ebook img

ApJ manuscript on the origin of the zodiacal cloud PDF

22 Pages·2010·1.56 MB·English
by  
Save to my drive
Quick download
Download
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview ApJ manuscript on the origin of the zodiacal cloud

Home Search Collections Journals About Contact us My IOPscience COMETARY ORIGIN OF THE ZODIACAL CLOUD AND CARBONACEOUS MICROMETEORITES. IMPLICATIONS FOR HOT DEBRIS DISKS This article has been downloaded from IOPscience. Please scroll down to see the full text article. 2010 ApJ 713 816 (http://iopscience.iop.org/0004-637X/713/2/816) The Table of Contents and more related content is available Download details: IP Address: 65.241.78.2 The article was downloaded on 29/03/2010 at 22:00 Please note that terms and conditions apply. TheAstrophysicalJournal,713:816–836,2010April20 doi:10.1088/0004-637X/713/2/816 (cid:2)C2010.TheAmericanAstronomicalSociety.Allrightsreserved.PrintedintheU.S.A. COMETARYORIGINOFTHEZODIACALCLOUDANDCARBONACEOUSMICROMETEORITES. IMPLICATIONSFORHOTDEBRISDISKS DavidNesvorny´1,2,PeterJenniskens3,HaroldF.Levison1,2,WilliamF.Bottke1,2,DavidVokrouhlicky´4, andMatthieuGounelle5 1DepartmentofSpaceStudies,SouthwestResearchInstitute,1050WalnutSt.,Suite400,Boulder,CO80302,USA 2CenterforLunarOrigin&Evolution,NASALunarScienceInstitute,Boulder,CO80302,USA 3CarlSaganCenter,SETIInstitute,515N.WhismanRoad,MountainView,CA94043,USA 4InstituteofAstronomy,CharlesUniversity,VHolesˇovicˇka´ch2,18000Prague8,CzechRepublic 5MuseumNationald’HistoireNaturelle,LaboratoiredeMine´ralogieetCosmochimieduMuse´um,61rueBuffon,75005Paris,France Received2009October26;accepted2010February28;published2010March26 ABSTRACT The zodiacal cloud is a thick circumsolar disk of small debris particles produced by asteroid collisions and comets. Their relative contribution and how particles of different sizes dynamically evolve to produce the observed phenomena of light scattering, thermal emission, and meteoroid impacts are unknown. Until now, zodiacal cloud models have been phenomenological in nature, composed of ad hoc components with properties not understood from basic physical processes. Here, we present a zodiacal cloud model based on the orbital propertiesandlifetimesofcometsandasteroids,andonthedynamicalevolutionofdustafterejection.Themodel isquantitativelyconstrainedbyInfraredAstronomicalSatellite(IRAS)observationsofthermalemission,butalso qualitativelyconsistentwithotherzodiacalcloudobservations,withmeteorobservations,withspacecraftimpact experiments,andwithpropertiesofrecoveredmicrometeorites(MMs).WefindthatparticlesproducedbyJupiter- familycomets(JFCs)arescatteredbyJupiterbeforetheyareabletoorbitallydecouplefromtheplanetanddrift downto1AU.Therefore,theinclinationdistributionofJFCparticlesisbroaderthanthatoftheirsourcecomets and leads to good fits to the broad latitudinal distribution of fluxes observed by IRAS. We find that 85%–95% of the observed mid-infrared emission is produced by particles from JFCs and <10% by dust from long-period comets.TheJFCparticlesthatcontributetotheobservedcrosssectionareaofthezodiacalcloudaretypicallyD ≈ 100μmindiameter.Asteroidaldustisfoundtobepresentat<10%.Wesuggestthatspontaneousdisruptionsof JFCs,ratherthantheusualcometaryactivitydrivenbysublimatingvolatiles,isthemainmechanismthatliberates cometaryparticlesintothezodiacalcloud.Theejectedmmtocm-sizedparticles,whichmayconstitutethebasic grain size in comets, are disrupted on (cid:2)10,000 yr to produce the 10–1000 μm grains that dominate the thermal emissionandmassinflux.BreakupproductswithD >100μmundergoafurthercollisionalcascadewithsmaller fragmentsbeingprogressivelymoreaffectedbyPoynting–Robertson(PR)drag.UponreachingD <100μm,the particles typically drift down to <1 AU without suffering further disruptions. The resulting Earth-impact speed and direction of JFC particles is a strong function of particle size. While 300 μm to 1 mm sporadic meteoroids are still on eccentric JFC-like orbits and impact from antihelion/helion directions, which is consistent with the aperture radar observations, the 10–300 μm particles have their orbits circularized by PR drag, impact at low speeds,andarenotdetectedbyradar.OurresultsimplythatJFCparticlesrepresent∼85%ofthetotalmassinflux atEarth.Sincetheiratmosphericentryspeedsaretypicallylow(≈14.5kms−1 meanforD = 100–200μmwith ≈12kms−1 beingthemostcommoncase),manyJFCgrainsshouldsurvivefrictionalheatingandlandonEarth’s surface. This explains why most MMs collected in antarctic ice have primitive carbonaceous composition. The presentmassoftheinnerzodiacalcloudat<5AUisestimatedtobe1–2×1019 g,mainlyinD = 100–200μm particles. The inner zodiacal cloud should have been >104 times brighter during the Late Heavy Bombardment (LHB) epoch ≈3.8 Gyr ago, when the outer planets scattered numerous comets into the inner solar system. The bright debris disks with a large 24 μm excess observed around mature stars may be an indication of massive cometarypopulationsexistinginthosesystems.Weestimatethatatleast∼1022,∼2×1021,and∼2×1020 gof primitive dark dust material could have been accreted during LHB by the Earth, Mars, and Moon, respectively. Keywords: comets:general–minorplanets,asteroids:general–zodiacaldust amountofscatteredlight(Hong1985;Kniessel&Mann1991; 1.INTRODUCTION Ishiguroetal.1999;Hahnetal.2002),theDopplershiftsofthe The zodiacal cloud is a dynamic assembly of meteoroids in solar Mg i Fraunhofer line (Hirschi & Beard 1987; Mukai & bound orbits around the Sun. The orbits depend on particle Mann1993;Clarkeetal.1996;Reynoldsetal.2004),andthe size, location in the cloud, and the type of parent body. more easily to interpret thermal emission observed in various Interstellardustparticlesthatpassthroughthesolarsystemare linesofsight(Kelsalletal.1998;Marisetal.2006).Particularly notconsideredinthispaper,noraresmallmeteoroidfragments good scattered light observations came from the Clementine that move out of the solar system on hyperbolic orbits (“beta- mission(Hahnetal.2002),whilethermalinfraredobservations meteoroids”). aremostlyfromtheInfraredAstronomicalSatellite(IRAS;Low Traditionally, the zodiacal cloud has been described with etal.1984;Hauseretal.1984;Goodetal.1986;Sykes1990), phenomenological models of dust distributions to explain the the Cosmic Background Explorer (COBE; Reach et al. 1995; 816 No.2,2010 COMETARYORIGINOFTHEZODIACALCLOUD 817 Kelsall et al. 1998), the Midcourse Space Experiment (MSX; all current satellite impact models use meteoroid velocity Priceetal.2003),theInfraredSpaceObservatory(ISO;Fixsen distributions (both magnitude and spatial direction) derived &Dwek2002;Leinertetal.2002;Reachetal.2003;Mueller from meteor observations (Taylor & Elford 1998; Jones & et al. 2005), and the Spitzer Space Telescope (Bhattacharya & Brown1993;Brown&Jones1999;Brown&Campbell-Brown Reach2004;Reachetal.2007). 