, 419, 897Van der Tak, F.F.S., Black, J.H., Sch¨oier, F.L., Jansen, D.J., vanDishoeck, E.F. 2007, A&A 468, 627Wang, M., Henkel, C., Chin, Y.-N., Whiteoak, J. B., Hunt C., M.,Mauersberger, R., Muders, D., 2004, A&A, 422, 883Wilson, T. L. & Rood, R., 1994, ARA&A, 32, 191Wilson, T. L., & Matteucci, F., 1992, A&AR 4, 1
TABLE 2Chronological Summary of observations carried out for NGC 1068
Molecular Telescope Observing Config. a Frequency Band Zenith T sys
RMS Synthesized BeamLine Dates at rest Opacity Noise b major × minor,P.A. c (YYYY-MM) (GHz) at 225 GHz (K) (mJy) ( ′′ × ′′ , ◦ ) CO(J=2–1) SMA 2008-01,2008-02 EX 220.399 LSB 0.10-0.20 100-150 12 1.0 × d & C O(J=2–1) 219.560 LSBHCO + (J=4–3) SMA 2007-08,2007-09 EX 356.734 USB 0.06-0.13 200-400 32 e × f CO(J=3–2) SMA 2007-09 EX 345.796 LSB 0.06-0.13 200-400 51 1.0 × f HCO + (J=3–2) SMA 2006-11 EX 267.558 LSB 0.06-0.15 100-150 33 1.0 × f HCN(J=3–2) SMA 2006-01 EX 265.886 LSB 0.05-0.10 100-150 24 e × e,f SMA 2006-10,2006-11 VEX LSB 0.10-0.22 100-200 0.53 × d,e CO(J=3–2) SMA 2005-10 C 356.734 USB 0.05-0.06 200-300 39 2.4 × d CO(J=1–0) PdBI 2003-02 A 110.201 SSB n.a. 150-300 1.7 2.5 × f CO(J=2–1) PdBI 2003-02 A 230.538 SSB n.a. 200-600 4 1.0 × e a SMA configurations: C=compact (baselines up to 70 m), EX=extended (baselines up to 220 m), VEX=very extended (baselinesup to 500 m); PdBI configurations: A=most extended (baselines up to 500m) b in 17 km s − wide channels c P.A. is measured from North to East d using uniform weighting e combined for all tracks f using natural weighting TABLE 3Continuum Parameters for NGC 1068. λ Synth. Beam ∆ α a ∆ δ a Peak Flux Deconv. Sizemajor × minor,P.A. Flux Density b major × minor,P.A.( ′′ × ′′ , ◦ ) ( ′′ ) ( ′′ ) (mJy/beam) (mJy) ( ′′ × ′′ , ◦ )1.4 mm 1.0 × ± ± ± ± ± × (0.7 ± ± × ± ± ± ± ± × (0.5 ± ± c × ± ± ± ± ± × (0.4 ± ± d × ± ± ± ± ± e µ m 2.1 × − ± − ± ± ±
11 (1.1 ± × (0.8 ± ± µ m (NA) 1.0 × ± ± ± ± ± × (0.8 ± ± µ m (UN) 0.6 × ± − ± ± ± ± × (0.8 ± ± a The offsets are with respect to α J2000 =02h42m40.70s and δ J2000 =-00 ◦ ′ . ′′ α J2000 =02h42m40.709s and δ J2000 =-00 ◦ ′ . ′′
95 (e.g. Gallimore et al. 2004; Krips et al. 2006).Positional errors are of pure statistical nature and were derived from the Gaussian fit to the data. They do not include absolutepositional uncertainties from the calibration, which are estimated to be ∼ . ′′ b Flux errors are purely statistical and do not account for uncertainties of the flux calibration. The latter are estimated to beof the order 10-20% (see text). c Averaged continuum emission derived from the HCN(J=3–2) (vex+ext) and HCO + (J=3–2) observations (ext). Data weremapped using natural weighting (NA). d Averaged continuum emission derived from the HCN(J=3–2) observations alone (vex+ext) using uniform weighting (UN). e Here, only a circular Gaussian fit has been carried out, while for the rest an elliptical Gaussian has been fitted to the data(see text). illimeter observations of NGC 1068 13
TABLE 4Individual components of the molecular line emission in NGC 1068.
