Odin observations of ammonia in the Sgr A +50 km/s Cloud and Circumnuclear Disk

0.14 MB·

by  Aa. Sandqvist

Astronomy&Astrophysicsmanuscriptno.nh3revrev (cid:13)c ESO2017
January11,2017
Odin observations of ammonia in the SgrA +50 kms−1 Cloud and
⋆
Circumnuclear Disk
Aa.Sandqvist1,Å.Hjalmarson2,U.Frisk3,S.Lundin4,L.Nordh1,M. Olberg5,andG. Olofsson1
1 StockholmObservatory,StockholmUniversity,AlbaNovaUniversityCenter,SE-10691Stockholm,Sweden
e-mail:[email protected]
2 Department of Earth and Space Sciences, Chalmers University of Technology, Onsala Space Observatory, SE-439 92 Onsala,
Sweden
7 3 OmnisysInstrumentsAB,SolnaStrandva¨g78,SE-17154Solna,Sweden
1 4 OHBSweden,POBox1269,SE-16429Kista,Sweden
0 5 OnsalaSpaceObservatory,ChalmersUniversityofTechnology,SE-43992Onsala,Sweden
2
Received<date>;accepted<date>
n
a
J ABSTRACT

0
Context.TheOdinsatelliteisnowintoitssixteenthyearofoperation,muchsurpassingitsdesignlifeoftwoyears.Oneofthesources
1
whichOdinhasobservedingreatdetailistheSgrAComplexinthecentreoftheMilkyWay.

Aims.TostudythepresenceofNH intheGalacticCentreandspiralarms.
] 3
A Methods.Recently,Odinhasmadecomplementaryobservationsofthe572GHzNH3 linetowardstheSgrA+50 kms−1 Cloudand
CircumnuclearDisk(CND).
G Results.Significant NH emissionhasbeenobservedinboththe+50 kms−1 CloudandtheCND.ClearNH absorptionhasalso
3 3
. beendetectedinmanyofthespiralarmfeaturesalongthelineofsightfromtheSuntothecoreofourGalaxy.
h Conclusions.Theverylargevelocitywidth(80 kms−1)oftheNH emissionassociatedwiththeshockregioninthesouthwesternpart
p 3
oftheCNDmaysuggestaformation/desorptionscenariosimilartothatofgas-phaseH Oinshocks/outflows.
- 2
o Keywords.Galaxy:centre–ISM:individualobjects:SgrA–ISM:molecules–ISM:clouds
r
t
s
a
[ 1. Introduction for the fractional abundance ratio of [O2/H2], averaged over a
1 ThecoreoftheMilkyWayGalaxyisaregionofgreatcomplex- e9t-aarlc.m20in08re)g-iosnig,nwifiascafnotuanmdotounbtesXo(fOH2)≤O,1C.2O×a1n0d−7C(iSwanedreqvdies-t
v ity containing a wide variety of physical environments. At the 2
tected in many regions of the SgrAComplex (Karlsson et al.
6 verycentreresidesafour-million-solar-masssupermassiveblack
2013).Unfortunately,therewereanumberofinstabilitiesinthe
3
hole, whose non-thermal radio continuum signature is called
5 SgrA∗. Orbiting around it, at a distance of one to a few pc, 572GHzNH3 receiverduringthose observationperiodswhich
2 with a velocity of about 100 km/s, is a molecular torus called preventedusfromobtainingresultsforthisNH3transitioninthe
0 coreregionoftheGalaxy,althoughsomeresultswereobtained
the Circumnuclear Disk (CND). The CND has a mass of 104
. forthespiralarmfeatures.Theseinstabilitieswereduetoaloss
1 to 105 M and a gas temperature of several hundred degrees.
