163 Pages
English

Study of long-lived radioactive sources in the Galaxy with INTEGRAL-SPI [Elektronische Ressource] / Wei Wang

-

Gain access to the library to view online
Learn more

Description

Technischen Universität MünchenMax-Planck-Institut für extraterrestrische PhysikGarching bei MünchenStudy of Long-lived RadioactiveSources in the Galaxy withINTEGRAL/SPIWei WangVollst"andiger Abdruck der von der Fakultät für Physik der Technischen UniversitätMünchen zur Erlangung des akademischen Grades einesDoktors der NaturwissenschaftengenehmigtenDissertation.Vorsitzender: Univ.-Prof. Dr. Andrzej J. BurasPrüfer: 1. Hon.-Prof. Dr. Günther Hasinger2. Univ.-Prof. Dr. Stephan PaulDie Dissertation wurde am 30.08.2007 bei der Technischen Universität Müncheneingereicht und durch die Fakultät für Physik am 27.09.2007 angenommen.12Physik-DepartmentDoktorarbeitStudyofLong-livedRadioactiveSourcesintheGalaxywithINTEGRAL/SPIWei Wang27 September, 2007Technische Universit? t M chenMax-Planck-Institut f r extraterrestrische PhysikAbstract26 60Two long-lived radioactive isotopes, Al and Fe in the Galaxy are studied with the26 60high spectral resolution INTEGRAL spectrometer (SPI). Al and Fe have the similarhalf-life ( million years) and astrophysical origins in the Galaxy. Their nucleosyn-thesis and ejection into the interstellar medium (ISM) are dominated by massive starsand the subsequent core-collapse supernova explosions. Detections of these isotopesprovide evidence that nucleosynthesis is ongoing in the Galaxy. And their line shapesre ect the dynamics of the ejected isotopes in the interstellar medium and then probe26 60properties of ISM.

Subjects

Informations

Published by
Published 01 January 2007
Reads 29
Language English
Document size 3 MB

TechnischenUniversitätMünchen
Max-Planck-InstitutfürextraterrestrischePhysik
GarchingbeiMünchen

StuSdoyurocfeLsoinngt-hlieveGadlRaxaydiowiatchtive
INTEGRAL/SPI

WeiWang

Vollst"andigerAbdruckdervonderFakultätfürPhysikderTechnischenUniversität
MünchenzurErlangungdesakademischenGradeseines
DoktorsderNaturwissenschaften
genehmigtenDissertation.

Vorsitzender:Univ.-Prof.Dr.AndrzejJ.Buras
Prüfer:1.Hon.-Prof.Dr.GüntherHasinger
2.Univ.-Prof.Dr.StephanPaul

DieDissertationwurdeam30.08.2007beiderTechnischenUniversitätMünchen
eingereichtunddurchdieFakultätfürPhysikam27.09.2007angenommen.

1

2

-DPartmentephysik

Doktorarbeit

StudyofLong-livedRadioactiveSourcesin
INTEGRAL/SPIwithGalaxythe

angWeiW

2007,eptemberS27

TechnischeUniversitätMüchen

Max-Planck-InstitutfürextraterrestrischePhysik

Abstract

Twolong-livedradioactiveisotopes,26Aland60Feinthe26Galaxy60arestudiedwiththe
highspectralresolutionINTEGRALspectrometer(SPI).AlandFehavethesimilar
half-life(∼millionyears)andastrophysicaloriginsintheGalaxy.Theirnucleosyn-
andthesistheandsubsequentejectionintothecore-collapseinterstellarsupernovmediuma(ISM)explosions.areDetectionsdominatedofbythesemassiveisotopesstars
proreflectvidetheevidencedynamicsthatofthenucleosynthesisejectedisotopesisinongoingtheintheinterstellarGalaxy.mediumAndtheirandlinethenprshapesobe
propertiesofISM.Observationsofdiffuse26Aland60Feemissionandtheirlineshapes
intifictheobjectivinneresofGalaxythisandthesis.activenearbystar-formationregionswithSPIaremainscien-
26gionWith(|l|the<30four◦,|-yb|ear<SPI10◦)data,usingwemodelobtainthefittingsAlwiththespectrumdifforferenttheskyinnertracerGalaxymodelsre-
including26AlCOMPTELmaps,HI,CO,EGRETskysurveymaps,freeelectronden-
sitydistributionmodels,theRobinyoungdiskmodel,anexponentialdiskmodel,
ahomogenousdiskmodel(constantalonglongitudes,exponential-likeinlatitudes,
scaleheight200pc).Thefittedparametersofderived26Alspectrafromdifferentdis-
withintributionerrormodels:bars.Diflineferentcentroid,skyfluxdistributionandwidth,modelsarewdoellnotafconsistentfect26Alwithintensityeachotherand
26(line3.0±0.2shapes)×for10−4thephinnercm−2sGalaxy−1.rad−The1whichmeasuredisAlconsistentintensitywithforthetheCOMPTELinnerGalaxyresults.is
The26Allinecentroidisdeterminedat1809.0±0.1keV,whichishigherthanthelabo-
shoratorwyavnarralueowforthefeature,26Alwithlineanofintrinsic1808.65±line0.07widthkeV.ofThe0.5±26Al0.4kelineV.forThisthe26AlinnerlineGalaxywidth
formodesttheinnerinterstellarGalaxyis-mediumconsistentturbulencewith(<expectation200kmofs−1)Galacticaroundrotationthe(sour<0.9ceskeofV)26Aland.
OurresultsareconsistentwiththepreviousreportsbyHEAO-CandRHESSI,buta
verBasedybronoad3DlinedistributionfeaturereportedmodelsbyofGRISGalacticis26clearlyAl,weruledcanoutbobtainyouraGalacticpresent26SPIAlresult.mass
Rof0(=2.78.5±0.6kpc,)Mwhich,wherecanbeweconvhaveertedtakentoathedistancecore-collapseofthesuperSunnovtoatherateofGalactic1.90±center0.95,
26correspondingtoastar-formationrate(SFR)of(◦3.8±1.9)Myr−1intheGalaxy26.
AlspectraalongtheGalacticplane(|l|<60)canbeobtainedwithSPI.Alline
centr26oidenergyshiftsalonglongitudesaredetected◦asexpected◦byGalacticrotation.
The1808.66AlkeVline,wellcentroidconsistentforthewithGalacticthelaboratorcenter(y|l|ener<5gy.,|b|Relativ<10eto)isthedeterlinecentrminedoidatof∼

3

theGalacticcenter,26Allinesshowaredshiftof∼0.1−0.2keVforpositivelongitudes
(5◦<l<40◦),andablueshiftof∼0.5−0.9keVfornegativelongitudes(−40◦<l<
−5◦).However,Galacticrotationpredictssymmetriclineenergyshiftsof±(0.2−0.4)
keVforpositiveandnegativelongitudesfromoursimulationsassumingthatGalactic
26Alfollowsthefreeelectrondensitydistributionmodelsbasedonpulsardispersion
measurements.Observedasymmetryof26AllineenergyshiftswithSPIcannotbe
explainedonlybyGalacticrotation.Othereffectsnotconsideredinsimulationsmay
contributetothisasymmetry:theBarstructureintheGalacticcenter;bulkmotionsof
nearby26Alsources.
The26AllinewidthandintensityalongtheGalacticplanearealsodetermined.26Al
linesfordifferentlongitudesshowthenarrowlinefeaturesexceptforthedirectionof
20◦<l<40◦.Possiblebroadlinefeatureinthisregionmayreflectthekinematicsof
nearby26AlsourcesintheSagittariusarm.Brightnessof26Alemissionalongtheplane
alsoshowasymmetries:the4thquadrantisbrighterthanthe1stquadrant,withthe
26Alfluxratioof∼1.3for|l|<60◦,and∼1.2for|l|<30◦.Insimulationsof26Al
emissionintheinnerGalaxy,threedifferentfreeelectrondensitydistributionmodels
areadopted:allcomponents;allexceptthethickdiskandtheGalacticcenter;only
thethindiskandspiralarms.Fortwoformermodels,noasymmetriesof26Alintensity
areexpected.Whileforthethirdmodel,26Alintensityforthe4thquadrantobviously
appearhigherthanthatforthe1stquadrant,andthepredictedratioisconsistentwith
theSPIobservations.Therefore,wedrawaconclusionthatthefree-electrondensity
modelofonlythethindiskandspiralarmswouldreflecttherealdistributionof
26.AlGalacticWealsoderived26AlspectraalongGalacticlatitudeswiththeSPIdata.Possible
26Alemissionforhighlatitudes(|b|>5◦)shouldoriginatefromlocalstar-formation
systemsintheGouldBelt,while26Alinlowlatitudes(|b|<5◦)wouldbedominated
bythelarge-scaleoriginintheGalacticdisk.26Alemissionforlowlatitudes(|b|<5◦)
issignificantlydetectedasexpected.The26Allineforthediskshowsanarrowfeature.
Theasymmetryof26Albrightnessforthe1stand4thquadrantsisalsodetectedfor
thediskcomponent.26Allinecentroidenergyshiftsalongthediskareevident,but
theasymmetryofenergyshiftsisstillinconsistentwiththeexplanationofGalactic
rotation.Alargeblueshiftforthenegativelongitudes(−40◦<l<−10◦)maybealso
attributedtothecontaminationofnearby26Alsources.Weak26Alemissionisdetected
inthepositivelatituderegionof5◦<b<20◦withablueshiftof∼0.7keV,butno26Al
signalisdetectedinthenegativelatituderegion−20◦<b<−5◦.Furtheranalysis
showsthat26Alemissionfor5◦<b<20◦iscontributedonlybythenegativelongitude
part,wherethenearestactivestarformationregion,theSco-CenOBassociationinthe
located.isbeltGouldDetectionsof26Alfromnearbystar-formationregionsarethegoodprobeforthe
massivestaroriginofGalactic26Alandkinematicsof26AlejectainISM.26Alspectrafor
threeregions(Cygnus,VelaandSco-Cen)areobtainedwiththepresentSPIobserva-
tions.Thereported26AlfluxfortheCygnusregionis∼(7.0±1.2)×10−5phcm−2s−1

4

26withwithoutalinesignificancecentroidlevelshift.of26∼Al6σ.emissionTheAlforlineVelaforisCynotgnussignificantlyappearsadetectednarrowwithfeature,SPI,
withameasuredfluxof(4.0±1.2)×10−5phcm−2s−1.The26AlspectrumforVela
appearsabroadlinefeature(intrinsicwidthofFWHM∼2keV),implyingthatsuper-
nova26explosionsandwindsfromWolf-RayetstarsinVelaha26vestrongeffectsonISM
nearlatitudes.AlWsoureces.foundTheaS26Alco-Cenlinefluxregionofw(6.3ould±2.0)contribute×10−5tophcmAl−2s−emission1.Thetowardspectrumhigh
enerappearsgyofathenarr26oAlwline.featureThisandhassignificantalargeblueshiftblueshiftimpliesof∼that1.526keAlVsourrelativceseintoStheco-Cenrest
aredominatedbythecomponentof26Alejectawithabulkvelocityof∼200kms−1
us.darwtoFinally,wesearchedforthegamma-raysignalfromtheotherlong-livedradioactive
isotope60FeintheGalaxywiththe2.5yearSPIdata.Thedetectionof60Fedecay
gamma-raylinesatboth1173and1333keVfromtheGalaxyarereportedwith60our
linefluxmeasurements,fromthewithinneraGalaxysignificanceregionof4.9isσ(after4.4±0.9)combining×10−5twphocmlines.−2sThe−1aradv−erage1.FromFe
60theFe/26sameAlofobser(14.8v±ations6.0)and%.Theanalysispresentproceduretheoreticalappliedpredictionsto26A,arewestillfindaconsistentfluxratiowithof
ourresultontheratioof60Fe/26Al.The60Fesignalsaretooweaktodetermineline
shapedetails;itappearsthataGaussianwiththewidthoftheinstrumentalresolu-
tioncanfitthedatawell,implyingthatbroadeningof60Felinesfromastrophysical
pralsoocessessearchisfornot60Fesignificant.emissionInfroromderthetoCyinvgnusestigateandVtheelavregions,ariationsoandverdothenotGalaxyfind,60wFee
signals.UsingmoreSPIdatainfuture,wecouldimprovespectralresolutionof26Alinthe
innerGalaxyand26Alspectraalongtheplane,andthenconfirm26Allineshapesfor
Cygnus,VelaandSco-Cen,furtherobtain26Al60spectrafromsomeothernearbystar-
formationregions,likeOrionandCarina.ForFestudies,thesignificancelevelof
of60detectionsFesignalsforthefortheinner1stGalaxyand4thwillbequadrantspushedcanforbewarprdsobedagain,withandmorepossibleSPIdifdata.ferences

5

6

Contents

Introduction11.1Cosmicnucleosynthesisandnuclearastrophysics.............
1.2Gamma-raylineastronomy..........................
441.3Ti........................................
1.4511keVemissionfromelectron-positronannihilation...........
1.4.1PositronsfrommillisecondpulsarsintheGalacticcenter.....
261.5All-skyobservationsofAlbyCGRO/COMPTEL.............
261.6TheoriginofAlintheGalaxy........................
1.6.1Corecollapsesupernovae.......................
1.6.2Wolf-Rayetstars.............................
1.6.3Novae..................................
1.6.4AsymptoticGiantBranch(AGB)stars................
1.6.5Interactionsofcosmicrayswiththeinterstellarmedium.....
601.7TheoriginofFe................................
1.7.1Massivestars..............................
1.7.2Supernovae...............................
1.8TheINTEGRALspectrometer(SPI)......................

