Planck intermediate results. XXV. The Andromeda Galaxy as seen by Planck

24 Pages·2017·5.29 MB·English

Planck intermediate results. XXV. The Andromeda
Galaxy as seen by Planck
P. A. R. Ade, N. Aghanim, M. Arnaud, M. Ashdown, J. Aumont, C.
Baccigalupi, A. J. Banday, R. B. Barreiro, N. Bartolo, E. Battaner, et al.
To cite this version:
P. A. R. Ade, N. Aghanim, M. Arnaud, M. Ashdown, J. Aumont, et al.. Planck intermediate results.
XXV. The Andromeda Galaxy as seen by Planck. Astronomy and Astrophysics - A&A, 2015, 582,
pp.A28. ￿10.1051/0004-6361/201424643￿. ￿cea-01383743￿
HAL Id: cea-01383743
https://hal-cea.archives-ouvertes.fr/cea-01383743
Submitted on 19 Oct 2016
HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est
archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents
entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non,
lished or not. The documents may come from émanant des établissements d’enseignement et de
teaching and research institutions in France or recherche français ou étrangers, des laboratoires
abroad, or from public or private research centers. publics ou privés.
A&A582,A28(2015) Astronomy
DOI:10.1051/0004-6361/201424643 &
(cid:13)c ESO2015 Astrophysics
Planck intermediate results
XXV. The Andromeda galaxy as seen by Planck
PlanckCollaboration:P.A.R.Ade82,N.Aghanim55,M.Arnaud69,M.Ashdown65,6,J.Aumont55,C.Baccigalupi81,
A.J.Banday90,10,R.B.Barreiro61,N.Bartolo28,62,E.Battaner91,92,R.Battye64,K.Benabed56,89,G.J.Bendo88,
A.Benoit-Lévy22,56,89,J.-P.Bernard90,10,M.Bersanelli31,48,P.Bielewicz78,10,81,A.Bonaldi64,L.Bonavera61,
J.R.Bond9,J.Borrill13,84,F.R.Bouchet56,83,C.Burigana47,29,49,R.C.Butler47,E.Calabrese87,J.-F.Cardoso70,1,56,
A.Catalano71,68,A.Chamballu69,14,55,R.-R.Chary53,X.Chen53,H.C.Chiang25,7,P.R.Christensen79,34,
D.L.Clements52,L.P.L.Colombo21,63,C.Combet71,F.Couchot67,A.Coulais68,B.P.Crill63,11,A.Curto61,6,65,
F.Cuttaia47,L.Danese81,R.D.Davies64,R.J.Davis64,P.deBernardis30,A.deRosa47,G.deZotti44,81,
J.Delabrouille1,C.Dickinson64,J.M.Diego61,H.Dole55,54,S.Donzelli48,O.Doré63,11,M.Douspis55,A.Ducout56,52,
X.Dupac36,G.Efstathiou58,F.Elsner22,56,89,T.A.Enßlin75,H.K.Eriksen59,F.Finelli47,49,O.Forni90,10,M.Frailis46,
A.A.Fraisse25,E.Franceschi47,A.Frejsel79,S.Galeotta46,K.Ganga1,M.Giard90,10,Y.Giraud-Héraud1,
E.Gjerløw59,J.González-Nuevo18,61,K.M.Górski63,93,A.Gregorio32,46,51,A.Gruppuso47,F.K.Hansen59,
D.Hanson76,63,9,D.L.Harrison58,65,S.Henrot-Versillé67,C.Hernández-Monteagudo12,75,D.Herranz61,
S.R.Hildebrandt63,11,E.Hivon56,89,M.Hobson6,W.A.Holmes63,A.Hornstrup15,W.Hovest75,
K.M.Huffenberger23,G.Hurier55,F.P.Israel86,A.H.Jaffe52,T.R.Jaffe90,10,W.C.Jones25,M.Juvela24,
E.Keihänen24,R.Keskitalo13,T.S.Kisner73,R.Kneissl35,8,J.Knoche75,M.Kunz16,55,3,H.Kurki-Suonio24,42,
G.Lagache5,55,A.Lähteenmäki2,42,J.-M.Lamarre68,A.Lasenby6,65,M.Lattanzi29,C.R.Lawrence63,R.Leonardi36,
F.Levrier68,M.Liguori28,62,P.B.Lilje59,M.Linden-Vørnle15,M.López-Caniego36,61,P.M.Lubin26,
J.F.Macías-Pérez71,S.Madden69,B.Maffei64,D.Maino31,48,N.Mandolesi47,29,M.Maris46,P.G.Martin9,
E.Martínez-González61,S.Masi30,S.Matarrese28,62,39,P.Mazzotta33,L.Mendes36,A.Mennella31,48,
M.Migliaccio58,65,M.-A.Miville-Deschênes55,9,A.Moneti56,L.Montier90,10,G.Morgante47,D.Mortlock52,
D.Munshi82,J.A.Murphy77,P.Naselsky79,34,F.Nati25,P.Natoli29,4,47,H.U.Nørgaard-Nielsen15,F.Noviello64,
D.Novikov74,I.Novikov79,74,C.A.Oxborrow15,L.Pagano30,50,F.Pajot55,R.Paladini53,D.Paoletti47,49,
B.Partridge41,F.Pasian46,T.J.Pearson11,53,M.Peel64,(cid:63),O.Perdereau67,F.Perrotta81,V.Pettorino40,F.Piacentini30,
M.Piat1,E.Pierpaoli21,D.Pietrobon63,S.Plaszczynski67,E.Pointecouteau90,10,G.Polenta4,45,L.Popa57,
G.W.Pratt69,S.Prunet56,89,J.-L.Puget55,J.P.Rachen19,75,M.Reinecke75,M.Remazeilles64,55,1,C.Renault71,
S.Ricciardi47,I.Ristorcelli90,10,G.Rocha63,11,C.Rosset1,M.Rossetti31,48,G.Roudier1,68,63,
J.A.Rubiño-Martín60,17,B.Rusholme53,M.Sandri47,G.Savini80,D.Scott20,L.D.Spencer82,V.Stolyarov6,85,66,
R.Sudiwala82,D.Sutton58,65,A.-S.Suur-Uski24,42,J.-F.Sygnet56,J.A.Tauber37,L.Terenzi38,47,L.Toffolatti18,61,47,
M.Tomasi31,48,M.Tristram67,M.Tucci16,G.Umana43,L.Valenziano47,J.Valiviita24,42,B.VanTent72,P.Vielva61,
F.Villa47,L.A.Wade63,B.D.Wandelt56,89,27,R.Watson64,I.K.Wehus63,D.Yvon14,A.Zacchei46,andA.Zonca26
(Affiliationscanbefoundafterthereferences)
Received21July2014/Accepted5September2015
ABSTRACT
TheAndromedagalaxy(M31)isoneofafewgalaxiesthathassufficientangularsizeontheskytoberesolvedbythePlancksatellite.Planck
has detected M31 in all of its frequency bands, and has mapped out the dust emission with the High Frequency Instrument, clearly resolving
multiplespiralarmsandsub-features.Weexaminethemorphologyofthislong-wavelengthdustemissionasseenbyPlanck,includingastudyof
itsoutermostspiralarms,andinvestigatethedustheatingmechanismacrossM31.Wefindthatdustdominatingthelongerwavelengthemission
(>∼0.3mm)isheatedbythediffusestellarpopulation(astracedby3.6µmemission),withthedustdominatingtheshorterwavelengthemission
heatedbyamixoftheoldstellarpopulationandstar-formingregions(astracedby24µmemission).Wealsofitspectralenergydistributionsfor
individual5(cid:48)pixelsandquantifythedustpropertiesacrossthegalaxy,takingintoaccountthesedifferentheatingmechanisms,findingthatthereis
alineardecreaseintemperaturewithgalactocentricdistancefordustheatedbytheoldstellarpopulation,aswouldbeexpected,withtemperatures
rangingfromaround22Kinthenucleusto14Koutsideofthe10kpcring.Finally,wemeasuretheintegratedspectrumofthewholegalaxy,
whichwefindtobewell-fittedwithaglobaldusttemperatureof(18.2±1.0)Kwithaspectralindexof1.62±0.11(assumingasinglemodified
blackbody),andasignificantamountoffree-freeemissionatintermediatefrequenciesof20–60GHz,whichcorrespondstoastarformationrate
ofaround0.12M yr−1.Wefinda2.3σdetectionofthepresenceofspinningdustemission,witha30GHzamplitudeof0.7±0.3Jy,whichisin
(cid:12)
linewithexpectationsfromourGalaxy.
