The NH$_2$D hyperfine structure revealed by astrophysical observations

0.13 MB·

by  F. Daniel

Astronomy&Astrophysicsmanuscriptno.NH2D˙spectro-7 (cid:13)c ESO2016
January5,2016
LettertotheEditor
The NH D hyperfine structure revealed by astrophysical
2
⋆
observations
F.Daniel1,2,L.H.Coudert3,A.Punanova4,J.Harju4,5,A.Faure1,2,E.Roueff6,O.Sipila¨4,P.Caselli4,R.Gu¨sten7,
A.Pon4,8,andJ.E.Pineda4
1 Univ.GrenobleAlpes,IPAG,F-38000Grenoble,France
2 CNRS,IPAG,F-38000Grenoble,France
6 3 LISA,UMR7583CNRS-UniversitesParisEstCre´teiletParisDiderot,61AvenueduGe´ne´raldeGaulle,F-94010Cre´teil,France
1 4 MaxPlanckInstituteforExtraterrestrialPhysics(MPE),Giessenbachstr.1,85748Garching,Germany
0 5 DepartmentofPhysics,P.O.Box64,00014UniversityofHelsinki,Finland
2 6 LERMA,ObservatoiredeParis,PSLResearchUniversity,CNRS,UMR8112,F-75014,Paris,France
7 MaxPlanckInstituteforRadioastronomie,AufdemHu¨gel69,53121Bonn,Germany
n 8 DepartmentofPhysicsandAstronomy,TheUniversityofWesternOntario,London,Canada,N6A3K7
a
J Received;accepted

2
ABSTRACT

]
A Context.The1 -1 linesoforthoandpara–NH D(o/p-NH D),respectivelyat86and110GHz,arecommonlyobservedtoprovide
11 01 2 2
G constraintsonthedeuteriumfractionationintheinterstellarmedium.Incoldregions,thehyperfinestructureduetothenitrogen(14N)
nucleusisresolved.Todate,thissplittingistheonlyonewhichistakenintoaccountintheNH Dcolumndensityestimates.
. 2
h Aims.We investigate how the inclusion of the hyperfine splitting caused by the deuterium (D) nucleus affects the analysis of the
p rotationallinesofNH2D.
- Methods.Wepresent30mIRAMobservationsoftheabovementionedlines,aswellasAPEXo/p-NH2Dobservationsofthe101-000
o linesat 333 GHz. The hyperfine patternsof theobserved lineswerecalculated taking into account thesplitting induced by theD
r nucleus.TheanalysisthenreliesonlineliststhateitherneglectortakeintoaccountthesplittinginducedbytheDnucleus.
t
s Results.Thehyperfinespectraarefirstanalyzedwithalinelistthatonlyincludesthehyperfinesplittingduetothe14Nnucleus.We
a findinconsistenciesbetweenthelinewidthsofthe1 -0 and1 -1 lines,thelatterbeinglargerbyafactorof∼1.6±0.3.Suchalarge
01 00 11 01
[ differenceisunexpectedgiventhetwosetsoflinesarelikelytooriginatefromthesameregion.Wenextemployanewlycomputed

linelistfortheo/p-NH Dtransitions,wherethehyperfinestructureinducedbybothnitrogenanddeuteriumnucleiisincluded.With
1 2
thisnewlinelist,theanalysisofthepreviousspectraleadstolinewidthswhicharecompatible.
v
Conclusions.Neglecting thehyperfine structure owing to D leads tooverestimate the linewidths of the o/p-NH D lines at 3mm.
2 2
The error for a cold molecular core is about 50%. This error propagates directly to the column density estimate. It is therefore
6
recommendedtotakeintoaccountthehyperfinesplittingscausedbyboththe14NandDnucleiinanyanalysisrelyingontheselines.
1
0 Keywords.Astrochemistry—Radiativetransfer—ISM:molecules—ISM:abundances
0
.