Molecular Component ∆ α a ∆ δ a Vel. Integrated Spatially IntegratedLine Peak Intensity a Intensity a ( ′′ ) ( ′′ ) (Jy beam − km s − ) (Jy km s − )HCN(J=3–2) E-knot +1.0 ± ± ± ± − ± ± ± ± + (J=3–2) E-knot +0.9 ± ± ± ± − ± ± ± ± + (J=4–3) E-knot +0.9 ± ± ± ± CO(J=3–2) E-knot +1.1 ± − ± ±
20 1330 ± − ± − ± ±
10 720 ± CO(J=3–2) E-knot +1.1 ± ± ± ± CO(J=2–1) E-knot +1.0 ± − ± ±
10 290 ± − ± − ± ± ± CO(J=2–1) E-knot +0.6 ± ± ± ± − ± ± ± ± O(J=2–1) E-knot +0.3 ± ± ± ± CO(J=1–0) E-knot +1.0 ± − ± ± ± − ± − ± ± ± CO(J=1–0) E-knot +0.5 ± ± ± ± − ± − ± ± ± a The parameters were determined by fitting a one- or two-component elliptical Gaussian profile to the uv -data of each line.Errors include the statistical uncertainties from the Gaussian fit and those from the calibration ( ∼ TABLE 5Molecular line parameters derived from the different line spectra of NGC 1068.
Molecular Velocity Line Line Vel. Integrated Vel. Integrated SD BeamLine Offset a,b,c
Flux a Width a,b,d
Intensity a,b
SD Intensity e (“)(km s − ) (Jy) (km s − ) (Jy km s − ) (Jy km s − )HCN(J=3–2) − ±
10 0.63 ± ±
30 150 ±
10 190 ±
10 9.5 ′′ HCO + (J=3–2) − ±
10 0.23 ± ±
50 50 ± ± ′′ HCO + (J=4–3) − ±
10 0.24 ± ±
40 60 ± ±
10 14 ′′ CO(J=3–2) − ± ± ±
10 2130 ± ±
300 14 ′′ CO(J=3–2) − ±
10 0.43 ± ±
30 100 ±
10 170 ±
20 14 ′′ CO(J=2–1) − ± ± ±
20 529 ± ± ′′ CO(J=2–1) − ±
10 0.50 ± ±
10 30 ± ± ′′ C O(J=2–1) +3 ±
10 0.10 ± ±
10 5.1 ± · · · · · · CO(J=1–0) +3 ± ± ±
10 120 ± ±
80 21 ′′ CO(J=1–0) +8 ±
10 0.009 ± ±
30 1.3 ± ± ′′ a The line parameters have been determined by fitting a single Gaussian line to the (spatially integrated) spectrum for eachmolecule. The line emission has been thereby integrated over the central 4 ′′ in NGC 1068. b statistical error from the Gaussian fit only. c with respect to v LSR =1137 km/s. d Full Width at Half Maximum (FWHM) e single dish (SD) integrated intensities as measured with the IRAM 30m and the JCMT telescope in the central 10-30 ′′ ofNGC 1068 (taken from: Israel 2009; P´erez-Beaupuits et al. 2009; Krips et al. 2008). The values were converted to Jansky scaleusing S[Jy]/T mb [K]=4.71 (30m) and S[Jy]/T mb [K]=15.6 (JCMT). K r i p s e t a l. TABLE 6Molecular line ratios for NGC 1068 a . (X[K]/Y[K]) Y CO CO C O HCN HCO + X J=1–0 J=2–1 J=3–2 J=1–0 J=2–1 J=3–2 J=2–1 J=1–0 J=2–1 J=3–2 J=1–0 J=3–2 J=4–3E-knot CO J=1–0 · · · ± ± ± · · · · · · · · · · · · · · · · · · · · · · · · · · · J=2–1 2.9 ± · · · ± · · · ± · · · ± · · · · · · ± · · · ± · · · J=3–2 6.0 ± ± · · · · · · · · · ± · · · · · · · · · ± · · · ± · · · CO J=1–0 0.05 ± · · · · · · · · · ± ± · · · · · · · · · · · · · · · · · · · · · J=2–1 · · · ± · · · ± · · · ± ± · · · · · · ± · · · ± · · · J=3–2 · · · · · · ± ± ± · · · · · · · · · · · · ± · · · ± · · · C O J=2–1 · · · ± · · · · · · ± · · · · · · · · · · · · · · · · · · · · · · · · HCN J=1–0 · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · ·
J=3–2 · · · ± ± · · · ± ± · · · · · · · · · · · · · · · · · · · · · HCO + J=1–0 · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · ·