Keywords.galaxies:individual:Messier31–galaxies:structure–galaxies:ISM–submillimeter:galaxies–radiocontinuum:galaxies
(cid:63) Correspondingauthor:M.Peel,[email protected]
ArticlepublishedbyEDPSciences A28,page1of23
A&A582,A28(2015)
1. Introduction than 160µm was primarily heated by star-forming regions
(henceforthabbreviatedas“SFRdust”),whiledust-dominating
The infrared (IR) and submillimetre (submm) wavelength do-
emission at wavelengths over 250µm may be primarily heated
mainisparticularlyusefulforunderstandingtheprocessesdriv-
by the total stellar populations, including evolved stars in the
ing star formation in various galactic environments, since dust
discsandbulgesofthegalaxies(henceforthabbreviatedasinter-
grainsre-emitinthisfrequencywindowtheenergythathasbeen
stellarradiationfielddust,or“ISRFdust”.Planck observations
absorbed from the UV-optical starlight. Our view of the global
of the SED of M31 allow us to examine dust heating within
innerandouterdiskstarformationandISMpropertiesofthespi-
the galaxy at frequencies much lower than was possible with
ralgalaxyinwhichweliveislimitedbyourpositioninsidethe
Herschel. Once the dust heating sources are identified empiri-
Galaxy,butournearestspiralneighbour,theAndromedagalaxy callyandtheSEDisseparatedintodifferentthermalcomponents
(alsoknownasMessier31),offersthebestviewoftheenviron-
based on the heating sources, it will be possible to more accu-
mentaleffectswithinanentiregalaxy,particularlybecauseofits
ratelymeasurethetemperatureofthecoldestdustwithinM31,
largeangularextentonthesky.
which is also critically important to properly estimate the dust
M31hasbeenextensivelystudiedatIR/submmwavelengths
mass.
with data from the Infrared Astronomical Satellite (IRAS,
Non-thermal emission from M31 can also be measured at
Habingetal.1984;Walterbos&Schwering1987),theInfrared
thelowestfrequenciescoveredbyPlanck.Synchrotronemission
Space Observatory (ISO, Haas et al. 1998), the Spitzer Space
fromM31wasdiscoveredintheearlydaysofradioastronomy
Telescope (Barmby et al. 2006; Gordon et al. 2006; Tabatabaei
(Brown&Hazard1950,1951).Ithassubsequentlybeenmapped
& Berkhuijsen 2010), and, most recently, the Herschel Space
at low frequencies (Beck et al. 1998; Berkhuijsen et al. 2003),
Observatory (Fritz et al. 2012; Groves et al. 2012; Smith et al.
and associated emission has even been detected at gamma-ray
2012; Ford et al. 2013; Draine et al. 2014; Viaene et al. 2014;
frequencies(Abdoetal.2010).However,thesynchrotronemis-
Kirketal.2015).ExceptfortheHerscheldata,theseIRobserva-
sion has not been studied at higher frequency. Free-free emis-
tionshavebeenrestrictedtoobservingthepeakofthedustemis-
sion is also expected from M31. This emission may be used
sioninthefarinfrared(FIR)aswellasmid-infrared(MIR)emis- to measure star formation rates in a way that is not affected by
sionfrom>∼100Kdustandpolycyclicaromatichydrocarbons.In dustextinctionorreliantuponassumptionsaboutthedustheat-
contrast,duetothelargeangularsizeofM31ithasbeenpartic-
ing sources (e.g., see Murphy et al. 2011). Free-free emission
ularly difficult to map the entire galaxy at wavelengths longer
from M31 was first seen by Hoernes et al. (1998), as well as
than 500µm, which is needed to constrain the Rayleigh-Jeans
Berkhuijsen et al. (2003) and Tabatabaei et al. (2013). Planck
side of the dust emission and the contribution of non-thermal
dataprovidetheopportunitytocharacterizethehigh-frequency
emission sources to the spectral energy distribution (SED). In
radioemissionforthefirsttime.
fact, the submm data for nearby spiral galaxies that have been
Section2ofthispaperdescribesthePlanckdata,andSect.3
published(e.g.,Dunneetal.2000;Stevensetal.2005;Daleetal.
theancillarydatathatweusehere.Wediscussthemorphology
2007)havehadlowsignal-to-noiselevels,havebeenbiasedto-
ofthedustasseenbyPlanckinSect.4,thecolourratiosandthe
wardsinfrared-brightsources, orhaveonlycoveredthe centres
implicationstheyhaveonthedustheatingmechanisminSect.5,
ofgalaxies. andthespectralenergydistributionson5(cid:48) scalesinSect.6and
Data from Planck (Tauber et al. 2010)1, which range
forthewholeofM31inSect.7.WeconcludeinSect.8.
from28.4to857GHz(10.5to0.35mm)withangularresolution
between31(cid:48) and5(cid:48),allowustoexaminetheRayleigh-Jeanstail
ofthedustSEDandthetransitionintofree-freeandsynchrotron 2. Planck data
emission at longer wavelengths. Moreover, since Planck ob-
Planck (Tauberetal.2010;PlanckCollaborationI2011)isthe
served the entire sky at high sensitivity, its High Frequency
thirdgenerationspacemissiontomeasuretheanisotropyofthe
Instrument (HFI, Lamarre et al. 2010) provides high signal-to-
cosmic microwave background (CMB). It observes the sky in
noise maps of M31 that extend to the outermost edges of the
nine frequency bands covering 30–857GHz (10.5 to 0.35mm)
galaxy.
with high sensitivity and angular resolution from 31(cid:48) to 5(cid:48).