2003), which pertain to much bigger particles than typically Thephenomenologicalmodelssuccessfullydescribethesize, encounteredbysatellites. spatial and velocity distributions of dust particles in the solar In this paper, we investigate what fraction of the zodiacal system(e.g.,Gru¨netal.1985;Divine1993;Dikarevetal.2004). cloudisduetocometaryversusasteroidaldustbycalculatingthe They are particularly useful for accessing the satellite impact evolutionofdustparticlesundersolarradiationforcesandplan- hazard,designingspacecraftimpactexperimentsandstudiesof etaryperturbations(includingresonancesandcloseencounters), extrasolaremissionsourcessuchasthecosmicmicrowaveback- ejectedfrommodelpopulationsofallpotentialsources(notjust ground(e.g.,Kelsalletal.1998).Thephemenologicalmodels, representativeexamples).Thesourcepopulationsincludeaster- however, fall short in answering basic questions related to the oids,activeandmostlydormantJupiter-familycomets(JFCs), originofthezodiacalcloud,itstemporalbrightnessvariability, Halley-typecomets(HTCs),andlong-periodOort-cloudcomets and the provenance of interplanetary particles collected at the (OCCs).Inrecentyears,muchinsightwasgainedintothedy- Earth.Consequently,theoriginofthezodiacalcloud,interplane- namicalcharacteristicsofthesepopulationsandthenumberof tarydustparticles(IDPs)collectedinEarth’sstratosphere(e.g., asteroids and comets that can contribute dust to the zodiacal Love & Brownlee 1993), and micrometeorites (MMs) on the cloud(e.g.,Levison&Duncan1994,1997;Jedicke&Metcalfe ground (Taylor et al. 1996; Engrand & Maurette 1998; Farley 1998; Wiegert & Tremaine 1999; Dones et al. 2004; Francis etal.1998,2006;Genge2006)isstillamatterofconsiderable 2005; Gladman et al. 2009). At the same time, it was realized debate. This limits our ability to link the detailed laboratory that mostly dormant JFCs are the main source of meteoroid studiesofIDPsandMMstothepropertiesoftheirparentbod- streamsintheinnersolarsystem(Jenniskens2006,2008)and ies, and to use the zodiacal cloud as a valuable reference for responsible for the antihelion/helion sources in the sporadic studiesoftheexozodiacaldebrisdisks. meteoroidbackground(Jenniskens2006;Wiegertetal.2009). Detailed dynamical models can be more useful in this Here,wecouplethesenewinsightstodynamicalbehaviorof context. At the root of dynamical models are the physical cometaryandasteroidaldustinordertoevaluatethecontribution properties of interplanetary dust, such as density, geometric ofvarioussourcestothethermalemissionofzodiacaldustand albedo, elemental composition, mineralogy, tensile strength, the influx of MMs.We show thatJFCs are the main source of heat capacity, etc. (e.g., Dumont & Levasseur-Regourd 1988; zodiacaldustinside5AUandthemostlikelysourceoftheMMs McDonnell&Gardner1998;Gustafson1994;Gustafsonetal. foundonEarth.Ourresultsalsoprovidequantitativeconstraints 2001;Levasseur-Regourdetal.2001),whichdeterminethebe- ondustlifetimes,influxrates,andvelocitydistributionsdirectly haviorofparticlesininterplanetaryspace(e.g.,planetarypertur- from the known abundances of meteoroid parent bodies. The bations,collisions,sublimation,sputtering)andtheirinteraction resultsareusedtoquantifythepropertiesofzodiacaldustcloud with a detector (e.g., ablation of MMs in Earth’s atmosphere, in the past and discuss implications for studies of exozodiacal 3Heretention,thermalradiation,lightscattering).Indynamical debrisdisks. models,theindividualparticlesaretrackedbynumericalcodes TosetupthestageforourmodelingdescribedinSection3,we astheyevolvebyvariousprocessesfromtheirsources(assumed discuss IRAS observations of the zodiacal cloud in Section 2. to be, e.g., asteroids, comets, satellites or Kuiper Belt objects) Results are presented in Section 4. In Section 5, we estimate tosinks(e.g.,whentheysublimate,disrupt,impactorleavethe thecurrentandhistoricalterrestrialaccretionratesofdustand solarsystem).Insightsintotheoriginofthezodiacalcloudcan discuss the implications of our work for studies of MMs and be obtained by calibrating the results of dynamical models on debris disks. Comet disruptions/splitting events are reviewed observations. in Section 6. We suggest that they are the main mechanism Untilnow,detaileddynamicalmodelshavebeenonlydevel- bywhichthecometaryparticlesareliberatedfromtheirparent oped for asteroidal dust to explain the origin of the zodiacal bodies into the zodiacal cloud. The mass-loss rate from JFCs dustbands(e.g.,Dermottetal.1984;Groganetal.1997,2001; mustbelargeenoughtosupplytheIDPcomplexagainstlosses Reach et al. 1997; Nesvorny´ et al. 2006; Vokrouhlicky´ et al. (Lisse2002).Previousworkandoriginofdustparticlesbeyond 2008) and trapped dust in Earth’s Langrange points first seen JupiterarediscussedinSections7and8,respectively. inIRASobservations(Dermottetal.1994b).Ithasbeenestab- 2.CONSTRAINTS lishedthatthedustbandsoriginatefromtheyoungestasteroid families (Nesvorny´ et al. 2003, 2008). However, claims that Ourprimaryconstraintsaretheobservationsofthezodiacal asteroids are a major if not dominant source of zodiacal dust, cloudbyIRASwhichhavebeenconfirmedbyCOBEandSpitzer by assuming that all main belt asteroids contribute dust (e.g., (Hauseretal.1984;Lowetal.1984;Kelsalletal.1998;Sykes Dermott et al. 1995; Durda & Dermott 1997; Kortenkamp & etal.2004).IRASmeasuredmid-infrared(MIR)fluxesinfour Dermott1998),haveremainedindoubt. filters with effective wavelengths of 12, 25, 60, and 100 μm. Models of the zodiacal cloud need not only explain line- Thesefilterscanbeusedaswindowsintothedustdistribution of-sight properties, but also the observed influx of meteors atdifferentdistancesfromtheSun.Measurementsinthe12μm (see Ceplecha et al. 1998; Jenniskens 2006, for a review) and IRAS band are mainly