2DataanalysesofINTEGRAL/SPIatMPE
2.1INTEGRAL/SPIobservations.........................
2.2Dataselectionandassembly..........................
2.3Backgroundmodelling.............................
2.4SpectralresponseofSPI............................
2.5Extractingspectrafromthesky:diffuseemissionandpointsources..
263AlemissionandlineshapesintheGalaxy
3.1Datapreparationandanalysis.........................
263.2AlemissionandlineshapesintheinnerGalaxy.............
263.3AstrophysicaloriginsofAllinewidth...................
263.4ComparisonofthedifferenttracermapsofAlsourcesintheGalaxy.
263.5Almass.....................................
263.6AllineshapesforthedifferentlongitudesalongtheGalacticPlane..
3.6.1Differencesinthe1stand4thquadrants..............
3.6.2RevisitingGalacticrotationeffect..................

910111315192326262831323233333436

414347475253575962676971767982

7

Contents4

5

6

8

263.6.3Alemissioninsmalllongitudeintervals..............
263.7LatitudestudiesofAlemission.......................
3.81809keVemissionintheinnerGalaxy:simulationsversusobservations
263.8.1ModelsofGalacticAl........................
263.8.2Alintensityandlineshapes:simulationsversusobservations.

26Alsourcesandlineshapesinstarformationregions
4.1Cygnusregion..................................
4.2Velaregion....................................
4.3Sco-Cenregion.................................

60DiffuseFeemissionintheGalaxybySPI
5.1Datapreparationandanalysis.........................
605.2ObservationsoftwoFelinesat1173and1333keVintheGalaxy...
605.2.1DetectionsofFefromtheinnerGalaxy..............
605.2.2SearchingforFesignalfromCygnusandVela..........
26605.3TheratioofFe/Al.............................

perspectivesandSummary

91969999103

105106110113

117118124124128128

133

Introduction1

RadioactivitywasfirstdiscoveredbytheFrenchphysicistHenriBecquerelmorethan
onecenturyago,whenheobservedtheimageofthephotographicplatefoggedby
andexposurethetometalselfradiationcanfremitom"light"uraniumsalts,(Becquerelandrathenys),foundwiththethatallintensityuraniumproportionalcompoundsto
Tthewoyearamountlaterof,Pierreuranium,andMarieindependentCurieoncoineditsthechemicalterm"radioactivcompositione"[forthoseBecquerel,elements1896].
whichemittedBecquerelrays.In1903,thesethreephysicistssharedtheNobelPrizein
researPhysics,ches"inontherecognitionradiationofthephenomenaextraordinardiscoyvsereredvicesbythePryhaofessorveHenrirenderedbyBecquerel."theirjoint

In1899,Rutherforddiscoveredthatthreedifferentkindsofradiationareemittedin
theseradioactivesubstances,andfirstlycalledthese"alpha(α)","beta(β)"and"gamma
(γ)"rays[Rutherford,1899].Inthelateryears,αrayswereidentifiedwiththenucleiof
helium[Rutherford,1905],andβrayswereelectrons[Becquerel,1900].Ittookalong
timeuntilpeoplerealizedthatγrayshaveallthepropertiesofenergeticelectromag-
neticradiation,nowcalledhighenergyphotons[Allen,1911,Compton,1929].

statesAtofpresent,atomicwenucleiunderstandthroughthatelectrowradioactiveakedecainteractions.ysaretransitionsMeasurementbetwofeendifradioactivferente
decalogicalysisclockanisimportantwidelytoolusedofinthelifeexperimentalanddifphysics.ferentOnscientificEarth,fields,e.g.radioactivitytheevasaolutiongeo-
oftheEarthandthecrust,biology,archeology,meteoritics.Inthedistantuniverse,ra-
dioactivedecayscanbeenstudiedbymeasuringγ-raylinespectraofcelestialsources.
Theγ-raylinescanidentifytheindividualisotope,andtheabundanceoftheseiso-
topesparticularcan,γbe-rayquantifiedlinephotonswiththearenearlymeasurementtransparentofγ-raforythelineunivintensityerse,withofthenoskyabsor.Inp-
tionbytheverydensemolecularcloudsandinterstellarmedium.Sodetectionsof
theseγ-raylinesareapowerfultooltostudythecosmicabundanceofradioactive
isotopes,whichistheoriginalscientificmotivationofthisthesis.Cosmicradioactive
isotopesareproducedthroughstellarnucleosynthesisandsupernovae(SNe),ejected
reintoviethewsbyinterstellarPrantzosandmediumDiehlby1996winds,,DiehlandandexplosionTimmespr1998ocesses).(alsoseedetailsinthe

9

1oductionIntr1.1Cosmicnucleosynthesisandnuclearastrophysics

TheUniverseisahugefactoryoftheelements.Just3minutesafterthebigbang,
primordialnucleosynthesisoccurred(originalideabyGamow1946).Thebigbangnu-
cleosynthesis(BBN)producesthepresentabundanceoflightelements:D(Deuterium),
3He,4He,7Li.Buttheheavierelements(A>12)cannotbeproducedduringBBN.Some
lightelements,e.g.B,Beand6Li,areproducedbycosmicrayspallationprocesses.But
theheavierelementsaremainlyproducedbynucleosynthesisinstars(seedetailsin
areviewbyWallersteinetal.1997,andagoodtextbookbyArnett1996,alsoreferto
theearlyexcellentpaperbyBurbidge,Burbidge,Fowler,andHoyle1957).
Starscanproducetheheavyelementsthroughfusionprocessesduetotheveryhigh
temperatureintheirinteriors.Hydrogenburningisfirsttriggered,andheliumispro-
ducedthroughthep−p12chainandCNOcycle.Thenheliumburningfollowsin16higher
temperature,producingC,andαcapturescanextendnucleosynthesistoO,and
20Ne,24Mg,etc.,uptotheverystable,doublemagicnucleus40Ca.Becauseexper-
imentsshowthatthe16O(α,γ)20Nerateisveryslowinstellarinteriors,atpresent
peoplegenerallybelievethattheformationofnucleiheavierthan16Oproceedssuc-
cessfullybycorecarbon,neon,oxygenandsiliconburning.Aftersiliconburning,
freedalphas,protons,andneutronsbuildnucleiuptotheregionofthemoststable
nuclei,theiron(e.g.54Fe,56Fe)abundancepeak.Ifthebuildupoccursrapidly,with
increaseofneutronexcess,thedominantnucleuswillbe56Ni.Theirongroupnuclei
aremoststableinthenature,sothefurtherfusion(burnings)stops.Massivestars(e.g.
M>8M)willendtheirlifethroughsupernovaexplosions.Supernovanucleosynthe-
sisistheonlyoriginoftheheavierelementsthantheirongroup.Duetotheneutron
richenvironmentinsupernovaexplosions,thenucleosynthesisproceedsbyneutron
captures(rapidcapture,called"r-process";slowone"s-process").Ther-processesand
s-processescouldproducenearlyalltheheavyelementsabovetheiron,e.g.,60Feand
famousradioactiveisotopes235,238U.Proton-richisotopesarerareinthenature(e.g.,
26Al),theycanbeproducedthroughprotoncapturesintheproton-richenvironment.
Starsproducetheelements,andcanejectedthemtotheinterstellarmedium(ISM)
bystellarwindsandsupernovaexplosions.Theinterstellarmediumwouldcooland
collapseintodenseclouds,andthenformthenewstars.Thenacosmicelementcycle
finishes,andanewonebegins.Cosmicnucleosynthesisbecomesthefoundationof
thelifeandthepresenthumanworld.
Cosmicelementabundancecouldbemeasuredpreciselybyastronomicalobserva-
tionsoverawiderangeoftheelectromagneticspectrum,speciallyintheopticalband
nowadays.Forexample,theHubbleSpaceTelescope,andlarge-apertureearth-based
telescopeshaveyieldsveryhighresolutionspectraofstars,galaxiesandinterstellar
medium.Thehighresolutionspectralanalysescandeterminecosmicabundanceof
mostelementsandtheirisotopes.Anyway,optical,infrared,andultravioletobserva-
toriesprovideonlyapartoftheinformation.Observationsofgamma-raylinesfrom
radioactiveisotopesopenanewwindowofobservationsandguideourunderstand-

10

1.2Gamma-raylineonomyastringofnucleosynthesisinstars.Theobservationalconstraintschallengetheorists,and
pushthemtododetailedcalculationsofstellarevolutionandnucleosynthesis.At
present,newastronomicalobservations,newnuclearexperimentalcapabilities,and
newcomputationalpowerpresentanopportunityforanimportantadvanceinour
nucleosynthesis.cosmicofwledgeknoTherearegoodnewsformodernnuclearastrophysics.(1)Awealthofnewobserva-
tionsrequiredetailednucleardataforacredibleexplanation.(2)Newacceleratorsand
detectiontechniquesprovideanunprecedentedandgrowingcapabilityforproviding
theneedednucleardata.(3)Exponentiallygrowingcomputationalpowermakesit
possibletoincludetheresultingmicrophysicsinsimulationsofstellarevolutionand
thesequenceofeventsleadingtonovae,X-raybursts,andsupernovaexplosions.The
enhancedprogramsfornuclearastrophysics,experimentalandtheoretical,willgreatly
advanceourunderstandingofthecosmos.

astronomylineGamma-ray1.2

Detectionsofgamma-raylinesprovideaspecialwindowtodeterminethecosmic
abundanceofradioactiveisotopes.Radioactiveisotopesareco-productswithstable
nucleiofnucleosynthesis.Theyaregenerallyproducedinstellarinteriors,supernovae,
novae,andinterstellarmedium.Theseradioactiveisotopesdecaywithemissionof
characteristicgamma-raylines.Measurementsofthesegamma-rayscanrelatetoex-
istenceoftheirparentisotopes.Gamma-raylineemissioncanpenetratedensemass
layers,sothedetectionoftheselinesaremoredirectthanthemeasurementsofphoto-
sphericorinterstellarabsorptionoremissionlinesintheopticalband.Observations
ofgamma-raylinesprovideapossibletooltoprobenucleosynthesisinearlyevolution
stagesofsupernovaandnovae,whengamma-rayemissionbecomestransparentafew
explosions.afterysdaGamma-raylinesarealsodetectedfromnuclearexcitationinastrophysics.Atomic
nucleiinexcitedstatesareproducedinenergeticcollisionsininterstellarspaceandin
thevicinityofcompactobjects.Typicalexcitationenergiesofatomicnucleiareinthe
MeVregime,prominentexamplesbeingtheCfirstexcitedlevelat4.4MeV,orOat6.1
MeV.IntheouterregionsoftheSun,suchcollisionsoccurprominentlyduringsolar
flares.Ininterstellarspace,cosmicrayinteractionswithambientinterstellargasresults
inde-excitationgamma-raylines.Nearneutronstars,neutroncapturelineat2.2MeV
ortheredshiftedonecouldbedetected(e.g.seeMcConnelletal.1997).Thislinecan
beverystrong[Kuzhevskij,Kuznetsov,andTroitskaia,1998]andwasalsodetectedin
solarflares[Terekhovetal.,1993].
Radioactiveisotopeshavethedifferentdecaylifetimes.Ifthedecaylifetimeistoo
short,theemissionsiteisnottransparentforgamma-raylines,wecannotobservethe
emission.Soonlyasmallnumberofradioactiveisotopeswitharelativelonglifetimes,
arerelatedtothegamma-raylinemeasurementsofcosmicnucleosynthesis.Wehave

