changes in the physical environment of the inner …

source: https://doi.org/10.7892/boris.96926 | downloaded: 20.4.2022

CHANGESINTHEPHYSICALENVIRONMENTOFTHEINNERCOMAOF67P/CHURYUMOV-GERASIMENKOWITHDECREASING

HELIOCENTRICDISTANCE

D. BODEWITS1§, L. M. LARA2, M. F. A’HEARN1,3, F. LA FORGIA4, A. GICQUEL5, G. KOVACS5, J. KNOLLENBERG6, M. LAZZARIN4, Z.–Y. LIN (�� )7, X. SHI5, C. SNODGRASS8, C. TUBIANA5, H. SIERKS5, C. BARBIERI4, P. L. LAMY9, R. RODRIGO10,11, D. KOSCHNY12, H. RICKMAN13,

H. U. KELLER14, M. A. BARUCCI15, J.–L. BERTAUX16, I. BERTINI17, S. BOUDREAULT5, G. CREMONESE18, V. DA DEPPO19, B. DAVIDSSON13, S. DEBEI20, M. DE CECCO21, S. FORNASIER15, M. FULLE22, O. GROUSSIN9, P.J. GUTIÉRREZ2, C. GÜTTLER5, S. F. HVIID6, W.-H. IP23, L. JORDA9, J.-R. KRAMM5, E. KÜHRT6, M. KÜPPERS24, J. J. LÓPEZ-MORENO2, F. MARZARI4, G. NALETTO25,4,19, N. OKLAY5, N. THOMAS26, 27, I. TOTH28, J.-B. VINCENT5

1. DepartmentofAstronomy,UniversityofMaryland,CollegePark,MD20742-2421,USA§Correspondingauthor.Email:[email protected]. InstitutodeAstrofisicadeAndalucia-CSIC,GlorietadelaAstronomia,18008Granada,Spain3. GaussProfessor,AkademiederWissenschaftenzuGöttingen,37077Göttingen,Germany4. UniversityofPadova,DepartmentofPhysicsandAstronomy,Vicolodell'Osservatorio3,35122,Italy5. Max-PlanckInstitutfürSonnensystemforschung,Justus-von-Liebig-Weg,337077Göttingen,Germany6. DeutschesZentrumfürLuft-undRaumfahrt(DLR),InstitutfürPlanetenforschung,Rutherfordstrasse2,

12489Berlin,Germany7. InstituteofAstronomy,NationalCentralUniversity,Chung-Li32054,Taiwan8. PlanetaryandSpaceSciences,DepartmentofPhysicalSciences,TheOpenUniversity,MiltonKeynes,MK7

6AA,UK9. AixMarseilleUniversité,CNRS,LAM(Laboratoired'AstrophysiquedeMarseille),UMR7326,13388

Marseille,France10. CentrodeAstrobiología,CSIC-INTA,28850TorrejóndeArdoz,Madrid,Spain11. InternationalSpaceScienceInstitute,Hallerstraße6,3012Bern,Switzerland12. ScientificSupportOffice,EuropeanSpaceResearchandTechnologyCentre/ESA,Keplerlaan1,Postbus

299,2201AZNoordwijkZH,TheNetherlands13. DepartmentofPhysicsandAstronomy,UppsalaUniversity,Box516,75120Uppsala,Sweden14. InstitutfürGeophysikundextraterrestrischePhysik(IGEP),TechnischeUniversitätBraunschweig,

Mendelssohnstr.3,38106Braunschweig,Germany15. LESIA-ObservatoiredeParis,CNRS,UniversitePierreetMarieCurie,UniversiteParisDiderot,5,PlaceJ.

Janssen,92195MeudonPrincipalCedex,France16. LATMOS,CNRS/UVSQ/IPSL,11Boulevardd'Alembert,78280Guyancourt,France17. CentrodiAteneodiStudiedAttivitáSpaziali"GiuseppeColombo"(CISAS),UniversityofPadova,Via

Venezia15,35131Padova,Italy18. INAF,OsservatorioAstronomicodiPadova,Vicolodell'Osservatorio5,35122Padova,Italy19. CNR-IFNUOSPadovaLUXOR,ViaTrasea,7,35131Padova,Italy20. DepartmentofIndustrialEngineering,UniversityofPadova,ViaVenezia,1,35131Padova,Italy21. UniversityofTrento,viaSommarive,9,38123Trento,Italy22. INAF-OsservatorioAstronomicodiTrieste,ViaTiepolo11,34014Trieste,Italy23. InstituteforSpaceScience,NationalCentralUniversity,32054Chung-Li,Taiwan24. OperationsDepartment,EuropeanSpaceAstronomyCentre/ESA,P.O.Box78,28691Villanuevadela

Cañada(Madrid),Spain25. DepartmentofInformationEngineering,UniversityofPadova,ViaGradenigo6/B,35131Padova26. PhysikalischesInstitutderUniversitätBern,Sidlerstr.5,3012Bern,Switzerland27. CenterforSpaceandHabitability,UniversityofBern,3012Bern,Switzerland28. MTACSFKKonkolyObservatory,KonkolyThegeM.ut15/17,HU1525Budapest,Hungary

BODEWITSETAL.–PHYSICALPROCESSESINTHEINNERCOMAOF67P(ARXIV) 2

ABSTRACT

The Wide Angle Camera of the OSIRIS instrument on board the Rosetta spacecraft isequippedwithseveralnarrowbandfiltersthatarecenteredontheemissionlinesandbandsofvariousfragmentspecies.Theseareusedtodeterminetheevolutionoftheproductionandspatial distribution of the gas in the inner coma of comet 67Pwith time and heliocentricdistance, here between 2.6 – 1.3 AU pre-perihelion. �Our observations indicate that theemissionobservedintheOH,OI,CN,NH,andNH2filtersismostlyproducedbydissociativeelectronimpactexcitationofdifferentparentspecies.WeconcludethatCO2ratherthanH2Oisasignificantsourceofthe[OI]630nmemission.Astrongplume-likefeatureobservedintheinCNand[OI]filtersispresentthroughoutourobservations.ThisplumeisnotpresentinOHemissionandindicatesalocalenhancementoftheCO2/H2Oratiobyasmuchasafactorof3.WeobservedasuddendecreaseinintensitylevelsafterMarch2015,whichweattributetodecreasedelectrontemperaturesinthefirstkilometersabovethenucleussurface.

6figures,6tablesKeywords:comets:individual67P/Churyumov-Gerasimenko–techniques:imageprocessing–plasmas–radiationmechanisms:non-thermal–molecularprocesses–atomicprocesses


1. INTRODUCTION

CometsareconsideredrelativelypristineleftoversfromtheearlydaysofourSolarSystem.Theyaredistinguishedfromotherminorbodiesbyacomaofgasanddustproducedwhenices retained from the formation of the solar system sublime. Understanding theconnection between the coma and the comet’s nucleus is critical because observationsrarely detect the nucleus directly, and its properties must often be inferred frommeasurements of the surrounding coma. Because measurements of the coma do notnecessarily represent the characteristics of the nucleus due to spatial, temporal andchemicalevolutionof theemittedmaterial,projecting thecomaobservationsback to thenucleus requires an understanding of the processes that induce changes in the coma.Compositional studies must take into account chemical reactions and photolysis todetermine how the molecular abundances measured in the coma relate to the bulkcompositionofthenucleus. Direct,high-resolutionobservationsofcometarynucleiarerare,comingonlyfromspacecraft encounters. Connecting thesemeasurements to the coma provides a valuablemeansofevaluating the techniquesused insituationswhere thenucleuscannotbeseen.ImagesobtainedbyGiotto,Vega,DeepSpace1andStardustshoweddetailsoftheirtargets’nuclei(Kelleretal.1987;Soderblometal.2002;Brownleeetal.2004;Veverkaetal.2013),but in each case, coma observations were limited to the continuum around closestapproach.TheDeepImpactspacecraftwasthefirsttoobserveboththegasanddustcomaeof comets 9P/Tempel 1 and103P/Hartley 2 through amulti-wavelength filter set,whilemonitoringeachcometformonthsaroundcloseapproach(A'Hearnetal.2005;A'Hearnetal.2011). OrbitingthecometsinceAugust2014,theEuropeanSpaceAgency’sRosettamissionhas allowed an unprecedented study of the activity and evolution 67P/Churyumov-Gerasimenko.The largeheliocentricdistanceof thecomet, its lowactivity levels,and thecloseproximityofthespacecrafttoitssurfaceallowustosampleanenvironmentthathasnever been studied before and that is not accessible for observations from Earth. Thispaper describes observations of the dust and gas in the coma of 67P/Churyumov-Gerasimenko acquired byRosetta’s Optical, Spectroscopic, and Infrared Remote ImagingSystem(OSIRIS).Fragmentspeciesarerelativelybrightandemitinwavelengthsaccessiblefrom theground.OSIRIS’narrowband filtersprovidean important link togroundbasedobservations, and help to connect our detailed knowledge of 67P/Churyumov-Gerasimenkotothewiderpopulationofcomets.