⊙ of phase lock for this receiver early in the mission. However,
0 Beyondthis, thereexistsalargeMolecularBeltconsistingpre-
7 dominantly of two GMCs, called the +50 and the +20 kms−1 withappropriatecenteringoftheobservingfrequencyandmany
1 frequencycalibrations,performedusinganearbytelluricozone
clouds.BothGMCsaremassive,about5×105M ,withaden-
: ⊙ line,itispossibletocorrectforthislackofphaselock.Thisshort
v sity104 -105 cm−3, gastemperature80-100K,anddusttem-
papernowreportsonnewsuccessfulOdinNH observationsin
Xi perature20-30K(e.g.Sandqvistetal.2008).Generalreviews theSgrA+50 kms−1 Cloud,andinthesouthw3esternregionof
oftheGalacticCentrehavebeenpresentedbye.g.Mezgeretal.
theCNDwhichisacomplexshockedregioncontaininganum-
r
(1996)andMorris& Serabyn(1996),with an up-to-dateintro-
a berofinteractingcomponents(Karlssonetal.2015).
ductiontotheSgrAComplexgivenbyFerrie`re(2012).
TheOdinsatellite(Nordhetal.2003;Frisketal.2003)has
2. Observations
surveyedtheSgrAComplexinanumberofdifferentmolecules.
WhileitwasunsuccessfulinsearchingforO2-a3σupperlimit The Odin observations of the o-NH (1 − 0 ) line at
3 0 0
572.4981GHztowardstheSgrA+50 kms−1 CloudatJ2000.0
Sendoffprintrequeststo: 17h45m51s.7,−28◦59′09′′wereperformedinApril2015,withan
AageSandqvist,e-mail:[email protected]
ON-sourcetotalintegrationtimeof27.4hours.TheNH obser-
⋆ Odin is a Swedish-led satellite project funded jointly by the 3
vationstowardsthesouthwestern(SW)lobeoftheSgrACNDat
Swedish National Space Board (SNSB), the Canadian Space Agency
J2000.017h45m39s.7,−29◦01′18′′wereperformedinApril2016,
(CSA),theNationalTechnologyAgencyofFinland(Tekes),theCentre
withanON-sourcetotalintegrationtimeof26.7hours.Thesys-
National d’Etudes Spatiales(CNES),France and the European Space
Agency (ESA). The former Space division of the Swedish Space temtemperatureswere3300and3400K,respectively.Thehalf-
Corporation,todayOHBSweden,istheprimecontractor,alsorespon- powerbeamwidthofOdin at the NH3 frequencyis 2.′1 and the
sibleforOdinoperations. main beam efficiency is 0.89 (Frisk et al. 2003). The backend
1
Aa.Sandqvistetal.:OdinobservationsofNH towardstheSgrAComplex
3
spectrometerwasa1050MHzAOSwithachannelresolutionof
1MHz.FormoredetailsofOdinGalacticCentreobservations,
seeKarlssonetal.(2013).
3. ResultsandDiscussion
Ourtwo NH profiles,obtainedtowardsthe+50 kms−1 Cloud
3
andtheCNDSW,arepresentedinFigs.1and2,wheretheyhave
beensmoothedtoachannelresolutionof1.6 kms−1.Gaussian
analysiswasperformedonbothprofilesbyvisualestimationof
allthreeparameters(intensity,velocity,andvelocityhalfwidth)
foreachemissionandabsorptioncomponentandthefitwasthen
executedinanunbiasedmanner.TheGaussiananalysisfittings
arealsopresentedinFigs.1and2,andinTables1and2where
thelisteduncertaintiesareatthe1σlevel.Thedominantimpres-
sion in the figures is the clear emission features at positive ve-
locities,whichariseintheSgrAComplex(e.g.Sandqvist1989).
Mostofthenumerousweakerabsorptionfeaturesatnegativeve-
locities are caused by the various Galactic spiral arm compo-
Fig.1.Odinobservationsofthe572GHzNH linetowardsthe
nentsalongthelineofsightfromtheSuntotheGalacticCentre 3
SgrA+50 kms−1 Cloud(blackprofile);thechannelresolution
and are identifiable by their radial velocities (Sandqvist et al.
is1.6 kms−1.TheredprofileistheresultoftheGaussiananal-
2015). A few of these features are near the limit of detectabil-
ysis.
ity,butcorrespondtoclearlyidentifiablesignaturesobservedin
e.g.OH,H OandCOlinespresentedbyKarlssonetal.(2013).