11

1oductionIntrTable1.1:Radioactiveisotopesforgamma-raylineastronomy

7Beisotopes77meandlifetimes7decaBey→7chainLi∗γ478-rayenergy(keV)nosourvaeces
22Na3.8yr22Na→22Ne∗+e+1275novae
26Al1.04×106yr26Al→26Mg∗+e+1809SNe,novae,
starsemassiv44Ti89yr44Ti→44Sc∗68,78,SNe
→44Ca∗+e+1157
56Ni111d56Ni→56Co∗158,812,SNe
→56Fe∗847,1238
57Co390d57Co→57Fe∗122SNe
60Fe2×106yr60Fe→60Co∗59,1173,1333SNe
e+104−107yre++e−→Ps511,<511SNe,pulsars,
materdarkγγ→

listedtheseinterestedisotopesinTable1.1.Forlong-livedradioactiveisotopes,the
emissionfromseveralormanynucleosynthesiseventsissuperimposed,e.g.26Aland
60Fe.Gamma-raylinesemissionsfrom26Aland60Fecanprobetheyoungsource
populations,andmaycorrelatetothestructureoftheGalaxy,whicharethetargets
thesis.thisinstudiedGamma-raylineastronomyhasbeengreatlydevelopedinthelasthalfcentury.The
firstgamma-raylineobservationswerefromOSO-3,OSO-7in1960’s,whichrevealed
thestrong2.223MeVlinefromsolarflares[Brandt,1969].Thislineresultsfromthe
formationofdeuteriumviatheunionofaneutronandproton;inasolarflarethe
neutronsappearassecondariesfrominteractionsofhigh-energyionsacceleratedin
theflareprocess.Andthefieldofgamma-raylineastronomytookgreatleapsfor-
wardwiththeHEAOC,theCOS-B(1975-1982)andtheSolarMaximumMission
(launchedin1980)satellites.TheHEAOCfirstdetectedthe1809keVlinefromra-
dioactive26AlintheinnerGalaxy[Mahoneyetal.,1982].In1977,NASAannounced
planstobuilda"greatobservatory"forgamma-rayastronomy.TheComptonGamma-
RayObservatory(CGRO)wasdesignedtotakeadvantageofthemajoradvancesin
detectortechnologyduringthe1980s,andwaslaunchedin1991.TheCOMPTELtele-
scopeaboardCGROfirstmappedthe1809keVγ-raylineemissionofthewholesky
[Diehletal.,1995a,Plüschkeetal.,2001],andfirstdetectedthestrongandbroad44Ti
linearound1157keVfromayoungsupernovaremnant(SNR)CasA[Iyudinetal.,
1994].TheOSSEtelescopeaboardCGROalsofirstlymappedthe511keVannihilation
lineemissionintheinnerGalaxy[Purcelletal.,1997].Currently,themainspace-based
gammarayobservatoryistheINTErnationalGamma-RayAstrophysicsLaboratory

12

ani - Pltrd - Plot of .RFR files : pltrd_/GRO Counts / bin 2001000

/u/c/ani/cas/sumres_cas_ph1+2+3+4+5_total.fit.ps ( ), 03-SEP-96 / 11:48:53, P 1 Cas A COMPTEL Phase 1-5 1.344iT-1008001000120014001600 Energy (keV) 1800
with radius 5.00 deg SRC Dataset ID: MPE-RFR-20663 chi sq.=36.33, 43 d.o.f. Sky circle 1 centered at Figure1.1:ThespectrumofLatitude ......: -2.10 deg Longitude .....: 111.70 deg 1157keVγ-raylinefromCasAdiscov Number of selected events: 26500 eredbyCOMPTEL44
ha(frveomthevIyudinelocitiesetal.of1997∼).EHA=0-180 deg; with radius 2.00 deg 10000Thekmbrsoad−1.linefeaturesuggeststhat theejectaofTi
Number of selected events: 21253 Sky circle 2 centered at Longitude .....: ???.70 deg resolution(INTEGRAL).enoughThetospectrresolveometerLatitude ......: >40.10 deg astraboarophysicaldlinesINTEGRALandallo(SPI)wprspectrovides oscoptheyinhighthespectralregime
ofgamma-rays.Theprogressandpresentresultsoftwoimportantlong-livedradioac-
tiveisotopes,26Aland60Fewillbediscussedinmoredetailsinthefollowing.Here,
wefirstbrieflypresent44therecentdetections−+ofgamma-raylinesfromtheothertwo
interestingisotopes:Tiandpositrons(eeannihilations).

44Ti1.344Tiisashort-livedradioactiveisotopeswithameanlifeof89years.Intheories,the
mostplausiblecosmicenvironmentforproductionof44Tiistheα-richfreeze-outfrom
high-temperatureburningnearthenuclearstatisticalequilibrium(seeWoosleyetal.
1973collapse;Arnettsuper1996nov).ae.TheSoitrequiredisbelievhighedvthataluesforcore-collapsetheentropysupermanoyvbeaefounddominateincore-the
44prturies.oduction44TiofdecayisradioactivbelieevedTi,tobewhichtheshouldonlybesourcevisibleofinthestablegamma-raisotopeysina44feCa.w44cen-Ti
gamma-raysreflectthecurrentrateofsupernovaeduetoitsshortdecaytimescale,so
thepresentdetectionsshouldbedirectlyrelatedtothecurrentpopulationofmassive
stars.Therearethreeγ-ray44lineswhichcanbeusedtodetectthedecayof44Ti:44the
68and78keVlinesfromtheScde-excitationcascadeandthe1157keVlineasCa
decaystoitsstablegroundstate.
Supernovaobservationshavedemonstratethattheseeventsaretheoriginof44Ti

13

/u/c/ani/cas/sumres_cas_ph1+2+3+4+5_total.fit Energy (keV)

1oductionIntr-410

-1 keV-510-2 cm-1 Counts s-610

-710

100Energy (keV)

Figure1.2:ThespectrumofCasAfromINTEGRAL/IBISandthebest-fitmodel(solid
redline,fromRenaudetal.2006).Thetwolinesatpositionsof67.8±1.6and
77.4±1.4keVareclearlydetected,andidentifiedasthesignatureof44Ti.

production.The1157keVγ-raylinefollowing44Tidecayhasbeendetectedinthe
340-yearoldGalacticSNRCasAwithCOMPTEL(seeFigure1.1,Iyudinetal.1994;
Schönfelderetal.2000).COMPTEL’ssurvey[Duprazetal.,1997,Iyudinetal.,1999]
haveresultedinanothercandidatesource,theso-calledVelajuniorSNR(GROJ0852-
4642,Iyudinetal.1998).YetCasAseemstobetheonlysourcewiththeestablished
44Tidetections.Thetwolow-energylinesat68and78keVfromCasAhavealsobeen
clearlydetectedbyBeppo-SAX[Vinketal.,2001],andconfirmedbyINTEGRAL/IBIS
(Figure1.2,Renaudetal.2006)andINTEGRAL/SPI(Martinetal.2007,inprepara-
tion).AndINTEGRAL’sinner-GalaxysurveywiththeIBISdidnotrevealanewsource
inthisregion[Renaudetal.,2004,2006].Furthermore,SN1987A’slatelightcurve,ob-
servedindetailsovermorethan15years,appearspoweredbyasimilaramountof
(0.2−2)×10−4M44Ti,frommodellingofradioactiveenergydepositionandphoton
transportintheSNR[Woosleyetal.,1989,FranssonandKozma,2002].
44Tiliveslongenoughsothatitshouldhavebeendetectedfromseveralrecent
Galacticsupernovaeiftheseoccuratarateof∼2percentury;butwiththeexcep-
tionofCasA(Iyudinetal.1994;Renaudetal.2004,2006),nootherSNRshave
beendetectedyet.AnalysisofCOMPTELdatasupporteda5σdetectionofCas
Aat(4.2±0.9)×10−5photoncm−2s−1inthe1157keVline(Iyudinetal.1997),
implying∼2.4×10−4Mof44Ti.Measurementsofthe68and78keVlinesby

14

1.4511Vkeemissionomfron-positrelectronannihilationBeppo-SAX,IBISandSPIaboardINTEGRALalsosuggestedtheejected44Timass
of∼(1.0−2.5)×10−4M.However,solarmetallicityTypeIIandTypeIbsuper-
novamodelsindicate∼(3−6)×10−5Myieldsof44Ti(e.g.Woosley&Weaver1995;
44TiThielemannobservetationsal.fr1996om;CasRauscherAasaetal.puzzle.2002;TheoristsLimongithen&Chiefsuggestedfi2003).thatThisleaasymmetriesvesthe
inthesupernovaexplosionmechanismcoulddriveanenhancementof44Tiintheejecta
son(e.g.betwNagatakieentheetal.obser1997v;ationMaedalimit&of44TNomotoiinCas2003A).andFigurethe1.3haspredictionsshownofthedifferentcompari-su-
pernovaexplosionmodelsontheejectedmassesof44Tiand56Ni.Thestandardsuper-
no2003va]modelspredictedwithalowersymmetricalamountof44Texplosionsi,cannot[Rauscherexplainettheal.,case2002of,CasLimongiA;andnon-sphericalChieffi,
modelsseemedtoresolvethe44TipuzzleinCasA[MaedaandNomoto,2003].SoCas
AmaybeaspecialcaseintheGalacticsupernovae,e.g.,possiblyanasymmetrical
TypeIbsupernovaexplosion.

1.4511keVemissionfromelectron-positronannihilation

PositronswithalowenergyaroundafewtensofeV,couldeitherannihilatedi-
rectlywithelectronsorformpositroniumbychargeexchangeorradiativecapture
[Bussard,Ramaty,andDrachman,1979].Theannihilationsproducethe511keVline
andpositroniumcontinuumbelowtheenergyof511keV.
PositronswilllosetheirenergymainlythroughviaCoulombinteractionswiththe
interstellarmedium.TheenergylosstimescaleofCoulombinteractionsinthemedium
τcdependsonthepositronenergy,themediumdensity,andthedegreeofionization.
Forthepositronswithenergyaround1-100MeV,theCoulombinteractioncooling
timescaleisestimatedasτc∼105n−1yr(withinafactorof2,Wangetal.2006),and
nisthemediumnumberdensityinunitsofcm−3.AndthemediumintheGalaxyis
quitecomplicated.Thentheenergylosstimescalescouldvaryfromτc∼103yrinthe
molecularclouds(n>102cm−2),τc∼105yrforthetypicalwarminterstellarmedium
(n∼1cm−2),andτc>107yrforthehotgas(n∼10−2cm−2).
Thepotentialpositronsourcesinclude:neutronstarsorblackholes[LingenfelterandRamaty,
1983];56Coβ-decaysinsupernovaremnants[Ellison,Jones,andRamaty,1990];other
radioactivenuclei(e.g.26Al)formedbynucleosynthesisinsupernova,nova,redgiants,
andWolf-Rayetstars[Ramatyetal.,1979];cosmicrayinteractionswiththeinterstellar
medium[Kozlovskyetal.,1987];pairproductionbygamma-rayphotonsinteracting
withstarlightphotonsintheinterstellarmedium[Mastichiadis,Protheroe,andStephen,
1991]andpulsarsandpulsarwinds[Sturrock,1971,Chi,Cheng,andYoung,1996],
probablygamma-raybursts(GRB,LingenfelterandHueter1984).
Sincethefirstdetection[JohnsonandHaymes,1973]andsubsequentidentification
[Leventhaletal.,1978]oftheGalactic511keVannihilationline,theoriginofthegalac-
ticpositronshasbecomealivelytopicofscientificdebate.Thefirstresultofthe

15

1oductionIntrFigure1.3:Yieldof44Tivsyieldof56Ni,frommodelsandobservations(from
Diehletal.2006b).ModelresultsarefromLimongiandChieffi(2003,filleddots,
withlargevariationsinyieldsduetovariationsinbothstellarmass-from15to35
M-andexplosionenergy),Rauscheret51al.(2002,crosses,forstarsinthe15to
25Mrangeandexplosionenergiesof10ergs)andMaedaandNomoto(2003,
asterisks);thelatterconcernaxisymmetricexplosionsin25and40Mstars,pro-
ducinghigh44Ti/56Niratios.44TidetectedinCasAappearsasahorizontal
shadedband(assumingthatitsdecayratehasnotbeenaffectedbyionisationin
theCasAremnant,otherwiseitsabundanceshouldbelower).Theamountof
44TiinSN1987Aisderivedfromitslateopticallightcurve.The44diagonal56dotted
lineindicatesthesolarratioofthecorrespondingstableisotopes(Ca/Fe)
].2004,Prantzos[

16

1.4511Vkeemissionomfron-positrelectronannihilationFigure1.4:All-skydistributionof511keVlineemissionobservedbyINTEGRAL/SPI
(fromKnödlsederetal.2005).