1
0 1. Introduction CDMS2), the line lists of the NH D microwave spectra are
2
6
based on the measurements by DeLucia&Helminger (1975),
1 Thefirstfirmidentificationofsinglydeuteratedinterstellaram- Cohen&Pickett(1982)andFusinaetal.(1988).Thesedatasets
:
v monia, NH2D, was reported by Olbergetal. (1985) towards onlyconsiderthenitrogenquadrupolehyperfinestructure.Inthis
i three molecular clouds at 86 and 110 GHz, the frequencies of letter,anewtheoreticalanalysisoftheNH Dhyperfinestructure
X 2
the JKa,Kc=111-101linesofo/p-NH2D,respectively.Theidentifi- is presented which takes into account the nitrogen quadrupole
r cationwasunambiguousthankstothenarrowlinewidthswhich interaction, the deuteron quadrupole interaction and the nitro-
a
allowedtoresolvethehyperfinesplittingduetothenitrogennu- gen spin-rotation interaction. It is based on the measurements
cleus. These two lines have since been detected in a variety of by Kukolich (1968), Cohen&Pickett (1982), and Fusinaetal.
cold astronomical sources and they are routinely employed to (1988). We show thatthe inclusionof the three couplingterms
derivethe(D/H)ratioinammonia(seee.g.Roueffetal.2005). intheanalysisprovidesasimpleexplanationfortheoriginofdif-
Thisratioisasensitivetracerofthephysicalconditionsandpro- ferentlinewidthsinthe1 -0 and1 -1 linesofNH D,asob-
01 00 11 01 2
videsstrongconstraintstodeuteriumfractionationmodels. servedrecentlytowardstheprestellarcoreH-MM1.Historically,
In the previous studies by Coudert&Roueff (2006, the first astronomicalobservationresolvingthe hyperfinesplit-
2009) and in the current spectroscopic databases (JPL1 and ting owing to the D quadrupole coupling was presented by
Caselli&Dore(2005)forDCO+.
Thepaperisorganizedasfollows.InSect.2,wedescribethe
⋆ BasedonobservationscarriedoutwiththeIRAM30mTelescope. IRAMandAPEXobservationsoftheo/p–NH Dlinesdetected
2
IRAM is supported by INSU/CNRS (France), MPG (Germany) and towards H-MM1. Sect. 3 gives the details of the spectroscopy
IGN(Spain).
1 spec.jpl.nasa.gov 2 http://www.astro.uni-koeln.de/cdms/
1
Danieletal.:NH Dhyperfinestructure
2
calculations. In Sect. 4, we discuss the impact of the new line consistsofmicrowaveandfarinfraredtransitionsinvolvingthe
list on the derivation of radiative transition parameters and we twoinversionsubstatesofthegroundvibrationalstate.Thisfirst
concludeinSect.5. setincludes174microwavetransitions(Cohen&Pickett1982)
and 297 far infrared transitions (Fusinaetal. 1988) for which
no hyperfine structure is resolved. The second set involves the
2. Observations
76microwavetransitionslistedinTableIXofCohen&Pickett
(1982) forwhich the nitrogenatom quadrupolecouplingstruc-
Thetargetof thepresentNH D observationsis thedense,star-
2
ture is resolved. The last data set comprises the 21 hyperfine
less core H-MM1 lying in the eastern part of Lynds 1688 in
componentsmeasured by Kukolich (1968) for the para 4 -4
Ophiuchus (Johnstoneetal. 2004; Pariseetal. 2011). The po- 14 04
sition observed,α=16h27m59s.0, δ=−24◦33′33′′, was chosen as rotation-inversiontransitionat25023.8MHz.Forthislastdata
set, the quadrupolecoupling structure due to both the nitrogen
the H column density peak, as derived from pipeline-reduced
2
anddeuteriumatomsisresolved.
Herschel far-infrared maps. These were downloaded from the
HerschelScienceArchive3.
Rotation-inversionenergieswerecomputedwiththehelpof
asemi-rigidrotatorHamiltonianforbothinversionsubstatesand
2.1.APEX12m a second order Coriolis coupling term between these substates
(Cohen&Pickett1982).Molecule-fixedcomponentsofthehy-
The ground-state 101–000 transitions of o/p-NH2D at ∼333 perfinecouplingtensorsarewrittenusingtheIAMaxissystem
GHz were observed using the upgraded version of the First ofthisreference.Quadrupoleandmagneticspin-rotationhyper-
Light APEX4 Submillimeter Heterodyne instrument (FLASH; fine couplings were taken into account leading to a hyperfine
Heymincketal. 2006) on APEX (Gu¨stenetal. 2006). The two Hamiltoniandependingonfourcouplingconstants:two forthe
lines separated by about 40 MHz were covered by one of the quadrupole coupling of the nitrogen and deuterium atoms and
MPIfR Fast Fourier Transform Spectrometers (XFTTS) con- twoforthemagneticspin-rotationcouplingofthesameatoms.