2
The rms noise levels in the two profiles are 9 and 11 mK, re-
spectively,sotheintensitiesoftheweakabsorptionfeaturesare
predominantlygreaterthan3σ.
The very fact that we are only observing ground-state am-
moniaabsorptionfeaturesatnegativevelocities,andseenovisi-
bleemission,againstveryweakcontinuumbackgroundsources
(only210 and 140mK) implies thatthe excitationtemperature
(T )ofthisgasmustbeclosetothetemperatureofthecosmic
ex
microwavebackground(T =2.725K),whichmeansthatthe
CMB
dominantpartoftheammoniapopulationisresidinginthelow-
est state. This also means that the absorbing gas regions must
have a density severalordersof magnitudelower than the crit-
ical density of the ammonialine (5×107 cm−3), which is also
consistentwiththeirlowH columndensities(seeTable3)and
2
lowvisualextinction.FromtheGaussiananalysisoftheabsorp-
tion features,we cannow obtaincolumndensities, N(NH ), of
3
ortho-NH inamannersimilartoKarlssonetal.(2013):
3
N(NH )=3.69×1012τ (NH )∆V (1)
3 o 3 FWHM
Fig.2.Odinobservationsofthe572GHzNH linetowardsthe
3
where∆V isthelinefullwidthathalfmaximumandτ is southwesternpartof theSgrA CircumnuclearDisk (blackpro-
FWHM o
thecentralopticaldepthoftheGaussianfittedtotheabsorption file);thechannelresolutionis1.6 kms−1.Theredprofileisthe
featureand(forT =T )canbecalculatedas resultoftheGaussiananalysis.
ex CMB
T∗
τ =−ln 1+ A . (2) of RADEX 1 (van der Tak et al. 2007). Assuming a gas tem-
o
" Tcont# perature of T = 80 K and a gas density of n = 104 cm−3
H2
(Walmsleyetal.1986;Sandqvistetal.2008)andalinewidthof
InEq.(2), TA∗ istheantennatemperatureoftheabsorptionline 39 kms−1,wegetN(NH3)=1×1016cm−2.Thisiscomparable
intensity and is a negative number and Tcont is the background tothevalueof2×1016deducedbyMills&Morris(2013)from
continuumtemperature,which is 210 and 140 mK for the +50
their observations of 14 higher NH transitions in this cloud,
kms−1 Cloud and the CND SW, respectively (Karlsson et al. 3
where they estimate that about 10% of the NH column origi-
3
2013).TheresultsoftheGaussiananalysisoftheabsorptionfea-
nates from a high temperature (≈ 400 K) cloud component. It
turesaresummarizedinTables1and2. is,however,largerthanthevalueof3×1015 cm−2 obtainedby
Themajoremissionfeatureinthe+50 kms−1 profile,at47
Herrnstein&Ho(2005)fromtheirobservationsofthe23GHz
kms−1 originatesinthe+50 kms−1 Cloud.TheGaussiananal- inversion line. For the +50 kms−1 Cloud, our column density
ysis yielded a T∗ = 266 mK and a line width of 39 kms−1.
A valuethenyieldsano-NH3abundancewithrespecttohydrogen,
Correctingforthebeamefficiency(0.89)weobtainamainbeam X(NH )= N(NH )/N(H ),of4×10−8,ifweassumeamolecu-
3 3 2
brightness temperature of 298 mK. In order to obtain an NH
3
columndensityforthecloud,wehaveusedtheon-lineversion 1 http://var.sron.nl/radex/radex.php
2
Aa.Sandqvistetal.:OdinobservationsofNH towardstheSgrAComplex
3
Table1.Gaussianfitstothe+50 kms−1 CloudNH profile.