2. OBSERVATIONS

OSIRIS consists of twobore-sighted cameras: anarrowangle camera (NAC, fieldof view2.2x2.2 degrees) and a wide angle camera (WAC, field of view 11.4x12.11 degrees)(Kelleretal.2007).TheWACisbestequippedtostudythecoma; its twelvenarrowbandandtwomediumwidthfiltersallowcolordiscriminationandtheimagingofemissionlinesand bands from gas and continuum in optical wavelengths (250–750 nm).We typicallymonitoredgasanddustactivitywiththeWAC,aboutonceeverytwoweeksforheliocentricdistancesgreaterthan2AU,andonceperweekbetween2and1.3AUpre-perihelion.The


Table1–Observinglog.SEQUENCE Date

(UTC)Time(UTC)

rh(AU)

Δrh(km/s)

Range(km)

Phase(deg)

MTP12/DEEPVAC 2015Jan24 11:35AM 2.47 -12.8 27.9 92.8MTP14/DEEPVAC 2015Mar12, 1:27AM 2.12 -13.3 80.7 50.6MTP15/STP051 2015Apr14 4:25PM 1.88 -13.2 170 74.5MTP16/STP055 2015May12 2:11AM 1.66 -12.5 155 71.6MTP17/DEEPVAC 2015June3 8:04AM 1.51 -10.9 232 89.3MTP18/VAC 2015July3 8:19AM 1.34 -7.77 168 89.7nominal sequence had a set of observations once per hour for a full comet rotation(~12h25mduringtheobservationsdescribedhere;Kelleretal.2015).AsbothRosettaand67P were coming closer to Earth, the data volume available increased, allowing us toincreasetheobservingcadencetoonesetofobservationsevery20min,for14h(toavoidaliasingwith thecomet’srotationperiod).Exposure timesareoptimized toachievegoodsignal-to-noise in the coma, often resulting in saturation on the nucleus and in theapparition of internal reflection artifacts (ghosts) appearing on the CCD images as acircular‘blob’totherightofthenucleus(Kelleretal.2007).Theacquiredimagesarescaledwithbinning(typically4x4 forcomaobservations)andobservinggeometry(distancetothe comet, distance to the Sun). To establish the connection with the nucleus, short-exposure,2x2-binnedimageswereacquiredwiththe375nmand610nmfiltersalongwiththecoma images.For thispaper,we limitourselves todataacquiredduring fourperiodswhenRosettaconductedtheso-called‘VolatileActivityCampaigns’.Thesecampaignsweremulti-instrument observations specifically designed to study the gas in the coma of 67P(includingOSIRIS,VIRTIS,MIRO,andAlice),andtwomorededicatedOSIRIScampaignstoprovide better temporal coverage. The observations discussed in this manuscript wereacquiredbeforethecomet’sperihelion,between2015January24andJuly3(seeTable1).During this period, the comet’s distance to the Sun decreased from 2.48 AU to 1.34 AU.BecauseRosettawasclosetothesurfaceinJanuary,thenucleusfillsasignificantpartofthefieldofview,whereasduringthe laterobservationstheWACmapsamuch largerpartofthecoma.Fortheanalysisdescribedbelow,weusedimagesacquiredattheapproximatesamediurnalphase.

3. ANALYSIS

3.1DataReductionandImageProcessing

All images are pre-processed using the standard OSIRIS pipeline (Tubiana et al.2015),which includes bias and dark subtraction, flat fielding, conversion from electronyieldtoradianceunits(Wm−2sr−1nm−1),andbad-pixelmasking.

3.1.1.Comagasemissionsinthefilters

Thenarrowbandfiltersweredesignedtosampleeitheremissionlinesandbandsofspecificgasesorcontinuumlightatnearbywavelengths,butinevitablyalsosampletheemissionofothermoleculeswithlinesthatfallwithinthenarrowbandfilters’passbandsatvarious


Table2–CharacteristicsoftheWAC’sfilters(Kelleretal.2007).

Filter+Name λcentral Width Comments (nm) (nm)

21Green 537.2 63.2 IncludesC2SwanΔν=-1,0.31UV245 246.2 14.1 Isolatedpinholes.41CS 259.0 5.6 Severaloverlappingpinholes.

SamplesCSA1Π–X1Σ+(0-0).51UV295 295.9 10.9 61OH 309.7 4.1 Severalpinholes,onestrong(at

thelowerleftquadrantoftheCCDimages).SamplesOHA2Σ+–X2Πi(0-0).

71UV325 325.8 10.7 Manyoverlappingpinholes.81NH 335.9 4.1 SamplesNHA3Π1–X3Σ-(0-0).

IncludesOH+(A3Π–X3Σ-).12Red 629.8 156.8 Broadbandfilter.13UV375 375.6 9.8 MinorcontributionC3CometHead

Group.14CN 388.4 5.2 SamplesCNvioletsystem,minor

contributionC3CometHeadGroup.IncludesCO2+(Ã2Π–X3Σ-).

15NH2 572.1 11.5 SamplesNH2Ã2A1àX2B1(0,10,0).

16Na 590.7 4.7 CoversbothNaD1andD2doublets;includesC2SwanΔν=-2

17OI 631.6 4.0 Samples[OI]1D–3P630nmlineonly.IncludesNH2Ã2A1àX2B1(0,8,0).

18Vis610 612.6 9.8 IncludesNH2Ã2A1àX2B1(0,9,0).levels. A summary of the characteristics of theWAC’s filters and of themost prominentemissionfeatureswithintheirpassbandsisgiveninTable2.

TheWAC canmap the distribution of water with its OI and OH filters. The OI filtercoverstheforbiddentransitionsfromtheOI(2p4)1Dstatetothegroundstate.TheOI(2p4)1Dstate ispopulateddirectlybyphotodissociationofH2Omolecules,as is theOI(2p4)1Sstate,whichrelaxesmostly(95%)bydecay intothe1Dstate(c.f.Cochran2008).TheOHfiltercoversthe(0-0)bandoftheA2Σ+–X2ΠtransitionofOH,centeredatabout308.5nm,which is excited almost entirely by fluorescence of sun light (c.f. Schleicher & A'Hearn1988).AsmallfractionofthephotodissociationofH2OalsoleadsdirectlytothepopulationofOH inhigh rotational statesof theA2Σ+electronic state (A'Hearnetal.2015),but theresultingemission fallsoutside thepassbandof theWAC’sOH filter.TheWAC’sCN filtercoversemissionfromB2Σ+–X2Σ+(0,0)transitionsaround388nm.TheNH2Ã2A1àX2B1(0,10,0)bandisverywideandhowmuchoftheemissionfallswithintheNH2filter’s


Table3–ContinuumremovalfactorsFfortheWACgasfiltersassumingdifferentcontinuumreddening(%per100nmbetween375nmand610nm).ΔFistherelative,error-propagated1-sigmastandarddeviationsoftheremovalfactors.

FILTERS RemovalFactor ΔF 0% 10% 20% 30% (%)OI/UV375 1.570 2.026 2.623 3.438 5CN/UV375 1.021 1.034 1.053 1.077 6OH/UV375 0.479 0.444 0.397 0.332 10NH/UV375 0.829 0.790 0.740 0.671 8NH2/UV375 1.738 2.124 2.630 3.321 5Na/UV375 1.673 2.080 2.614 3.343 8Vis610/UV375 1.648 2.127 2.753 3.609 5

passbanddependson theheliocentricdistanceandvelocity.TheNH filter covers theNHA3Π1–X3Σ-(0-0)transition.WhiletheWACisequippedwithCSandNafilters,thosetwofilterswereusedonlysporadicallyduringthefirsthalfofthemissionbecausethelowSNRand pinholes (CS) and contamination by C2 emission (Na) hampers the interpretation ofobservationsmadewiththesetwofilters.

The WAC is also equipped with several filters that can sample the continuum.Comparing the 375nm and 610nm narrowband filters illustrates the level ofcontaminationofthelatterfilterbygasemissionfeatures.InWACobservationsof16Cyg,asystem of two solar analogs, we measured a flux ratio Vis610/UV375 = 1.648 ± 0.08,whereasforthedustsurrounding67Pthisratiowasbetween2.8and5.6(Jan/March)and2.5–3.3(June/July).Suchvalueswouldrequireareddeningbetween20-40%per100nmbetween the two filters, while we determined from the NAC images that the averagereddeningof thedustwasbetween375and610nmwasaround18%per100nm (Sec.3.1.2.).Thissuggest that in the January/Marchdata,upto50%of the flux in the610nmfiltermight come from gaseous emission, probablymostly due to the emission from theNH2Ã2A1àX2B1(0,9,0)andC2d3Πg–a3Πu (Δv=-2)transitions.Aswillbediscussedbelow, the Vis610/UV375 ratio measured in June and July is close to what is expectedbased on the NAC reddening measurement, suggesting that the relative contribution ofgaseousemissiontothefluxmeasuredintheVis610decreased.

To furtherassess thegas contamination inother filters,we show the transmissionoftheWACfiltersonahigh-resolutionspectrumofcomet122P/DeVico;Cochran2002);Fig1a).Thespectrumofthiscometisnotaparticularlygoodproxyforthatof67P;thecometwasobservedveryclosetotheSun(0.7AU),wasextremelydust-poor,andthetwocometshavedifferentcompositions(67Pisdepletedinitscarbonchainmoleculeswhile122Phasa‘typical’composition;Fink2009).However,thehigh-qualityspectrumandlinecataloguedemonstratetheextentofthecontaminationandtheidentificationofthegasesresponsiblefor it. An example is shown in Fig 1b, where we show the transmission of the OI filteroverlaid on comet 122P/DeVico’s spectrum. The filterwas designed to sample emissionfromthe[OI]1Dà 3P lineat630nm,butalsocontainsseveralemissionlinesoftheNH2molecule,notablytheÃ2A1àX2B1(0,8,0).TheWACisequippedwithanarrowbandfilterdesignedspecificallytoobservetheNH2Ã2A1àX2B1(0,10,0)band,whichcanbeusedto


Fig.1a–OSIRIS/WACfiltertransmissionprofilesoverlaidonahigh-resolutionspectrumofcomet 122P/DeVico (Cochran et al. 2002). Profiles are not convolvedwith the quantumefficiencyofthedetector.

Fig.1b–Asabove-closeupontheOIfiltertransmission.


remove NH2 emission from the 610 nm (NH2 Ã 2A1à X 2B1 (0,9,0) transition) and OInarrow band filters (NH2Ã 2A1à X 2B1 transition’s (0,8,0)) if the relation between thethree different emission bands of NH2 is known. To calculate the contribution of NH2emissiontothefluxmeasuredintheNH2,Vis610,andOIfiltersweweightedthearchivalspectrumS(λ)of122PwithOSIRIS’CCDquantumefficiencyQ(λ),thereflectivityR(λ),andtransmissionofitsfiltersT(λ):

!′ # = &(()∙+(()∙,(()∙- ( ∙.(

&(()∙+(()∙,(()∙.( [1]

ForcometDeVicowefindthattheratiobetweenNH2emissionintheOIandNH2filtersis1.389,and0.45fortheratiobetweenNH2emissionintheVis610andNH2filters.Usingthisbandratio,wefindthatNH2typicallycontributedbetween10-20%tothefluxmeasuredinthe OI filter (after continuum subtraction); however, as we will discuss in Sec. 5, thisassumes that band ratios between photo fluorescent excitation and electron impactexcitationaresimilar,whichmaynotbethecase.