nected to the 0.8 mm receiver. The original spectral resolution Using equationssimilar to Eqs. (3) and (17) of Thaddeusetal.
at 333 GHz is about 35 ms−1; the spectra shown in this paper (1964), these four constants can be expressed as diagonal ma-
areHanningsmoothedtoaresolutionof69ms−1(76kHz).The trix elements of four operators involving four rank two hy-
APEX beam size (FWHM) is ∼20′′ at 333 GHz. The observa- perfine coupling tensors: the zero trace χN and χD, describing
tions were carried out between 29 and 31 May, 2015 in stable quadrupole coupling of the nitrogen and deuterium atoms, re-
and fairly good weather conditions (PWV 0.7-1.2 mm), using spectively,andCN andCD,correspondingtothemagneticspin-
positionswitchingforskysubtraction.Theaveragesystemtem- rotationcouplingofthesameatoms.
peratureat333 GHz was 260 K. The resulting RMS noise ata
resolutionof69ms−1is0.017KontheT∗ scale. Hyperfine energies were calculated taking the coupling
A
scheme: J + I = F and F + I = F where J is the rota-
N 1 1 D
tionalangularmomentum,andI and I are angularmomenta
N D
2.2.IRAM30m
of the nitrogen and deuterium nuclei, respectively. The corre-
The1 -1 linesofo/p-NH Dat∼86and∼110GHzwereob- spondingcoupled basis set functions|J,IN;F1,ID;F,MFi were
11 01 2
servedattheIRAM30mtelescopeusingtheEMIR090receiver5 usedtosetupthehyperfineHamiltonianmatrix.Matrixelements
weretakenfromThaddeusetal.(1964).
andtheVESPAautocorrelator.Thespectralresolutionofthisin-
strument, 20 kHz, is 68 ms−1 at 86 GHz and 53 ms−1 at 110
Theinversion-rotationtransitionsbelongingtothefirstdata
GHz. At these frequencies,the beam sizes of the telescope are
set were analyzed first allowing us to determine a set of spec-
29′′ and23′′,respectively.Theobservationswereperformedon
troscopic parameters analogous to that listed in Table III of
July5,2015,inacceptableweatherconditions(PWV8–10mm).
Cohen&Pickett(1982).Componentsofthefourhyperfinecou-
The observing mode was position switching. The integration
pling tensors were then fitted to the frequencies of transitions
times for the o/p-NH D lines were 15 and 22 minutes, and the
2 belongingtothesecondandthirddatasetsevaluatingthehyper-
average system temperatures at 86 and 110 GHz were 170 K
fine couplingconstantswith theeigenfunctionsretrievedin the
and250K, respectively.TheresultingRMS noise levelsof the
first analysis. For lines belongingto the first (second) data set,
o/p–NH Dspectrawere0.07Kand0.08KontheT∗ scale.
2 A an experimentaluncertaintyvalue of 0.1MHz (1 kHz) was as-
sumedandcompareswellwitharootmeansquaredeviationof
the observed minus calculated residual of 8.4 kHz (1.5 kHz).
3. Hyperfinepatterncalculations
Table 1 reports the values obtained for fitted components of
JustlikeinNH3,thenitrogennucleusinNH2Dcantunnelacross the hyperfine couplingtensors. These componentsare given in
theplanemadebytheHandDnuclei.Eachrotationaltransition theaxissystemofCohen&Pickett(1982).Duetothefactthat
isthussplitintoadoubletbythisinversionmotion.Theresulting the rotation-inversion transition measured by Kukolich (1968)
rotation-inversionlevelsareeithersymmetricoranti-symmetric is characterized by ∆J = 0, calculated hyperfine frequencies
undertheexchangeofthetwoprotonsandtheywilleithercor- mainly depend on the sum χD +χD. For this reason, only χD
xx yy yy
respondtoparaorortholevel. wasvariedintheanalysisandχD wasconstrainedtoavaluere-
xx
Hyperfine patterns were calculated from a fit of high-
trievedfromthedeuteroncouplingconstant(eq Q) reportedby
ξ D
resolution data that can be divided into three sets. The first set Kukolich (1968) and using the fact that angle between the ND
3 www.cosmos.esa.int/web/herschel/science-archive bondandthez-axis(Cohen&Pickett1982)is78.98◦.Magnetic
4 This publication is based on data acquired with the Atacama spin-rotation coupling effects could only be retrieved for the
Pathfinder Experiment (APEX). APEX is a collaboration between nitrogen atom. As all three diagonal components of the corre-
the Max-Planck-Institut fur Radioastronomie, the European Southern sponding tensor could not be determined separately, they were
Observatory,andtheOnsalaSpaceObservatory constrained to be equal, as for the normal species (Kukolich
5 www.iram.es/IRAMES/mainWiki/EmirforAstronomers 1967).
