VALES: III. The calibration between the dust continuum and interstellar gas content of star-forming galaxies
T. M. Hughes, E. Ibar, V. Villanueva, M. Aravena, M. Baes, N. Bourne, A. Cooray, L. J. M. Davies, S. Driver, L. Dunne, S. Dye, S. Eales, C. Furlanetto, R. Herrera-Camus, R. J. Ivison, E. van Kampen, M. A. Lara-López, S. Maddox, M. J. Micha?owski, I. Oteo, D. Smith, M. W. L. Smith, E. Valiante, P. van der Werf, S. Viaene, Y. Q. Xue
MMNRAS , 1–5 (2017) Preprint 15 October 2018 Compiled using MNRAS L A TEX style file v3.0
VALES: III. The calibration between the dust continuumand interstellar gas content of star-forming galaxies
T. M. Hughes (cid:63) , E. Ibar , V. Villanueva , M. Aravena , M. Baes , N. Bourne ,A. Cooray , L. J. M. Davies , S. Driver , , L. Dunne , , S. Dye , S. Eales ,C. Furlanetto , , R. Herrera-Camus , R. J. Ivison , , E. van Kampen ,M. A. Lara-L´opez , S. Maddox , , M. J. Micha(cid:32)lowski , I. Oteo , , D. Smith ,M. W. L. Smith , E. Valiante , P. van der Werf , S. Viaene , , Y. Q. Xue Instituto de F´ısica y Astronom´ıa, Universidad de Valpara´ıso, Avda. Gran Breta˜na 1111, Valpara´ıso, Chile N´ucleo de Astronom´ıa, Facultad de Ingenier´ıa, Universidad Diego Portales, Av. Ej´ercito 441, Santiago, Chile Sterrenkundig Observatorium, Universiteit Gent, Krijgslaan 281-S9, Gent 9000, Belgium Institute for Astronomy, University of Edinburgh, Royal Observatory, Edinburgh EH9 3HJ, UK Department of Physics and Astronomy, University of California, Irvine, CA 92697, USA International Centre for Radio Astronomy Research, University of Western Australia, Crawley WA 6009, Australia School of Physics and Astronomy, University of St Andrews, North Haugh, St Andrews KY16 9SS, UK School of Physics and Astronomy, Cardiff University, The Parade, Cardiff CF24 3AA, UK School of Physics and Astronomy, University of Nottingham, University Park, Nottingham NG7 2RD, UK CAPES Foundation, Ministry of Education of Brazil, Bras´ılia/DF 70040-020, Brazil Max-Planck-Institut f¨ur extraterrestrische Physik, Giessenbachstraße, 85748 Garching, Germany European Southern Observatory, Karl-Schwarzschild-Strasse 2, 85748, Garching, Germany Instituto de Astronom´ıa, Universidad Nacional Autonoma de M´exico, A.P. 70-264, 04510 M´exico, D.F., M´exico Centre for Astrophysics Research, University of Hertfordshire, Hatfield, Hertfordshire AL10 9AB, UK Leiden Observatory, Leiden University, PO Box 9513, 2300 RA Leiden, The Netherlands CAS Key Laboratory for Researches in Galaxies and Cosmology, Center for Astrophysics, Department of Astronomy,University of Science and Technology of China, Chinese Academy of Sciences, Hefei, Anhui 230026, China
Accepted 2017 February 23. Received 2017 February 22; in original form 2017 February 18.