Planck’s large-scale map of the region provides the oppor-
The Low Frequency Instrument (LFI; Mandolesi et al. 2010;
tunitytostudydustheatinginM31outtotheopticalradiusof
Bersanelli et al. 2010; Mennella et al. 2011) covers the 30, 44,
thegalaxy.PriorinvestigationswithIRASofdustheatinginour
and70GHz(10.5,6.8,and4.3mm)bandswithamplifierscooled
Galaxyandothershadproducedseeminglycontradictoryresults,
to 20K. The High Frequency Instrument (HFI; Lamarre et al.
with some studies indicating that dust seen at 60–100µm was
2010;PlanckHFICoreTeam2011a)coversthe100,143,217,
heated primarily by star-forming regions (Devereux & Young
353,545,and857GHz(3.0,2.1,1.38,0.85,0.55,and0.35mm)
1990; Buat & Xu 1996) and others demonstrating that evolved
bandswithbolometerscooledto0.1K.Polarizationismeasured
stellarpopulationscouldpartiallycontributetoheatingthedust
inallbutthehighesttwobands(Leahyetal.2010;Rossetetal.
seenbyIRAS(e.g.,LonsdalePersson&Helou1987;Walterbos
2010). A combination of radiative cooling and three mechani-
& Schwering 1987; Sauvage & Thuan 1992; Walterbos &
cal coolers produces the temperatures needed for the detectors
Greenawalt1996).MorerecentobservationswithHerschelofa
andoptics(PlanckCollaborationII2011).Twodataprocessing
number of galaxies, including M81, M83, NGC 2403 (Bendo
centres (DPCs) check and calibrate the data and make maps of
et al. 2010, 2012a) and M33 (Boquien et al. 2011), demon-
the sky (Planck HFI Core Team 2011b; Zacchei et al. 2011).
strated that dust-dominating emission at wavelengths shorter
Planck’ssensitivity,angularresolution,andfrequencycoverage
makeitapowerfulinstrumentforGalacticandextragalacticas-
1 Planck (http://www.esa.int/Planck) is a project of the
trophysics as well as cosmology. Early astrophysics results are
EuropeanSpaceAgency–ESA–withinstrumentsprovidedbytwosci-
giveninPlanckCollaborationVIII–XXVI(2011),basedondata
entificConsortiafundedbyESAmemberstates(inparticularthelead
countries:FranceandItaly)withcontributionsfromNASA(USA),and taken between 13 August 2009 and 7 June 2010. Intermediate
telescopereflectorsprovidedinacollaborationbetweenESAandasci- astrophysicsresultsarepresentedinaseriesofpapersbetween
entificConsortiumledandfundedbyDenmark. themajordatareleases.
A28,page2of23
PlanckCollaboration:AndromedaasseenbyPlanck
CMB map to subtract the CMB from the Planck data, as from
+43:00
a visual inspection this appears to be the cleanest map of the
-4 .0 -3 .5 -4 .0 -3 .5 CMBinthisregionfromthefourmethods(seeFig.1).TheNILC
log(T (K)) log(T (K))
+42:00 RJ RJ andSEVEMmapsareparticularlycontaminatedbyemissionfrom
M31; we also use the NILC map to test the impact of residual
+41:00 foregroundemissionbeingsubtractedoutofthemapsalongwith
theCMB.Figure1showsthe217GHz(1.38mm)mappre-and
+40:00 post-CMB subtraction, along with the four CMB maps of the
217 GHz (1.38 mm) 217 GHz (1.38 mm) M31region.
No CMB subtraction CMB-subtracted
The maps are converted from CMB temperature (T ) to
+43:00 Rayleigh-Jeansbrightnesstemperature(TRJ)usingthesCtaMnBdard
000) +42:00 -4 0T0CMB0 (µK40 )0 -4 0T0CMB0 (µK40 )0 acloseoffiucsieentthseansomdeisncarlibfreedquinenPcileasncfkorCthoellabbaonrdast,ioannd(2(0w1h3e)n; fiwte-
2
J ting a spectral model to the data) colour corrections, depend-
n (
o ing on the spectra of the emission. The 100 and 217GHz (3.0
nati +41:00 and1.38mm)PlanckbandsincludeCOemission.TheCOemis-
ecli sionfromM31hasbeenmappedwithground-basedtelescopes
D +40:00
(e.g., Koperetal.1991; Dame et al. 1993; Nieten et al. 2006),
SMICA CMB map Commander CMB map
but these do not include the full extent of M31 that is consid-
+43:00 eredhere.TheCOemissionistooweaktobereliablydetectedin
thefull-skyPlanckCOmaps(PlanckCollaborationXIII2014)3.
-4 00 0 40 0 -4 00 0 40 0
T (µK) T (µK) WedonotcorrectfortheCOemissioninthemaps;insteadwe
+42:00 CMB CMB
omit the 217GHz (1.38mm) channel from the SED fitting in
Sect.6,andwecomparethelevelofCOemissionexpectedfrom
+41:00 theground-basedCOmapofNietenetal.(2006)(describedin
Sect.3.1)withtheemissionattributabletoCOintheintegrated
+40:00 PlanckmeasurementsofM31inSect.7.
SEVEM CMB map NILC CMB map To assess the uncertainty in the Planck maps, we estimate
the instrumental noise and cirrus contamination by measuring
00 :50 00: 40 00: 50 00 :40
the scatter of flux densities in an adjacent background region
Right Ascension (J2000)
of the Planck maps (see Sect. 6 for details). We conservatively
Fig.1. Top left: M31 at 217GHz (1.38mm), with no CMB subtrac- assume calibration uncertainties of 10% for 857 and 545GHz
tion. Top right: the SMICA CMB-subtracted 217GHz (1.38mm) map.
(350and550µm)and3%forallotherPlanck frequencies(see
Middle left: the SMICA map of the CMB in the same region. Middle
PlanckCollaboration2013).
right: Commander CMB map. Bottom left: SEVEM CMB map. Bottom
For the integrated spectrum analysis in Sect. 7, the Planck
right:NILCCMBmap.
full-sky maps are used directly in HEALPix4 format (Górski
et al. 2005). For the higher resolution analyses, however, we
InthispaperweusePlanck datafromthe2013distribution use “postage stamp” 2D projected maps centred on M31. To
of released products (Planck Collaboration I 2014), which can conserve the photometry of the data whilst repixelizing, we
beaccessedviathePlanckLegacyArchiveinterface2,basedon use the Gnomdrizz package (Paradis et al. 2012a) to create
thedataacquiredbyPlanckduringits“nominal”operationspe- the postage stamp maps from the HEALPix data; since the
riod from 13 August 2009 to 27 November 2010. In order to HEALPixGnomviewfunctionusesnearest-neighbourinterpola-
studyM31inthePlanckmaps,theCMBneedstobesubtracted. tion, it does not necessarily conserve the photometry, although
CMBmaps inthe M31 regionare showninFig. 1.The CMB- wetestedthatthereisnosignificantdifferenceinthiscase.The
subtracted Planck maps are presented in Fig. 2 at their native resulting postage stamp maps in equatorial coordinates are of
resolution and their properties are summarized in Table 1. We size 250(cid:48) ×250(cid:48) with 0.5(cid:48) ×0.5(cid:48) pixels, centred on RA 10◦.68,
use various combinations of the Planck maps throughout this Dec41◦.27(l=121◦.2,b=−21◦.6).
paper,andallmapsareusedtoqualitativelystudytheemission Whenquantitativelyanalysingthedata,wefirstsmoothtoa
atallPlanckfrequenciesinSect.4.Mapsat217GHzandhigher commonresolutionofeither5(cid:48) (at217GHzandabove/1.38mm
frequencies(1.38mmandshorterwavelengths)areusedtoquan- orlower,wherethedatahaveanativeresolutionof4.(cid:48)39–4.(cid:48)87)
titativelyinvestigatetheemissioninSects.5and6,andallfre- or 1◦ (at all frequencies) by convolving the map with a circu-
quenciesareusedtomeasurethetotalemissioninSect.7. larGaussianbeamwithafull-widthathalf-maximum(FWHM)
CMB subtraction is particularly important for M31, of θ = (cid:16)θ2 −θ2 (cid:17)1/2, where θ is the desired FWHM
conv new old new
given the similarities in angular size between M31 and the
andθ isthecurrentFWHMofthemaps.Forsomeofthelater
old
anisotropies in the CMB. Additionally, there is an unfortu- analysis,wealsorepixelizethe5(cid:48) resolutiondatainto5(cid:48) pixels
nately large (approximately290µK) positive CMB fluctuation (see Sect. 5), while the 1◦ data are analysed at their native
at the southern end of M31, which can be clearly seen in
the CMB map panels of Fig. 1. As part of the Planck 2013 3 There is no detection in the Planck Type 1 CO map (Planck
delivery, CMB maps from four component separation tech- Collaboration XIII 2014); the morphology is not consistent with the
niques were released, namely maps from the Commander, knownstructureintheType2map,andalthoughthereisadetectionin
NILC,SEVEMandSMICAcomponentseparationmethods(Planck theType3mapandtheringmorphologyisvisible,thedetectiondoes
Collaboration XII 2014). We specifically use the SMICA nothaveahighsignal-to-noiseratioandmaybecontaminatedbydust
emission.