2
Danieletal.:NH Dhyperfinestructure
2
Table1:Hyperfineparameters.
Parameter Value Parameter Value
χN/MHz 1.906(84) χD/kHz 274.67a
xx xx
χN/MHz 2.040(98) χD/kHz −114.9(15)
yy yy
CN/kHz 4.993(100)b
xx
Notes.Numbersinparenthesesareonestandarddeviationinthesame
unitsasthelastdigit.(a)Constrainedvalue.(b)Thethreediagonalcom-
ponentsofthistensorwereconstrainedtobeequal.
Theresultsoftheaboveanalysiswereusedtopredicthyper-
fine patterns of other rotation-inversion transitions, evaluating
hyperfineintensitieswithEq.(29)ofThaddeusetal.(1964).We
note that hyperfineeffects due to the two hydrogenatomsmay
beimportantfororthotransitions.Theseeffectswereevaluated Fig.1:Theleftcolumnshowsthefitoftheo-NH Dlinesat86
2
taking into accountthe spin-spin coupling,calculated from the GHz (upperpanel)and333GHz (lower panel)with the hyper-
equilibriumstructureofthemolecule,andthespin-rotationcou- finestructureduetotheDnucleus.Therightcolumnshowsthe
pling, evaluatingthe coupling constant for J=1 from Kukolich fitofthep-NH Dlinesat110GHz(upperpanel)and333GHz
2
(1967) and Garveyetal. (1976). Inclusion of these additional (lowerpanel).Notethatfitsofsimilarqualityareobtainedifthe
hyperfinecouplingswerefoundtoaffectmarginallythelinepa- hyperfinestructureduetoDisomitted.
rametersand in particular,the linewidth is atmostaltered bya
fewpercents(seeSection4).Therefore,theseeffectshavebeen
neglectedinthefollowing.
the 14N nucleus. Doing so, the HFS method applied to the H-
MM1observationsdescribedintheprevioussectionleadtode-
4. Modelling rivethefollowingparametersforthep-NH2Dlines:
The HFS method of the CLASS software6 allows to quickly – 110GHz:τtot =2.2±0.6and∆v=0.33±0.02kms−1
analyse hyperfinespectra. In particular,it givessome basic pa- – 333GHz:τtot =2.0±0.8and∆v=0.20±0.02kms−1
rametersofthelinesasthewidth∆voftheindividualtransitions
andfortheo-NH Dlines:
or the opacity summed over all the hyperfine components τ . 2
tot
The total column density of a molecule can be inferred from – 86GHz:τ =5.1±0.3and∆v=0.37±0.01kms−1
tot
theparametersobtainedwiththeHFSfitusingrelation(seee.g. – 333GHz:τ =2.2±0.3and∆v=0.24±0.01kms−1
tot
Bacmannetal.2010;Mangum&Shirley2015)
Differenttransitionsofamoleculeshouldhavesimilarintrinsic
8πν3 Q(T ) exp Eu linewidths if they originate from the same region of the cloud.
N = c3 guAeuxl exp(cid:16)kb(cid:16)hTkνebxT(cid:17)ex−(cid:17)1Z τuldv. (1) Ifilinlnleeasdst:wroiiftphthhdyeisffisceoarulernsctoecurhricatiercbsa,olutdhreissndscieotninesdsitiaytiroeonrfotiresmmnepodetrinnateducireffesesgrareranidltyipefanurttsls-,
As explained in Bacmannetal. (2010), the integration of the
ofthecloud.Thefactor∼1.6±0.3foundbetweenthelinewidths
opacity over velocity for the u → l component can be related
ofthe 1 -1 and 1 -0 transitionsof o/p-NH D is, however,
11 01 01 00 2
tothetotalopacityτ givenbytheHFSfit,through
tot puzzling because all four transitions have high critical densi-
ties(7104–7105cm−3),alltheobservationshavesimilarspatial
Z τuldv∝τtotsul∆v, (2) resolutions (see Sect. 2), and finally, because NH2D should be
stronglyconcentratedonthecentreofthecoreforchemicalrea-
where s is the line-strength associated with the isolated hy- sons.Wewouldthusexpecttheselinestoprobethesamevolume
ul
perfinetransition.Thelinewidthwhichentersthisexpressionis ofgasandweshould,inprinciple,derivesimilarvaluesfor∆v.