ABSTRACT
We present the calibration between the dust continuum luminosity and interstellargas content obtained from the Valpara´ıso ALMA Line Emission Survey (VALES)sample of 67 main-sequence star-forming galaxies at . < z < . . We use CO(1–0)observations from the Atacama Large Millimetre/submillimetre Array (ALMA) to tracethe molecular gas mass, M H , and estimate the rest-frame monochromatic luminosityat 850 µ m, L ν , by extrapolating the dust continuum from MAGPHYS modellingof the far-ultraviolet to submillimetre spectral energy distribution sampled by theGalaxy And Mass Assembly (GAMA) survey. Adopting α CO = 6.5 (K km s − pc ) − ,the average ratio of L ν / M H = (6.4 ± × erg s − Hz − M − (cid:12) , inexcellent agreement with literature values. We obtain a linear fitof log (cid:0) M H / M (cid:12) (cid:1) = ( . ± . ) log ( L ν / erg s − Hz − ) − ( . ± . ) . We providerelations between L ν , M H and M ISM when combining the VALES and literaturesamples, and adopting a Galactic α CO value. Key words: galaxies: ISM – ISM: lines and bands – submillimetre: galaxies
Disentangling the physical processes contributing to thedecline in the overall cosmic star formation rate density ( ρ SFR ) (cid:63) E-mail: [email protected] since the observed peak at z ∼ © a r X i v : . [ a s t r o - ph . GA ] F e b T. M. Hughes et al. gas component (see e.g. Carilli & Walter 2013, and referencestherein). However, the linear relationship between the 21-cm line brightness and the column density of gas breaks foroptically thick gas (Braun et al. 2009). Furthermore, the ‘ α CO factor’, the constant of proportionality between the mass ofthe molecular phase and the CO line emission, typically fromthe J=1–0 or J=2–1 line, is highly uncertain with a possibledependence on gas-phase metallicity (Wilson 1995; Israel2005), galaxy kinematics, and excitation conditions (Solomon& Vanden Bout 2005). The standard CO/21-cm method mayalso overlook a significant fraction of lower column densitymolecular gas which is not CO bright and so traced byneither line (Abdo et al. 2010; Planck Collaboration et al.2011a). Technologically, it remains impossible to detect theH i line from galaxies at z > Herschel
SpaceObservatory (Pilbratt et al. 2010) with the PhotodetectorArray Camera and Spectrometer (PACS; Poglitsch et al.2010) and the Spectral and Photometric Imaging REceiver(SPIRE; Griffin et al. 2010) were jointly capable of detectingthe far-infrared (FIR) to submillimetre (submm) continuumemission originating from the dust component in six wave-bands (70 to 500 µ m) with significantly higher sensitivity andangular resolution than previous FIR/submm experiments,making it possible to derive a calibration between thedust emission and the ISM mass, M ISM (Eales et al. 2012;Magdis et al. 2013), though the calibration is dependenton an accurate knowledge of the dust temperature.Most recently, Scoville et al. (2016) used a calibrationbetween the dust continuum at λ = µ m and the moleculargas content to infer the properties of higher redshift ( z ≤ )galaxies. The empirical calibration was obtained considering Planck observations of the Milky Way (Planck Collaborationet al. 2011b,c) and samples of low redshift star-forming galax-ies (Dale et al. 2005; Clements et al. 2010), ultraluminousinfrared galaxies (ULIRGs), and higher redshift ( z = µ m luminosity per unit ISM mass, the calibration basedon the sample of 70 star-forming galaxies, SMGs and ULIRGs,gave L ν / M H = (6.7 ± × erg s − Hz − M − (cid:12) . Byapplying their calibration to ALMA observations of galaxiesin three redshift bins up to z = . , Scoville et al. concludethat starburst galaxies above the main sequence are largelythe result of having greatly increased gas masses rather thanan increased efficiency of converting gas to stars, with star-forming galaxies at z > exhibiting ∼ ∼ × ) at estimating theISM mass than molecular line observations and applicableto more readily obtainable continuum observations at higherredshift, the method assumes a solar metallicity and somay not apply to lower mass, metal-poor galaxies at higher VALES sampleSF galsULIRGSSMGs L CO = 3 . × − L ν L CO = 2 . × − L ν L CO = 3 . × − L ν Scoville et al.:VALES:Combined: L ν [erg s − Hz − ] L C O [ K k m s − p c − ] Figure 1.