511keVemissionmorphologywasobtainedbyOSSEaboardCGRO[Purcelletal.,
1997,Chengetal.,1997,Milneetal.,2000,2001],andobservationswererestrictedto
theinnerGalaxy.WiththelaunchofINTEGRALsatellite,oneoftheprimeinstru-
ments,SPIallowsthedetailedstudiesofpositronannihilationfeatures,combining
high-resolutionspectroscopy(R∼250at511keV)withrelativelygoodangularreso-
lution(2.5◦FWHM).TherecentSPIresultsbasedontwo-yeardatafirstlypresented
anall-skymapof511keVelectron-positronannihilationemission[Knödlsederetal.,
2005].Thespatialdistributionof511keVlineemissionappearscenteredontheGalac-
ticcenter(GC,bulgecomponent,withradiusof6◦−8◦),withnoorweakcontribution
fromadiskcomponent(seeFigure1.4;Teegardenetal.2005,Knödlsederetal.2005)1,
andnoevidenceforadditionalpoint-like511keVlineemissiondowntoafluxlimitof
1.4×10−4phcm−2s−1.Thehigh-resolutionspectralanalysesshowastrong511keV
linewithapositroniumcontinuum(Figure1.5),suggestingthatthedominantfrac-
tionofpositrons(∼94%)formpositroniumbeforeannihilation[Churazovetal.,2005,
Jeanetal.,2006].Thehighluminosityof511keVemissionwithapositroniumcon-
tinuumcomponentsuggeststhatthepositroninjectionrateisupto1043e+s−1within
∼6◦oftheGalacticcenter.
Inthepresentviewpoints,thebulge-dominatedmorphologyofthe511keVlinemay
indicatethatthepositronsourcepopulationcouldbeanoldstellarpopulation.How-
ever,somecandidatepositronsources,likesupernovae,26Alsources,pulsars,cosmic
rayinteractions,showadisk-likedistribution.Thelow-massX-raybinaries(LMXBs)
1Recently,thesignificantdiskcomponentfrom511keVlineemissionwasreportedfromnear4-yearSPI
observations(Weidenspointneretal.2007inpreparation).

17

:5.1FigureoductionIntr118

Gaussian

a

omfr

contributions

with

SPI

yb

measured

spectrum

the

of

Fit

line,anortho-positroniumcontinuumandapower-lawGalacticcontinuum(from

inthewingsoftheline.

Jeanetal.2006).AsingleGaussiandoesnotgiveagoodfittothefluxmeasured

1.4511Vkeemissionomfron-positrelectronannihilationandTypeIasupernovaecouldbethecandidatepositronsources,buttheymaynotpro-
ducetheenoughpositroninjectionrateintheGalacticcenter.Asaresult,Casséetal.
[2004]arguesthatnormalsupernovaexplosionintheGCcannotcontributetothe
requiredpositrons,buthypernovae(TypeIcsupernovae/gamma-raybursts)inthe
Galacticcentermaybepossiblesourcesofgalacticpositrons.Inaddition,annihila-
tionsoflightdarkmatterparticlesintoe±pairs[Boehmetal.,2004]havealsobeen
alsosuggestedproposedasthethatthepotentialcontinuousoriginofcapturethe511ofkestarsVblineytheinthesuperGC.massivChengeetblackal.[hole2006at]
SgrA*couldexplainthemorphologyandintensityofthee±annihilationline.But
thesepositronsscenariosoverthemaygalacticrequirebulgeturbulentwithindifthefusionpositrpronocesseslifetimewhich(∼are104able−to107difyr).fuseIttheis
alsosuggestedthatamillisecondpulsar(MSP)populationexistsintheGalacticcen-
ter[Wangetal.,2005a,Wang,2006],andpositronsfromwindsofthesemillisecond
pulsarscanprovideenoughpositroninjectionrate[Wangetal.,2006].

1.4.1PositronsfrommillisecondpulsarsintheGalacticcenter
Millisecondpulsarsareoldpulsarswhichcouldhavebeenmembersofbinarysys-
temsandbeenrecycledtomillisecondperiods,havingformedfromlowmassX-ray
binariesinwhichtheneutronstarsaccretedsufficientmatterfromeitherwhitedwarf,
evolvedmainsequencestarorgiantdonorcompanions.Thecurrentpopulationof
theserapidlyrotatingneutronstarsmayeitherbesingle(havingevaporateditscom-
panion)orhaveremainedinabinarysystem.Inobservations,theydistributeintwo
populations:theGalacticfield(1/3)andglobularclusters(2/3,seedetailsinWang
).2006OnemotivationforamillisecondpopulationintheGalacticcentercomesfromthe
GeVspectrumtowardtheGalacticcenterdetectedbyEGRETonboardtheCompton
GRO[Mayer-Hasselwanderetal.,1998].Thephotonspectrumcanbewellrepresented
byabrokenpowerlawwithabreakenergyat∼2GeV(seeFigure1.6).Somemodels,
e.g.gamma-raysrelatedtothemassiveblackhole,inverseComptonscattering,and
mesonicdecayresultingfromcosmicrays,aredifficulttoproducethehardgamma-
rayspectrumwithasharpturnoveratafewGeV.However,thegamma-rayspectrum
towardtheGalacticcenterissimilarwiththegamma-rayspectrumemittedbymiddle-
agedpulsars(e.g.VelaandGeminga)andmillisecondpulsars(Wangetal.2005).
InFigure1.6,wecanseethatthesuperposedspectrumof6000MSPscouldsignifi-
cantlycontributetotheobservedGeVspectrum(Wangetal.2005).Wang(2006)also
suggestedthatnon-thermalX-rayemissionfrompulsarwindnebulaeoftheseMSPs
cancontributetoweakunidentifiedX-raysourcesdiscoveredbydeepChandraX-ray
surveysoftheGalacticcenter(afieldof17×17,Munoetal.2003).
Firstly,wewillarguethatmillisecondpulsarswoulddominatethepulsarpopulation
intheGalacticcenterregion.ThebirthrateofnormalpulsarsintheMilkyWayis
about1/150yr(Arzoumanian,Chernoff,&Cordes2002).Asthemassintheinner20

19

120

oductionIntrThe

fusedif

ygamma-ra

inom(fr

theWom

GalacticangW

centerThe).

regionsolidThe

withindashedand

5.1dashed

Figure1.6:Thediffusegamma-rayspectrumintheGalacticcenterregionwithin1.5◦
andthe511keVlineemissionwithin6◦(fromWang2006).Thesolidanddashed

spectrasimulatedthearelines

magnetic

field

distributions

in

of6000MSPsaccordingtothedifferentperiodand

globular

clusters

and

the

Galactic

field

.elyrespectiv

1.4511Vkeemissionomfron-positrelectronannihilationpcoftheGalacticcenteris∼108M[Launhardtetal.,2002],thebirthrateofnormal
pulsarsinthisregionisonly10−3ofthatintheentireMilkyWay,or∼1/150000yr.
Wenotethattheratemaybeincreasedtoashighas∼1/15000yrinthisregionif
thestarformationrateinthenuclearbulgewashigherthanintheGalacticfieldover
last107−108yr(seePfahletal.2002).Fewnormalpulsarsarelikelytoremainin
theGalacticcenterregionsinceonlyafraction(∼40%)ofnormalpulsarsinthelow
velocitycomponentofthepulsarbirthvelocitydistribution[Arzoumanianetal.,2002]
wouldremainwithinthe20pcregionoftheGalacticcenterstudiedbyMunoetal.
(2003)ontimescalesof∼105yrs.Thenumberofgamma-raymaturepulsarsisnot
higherthan10accordingtothebirthrateofpulsarsintheGalacticcenter.Sonormal
pulsarsarenotlikelytobeamajorcontributor.
Ontheotherhand,theremayexistapopulationofoldneutronstarswithlowspace
velocitieswhichhavenotescapedtheGalacticcenter[BelczynskiandTaam,2004].
Suchneutronstarscouldhavebeenmembersofbinarysystemsandbeenrecycled
tomillisecondperiods,havingformedfromlowmassX-raybinariesinwhichtheneu-
tronstarsaccretedsufficientmatterfromeitherwhitedwarf,evolvedmainsequence
starorgiantdonorcompanions.Thecurrentpopulationofthesemillisecondpulsars
mayeitherbesingleorhaveremainedinabinarysystem.Thebinarypopulationsyn-
thesisintheGalacticcenter(Taam2005,privatecommunication)showsmorethan200
MSPsareproducedthroughrecyclescenarioandstayintheMuno’sregion.
Millisecondpulsarsasthepositronsources
Weconsiderthattheelectron-positronpairproductionoccursinthepulsarouter-
MSPs.inregionmagnetosphericIthasbeenproposedthatthereisastrongmultipolemagneticfieldnearthestellar
surface,althoughaglobaldipolemagneticfieldgivesagooddescriptionofthemag-
neticfieldfarfromthestar[RudermanandSutherland,1975,Ruderman,1991].The
typicalradiusofcurvaturelofthelocalmagneticfieldisontheorderofthecrust
thicknessofthestar(i.e.l∼105cm),whichismuchlessthanthedipoleradiusofcur-
vatureofdipolefieldcomponentnearstellarsurface.Therelationbetweenthelocal
multipolemagneticfieldanddipolefieldcanbegivenby[ZhangandCheng,2003]
BsBd(R)3,(1.1)
lwhereBdisthedipolemagneticfieldofapulsar,Ristheradiusofneutronstars.For
MSPs,typicallyBd∼108−109G,Bs∼1011−1012Gwhichismuchlowerthanthe
quantumcriticalmagneticfieldBq∼4.4×1013G,sopaircascadesarealsoefficientin
field.multipolelocaltheThepairproductionmechanismisasynchrotronphotoncascadeinastrongmag-
neticfield.Photonswillbeconvertedintoe±pairsinthelocalmagneticfieldwhen
theirenergysatisfies(Ruderman&Sutherland1975)
E≥Ecrit≡2mec2Bq(R)−3.(1.2)
lB15d

)2.1(

21

oductionIntr1Theprimarye±fromtheouter-gaphavetheenergyEp=γpmec2=5.7×1012P1/3eV,
sogenerally,theenergiesofprimarycurvaturephotonsandsecondarysynchrotron
photonsarehigherthanEcrit,aphoton-electroncascadewillstartanddevelopuntil
thisconditionfails.Attheendofacascade,eachincomingprimaryelectron-positron
canproduce,onaverage,
ENe±=Ecritp=1.9×103Bd,9P1/3(lR)3,(1.3)
andthenthetotalpairproductionratecanbeestimatedas
N˙e±=fN˙Ne±=2×1033fBd10,9/7P−8/21(lR)30/7s−1,(1.4)
wheref5.5P26/21B12−4/7isthefractionsizeoftheoutergap,and
N˙=2.7×1027P−2Bd,9(R)3s−1(1.5)
listheprimaryelectron-positronspassingthroughthepolar8gap[GoldreichandJulian,
rate1969].forTaakingMSP:theN˙±∼typical5×1037eparameters+s−1(WP=ang3etms,al.Bd2006=).3×10G,thepositroninjection
Sincethesepairsearecreatedclosetothestellarsurfaceandthefieldlinesarecon-
verstellarging,surface.onlyaHosmall[1986]fractionshowmaedythatkeepthemolossvingconetowarfordthethesestarpairsandwillannihilateapproachonπ/the2,
inotherwords,mostpairswillbereflectedbythemagneticmirroringeffectandthen
movetowardthelightcylinder.Theseparticleswillflowoutwiththepulsarwindand
beacceleratedbythelow-frequencyelectro-magneticwave.
ThenhowmanyMSPsintheregionofannihilationemissions?Figure1.6hasshownthat
6000MSPscan◦contributetogamma-rayswith1.5◦,andthediffuse511keVemission
hatheveanumbersize∼of6.MSPsWebydo6000not×kno(6w◦/the1.5◦)2∼distribution105,whereofwMSPseinassumethetheGC,sonumberwejustdensityscale
ofMSPsmaybedistributedasρMSP∝rc−1,wherercisthescalingsizeoftheGC.42Then+
sa−1totalwhichpositrisonconsistentinjectionwithratefrtheomthepresentobsermillisecondvationalpulsarconstraints.populationisWhat’s∼5×more,10eour
scenarioofamillisecondpulsarpopulationaspossiblepositronsourcesintheGChas
thesomepradvoblemantagesofthetostrexplainongtheturbulentdiffusedifmorfusionphologywhichofis511kerequiredVlinetodifemissionsfuseallwithoutthese
positronstoafewhundredpc.
Therearemanypossiblepositronsourcesatpresent.Thus,howcouldwedistin-
canguishtheestimatemodeltheofatypicalmillisecondspatialdifpulsarfusionscalepopulationofpositrfromonsotherinthemodels?magneticFirstlyfield,wofe
theGC,whichisgivenbyλdiff∼(rLct)1/2(Wangetal.−52006),whererL≈Ee/eBis
thefieldLarinmortheGCradius,[ELaRosaeisettheal.,ener2005gy].ofThepositravons,erageB∼cooling10Gtimeisttheofavpositrerageonsinmagneticthe
22