sensitive to distributions of particles the impact rate of meteoroids on satellites (Love & Brownlee at ∼1–2 AU, while the 25 and 60 μm band measurements 1993).Untilnow,modelsthatweredevelopedtoexplainthese preferentially detect thermal emission from larger distances. dynamical phenomena (e.g., Gru¨n et al. 1985; Divine 1993; IRASobservationsinthe100μmbandarelessusefulforprobing Staubachetal.1997;Dikarevetal.2004)werebasedonadhoc thethermalradiationofdustparticlesintheinnersolarsystem. populations of meteoroids in various types of orbits without IRAS showed that the MIR brightness of the zodiacal cloud a dynamical underpinning to sources and sinks. Moreover, peaksattheeclipticandhasbroadwingsthatextendalltheway 818 NESVORNY´ ETAL. Vol.713 Figure1.UppersolidlinesineachpanelshowIRASscan180_24(seeTable3inNVBS06)thathasbeensmoothedbyalow-passfiltertoremovepointsourcesand instrumentalnoise.Differentpanelsshowfluxesat12,25,60,and100μmIRASwavelengths.Thebottomsolidlinesshowthecontributionofthreemainasteroiddust bandstotheobservedfluxes.AccordingtoNVBS06,thesedustbandscontributetotheobservedfluxesby≈9%–15%within10◦totheecliptic,and<5%overall. Thestrongsignalat100μmbetweenlatitudesb≈−80◦andb≈−30◦isthegalacticplaneemission(alsoapparentat60μm).FigurefromNVBS06. to the ecliptic poles (Figure 1). The variation with the ecliptic bygravitationalperturbationsfromJupiter(Dermottetal.1995, longitude is minimal indicating that the zodiacal cloud is a 2001).Sincewearenotinterestedinthesedetailedfeaturesof nearlysymmetricaldiskofcircumsolarparticlesthatisroughly the zodiacal cloud here, we removed them by combining the centeredattheSun(e.g.,Staubachetal.2001).Thesignificant selectedIRASscansintoarepresentativeprofile.Thiswasdone flux received from the ecliptic poles (∼1/3–1/4 of the ecliptic byfirstshiftingthescansbyasmallvalueinlatitude(<2◦),so flux) also suggests that the cloud must be rather thick in the that the peak of emission was centered exactly at b = 0, and normaldirectiontotheeclipticplane. calculatingtheaveragefluxfromallselectedscansasafunction Following Nesvorny´ et al. (2006, hereafter NVBS06), we ofb.Weendedupwithprofilesshowingthemeanfluxesat12, selectedseveralrepresentativeIRASscansthatmetthefollowing 25, and 60 μm wavelengths as a function of ecliptic latitude conditions.(1)Weusedscansthatcoveredacontinuousrange (Figure2).Theseprofilesrepresentthemainconstraintsonthe of ecliptic latitudes from b < −80◦ to b > 80◦. (2) We work described here. Additional constraints are discussed in did not use scans that had gaps created when the telescope Section4.3. skipped over bright sources. (3) We did not use scans that showed strong emission from extrasolar sources such as the 3.MODEL galacticplane,galacticcirrus,pointsources,etc.(4)Werequired that the selected scans covered all values of ecliptic longitude Our model of the zodiacal cloud has four parts: we (1) and the available range of solar elongations, l(cid:6). Tables 2 and define the initial orbital distributions of particles for different 3 in NVBS06 list the basic information about the selected sources (asteroid and comet populations), (2) track the orbital scans. evolutionofparticleswithvarioussizesfromsourcestosinks, Inthiswork,weonlyconsiderscanswithl(cid:6) ≈ 90◦.Thisis (3)determinethethermalinfraredemissionfromthesesynthetic mainly done because the detail variation of brightness with l(cid:6) particle distributions, and (4) model the detection of their is difficult to characterize as it requires an appropriate model emission by IRAS. These model components are described forthecollisionaldisruptionofparticles(NVBS06).Wedonot below. includetheeffectsofcollisionsbetweenparticlesinourmodel To simplify things, we do not initially account for the (except in Section 4.2, where a simple model for collisions is collisional disruptions of dust particles in the model. This is a used).Theconsideredvalueofl(cid:6)isinthemiddleoftheavailable majorassumptionwhichweverifyinSection4.2.Forexample, rangeandthusbestrepresentsIRASobservations. NVBS06 showed how the collisional disruption of particles, Thezodiacalcloudisknowntobewarpedandhaveacenter andproductionofsmallerdaughterproducts,affectthespatial thatisslightlyoffsetfromtheSun.Thesefeaturesareproduced distributionofparticlepopulationsintheasteroidaldustbands. No.2,2010 COMETARYORIGINOFTHEZODIACALCLOUD 819 dustbandshavebeenidentified(Sykes1990).Sincethesedust bandsaremuchfainterthanthethreemaindustbands,Nesvorny´ etal.resultssettheupperlimitonthecontributionofidentified asteroidbreakupstothezodiacalcloud. Theorbitaldistributionoftheasteroidbeltsource(2)ismod- eledbyusingtheobservedorbitaldistributionofasteroidswith D > 15 km. We obtained this distribution from the ASTORB catalog (Bowell et al. 1994). This sample should be complete andunbiased(Jedicke&Metcalfe1998;Gladmanetal.2009). TheorbitaldistributionofasteroidswithD <15km,whichmay beabetterproxyfortheinitialdistributionofdustproducedin mainbeltcollisions(Sykes&Greenberg1986),isroughlysim- ilartothatoflargeasteroids.Thus,theorbitaldistributionthat weuseshouldbeareasonableassumptionfor(2). The orbital distributions obtained by numerical integrations oftestparticletrajectoriesfromindividualcometswouldsuffer Figure2.MeanIRASprofilesat12,25,and60μmwavelengths.Tomakethese from the heavily biased comet catalogs. Moreover, there are profiles,theselectedIRASscanswerecenteredattheecliptic,smoothedbya hundreds of known comets, each producing dust at a variable low-passfilter,andcombinedtogether.Thegrayrectanglesatl < −78◦ and 40◦<l<70◦blockthelatituderangewherethemeanfluxesweresignificantly andtypicallyunknownrate.Inaddition,itisalsopossiblethat the present zodiacal dust complex contains particles from lost affectedbythegalacticplaneemission.Wedonotusetheexcludedrangein thiswork.Theuncertaintiesofthemeanfluxvaluesarenotshownherefor parentsassuggestedbytheorphanedType-IItrails(Sykes1990) clarity; they are too small to clearly appear in the plot. The characteristic andidentificationofmeteoroidstreamswithdisruptedJFCs(see errors at different wavelengths averaged over latitudes are σ12μm = 0.59 Jenniskens(2008)forareview).Inviewofthesedifficulties,we MJysr−1,σ25μm =1.1MJysr−1,andσ60μm =2.7MJysr−1.Theyincrease resortedtothefollowingstrategy. withwavelengthduetothelargerroleofgalacticemissionatlongerwavelengths. The