The correlation between L ν and L (cid:48) CO found forgalaxies in our VALES sample observed with ALMA (Villanuevaet al. 2017; Hughes et al. 2016). For galaxies with CO detections(blue circles), we show the average ratio (black dashed line) andcompare to the mean value (black dotted line) found for the low- z samples of star-forming galaxies (SF; squares), ultraluminousinfrared galaxies (ULIRGs; triangles), and submillimetre galaxies(SMGs; diamonds) studied in Scoville et al. (2016). The averageof these combined samples is superimposed (dashed-dotted line). redshifts. It is crucial to test the robustness of this calibrationto ensure that any evolution with redshift is in fact physical.In this Letter, we present the calibration between thedust continuum and molecular gas content derived frommeasurements of L ν , M H and M ISM for an expanded, ho-mogeneous sample of 67 main-sequence star-forming galaxiesat . < z < . in the Valpara´ıso ALMA Line EmissionSurvey (VALES; Villanueva et al. 2017; Hughes et al. 2016),based on a combination of Band 3 CO(1–0) observationstaken with the Atacama Large Millimetre/submillimetreArray (ALMA) and FUV-submm photometry from theGAMA survey (Driver et al. 2016; Wright et al. 2016).We adopt a Λ CDM cosmology with H = km s − Mpc − , Ω M = . and Ω Λ = . . Our sample of galaxies was originally drawn from the
Herschel
Astrophysical Terahertz Large Area Survey (Ealeset al. 2010; Valiante et al. 2016; Bourne et al. 2016), a
Herschel programme capable of providing a sufficient numberof far-IR bright galaxies over ∼
600 deg with a wealth ofhigh-quality ancillary data. From the three equatorial fieldsspanning ∼
160 deg covered by H -ATLAS, galaxies wereselected based on the following criteria: (1) a flux of S µ m >
150 mJy; (2) no neighbours with S µ m > mJy ( σ )within 2 arcmin from their centroids; (3) an unambiguousidentification ( reliability > MNRAS , 1–5 (2017)
ALES: III. Dust–gas calibration L3 r -band radius < (cid:48)(cid:48) , i.e. smallerthan the PACS spectroscopic field of view; (5) high-qualityspectroscopic redshifts ( zqual >
3) from the Galaxy andMass Assembly survey (GAMA; Liske et al. 2015); and (6) aredshift between 0.02 < z < × M (cid:12) ,SFRs between 0.6 and 100 M (cid:12) yr − , and metallicities of . < + log (O/H) < . (see Villanueva et al. 2017). We exploit our VALES observations targeting the CO(1–0)line in Band 3 for 67 galaxies obtained during cycle-1 and-2. Villanueva et al. (2017) present the observations, datareduction and a detailed characterisation for the completesample. All observations were reduced homogeneously withinthe
Common Astronomy Software Applications (CASA;McMullin et al. 2007) using a common pipeline, developedfrom standard pipelines, for calibration, concatenation andimaging, with standard bandpass, flux and phase calibrators.Velocity-integrated CO(1–0) flux densities, S CO ∆ v , in unitsof Jy km s − were obtained by collapsing the cleaned,primary-beam-corrected data cubes between ν obs − ν FWHM and ν obs + ν FWHM , and fitting these cubes with a Gaussian.We detect > % (49 of 67) of the targets with a > σ peakline detection. We estimate upper limits as 5 × the measuredRMS from the collapsed cubes set at 100 km s − spectralresolution and adopting ν FWHM = 250 km s − . All of our galaxies are present in the GAMA PanchromaticData Release (Driver et al. 2016) that provides imaging forover 230 deg with photometry in 21 bands extending fromthe far-ultraviolet to far-infrared from numerous facilities,currently including: GALaxy Evolution eXplorer (GALEX),Sloan Digital Sky Survey (SDSS), Visible and InfraredTelescope for Astronomy (VISTA), Wide-field InfraredSurvey Explorer (WISE), and Herschel . These data areprocessed to a common astrometric solution from whichhomogeneous photometry is derived for ∼
221 373 galaxieswith r < µ m isavailable for each galaxy. L ν – M ISM
CALIBRATION3.1 Estimating the dust continuum luminosity