1.5All-skyobservationsof26AlbyCGRO/COMPTEL6loGCwis∼angular10years,resolutionsotheofcharacteristicSPI/INTEGRALdiffusion(about2scaleisdegrees),aboutwe1canpc.assumeBecausethatofthethe
positrTherefore,onswannihilateepredictinthethatlocaltheregionspatialastheirintensitysources,distributioni.e.theofthemillisecondannihilationpulsars.lines
theshouldGC.WfolloewcouldthespatialassumethedistributionspatialofMSPsdistributionifaofMSPsmillisecondshouldpopulationfollowtheexistsmassin
becausedistributiontheofprtheoperGCmotionthoughvwelocityedoofnotMSPsknowishowrelativwellelytheloyw,wfolloewcouldeachotherreasonably.But
sourassumecesthatoriginatethetwinotheMSPdistributionspopulation,arequitethe511closeketoVeachannihilationother.lineThenifintensitythepositrwouldon
follooriginatewthefrommasssuper(e.g.novstars)aeorhyperdistributionnovae,ofthethe511GalactickeVlinecenteremssionregion.Ifcouldthefollopositrwtheons
scenario,distributiontheofmassivannihilationestarsemissionsanddensemaymolecularcorrelatetoclouds.thedarkForthematterlightdensitydarkprmatterofile.
Discriminationofthesepossiblecorrelationsmaybetestedinthefuturehighresolu-
ations.vobsertion

1.5All-skyobservationsof26AlbyCGRO/COMPTEL

The1809keVgamma-raylinefromradioactive26Alwithitsdecaytimeof1mil-
lionyearscanbeusedasthetraceroftherecentnucleosynthesisactivityinthe
Galaxy.RamatyandLingenfelter[1979]firstpredictedthe1809keVγ-raylineflux
of∼10−4phcm−2s−1rad−1fromtheinnerGalaxy,combiningthesolar27Alabun-
dancewithanestimateoftheisotoperatio26Al/27Al(∼10−5)fromsupernovae.
The1809keVγ-raylineemissionwasfirstdetectedwiththeGespectrometeron
theHEAO-Cspacecraft(Mahoneyetal.1982).Thisdetectionwasconfirmedbythe
measurementofGalactictransitsthroughthefieldofviewbytheNaIspectrometer
ontheSMMspacecraft[Shareetal.,1985].Thefollowingballoon-borneexperiments
measuredtheintensityofthe26Alemissionandderivedsomeinformationonits
angulardistribution(seevonBallmoosetal.1987,SchoenfelderandVarendorff1991,
Durouchouxetal.1993,PrantzosandDiehl1996).Andthe26Alfluxesderivedfrom
thesedifferentmeasurementsdependontheassumedGalacticdistributionof26Alline
emission.Forexample,theyfound(1.1−4.6)×10−4phcm−2s−1forapointsource
attheGalacticcenter,and(3.9−5.4)×10−4phcm−2s−1fromtheinnerGalaxyfora
flatsupernovamodel.Somappingthe1809keVγ-raylineemissionoftheGalaxycan
provideinsightintothenatureof26Alsources,andprecisedeterminationofthe26Al
flux.TheCOMPTELimagingtelescopeaboardtheComptonObservatoryperformedthe
firstsurveyof26Alγ-raylineemissioninthewholeGalaxy.COMPTELcoveredthe
energyrangeof1–30MeV,withanenergyresolutionof140keV(FWHM)around
1809keVandanangularresolutionof3.8◦(seeSchoenfelderetal.1993).COMPTEL

23

1oductionIntrhastheenoughsensitivitytostudytheoriginof26Al.
Diehletal.(1995a)analyzedthefirst-yearCOMPTELsurveydata,whichhas
coveredthewholeGalacticplane.Theyappliedtwoimage-generationalgorithms
totheCOMPTELdata:theMaximumLikelihoodmethod[deBoerandet.al.,1991,
Bloemenetal.,1994]whichconvolvesapointsourceinthedataspacewiththe3-
dimensionalpoint-spreadfunction(PSF),testingforthestatisticalsignificanceofthe
point-sourcehypothesisthroughoutthesky;theMaximumEntropymethod(Stronget
al.1991;Strong2003)whichextractsa1809keVskyimage,usingthe3-DPSF,by
maximizingtheoverallimageentropy(i.e.minimizingitsnewinformationcontent)
untilsatisfactoryagreementwiththedataisachieved.Thesetwomethodsyieldthe
mapsof1809keVemissionalongtheGalacticplanewithstructuredemissionovera
widelongituderangeandamarkedasymmetryrelativetotheGalacticcenter.The
smoothintensitydistributionexpectedfromthenovaorigincannotreconcilethedata.
Overlaysofthe26AlemissionmapwiththepositionsofsupernovaremnantsandWolf-
Rayet(WR)starsintheGalaxysuggestthatthelarge-scalestructureof26Alemission
intheGalaxymaybecontributedtomassivestars.AndtheVelaregion(theVelaSNR)
showsevidenceforasingleidentifiedclose-by26Alsource[Diehletal.,1995b].So
thefirst26AlemissionmapoftheGalacticPlanefavorsthedominant26Aloriginfrom
massivestars,presumedlycore-collapsesupernovaeorWolf-Rayetstars.
Thefirst1809keVall-skymapwaspresentedbasedonthefirstthreeyearsofCOMP-
TELobservations[Oberlacketal.,1996].Thisimagewasalsobasedonthemaximum
entropydeconvolutionmethod[Strongandet.al.,1991,Strong,2003]applyinganad-
jacentenergybackgroundmodeltotheindividualobservationperiods.Theall-sky
mapconfirmedthenon-localcharacterofthedetected26AlemissioninthefirstGalac-
ticplanesurvey(Diehletal.1995a).Later,thisfeaturewasconfirmedinthe5-year
COMPTELimage[Oberlack,1997].Mostofthe26Alemissionisattributedtoyoung,
massivestarsandstar-formationregions.
Knödlsederetal.[1999a]introducedamulti-resolutionregularizedexpectationmax-
imizationalgorithm(MREM)usingawaveletfilteringfornoisesuppression,andap-
pliedittotheCOMPTELdata.TheMREMimagesaremuchlessstructuredthan
themaximumentropymaps,butthepreviouslyreportedmainemissionfeaturescan
beconfirmed.TheMREMapproachmaybesomewhatconservativewithrespectto
imagestructures,whereasthemaximumentropyimagesmaystillincludeartifacts.
Basedonthe9-yearCOMPTELobservations(fromthelaunchofCGROinspring
1991totheendofthemission,earlysummer2001),Plüschkeetal.(2001)obtained
1809keVallskymapswithtwoimagingmethods:theMaximumEntropymethod
andMREM.Thebackgroundwasalsomodelledonthebasisofanadjacentenergyap-
proach.Bothimagereconstructions(seeFigures1.7and1.8)showanextended◦Galac-
tic◦ridgeemissionmostlyconcentratedtowardstheGalacticcenterregion(−30<l<
30),plusanemissionfeatureintheCygnusregion,andalow-intensityridgealong
theCarinaandVelaregions.Thesefeaturesconfirmthepreviouslyreportedemission
structures.Inaddition,themaximumentropyimageshowssomelow-intensityfea-

24

1.5All-sky

observationsof26AlbyCGRO/COMPTELMaximum EntropyIteration 7

Maximum EntropyS.PlüschkeIteration 71γ−Intensity ph cm−2 sr−1 s−1 x 10−3
00.000.400.801.201.602.002.402.80

Figureobser1.7:vedThewithmaximum-entrCOMPTELopovyer9all-skyyearsimage(fromofthePlüschkeGalactic2001).1809keVlineemission

MREMconverged

MREMS.Plüschkeconverged1γ−Intensity ph cm−2 sr−1 s−1 x 10−3
00.000.410.821.231.632.042.45

2.86

Figure1.8:TheMREMall-skyimageoftheGalactic1809keVlineemissionobserved
byCOMPTELover9years(fromPlüschke2001).

25

1oductionIntrturesinthelongituderangebetween110◦and270◦,e.g.theOrionregion(Figure1.7).
NeartheGalacticcenterregion,theimageshowsapossibleemissionfromthenearby
Sco-Cenregion.Alsoatlatitudesbeyond±30◦,someoftheselow-intensitystructures
arevisible,whichmaybeartifacts,subjecttofurtherstudies.
Insummary,fromtheall-sky26AlemissionimagebyCOMPTEL,theobserved1809
keVγ-raylineisascribedtotheradioactivedecayof26Alintheinterstellarmedium.
26Alhasbeenfoundtobepredominantlysynthesizedinmassivestarsandtheirsub-
sequentcore-collapsesupernovae.Furthermore,26Alfluxenhancementsaredetected
alignedwithregionsofrecentstarformation,suchasapparentlyobservedinthe
CygnusandVelaregions.

1.6Theoriginof26AlintheGalaxy
26Alisanunstablenucleus,producedalmostexclusivelybyprotoncaptureon25Mgin
asufficientlyhotenvironment[Woosley,1986],mainlydestroyedbytheβ+decayinto
26Alsincethecompetingdestructionprocess,i.e.,the26Al(p,γ)27Sireaction,becomes
efficientforT>5×107K,andtheendofcentraltheHburningbarelyreachessucha
temperature(seethereactionchainsshowninFigure1.9).Inaddition,the26Alfreshly
synthesizedmustbeejectedintotheinterstellarmediumbeforeitisdestroyedinsitu.
Soitssynthesisoccurs,essentiallyinthreedifferentspecificenvironments:i.e.,the
coreHburning,theCandNeconvectiveshells,andtheexplosiveNeburning.For
thehydrostaticnucleosynthesisinthecoreofstarswithconvectiveenvelopes,thefresh
26Alrequirestobeconvectedawayfromthehotinnerburningregionsufficientlyfastto
preventdestruction,andejectedbystrongstellarwinds.26Alcanbealsoproducedby
nuclearreactionsoflow-energyheavycosmicraysintheinterstellarmedium[Clayton,
].1994Thepresenttheoreticalknowledgeof26AloriginintheGalaxywillbepresentedas
follows(seemoredetailsinPrantzos&Diehl1996).Here,wediscussfivepossibleori-
ginsof26Alseparately:core-collapsesupernovae;Wolf-Rayetstars;novae;asymptotic
giantbranch(AGB)stars;andcosmic-raynuclearreactionsintheinterstellarmedium.

supernovaecollapseCore1.6.1Massivestars(e.g.M>8M)endupascore-collapsesupernovae(TypeIIandType
Ib/cevents).Thirtyyearsago,ithasbeensuggestedthat26Aliscreatedincore-
collapsesupernovae[RamatyandLingenfelter,1977,Arnett,1977].Explosive26Al
nucleosynthesisistriggeredbytheshockwaveintheNeburningshell(Woosley&
Weaver1980)and26Alproductioncanbeenhancedbyneutrino-inducednuclearreac-
tions[Woosleyetal.,1990].
Significant26Alproductionoccursinthelatepre-supernovaphases,when26Alsyn-
thesizedintheC/Neconvectiveshellislocatedcloseenoughtotheironcorethatis

26

γ21Ne(p,)22Na
(β+)(β+)

2122NaNe

(p,γ)(p,γ)

1.6Theoriginof26AlintheGalaxy25(p,γ)26(p,γ)27
SiMgAl(β+)(β+)

25Al26Mg(β+)

(p,γ)(p,γ)