2 http://pla.esac.esa.int/pla/ 4 http://healpix.jpl.nasa.gov
A28,page3of23
A&A582,A28(2015)

+43:00

-3. 75 -3. 2 5 - 4 .0 -3 .5 -4 .0 -3 .5
log(T (K)) log(T (K)) log(T (K))
+42:00 RJ RJ RJ
+41:00
+40:00
28.4 GHz (10.5 mm) 44.1 GHz (6.8 mm) 70.4 GHz (4.3 mm)
+43:00

0) -4 . 25 -3. 75 -4. 25 -3. 7 5 -4 .0 -3 .5
0 log(T (K)) log(T (K)) log(T (K))
0 +42:00 RJ RJ RJ
2
J
(
n
o
ati +41:00
n
cli
e
D +40:00
100 GHz (3.0 mm) 143 GHz (2.1 mm) 217 GHz (1.38 mm)
+43:00

-4 .0 -3 .5 - 4 .0 -3 .5 -3 .0 - 4 .0-3 .5-3 . 0
log(T (K)) log(T (K)) log(T (K))
+42:00 RJ RJ RJ
+41:00
+40:00
353 GHz (850 µm) 545 GHz (550 µm) 857 GHz (350 µm)
00 :50 00: 40 00: 50 00: 40 00: 50 00: 40
Right Ascension (J2000)
Fig.2. Maps of M31 in total intensity from Planck (after CMB subtraction). Top to bottom, left to right: Planck 28.4, 44.1, and 70.4GHz;
Planck 100, 143, and 217GHz; Planck 353, 545, and 857GHz. All plots have units of Kelvin (T ), have a 1◦ equatorial graticule overlaid,
RJ
are250(cid:48)×250(cid:48)with0.5(cid:48)×0.5(cid:48)pixelsandarecentredonRA10◦.68,Dec41◦.27,withnorthupandeasttotheleft.
HEALPixN resolution.ThePlanckbeamsarenotsymmetric ameterof190.(cid:48)5±0.(cid:48)1andamajor-to-minorratioof3.09±0.14
side
at the roughly 20% level (e.g., see Zacchei et al. 2011; Planck (deVaucouleursetal.1991).FollowingXu&Helou(1996),we
HFI Core Team 2011b), and the ancillary data sets used (see assumethatithasaninclinationanglei=79◦andapositionan-
Sect.3)willalsohavenon-Gaussianbeams;smoothingthedata gleof37◦withrespecttoNorth(bothinequatorialcoordinates).
to a common resolution reduces the effect of the asymmetry.
However, a residual low-level effect will still be present in this
3.1. The5(cid:48) dataset
analysis, for example in terms of introducing some correlation
betweenadjacent5(cid:48)pixels.
WeusesixdatasetsinadditiontothePlanckdatainour5(cid:48) res-
olution study. At high frequencies, we use IRAS, Spitzer,
Herschel, and ISO data to trace the shorter wavelength dust
3. Ancillarydata

emission. At lower frequencies, we use ground-based H and
Theancillarydatathatweuseinthispaperfallundertwocate- CO data sets to trace the gas within the galaxy. To match the
gories.Forthehigherresolutionspatialanalysis,weneedobser- pixelizationandresolutionoftheseotherdatatothePlanckdata,
vationswithresolutionequaltoorgreaterthanthePlanck high weregridthedataonto0.25(cid:48) pixelsinthesamecoordinatesys-
frequencyresolutionof5(cid:48);thesearedescribedinSect.3.1.For temandskyregionasthePlanckdataandthensmooththemto
theintegratedSED,wecanmakeuseoflarge-scalesurveydata acommonresolutionof5(cid:48).Furtherrepixelizationto5(cid:48) pixelsis
witharesolutionof1◦ orhigher;wedescribethesedatasetsin thencarriedoutfortheanalysisinlatersections.
Sect.3.2.AllofthedatasetsaresummarizedinTable1. To trace dust emission in the infrared, we use 24, 70,
Wealsomakeuseofancillaryinformationaboutthedistance and160µmdataoriginallyacquiredbyGordonetal.(2006)us-
andinclinationofM31.M31isatadistanceof(785±25)kpc ingtheMultibandImagingPhotometerforSpitzer(MIPS;Rieke
(McConnachie et al. 2005), with an optical major isophotal di- etal.2004)ontheSpitzerSpaceTelescope(Werneretal.2004).
A28,page4of23
PlanckCollaboration:AndromedaasseenbyPlanck
Table1.Sourcesofthedatasetsusedinthispaper,aswellastheirfrequency,wavelength,resolution,calibrationuncertainty,andrmson5(cid:48)scales
(fordatawith5(cid:48)resolutionorbetteronly;seelater).