In the absence of measurements of the dust continuumat 850 µ m, we adopt an estimate of the L ν basedon an extrapolation of the modelled SED. Our primary http://casa.nrao.edu/index.shtml http://cutout.icrar.org/panchromaticDR.php approach to estimate L ν exploits the FUV–FIR/submm H -ATLAS/GAMA photometry available for all our galaxiesmodelled with the Bayesian SED fitting code, MAGPHYS(da Cunha et al. 2008). The code fits the panchromaticSED, giving special consideration to the dust–energy balance,from a library of optical and infrared SEDs derived froma generalised multi-component model of a galaxy. TheFIR/submm dust emission is modelled with five modifiedblack bodies, of which two components have variable temper-atures representing thermalised cold and warm dust and theother three components represent hot dust at 130, 250 and850 K. Two geometries describe the dust distribution: birthclouds of new stars contain only warm and hot circumstellardust, whereas all five dust components may contribute to thedust in the diffuse ISM. As our focus is solely on estimatingthe rest frame 850 µ m continuum luminosity, we refer thereader to Driver et al. (2017) for details of the completeanalysis of the MAGPHYS modelling of all the GAMASEDs, yet note that Villanueva et al. (2017) demonstrate howthe stellar masses, IR luminosities, L IR , and SFRs derivedfrom MAGPHYS are consistent within the uncertainties toempirical estimates found in Ibar et al. (2015).Using the best-fit SEDs, we calculate the median modelflux between 800 and 900 µ m, S ν , and convert this flux– that ranges from 1 to 15 mJy – into a monochromaticrest-frame luminosity, L ν , in units of erg s − Hz − , via L ν = . × S ν ( Jy ) ( + z ) − D L K erg s − Hz − (1)where D L is the luminosity distance in Mpc, and K is the K -correction given by Eqn. 2 in Dunne et al. (2011), followingtheir exact methodology . In addition to FUV–FIR/submmSED modelling via MAGPHYS, we also examine the resultsof fitting the five H -ATLAS PACS/SPIRE photometricbands with a one-component modified blackbody as originallypresented by Hildebrand (1983), assuming a power-law dustemissivity and either keeping the spectral index β as a freeparameter or fixing the value at 1.8 (e.g. Galametz et al. 2012).In both cases, our best-fit model fluxes are consistent andproduce results that support the conclusions reached withthe MAGPHYS SED fitting results. We then compute theuncertainty in L ν from the standard deviation of the threeluminosity values we obtain from modelling the SEDs withMAGPHYS and the two fits with one-component modifiedblack bodies adopting variable and fixed spectral indices. From our velocity-integrated CO(1–0) flux densities, S CO ∆ v ,in units of Jy km s − , we calculate the CO line luminosity, L (cid:48) CO , in units of K km s − pc following Eqn. 3 of Solomon &Vanden Bout 2005, given as L (cid:48) CO = . × S CO ∆ v ν − D ( + z ) − , (2)where ν obs is the observed frequency of the emission line inGHz. The values for L (cid:48) CO are in the range of ( . − . ) × K km s − pc , with an average value of ( . ± . ) × K km s − pc . The CO line luminosity can then be converted The mean scatter between L ν calculated via the methodpresented in Appendix A of Scoville et al. (2016) and that usedhere is ±000
In the absence of measurements of the dust continuumat 850 µ m, we adopt an estimate of the L ν basedon an extrapolation of the modelled SED. Our primary http://casa.nrao.edu/index.shtml http://cutout.icrar.org/panchromaticDR.php approach to estimate L ν exploits the FUV–FIR/submm H -ATLAS/GAMA photometry available for all our galaxiesmodelled with the Bayesian SED fitting code, MAGPHYS(da Cunha et al. 2008). The code fits the panchromaticSED, giving special consideration to the dust–energy balance,from a library of optical and infrared SEDs derived froma generalised multi-component model of a galaxy. TheFIR/submm dust emission is modelled with five modifiedblack bodies, of which two components have variable temper-atures representing thermalised cold and warm dust and theother three components represent hot dust at 130, 250 and850 K. Two geometries describe the dust distribution: birthclouds of new stars contain only warm and hot circumstellardust, whereas all five dust components may contribute to thedust in the diffuse ISM. As our focus is solely on estimatingthe rest frame 850 µ m continuum luminosity, we refer thereader to Driver et al. (2017) for details of the completeanalysis of the