orbital distribution of JFCs was taken from LD97 who followedtheevolutionofbodiesoriginatingintheKuiperBelt Theworkdescribedhere,however,islesssensitivetocollisional as they are scattered by planets and evolve in small fractions processesbecausetheoverallgrossshapeofthezodiacalcloud into the inner solar system. Starting at the time when the should be mainly controlled by the orbital distribution of its comets’periheliondistance,q,firstdropsbelow2.5AUinthe source population(s), violent dynamics of particles moving on LD97 simulations, we include it as an active JFC in our list planet-crossing orbits, and the effects of Poynting–Robertson of source objects. LD97 showed that the orbital distribution (PR)drag.Thesefeatures/processesareincludedinourmodel of visible JFCs obtained in this way nicely approximates aswedescribebelow. the observed distribution. Moreover, LD97 argued that the inclinationdistributionofnewJFCs(reachingq < 2.5AUfor 3.1.Sources the first time) is relatively narrow. The inclination distribution widensatlatertimesastheJFCorbitsbecomemorespreadby Ourmodelstartswiththesourcepopulationofobjects.This Jupiterencounters. may be either of the following: (1) individual asteroid groups Bycomparingthewidthofthemodelinclinationdistribution suchastheKarin,Veritas,orBeagleasteroidfamilies(NVBS06; with that of the observed JFCs, LD97 were able to estimate Nesvorny´ et al. 2008), (2) asteroid belt as a whole, (3) active the fading lifetime of JFCs, t , defined as the characteristic JFC JFCsdefinedbytheirdebiasedorbitaldistributionobtainedfrom time between their first and last apparitions. They found that Levison & Duncan (1997, hereafter LD97) and their physical t = 12,000 yr with a 90% confidence interval 3000 < JFC lifetime, (4) JFCs with orbital distribution similar to (3) but t <30,000yr.SeeFigure3foranillustrationofthesteady- JFC dynamically evolved beyond their nominal active lifespan, (5) state JFC orbit distribution for t = 12,000 yr. This is our JFC HTCs, and (6) OCCs. These source populations are described initialdistributionofparticlesproducedbyactiveJFCs.Weuse inamoredetailbelow. t asafreeparameterinourmodelwithvalues extending to JFC We ignore the contribution of interstellar dust particles t =100,000yrtoaccountforthepossibilitythatanimportant JFC becausethermalemissionfromthesesmallparticles(diameter dustcomponentmaybeproducedbyolddormantJFCsasthey D <1μm)wouldcreatediagnosticspectralfeaturesintheMIR spontaneouslydisrupt. wavelengthsthatarenotobserved(e.g.,Reachetal.2003).More OurmodelsfortheorbitaldistributionsofHTCsandOCCs specifically, observations of the MIR spectrum of the zodiacal are simpler than the one described for JFCs above, because cloudbyReachetal.(2003)suggestthatthebulkofthezodiacal we do not have in hand an appropriate numerical model that cloudisproducedby∼10–100μmparticles.Wealsoignoredust we could use with confidence. Fortunately, this is not a major producedbydisruptivecollisionsintheKuiperBelt(hereafter limitationfactorinthisworkbecauseourmainresultsdescribed KB dust) because Landgraf et al. (2002) and Moro-Mart´ın & in Section 4 are not sensitive to the detailed properties of the Malhotra(2003)showedthatKBdustparticlesshouldrepresent HTCsandOCCspopulations. onlyaminorcontributiontotheinnerzodiacalcloud. ForHTCs,weassumethatthedifferentialdistributionofthe Resultsfor(1)weretakenfromNVBS06andNesvorny´ etal. periheliondistance,dN(q),isdN(q)∝qdq,andsetanupper (2008)whofoundthatthethreemaindustbandsdiscoveredby limitofqat3AU.HTCstypicallybecomevisible/activeonlyif Lowetal.(1984)originatefromtheKarin,Veritas,andBeagle theyreachq <3AU.Similarly,thecumulativesemimajoraxis asteroidfamilies.Accordingtotheseresults(Figure1),thethree distribution of HTCs, N(< a), is assumed to be N(< a) ∝ a maindustbandsrepresentonly≈9%–15%ofthezodiacalcloud with an upper cut on a at 50 AU. The differential inclination emissionattheeclipticand<5%overall.Aboutadozenother distribution dN(i) is taken from Levison et al. (2006). This 820 NESVORNY´ ETAL. Vol.713 Theindividualcometsinourmodelareassumedtocontribute inroughlythesameproportiontothecircumsolardustcomplex. The reasoning behind this assumption is that if an individual super-cometwerethedominantsourceofcircumsolardust,the zodiacalcloudwouldnothavesuchasmoothandsymmetrical structure. In fact, the observed smooth structure of the zodi- acal cloud probably implies a source population that contains numerousobjectsthatarewellmixedintheorbitalspace. 3.2.OrbitEvolution The orbits of particles were tracked using the Wisdom– Holman map (Wisdom & Holman 1991) modified to include effects of radiation forces (Burns et al. 1979; Bertotti et al. (cid:9) 2003).TheaccelerationF onaparticleduetotheseforcesis (cid:2)(cid:3) (cid:4) (cid:5) ˙ (cid:9) (cid:9) F(cid:9) =βGm(cid:6) 1− R R − V , (2) R2 c R c (cid:9) (cid:9) where R is the orbital radius vector of the particle, V is its velocity,Gisthegravitationalconstant,m(cid:6) isthemassofthe Sun,cisthespeedoflight,andR˙ = dR/dt.Theacceleration (2) consists of the radiation pressure and velocity-dependent Figure3.OrbitaldistributionofJFCsfromtheLD97model.Panelsshowthe PR terms. Parameter β is related to the radiation pressure periheliondistance(a),andinclination(b),asfunctionsofthesemimajoraxisfor coefficient,Q ,by JFCswithq<2.5AUandtJFC=12,000yr.SeeSection3.1forthedefinition pr oftJFC.The2:1and3:2mean-motionresonanceswithJupitercorrespondtothe Q gapsinthedistributionata ≈3.3and3.96AU,respectively.Theinclination β =5.7×10−5 pr, (3) distributionofJFCsshownhereisremarkablysimilartothatobtainedbyDi ρs Sistoetal.