Source ν[GHz] λ[mm] Res. Unc.σ(5(cid:48),Jy) Analysis Reference
Haslam 0.408 734 60.(cid:48) 10% ... L Haslametal.(1982)
Dwingeloo 0.820 365 72.(cid:48) 10% ... L Berkhuijsen(1972)
Reich 1.4 214 35.(cid:48) 10% ... L Reichetal.(2001)
WMAP9-year 22.8 13 49.(cid:48) 3% ... L Bennettetal.(2013)
Planck 28.4 10.5 32.(cid:48)24 3% ... L PlanckCollaborationII(2014)
WMAP9-year 33.0 9.0 40.(cid:48) 3% ... L Bennettetal.(2013)
WMAP9-year 40.7 7.4 31.(cid:48) 3% ... L Bennettetal.(2013)
Planck 44.1 6.8 27.(cid:48)01 3% ... L PlanckCollaborationII(2014)
WMAP9-year 60.7 4.9 21.(cid:48) 3% ... L Bennettetal.(2013)
Planck 70.4 4.3 13.(cid:48)25 3% ... L PlanckCollaborationII(2014)
WMAP9-year 93.5 3.2 13.(cid:48) 3% ... L Bennettetal.(2013)
Planck 100 3.0 9.(cid:48)65 3% ... L PlanckCollaborationVI(2014)
Planck 143 2.1 7.(cid:48)25 3% ... L PlanckCollaborationVI(2014)
Planck 217 1.38 4.(cid:48)99 3% 0.015 LH PlanckCollaborationVI(2014)
Planck 353 0.85 4.(cid:48)82 3% 0.05 LH PlanckCollaborationVI(2014)
Planck 545 0.55 4.(cid:48)68 7% 0.13 LH PlanckCollaborationVI(2014)
HerschelSPIREPLW 600 0.50 0.(cid:48)59 4% 0.09 H Fritzetal.(2012)
Planck 857 0.35 4.(cid:48)33 7% 0.31 LH PlanckCollaborationVI(2014)
HerschelSPIREPMW 857 0.35 0.(cid:48)40 4% 0.23 H Fritzetal.(2012)
HerschelSPIREPSW 1200 0.25 0.(cid:48)29 4% 0.29 H Fritzetal.(2012)
COBE-DIRBE 1249 0.240 40.(cid:48) 13% ... L Hauseretal.(1998)
ISO 1763 0.175 1.(cid:48)3 10% 0.8 H Haasetal.(1998)
SpitzerMIPSB3 1870 0.16 0.(cid:48)63 12% 0.53 H Gordonetal.(2006),Bendoetal.(2012b)
HerschelPACS 1874 0.16 0.(cid:48)22 4% 2.4 H Fritzetal.(2012)
COBE-DIRBE 2141 0.14 40.(cid:48) 13% ... L Hauseretal.(1998)
COBE-DIRBE 2997 0.10 40.(cid:48) 13% ... L Hauseretal.(1998)
HerschelPACS 2998 0.10 0.(cid:48)21 4% 2.5 H Fritzetal.(2012)
IRAS(IRIS)B4 3000 0.10 4.(cid:48)3 13% 0.37 LH Miville-Deschênes&Lagache(2005)
SpitzerMIPSB2 4280 0.07 0.(cid:48)3 10% 0.12 H Gordonetal.(2006),Bendoetal.(2012b)
COBE-DIRBE 5000 0.06 40.(cid:48) 13% ... L Hauseretal.(1998)
IRAS(IRIS)B3 5000 0.06 4.(cid:48)0 13% 0.12 LH Miville-Deschênes&Lagache(2005)
IRAS(IRIS)B2 12000 0.025 3.(cid:48)8 13% 0.06 LH Miville-Deschênes&Lagache(2005)
SpitzerMIPSB1 12490 0.024 0.(cid:48)1 4% 0.012 H Gordonetal.(2006),Bendoetal.(2012b)
IRAS(IRIS)B1 25000 0.012 3.(cid:48)8 13% 0.04 LH Miville-Deschênes&Lagache(2005)
SpitzerIRAC 83000 0.0036 1.(cid:48)(cid:48)7 3% ... H Barmbyetal.(2006),SpitzerScienceCenter(2012)
DRAOH 1.4 ... 0.(cid:48)37 5% ... H Cheminetal.(2009)
IRAMCO 115. ... 0.(cid:48)38 15% ... H Nietenetal.(2006)
Notes.TheAnalysiscolumnindicateswhetherthedatasethasbeenusedinthe5(cid:48)high-resolution(H)and/or1◦low-resolution(L)analysis.The
firstpartofthetabledescribesthecontinuumdatasets,andthesecondpartdescribesthespectrallinedatasets.
The 24, 70, and 160µm data have point spread functions with sideofthemasktoassessthermsuncertaintyintheSpitzerdata,
FWHM of 6(cid:48)(cid:48), 18(cid:48)(cid:48), and 38(cid:48)(cid:48), respectively. The data were pro- includingthermalnoise,residualcirrusandotherlarge-scalefea-
cessedusingtheMIPSDataAnalysisTools(Gordonetal.2005) turesthatcontributetotheuncertaintyon5(cid:48)scales.
as well as additional data processing steps described by Bendo Weadditionallyutilisethe3.6µmdataproducedbyBarmby
et al. (2012b). The data processing includes the removal of the et al. (2006) using the Spitzer Infrared Array Camera (IRAC;
effect of cosmic ray hits and drift in the background signal. Fazioetal.2004)totracetheolderstellarpopulationinthebulge
The background is initially subtracted from the individual data ofM31;thesedatahavearesolutionof1(cid:48).(cid:48)7.
framesbycharacterizingthevariationsinthebackgroundsignal We also include Herschel data from the Herschel
as a function of time using data that fall outside of the galaxy. Exploitation of Local Galaxy Andromeda (HELGA) survey
This removes large-scale background/foreground structure out- (Fritz et al. 2012). These data are at 500, 350, 250, 160
sidethegalaxy,includingthecosmicinfraredbackground(CIB) and100µm(600,857,1200,1874and2998GHz)andhaveres-
andzodiacallight,butresidualforegroundcirrusstructuresand olutionsbetween0.(cid:48)21and0.(cid:48)59(Lutz2012;Valtchanov2014).
compact sources remain in the data. Any residual background We re-align the data to match the Spitzer 24µm data using
emissioninthefinalmapsismeasuredoutsidetheopticaldiscof three compact sources that are seen at all Herschel frequen-
thegalaxyandsubtractedfromthedata.Foradditionaldetailson cies.TheHerscheldatahaveafluxcalibrationuncertaintyof4%
thedatareductionprocess,seethedescriptionofthedatareduc- (HerschelSpaceObservatory2013;Bendoetal.2013).Forthe
tionbyBendoetal.(2012b).PSFcharacteristicsandcalibration integrated SED, we smooth the Herschel data to 1◦, and then
uncertaintiesaredescribedbyEngelbrachtetal.(2007),Gordon regrid these data onto an oversampled N = 16384 HEALPix
side
etal.(2007),andStansberryetal.(2007).Weusethedataout- map,whichwasthenresampleddowntoN =256.
side
A28,page5of23
A&A582,A28(2015)
We also make use of several data sets for comparison pur- 4. Infraredmorphology
poses, namely: the IRAS data at 12, 25, 60, and 100µm from
Miville-Deschênes & Lagache (2005); the ISO 175µm data ThePlanckmapsofM31areshowninFig.2.At100GHzand
from Haas et al. (1998, priv. comm.); the H emission map allhigherPlanckfrequencies(3.0mmandshorterwavelengths),
from Chemin et al. (2009); and the 12CO J = 1→0 map the prominent 10kpc dust ring of Andromeda can clearly be
of M31 from Nieten et al. (2006). These are described in seen – as expected based on previous infrared observations of
AppendixA. M31–aswellasanumberofotherextendedfeatures.
At high frequencies, where we have the highest signal-to-
noise and highest spatial resolution, we can see features much
3.2. The1◦ dataset further out than the 10kpc ring. The top panels of Fig. 3 show
thePlanck857GHz(350µm)maplabelledtoshowthelocations
of the key features in the dust emission; the bottom-left panel
WhenlookingattheintegratedspectrumofM31,wemakeuse
shows the Herschel 857GHz (350µm) data, and the bottom-
of a number of large-area, low-resolution surveys. These are 
rightpanelofFig.3showstheH mapfromCheminetal.(2009)
publicly available in HEALPix format. We convolve the data
for comparison. A cut through the Planck 857 and 353GHz
sets to a resolution of 1◦ to match the resolution of the low-
(350 and 850µm) maps is shown in Fig. 4. To the south, a to-
frequency data sets and perform the analysis directly on the
tal of four spiral arm or ring structures can clearly be seen –
HEALPixmaps.