J=3–2 · · · ± ± · · · ± ± · · · · · · · · · ± · · · · · · ± · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · ± · · · W-knot CO J=1–0 · · · ± ± ± · · · · · · · · · · · · · · · · · · · · · · · · · · · J=2–1 1.5 ± · · · ± · · · ± · · · · · · · · · · · · ± · · · ± · · · J=3–2 4.0 ± ± · · · · · · · · · ± · · · · · · · · · ± · · · ± · · · CO J=1–0 0.02 ± · · · · · · · · · ± · · · · · · · · · · · · · · · · · · · · · · · · J=2–1 · · · ± · · · ± · · · · · · · · · · · · · · · ± · · · ± · · · J=3–2 · · · · · · ± · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · C O J=2–1 · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · ·
HCN J=1–0 · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · ·
J=3–2 · · · ± ± · · · ± · · · · · · · · · · · · · · · · · · ± · · · HCO + J=1–0 · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · ·
J=3–2 · · · ± ± · · · ± · · · · · · · · · · · · ± · · · · · · · · · J=4–3 · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · ·
Total CO J=1–0 · · · ± ± ± · · · · · · · · · · · · · · · · · · · · · · · · · · · J=2–1 2.0 ± · · · ± · · · ± · · · ± · · · · · · ± · · · ± · · · J=3–2 4.0 ± ± · · · · · · · · · ± · · · · · · · · · ± · · · ± · · · CO J=1–0 0.04 ± · · · · · · · · · ± ± · · · · · · · · · · · · · · · · · · · · · J=2–1 · · · ± · · · ± · · · ± · · · · · · · · · ± · · · ± · · · J=3–2 · · · · · · ± ± ± · · · · · · · · · · · · ± · · · ± · · · C O J=2–1 · · · ± · · · · · · ± · · · · · · · · · · · · · · · · · · · · · · · · HCN J=1–0 · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · ·
J=2–1 · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · ·
J=3–2 · · · ± ± · · · ± ± · · · · · · · · · · · · · · · ± · · · HCO + J=1–0 · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · · ·
J=3–2 · · · ± ± · · · ± ± · · · · · · · · · ± · · · · · · ± · · · · · · · · · · · · · · · ± · · · · · · · · · · · · · · · ± · · · a The line ratios were derived by spatially averaging over the respective region from the velocity intergated line ratio maps from Fig. 12. The errors denote thereby thestandard deviation from the averaged values. Please note, that in some cases, the line ratios might vary from those estimated from Tables 4 and 5. See text for a discussion. illimeter observations of NGC 1068 15
Fig. 1.—
Continuum emission of NGC 1068 at λ =1.4 mm (Fig. 1a; black contours), 1.3 mm (Fig. 1a-c; grey scale and grey contours),1.0 mm (Fig. 1b; black contours) and 850 µ m (Fig. 1c; black contours), observed with the SMA and the IRAM PdBI. The white crossdenotes the position of the AGN measured by Gallimore et al. (2004, G04). The contours of the 1.3 mm continuum emission (PdBI)start at 5 σ =4 mJy in steps of 1 σ . a) The contours of the 1.4 mm continuum emission (SMA) start at 5 σ =4 mJy in steps of 1 σ . b) Thecontours of the 1.0 mm continuum emission (SMA, NA) start at 3 σ =1.6 mJy in steps of 1 σ , while the contours of the uniformaly mapped1.0 mm continuum emission run from 3 σ =2.3 mJy in steps of 1 σ . c) The contours of the 850 µ m continuum emission (SMA, NA) start at3 σ =2.4 mJy in steps of 1 σ , while the contours of the uniformaly mapped 850 µ m continuum emission run from 3 σ =2.6 mJy in steps of 1 σ . Fig. 2.—