associated to thermal and non-thermal processes, i.e. the mo- BytakingintoaccountthesplittinginducedbytheDnucleus
tionofmoleculesatmicroscopic(temperature)andmacroscopic (seeTable2and3),theparametersderivedfromtheHFSmethod
(turbulence)scales. Such an expressionis routinelyused in as- are,forp–NH2D:
trophysicalapplications.IntheparticularcaseofNH D,sucha
2 – 110GHz:τ =2.2±0.7and∆v=0.23±0.02kms−1
relation is often applied to the analysis of the 86 or 110 GHz tot
– 333GHz:τ =1.4±0.8and∆v=0.19±0.02kms−1
lines, with the aim to putconstraintson the deuteriumfraction tot
(seee.g.Olbergetal.1985;Tine´etal.2000;Roueffetal.2005; andfortheo-NH D:
2
Busquetetal.2010;Fontanietal.2015).
InthecaseofNH2D,thehyperfinestructureinducedbythe – 86GHz:τtot =5.2±0.5and∆v=0.24±0.01kms−1
Dnucleusisnotresolvedinastrophysicalmediasincethebroad- – 333GHz:τtot =2.3±0.4and∆v=0.23±0.01kms−1
eningofthelinesduetonon-thermalmotionsislargerthanthe
Thecorrespondingfitsoftheo/p-NH Dhyperfinetransitionsare
hyperfine splitting. Hence, it seems reasonable to analyze the 2
compared to the observationsin Fig. 1. For both spin isomers,
linesjusttakingintoaccountthehyperfinestructureinducedby
we find that the differences between the linewidths are largely
6 the HFS acronym stands for ”HyperFine Structure” reduced,thedifferencesbeingnowatmost∼20%.Inparticular,
and a description of the method is available at wefindthattheeffectismostimportantforthe86and110GHz
www.iram.es/IRAMES/otherDocuments/postscripts/classHFS.ps lineswhilethetwolinesat333GHzhavesimilarlinewidthswith
3
Danieletal.:NH Dhyperfinestructure
2
111 101 p-NH2D o–NH2D
F1 F F1’ F’ ν(MHz) Aul(s−1) ν(MHz) Aul(s−1)
0 1 1 0 110151.982 1.95(6) 85924.691 9.23(7)
0 1 1 2 110152.040 9.18(6) 85924.749 4.35(6)
0 1 1 1 110152.072 5.20(6) 85924.781 2.46(6)
0 1 2 2 110152.565 6.21(8) 85925.273 2.93(8)
0 1 2 1 110152.662 1.52(7) 85925.370 7.17(8)
2 1 1 0 110152.935 1.99(6) 85925.644 9.43(7)
2 2 1 2 110152.954 6.05(7) 85925.662 2.87(7)
2 3 1 2 110152.980 4.44(6) 85925.688 2.10(6)
2 2 1 1 110152.986 2.52(6) 85925.694 1.20(6)
2 1 1 2 110152.993 4.85(8) 85925.702 2.30(8)
2 1 1 1 110153.025 3.36(6) 85925.734 1.59(6)
2 2 2 2 110153.478 8.81(6) 85926.186 4.18(6)
0 1 0 1 110153.484 2.82(10) 85926.191 1.33(10)
2 3 2 2 110153.504 1.07(6) 85926.212 5.09(7)
Fig.2:Line-strengths,normalizedandcenteredonrestfrequen- 2 1 2 2 110153.517 2.72(6) 85926.225 1.29(6)
cies,forthe86GHzand333GHzlinesofo–NH D.Inblack,the 1 1 1 0 110153.534 1.58(6) 85926.243 7.47(7)
2
2 2 2 3 110153.536 2.09(6) 85926.244 9.92(7)
spectroscopy accountsfor the coupling with D while the spec-
2 3 2 3 110153.562 1.10(5) 85926.270 5.23(6)
troscopyobtainedjustaccountingforNisshowninred. 1 2 1 2 110153.574 2.92(6) 85926.282 1.38(6)
2 2 2 1 110153.576 2.48(6) 85926.284 1.18(6)
1 0 1 1 110153.580 2.73(6) 85926.288 1.30(6)
1 1 1 2 110153.592 2.10(6) 85926.301 9.96(7)
1 2 1 1 110153.606 1.04(6) 85926.314 4.91(7)
orwithouttakingintoaccountthehyperfinestructureinducedby 2 1 2 1 110153.615 8.30(6) 85926.323 3.94(6)
theDnucleus.Thisdifferencebetweenthelinescomesfromthe 1 1 1 1 110153.625 1.15(6) 85926.333 5.42(7)
1 2 2 2 110154.098 1.45(6) 85926.806 6.87(7)