(2009;theirFigure10). where radius s and density ρ of the particle are in cgs units. distributioncanbeapproximatelydescribedby PressurecoefficientQprcanbedeterminedusingtheMietheory (Burnsetal.1979).WeusedQ =1whichcorrespondstothe pr dN(i)∝sin(i)e−12(σi)2di (1) geometricalopticslimitwheresismuchlargerthantheincident- lightwavelength.Weassumedthatthesolar-winddragforcehas withσ =30◦. thesamefunctionalformasthePRtermandcontributesby30% According to√Francis (2005), the long-period comets have tothetotaldragintensity(Gustafson1994). dN(q) ∝ (1+ q)dq forq < 2AU.Forq > 2AU,Francis’ WeusedparticleswithdiameterD =2s =10,30,100,200, study predicts dN(q) being flat or declining while we would 300, 1000 μm and set their bulk density to ρ = 2 g cm−3. expecttheperiheliondistributiontoincreasewithq.Itprobably For comparison, Love et al. (1994) reported ρ ≈ 2 g cm−3 just shows that the distribution is not well constrained for for stratospheric-collected IDPs, while McDonnell & Gardner q > 2 AU. We use dN(q) ∝ 2.41(q/2)γ dq for q > 2 AU (1998)foundmeanρ =2–2.4gcm−3fromtheanalysisofdata with 0 < γ (cid:3) 1. The semimajor axis values of OCCs are set collectedbytheLDEFandEurecasatellites.Ontheotherhand, between 10,000 and 50,000 AU, which is known as the Oort densityof1gcm−3hasbeenoftenassumedforcometarymatter spike (Wiegert & Tremaine 1999). We also use a = 1000 AU (e.g.,Joswiaketal.2007;Wiegertetal.2009).Gru¨netal.(1985) to check on the dynamical behavior of dust particles launched suggested that 20%–40% of particles may have low densities fromthereturningOCCs.TheinclinationdistributionofOCCs whereas most meteoroids have ρ = 2–3 g cm−3. Our results issettobeuniformlyrandombetween0◦and180◦. maybeeasilyrescaledtoanyρ valueandweexplicitlydiscuss Wetestedtwodifferentmethodstolaunchparticlesfromtheir theeffectofρ whereveritisappropriate. source objects. In the first method, chosen to approximate the The particle orbits were numerically integrated with the ejectionofparticlesfromactivecomets,welaunchedparticles swift_rmvs3 code (Levison & Duncan 1994) which is an attheperihelionwhenthemeananomalyM =0.Inthesecond efficient implementation of the Wisdom–Holman map and method, we launched particles along the orbit with uniform which, in addition, can deal with close encounters between distribution in M. This second method is more appropriate for particles and planets. The radiation pressure and drag forces the asteroidal particles and for particles produced by comet wereinsertedintotheKeplerianandkickpartsoftheintegrator, disruptions. Indeed, the identified comet disruptions do not respectively. The change to the Keplerian part was trivially seem to be correlated in any way with the perihelion passage done by substituting m(cid:6) by m(cid:6)(1−β), where β is given by (e.g., Weissman 1980). The two ejection methods produced Equation(3). comparableresults.Giventhatdisruptionsshouldrepresentthe Thecodetrackstheorbitalevolutionofaparticlethatrevolves main mass loss in comets (see Section 6), we use the second aroundtheSunandissubjecttothegravitationalperturbations methodforallresultspresentedinSection4.Tosimplifythings, of seven planets (Venus to Neptune) until the particle impacts we neglect the ejection velocities of dust particles from their a planet, is ejected from the solar system, or drifts to within parent object and assume that they initially follow the parent 0.03 AU from the Sun. We removed particles that evolved to object’sorbitmodifiedbytheradiationpressure. R <0.03AUbecausetheorbitalperiodforR (cid:2)0.03AUisnot No.2,2010 COMETARYORIGINOFTHEZODIACALCLOUD 821 properly resolved by our integration time step. In Section 4.2, thermallyradiatingparticlewithdiameterD.S(R,L,B)defines wealsoconsidertheeffectofcollisionaldisruptionbyremoving thespatialdensityofparticlesasafunctionoftheheliocentric particles from the simulations when they reach their assumed distance, R, longitude, L, and latitude, B (all functions of r physicallifespan.Wefollowed1000–5000particlesforeachD, as determined by geometry from the location and pointing source,andparametervalue(s)thatdefinethatsource. direction of the telescope). N(D;a,e,i) is the number of particleshavingeffectivediameterDandorbitswithsemimajor 3.3.ThermalEmissionofParticles axis,a,eccentricity,e,andinclination,i. WeevaluatetheintegralinEquation(6)bynumericalrenor- Particles were assumed to be isothermal, rapidly rotating malization(seeNVBS06).F(D,r)iscalculatedasdescribedin spheres.Theabsorptionwasassumedtooccurintoaneffective Section 3.3. N(D;a,e,i) is obtained from numerical simula- crosssectionπs2,andemissionoutof4πs2.Theinfraredflux tionsinSection3.2.S(R,L,B)usestheoreticalexpressionsfor density (per wavelength interval dλ) per unit surface area at the spatial distribution of particles with fixed a, e, and i, and distancerfromathermallyradiatingparticlewithradiussis randomizedorbitallongitudes(Kessler1981;NVBS06). s2 Weassumethatthetelescopeislocatedat(xt =rtcosφt,yt = F(λ)=(cid:8)(λ,s)B(λ,T) , (4) r sinφ,z = 0) in the Sun-centered reference frame with r2 rt =1At Ut.Itsviewingdirectionisdefinedbyaunitvectorwith t where (cid:8) is the emissivity and B(λ,T) is the energy flux at components(xv,yv,zv).InEquation(6),thepointingvectorcan (λ,λ+dλ)persurfaceareafromablackbodyattemperatureT: bealsoconvenientlydefinedbylongitudelandlatitudebofthe pointingdirection,wherex =cosbcosl,y =cosbsinl,and v v B(λ,T)= 2πλh5c2 (cid:6)ehc/λkT −1(cid:7)−1. (5) lz(cid:6)v ==si9n0b◦.AansddecsaclrciubleadteinthSeectthioenrm2,awl eflufixxtohfevsoalraiorueslopnagrattiicolne populations as a function of b and wavelength. The model Inthisequation,h=6.6262×10−34JsisthePlanckconstant, brightness profiles at 12, 25, and 60 μm are then compared c = 2.99792458 × 108 m s−1 is the speed of light, and withthemeanIRASprofilesshowninFigure2. k =1.3807×10−23 JK−1istheBoltzmannconstant.Weused To check our code, known as Synthetic InfraRed Telescope (cid:8) =1whichshouldberoughlyappropriateforthelargeparticles (SIRT), we programmed a particle version of the algorithm, usedinthiswork.SeeNVBS06foramoreprecisetreatmentof whichshouldbeinmanywayssimilartothecorealgorithmin (cid:8)(λ,s)fordustgrainscomposedofdifferentmaterials. SIMUL (Dermott et al. 1988; see also Dermott et al. 2001). TodeterminethetemperatureofaparticleatdistanceRfrom The particle version inputs the orbital elements of particles the Sun, we used the temperature variations with R that were obtained in the orbital simulations (Section 3.2) and produces proposed by different authors from spectral observations of their orbit clones that are uniformly spread over 2π in mean the zodiacal cloud (e.g., Dumont & Levasseur-Regourd 1988; anomalyM.Thus,everytestparticleisassumedtotraceacloud Renardetal.1995;Leinertetal.2002;Reachetal.2003).For of real particles having the same orbit as the test particle but example,Leinertetal.