the 10kpc ring (“F8”), as well as a second structure just out-
At low frequencies, we use the low-resolution radio maps side of the ring (“F9”) and the more distant 21kpc (“F10”)
of the sky at 408MHz (73.4cm; Haslam et al. 1981, 1982), and 26kpc (“F11”) arms. These features have also been iden-
820MHz, (36.5cm; Berkhuijsen 1972) and 1.4GHz (21.4cm; tifiedbyFritzetal.(2012),andcanalsobeseeninthepaneldis-
Reich1982;Reich&Reich1986;Reichetal.2001).Weassume playing the Herschel 857GHz data. We see a hint of the emis-
that there is a 10% uncertainty in these maps, which includes sion at 31kpc (“F12”), which is also seen in H emission, but
both the uncertainty in the flux density calibration (between 5 weareunabletoconfirmitasbeingpartofM31,becauseofthe
and 10% depending on the survey) and also uncertainties aris- largeamountofsurroundingcirrusemissionfromourGalaxy.To
ingfromthemorphologyofthesurroundingstructureinteracting thenorth,weseethreesetsofspiralarmstructures(“F4”,“F3”,
with the aperture photometry technique we use. This choice of and“F2”)andawispofemissionatthenorthernmostendofthe
uncertaintyhasbeenshowntobereasonableforGalacticanoma- galaxy(“F1”,confirmedbycross-checkingagainstHemission,
lous microwave emission (AME) clouds (Planck Collaboration as shown in the top-right panel) before running into confusion
Int. XV 2014). For the 408MHz (73.4cm) map, we also add fromGalacticemission.Thesefeaturescanbeseenmostclearly
3.8Jy to the uncertainty to take into account the baseline stria- at857and545GHz(350and550µm),butarealsoclearlyvisi-
tions in the map, which are at the level of ±3K (Haslam et al. bleinthe353,217and143GHz(0.85,1.38and2.1mm)maps
1982;PlanckCollaborationInt.XV2014).Allofthemapshave afterCMBsubtraction.The10kpcringcanstillbeseenclearly
beencalibratedonangularscalesofaround5◦,andconsequently at100GHz,butatthatfrequencythemoreextendedstructureis
thedifferencebetweenthemainandfullbeams(thelatterinclud- notvisible.Wewillusetheterminologyof“ring”forthe10kpc
ing sidelobes) needs to be taken into account when measuring ring,sinceithasbeenclearlydemonstratedtobeanear-complete
the flux densities of more compact sources. This is significant ring(e.g.,seeHaasetal.1998),and“arm”forallotherstructures
for the 1.4GHz (21.4cm) map, where the factor for objects on thatmaynotcompletelycirclethegalaxy.
1◦ scales is approximately 1.55. As M31 is on an intermedi-
Anareaofparticularlystronglong-wavelengthemissioncan
ate scale, we adopt an intermediate correction factor of 1.3 ±
be clearly seen at the southern end of the 10kpc ring, down
0.1,andalsoincludetheuncertaintyinthiscorrectionfactorin
to frequencies of 100GHz (3.0mm; see Fig. 2; in Fig. 3 it is
the flux density uncertainty. We assume that the value and un-
just to the right of “F8” and is also marked as “S6”). This
certainty on the ratio for the 408MHz (73.4cm) and 820MHz
has the highest contrast with the rest of M31 at frequencies
(36.5cm) maps is small, and hence will be well within the ex-
of 143 and 217GHz (2.1 and 1.38mm); there are also hints of
istingcalibrationuncertainties,aspere.g.,PlanckCollaboration
it down to 70.4GHz (4.3mm). This is probably what is seen
Int. XV (2014). We also make use of the integrated flux densi-
at the highest frequencies of the WMAP data and in Planck
tiesfromhigher-resolutionsurveyscollatedbyBerkhuijsenetal.
data by De Paolis et al. (2011, 2014); i.e., the asymmetrical
(2003)forcomparisontothoseextractedfromthemaps.
emission between the north and south parts of M31 that they
Atintermediatefrequencies,inadditiontoPlanck-LFIdata, detect is caused by either varying dust properties across the
weusethedeconvolvedandsymmetrized1◦-smoothedWMAP galaxyorCMBfluctuations,ratherthanbeingcausedbygalactic
9-yeardataat22.8,33.0,40.7,60.7,and93.5GHz(13,9.0,7.4, rotation.
4.9,and3.2mm; Bennettetal.2013)5.Whenfittingamodelto
Within the 10kpc ring, several features are also visible. A
these data, we apply colour corrections following the recipe in
brightspotinsidetheGalacticringtothenorthismarkedas“F5”
Bennettetal.(2013),andweconservativelyassumeacalibration
in the top-right panel of Fig. 3. This region corresponds to a
uncertaintyof3%.
limbofanasymmetricspiralstructurewithinthe10kpcringthat
At higher frequencies, we include the low-resolution hasalsobeenseeninthehigher-resolutionSpitzerdata(Gordon
COBE-DIRBE data at 1249, 2141, and 2997GHz (240, 140, etal.2006).Planckdoesnotdetecttheemissionfromthecentral
and 100µm; Hauser et al. 1998), in addition to the IRAS data nucleus,despitetheprominenceofthisregioninmapsofM31
described in Appendix A. We assume that these data have an athigherfrequencies.Thisimpliesthatthemajorityofthedust
uncertaintyof13%.WealsousetheHerscheldataasdescribed in this region is at a higher dust temperature than average –
above. an implication which will be explored in later sections. Planck
does, however, detect a compact object to the right of the nu-
5 http://lambda.gsfc.nasa.gov/product/map/dr4/maps_ cleus(markedas“F7”)thathasalsobeenseenwithIRAS(Rice
band_smth_r9_i_9yr_get.cfm 1993),ISO(Haasetal.1998),andSpitzer(Gordonetal.2006);
A28,page6of23
PlanckCollaboration:AndromedaasseenbyPlanck

+43:00

0) F1 F2 -l4o .g0(T-3 .(5K)) - 4 .l0og(-T3 .5 (K-)3) .0
00 +42:00 F3 RJ S1 RJ
2 F6 S2
J F4
( B3 F5 S3
n F7
o
ati +41:00 S5 S4
n
cli F8 S6
e F9
D +40:00 Planck Planck Fig.3. Top-left: the 857GHz (350µm) Planck
857 GHz F10 857 GHz map of M31, highlighting the extension of
(350 µm) F11 (350 µm) the dust out to large distances. Fields of in-
F12
terest are labelled with “F,” and circled in
+43:00 some cases. The position of the radio source
B3 0035+413 is also shown (labeled B3).
000) +42:00 F1 F3F2 l o- 5g .0(Jy a-4r c.0sec- 2 ) 1 .0log(T2R .J0 (K)) Tscoapl-erigchhot:sethnet8o57hiGghHlzig(h3t5t0hµemb)rimghatpe,rweimthisa-
2 F6 sion. Circles labelled with “S” indicate the
J F4 sources included in the Planck ERCSC (see
( B3
n Appendix B); two fields in the centre re-
o
ati +41:00 gion are also labelled with “F”. Bottom-left:
n the857GHz(350µm)Herschelmapconvolved
cli F8 tothesameresolutionasthePlanckdata,with
e F9 the same fields of interest labelled as in the
D +40:00 Herschel
857 GHz F10 HI Planck map above. Bottom-right: a map of
(350 µm) F11 the H emission from M31 (Chemin et al.
F12
2009)convolvedtothesameresolutionasthe
00: 50 00: 40 00: 50 00: 40 Planck 857GHz (350µm) map, with 5, 10,
and15kpcringsoverlaidingreen,andthecut
Right Ascension (J2000) Right Ascension (J2000)
usedinFig.4inblue.
9 oftheSunyaev-ZeldovicheffectinthehaloofM31(e.g.,Taylor
F4
etal.2003)unlesstheemissionisproperlytakenintoaccountor
8 F8
suitablymasked.