Velocity integrated line emission of CO(J=3–2) ( a-g ), HCN(J=3–2) ( a ), HCO + (J=3–2) ( b ), HCO + (J=4–3) ( c ), CO(J=2–1)( d ), CO(J=2–1) ( e ), CO(J=3–2) ( f ), C O(J=2–1) ( g ), CO(J=1–0) ( h ) and CO(J=1–0) ( i ) in NGC 1068, observed with the SMAand the IRAM PdBI. Contour levles are: CO(J=3–2) – from 10 σ by 6 σ with 1 σ =4.8 Jy km s − ; HCN(J=3–2) – from 3 σ by 1 σ with1 σ =2.6 Jy km s − ; HCO + (J=3–2) – from 2 σ by 1 σ with 1 σ =3.4 Jy km s − ; HCO + (J=4–3) – from 3 σ by 1 σ with 1 σ =2.4 Jy km s − ; CO(J=2–1) – from 5 σ by 5 σ with 1 σ =1.2 Jy km s − ; CO(J=2–1) – from 3 σ by 1 σ with 1 σ =0.9 Jy km s − ; CO(J=3–2) – from 5 σ by1 σ with 1 σ =3.5 Jy km s − ; C O(J=2–1) – from 3 σ by 1 σ with 1 σ =0.9 Jy km s − ; CO(J=1–0) – from 5 σ by 3 σ with 1 σ =0.4 Jy km s − ; CO(J=1–0) – from 1 σ by 1 σ with 1 σ =0.1 Jy km s − . illimeter observations of NGC 1068 17 Fig. 3.—
Spatially integrated spectrum of different molecular lines in NGC 1068. The single-line Gaussian fit is indicated with a dottedred line (parameters are listed in Table 5) while the multiple Gaussian fit is plotted with a dashed blue line.
Fig. 4.—
Spatially integrated spectrum of the CO(J=1–0), (taken from Schinnerer et al. 2000), CO(J=2–1) and CO(J=3–2) ( leftcolumn ) and HCN(J=3–2) and HCO + (J=4–3) emission over the E-knot ( dotted blue ) and W-knot ( solid red ) component of NGC 1068. illimeter observations of NGC 1068 19 Fig. 5.—
Iso-velocity maps of the different molecular lines observed in NGC 1068. The grey scale correspond to the velocity integratedemission of each line with the same contours as used in Fig. 2. The velocity contours are in steps of 10 km s − around the systemic velocityof NGC 1068. The grey lines indicate the cuts along which the position-velocity diagrams (see Fig. 10) were taken for the respectivemolecules ( CO(J=2–1), CO(J=2–1) & HCN(J=3–2)).
Fig. 6.—
Red- and blueshifted emission of NGC1068. Contours are in steps of a) from 7 σ by 5 σ with 1 σ =2.0 Jy km s − ; b) from 5 σ by5 σ with 1 σ =0.9 Jy km s − ; c) from 4 σ by 1 σ with 1 σ =0.5 Jy km s − ; d) from 3 σ by 1 σ with 1 σ =1.9 Jy km s − ; e) from 3 σ by 1 σ with1 σ =2.3 Jy km s − ; f) from 2 σ by 1 σ with 1 σ =1.4 Jy km s − . illimeter observations of NGC 1068 21 Fig. 7.—
Channel maps of the CO(J=2–1) emission in NGC 1068. Please note that the CO(J=2–1) data were resampled to matchthe spectral resolution of the CO(J=3–2) data and facilitate a comparison, especially with respect to the line-ratio channel map inFig. 13. Contour spacing is in steps of 5 σ =37.4 mJy beam − . We use a spectral resolution of ∼ LSR =1137 km s − of NGC 1068. Fig. 8.—
Channel maps of the CO(J=3–2) emission in NGC 1068. We use a spectral resolution of ∼ − and the original spatialresolution from the observations. Contour spacing is in steps of 5 σ =400 mJy beam − . The zero channel corresponds to v LSR =1137 km s − of NGC 1068. illimeter observations of NGC 1068 23 Fig. 9.—
Channel maps of the CO(J=2–1) emission in NGC 1068. We use spectral resolution of ∼ − . Contour spacing is insteps of 2 σ =11 mJy beam − . The zero channel corresponds to v LSR =1137 km s − of NGC 1068. Fig. 10.—
Position-velocity diagram of NGC 1068 for the CO(J=2–1) (grey scale), CO(J=3–2) (black contours, top figure) andHCN(J=3–2) (black contours; bottom figure) emission. illimeter observations of NGC 1068 25
Fig. 11.—
Parametrisation of the kinematics in the CND of NGC 1068 for the CO(J=2–1) emission (filled black data points) takenalong the slits at different position angles (at 0, 30, 60, 90, 120, and 150 ◦ ). σ is the average line-of-sight velocity dispersion, ǫ the rmsvariation of the velocity from point to point and ∆ the rotational velocity along each slit (see text in Section5.2 for a more detaileddescription). VNRot = violent non-rotators, CNRot = calm non-rotators and Rot = rotators. The filled grey data points were determinedfrom the model discussed in Section 5.2. Fig. 12.—