spectroscopyandisillustratedinFig.2fortheo-NH Dspiniso-
2 1 1 2 2 110154.117 5.18(6) 85926.825 2.46(6)
mer(thedescriptionthatfollowswouldbesimilarforp-NH D). 1 2 2 3 110154.156 5.63(6) 85926.864 2.67(6)
2
Inthisfigure,weseethatthenumberofhyperfinecomponentsis 1 0 2 1 110154.170 9.22(6) 85926.877 4.37(6)
1 2 2 1 110154.196 1.11(7) 85926.904 5.25(8)
limited for the 101-000 transitions.Additionally,for these lines, 1 1 2 1 110154.215 6.91(7) 85926.922 3.28(7)
thespreadinvelocityofthehyperfinecomponentsisreducedby 2 2 0 1 110154.397 2.64(8) 85927.104 1.25(8)
2 1 0 1 110154.437 1.24(7) 85927.143 5.85(8)
comparisontothespreadofhyperfinesinthe1 -1 transitions.
11 01 1 0 0 1 110154.991 4.59(6) 85927.698 2.18(6)
Hence,forthelattertransitions,takingintoaccountthecoupling 1 2 0 1 110155.017 5.40(6) 85927.724 2.56(6)
withDleadstoareduced∆v.Finally,weseethatforthe86and 1 1 0 1 110155.036 5.85(6) 85927.743 2.77(6)
110GHztransitions,theestimatesofτ aresimilarwiththetwo
tot
setsoflinelists.Asaresult,accordingtoEq.1and2,theerror
Table 2:Line list ofthe 1 -1 o/p-NH D transitions,with the
madeinthelinewidthestimatewilltranslatedirectlytothecol- 11 01 2
hyperfinestructure due to both the nitrogenand deuterium nu-
umn density estimate. In the case of the H-MM1 observations,
clei. The A coefficientsare givenin the formata(b)such that
thetwoestimateswilltypicallydifferbyafactorof∼1.5,which A =a10−bul.
iswellabovecalibrationerrors. ul
101 000 o-NH2D p–NH2D
5. Conclusion F1 F F1’ F’ ν(MHz) Aul(s−1) ν(MHz) Aul(s−1)
0 1 1 2 332780.875 4.68(6) 332821.618 4.37(6)
Wehaveobservedthe1 -1 and1 -0 linesofo/p-NH Dto- 0 1 1 1 332780.875 2.26(6) 332821.618 2.11(6)
11 01 01 00 2 0 1 1 0 332780.875 1.20(6) 332821.618 1.12(6)
wardsthe prestellarcore H-MM1,andcalculatedthe hyperfine 2 1 1 2 332781.695 3.14(7) 332822.439 2.93(7)
patternsoftheobservedtransitions.Wefoundthatwhenthehy- 2 1 1 1 332781.695 4.54(6) 332822.439 4.24(6)
2 1 1 0 332781.695 3.29(6) 332822.439 3.07(6)
perfinesplittinginducedbytheDnucleusisneglected(asdone
2 3 1 2 332781.735 8.14(6) 332822.479 7.60(6)
inpreviousstudies),thelineanalysisleadstoinconsistentresults 2 2 1 2 332781.793 1.59(6) 332822.537 1.48(6)
pertaining the linewidths of the two transitions of o/p-NH D. 2 2 1 1 332781.793 6.56(6) 332822.537 6.12(6)
2
1 1 1 1 332782.285 1.35(6) 332823.029 1.25(6)
For both spin isomers, the widths of the 1 -1 and 1 -0
11 01 01 00 1 1 1 2 332782.285 3.15(6) 332823.029 2.94(6)
linesdifferbyafactorof1.6±0.3iftheDcouplingisnottaken 1 1 1 0 332782.285 3.65(6) 332823.029 3.41(6)
intoaccount.Onthecontrary,thenewlinelistgivescomparable 1 2 1 1 332782.317 1.59(6) 332823.062 1.48(6)
1 2 1 2 332782.317 6.56(6) 332823.062 6.12(6)
linewidthsfor all the transitions. An error in the linewidth will 1 0 1 1 332782.375 8.14(6) 332823.120 7.60(6)
be transferredto the column density estimate, and the effect is
particularlypronouncedforthe1 -1 linesoforthoandpara-
11 01
NH2D at 86 and 110 GHz. In the case of H-MM1, the column Table 3: Same as Table 2 but for the 101-000 o/p-NH2D transi-
density estimates derived from these lines with or without the tions.