)MB 7 F5 AtPlanck’slowestfrequencies,M31isclearlydetectedbut
C
e (T 6 is not resolved, because of the low resolution at these frequen-
ur F3 cies. It is clearest at 28.4GHz (10.5mm), but can also be seen
erat 5 F9 at44.1and70.4GHz(6.8and4.3mm)–althoughatthesefre-
p
m quencies CMB subtraction uncertainty starts to become impor-
e 4
s t tant. The emission mechanism powering the source detections
es 3 F1 atthesefrequencieswillbediscussedinthecontextoftheinte-
n
Bright 2 F10F11F12 gratTedheSrEeDarienaSlseocts.e7v.eralsourcesinthefieldneartoM31that
1 arepresentinthemapsshownhere,includingthedwarfgalaxy
M110/NGC 205 (“F6”) and the blazar B30035+413 (“B3”),
0
-40 -30 -20 -10 0 10 20 30 40 additionally a number of components of M31 are included in
Galactocentric distance (kpc) thePlanck EarlyReleaseCompactSourceCatalogue(ERCSC;
Fig.4. Cut along the major axis of M31. Negative values are on Planck Collaboration VII 2011) and the Planck Catalogue of
the higher declination side of the galaxy. The red line is 857GHz CompactSources(PCCS; PlanckCollaborationXXVIII2014).
(350µm) and the blue line is 353GHz (850µm) rescaled by a factor WediscusstheseinAppendixB.
of 1000. Various spiral arm structures, including the outermost rings
at 20–30kpc, are clearly visible at both frequencies. Fields are num- 5. Colourratiosandimplicationsfordustheating
beredasperFig.3.
5.1. Colourratios
AnumberofrecentanalysesbasedonHerscheldatahaveused
thelattersuggeststhatthisemissionislocatedwhereabarand FIR surface brightness ratios to examine dust heating sources
thespiralarmstructuremeet. within nearby galaxies. Bendo et al. (2010) performed the first
Outside of the ring, we see some cirrus dust emission from such analysis on M81; subsequent analyses were performed
our own Galaxy, particularly at the higher Planck frequencies. on M33 by Boquien et al. (2011), on M83 and NGC2403 by
This is largely present on the north side of M31, towards the Bendo et al. (2012a), and on a wider sample of galaxies by
Galactic plane, but it can also be seen elsewhere, for example Bendo et al. (2015). These techniques have been shown to be
thereisanarcofcirrusemissionabovetheoutermostringsatthe morerobustatidentifyingdustheatingsourcesthancomparing
south end of M31. Depending on the temperature and spectral FIRsurfacebrightnessesdirectlytostar-formationtracers,since
propertiesofthisemission,itmaypresentproblemsforstudies brightnessratiosnormalizeoutthedustcolumndensity.
A28,page7of23
A&A582,A28(2015)

+42:30
+42:00 l og-5(J.5y arc-s5e.0c-2 ) l- o5g.0(Jy a-4rc.5sec-2 ) -4l o.5g(Jy -a4r.c0sec-2 ) -4l o.5g(Jy- 4a.r0csec-3-2 .)5 l- o4g.0(Jy a-3rc.5sec-2 )
+41:30
+41:00
+40:30
Planck Planck Planck Herschel Planck
217 GHz (1.38 mm) 353 GHz (850 µm) 545 GHz (550 µm) 600 GHz (500 µm) 857 GHz (350 µm)
+40:00
+42:30
00) +42:00 l- o4g.0(Jy a-3rc.5sec-2 ) l og(-J3y. 5arcse-3c.-02 ) l og(-J3y.5 ar-c3s.e0c-2 ) l og(-J3y.5 ar-c3s.e0c-2 ) l og(-J3y.5 ar-c3s.e0c-2 )
0
2
n (J +41:30
o
ati +41:00
n
ecli +40:30
D Herschel Herschel ISO Herschel Spitzer
857 GHz (350 µm) 1200 GHz (250 µm) 1763 GHz (175 µm) 1870 GHz (160 µm) 1870 GHz (160 µm)
+40:00

+42:30
+42:00 l og-(4J.y0 -a3r.c5s-e3c.-02 ) l o-g4(.J0y a-3rc.5sec-3-2 .)0 -5l o.0g(Jy a-4rc.0sec-2 ) -5l o.0g(-J4y. 5ar-c4s.e0c-2 ) l o-g5(.J5y -a5r.c0se-c4-.2 5)
+41:30
+41:00
+40:30
Herschel IRAS Spitzer IRAS IRAS
3000 GHz (100 µm) 3000 GHz (100 µm) 4280 GHz (70 µm) 5000 GHz (60 µm) 12490 GHz (24 µm)
+40:00
00:50 00:40 00:50 00:40 00:50 00:40 00:50 00:40 00:50 00:40
Right Ascension (J2000)
Fig.5.Maskedandpixelized5(cid:48) datasetorderedinincreasingfrequency(decreasingwavelength).Themapsare3◦×3◦ insize.Thereisaclear
changeofmorphologyasfrequencyincreases,withtheringsbeingmostprominentatthelowerfrequencies,andthenucleusbeinghighlightedat
higherfrequencies.
We studied the dust heating sources in M31 using surface in surface brightness. The bright feature at the southern end of
brightness ratios measured in all Planck data at frequencies M31(source“S6”inFig.3)isclearlypresentatallfrequencies,
of 217GHz and above (1.38mm and shorter wavelengths), as althoughitismorenotableatν < 1700GHz.Thereisachange
wellasinHerschel,ISO,Spitzer,andIRASdata.Weconvolved of morphology apparent as frequency increases, with the rings
all of the high-resolution data (marked with “H” in Table 1) beingmostprominentatthelowerfrequencies,andthenucleus
to 5(cid:48) resolution using a Gaussian kernel. We then rebinned the beinghighlightedathigherfrequencies.
datato5(cid:48)pixels;weselectedthissizebecauseitisthesamesize Figure 6 shows the surface brightness ratios between ad-
as the beam, and so the signal in each pixel should be largely jacent frequency bands in the combined Planck and ancillary
independent of the others. It is important to do this repixeliza- data set, where e.g., S /S denotes the colour ratio be-
545 353
tioninordertoavoidtheappearanceofartificialcorrelationsin tween545and353GHz(0.55and0.85mm).Theratiosoflower
thedata.Wemaskedoutdatafromoutsidetheopticaldiscofthe frequency bands (i.e., the Planck bands) appear to smoothly
galaxybyrequiringthepixelcentretobelessthan22kpcaway decrease with radius, although the data are somewhat noisy
from the centre of M31 (which is equivalent to a 1◦.42 radius at S /S . The 10kpc ring is hardly noticable at all in the
545 353
alongthemajoraxis).Toavoidpixelsstronglyaffectedbyfore- lowerfrequencyratiomaps.Inthehigherfrequencyratiomaps
groundorbackgroundnoise,weonlyuseddatafrompixelsthat (e.g., the S /S map), the ring is much more prominent,
4280 3000
had been detected in the 353GHz (850µm) image at 10 times andthecoloursalsoappearenhancedinacompactregionaround
thermsuncertaintyinthedata(measuredina10×10pixelre- thenucleus.
gionofskytothefarbottom-leftofthedatasetusedhere).We Theseresultsimplythatthedifferentfrequenciesaredetect-
also removed pixels at the top-left corner of the images, where ing dust heated by different sources. At higher frequencies, the
the data have been contaminated by bright cirrus structures in enhancedsurfacebrightnessratiosintheringandthenucleusin-
theforeground.Finally,weremovedpixelsattheedgeoftheop- dicatethatthedustisbeingheatedbythestar-formingregionsin
ticaldisc,whichprimarilysampleacombinationofbackground these structures. At lower frequencies, however, the ratios vary
emissionandemissionfromthewingsofthebeamsforsources more smoothly with radius, and the ring does not appear to be
in the dust ring. This leaves 126 independent data points. We as enhanced as in the data for the higher frequencies. This is
look at colour ratios in adjacent bands of the same telescope consistent with dust heating being dominated by the total stel-
to minimize the effect of different telescope beams on the re- larpopulation,includingstarsinthebulgeofthegalaxy,which
sults; the exception to this is the highest frequency compari- should vary smoothly with radius (as has been suggested for
son, where we have no alternative but to compare Spitzer with othergalaxiesbyBendoetal.2010,2012a).