Fluorescence efficiencies for NH2 have been calculated by Kawakita & Watanabe(2002).Theirmodelshowsthattheratiobetweenfluorescenceefficienciesofevenandoddtransitionsisstronglydependentontheheliocentricdistanceandheliocentricvelocity,butthattheratiobetweenevenbandsisconstant.NH2emissionintheNH2andOIfiltersisthuscoupled and can be removed with a constant factor that is independent of heliocentricdistance and velocity, and this factor should be the same for both 122P and 67P. TheVis610filterhoweverincludesemissionfromanoddband,(0,9,0),whichhasamuchmorecomplexrelationtotheemissionintheeven(0,10,0)bandsampledbytheNH2filter.WethereforedidnotattempttoremovethecontributionofNH2tothe imagesacquiredwiththeVis610filter.

Concluding,wedeemthefiltercenteredat375nmbestsuitedtoimagethecontinuumlight, although its transmission includes some emission lines from C3 (Ã 1Πuà X1Σg+)transitions.Therearefeweremissionlinesatshorterwavelengths,butthethreenear-UVcontinuumfilterscenteredon245,295,and325nmsufferheavilyfrompinholes(circulardefectsinthefilters’coatingsfromimpuritiesintheproductionprocess)whichrepresentan increased light flux at shorter and longer wavelengths than those in the filter widthbandpass.ThedisadvantageofusingtheUV375filtertoremovethecontinuumfromallthegasfiltersisthatsinceweuseonlyonefilter,wedonotmeasurecomacolors,andhavetoassumealinearreflectancewithrespecttothewavelength.Theassumedreddeninghasalargeimpactonthecontinuumremovalfactorusedwithincreasingwavelengthdifference.

3.1.2.Continuumremovalandreddening

Tomeasure the gaseous emission line flux, the contribution of the reflected continuumneeds tobe removed (seee.g.Farnhametal.2000).WeusedOSIRISobservationsof thesolaranaloguesofthe16Cygsystem(thetwostarsarenotspatiallyresolvedbytheWAC)todeterminetheratioofnarrowbandcontinuumfluxes inall filters forasolarspectrum.Theseratiosαaretheweightfactorsneededtoremovethefluxcontributionof‘grey’dustfrom the total flux Ftot measured in a narrow band using the flux F375measured in theUV375filterinordertomeasurethefluxcontributionbythegasofinterest:


/012 = /343 − 6 ∙ /789 [2]

The signal-to-noise ratio in the OH and NH filters is poor because UV fluxes from solaranalogs are low and the CCD’s quantum efficiency drops significantly below 400 nm(Magrinetal.2015).Inaddition,thelongexposuresrequiredinthesefiltersincreasesthechanceofcosmicraysaffectingthemeasurement,reducingthenumberofuseableimages.Averagefilterratiosand1-σstandarddeviationsaregiveninTable2.

TocalculatetheeffectofspectralreddeningP(inpercentper100nm)onthefluxratioswe first calculated the resulting, reddened spectrumbyweighting a solar spectrumwiththereddening: # asa linear functionofwavelengthλbetweenλ1=375.6nmandλ2=612.6nm(thecentralwavelengthsof twoof theWAC’scontinuumfilters),normalizedat494nm:

: # = 1 + 0.01 ∙ Ρ ∙ λ − 0.5 ∙ (λB − λC) [3]

We used the ratios of the fluxes of the reddened spectrum and the original solarspectrumtocalculatecontinuumremovalfactorsforincreasingreddening,whicharegivenin Table 3. To estimate the average reddening of the continuumof 67Pweperformed adeeperinvestigationofthecolorsofthecomausingNACmedium-bandfiltersobservations,forwhichthecontributionsofextraneousemissionlinesshouldbelessimportant.Forthis,wechooseadatasetthathadgoodcoverageofthecomaaroundthenucleususingalargenumberfilters.Theselectedimageshavebeenacquiredon2015February18infivefilters:Near-UV,Blue,Orange,Red,Infrared.Thecometwasatabout2.3AUfromtheSunandthephaseanglewasabout85deg.Thespacecraftwasat220kmfromthecometandtheNAC’sfieldofviewwasabout9kmaroundthenucleus.

We constructed a map of radial distance from the nucleus limb and computed theaverage coma surface brightness over an azimuthal region generally free of gas jets andghosts.AssumingthatinthebroadNACfilters(mostly~50nm)thecontributionofthegasemission issmallwithrespect tothedustcontribution,wederivedthecolorsof thedustcoma. We found that the coma has an average reddening of 19% per 100 nm in thewavelengthrange360–649nm,theclosesttothe375–610nmrangeobtainedwithWACfilters.

However, NAC imaging is not available concurrentwithmostWAC gas observations,anditmayormaynotsampletherelevantregionofthecomagiventheinstrument’ssmallfieldofview.Therefore,wedevisedasecond,empiricalmethodtodeterminethereddeningof the dust in the coma. We assumed that NH2 emission comes predominantly fromfragmentspeciesinthecoma.Thus,fortheimagesof67PacquiredwiththeNH2filtertheunderlyingcontinuumsubtractionwasdonebyvaryingthecontinuumremovalfactoruntiljets disappeared from the NH2 filter image leaving a rather isotropic NH2 coma. Thismethodtypicallyyieldsfactorsthatcorrespondto17–20%per100nmbetween375and610nm,somewhat largerthanreportedfromground-basedobservationsof thecomet inpreviousapparitions(11±2%per100nmbetween436and797nm;Tubianaetal.2011)but consistent with ground-based observations acquired during this apparition (20%betweenBandV;Snodgrassetal.2015).Wethereforeusedaconstantreddeningof18%per100nmforthedatadiscussedinthispaper.


3.2 ColumnDensitiesandProductionRates

Weidentifiedobservationsineachsequencethatwereacquiredatthesamediurnalphase,extracted surface brightness profiles in two directions, and measured the surfacebrightnessatafixedposition3kminthesunwarddirectionabovethesurface.

When the formationandexcitationprocessesof the fragment species areknown, themeasured surface brightnesses S(x,y) can be converted into column densities Ncol byassumingfluorescenceefficienciesgofeachspecies:

DE4F(G, I) =JK∙- L,M ∙N(

OP∙0 [4]

where Δλ is the FWHM of the filter (Table 2) and Ep the energy of the photons at thewavelengthoftheemissionfeatureconsidered.[OI]surfacebrightnessescanbeconvertedintoH2Ocolumndensitiesusingreactionratesforpromptexcitationofthe1Sand1Dstates(Bhardwaj&Raghuram2012),weighedbythebranchingratiosofthetransitionsleadingtoemissionat630nm(the2ndlineofthedoubletat636nmisattheedgeoftheOIfilterwherethetransmissionisbelow6%ofthepeaktransmission;seeFig.1b).TheCN,OH,andNH surface brightnesses were converted into column densities using publishedfluorescenceefficiencies (Table4).WehavecalculatedproductionratesusingastandardHasermodel for easier comparisonwith other observations. Assumed outflow velocitiesandlifetimesforparentsanddaughterscanbefoundintheAppendix(TableA1).Tobetterevaluatethecolumndensities,wealsocalculatedproductionrateswithamodifiedHasermodelthattakesthegasacceleration,collisionalquenchingofthelong-lived1Dstateoftheoxygenatom,andtheeffectoftheoxygenatommovingoutofthefieldofviewbeforeitcandecaytothegroundstateintoaccount.ThismodelisdescribedinAppendixA.

3.3 Uncertainties

The results are subject to several possible systematic uncertainties. The absolutecalibrationofOSIRISisbetterthan1%formostfilters,butthecalibrationconstantfortheOH filter (and continuum filters <300 nm) has an uncertainty of ~10% (Tubiana et al.2015). Bias levels are temperature dependent and have gradually changed over timebecausethespacecraftapproachedtheSun.Fromthe16Cygobservations(wherewecansee the background),we estimate that the bias level is now constrained towithin 1DN.Because the bias is individually determined for each hardware configuration, the errorremains1DNindependentofbinning.TheresultinggasdetectionstypicallyhadaSNRof4orbetterper4x4pixelat100pixelsfromthenucleus.Thecontinuumremovalisthelargestsystematic and statistical uncertainty in our data analysis. For example, in the dataacquiredinMarch,atadistanceof100pixels,i.e.,0.82km,fromthenucleus,thecontinuumcontributes10%tothetotalsignalinCN,~20%inOI,~30%inOH,andasmuchas65%ofthe signal in theNH2 filter.We have also tried to optimize the quality of the continuumremoval factors by averaging repeated observations of 16Cyg, and these coefficients arenow constrainedwithin 5–10% (Table 3). As discussed above, we only use the UV375filtertoobservethecontinuumemission,andhavetoinferthecolorofthecoma.Wedonotaccount for spatialgradientsand temporalvariations in thecolorof thecoma.Assumingthat thereddening is typicallybetween0–30%,differences inreddeningcouldresult in


uncertainties of <1% (CN), 8% (OI), 5% (OH), and 30% (NH2) in the resulting pure gasemissionaftercontinuumsubtraction.. Uncertainties in the assumed fluorescence efficiencies can affect the resultssystematically. Even more so, the use of those factors rests on the premise that thedominantprocessesarephoto-processes,whichwedeemnotbethecase(Sec.5).

4. RESULTSContinuumandcontaminantremovednarrowbandimagesforMarchandJuneareshownin Fig. 2, along with contextual images acquired with the 375 nm filter. All images areorientedsuchthatthedirectiontotheSunisupward.Ineachframe,theentirefieldofviewoftheWACisshown,whichsidescorrespondsto16and46kmprojectedatthenucleus,respectively. Ineachimageacircular featurecanbeseenontherightsideofthenucleus(saturatedpixelsaremaskedblackinourimageprocessing).These‘ghosts’arecausedbyinternal reflections from optical elements and are always placed on the detector at thesameplacewithrespecttothenucleus.Asmaller,weakerartifactcanbeseeninthesamedirectiondirectlynexttothenucleus.ThreeremnantpinholescanclearlybeseenintheOHimages.

Allcontinuumimagesshowmultiplebright,collimatedjetsonthesunwardsideofthenucleus.Inthetoprow,theMarchdata,thosejetsarenotpresentinanyofthecontinuum-subtractedgasimages.Atallepochsweseeaplume-likemorphologyperpendiculartothesunwarddirection.ThisfeatureisseenintheOIandCNfilter,butnotintheOH,NH,andNH2filters,wheretheobservedmorphologyislesspronouncedandenhancedtowardstheSun(Fig.2).