(2002)proposedthatT(R)=280/R0.36 different angular locations along the orbit. See Vokrouhlicky´ K near R = 1 AU from ISOPHOT spectra. We used T(R) = et al. (2008; their Section 2.5) for a technical description of T /Rδ K, where T is the temperature at 1 AU and δ is the algorithm. This procedure is based on the assumption that 1AU 1AU a power index. We varied T and δ to see how our results any concentration of particles with a specific M value would 1AU depend on these parameters. Values of T ≈ 280 K and be quickly dispersed by the Keplerian shear. We employ this 1AU δ = 0.5 correspond to the equilibrium temperature of large proceduretoimprovetheresolution.Withoutit,thenumberof darkparticles.Valuesδ <0.5wouldbeexpected,forexample, integrated test particles would be too small to obtain a useful for fluffy particles with small packing factors (e.g., Gustafson result. et al. 2001). Note that the power law is used here as a simple Tobeabletocomparetheresultsoftheparticlealgorithmwith empiricalparameterizationofthetemperaturegradient.Thereal SIRT,theparticlealgorithmmustalsousesmoothdistributions temperaturegradientislikelytobeamorecomplicatedfunction in perihelion longitude (cid:11) and nodal longitude Ω. This is oftheheliocentricdistance(andparticleproperties). achieved by generating additional orbital clones with (cid:11) and Ω uniformly spread over 2π. Figure 4 shows examples of the 3.4.SyntheticObservations results obtained from the particle algorithm and SIRT. The agreement between the two codes is excellent which gives us TocompareourresultswithIRASobservationsdescribedin confidencethatbothcodesworkproperly.WefindthattheSIRT Section 2, we developed a code that models thermal emission algorithmbasedontheKesslerdistributionismuchfasterthan from distributions of orbitally evolving particles and produces theparticleone.Forthisreason,weusetheoriginalSIRTcode infrared fluxes that a space-borne telescope would detect de- inthisstudy. pendingonitslocation,pointingdirection,andwavelength.See NVBS06foradetaileddescriptionofthecode. 4.RESULTS In brief, we define the brightness integral along the line of sightofaninfraredtelescope(definedbyfixedlongitudeland Our primary model parameters are the relative contribution latitudebofthepointingdirection)as ofdifferentsourcestothezodiacalcloud.Thetotalmodelflux (cid:8) (cid:8) (cid:8) asafunctionoftheeclipticlatitudeisobtainedas ∞ dadedi drr2 dDF(D,r)N(D;a,e,i)S(R,L,B), F (b)=α F +α F +α F +α F , a,e,i 0 D model AST AST JFC JFC HTC HTC OCC OCC (6) (7) where r is the distance from the telescope and F(D,r) is the whereαarecoefficientsthatsatisfyα +α +α +α = AST JFC HTC OCC infrared flux (integrated over the wavelength range of the 1,andF ,F ,F ,andF aremodelfluxesobtainedfor AST JFC HTC OCC telescope’s system) per unit surface area at distance r from a differentsources.Wenormalizethemsothattheeclipticmodel 822 NESVORNY´ ETAL. Vol.713 Figure 4. Comparison between results obtained from the particle algorithm Figure5.Comparisonofthe25μmprofilesproducedbydifferentsourceswith (dots)andSIRT(solidlines).Case1correspondstoasteroidalparticleswith IRASobservations.TheblacklineshowsourmeanIRASscanforl(cid:6) = 90◦. a =2.5AU,e=0.1,andi =10◦.Case2correspondstocometaryparticles The colored lines show profiles expected from different source populations: crossingEarth’sorbitwitha = 1AU,e = 0.5,andi = 50◦.Inbothcases, asteroids(green),JFCs(red),andHTCs(blue).TheOCCflux,notshownhere we assumed that particles have D = 100 μm and are distributed randomly forclarity,isanearlyconstantfunctionoflatitude.Themaximumfluxineach inΩ,(cid:11),andM.Thefluxat25μmwasnormalizedtoapopulationof1015 profilehasbeennormalizedto1.WeusedD=100μmandtJFC=12,000yr. particlesinCase1and2×1015particlesinCase2.Observationswithrt=1AU ThemaindifferencesbetweenprofilesarenotsensitivetotheexactchoiceofD andl(cid:6) =90◦wereassumed.InCase2,theparticlealgorithmshowsascatter andothermodelparameters. aroundtheexactsolutionduetotheroughresolutionofthedistributionnearthe telescope’slocation.Weused5×1010orbitclonesintheparticlealgorithm. fluxfromeachsourceisequaltothatofthemeanobservedflux atb =0.Coefficientsαthereforegivetherelativecontribution ofdifferentsourcesattheecliptic. The model flux profiles depend on the particle size, wave- length, and for JFCs also on the assumed value of t . As JFC described in Section 3.2, we tracked particles with 10 < D <1000μm.Thesedifferentsizesaretreatedindividuallyin Equation(7).Inparticular,wedonotattempttoconstructplau- sible size–frequency distributions for different sources. It is therefore assumed that a single characteristic particle size, or a narrow range of sizes, can be effectively used to model the observedMIRflux.Thisassumptionneedstobeverifiedlater. InSection4.1,wefirstconsideramodelwherethelifespan of particles is limited by their dynamical lifetime. Effects of Figure6.Dependenceoftheshapeof25μmprofilesproducedbyasteroidal particledisruptionsarediscussedinSection4.2. particlesonD.Thedashedlineshowsthemean25μmIRASprofileforl(cid:6)=90◦. Theuppersolidcurvesshowthemodelresultsforthesamewavelengthand elongation.Thebottomlinesshowtheresidualfluxobtainedbysubtractingthe 4.1.Collision-freeModel modelfluxfromthemeanIRASprofile.ResultsforD=30,100,and300μm asteroidalparticlesareshownwithslightlybroaderprofilescorrespondingto Figure 5 shows the 25 μm flux of D = 100 μm particles largerD.TheprofilesforD=10and1000μm,notshownhere,arenarrower produced by different source populations. Note that these than the ones for D = 30 μm. For D = 1000 μm, this is mainly due to profiles do not sensitively depend on the particle size (see theeffectsofdisruptivecollisionsthatdestroylargegrainsbeforetheycould evolvedownto1AU(seediscussioninSection4.2).Noneofthemodelprofiles Figures6and7fortheresultsfordifferentD).Instead,themain obtainedwithasteroidalparticlescanmatchIRASobservations. differences between results for different source populations in Figure5reflecttheinitialorbitdistributionofparticlesineach source and their orbit evolution. Therefore, these profiles can The profile produced by HTC particles is much broader helpustoidentifythesourcepopulation(s)thatcanbestexplain than the observed one (Figure 5). In this case, the magnitude IRASobservations. of the polar fluxes is ≈1/2 of that near the ecliptic. This The asteroidal particles produce a profile with a very sharp result reflects the very broad