3
V T∗ ∆V τ N(NH ) Source
A o 3
(kms−1) (mK) (kms−1) (cm−2)
−230 −36±8 7 0.19±0.05 (5±1)×1012 HighNegativeVelocityGas(HNVG)
−217 −41±7 9 0.22±0.04 (7±1)×1012 HNVG
−189 −29±8 7 0.15±0.04 (4±1)×1012 HNVG
−136 −39±8 11 0.21±0.04 (8±2)×1012 ExpandingMolecularRing(EMR)(nearside)
−111 −53±3 51 0.29±0.02 (55±3)×1012 EMR(nearside)
−48 −40±6 11 0.21±0.03 (9±1)×1012 3-kpcarm
−29 −19±9 7 0.10±0.04 (2±1)×1012 −30 kms−1arm
−3 −45±39 12 0.24±0.20 (10±9)×1012 Local/Sgrarm
47 266±29 39 ... (1±0.1)×1016 +50 kms−1Clouda
49 −96±28 18 0.23±0.07 (15±5)×1012 +50 kms−1Cloudb
186 28±8 7 ... ... ...
aFromRADEXanalysis,seetext
bSelf-absorption,seetext
larhydrogencolumndensityofN(H )=2.4×1023cm−2(Lis&
2
Carlstrom1994),whichwasdeterminedfromobservationsofa
totaldustcolumn.UsinginsteadthevalueofN(H )=1.6×1023
2
cm−2, which was determined from SEST observations of the
C18O(1-0)and(2-1)linesatthisposition,(Karlssonetal.2013),
wegetano-NH abundanceof6.3×10−8.
3
Theemissionprofileofthe+50 kms−1Cloudissignificantly
affected by self-absorption, as can be seen in Fig. 1. Also, the
Gaussian analysis resulted in a clear self-absorption feature at
49 kms−1(seeTable1)withanintensityof−96mK.Thecorre-
spondingopticaldepthandcolumndensityforthisfeaturewere
thendetermined,usingEqs.(2)and(1),assumingthattheself-
absorbing region of the cloud is on the near side of the +50
kms−1 Cloud and thus absorbing both the background contin-
uumof210mKandthebackgroundlineemissionof265mK,a
totalof475mK.
There are two overlapping emission features in the CND
SW profile seen in Fig. 2 and analysed in Table 2, namely
at 33 and 68 kms−1 with line widths of 27 and 81 kms−1,
respectively. In both of these features there are signs of self- Fig.3. Comparison of the Odin observations of the 572 GHz
absorptionat34and76 kms−1,respectively.Theopticaldepths NH3 line (magenta) and the 548 GHz H128O line (black) to-
of these self-absorptions will depend on the relative positions wards the southwestern part of the SgrA Circumnuclear Disk;
of the components along the line of sight. For the 34 kms−1 thechannelresolutionsare1.6and3 kms−1,respectively.
self-absorption we determine the optical depth, assuming that
it and its emission companion at 33 kms−1 are in front of the
emission componentat 76 kms−1, and that the self-absorption namelyagastemperatureof80Kanddensityof104cm−3.The
is on the near side of the 33 kms−1 component.Thusthe self- antennatemperatureof139mKconvertstoamainbeamtemper-
absorbingregionabsorbsthecontinuumemissionof140Kand atureof156mK;thelinewidthis27 kms−1.ThebestRADEX
both the 33 kms−1 emission and the blue overlappingwing of fitthenyieldsanNH columndensityof4×1015cm−2.