242028272319F(p,γ)Ne(p,α)Na(p,γ)Mg(p,α)Al(p,γ)Si
Ne−NaMg−Al

Figureby1.9:dashedReactionscirclesof(fromtheRolfsNe-NaandandRodneMg-Aly1994chains.).Unstableisotopesaredenoted

partiallydestroyedbythepassageoftheshockwave.AsfortheCshell,itstypical
temperature[log(T)<9.08]doesnotallowingeneralasubstantialproductionof26Al
.ButafterthecentralSiburning,thestrongcontractionandheatingoftheinnercore
thatprecedesthefinalgravitationalcollapse,inducesastrongtemperatureincreaseof
theC-burningshell[log(T)∼9.255].Ifatthis26stage,suchaburning25occursinaneffi-
cientlyconvectiveshell,asubstantialamountofAlisproduced.Mgthatentersin
the26AlproductioncomesdirectlyfromtheinitialCNOabundance,andprotonsare
mainlyproducedbythe12C(12C,p)23Naandthe23Na(α,p)26Mgprocesses.Contrary
tovariationinCburning,26AlisalwaysproducedbyNeburningbothinthecenter
andtheshell.Asubstantialamountof25MgisleftunburnedintheCshell,protons
neededtoactivatethe25Mg(p,γ)26Alreactionmainlycomefromthe23Na(α,p)26Mg.
Andthemosteffectiveprotonpoisonisthe23Na(p,α)20Ne.Inradiativeenvironment
the26Alequilibriumabundancedependsonthelocalbalancebetweenproductionand
destruction,sothatonlythepresenceofanefficientconvectiveshellcouldactasa
preservationbuffer.Sincemostofstarsdonotreachtheonsetoftheexplosionwith
anefficientNeconvectiveshell,ingeneralonlysmallequilibriumabundanceof26Al
locatedwiththeradiativeNeburningshellispresentatthebeginningofthecorecol-
lapse.Thereareexceptionsfor14Mand15Mmodels(seetrianglesinFigure1.10).
StarswithinthismassintervalarecharacterizedbythelackofaCconvectiveshell.
AttheendofthecentralSiburning,ANeconvectiveshellformsveryclosetothe
regionwhereiswaspreviouslyefficientinaCconvectiveshell.Duringthelaststrong

27

1oductionIntrcontractionconsequenceofoftheprcore,oducingsuchaahugeNeconvamountectivofe26shellAl.Inpenetratesgeneral,atheC-richlargerregionthemass,withthethe
26etlageral.2002amount;ofLimongiAl&thatChiefsurfiviv2006es).theTheexplosiontriangles(WinoosleFigurey&1.W9eavshoerw1995the;26AlRauscheryield
producedinC/Neconvectiveshellsthatsurvivedtotheexplosionasafunctionofthe
mass.initial26Alisproducedduringexplosionatatypicaltemperatureoftheorderof∼2.3
billiondegrees.SuchaconditionoccurswithintheCconvectiveshellandthemain
processthatcontrolsitsproductionisalsothe25Mg(p,γ)26Alprocess,whileitsde-
struction25isnowcontrolled,roughlyparitheticallybythe(24n,p)and(n,α)processes.
TheMgnowcomesmainlyfromthe(n,γ)captureonmg,thisisotopebeinga
primaryoutcomeoftheCandNeburning.Theneutrondensitythatentersinboth
theproductionanddestructionof26Alisdeterminedbythecompetitionamongsev-
eralprocesses.Themainneutronproducersare(α,n)captureson26Mg,25Mg,21Ne,
and29Si,plusthe(p,n)captureon28Al,whilethemainneutronpoisonsarethe(n,γ)
captureson24Mg,16O,and20Ne.Theprotondensityisdeterminedbythecompetition
betweentheproductionviathe(α,p)captureson20Ne,24Mg,23Na,anddestruction
viathe(p,γ)captureson26Mg,20Ne,24Mg,27Aland25Mg,plusthe(p,n)reactionon
28Al.InFigure1.10,squarespresentthecontributionoftheexplosiveburningtothe
26.AlofsynthesisTheexplosive26Alyieldisgenerallyalittlehigherthanpre-supernovaproduction
withmodelsthemadifyhaferentvedifinitialferentmassespredictions,(Limongie.g.,&theChieffiexplosiv2006,esee26AlisFigurelow1.10).comparedButdiftoferentpre-
1993super].noTvaypicallypr,oductionthe26Alforyieldsstarsfrwithommassescore-collapsehighersuperthanno35vMae[rangeWeavfreromand2×W10−oosle5−y,
5×10−4Mwithinitialstarmassesof12−120M(LimongiandChieffi2006,alsosee
).10.1Figure

starsolf-RayetW1.6.226HyAldr.AtostaticcentralcoreHHburexhaustion,ninginthethe26mainAlissequencelocatedinstarsthecanHeprcoreoduceandlaringetheamountsregionofof
variableHleftbehindbytherecedingconvectivecore.SincetheHeburningeasily
andquicklydestroysthe26Al(viathe(n,α)and(n,p)reactions),theamountof26Al
synthesizedbycentralHburningandpossiblepreserveduptotheexplosionisjustthe
onelocatedintheH-richlayersplustheonelockedinthefractionoftheHecorethat
efwouldfective,not26beAlafwouldfectedbymostlytheHedecayburbeforening.Ifitthecouldbedredge-upejectedandbythethemass-lossexplosion.wereThesenot
twtimesoandphenomena:thenincreasedredge-uptheamountandof26mass-loss,Alejectedmayintoanticipatethesuchinterstellarejectionmedium.atearlier
Starswithmassesbetween10−35Mwillundergoadredge-upepisodethatdoes
notenterintotheHecore,andthemass-lossisweak.SincetheHeconvectiveshell

28

Figure

1:10.

26Al

yields

in

ferentdif

processes

1.6(C/Ne

Theoriginofeectivconv

windsintheWRstarphase)asafunctionoftheinitialstellar

calculationsbyLimongiandChieffi2006).

26Alinshells,

theGalaxyexplosion,

om(frmassrecent

29

1oductionIntrFigure1.11:Comparisonamongthe26Alyieldsasafunctionoftheinitialmassderived
bydifferentworkinthelasttenyears(thedatapointsofsolidcirclesaretaken
fromLimongiandChieffi2006).Seemoredetailsinthetext.

extendsalmostuptothebaseoftheHburningshell,onlyatinyamountof26Al
whichispresentintheregionofvariableHleftbytherecedingHconvectivecoreand
engulfedintheconvectiveenvelope,wouldbeejectedintotheinterstellarmedium.
Starsmoremassivethan35Mdonotshowdredge-upepisodes,butthemass-loss
issostrongthatasubstantialfractionoftheHecoreisejectedthroughstellarwinds,
i.e.,theW26olf-Rayetphase[vanderHuchtetal.,1988].FortheWolf-Rayetstars,alarge
amountofAlpresentintheHecoreisthuspreservedfromthedestructionand
ejectedintotheinterstellarmedium.Sothemain26AlproductionintheISMbefore
supernovaexplosionscomesfromthestellarwindsofWolf-Rayetstars.
Thelargeramountof26AlisproducedwiththehigherinitialmassesoftheWolf-
Rayetstars(Figure1.10).Theaverageyieldof26AlproductionduringtheWolf-Rayet
phaserangesfrom1×10−5−3×10−4Mforwithinitialstarmassesof35−120M
[LimongiandChieffi,2006,Palaciosetal.,2005].
Fromtheall-skysurveyof1809keVemissionbyCOMPTEL(Plüschkeetal.2001,
alsosee§1.3),the26Alintheinterstellarmediumispredominantlysynthesizedinmas-
sivestars,throughthestrongstellarwindsofWolf-Rayetstarsandcore-collapsesuper-
novaexplosions.Thetotal26Alyieldsduetothetwoprocesseshavebeencalculated
bydifferentauthors.Figure1.11displaysthecomparisonamongthe26Alyieldswith
afunctionofaninitialmassbythedifferentworks(e.g.,LimongiandChieffi2006;

30

1.6Theoriginof26AlintheGalaxyW1996;oosleyRauscherandWeaetval.er19952002;;LangerPalaciosetetal.al.19952005;).MeTheynetavetal.erage199726Al;yieldsThielemannproetvidedal.
byvariousw−4orkrangefrom(0.3−30)×10−5Mfor10−35Mstars,26andfrom
(1−10)×10Mforstarsabove35M.ThecalculationsofthetotalAlyield
wouldbedirectlycomparedwiththeobservationallimits.Knödlseder[1999]argued
thatmostof26AlintheGalaxycomesfromWRstars,andthestudyof26Alinthe
Cygnusregionisawaytoresolvethisargument.

Novae1.6.3826aAlfastprdeclineoductionfromrequiresmaximummoderatetemperaturepeak[Wtemperatures,ardandFoe.g.,wlerT,peak1980<].2×These10K,condi-and
tionsarecommonlyachievedinnovaoutbursts.In1980s,one-zonemodelcalcula-
tionsofexplosiveH-burningnucleosynthesiswithsolarorCNO-enhancedenvelopes
v[aeHillebrandtmightprandoducesufThielemannficient,amounts1982,Wof26iescherAltoetal.account,1986for]somesuggestedofthethatobservclassicaledmete-no-
theoriticbasisofanomaliesONeMgbutwwhiteoulddwnotarfstarsrepresent[WeissmajorandTGalacticruran,sour1990,ces.Nofar,NewShaviv,calculationsandonStarrfield,
2622,1991]prconcludingoducedthatlarthegeONeamountsnovofaelong-livmightedberadioactivimportantesournuclei,cesofsuchtheasNaGalacticand26AlAl.
AccordingtoNofar,Shaviv&Starrfield(1991),no26Alproductionoccursformodels
withStarrfieldTpeaket>al.2.7,×1993108],K.Butdemonstratingsomethecalculationscrucialrrefuteolethisplayedresultby[convPolitanoectionetal.to,carr1995y,
26prsomeotonAlcapturestotheisouterprev,ented.coolerlayFurtherersofmore,thetheenvprelop,oductionwhereofits26Albydestructionnovaethrisvoughery
sensitivadopted.etoONethenovinitialaeshouldcompositionbemoreoftheimportantenvelope26AlandsourtocesthethannuclearCOnovae,reactionbecauserates
seednucleifortheNe-NaandMg-Alcycles26(seeFigure1.9)arealmostabsentinCO
novae.Forthesamereason,theamountofAlsynthesizedinONenovaedependson
theinitialcompositionofthewhitedwarfcore.Someimprovementsinthenuclearre-
actionratessinceCaughlanandFowler[1988]wouldleadtoalower26Alproduction.
Recentpredicthythatdrtheodynamicejected26AlcalculationsmassbyofnoONevanovaeoutburstsranges[Jose,fromHer(0.3nanz,−1.7and)×Coc10,−81997M]
consideringvariouswhitedwarfmassesandaccretionrates.
26asThethemassamountofoftheAlunderlyingejectedintowhitethedwarfinterstellarincreases[mediumJosebetyal.,ONe1997no].vaeSothedecreaseslow-
masswhitedwarfsaremostlikelycandidatesfor26Alproduction,withthehigher
26Alproductionandhigherejectedmass.Butwhitedwarfslowerthan∼1.1Mare
expectedtobeCOwhitedwarf,whichareunabletoproduceimportantquantities
beof26∼Al2.×10−Then8Mthe.Themaximumpredictedejectioncontributionmassofof26AlnovbayoneoutburstsONetonovtheaevGalacticentw26ouldAl
ranges(0.1−0.4)M,whichissmallcomparedwiththepresentobservationallimits

31

1oductionIntr26(Mgal(INTEGRAL/SPIAl)∼2[MDiehl)etderival.,edb2006ya].COMPTELHence,novaemeasurementsrepresent[Diehlimportantetal.,26Al1995asour]andces
inthehypothesisGalaxyof,ybutoungcannotbepopulationstheasdominantmajorsourones,ceswhichoftheisGalacticconsistent26Alwith(seethe§1.3).accepted