ThecontinuumsubtractedNH2 imageshavepoorSNRandshowabroaddistributionwith little structure. For the data acquired in June and July, while assuming the samereddeningof18%per100nm(Sec.3.1.2),thejetsareover-subtractedintheOIdataandcannot be entirely removed from most of the other filters. We believe this to be aconsequenceofourapproachtothecontinuumsubtraction(i.e.assumingaconstantcolorthroughouttheinnercoma).ForeachoftheimagesinFig.2,thecolorscaleofthelookuptablewasadjustedtobestemphasizethemorphology.ItisclearthatwhilethemorphologyintheOHandNHimagesremainsisotropicanddiffuse,theconeinthe[OI]andCNimagesbecomesmuchfainterandlessdefinedovertime. Radialsurfacebrightnessprofileswereextractedfroma21-pixelwideboxbetweentheleftlimbofthenucleusandtheedgeoftheframe.Weextractedsurfacebrightnessesinboththehorizontalandvertical(sunward)direction,startingatthenucleus(orientationasinFig.2).The1-sigmauncertaintywasestimatedfromthestandarddeviationwithinthe21-pixelwide box. The resulting continuum subtracted profiles are shown in Fig. 3. ThefirsttwoepochsshowthebestSNR.Thesurfacebrightnessdecreasedbyafactorof5-10afterFebruary,resultinginvisiblepoorerSNR.FortheobservationsinApril,May,andJune,data within 1 km from the surface seems unreliable, with the exception of CN. NH2 isconsistently very noisy (as the plots have a logarithmic scale, negative values are notshown),anditwasprobablynotdetectedintheJuneandJuly2015.Onallsixepochs,thereis a clear difference between the shape of the profiles of OH andNH,which show a 1/rdrop-offwithdistance,andthe[OI]andCNprofiles,whichareflatwithinthefirst~8kmfromthesurface.InJanuaryandMarch,theOHprofilesinthesunwarddirectionarevery


FIG2.Lefttoright:shapemodelimageforreference,continuumsubtractedimagesacquiredwiththeOH,NH,NH2,OI,andCNfilters,andalong-exposureUV375filterimageforcomparison.Intheorientationusedhere,theSunisalwaystowardthetop,andthemainreflectionghostcanbeseentotherightofthenucleus.AllimagesaretheentirefieldofviewoftheWAC(11.4x12.1degrees).Thecolorscaleisdifferentforeveryfilter,butiskeptconstantthroughoutallepochswithineachfilter.similartothoseintheperpendiculardirection.Afterthat,theemissiononthesunwardsideofthecomaismuchstronger(factor2)thanthatinthedirectionorthogonaltothecomet-sunline.ThisasymmetryisalsopresentintheNHprofiles,albeitlesspronounced.IntheOIandCNprofiles,theemissionissomewhatstrongerintheorthogonaldirectionbecausetheplumesvisibleinFig.2extendtobothextractiondirections.OnthetwoepochswherewehaveagoodNH2detection,JanuaryandMarch,itsprofilesareveryflatandverydifferentfromtheNHprofile.


Fig.3–Surfacebrightnessprofiles.Bluelinesareprofilesinthehorizontaldirection(‘plume-ward’);redlinesareprofilesintheverticaldirection(sunward;cf.Fig.2A).Thethin,lightershadelines(pinkandcyan)indicatethe1-sigmauncertainties.

5.EMISSIONPROCESSES

There are numerous inconsistencies in our results if we assume standard cometaryphysics.First,columndensitiesandproductionratesderivedfromOSIRISimagesaremuchhigher than thosemeasured by other instruments on boardRosetta (MIRO, VIRTIS, andROSINA/COPS;Bieleretal.2015a;Bockelée-Morvan,D.etal.,2015;Fougereetal.,2016;).AssumingphotodissociationofH2Oasthemainsourceofformationandpromptemissionbyatomicoxygen,andphotodissociationandsubsequentfluorescentexcitationofOH,CN,andNHemission,wederivedcolumndensitiesandcalculatedglobalproductionratesusingstandardHasermodel(Fig.4,Tables5and6).

TherearecurrentlynocontemporaneousmeasurementsavailableoftheabundanceofCNandNH.AbundancesofNH3andHCNwere0.06%and0.09%withrespecttowater,measuredbyRosetta’sROSINAinstrumentat3.1AUonthesunwardsideofthecomet(LeRoyetal.2015).Fink(2009)measuredNH2andCNabundancesof0.19%and0.15%from


the ground at 1.35 AU from the Sun post-perihelion during the 1995 apparition. Tocalculatetheexpectedsurfacebrightnesslevels,weassumedabundancesof0.1%forHCNandNH3,andwaterproductionratesfromFougereetal.(2016).AsshowninFig.4a,waterproduction rates derived from theOH observations using the standardHasermodel areinitiallymorethenafactor300largerthenexpected.ProductionratesderivedfromCNand

FIG.4A–WaterproductionratesderivedfromOH(bluediamonds)and[OI](blacksquares),CNproductionrates(greencircles),andNHproductionrates(redtriangles),basedontheassumptionthatphotodissociationandfluorescencearethedrivingdestructionandexcitationmechanism.ThedashedlineindicatesH2OproductionratesderivedfromMIROmeasurements(Fougereetal.,2016);thedottedlineindicatesexpectedproductionratesofNH3andHCNassumingfixedabundancesof0.1%.

FIG.4B–AsFig4Abutproductionratesarenowcalculatedusinganenhancedmodelthatincludesaccelerationintheinnercoma.H2Oproductionratesderivedfrom[OI]emissionincludequenchingandtransporteffects.


Table4–Photonproductionefficienciesbasedonphotodissociationrates(OI)andfluorescencemodels(OH,CN,NH)inunitsofphotons/molecule.

DATE OI1 OH2 CN3 NH4 (photons/molecule/s)1/24/2015 9.95E-8 7.13E-5 0.010 1.89E-33/12/2015 1.35E-7 9.21E-5 0.014 2.56E-34/14/2015 1.72E-7 1.18E-4 0.018 3.29E-35/12/2015 2.20E-7 1.61E-4 0.023 4.20E-36/3/2015 2.85E-7 1.95E-4 0.029 5.09E-37/3/2015 3.59E-7 1.43E-4 0.041 6.68E-3

References:1)Bhardwaj&Raghuram2012;2)Schleicher2010;3)Schleicher&A'Hearn1988;Kawakita&Watanabe2002;4)Kimetal.1989.NHsurfacebrightnessesarealsolargebyfactorsof1000.Thischangesdramaticallywhenweassumeamorerealisticvelocity(eq.A2)intheinnercoma,seeFig.4b.DerivedproductionratesforH2OderivedfromOHarenowaboutoneorderofmagnitudetoolarge,thoseofNHandCNparentsbyabouttwoordersofmagnitude.ForOHandCN,theJulyresultsareconsistentwithexpectedproductionrates.NHemissionremainsaboutafactorof10toohigh.

Second,withtheWACweexpecttoobservetwowaterphotolysisproducts,OHand

OI. Using the standard Haser model, water production rates derived from OH seemconsistentlylargerthanthosederivedfrom[OI],byafactorof6inJanuaryandbyafactorof 30 in July. This situation is not resolved by our appended comamodel because boththere is no change in the relative columndensities ofOH andOI if both are assumed tocomefromH2O.

Third, all production ratesdrop significantlybetweenMarchand June, 2015.Thisvariation ismuch larger than thediurnalvariationof the totalwaterproduction (~25%;Gulkis et al. 2015) and not consistentwith the observed trend of increasing productionrateswithdecreasingheliocentricdistance.

Fourth, the luminosityprofilesareatoddswiththeobservedmorphology.Surfacebrightness profiles of parent species usually decrease with ~1/r close to the nucleus,whereas those of fragment species have shallower slopes (c.f. Combi et al. 2004). TheOSIRISsurfacebrightnessprofiles(Fig.3)howeverareflatfor[OI]andCN,forwhichthemorphologysuggestsapromptexcitationprocess.Incontrast,themorphologyseenintheOHandNHfilters–asymmetricdistributionaroundthenucleus–istypicalforafragmentspeciesthatgetsasignificantvectorialkickuponphotodissociationofaparentspecies.ThemorphologyofemissionintheOIandCNimagesresemblestheprojectionofaconeofgasandindicatesaparentdissociationprocessthatproducesthesefragmentsdirectlyintoanexcitedstate.

Wethereforeconcludethatbyadjustingourmodelstobetterdescribethephysicalprocessesintheinnercomawecanexplainsomeoftheobservations,butthatthedifferencesbetweenourobservationsandmodelresultsindicatethatphotodissociationandfluorescencearenotthedominantprocessesresultingintheOH,[OI],CN,andNHemissionobservedintheinnercoma.Instead,thefragmentsmightbefragmentsfrom

Table5–Surfacebrightnessmeasuredat1kmabovethesurfaceinthehorizontaldirection(orthogonaltosunwarddirection).Columndensitiesandproductionratesarederivedassumingphotoprocessesandareonlygiventoshowthediscrepancybetweenobservationsandexpectedproductionrates(Sec.5). SurfaceBrightness Logcolumndensities (10-6W/m2/sr) (molec./m2)Date OI OH CN NH NH2 OI OH CN NH1/24/2015 11.5±0.2 1.5±0.2 11.4±0.2 1.4±0.2 5.3±0.3 21.7 17.6 16.5 16.23/12/2015 12.1±1.0 11.0±1.3 16.5±0.2 7.4±0.5 4.8±1.4 21.6 18.4 16.5 16.84/14/2015 1.3±0.8 2.7±0.2 2.0±0.1 1.1±0.8 0±1 20.5 17.6 15.4 15.95/12/2015 2.4±0.9 2.7±0.4 3.7±0.2 2.0±0.5 0.9±1 20.6 17.5 15.6 16.06/03/2015 1.6±1.0 3.8±0.2 2.0±0.3 1.7±0.5 0.8±0.8 20.1 17.6 15.2 15.87/03/2015 3.1±1.6 6.4±0.3 4.4±0.3 2.7±0.3 1.1±2 20.5 17.9 15.4 15.9Table6–Productionratesderivedassumingphotoprocessesandareonlygiventoshowthediscrepancybetweenobservationsandexpectedproductionrates(Sec.5).ResultsareshownforastandardHasermodel,foranenhancedmodelthatincludesacceleration(incl.OIa),andforamodelthatincludesquenchingandtransportofOI(OIb)Theproductionratesareallfortheassumedparentsoftheobservedfragments,thelabelindicatesfromwhatfragmenttheywerederived(i.e.OIforH2O,OHforH2O,CNforHCN,andNHforNH3). StandardHaser Enhancedmodel Logprod.rates Logprod.rates