3
the68 kms−1emissionfeature,whichamountstoatotalsumof
PerhapsthemostinterestingfeatureisthebroadNH emis-
3
140+138+28=306mK,avalueweuseforTcontinEq.(2).For sion feature centered at 68 kms−1 with a line width of 81
the76 kms−1self-absorption,weassumethattheself-absorbing kms−1. Odin has detected a comparable very broad H18O ab-
regionisonthenearsideofthe68 kms−1 emissionregionand 2
sorption line at the same position (Karlsson et al. 2015) - see
both lie either in front of, or behind, the 140 K continuum re-
Fig. 3. The antennatemperatureof this NH feature is 46 mK,
3
gion.Thusarangefortheopticaldepthisobtained,usingeither
whichimpliesa mainbeamtemperatureof52mK.Applyinga
the red-shifted emission of the 68 kms−1 region at the appro-
RADEXstudy,andassumingakinetictemperatureof200Kand
priatevelocityof76 kms−1 whoseemissionvalueis45mK,or densityof4×104cm−3(Requena-Torresetal.2012),weobtain
140+45=185mK,thevaluesofwhichweuseforTcontinEq. a column density of o-NH of 7×1014 cm−2. The H column
3 2
(2). density for this region is 4×1022 cm−2 (Karlsson et al. 2015),
TherearetwomajoremissionfeaturesintheCNDSWpro- which would then yield an abundanceratio for o-NH with re-
3
fileandwehaveperformedaRADEXanalysisofboth.Firstwe spect to H of X[NH ] = 1.75× 10−8. Karlsson et al. (2015)
2 3
assume that the 33 kms−1 feature originates in the Molecular obtainedastrikinglyhighH Oabundanceof1.4×10−6 inthis
2
beltandwecanthusapplyitsphysicalpropertiestotheanalysis, shockregion.
3
Aa.Sandqvistetal.:OdinobservationsofNH towardstheSgrAComplex
3
Table2.GaussianfitstotheCNDSWNH profile.
3
V T∗ ∆V τ N(NH ) Source
A o 3
(kms−1) (mK) (kms−1) (cm−2)
−220 −43±8 7 0.36±0.08 (9±2)×1012 HNVG
−195 −34±5 17 0.28±0.05 (17±3)×1012 HNVG
−134 −49±12 4 0.43±0.13 (6±2)×1012 EMR(nearside)
−113 −31±6 15 0.25±0.05 (14±3)×1012 EMR(nearside)
−85 −44±8 8 0.38±0.08 (11±2)×1012 CND
−51 −43±7 9 0.37±0.07 (13±2)×1012 3-kpcarm
−22 −19±8 8 0.14±0.06 (4±2)×1012 −30 kms−1arm
3 −44±11 4 0.38±0.11 (6±2)×1012 Local/Sgrarm
33 139±47 27 ... (4±1)×1015 MolecularBelta
34 −64±50 13 0.23±0.19 (11±9)×1012 MolecularBeltb
68 46±6 81 ... (7±1)×1014 Shockregiona
76 −31±8 9 0.19±0.05to1.18±0.6 ((6±2)to(40±20))×1012 ??b
aFromRADEXanalysis,seetext
bSelf-absorption,seetext
Table3.NH andH Oabundancecomparisons
3 2
Source N(NH ) N(H ) X(NH ) X(H O)
3 2 3 2
(cm−2) (cm−2) (×10−9) (×10−9)
SgrAComplexComponents
+50 kms−1Cloud (1±0.1)×1016 1.6×1023a 63 40a
+50 kms−1Cloudredwing ... ... ... ≥1000a
MolecularBeltatCNDSW (4±1)×1015 ¡1.0×1023b ¿40 ...
ShockregionatCNDSW (7±1)×1014 4×1022c 18 1400c
+20 kms−1Cloud ... 1×1023b ... 20a
+20 kms−1Cloudredwing ... ... ... ≥800a
GalacticComponents
EMR (30±2)×1012d 2.5×1021e 12 ...