1.6.4AsymptoticGiantBranch(AGB)stars
Theintermediate-massstars(1.5≤M/M≤6)canevolvethroughathermally-
pulsingphasewiththestrongstellarwind,so-calledasymptoticgiantbranch(AGB)
stars.AGBstarsarecharacterizedbytwoburningshells,oneofhelium,andoneof
hydrogen,andbyadeep26convectiveenvelopeextendingfromabovetheH-burning
shelluptothesurface.AlcouldbeproducedinthreesitesofAGBstars:theH-
burningshellviaMg-Alchain;theHe-burningshellviaα-captureon22Ne,andat
thebaseoftheconvectiveenvelopeinthemostmassiveAGBstarsthatexperienceH-
burning.InAGBstars,26AlisefficientlyproducedbyH-burning,butdestructionby
n-capturereactionduringtheinterpulseandpulsephasesbecomeincreasinglymore
efficient[MowlaviandMeynet,2000].
Theejected26AlmassesbyAGBstarsdependonthetemperaturewhichdirectly
relatestotheinitialmass,andinitialcomposition.Recentcalculationsshowthat
thelowmassAGBstars(<4M)cannotsignificantlycontributeto26Alproduction
[MowlaviandMeynet,2000,KarakasandLattanzio,2003].TheAGBstarwithmasses
of4−6Mcanyield26Alintheregimeof(0.2−8)×10−8M(Karakas&Lattanzio
2003).TheAGBstarswithlowermetallicityproduceahigheramountof26Al.The
roughestimationofthecontributionbyAGBstarstotheGalactic26Alvariesfrom
0.01prediction−0.4Mof26Al(MobywlaviAGB&Mestars,ynetAGB2000).starsThoughcannotthebelarthegemainuncertainties26Alsourexistcesforinthethe
Galaxy.ButtheAGBstarscouldbebestcandidatestoexplaintheinferred26Al/27Al
ratiosrangingfrom∼10−4−10−2observedinmeteoriticgrains.Asaninteresting
science,modelspredictahigheramountof26Al(∼(1−2)×10−7M,Mowlavi&
26Meynet2000)aroundplanetarynebulae,whichcouldbepossiblecandidatesfordirect
future.indetectionAl

1.6.5Interactionsofcosmicrayswiththeinterstellarmedium
Theinteractionsofacceleratedparticleswithambientmattercanproduceavariety
ofgamma-raylinesfollowingthede-excitationofexcitednucleiinboththeambient
matterandtheacceleratedparticles.Nuclearreactionoflowenergyheavycosmic-ray
particleshavebeenproposedasanother26Alsourceprocess[Clayton,1994].Based
ontheCOMPTELmeasurementofexcited12CintheOrionmolecularcloudcomplex
[Bloemenetal.,1994],Clayton(1994)suggestedthatthiscouldbeanefficient26Al
sourceprocess.Thecrosssectionsforthe26Mg(H,n)26Aland28Si(H,ppn)26Alreactions
areofthesimilarmagnitudeastheoneforthe12C(H,p)∗12Creaction.Therefore,the

32

1.7Theoriginof60Feestimation−4fortheOrionregionofactivestarformationcorrespondstoan26Alyield
of∼10M.
TheGalactic26Alyieldisquiteuncertain,dependingonthefractionsofmolecular
cloudsirradiatedbylow-energycosmicrays.AndtheabsenceofsubstantialGalactic
plane4.4MeVemissiondueto12Cde-excitationsuggeststhatthisprocessisprobably
negligibleasaGalactic26Alsource[Ramaty,1996].

1.7Theoriginof60Fe

Theradioactiveisotope60Feisbelievedtobesynthesizedthroughsuccessiveneutron
captureson60Feisotopes(e.g.,56Fe)inaneutron-richenvironmentinsideHeshellsin
AGBbeforestarsor(duringFeistheirstoredfinalinevwhiteolutiondwarfstocoreandcannotcollapsebesuperejected),novaeand(seemassivFiguree1.stars,11).
6060FecanbealsosynthesizedinTypeIaSNe59[Woosley,1997].Itis59alsodestroyedbythe
Fe(n,γ)process.59Since−itsclosestparent,Feisunstable,theFe(n,γ60)processmust
competewiththeFe(γ)decaytoproduceanappreciateamountofFe.

starsMassive1.7.1Aneutron-richenvironmentinmassivestarsisrequiredtoproduce60Fe.Anda
temperatureoftheorderof2billiondegreesrepresentsanupperlimitforthesynthesis
of60Febecauseabovethistemperaturethe(γ,n)andthe(γ,p)photondisintegrations
ofboth59Feand60Febecometremendouslyefficient.Suchanoccurrencelimitsa
possible60FeproductiontotheHe,C,Neshellburningphases.
InHeburning,themainneutrondonoristhe22Ne(α,n)25Mgprocess,andatempera-
tureoftheorderof4×108Kwouldberequiredtoreachthethresholdneutrondensity
of3×1010ncm−3.IncentralHeburning,thetemperatureremainsbelow3×108K,
sothattheneutrondensityneverexceeds107ncm−3,andnoappreciableproduction
of60Feoccurs.InshellHeburning,thetemperaturecouldraisetoandabove4×108
K,thenalargeamountof60Femaybesynthesized.Forthestarswithmassesbelow40
M,thetemperatureatbaseoftheHeconvectiveshellneverraisesenoughtomake
the59Fe(n,γ)processcompetitivewithrespecttothe59Fe(β−)decay,sono60Feissyn-
thesizedinthesestars.SuchfeatureholdsuptothefirstmassthatbecomesaWR
star,whichexperiencessuchstrongmasslossthatitfirstlosesallH-richenvelopeand
thencontinueserodingtheHecoreuptothemomentofthecorecollapse.Inthese
stars,theHeconvectiveshellformswithintheregionofvariableHeabundance,then
aproblemarisesofwhethertheSchwarzschildcriterionortheLedouxoneisusedto
determineifaconvective60regionforms.InFigure1.13,Thesolidandopencirclesrefer
totheamountofejectedFefortwodifferentchoicesofthestabilitycriterioninthe
Heconvectiveshell,i.e.,theSchwarzschildandtheLedouxcriteria,respectively.C
burningbehavessimilarlytotheHeburning.Theneutrondonoristhe22Ne(α,n)25Mg

33

1oductionIntrprocess,withα-particlesprovidedbythe12C(12C,α)20Nereaction.InCshellburning,
thehightemperature(>109K)allowsalargeproductionofα-particleswhichtrans-
latesintoahighneutrondensityandhencealargeyieldof60Fe.Cconvectiveshell
couldproduceaconspicuousamountof60Feinstarsoftheinitialmassbelow40M
.Neburningmayproduce60Fe,butthelackofanextendedandstableconvective
shelllastinguptotheexplosionpreventsthebuildupofasignificantamountof60Fe
.Theaverageyieldofejected60Feinconvectiveshellsvariesfrom10−6−7×10−4
M,increasingwiththeinitialmassfrom11-120M[LimongiandChieffi,602006].
CalculationsbyLimongi&Chieffi(2006)alsosuggestedthedominantyieldofFein
massivestarsabove40McomesfromtheHeconvectiveshellburning,whichsen-
sitivelydependsonthemass-lossrate.Modelswithastrongmass-lossrate[Langer,
1989]wouldreducethe60FeproductionduringtheWRphasesintoalevelof∼10−5
M(seeFigure1.13).

Supernovae1.7.2Thelastepisodeofsynthesisof60Feoccurswhentheblastwavecrossesthemantle
ofthetemperaturestaronisitsofwatheyoroutderofduring2.2the×109K,core-collapseandhencesuperrnovoughlyainexplosions.thesameTheregionpeak
wheretheexplosivesynthesisof26Aloccurs.Theaverageyieldof60Feinexplosion
variesfrom10−5M(initialmassregionof11−50M)to4×10−5M(50−120M
).Figure1.12showsthetotalamountofejected60Feasafunctionoftheinitialstellar
masses(from11-120M)calculatedbyLimongiandChieffi[2006].Theresultsbythe
previouswork(initialmassesfrom12-40M,Woosley&Weaver1995;Rauscheret
60al.Fe2002yield)bareyalsoRauscherpresented(2002)forareacomparison.significantlyInlargerthethaninitialthemassyieldsrangeby<both40MW,ooslethey
&Weaver(1995)andLimongi&Chieffi(2006).InFigure1.12,thesolidandopen
circlesareobtainedbyadoptingthemass-lossratebyNugisandLamers[2000];the
starsareobtainedbytakingtheratebyLanger[1989].Themass-lossrateproposedby
astrLangeronger[1989]ismass-lossmuchrate,strongertheyhathanvetheNugissignificantlyandreduceLamersthe[200060Fe]prone.Woductionithadoptingduring
theWRphaseswiththeinitialmassesabove40M.
Inaddition,60Fecouldalsobeproducedinsubstantialamountsbyraresubtypesof
SNIa[Woosley,1997],whichwouldthenbepointsourcesof60Fegamma-rays.The
averageyieldof60FeinSNIais(1−5)×10−3Mdependingonmodels.The60Fe
pereventistypically100timesgreaterinahigh-densitywhitedwarfexplosionthan
inaTypeIIsupernova.Withaneventrate10−4yr−1,thecompositesignal60herecomes
from∼100pointsources.Asinglesourceat10kpcthatmade0.005MofFewould
bevisibleforseveralmillionyearsatafluxlevelof10−7phcm−2s−1.However,this
iswellbeyondthecapabilityofpresentgamma-raytelescopes.

34

1.7Theoriginof60Fe60Figuree.g.,1.5612:Fe,Theand59captionFeβof-decaFeyprwilloducedgeneratebyaneutrleakoninthecapture60Feprprocessesoduction.ofironisotopes

Figure1.13:Comparisonamongthe60Feyieldsasafunctionoftheinitialstellarmass
prciroclesvidedreferbytovtheariousamountauthorsof(frejectedom60FeLimongifor&twoChiefdiffiferent2006).choicesTheofsolidtheandstabilityopen
criterionintheHeconvectiveshell,i.e.,theSchwarzschildandtheLedouxcriteria,
60therespectivLangerely[.1989The]starsmass-lossrepresentratetheformassivamounteofstarsejected(WRFestarsofobtainedmassesbyfromadopting40-
[1201995M]and).SolidRauschersquaresetal.[and2002]intrianglestheareinitialthemassresultsrangefromof12W-oosle40yMand.Weaver

35

1oductionIntr(SPI)spectrometerINTEGRALThe1.8

TheINTErnationalGamma-RayAstrophysicsLaboratory(INTEGRAL)isanEuropean
(ESA)Gamma-RayObservatorySatelliteMissionforthestudyofcosmicgamma-ray
sourcesinthekeVtoMeVenergyrange[Winkleretal.,2003].INTEGRALwassuccess-
fullylaunchedfromBaikonurCosmodrome(Kazakhstan)onOctober17,2002using
aPrcentric,otonrwithocketanproapogeevidedofby153the000Russiankm,aSpaceperigeeAgencyof.9000Thekm,andINTEGRALa3dayorbitisperiodec-
[Jensenetal.,2003].
ThescientificobjectivesofINTEGRALconcentrateon:

(1)stellarnucleosynthesisviadetection,cartographyandfinespectroscopyofsources
such(nearbasythoseSNRs)of44andTi,26Alstructures,60Feand(Galacticthe511disk)keVpositremittingonradioactivannihilationeline;gamma-raylines
(2)compactobjectssuchaspulsars,neutronstarsandblackholesinbinaries,and
massiveblackholesinthecentersofactivegalacticnuclei;
(3)gamma-rayburststudiesusingfastdetectionandpreciselocalizationatthearcmin
leveltotriggerafterglowstudies.
withThePrincipalrepresentativInvesofestigatorsthe(PI’s)participatingandMissionagenciesSconstitutecientistsfortheINTEGRALINTEGRALStogethercience
WorkingTeam(ISWT).AllPIteams,providingthepayloadanddatacenter,consist
oflargeinternationalcollaborationsfromscientificinstitutesfromalmostall14ESA
memberstates,USA,Russia,CzechRepublicandPoland.
typeINTEGRALmission[wWasinklerconceivetal.ed,fr2003om].itsMostinitialofstudythetotalphaseinobser1989vingastimean(∼observ70%)atory-is
awardedastheGeneralProgrammetothescientificcommunityatlarge.Proposals
forobservationsareselectedontheirscientificmeritonlybyasingleTimeAllocation
AsaCommitteereturnto(TAC).thoseThesescientificselectedcollaborationsobservationsandaretheindividualbaseofthescientistsgeneralwhoprcontributedogramme.
tointhethedevINTEGRALelopment,SciencedesignWandorkingprTeamocurement(ISWT),ofaINTEGRALportionofandthewhototalarescientificrepresentedob-
servingtime,theguaranteedtime,willbeusedfortheirCoreProgrammeobservations.
INTEGRALdataarenotautomaticallyPublic.Withinoneyear,thedatabelongtothe
PIoftheobservation.
Timagerwomain(IBIS)withinstrumentsanangularareonboarresolutiondof12,INTEGRAL(seeallowingFigurefor1sour.14ce):thelocalizationINTEGRALwith
arcminprecisewithafieldofviewof9◦×9◦[Ubertinietal.,2003];theINTEGRAL
lutionspectrofometer2.5◦(SPI)withinawithfieldanofenerviegywof16resolution◦×16of◦.2.T5wokeVsmallat1.3MemonitorsVandareangularincludedreso-in
thepayload:JEM-XcomposedoftwoidenticalX-raytelescopesthatprovidecomple-

36

Figure.1:14The

trometer

four

SPI;

the

main

yX-ra

instruments

monitor

daboar

JEM-X;

8.1

The

INTEGRAL

INTEGRAL:

and

the

optical

the

ometerspectr

imager

camera

IBIS;

OMC.

the

(SPI)

spec-

37

ACS

subsystems

1

oductionIntr

:15.1Figure38

A

scintillator

are

highlighted.

plastic

mask,

coded

The

instrument.