(molec./s) (molec./s)Date OI OH CN NH OIa OIb OH CN NH1/24/2015 28.0 28.8 26.8 27.1 27.7 29.6 27.4 25.4 25.73/12/2015 27.9 29.5 26.7 27.6 27.6 29.9 28.1 25.3 26.34/14/2015 27.9 28.7 25.6 26.6 26.6 29.0 27.4 24.2 25.35/12/2015 27.1 28.5 25.7 26.4 26.7 29.2 27.2 24.3 25.46/03/2015 26.8 28.5 25.3 26.5 26.5 29.0 27.2 23.9 25.27/03/2015 27.0 28.8 25.5 26.5 26.7 29.1 27.5 24.1 25.2

FIG.5A–ModeledsurfacebrightnessesfromdifferentprocessesinthecomafortheJan24,2015observations.EID–electronimpactdissociation(dashedlines).PP–Photodissociation,Promptexcitation(solidlines).PF–Photodissociationandsubsequentfluorescentemission(dottedlines).Colorsindicatetheparentspecies:Blackandblue–H2Oproducts;Red–CO2products;Orange–COproducts;Magenta–O2;Green–HCN.Cyan–NH3.Thedash-dottedbrownlineshowsthesumofall[OI]emission.differentparentspeciesand/orformedbyotherprocesses.Wewilldiscussthisinfurtherdetailbelow.

5.1Waterfragments:OHandOI

As concluded above, the emission from [OI] and OH cannot be explained byphotodissociationofH2O, followedbyprompt emissionof [OI] orby fluorescenceofOH.The surface brightness profile of OH suggests its emission might be the product of aprocess thatproducesOHdirectly in theA2Σ+state.However, thedifferencebetween the[OI]andOHmorphologyindicatesthatatleastpartoftheemissionofthetwofragmentsisnotrelated.

Fromthemorphologyofthe[OI]emissionweconcludedthatit istheproductofaprocess that directly produces atomic oxygen in an excited state. Like H2O,photodissociation of CO2 and CO produces OI in the 1D and 1S states, resulting in [OI]emissionat630nm.AbundancesofCO2andCOvarygreatlybetweenthesummerand


FIG.5B–AsFig.5AabovebutfortheJuly3,2015observations.winter hemispheres. Here we assume that the sunlit side dominates the total gasproductionrate; there,abundanceratios fromthe ‘summerhemisphere’apply,whichareH2O:CO2:CO=100:2.7:2.5(LeRoyetal.2015).AssumingCO2andCOformationratesinto OI 1D and 1S from (Bhardwaj & Raghuram 2012), photodissociation of CO2 and COcontributes10%and<1%tothe[OI]630nmemissioncomparedtoH2O(100%).

The ROSINA instrument reported unexpectedly high abundances of O2, with anaverage of O2/H2O =�3.7 ± 1.5 %, with local abundances as high as 10% (Bieler et al.2015b).ThedistributionofO2inthecomasuggesteditisreleasedbythenucleusandthatitsreleaseiscorrelatedtotheoutflowofwater.BecausethephotodissociationofO2intoOI1D is very efficient and at these abundances (Huebner et al. 1992), it can contribute asmuch as an additional 25–60% to the [OI] 630 nm emission fromH2O. Thuswhile thephotodissociation of CO2, CO, and O2 molecules combined may produce as much [OI]emission as the photodissociation ofH2O, it cannot explain the factors of 20–40 of fluxexcessobservedbetweenJanuaryandMarch,norcanitexplaintheOHobservations.

WethenconsideredseveralprocessesthatmightproducethehighobservedsurfacebrightnessofbothOHand[OI]directlyfromH2O,includingdissociativerecombinationofH2O+, electron excitation of OI, and sputtering of water ice (c.f. Bhardwaj & Raghuram2012)butnoneofthosehavereactionratesthatexceedthatofphotoprocesses.Feldmanetal.(2015)concludedthatelectronimpactdissociationofH2OvaporproducedHIandOI


FIG.5C–ModeledsurfacebrightnessesfromdifferentprocessesinthecomafortheJan24,2015observations,butincludingtheeffectofcollisionalquenchingandtransport.EID–electronimpactdissociation(dashedlines).PP–Photodissociation,Promptexcitation(solidlines).PF–Photodissociationandsubsequentfluorescentemission(dottedlines).Colorsindicatetheparentspecies:Blackandblue–H2Oproducts;Red–CO2products;Orange–COproducts;Magenta–O2;Green–HCN;Cyan–NH3.Thedash-dottedbrownlineshowsthesumofall[OI]emission.emissionobservedbyRosetta/Alice in theFar-UV.Electron impactdissociationproducesOI in the 1D and 1S states, and OH in the A2Σ+ state. Those reactions typically haveappearance thresholds between 10-20 eV, suggesting that they are driven by the largepopulation of suprathermal electrons observed in by Rosetta’s Ion and Electron Sensor(IES;Clarketal.2015).Whileelectronswithenergiesof100eVandlargerwereobserved,the distribution falls off steeply >30 eV. At these impact energies, the cross section forelectronimpactproductionofOH(A2Σ+)is7x10-22m2(Avakyan&al1998)andthatofOI(1S+1D)is6x10-23m2(Bhardwaj&Raghuram2012).

To test whether electron impact dissociation can explain the observed surfacebrightnesswe added electron impact processes to ourmodel (Appendix A). In brief,weassumedasphericallyoutgassingnucleuswithwaterproductionratesfromtheempiricaltrend reportedbyFougere et al. (2016), and relative abundancesof 3% forCOandCO2,0.1%forHCNandNH3.Fortheelectrons,weassumedaradialdistributionthatdecreasedwiththeinverseofthedistancetothenucleusbasedonmeasurementsbytheRosetta


FIG.5D–AsFig.5CabovebutfortheJuly3,2015observations.

Plasma Consortium’s (RPC) Langmuir- andMutual Impedance Probes. For the electrondensity,weusedmeasurementswith theElectronand IonSensor (IES) (Broilesetal. inpress)whichweassumedto increase linearlywith theproductionrateandquadraticallywiththedecreaseinheliocentricdistance.Inaddition,weincludedtheeffectsofquenchingoftheOI1DstatebycollisionswithH2OmoleculesaswellastheeffectoftransportastheOIatommovesoutofthefieldofviewbeforeitcandecaytothegroundstate.

TheresultsofthemodelareshowninFig.5.Wefindthatat1kmfromthesurface,electron impact dissociation of H2O can produce up to a factor of 10more [OI] 630 nmemission than thephotodissociationofH2O intoexcitedOI 1D,andat least twoordersofmagnitude more OH emission than the photodissociation of H2O and subsequentfluorescentexcitationofOHbysunlight.However,themorphologyandsurfacebrightnessprofiles seen in the OI and OH filters is very different, which is surprising when weassumedtheyareboththeproductofelectronimpactdissociationofH2O.IntheOIimagesthereisaclearplumevisibletotheleft,whichisentirelyabsentinOHimages.IfthemainemissionintheplumeoriginatedfromelectronimpactdissociationofH2O,itshouldalsobepresentinOHemission.Ifelectronimpactdissociationdrivestheemissionof[OI]andOHin thecoma, then theatomicoxygenemission in theplume isprobablyproduced fromamoleculeotherthanH2O.


WearenotawareofexperimentalcrosssectionsfortheproductionofOI(1D),butbased on theoretical cross sections, Bhardwaj & Raghuram (2012) suggest that reactionrates forproducingOI (1D) fromCO2are asmuchas40 times larger than reaction ratesproducingOI(1D)fromH2O.ThereactionratesfortheproductionofexcitedoxygenatomsfromCOaremuch lower than thoseofH2O.Assuming theseelectron impactdissociationcross sections, a CO2/H2O abundance of 3%would imply that 60%of the observed [OI]emission comes fromCO2.This large contributionby electron impactdissociationofCO2explains thedifferencesbetweenboth themorphologyandsurfacebrightnessprofilesofthe [OI] and OH emission. The plume is visible in [OI] but not in OH. Its [OI] surfacebrightnessat the core is~2x that at similardistances in theambient coma, suggestingalocalenhancementoftheCO2abundance.

We do not expect our model to provide a fully realistic description of the innercoma. Predicted OH surface brightnesses are higher than observed, and quenching andtransportmayremoveasmuchas factor100 fromthedetectable [OI]surfacebrightness(Fig.5aandFig.5c).Inparticular,theasymmetricoutgassingof67Pwillaffecttheimpactoftheseprocesses.However,ourresultsdoconfirmthatelectronimpactdissociationcanindeed explain the emissionobserved inOSIRIS’OI andOH filters from January throughApril2015.

5.2Ammoniafragments:NHandNH2

ThephotodissociationofNH3resultsintheproductionofNH2(96%)andNH(~3%);mostoftheNHisthusnormallyagranddaughterproductofNH3(Huebneretal.1992).Electronimpact dissociation of NH3 has been relatively well studied and cross sections for theproduction ofNH2 (Ã 2A1) andNH (A3Π1) at 100 eV are both~2-3 x 10-22m2(Müller&Schulz1992), i.e. threetimessmallerthantheproductionofexcitedhydroxyl(TableA1).To our knowledge, a detailed study of the excitation of NH2 through this process is notavailable,leavingtherelativeintensitiesoftheNH2bandsthatcontaminateseveralfiltersanopenquestion(Sec.3.1.1).Surprisingly, thesurfacebrightnessprofilesofNH2andNHare very different (Fig. 3), which is inconsistent if the emission from both fragments isproduced by electron impact dissociation of NH3. We used our model to evaluate thesurface brightness expected from different emission processes. In all observations, theprofileofNHresembles thatofOH,suggesting that thesamephysicalprocesscauses theemission.TheprofileofNH2resembles thatofOIandCN in JanuaryandMarch,andwasprobablynotdetectedafterthat.