MAGPHYS modelling of all the GAMASEDs, yet note that Villanueva et al. (2017) demonstrate howthe stellar masses, IR luminosities, L IR , and SFRs derivedfrom MAGPHYS are consistent within the uncertainties toempirical estimates found in Ibar et al. (2015).Using the best-fit SEDs, we calculate the median modelflux between 800 and 900 µ m, S ν , and convert this flux– that ranges from 1 to 15 mJy – into a monochromaticrest-frame luminosity, L ν , in units of erg s − Hz − , via L ν = . × S ν ( Jy ) ( + z ) − D L K erg s − Hz − (1)where D L is the luminosity distance in Mpc, and K is the K -correction given by Eqn. 2 in Dunne et al. (2011), followingtheir exact methodology . In addition to FUV–FIR/submmSED modelling via MAGPHYS, we also examine the resultsof fitting the five H -ATLAS PACS/SPIRE photometricbands with a one-component modified blackbody as originallypresented by Hildebrand (1983), assuming a power-law dustemissivity and either keeping the spectral index β as a freeparameter or fixing the value at 1.8 (e.g. Galametz et al. 2012).In both cases, our best-fit model fluxes are consistent andproduce results that support the conclusions reached withthe MAGPHYS SED fitting results. We then compute theuncertainty in L ν from the standard deviation of the threeluminosity values we obtain from modelling the SEDs withMAGPHYS and the two fits with one-component modifiedblack bodies adopting variable and fixed spectral indices. From our velocity-integrated CO(1–0) flux densities, S CO ∆ v ,in units of Jy km s − , we calculate the CO line luminosity, L (cid:48) CO , in units of K km s − pc following Eqn. 3 of Solomon &Vanden Bout 2005, given as L (cid:48) CO = . × S CO ∆ v ν − D ( + z ) − , (2)where ν obs is the observed frequency of the emission line inGHz. The values for L (cid:48) CO are in the range of ( . − . ) × K km s − pc , with an average value of ( . ± . ) × K km s − pc . The CO line luminosity can then be converted The mean scatter between L ν calculated via the methodpresented in Appendix A of Scoville et al. (2016) and that usedhere is ±000 , 1–5 (2017) T. M. Hughes et al. VALES sampleSF galsULIRGSSMGs L ν [erg s − Hz − ] L ν / M H [ e r g s − H z − M − (cid:12) ] Figure 2.
The ratio of L ν to M H found for galaxies in ourVALES sample observed with ALMA (Villanueva et al. 2017;Hughes et al. 2016). From galaxies with CO detections (bluecircles), we show the average ratio (black dashed line) with the ± σ range (light grey region) and the best linear fit (black solidline) with ± σ confidence limits (dark grey region). The low- z samples of SF galaxies (squares), ULIRGs (triangles), and SMGs(diamonds) are shown together with the mean value (black dottedline), taken from Fig. 1 of Scoville et al. (2016). into the molecular gas mass (including the mass of He), M H , by assuming an α CO conversion factor (see Eqn. 5 inSolomon & Vanden Bout 2005). Our VALES galaxies havehigh stellar masses ( ≥ M (cid:12) ), thus avoiding metal-poorsystems in which the dust-to-gas abundance ratio is expectedto decrease nor where significant molecular gas exists withoutCO emission (see e.g. Bolatto et al. 2013).We first exclude from our analysis the merger/interactingsystems identified using a K-band-based morphologicalclassification as outlined in Villanueva et al. (2017). Tofacilitate a direct comparison with the results of Scovilleet al., we primarily adopt α CO = 6.5 (K km s − pc ) − forthe bulge- and disk-dominated galaxies with normal starformation. In our 43 normal galaxies with detected COemission, we derive M H values in the range of log M H / M (cid:12) = . − . with an median of . ± . . Finally, to estimatethe atomic hydrogen content, M H i , we use the H i –colourscaling relation given by Eqn. 4 of Zhang et al. (2009) withthe g − r colour and i –band surface brightness availablefrom the GAMA photometry. The H i mass ranges between log M H i / M (cid:12) = . − . with an average of . ± . andtypical errors of ± M ISM = M H i + M H using standard error propagation. L ν / M ISM calibration
Bringing these measurements together, we now examinethe calibrations between the dust continuum and gascontent found for the galaxies in our VALES sampleconsidering only those galaxies with CO line detections.The VALES sample exhibits a mean ratio S CO ∆ v / S ν of 1081 ±