Herscheldata.
The rebinned data are shown in Fig. 5. The 10kpc ring is
5.2. Determinationofdustheatingmechanism
present at all frequencies, but emission from the centre of the
galaxy becomes more prominent as the frequency increases. Tolinkthesurfacebrightnessratiostoheatingsources,wecom-
At1800GHz(170µm),bothsourcesareapproximatelysimilar paredtheratiostotracersofthetotalstellarpopulationandstar
A28,page8of23
PlanckCollaboration:AndromedaasseenbyPlanck
formation. As a tracer of the total stellar population, we used
+42:30 the Spitzer 3.6µm image, which is generally expected to be
+42:00 3.5 4.0 0 .0 0.5 1 .0
dominatedbytheRayleigh-Jeanstailofthephotosphericemis-
+41:30 sion from the total stellar population (Lu et al. 2003). While
+41:00 the 3.6µm band may also include roughly 1000K dust emis-
sionfromstar-formingregions(Mentuchetal.2009,2010),this
+40:30
effect is usually only a major issue in late-type galaxies with
+40:00 SP545/SP353 ESF/ETotal (SP545/SP353)
relativelystrongstarformationcomparedtothetotalstellarsur-
+42:30 face density. A high infrared-to-visible ratio or a dominant ac-
+42:00 3.0 3.5 0 .0 0.5 1 .0 tive galactic nuclei (AGN) could also contaminate the 3.6µm
emission; however, neither of these issues are present in M31.
+41:30
We used the Spitzer 24µm image as a star-formation tracer,
+41:00
as this has been shown to generally originate from SFR dust
+40:30 (Calzettietal.2005,2007;Prescottetal.2007;Zhuetal.2008;
+40:00 SP857/SP545 ESF/ETotal (SP857/SP545) Kennicuttetal.2009),althoughwecautionthatitispossiblefor
+42:30 some 24µm emission to originate from dust heated by the dif-
+42:00 2.5 3 .0 0 .0 0.5 1 .0 fuseinterstellarradiationfieldfromolderstars(e.g.,Li&Draine
2001; Kennicutt et al. 2009). The 24µm band also includes a
+41:30 small amount of stellar emission; this was removed by multi-
+41:00 plying the 3.6µm image by 0.032 and then subtracting it from
+40:30 the24µmimage,asdescribedbyHelouetal.(2004).Whilewe
+40:00 SH857/SH600 ESF/ETotal (SH857/SH600) coulduse24µmdatacombinedwithHαorultravioletemission
to trace both obscured and unobscured star formation (as sug-
+42:30 gestedby,e.g.,Leroyetal.2008andKennicuttetal.2009),the
00)+42:00 2.0 2.5 0 .0 0.5 1 .0 publicly-availabledatahaveproblemsthatmakethemdifficultto
0
n (J2+41:30 includeinouranalysis6.Bendoetal.(2015)havedemonstrated
o thatusingthe24µmdataasastar-formationtracerforthisanal-
ati+41:00
Declin+40:30 yansids2w4ilµlmyiestladr-rfeosrumltastitohnattararecesrim(toilawrittohiunsainbgouatc5o%m)p,ossoitgeivHeαn
+40:00 SH1200/SH857 ESF/ETotal (SH1200/SH857) theissueswiththeultravioletandHα wewillusesolely24µm
+42:30 emission as our nominal star-formation tracer. We also use the
+42:00 1.0 2.0 0 .0 0.5 1 .0 combination of the 24µm and far-ultraviolet (FUV) data, us-
ingtheforegroundstar-subtractedGALEXdatafromFordetal.
+41:30
(2013)andViaeneetal.(2014),tocheckwhetherthishasasig-
+41:00 nificant impact on our results. We return to the results of using
+40:30 thisstar-formationtracerlaterinthissection.
+40:00 SH1870/SH1200 ESF/ETotal (SH1870/SH1200) InFig.7,weshowthesurfacebrightnessratiosforthebinned
+42:30 dlaattiaonassaarfeungcivtieonnionfTthaebl3e.62a.nOdn2ly4µthmedSata.S/Statisticrsatoiontshheowres-
+42:00 0.4 0.8 0 .0 0.5 1 .0 a stronger correlation with the 24µm em4i2s8s0ion3t0h0a0n the 3.6µm
+41:30 emission,whichimpliesthatonlythedustdominatingemission
at ≥4280GHz (≤70µm) is heated by star-forming regions. All
+41:00
the other ratios are more strongly correlated with 3.6µm emis-
+40:30
sion, which demonstrates that the total stellar populations, and
+40:00 SH3000/SH1870 ESF/ETotal (SH3000/SH1870) notsolelythestarformingregions,areheatingthedustobserved
+42:30 below3000GHz(100µm).
+42:00 0.2 0.4 0 .0 0.5 1 .0 As an additional assessment of the heating sources of
the dust observed at these frequencies, we fit the S /S ,
+41:30 545 353
S /S ,S /S ,S ,S ,andS /S ra-
857 545 1200 857 1874/1200 3000/1870 4280 3000
+41:00 tiodatatothe3.6and24µmdatausingtheequation
+40:30
(cid:32) (cid:33)
+40:00 SS4280/SH3000 ESF/ETotal (SS4280/SH3000) ln Sν1 =αln(I +A I )+A . (1)
00:50 00:40 00:50 00:40 Sν2 SFR 1 stars 2
Right Ascension (J2000)
Thisequation,derivedfromtheStefan-BoltzmannlawbyBendo
Fig.6.Left:colourratioplotswith5(cid:48) pixels.The5(cid:48)pixelsareslightly
et al. (2012a), relates the surface brightness ratios of the dust
biggerthanthebeamsize,suchthatthesignaltheycontainislargelyin-
withthedustheatingsources.TheratioS /S ontheleftside
dependent.Thecolourratiosforthelowerfrequencyplotsdonottrace ν1 ν2
the local features; rather, they depend on the galactocentric distance.
6 The publicly-available Hα data, such as from the Virginia Tech
Thehigherfrequencyplots–inparticularthehighestfrequencyone–
Spectral-Line Survey, Dennison et al. (1998), contain artefacts from
tracethelocalstructuremuchmore.Right:imagesoftherelativecontri-
incompletely-subtracted foreground stars, as well as incompletely-
butionofstarformationtoheatingthedusttracedbythespecifiedpair
subtractedcontinuumemissionfromthebulgeofM31,andarethere-
offrequencies.ThesemapswerederivedusingEq.(2)andtheparam-
foreunsuitableforouranalysis.Thepublicly-availableultravioletdata
etersinTable2.Toptobottom:S /S ,S /S ,S /S ,
P545 P353 P857 P545 H857 H600 containmultiplebrightforegroundstars,aswellasemissionfromolder
S /S ,S /S ,S /S ,andS /S .
H1200 H857 H1874 H1200 H3000 H1874 S4280 H3000 stellarpopulationswithinM31.
A28,page9of23