Assuminganabundanceof0.1%withrespecttoH2O,weincludedelectronimpactexcitationandfluorescentemissioninourmodel.PhotodissociationofNH3intoNH2andNH cannot explain the emission for any of the observations (Fig. 4b), nor does electronimpact dissociation (Fig. 5c), which both fall about a factor 100 short. Upon furtherinvestigation, theonlyplausible explanation for theNH filterobservationsappears tobeemission from the OH+ (A3Π – X3Σ- 0-0) band. Electron impact dissociation on H2O canindeedproduceexcitedhydroxylionswithacrosssectionof8x10-24m2at100eV(Mülleret al.1993), or about 100 times smaller than that for the production of OH (A2Σ+). Theemissioncrosssections forneutraland ionizedOHbothdependstronglyon theelectronimpactenergyandtheratiobetweenthesurfacebrightnessesintheNHandOHfilterswillthereforedependsstronglyontheelectrontemperatureinthecoma.


5.3CN

Themorphologyof theCNemission resembles thatof [OI] at all six epochs, and like thewater group fragments, its surface brightness is higher than expected from photoprocesses,byasmuchasafactorofatleast200(Fig.4b).PromptexcitationofCNfollowingphotodissociation of HCN is not a very efficient emission mechanism (Fray et al. 2005;Bockelée-Morvan&Crovisier1985),andourmodelindeedsuggeststhatphotodissociationofHCNandsubsequentfluorescentexcitationofCNwouldproduceatleast50timesmorelight thanpromptexcitationofCN following thephotodissociationofHCN.Following thediscussionontheproductionof[OI]andOHemission,wethenevaluatedifelectronimpactdissociation of HCN into the CN (B2Σ) state can drive the production of excited CN.Qualitative results confirm that this reaction channel is important in electron impactdissociation of HCN (Nishiyama et al. 1979), but cross sections are not available in theliterature.Assumingcrosssectionsandabundancessimilar toNH3reactions forHCN,wewouldexpectsurfacebrightnessescomparable toNH in January(Fig5c), i.e.around10-8W/m2/srat1kmfromthesurface.Interestingly,owingtothelargefluorescenceefficiencyofCN,emissionlevelsfromthephotodissociationofHCNfollowedbyfluorescentexcitationofCNlikelyproducesanorderofmagnitudemorelightthanelectronimpactdissociationofHCN.DissociativeelectronimpactexcitationofHCNcannotexplainourobservations.

This suggests that the light observed in the CN filter in January is produced byanotherspecies.AccordingtoAjelloetal. (1971)about5%of the lightemitted followingionizing electron impact excitation of CO2 fallswithin the CN filter’s passband. This alsoexplains the similarity inmorphologybetween theOI andCN filter images.At an impactenergyof30eV,theemissioncrosssectionfortheentireCO2+(Ã2Π–X3Σ-)bandis4x10-21m2,thusresultinginanemissioncrosssectionof2x10-22m2,about4xsmallerthanthatofdissociative electron impact excitation of H2O into OH*, and about 10x smaller than theproductionofOI 1D fromthedissociativeelectron impactexcitationofCO2.Wenote thatwhilethe[OI]emissionisstronglyaffectedbyquenchingandtransporteffectbutthattheCO2+emissionisnot,whichmayexplainwhy[OI]surfacebrightnessesarecomparabletothosemeasured in the CN filter (Table 5). The expected CO2+ surface brightness can beestimatedfromtheCO2-to-[OI]profileinFig.5aandwillbeoforder10-5W/m2/sr,orabout100timesthatoffluorescentemissionfromCN.Photodissociationofaparentmoleculeandsubsequent fluorescent excitation of fragment CN can explain the observed surfacebrightness levelsafterMay2015,but thepersistentplumemorphologysuggests that theobservedemissionmaybeaproductofbothCNfluorescenceemissionandCO2+emissionfromelectronimpact.

6.TRENDSWITHPERIHELIONDISTANCETheproductionratesshown inFig.4suggest threedifferentepochs: increasingemissionbefore March 2015 following the heliocentric trend observed by MIRO, then a sharpdecreasebetweenmid-MarchandJune,2015,followedbyanincreaseafterJune2015.ThedropincomaemissionismorethananorderofmagnitudeandisseenintheOI,OH,CN,NH,andNH2filters.Overthecourseofourobservations,themorphologyoftheemissioninthedifferentfiltersdoesnotchangenoticeably(Fig.2).AfterMay,theOHandCNemission


FIG.6A–SDO/EVEintegratedirradiancesbetween30–100nmat1AU(blackline)andattheheliocentricdistanceof67P(blueline).Redandbluedotsindicatethetimesofobservationsdiscussedinthispaper.levelsareat the levelexpectedfromphotoprocessesalone(Fig4). If theelectrondensityevolvedwithourscalinglaw(linearwiththegasproductionandinversequadraticallywiththe heliocentric distance), we would expect 1000 times more OH emission than wasobserved(Fig.5d).Similarly,thepredicted[OI]emissionfromdissociativeelectronimpactofH2O andCO2would produce 10xmore emission thanwas observed in July.However,production rates derived from [OI] and NH surface brightnesses assumingphotodissociation still require unrealistically high production rates (Fig. 4), and themorphology in the CN filter still resembles that seen in the OI filter, suggesting thatcollisions with electrons still play a role in the emission observed after May 2015. WeproposethatthedropbetweenMarchandJune2015wasthuscausedbyachangeinthenumberortemperatureofprojectileelectronsavailable,andthatthisaffectstheemissioninthedifferentfilterindifferentwaysdependingontheenergydependenceoftherelevantelectronimpactdissociationprocesses.

Toexplain thedecreaseofemission,we firstconsider theproductionofelectrons.TheRosetta/RPCobservationsinFebruary,2015indicatethatthemainsourceofelectronswithin256kmof thenucleus is theneutralgas inthecoma(Edbergetal.2015).Thefarandextremeultraviolet(FUV/EUV)solarfluxdeterminesthephotoionizationratesofH2Oandothermoleculesinthecoma,andthusalsocontrolsthenumberofelectronsavailable.InspectionofdailyaveragedsolarUVspectraacquiredwithExtremeUltravioletVariabilityExperimentonboardtheSolarDynamicsObservatory(SDO/EVE;Woodsetal.2012).We


FIG.6B–Opticaldepthforlightwithawavelengthof30nm,calculatedforthedifferentobservingdates,JanuarytoJuly2015frombottomtotop.integrated the irradiance spectraover the range30–100nm1, thewavelengths that aremostefficientinproducingelectronsbyionizingwater(Huebneretal.1992;Budzienetal.1994). The results are shown in Fig. 6a. Over the course of our observations, the UVirradiance decreased gradually, and it varied by 10% at time scales of 1 or 2 weeks.Comparing the production rates andUV irradiances in Fig. 6a suggests there is no clearcorrelationbetweentheshort-termsolarvariations. InthisperiodthecometapproachedtheSunfrom2.5AUto1.3AU,aneffectmuchlargerthantheweeklyUVvariations. Withincreasinggasproductionrates,theopticaldepthoftheinnercomaincreasesand fewerphotoelectronsareproduced.To investigate theopticaldepthof thecoma,wecalculated H2O column densities using a Haser distribution (Appendix A), and assumedwater production rates from the empirical formula by Fougere et al. (2016; Fig. 4). Thedominant photons for ionization havewavelengths between 30 – 85 nm (Huebner et al.1992;Budzienetal.1994).Atthesewavelengths,H2Ohasphoto-absorptioncrosssectionsof1.0x10-21m2and1.6x10-21m2,respectively(Phillipsetal.1977).TheresultsareshowninFig.6b.Therelevantregionhereisbetween1and10kmfromthesurface.InJanuary,only1to10%ofphotonswerelostwithinthisregion.Thisincreasedto10to64%inJuly.

1WeusedSDO/EVElevel3version5dataacquiredwiththeMEGS-Binstrument,availableonlineathttp://lasp.colorado.edu/eve/data_access/evewebdata/products/level3/


TheelectronproductionthuslikelydecreasedintheregionseenbyOSIRIS,butnotenoughtoexplaintheobserveddropinemission. Along with the gas production rates, dust production rates also increasedsignificantlyoverthecourseofourobservations.BasedonCassiniobservations,ithasbeensuggestedthatnano-grainchargingcouldresultinsignificantelectrondepletion(Vigrenetal.2015;Nilsson,etal.2015b).Reactionratesforattachmentofelectronstograinsareverylow for suprathermal electrons and would mostly affect electrons with much lowerenergies than thoseresponsible for theelectron impactdissociation(>10eV).Wedonotexpectelectron-dustinteractionstoaffecttheobservedemission. Themost likely explanation of thedecrease in emission from the inner coma is asignificantdecreaseintheelectrontemperature,whichreducesreactionrates.Atenergiesof a feweV,when temperatures fall below the appearance thresholdsof thedissociativereactions,theexcitationofH2Omoleculesbecomesaneffectivecoolingprocess(Cravens&Korosmezey1986).Electronsgainenergybyphoto-ionizationprocesses,comet-solarwindinteractions(Clarketal.2015),andloseenergythroughcollisionalprocesses,forexample(Wegmannetal.1999):

!"# + ℎ& → !"#( + ) +12.3)/ !"# + ) → !" + #( 12 ) + ) −12.3)/ [6]

!"# + ) → #! + ! + ) −5.1)/ TheGiottoprobeobservedmaximumiondensity12000kmfromthenucleusof1P/Halley(Häberlietal.1995).Thiswasattributed toasteep increaseof theelectron temperaturewith distance to the nucleus, in turn caused by decreased cooling through collisionsbetweenelectronsandneutrals.

In the period January – July 2015, the neutral gas density increasedmuch faster(~rh–4.2)thantheincreaseofsolarradiation(~rh–2),whichshouldleadtoadecreaseintheelectron temperature. The electron impact dissociation processes that produce theemissionobservedwithOSIRISallhaverelativelyhighappearanceenergies(Avakyanetal.1998).Coolingof theelectronsbelowthoseenergiescould lead toanabruptdecrease intheemission.

Lastly, the observed decrease in the emissionmight be related to changes in theinteractionbetweencometandsolarwind.Solarwindelectronshavehighertemperaturesthanphoto-electronsand its interactionwith thecomamayheatelectrons (Häberlietal.1996; Clark et al. 2015). Initially, the solar wind could penetrate deep into the coma(Nillsonetal.2015a,b).Withincreasingproductionrateplasmaboundariesseparatedthesolarwind from the cometary gas (Rubin et al. 2015). By the end of February 2015, theRPC/Ion Composition Analyzer measured that the proton energy spectrum started tochange,indicativeofanincreasinginteractionbetweencometandsolarwind.