265 km s − , corrresponding to a mean L (cid:48) CO to L ν ratio of (2.91 ± × − in units of the luminosity ratiodimensions (see Fig. 1), which is in agreement with thatfound (3.02 × − ) for the three galaxy samples analysedin Scoville et al. (2016). In particular, the VALES galax-ies have properties more akin to the low- z normal star-forming galaxies and ULIRGs than the SMG sample. Afterconverting the CO luminosity into molecular gas mass, wefind average ratios of S ν / M H = (6.9 ± × − Jy M − (cid:12) and L ν / M H = (6.4 ± × erg s − Hz − M − (cid:12) , also inexcellent agreement with the mean values found by Scovilleet al. (2016) and with near-matching scatter.Although a constant ratio is appropriate to describethe average properties of the both the Scoville et al. andVALES samples across the luminosity range, there is a veryminor trend that galaxies with L ν > erg s − Hz − tend to lie on or above the average ratio (see Fig. 2).Galaxies with L ν fainter than this luminosity have slightlylower ratios than the average. Our results suggest thatadopting a constant L ν / M H ratio to estimate the ISMmass would underestimate M H in galaxies where L ν > erg s − Hz − (and vice versa) and so a linear fit (inlogarithmic space) may be more appropriate for the galaxies.For the VALES sample, we obtain log M H = ( . ± . ) log L ν − ( . ± . ) , (3) log M ISM = ( . ± . ) log L ν − ( . ± . ) , (4)in which M and L ν have units of M (cid:12) and erg s − Hz − ,respectively. This relation is valid between × < L ν < × erg s − Hz − for normal main-sequencestar-forming galaxies and is based on the assumption that α CO = 6.5 (K km s − pc ) − . Furthermore, we consider thecalibrations we obtain from combining the 70 galaxies ofScoville et al. (2016) with the 43 CO-detected star-forminggalaxies in our VALES sample. The L ν – M H calibrationfor this combined sample of 113 objects is then log M H = ( . ± . ) log L ν − ( . ± . ) (5)with a scatter of ∼ α CO and is not included in our error calculations. Wenote that although Scoville et al. (2014) include the H i masscontribution to M ISM (estimated as 50% of the H mass),Scoville et al. (2016) only consider the H mass component,therefore we do not include a L ν – M ISM calibration for thecombined sample. We summarise these best-fit relations andthe corresponding correlation coefficients in Table 1, in whichwe also present the relations we obtain when adopting theGalactic value of α CO = 4.6 (K km s − pc ) − for our sample. We have reported an updated calibration between the dustcontinuum and molecular gas content for an expanded sampleof 67 main-sequence star-forming galaxies at . < z < . drawn from the H -ATLAS, using gas mass measurementsfrom ALMA Band-3 CO(1–0) observations and estimates ofthe monochromatic luminosity at 850 µ m (rest-frame), L ν ,via an extrapolation of the dust continuum from MAGPHYSmodelling of the FUV to FIR/submm SED observed by theGAMA survey. Although we confirm an average L ν / M H MNRAS , 1–5 (2017)
ALES: III. Dust–gas calibration L5 Table 1.
The best-fit relations between the dust continuum lumi-nosity at 850 µ m (in units of erg s − Hz − ) and various parametersconsidered in this work, expressed as log y = m log L ν + c ,with the corresponding 1 σ errors. We also state the scatter in dex( σ ), Spearman ( ρ S ), and Pearson ( ρ P ) coefficients, where the valueshave probabilities P( ρ P ) and P( ρ S ) of > N = or 113. a Combinedsample of 113 galaxies in VALES and Scoville et al. (2016). y m c σ ρ P ρ S VALES sample only; α CO ≡ − pc ) − L (cid:48) CO ± ± M H / M (cid:12) ± ± M ISM / M (cid:12) ± ± a ; α CO ≡ − pc ) − L (cid:48) CO ± ± M H / M (cid:12) ± ± α CO ≡ − pc ) − L (cid:48) CO ± ± M H / M (cid:12) ± ± M ISM / M (cid:12) ± ± a ; α CO ≡ − pc ) − L (cid:48) CO ± ± M H / M (cid:12) ± ± ratio in close agreement with literature values (see Scovilleet al. 2016, and references therein), the linear fit givenby Eqn. 4 alleviates the issue that the ISM mass may beoverestimated for galaxies with lower continuum luminosities.Whilst we recommend using this best-fit calibration ratherthan the constant calibration for