Velocity integrated molecular line intensity ratios for NGC 1068 above a ≥ σ threshold. The white cross marks the po-sition of the mm continuum emission that is associated with the AGN. The different line ratios are: a) CO(J=3–2)-to- CO(J=1–0); b) CO(J=3–2)-to- CO(J=2–1), c) CO(J=2–1)-to- CO(J=1–0), d) CO(J=3–2)-to- CO(J=3–2), e) HCN(J=3–2)-to- CO(J=3–2), f) HCO + (J=3–2)-to- CO(J=3–2), g) HCN(J=3–2)-to-HCO + (J=3–2), h) HCO + (J=3–2)-to-HCO + (J=4–3), i) CO(J=2–1)-to- CO(J=2–1), j) CO(J=2–1)-to-HCN(J=3–2), k) CO(J=2–1)-to-HCO + (J=3–2), l) CO(J=2–1)-to-C O(J=2–1), m) CO(J=1–0)-to- CO(J=2–1), n) CO(J=1–0)-to- CO(J=1–0), o) CO(J=2–1)-to- CO(J=3–2), p) CO(J=1–0)-to- CO(J=1–0) illimeter observations of NGC 1068 27
Fig. 13.— CO(J=3–2)-to- CO(J=2–1) line ratio above a 4 σ -threshold for each line. Fig. 14.—
Spatially averaged line ratio for the eastern and western knot as function of velocity, derived from Fig. 13. The error bars denotethe variance of each averaged value, which does not exceed 50% in most cases. The missing values for velocities −
120 and −
200 km s − and +100 to +140 km s − are due to the lack of emisison in the respective knot. The dashed lines represent the median of the line ratiosfor the two knots. illimeter observations of NGC 1068 29 Fig. 15.—
Spectral energy distribution of the continuum emission in NGC 1068, based on data from this paper, Krips et al. (2006) ( blackcrosses ), and H¨onig et al. (2008) ( grey crosses ). The dotted line respresents the model for the radio continuum emission (either electronscattered synchrotron emission ( a ), synchrotron emission ( b ), and free-free absorption ( c )), the dashed line represents the model for theIR data (a two-temperature grey body a-c ), and the solid line represents the composite of both ( a-c ). The data observed at an angularresolution below 1 ′′ are additionally marked with a circle. Fig. 16.—
Velocity channel maps of the CO model compared to the CO(2–1) emission. illimeter observations of NGC 1068 31
Fig. 17.—
Spectrum of the CO model compared to the CO(2–1) emission.
Fig. 18.—
Moment maps of the CO model compared to the CO(2–1) emission. The velocities are plotted in steps of 20 km s − for theMoment-1 and Moment-2 maps. illimeter observations of NGC 1068 33 Fig. 19.—
Blue- and Redshifted emission of the CO model compared to the CO(2–1) emission.
Fig. 20.—
Position-velocity diagram of the CO model compared to the CO(2–1) emission.
Fig. 21.— χ -fit results obtained from the RADEX simulations of the excitation conditions of the molecular gas. Shown are four different[ CO]/[ CO] abundance ratios (=10,26,52,110) for three different CO column densities respectively. The middle panel shows the best χ -fit found for each abundance ratio. illimeter observations of NGC 1068 35 Fig. 22.— χ -fit results obtained from the RADEX simulations of the excitation conditions of the molecular gas. Shown are four different[HCN]/[HCO + ] abundance ratios around the standard galactic value of [HCN]/[HCO + ] ≃
10 for three different HCN column densitiesrespectively. The middle panel shows the best χ2