3-kpcArm (11±2)×1012d 4.0×1021 f 3 ¿3 f
−30 kms−1Arm (3±1)×1012d 1.9×1021 f 2 30a
Local/SgrArm (8±5)×1012d 6.4×1021 f 1 ¿3 f
a Karlssonetal.(2013);b valueforthe+20 kms−1 CloudfromKarlssonetal.(2013);c Karlssonetal.(2015);d meanoftherespectivevalues
inTables1and2;e deducedfromthe13CO(1−0)profileinFig.1(a)byBallyetal.(1988)andtheconversionfactorgivenbySofue(1995); f
Sandqvistetal.(2003)
A summary of the derived NH abundances for different magnitudesofvisualextinction2 andratherlowclouddensities
3
componentsin the SgrAComplexandGalactic featuresis pre- (socalledtranslucentclouds).However,itseemsthatachemical
sented in Table 3, togetherwith H O abundances,where avail- modelcombininggas-phaseand grainsurface reactionscan do
2
able.Itappearsthattheortho-ammoniaabundancesof(2−6)× thejobatatemperatureof30-50Kandadensityof≈103cm−3,
10−8 observed by us in the SgrA molecular cloud regions(the butonlyiftherelevantspeciesreallyaredesorbedasaresultof
+50 kms−1 Cloud, as well as the Molecular Belt) can be ac- exothermicsurfacereactions(Perssonetal.2014;theirFig.C3,
commodated in current chemical models for dense gas clouds and the translucent gas part of Fig. C2). A remaining concern
(PineaudesForetsetal.1990),especiallysoifgrainsurfacere- maybethatthesimultaneousmodelabundancesofNHandNH
2
actionsandvarioussubsequentdesorptionprocessesareconsid- are not consistentwith those observed by Persson et al. (2010,
ered(Perssonetal.2014;densecloudpartoftheirFig.C2).Our 2012).
detectionsofNH atthemuchlowerabundanceof(1−3)×10−9 As discussed in some detail by Karlsson et al. (2013), the
3
in spiral arm clouds are not so easy to understand in terms of gas-phase water abundances determined for the Sgr A +50
chemical models. Similarly low abundanceshave been derived kms−1 and +20 kms−1 molecular clouds are considerably en-
by Persson et al. (2010, 2012) from Herschel observations of hancedcomparedtothesituationincoldcloudcoreswherethe
spiral arm NH absorption regions in the directions of W49N watermainlyresidesasiceonthecoldgrainsurfaces.Similarly
3
and G10.6−0.4 (W31C), and by Wirstro¨m et al. (2010) from high or even more enhanced water abundanceshave been esti-
Odin observations in the direction of SgrB2. The NH abun- mated in lower density spiral arm clouds observed in absorp-
3
dancesindiffuseortranslucentclouds,estimatedbyLisztetal. tionagainstSgrAbyKarlssonetal.(2013),andagainstSgrB2
(2006) from23 GHz inversionline absorptionagainstcompact using Odin by Wirstrm et al. (2010) and using Herschel Space
extragalactic sources, are also similarly low. The low H col- ObservatoryalsobyLisetal.(2010).Theelevatedgasphasewa-
2
umndensitiesofthespiralarmcloudscorrespondtoonlyafew
2 usingN(H )/A ≈0.94×1021cm−2mag−1,Bohlinetal.(1978)
2 v
4
Aa.Sandqvistetal.:OdinobservationsofNH towardstheSgrAComplex
3
ter abundancesdefinitely require desorptionof water ice, most Requena-Torres,M.A.,Gu¨sten,R.,Weiss,A.,etal.2012,A&A,542,L21
likely handled by PDR modeling including grain surface reac- Sandqvist,Aa.1989,A&A,223,293
tions(cf.Hollenbachetal.2009)Thereleaseofwatermolecules Sandqvist,Aa.,Bergman,P.,Black,J.H.,etal.2003,A&A,402,L63
Sandqvist,Aa.,Larsson,B.,Hjalmarson,Å.,etal.2008,A&A,482,849
formedon colder grainsurfacesmay also explain the ortho-to-
Sandqvist,Aa.,Larsson,B.,Hjalmarson,Å.,etal.2015,A&A,584,A118
paraH2O abundanceratio lowerthan3 verylikely observedin Sofue,Y.1995,PASJ,47,527
thespiralarmsagainstSgrB2(Lisetal.2010). vanderTak,F.F.S.,Black,J.H.,Scho¨ier,F.L.,Jansen,D.J.,&vanDishoeck,
The very high gas-phase water abundance determined for E.F.2007,A&A,468,627
Walmsley,C.M.,Gu¨sten,R.,Angerhofer,P.,&Mundy,L.1986,A&A,155,129
the shock regionat CND SW by Karlsson et al. (2015) is sim-
Wirstro¨m,E.S.,Bergman,P.,Black,J.H.,etal.2010,A&A,522,A19
ilar to that found in the red-ward high-velocity wings of the
Sgr A molecular clouds, and likely results from shock heat-
ing causing release of pre-existing grain surface water, possi-
bly combined with high temperature shock chemistry. (cf. dis-
cussion in Karlsson et al. 2013). The very large velocitywidth
(80 kms−1) of the NH emission associatedwith the shockre-
3
gion at the CND SW (Fig. 2 and Table 2) may suggest a for-
mation/desorption scenario similar to that of gas-phase H O
2
in shocks/outflows, where the actual ammonia abundance is
achievableinchemicalmodelsforquiescentmolecularcloud(as
discussedpreviously).