estimating the gas contentfrom dust continuum observations of main-sequence galaxiesat high redshift, we stress that the largest uncertainty in thiswork remains in the α CO factor.Whichever value of α CO we choose to adopt, usingthese Galactic-type values assumes that the CO emissioncomes from viriliazied molecular clouds bound by self-gravity.However, it remains possible that the CO line emission mayactually trace material bound by the total potential of thegalactic center consisting of a mass of stars and dense gasclumps equal to the dynamical mass, M dyn , and a diffuseinterclump medium the CO emitting gas of mass M gas . Inthis case, M gas = M dyn ( α CO L (cid:48) CO ) (see Eqn. 6 in Solomon &Vanden Bout 2005), meaning the usual relation of α CO willbe changed if a fraction of the CO emission in our galaxiesoriginates from an intercloud medium bound by the galaxypotential. Unfortunately, we currently have too few normaldisk-dominated galaxies with spatially-resolved CO emission(7 sources in total) to robustly identify whether such a changeto the α CO factor is warranted in our sample. We would alsorequire more accurate estimates covering a greater dynamicrange of metallicity in order to test the effect of the α CO dependency on metallicity. In future VALES studies, weaim to use ALMA and MUSE to further investigate therobustness of the calibration between the dust continuumand molecular gas content with an α CO constrained by 3Dkinematical modelling for a larger sample of resolved galaxies. ACKNOWLEDGMENTS
TMH and EI acknowledge CONICYT/ALMA funding Program inAstronomy/PCI Project N ◦ :31140020. MA acknowledges partialsupport from FONDECYT through grant 1140099. DR acknowl-edges support from the National Science Foundation under grantnumber AST-1614213 to Cornell University. LD, SJM and RJIacknowledge support from European Research Council AdvancedInvestigator Grant COSMICISM, 321302; SJM and LD are alsosupported by the European Research Council Consolidator Grant CosmicDust (ERC-2014-CoG-647939). YQX acknowledges sup-port from grants NSFC-11473026 and NSFC-11421303. This paperuses the following ALMA data: ADS/JAO.ALMA
REFERENCES
Abazajian K. N., et al., 2009, ApJS, 182, 543Abdo A. A., et al., 2010, ApJ, 710, 133Bolatto A. D., Wolfire M., Leroy A. K., 2013, ARA&A, 51, 207Boselli A., Lequeux J., Gavazzi G., 2002, A&A, 384, 33Bourne N., et al., 2016, MNRAS, 462, 1714Braun R., et al., 2009, ApJ, 695, 937Carilli C. L., Walter F., 2013, ARA&A, 51, 105Clements D. L., Dunne L., Eales S., 2010, MNRAS, 403, 274Dale D. A., et al., 2005, ApJ, 633, 857Driver S. P., et al., 2016, MNRAS, 455, 3911Driver S. P., Andrews S. K., et al. 2017, MNRAS, in preparation.Dunne L., et al., 2000, MNRAS, 315, 115Dunne L., et al., 2011, MNRAS, 417, 1510Eales S., et al., 2010, PASP, 122, 499Eales S., et al., 2012, ApJ, 761, 168Galametz M., et al., 2012, MNRAS, 425, 763Griffin M. J., et al., 2010, A&A, 518, L3Hildebrand R. H., 1983, QJRAS, 24, 267Hughes T. M., et al., 2016, preprint, ( arXiv:1611.05867 )Ibar E., et al., 2015, MNRAS, 449, 2498Israel F. P., 2005, A&A, 438, 855Liske J., et al., 2015, MNRAS, 452, 2087Madau P., Dickinson M., 2014, ARA&A, 52, 415Magdis G. E., et al., 2013, A&A, 558, A136McMullin J. P., Waters B., Schiebel D., Young W., Golap K., 2007,in Shaw R. A., Hill F., Bell D. J., eds, Astronomical Societyof the Pacific Conference Series Vol. 376, Astronomical DataAnalysis Software and Systems XVI. p. 127Pilbratt G. L., et al., 2010, A&A, 518, L1Planck Collaboration et al., 2011a, A&A, 536, A19Planck Collaboration et al., 2011b, A&A, 536, A21Planck Collaboration et al., 2011c, A&A, 536, A25Poglitsch A., et al., 2010, A&A, 518, L2Scoville N., et al., 2014, ApJ, 783, 84Scoville N., et al., 2016, ApJ, 820, 83Scoville N., et al., 2017, preprint, ( arXiv:1702.04729 )Solomon P. M., Vanden Bout P. A., 2005, ARA&A, 43, 677Valiante E., et al., 2016, MNRAS, 462, 3146Villanueva V., Ibar E., Hughes T. M., et al. 2017, in preparation.Wilson C. D., 1995, ApJ, 448, L97Wright A. H., et al., 2016, MNRAS, 460, 765Zhang W., et al. 2009, MNRAS, 397, 1243da Cunha E., Charlot S., Elbaz D., 2008, MNRAS, 388, 1595This paper has been typeset from a TEX/L A TEX file prepared bythe author.MNRAS000