SUMMARYANDCONCLUSION

ThisstudyusesdatafromRosetta’sOSIRIScamerasystemandpresentsobservationsoftheinner comaof67P/Churyumov-Gerasimenkoacquiredwithnarrowband filters centeredon the emission features of OH, OI, CN, NH, and NH2. The observations explore a newregime in cometary science: the inner coma of a low-activity comet at large heliocentric


distances,andtheRosettamissionallowsustostudyhowthisenvironmentchangedovertime. We developed an extensive image reduction procedure and applied this toobservationsacquiredbetween Januaryand July,2015,when thecomet’sdistance to theSundecreasedfrom2.5to1.3AU.

Theobservations inall filtersatallepochs indicatedsurfacebrightnesses thatareone or more orders of magnitude higher than can be explained by photodissociation.Instead,theemissionis likelytheresultofdissociativeelectronimpactexcitation,and/ordifferentspecies.TheOHemissioncanbeattributedtoelectrondissociativeexcitationofH2O, and theobserved intensities are generally consistentwithneutral columndensitiesandelectrondensitiesobservedbyotherinstrumentsonboardRosetta.Weattributemostofthe[OI]630nmemissionintheinnercomatoelectronimpactdissociativeexcitationofCO2,whichhasamuchlargercrosssectionthanH2O.TheemissiondetectedintheNHfilteris likely the result of electron dissociative excitation that produces OH+ ions from H2Omolecules directly in an excited state. The surface brightness levels observed in the CNfilter and the similarity between the morphology observed in the CN and [OI] filterssuggests that the emission is the product of both electron impact dissociative excitationproducedCO2+ ions andof fluorescent emissionbyCN radicals. Follow-up studies of thecorrelationbetween theemission seen inCNvs. [OI], andOHvs.NHmightprovideveryinterestingwindowsonthedifferentreactionchannelsfromimpactonthesamespecies.

TheintensityoftheemissionintheinnercomadecreasedbetweenMarchandJune,despiteincreasinggasproductionratesasthecometapproachedtheSun.Themostlikelyexplanation is that because the gas production rates increased much faster than theionizingsolarradiation,collisionsbetweenelectronsandneutralwatermoleculesloweredelectrontemperaturesintheinnercomabelowtheactivationthresholdofthedissociativeimpactexcitationreactions.Theincreaseofopticaldepthanddeflectionofthesolarwindmayhavefurthercontributedtotheelectroncooling.

OurresultsthusshowthatthenarrowbandfiltersonRosetta/OSIRIScanbeusedtoremotely study the interactionbetweenelectrons and theneutral gas in the inner coma.However,thelackofexperimentalcrosssectionshamperstheinterpretationofourresults;mosturgentlyneededarethosefortheproductionofOI1DfromCO2andH2O,andofCNfromHCN).Inaddition,wehaveidentifiedtheneedformodelsthatcombinethephysicaland chemical processes of cometary gases with plasma characteristics such as electrontemperatures.

Impactexcitationmaymatterindifferentplanetaryenvironmentsthroughdifferentinteraction mechanisms (photo-electrons, solar wind protons and/or electrons). First,since the resulting emission traces the distribution of a parent species, electron impactdissociationcouldexplaintheexcitationoffragmentswheneverasteep,radialgradientisseenintheemissionmorphologyofthefragmentspecies,suchasinthelarge-scalejet-likestructuresaroundcomets(A’Hearnetal.1986).Second,underconditionssimilartothatof67P, dissociative electron impact excitation can lead to significant emission by fragmentspecies.ThismightleadtodetectableemissionfromMainBeltCometsorevenCeres–ortoan overestimate of production rates if only photo-processes are assumed to drive theemissionoffragmentspecies.

Comet67PreacheditsperiheliononAugust13,2015,ataheliocentricdistanceof1.24AU.TheRosettaspacecraftwillcontinuetoorbititsnucleusandstudyhowitscomaevolveswhilethecometmovesawayfromtheSunagain.


Acknowledgements: The authors wish to thank Mike Combi, Mike Mumma, and theRosetta/Aliceteamforhelpfuldiscussionsregardingelectronimpactprocessesintheinnercoma. Rosetta is an ESA mission with contributions from its member states and NASA.OSIRISwasbuiltbyaconsortiumof theMax-Planck-Institut fürSonnensystemforschung,Göttingen,Germany,theCISAS,UniversityofPadova,Italy,theLaboratoired’AstrophysiquedeMarseille, France, the Instituto de Astrofísica de Andalucia, CSIC, Granada, Spain, theResearchandScientificSupportDepartmentoftheEuropeanSpaceAgency,Noordwijk,TheNetherlands,theInstitutoNacionaldeTécnicaAeroespacial,Madrid,Spain,theUniversidadPolitéchnica de Madrid, Spain, the Department of Physics and Astronomy of UppsalaUniversity, Sweden, and the Institut für Datentechnik und Kommunikationsnetze derTechnischen Universität Braunschweig, Germany. The support of the national fundingagenciesofGermany(DLR),France(CNES), Italy(ASI),Spain(MEC),Sweden(SNSB),andtheESATechnicalDirectorateisacknowledged.ThisworkwasalsosupportedbyNASAJPLcontract1267923totheUniversityofMaryland(M.F.A’H.andD.B.).M.F.A’H.isalsoaGaussProfessor of the Akademie derWissenschaften zu Göttingen andMax-Planck-Institut fürSonnensystemforschung(Germany).WegratefullyacknowledgeuseofSDO/EVEdata.


REFERENCES

A'Hearn,M.F.,Hoban,S.,Birch,P.V.,etal.,1986.Nature,324,pp.649.A'Hearn,M.F.,Belton,M.J.S.,Delamere,W.A.,etal.,2005.Science,310(5),pp.258.A'Hearn,M.F.,Belton,M.J.S.,Delamere,W.A.,etal.,2011.Science,332(6),pp.1396.A'Hearn,M.F.,KrishnaSwamy,K.S.,Wellnitz,D.D.,andMeier,R.,2015.AJ,150(1),p.5.Atkinson,R.,Baulch,D.L.,Cox,R.A.,etal.,1997.JournalofPhysicalandChemicalReference

Data,26(6),pp.1329.Avakyan,S.V.II'in,R.N.,Lavrov,V.M.,andOgurtsov,G.N.,1998.Collisionprocessesand

excitationofUVemissionfromplanetaryatmosphericgases:ahandbookofcrosssections.Ed.S.V.Avakyan.(Amsterdam,Netherlands:GordonandBreachScientificPublisers).

Bhardwaj,A.&Raghuram,S.,2012.ApJ,748(1),p.13.Bieler,A.,Altwegg,K.,Balsiger,H.,etal.,2015a.Astron.Astroph.,583,p.A7.Bieler,A.,Altwegg,K.,Balsiger,H.,etal.,2015b.Nature,526(7575),pp.678.Bockelée-Morvan,D.&Crovisier,J.,1985.Astron.Astroph.,151,pp.90.Bockelée-Morvan,D.,Debout,V.,Erard,S.,etal.,2015.Astron.Astroph.,583,P.A6.Brownlee,D.E.,Horz,F.,Newburn,R.L.etal.,2004.Science,304(5),pp.1764–1769.Budzien,S.A.,Festou,M.C.&Feldman,P.D.,1994.Icarus107,p164.Clark,G.,Broiles,T.W.,Burch,J.L.,etal.,2015.AstronomyandAstrophysics,583,p.A24.Cochran,A.L.,2002.Icarus,157(2),pp.297.Cochran,A.L.,2008.Icarus,198(1),pp.181.Combi,M.R.,Harris,W.M.&Smyth,W.H.,2004.inCometsII,ed.M.C.Festou,H.U.Keller,&

H.A.Weaver(Tucson,AZ:Univ.ArizonaPress),�p.523.Cravens,T.E.&Korosmezey,A.,1986.PSS,34,10,pp.961.Edberg,N.J.T.,Eriksson,A.I.,Odelstad,E.,etal.,2015.GRL,42(1),pp.4263–4269.Farnham,T.L.,Schleicher,D.G.&A'Hearn,M.F.,2000.Icarus,147,p.180.Feldman,P.D.,A'Hearn,M.F.,Bertaux,J.-L.,etal.,2015.Astron.Astroph.,583,p.A8.Fink,U.,2009.Icarus,201(1),pp.311–334.Fougere,N.,Altwegg,K.,Berthelier,J.-J.,etal.,2016.AstronomyandAstrophysics,588,

p.A134.Fray,N.,Benilan,Y.,Cottin,H.,Gazeau,M.-C.,andCrovisier,J.,2005.PSS,53(1),pp.1243–

1262.Gulkis,S.,Allen,M.,VonAllmen,P.,etal.,2015.Science,347(6),p.aaa0709.Häberli,R.M.etal.,1995.Astron.Astroph.,297,pp.881Häberli,R.,Altwegg,K.,Balsiger,H.,andGeiss,J.,1996.JournalofGeophysicalResearch,

101(A7),pp.15579.Huebner,W.F.,Keady,J.J.&Lyon,S.P.,1992.Astroph.Sp.Sci.,195,p.1.Kawakita,H.&Watanabe,J.-I.,2002.Ap.J.,572(2),pp.L177.Keller,H.U.,Delamere,W.A.,Huebner,W.F.,etal.,1987.Astron.Astroph.,187,pp.807–823.Keller,H.U.,Kueppers,M.,Fornasier,S.,etal.,2007.Sp.Sci.Rev.,128,p.433.Keller,H.U.,Mottola,S.,Skorov,Y.,andJorda,L.,2015.Astron.Astroph.,579,p.L5.Kim,S.J.,A'Hearn,M.F.&Cochran,W.D.,1989.Icarus,77,p.98.LeRoy,Altwegg,K.,Balsiger,H.,etal.,2015.Astron.Astroph.,583,p.A1.