4. Conclusions
We have successfully observed NH emission from the SgrA
3
+50 kms−1 Cloud and the CND using the Odin satellite. The
verylargevelocitywidth(80 kms−1)oftheNH emissionasso-
3
ciatedwiththeshockregioninthesouthwesternpartoftheCND
may suggest a formation/desorptionscenario similar to that of
gas-phase H O in shocks/outflows. In addition, clear NH ab-
2 3
sorptionhasalsobeenobservedinthespiralarmfeaturesalong
thelineofsighttotheGalacticCentre.
The high quality Odin NH spectra, observed in 2015 and
3
2016, presented in this short paper, together with the H O and
2
H18OspectraofsimilarlyhighqualityobservedbyOdintowards
2
CometC/2014Q2(Lovejoy)inJanuary-February2015(Biver
et al. 2016) were obtained 14-15 years after the launch of the
Odin satellite. Thisdemonstratesto ourdelight(andalso some
surprise) that it is possible to build a comparativelycheap, but
still rather complicated, satellite observatory(with five tunable
heterodynereceiversandamechanicalcoolingmachine)which
can remain in high quality operationfor 15 years or more. We
shouldherenotethatthe Odinsatellite observatorysincemany
yearshasbeenoperatedpracticallyfull-timeinglobalmonitor-
ingoftheterrestrialatmosphere(theOdinaeronomymode).
References
Bally,J.,Stark,A.A.,Wilson,R.W.,&Henkel,C.1988,ApJ,324,223
Biver,N.,Moreno,R.,Bockele´e-Morvan,D.,etal.2016,A&A,589,A78
Bohlin,R.C.,Savage,B.D.,&Drake,J.F.1978,ApJ,224,132
Ferrie`re,K.2012,A&A,540,A50
Frisk,U.,Hagstro¨m,M.,Ala-Laurinaho,J.,etal.2003,A&A,402,L27
Herrnstein,R.M.,&Ho,P.T.P.2005,ApJ,620,287
Hollenbach,D.,Kaufman,M.J.,Bergin,E.A.,&Melnick,G.J.2009,ApJ,690,
1497
Karlsson,R.,Sandqvist,Aa.,Hjalmarson,Å.,etal.2013,A&A,554,A141
Karlsson,R.,Sandqvist,Aa.,Fathi,K.,&Martin,S.2015,A&A,582,A118
Lis,D.C.,&Carlstrom,J.E.1994,ApJ,424,189
Lis,D.C.,Phillips,T.G.,GoldsmithP.F.,etal.2010,A&A521,L26
Liszt,H.S.,Lucas,R.,&Pety,J.2006,A&A,448,253
Mezger,P.G.,Duschl,W.J.,&Zylka,R.1996,A&AR,7,289
Mills,E.A.C.,&Morris,M.R.2013,ApJ,772,105
Morris,M.,&Serabyn,E.1996,ARA&A,34,645
Nordh,H.L.,vonSche´ele,F.,Frisk,U.,etal.2003,A&A,402,L21
Persson,C.M.,Black,J.H.,Cernicharo,J.,etal.2010,A&A,521,L45
Persson,C.M.,DeLuca,M.,Mookerjea,B.,etal.2012,A&A,543,A145
Persson,C.M.,Hajigholi,M.,Hassel,G.E.,etal.2014,A&A,567,A130
PineaudesForets,G.,Roueff,E.,&Flower,D.R.1990,MNRAS242,512
5