inclination distribution of HTCs peak centered at the ecliptic. The emission from asteroidal (Section3.1).ApotentiallysignificantcontributionofHTCsto particles near the ecliptic poles is relatively faint. The polar the zodiacal cloud is also problematic because the two large emission comes from the particles that evolved by PR drag HTCs, 109P/Swift-Tuttle and 1P/Halley, tend to librate about froma > 2AUtoR = 1AU.Whilemostasteroidalparticles mean-motion resonances, causing relatively stable orbits for indeed reach 1 AU, they pass too briefly near b = ±90◦ to longperiodsoftime.Thus,dustreleasedbyHTCsisexpected produce important polar fluxes. This is why most radiation is to be concentrated along certain locations on the sky making received from b ∼ 0, where the telescope collects the thermal it difficult to explain the smooth profile of the zodiacal dust. emission of particles over a wide distance range. A broader Note also that Altobelli et al. (2007) have not detected HTC distributionoforbitalinclinationsisapparentlyneededtomatch particleimpactsintheCassinidustexperiment,indicatingthat IRASmeasurements. HTCparticlesarerelativelysparse. No.2,2010 COMETARYORIGINOFTHEZODIACALCLOUD 823 orbitallydecouplefromtheplanetanddriftdownto1AU.This results in a situation where the inclination distribution of JFC particlesissignificantlybroader(≈20◦ meaniforR < 3AU) thanthatoftheirsourceJFCs.ThisexplainsFigure5andshows limitations of the arguments about the zodiacal cloud origin basedonthecomparativeanalysisofsources(e.g.,Hahnetal. 2002). WewillnowaddressthequestionofhowtheMIRfluxesfrom theJFCparticlesdependonDandt .Wedefine JFC (cid:8) 1 π/2 [F (b)−F (b)]2 η2 = IRAS model db, (8) π −π/2 σ2(b) whereF isthemeanIRASflux,σ isthestandarddeviation IRAS of F determined from the spread of representative IRAS IRAS scansforeachb(Section2),andF = F (i.e.,α = 1 model JFC JFC andα = α = α = 0inEquation(7)).Notethatthe AST HTC OCC integrationinEquation(8)issettoavoidtheintervalsinbwith stronggalacticemission. While the definition of η2 in Equation (8) is similar to the usualχ2 statistic(e.g.,Pressetal.1992),wewillnotassigna rigorous probabilistic meaning to the η2 values obtained from Equation (8). This is mainly because it is not clear whether the σ values computed in Section 2 from the IRAS data can adequately represent the measurement errors. Instead, we will Figure7.Dependenceoftheshapeof25μmprofilesproducedbyJFCparticles use Equation (8) only as an indication of whether a particular onDandtJFC.ThepanelsshowresultsfordifferenttJFC:(a)tJFC=12,000yr, modelismorereasonablethananotherone.Modelswithη2 (cid:2)1 d(ba)shtJeFdCli=ne3in0e,0a0ch0pyarn,e(lc)shtoJFwCs=the5m0,e0a0n025yrμ,manIdR(AdS)ptJrFoCfil=efo1r0l0(cid:6),0=0090y◦r..TThhee wgiivlelsbeη2gi=ven5.p1r.iority.Forareference,theJFCmodelinFigure5 uppersolidcurvesshowthemodelresultsforthesameelongation.Thebottom linesshowtheresidualfluxobtainedbysubtractingthemodelfluxesfromthe We calculated η2 as a function of D and tJFC to determine meanIRASprofile.ResultsforD=10,30,100,300,and1000μmareshownin which values of these parameters fit IRAS observations best. eachpanelwithbroaderprofilescorrespondingtolargerD.Someofthesemodel Wefoundthatthebestfitswithη2 < 10occurfor30 (cid:3) D (cid:3) pDro>file3s00doμmno,tamndatDch=IR1A0Sμombscearvnabtieonclsewarelyll.ruSlpeedcoifiuct.ally,tJFC >50,000yr, 100μmandtJFC (cid:3)30,000yr. Figure 7 illustrates how the shape of the 25 μm profile producedbyJFCparticlesdependsonDandt .Theprofiles JFC TheOCCparticles,whichhaveanearlyisotropicinclination become wider with increasing D and t values. For t = JFC JFC distribution, produce the MIR flux that is constant in latitude 12,000yr,thebestresultswereobtainedwithD = 30μmand (not shown in Figure 5). Therefore, the ecliptic and polar D = 100 μm (η2 = 3.2 and 5.1, respectively). The profiles fluxes from OCC particles are roughly the same and do not forD = 10μmaretoonarrowandclearlydonotfitdatawell match observations. We conclude that a single-source model (η2 =70),whilethoseforD =1000μmareslightlytoowide witheitherasteroidal,HTC,orOCCparticlescannotmatchthe (η2 =58).Wealsofoundthattherearenoreallygoodfitswith observedprofileofthezodiacalcloud. t > 30,000 yr, because the profiles are too broad near the JFC WeareleftwithJFCs.Itisnotablethatthewidthandshapeof eclipticindependentlyoftheparticlesize. theJFCprofileinFigure5closelymatchobservations.Theonly Thebestsingle-sourcefitsdiscussedabovehaveη2 >1which slightdifferenceisapparentforlargeeclipticlatitudeswherethe is not ideal. According to our additional tests this is probably modelflux,shownhereforD =100μmandtJFC =12,000yr,is notduetothecoarseresolutionandstudiedrangeofDandtJFC. slightlyweakerthantheonemeasuredbyIRAS.Wewilldiscuss Instead, this may point to (1) a subtle problem with our JFC thissmalldifferencebelowandshowthatitcouldbeexplained model,and/or(2)thepossibilitythatadditionalminorsources if: (1) slightly smaller JFC particles were used, and/or (2) the (such as asteroids, OCCs, or HTCs) should be included in the zodiacalcloudhasafaintisotropiccomponent.Wethusbelieve model.Option(1)isdifficulttotestunlessabettermodelofthe thatthecloseresemblanceofourmodelJFCprofilewithIRAS JFCpopulationbecomesavailable.Here,weconcentrateon(2) dataisastrongindicationthatJFCsarethedominantsourceof becauseanampleevidenceexists(e.g.,for∼10%near-ecliptic particlesinthezodiacalcloud. asteroid contribution from asteroid dust band modeling) that SinceasteroidsandactiveJFCshavesimilarinclinationdis- theseadditionalsourcesmayberelativelyimportant. tributions (Hahn et al. 2002), it may seem surprising that JFC We start by discussing the results obtained by assuming particlesproducesubstantiallywiderMIRfluxprofilesthanas- that the zodiacal cloud has two sources. The motivation for teroidal particles. By analyzing the results of our numerical considering the two-source model was the following. First, integrations,wefindthattheencounterswithterrestrialplanets we wanted see whether a combination of two sources could and secular resonances are apparently not powerful enough to successfully fit the observed profile. Second, we attempted to significantlyaffecttheinclinationdistributionofdriftingaster- place upper limits on the contributions of asteroid, OCC, and oidal particles. The inclination distributions of the asteroidal HTC sources. While it is obvious that models with more than particles and their source main belt asteroids are therefore es- two sources can be tuned to fit the data better, it is not clear sentially the same (≈10◦ mean i). On the other hand, we find whethermorethantwosourcesareactuallyrequired.Ourtwo- thatJFCparticlesarescatteredbyJupiterbeforetheyareableto sourcemodelswereusedtotesttheseissues. 