Magrin,S.,LaForgia,F.,DaDeppo,V.,etal.,2015.Astron.Astroph.,574,p.A123.McConkey,J.W.,Malone,C.P.,Johnson,P.V.,etal.,2008.PhysicsReports,466(1),pp.1–103.Müller,U.&Schulz,G.,1992.J.Chem.Phys,96(8),p.5924.Müller,U.,Bubel,Th.,andSchulz,G.,199#.Z.Phys.D,25,p.167.Nilsson,H.,StenbergWieser,G.,Behar,E.,etal.,2015a.Science,347(6220),pp.aaa0571.Nilsson,H.,StenbergWieser,G.,Behar,E.,etal.,2015b.Astron.Astroph.,583,P.A20.Nishiyama,I.,Kondow,T.&Kuchitsu,K.,1979.Chem.Phys.Lett.,62(3),pp.462–465.Phillips,E.,Lee,L.C.&Judge,D.L.,1977.JournalofQuantitativeSpectroscopy,18,pp.309Rubin,M.,Gombosi,T.I.,Hansen,K.C.,etal.,2015.EMP,116,p141-157.p14Schleicher,D.G.,2010.AJ,140,p.973.Schleicher,D.G.&A'Hearn,M.F.,1988.ApJ,331,p.1058.Snodgrassetal.,2015,Astron.Astroph.,588,A80.Soderblom,L.A.,Becker,T.L.,Bennett,G.etal.,2002.Science,296(5),pp.1087.Streit,G.E.,Howard,C.J.,Schmeltekopf,A.L.,Davidson,J.A.,andSchiff,H.I.,1976.J.Chem.

Phys.,65,pp.4761.Tubiana,C.,Guettler,C.,Kovacs,G.,etal.,2015.Astron.Astroph.,583,p.A46.Tubiana,C.,Böhnhardt,H.&Agarwal,J.,2011.Astron.Astroph.527,pp.A113.Veverka,J.,Klaasen,K.,A'Hearn,M.,etal.,2013.Icarus,222(2),pp.424–435.Vigren,E.,Galand,M.,Lavvas,P.,andEriksson,A.I.,2015.ApJ,798,p130.Weaver,H.A.,Feldman,H.A.,A'Hear,M.F.,etal.,1999.Icarus,141,p.1.Wegmann,R.,Jockers,K.&Bonev,T.,1999.PSS,47,p.745.Woods,T.N.,Eparvier,F.G.,Hock,R.,etal.,2012.SolarPhysics,275(1),pp.115.


APPENDIXA:EMISSIONMODELToaidinourinterpretationofourresults,wehavedevelopedabasiccomamodel,summarizedbelow.Atruemodelwouldcombinearealisticdistributionoftheneutralgas,thecometaryandsolarwindplasma,andallrelevantchemicalandphysicalreactions,andthecomplexityofsuchamodelisoutsidethescopeofthispaper.

A1.NeutraldensitymodelTotalwaterproductionratesof67PwereestimatedusingtheempiricalrelationderivedbetweenwaterproductionQ(H2O)andheliocentricdistancerhbyFougereetal.(2016):

Q(H2O)~1.02x1028rh-4.2 (molecules/s) [A1]Weassumedconstant abundancesofCO2/H2O=3%,CO/H2O=3%,O2/H2O=3.8%, andHCN/H2O = 0.1% (LeRoy et al 2015; Fougere et al. 2016; Bieler et al. 2015). For thedistributionoffragmentspeciesweassumedastandardHasermodel(Festou1981,Combietal.2004).Thephotodissociationratesof thedifferentgasesare listed inTableA1,andthosewere scaledwith theheliocentricdistance as1/rh2.Weassumed that all gases areacceleratedtoabulkvelocityvgwithinadistanced=200kmfromthecenterofthenucleus(Combietal.2004)usingtheempiricalrelation:

&6 = 0.85 ∗ ;<=2 " ∗ 1 − exp =A

BCDE (km/s) [A2]

Weassumedthesamebulkoutflowgasvelocityforbothparentandfragmentspecies.

A2.ElectronsThe electron density in the coma was measured by the Langmuir (LAP) and mutualimpedanceprobes(MIP)onRosettabetween2015,February4–28(2.3AUfromtheSun;Edberg et al. 2015). For the electron density we assumed that the number density andradial distributionmeasured by LAP/MIP scaled linearlywith thewater production rateandthephotoionizationrate(thusthesquareoftheheliocentricdistance):

FG =HIJK2CJL ∗

2CMN

A ∗ OPQJ(m-3) [A3]

wherenedenotestheelectrondensityanddthedistancefromthecomet’ssurface.Ratherthanassumingatemperaturedistribution,weassumedafixedelectrontemperatureof30eV (3.3 x 106 m/s) for electron fluxes and for the cross sections of electron impactreactions. To calculate the emission resulting from dissociative electron impactdissociation,wemultipliedtheneutraldensitiesinthecoman(d)atadistancedfromthenucleus with the local electron density ne(d), electron velocity ve, and emission crosssectionσorwiththereactionratewhereapplicable(TableA1):

R S = T A ∗TU(A)∗VU∗WXY (ph/s/m3) [A4]

This is then integrated over the line of sight (from the position of the spacecraft Δ toinfinity) to produce the surface brightnesses B(Z)as a function of the angle betweenspacecraft-cometlinealongthecomet-Sunline:


[(Z) = \ AXY ]S

^∆ (ph/s/m2/sr) [A5]

Theradialdistancedfromthenucleusinanypositioninthecomet-sun-spacecraftplanecanbecalculatedusingthecosinerule.

A3.OIquenchingandtransportThe[OI]630nmemissionisdifferentfromtheotheremissionfeaturesbecauseitisaforbiddenline,andweexpectquenchingoftheOI1DstatebythereactionOI1D+H2Oà2OHtodecreasethe[OI]630nmemissionwithincreasingproductionrates.TheOI1Dstatehasalonglifetimeof101s(Atkinsonetal.1997).ToassesstheeffectofthedeactivationofOI1DthroughcollisionswithH2OonthesurfacebrightnessprofilewecombinedtheHasermodelfromSec.A1andadoptedexperimentalratecoefficientsof2x10-19m3/s/molecule(Streitetal.1976).TocorrectforexcitedOIatomslostthroughcollisionalquenching,wecalculateaneffectivedensityn’(d)byweightingtheneutralgasdensityfunctionn(d)fromequationA3byacorrectionfactor.AssumingaquenchingrateR(TableA1),thedistancel(d)anOIatommovingwithvelocityvdcantravelbeforecollidingis:

`(S) = Vab∗T(A) [A6]

Nowthefractionofatomsthatcanemitandthusisnotquenchedisgivenby:

Fc S = F S ∗ 1 − exp =d(A)Va∗e

[A7]

whereτisthelifetimeoftheOI1Dstate(101s),andn(d)thedensityofneutralwateratadistancedofthesurface.TheeffectofquenchingonthesurfacebrightnessisshowninFig.A1.Weassumedafixeddistancebetweencometandspacecraftof100kmtoevaluatetheeffectofquenchingandhaveaddedthe[OI]emissionfromallH2OandCO2processes(thedominantcontributionstotheOIemission,seeFig.5inthemaintext).InJanuary,onlythefirstkilometerisaffectedbyquenchingintheinnercoma.InJuly,theinner3kilometersareaffected,andat1kmapproximatelytheobservedsurfacebrightnessisapproximately50%oftheemittedlight. Inaddition,becauseofthelonglifetimesofthe1Dstate,onlypartoftheatomsdecaytothegroundstatewithinthefieldofview.Weaccountforthiswithasecondcorrectionfactor:

Fcc S = F′ S ∗ 1 − exp =AVa∗e

[A8]

Theresults,againforafixedspacecraftdistanceof100km,areshowninFigs.5cand5d.TheeffectoftheOI1Dlifetimeisdramaticandflattensthesurfaceemissioninthefirstfewkilometersaroundthenucleus.

A4.OpticaldepthWith increasing gas production rates, the optical depth of the inner coma increases andfewer photoelectrons are produced. To investigate the optical depth of the coma, wecalculatedH2OcolumndensitiesusingaHaserdistribution(Sec.3.3),andassumedwaterproductionratesfromtheempiricalformulabyFougereetal.(2016;seealsoFig.4a).Thedominant photons for ionization havewavelengths between 30 – 85 nm (Huebner et al.1992;Budzienetal.1994).Atthesewavelengths,H2Ohasphoto-absorptioncrosssectionsof1.0x10-21m2and1.6x10-21m2,respectively(Phillipsetal.1977).Theresultsareshown


inFig.6b.Therelevantregionhereisbetween1and10kmfromthesurface. InJanuary,only1to10%ofphotonswerelostwithinthisregion.Theeffectsoftheincreasingopticaldepth on the neutrals and electrons in the inner coma is not included in the surfacebrightnessmodel.


TableA1.Assumedreactionrates(atrh=1AU)andcrosssectionsReaction Product Crosssectionorrate Reference

H2O+hv any 1.04x10-5s-1 Combietal.2004H2O+hv OH+hv 8.5x10-4s-1 Combietal.2004H2O+hv OI(1S+1D) 8.6x10-7s-1 Bhardwaj&Raghuram2012CO2+hv any 2.2x10-6s-1 Weaveretal.1999CO2+hv OI(1D) 1.9x10-6s-1 Bhardwaj&Raghuram2012CO+hv any 3.3x10-6s-1 Huebneretal.1992CO+hv OI(1D) 9.1x10-8s-1 Bhardwaj&Raghuram2012O2+hv any 4.5x10-6s-1 Huebneretal.1992O2+hv OI(1D) 4.0x10-6s-1 Huebneretal.1992NH3+hv any NH3+hv NH HCN+hv any 7.5x10-5s-1 Huebneretal.1992HCN+hv CN 7.5x10-5s-1 Huebneretal.1992HCN+hv CN(B2Σ+) 4.5x10-5s-1 Frayetal.2005H2O+e- OH(A2Σ+) 8.5x10-22m2 Avakyanetal.1998H2O+e- OI(1S+1D) 9.7x10-16m3s-1 Bhardwaj&Raghuram2012CO2+e- OI(1S+1D) 5.3x10-14m3s-1 Bhardwaj&Raghuram2012CO+e- OI(1S+1D) 2.9x10-16m3s-1 Bhardwaj&Raghuram2012O2+e- OI(1S+1D) 1x10-14m3s-1 EstimatedfromAvakyanetal.1998O2+e- Xq+ 1.5x10-20m2 Straubetal.1996NH3+e- NH(A3Π1) 2.8x10-22m2 Müller&Schulz1992OI1D+H2O 2OH 2x10-16m3/s/molecule Streitetal.1976

changes in the physical environment of the inner …

Documents