hyperfinestructureowingtoDdifferbyafactorof∼1.5.
Acknowledgements. This work has been supported by the Agence Nationale
delaRecherche(ANR-HYDRIDES),contractANR-12-BS05-0011-01andby Caselli,P.&Dore,L.2005,A&A,433,1145
theCNRSnationalprogram“Physico-ChimieduMilieuInterstellaire”.AP,PC Cohen,E.A.&Pickett,H.M.1982,J.Mol.Spectrosc.,93,83
and JP acknowledge the financial support of the European Research Council Cohen,E.A.&Pickett,H.M.1982,JournalofMolecularSpectroscopy,93,83
(ERC;projectPALs320620).JHacknowledgessupportfromtheMPEandthe Coudert,L.H.&Roueff,E.2006,A&A,449,855
AcademyofFinlandgrant258769.PartialsalarysupportforAPwasprovidedby Coudert,L.H.&Roueff,E.2009,A&A,499,347
aCanadianInstituteforTheoreticalAstrophysics(CITA)NationalFellowship. DeLucia,F.C.&Helminger,P.1975,JournalofMolecularSpectroscopy,54,
200
Fontani,F.,Busquet,G.,Palau,A.,etal.2015,A&A,575,A87
References Fusina, L., Di Lonardo, G., Johns, J. W. C., & Halonen, L. 1988, J. Mol.
Spectrosc.,127,240
Bacmann,A.,Caux,E.,Hily-Blant,P.,etal.2010,A&A,521,L42 Garvey,R.M.,Lucia,F.C.D.,&Cederberg,J.W.1976,Molec.Phys.,31,265
Busquet,G.,Palau,A.,Estalella,R.,etal.2010,A&A,517,L6 Gu¨sten,R.,Nyman,L.Å.,Schilke,P.,etal.2006,A&A,454,L13
4
Danieletal.:NH Dhyperfinestructure
2
Heyminck, S.,Kasemann, C.,Gu¨sten, R.,deLange, G.,&Graf,U.U.2006,
A&A,454,L21
Johnstone,D.,DiFrancesco,J.,&Kirk,H.2004,ApJ,611,L45
Kukolich,S.G.1967,Phys.Rev.,156,83
Kukolich,S.G.1968,J.Chem.Phys.,49,5523
Mangum,J.G.&Shirley,Y.L.2015,PASP,127,266
Olberg,M.,Bester,M.,Rau,G.,etal.1985,A&A,142,L1
Parise,B.,Belloche,A.,Du,F.,Gu¨sten,R.,&Menten,K.M.2011,A&A,528,
C2
Roueff,E.,Lis,D.C.,vanderTak,F.F.S.,Gerin,M.,&Goldsmith,P.F.2005,
A&A,438,585
Thaddeus,P.,Krisher,L.C.,&Loubser,J.H.N.1964,J.Chem.Phys.,40,257
Tine´,S.,Roueff, E.,Falgarone, E.,Gerin, M.,&Pineau des Foreˆts, G.2000,
A&A,356,1039
5
This figure "nh2d_observations.png" is available in "png"(cid:10) format from:
http://arxiv.org/ps/1601.00162v1