824 NESVORNY´ ETAL. Vol.713 Figure 8. Examples of fits where we modeled the zodiacal cloud as having twosources.Fluxesat25μmareshown.(a)Thebest-fitmodelwithasteroid andOCCsources.ThismodeldoesnotfitIRASobservationswell.Themodel profileistoonarrowneartheeclipticandtoowideoverall.(b)Ourbesttwo- sourcemodel.Here,weusedαJFC =0.97,αOCC =0.03,D =100μm,and tJFC=12,000yr.Thefaintisotropiccomponentimprovesthefitqualitysothat Figure9.ModelconstraintsonthecontributionofasteroidandOCCparticles η2=0.36in(b).Thismaysuggestthatthezodiacalcloudcontainsasmallbut tothezodiacalcloud.Here,weusedathree-sourcemodelwithD =100μm significantfractionofOCCparticles. and αHTC = 0. For a range of the αOCC and αAST values, we set αJFC = 1−αOCC−αAST,andcalculatedη2(Equation(8))foreachmodel.Thecontours showη2=3,10,and30.Theshadedareadenotestheparametersofourbest-fit In the first test, we used the two-source model with aster- modelswithη2 <1.ThesemodelshaveαOCC <0.13andαAST <0.22thus oids and OCCs (i.e., α = α = 0 in Equation (7)). We placinganupperlimitonthenear-eclipticcontributionofasteroidandOCC JFC HTC particles. found that this particular model produces unsatisfactory fits (η2 > 10) to IRAS observations for all particle sizes consid- ered here (10–1000 μm) (Figure 8(a)). The model profile is smallasteroid/OCCcomponents(αAST (cid:2)0.2andαOCC (cid:2)0.1). significantlynarrowerneartheecliptic,wheretheasteroidcom- To verify this conclusion, we considered three-component ponentprevails,andistoowidefor|b| (cid:4) 40◦,wheretheOCC models with αHTC = 0 and used αJFC, αAST, and αOCC as component prevails. This happens mainly due to the fact that free parameters. We found that the best two-source fit with theasteroiddustisconfinedtotheeclipticplaneandproduces αJFC =0.97andαOCC =0.03cannotbesignificantlyimproved averynarrowprofileneartheecliptic(Figure5).Wealsofind byincludingasmallasteroidcontribution.Similarly,thefitwith it unlikely that two so distinct populations of objects, such as αJFC = 0.9 and αAST = 0.1 cannot be improved by adding a themainbeltasteroidsandOCCs,wouldhavecomparabledust smallOCCcontribution.Thus,theasteroid/OCCcontributions production rates. Thus, we believe that the two-source model cannot be constrained independently because their effects on withasteroidandOCCdustcanbedismissed. the combined profiles can be compensated by adjusting the D Inthesecondtest,wesetαOCC = αHTC = 0andconsidered andtJFCvaluesofthedominantJFCparticles. models of the zodiacal cloud with the JFC and asteroid com- If we set the parameters of the dominant JFC particles to ponents. We found that a small contribution of asteroid dust be D = 100 μm and tJFC = 12,000 yr, however, the αAST can improve the fits. For example, η2 = 0.92 for the D = and αOCC values can be constrained much better (Figure 9). 30 μm JFC particles with αJFC = 0.9 and tJFC = 12,000 yr, For example, models with η2 < 1 require that αAST < 0.22 and D = 100μm asteroidal particles with αAST = 0.1. This and αOCC < 0.13. Thus, under reasonable assumptions, the represents a significant improvement from η2 = 3.2 that we contributionofasteroidparticlestothenear-eclipticIRASfluxes obtainedforasingle-sourcemodelwithJFCparticlesonly.Val- isprobably<10%–20%,inagreementwiththeresultsobtained ues α (cid:4) 0.3 are clearly unsatisfactory because η2 > 10 in NVBS06 from modeling of the asteroid dust bands. This AST for α > 0.3. Also, η2 > 3.1 for α > 0.2 which shows means that asteroid dust contributes only by <10% to the AST AST that the fit does not improve when we add a (cid:4)20% asteroid overallzodiacaldustemissionatMIRwavelengths.Thethermal contribution. These results suggest that a very large asteroid emission of OCC particles can constitute as much as ∼10% contributiontothezodiacalcloudcanberuledout.Thisagrees of the near-ecliptic emission with ≈5% providing the best fits withtheconclusionsofNVBS06whofoundthatα (cid:2) 0.15 (Figure 9). When integrated over latitude, the overall OCC frommodelingofthemainasteroiddustbands. AST componentinthezodiacalcloudislikelytobe(cid:2)10%.6 Finally,thetwo-componentmodelwiththeJFCandisotropic For the sake of consistency with the results suggested from OCC sources can fit the data very well (Figure 8(b)). With modelingoftheasteroiddustbands(seeSection3.1;NVBS06), D =100μmandt =12,000yr,correspondingtooneofour weimposeasmallasteroidcontributionintheJFC/OCCmodel. JFC bestsingle-sourcefitswithJFCs,α =0.97andα =0.03, Figure10showsourpreferredfitatdifferentIRASwavelengths. JFC OCC wefindthatη2 =0.36,byfarthebestfitobtainedwithanytwo- The η2 values of this fit in different wavelengths are the sourcemodel.AsFigure8(b)shows,thefitisexcellent.Wemay following:0.29at12μm,0.35at25μm,and0.06at60μm.This thusfindanevidenceforasmallcontributionofOCCparticles is very satisfactory. Since our model does notinclude detailed to the zodiacal cloud. A much larger OCC contribution is not emissivity properties of dust grains at different wavelengths supported by the data because the fit gets significantly worse forαOCC > 0.1.Forexample,η2 > 10forαOCC > 0.15which 6 Notethatweareunabletodistinguishinthemodelbetweentheuncertainty isclearlyunsatisfactory.AlargecontributionofOCCparticles ininstrumentalisotropicbrightnessofIRASmeasurementsandtheOCC canthereforeberejected. contribution.Wealsodonotknowhowmuchofthebrightnessofthefaintest partoftheskyiscosmological(orgalactic).Forthesereasons,theOCC We propose based on the results described above that the contributiontothezodiacalcloudisdifficulttocalibratefromIRAS zodiacal cloud has a large JFC component (αJFC (cid:4) 0.9), and observationsalone.

Description:
ABSTRACT. The zodiacal cloud is a thick circumsolar disk of small debris particles produced by asteroid collisions and comets. Their relative
See more

The list of books you might like

Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.