SPI

the

of

wvie

yawcut-a

brief

anticoincidence

subassembly

(PSAC),

Ge

camera

and

1.8TheINTEGRALspectrTable1.2:KeyparametersoftheINTEGRALspectrometerSPI

ometer(SPI)alueVParameterEnergyrange18keV–8MeV
Detector19Gedetectors,cooledat85K
2cm500areaDetectorFieldofview(fullycoded)16◦(cornertocorner)
Spectralresolution(FWHM)3ke◦Vat1.8MeV
Angularresolution(FWHM)2.5(pointsources)
Continuumsensitivity(3σ,106s5.5×10−6phcm−2s−1keV−1at100keV
Linesensitivity(phcm−2s−1,3σ,106s3.3×10−5at100keV,2.4×10−5at1MeV
Sourcelocation(radius)≤1.3◦(dependingonS/N)
Absolutetimingaccuracy(3σ)≤200µs
kg1309MassPower(max/average)385W/110W

mentarydatafrom1to15keV;OMC,anopticalcamera(Winkleretal.2003).
moreHere,detailsweonlyinVproedrennevideaetal.brief2003;descriptionRoquesofetal.the2003spectr).SPIometerconsistsSPI.of(Figurethe1follo.15,wingsee
mainsubsystems[Vedrenneetal.,2003]:

(1)thecamera,composedof19highpuritygermaniumdetectors(GeD)andtheir
onics;electrassociated(2)apreamplifierstwo-stageto215coolingK;thesystem:activethestagepassivcoolsethestageGecoolsarraythedocrwnytoostat85–90housingK;andthe
(3)apulseshapediscriminationsystemwhichallowsdiscriminationbetweensingle
andmulti-siteinteractionsinoneGedetector;
(4and)anaactivplasticescintillatoranticoincidenceshieldanticoincidence(ACS)madesubassemblyof91Bi3(PSAC);Ge4O12(BGO)crystalblocks
(5)adigitalfront-endelectronics(DFEE)providingtheSPIinternaltiming,thevarious
(6)acoded(anti)coincidencemaskwhichfunctions,allowsandtheimagingprimarofythedatasky.encoding;
VAedrennesummaretyal.of2003key).instrumentparametersofSPIispresentedinTable1.2(alsosee
Atpresent,SPIhasbeenfunctioningwellforfouryears.Observationsandscience
analysesofSPIdatawillbepresentedindetailsinChapter2.Gedetectorsallowfor
highindividualspectralgamma-raresolutionylinesof∼and2.5ketheirVat1shapes,MeV,e.g.thesuitable511forkeVastrline,ophysicalγ-raylinesstudiesfromof

39

1oductionIntr

radioactivities

40

of

44

Ti,

26Al

and

60Fe

.

2DataanalysesofINTEGRAL/SPIatMPE

TheSPIspectrometeronaboardINTEGRALfeaturesacamerasystemwith19Ge
detectors,imagingphotonsthroughatungstencodedmask.Thoughbackgroundis
reduceddominatesbytheanoverallanticoincidencegamma-raydetectorsignals.systemsurroundingthem,thebackground
prTheoximatelycoded50%maskofthecastsacamerashadowareaontoforathepointcamerasourceplane,intheefskyfectiv.elyVariationoccultingoftheap-
camerapointingaroundthesourcedirectioninaditherpatternisusedtocollecta
databaseofshadowgramswhichcanbedeconvolvedtofindthesourcelocationalso
inthepresenceofalargebackgroundsignal.ThoughSPIsupportsγ-rayimaging,it
kehasVatbeen1MeV,optimizedAttiéetforal.γ-ra2003y,lineRoquesspectretal.oscopy2003).withInathishighthesis,energywealsoresolution(concentrate∼2.5
onthespectroscop26yanalysis60oftheSPIdata,withstudiesofγ-raylinesfromcosmic
radioactivityofAlandFeintheGalaxy.
CalibrationcampaignsofthespectrometerSPIhavebeenperformedbeforelaunch
inarea,orderthetospectraldetermineresolutiontheandinstrumenttheangularcharacteristics,resolutionsuch[Attiéasettheal.ef,2003fectiv].edetectionAbsolute
determents.Atmination1MeofV,thetheefeffectivfectiveeareaareaishas65cmbeen2foraobtainedpointfromsourceonsimulationstheopticalandaxis,measure-the
spectralresolution∼2.3keV.◦Theangularresolutionisbetterthan2.5◦andthesource
separationcapabilityabout1.Temperaturedependantparameterswillrequireper-
usedmanentastracersin-flighttoextractcalibration.theNeutrbackgronoroundspallationcomponentinducedofsomebackgrlinesoundofastrlineswillophysicalbe
interest.enceTheanalysisINTEGRAL[CourSvcienceoisieretDataal.,2003Center],i.e.,(ISDC)anprorovidesganizedthearchivbasiceoftheinfrastructuredata,andforthesci-
associatedsoftwaretoolstoprepare,execute,andviewdataanalysissteps.TheISDC
specificinfrastructurealgorithmswaswerepreparedisolatedinainmost“instrument-specificinstrument-independentsoftware”way,modules.andinstrument-Standard
cessinganalysishastasks,beenstartingpreparedfromattheISDCpointingthrsetoughscriptsdefinition,whichandrperforoutinelymaendingpipelineinofimagespro-
and/orpointsourcespectra,throughe.g.,theSPIROSsoftware[SkinnerandConnell,
2003].Suchanalysiswilladequatelyaddresspointsourceswithcontinuumspectra
withintheinnerfieldofview.Andγ-raylinespectralanalysisanddiffuseemission
wdataouldhaanalysisvelesswmethodsandell-conditionedalgorithmsanalysisatprISDCoblems.wasThepresentedgeneralinDiehldescriptionetal.[of2003thea].SPI

41

242

DataanalysesofINTEGRAL/SPIatMPEISDC-Prepared DataInstrumental Parameters
Background
Energy CalibrationParameters
Science Baseline Data
Pointings
Good-Time IntervalsBackground Templates
Dead Time
Corrected Eventsfrom Tracer Candidates
e.g. GEDSAT Rates
Data Selection and AssemblyInstrumental Response
Function
(spiselectscw)Degradation
Excluding Anomalies
Data of AdjacentAdjacent Energy
Science DataEnergiesTemplate
per Detector
per Pointing
plus Dead Time and
Pointing InformationBackground Modelling(spiorthomodel)
Background Models: Multiple Components

Spectral Extraction from Model Fitting ( spimodfit)
Spectra from the Sky: Diffuse Emission and Point Sources
Further Analysis: Imaging and Timing

Figure2.1:SchemefortheSPIdataanalysissoftwaresystemdevelopedatMPE.

2.1INTEGRAL/SPIobservationsInthischapter,wewillgiveanoverviewoftheSPIdataanalysissoftwaresystem
developedatMax-Planck-InstitutfürextraterrestrischePhysik(MPE,Figure2.1).This
sionanalysiscomponentssystemhasfrombeenlargeoptimizedobservtoationstudysetscovindividualeringfourγ-rayyearslinesofanddata.diffuseemis-
Thedataanalysisstepsare(alsoseeFigure2.1):
(1)Assemblyofdataselectedtobefreeofcontaminationby,e.g.,solar-flareevents.
(2)Modellingtheinstrumentalbackgroundoverthefulltimeperiodofobservations
database.INTEGRALfulltheusinga(3)modelFittingofthecelestialmeasuredgamma-radatayinnarremission,owenerfoldedgybinsthrwithoughthethebackgrinstrumentaloundmodelresponseand
intoenergythebindatathenspaceofcomprisesthethemeasurement;spectrumoftheobserfittedvedskyamplitudeofemission.thecelestialmodelper
(4)Derivinglineparametersofthecelestialsignalincludingpointsourcesanddif-
emission.fusetionIninthethefollowing,correspondingwewillsoftwdiscussareinutilitiesdetailswithintheanalysisMPE’sstepsandINTEGRAL/SPItheirdataimplementa-analy-
sissystem.Firstly,wewillintroducetheobservationswithSPI.

observationsINTEGRAL/SPI2.1

INTEGRAL/SPIisaspectrometerdevotedtotheskyobservationinthe18keV–8MeV
energyrangeusing19Gedetectors(GeD).Theperformanceofthecryogenicsystem
isnominalandallowstocoolthe19kgofGedownto85Kwithacomfortablemargin.
TheenergyresolutionofthewholeGecamerais∼2.5keVat1MeV[Roquesetal.,
2003,Attiéetal.,2003].However,thisresolutiondegradeswithtimeduetocosmic
particleirradiationinspace.Andtheinstrumentisperiodicallyshutoffforafewdays
whileannealing(byheatingfromcryogenictemperaturesto100◦C)isappliedtothe
detectorstorestoretheenergyresolution(i.e.Figure2.2)byheatingofthecosmic-
ray-induceddefects[Roquesetal.,2003,Leleuxetal.,2003].Thesuccessofannealing
highdependsontheinitialdamageandontheannealingtemperatureandduration.
SPIGedetectorsareheatedbyfourresistorsgluedontheGedetectorarraysupport.
FromenergyresolutionevolutionwithtimeinFigure2.2,speciallythechangesafter
eachannealing,therecoverylooksnearlyperfect.TheannealingcapabilityofSPIisa
constraintonthedesignofthecamerabutnecessarytoensuregoodperformancefor
alonglifetimemission(>5years).
Primaryγ-raysignalsaregamma-rayinteractionsinthe19Gedetectormodules
ofthecamera,translatedindetectorsignalamplitude,shapeandrelativetimeamong
detectorunits.Triggersofjustoneofthe19detectorsarecalledsingleevents(SE,pseudo-
detectorIDs0–18),withthreesubclassesdistinguished(inpseudo-detectorIDs85–141),
dependingonthesuccessofthepulseshapedetermination.Abasiceventmessage
holdsdetector-ID,triggertime,signalamplitude,andmeasuredpulseshapeinforma-

43

:244

.olu-vre

2INTEGRAL

Figurethe

of

MPEunits

atin

INTEGRAL/SPItime

ofwith

analysesolutionve

Dataresolution

2gyEner

andtion(3dadurationsys)arinoundunits1800ofkehours.V,andThissixenerannealinggyprresolutionocessesisaredeternotedminedwithfromdatesthe

spectrumforeachdetectorindividuallyandthenaveragedoverall19detectors.

2.1INTEGRAL/SPIobservationsFigure2.3:AveragedSPIrawspectrumdominatedbybackgroundandinstrumental
lines[Weidenspointneretal.,2003].SingleGeDevent(SE+PE)anddoubledetec-
torevent(ME2)spectraareshownseparately(fromJeanetal.2003).

tion.Ifthepulseshapesarederived,thesearethencalledpulse-shape-discrimination
events(PE).Detectortriggerswhichoccursimultaneouslyinmorethanoneofthe19
detectordetectorsIDswithin19–a84).Thecoincidenceymayinterarisevalfrofom350annsareinteractioncalledcascademultipleofaeventssingle(ME,primarpseudo-y
photon.Asubclass,doubleGedetectorevents(ME2,pseudo-detectorIDs19–60)trigger
twoheightsGearedetectortransmittedelements.togetherForME,withthetherelatividentifiersearrivofaldetectorstimesininvolvdetectors.edandTheseallpulsemul-
realtiplesefdetectorsfectivelyforimprconstituteovedvirtualsensitivitydetectorandmodulesangular,whichresolutioncanbeatusedhighertogetherenergies.withThe19
countratioofMEtoSEriseswithincreasingenergies,being40%at2MeV(see
).3.2Figure2003All).evThisentsaresystemprproocessedvidesbeyvtheentdigitaltimingfrwithont102end.4µselectraccuracyonics,ev(DFEE,entVrejectionedrenneetusingal.
themonitored.ACSvetoThesignal.dataprTheeocessingventselectrarealsoonicscounted(DPE)andreceivesthefromdeadthetimeDFEEisperallthemanentlydata
neededdistributionforofsuch19onboarGeDsdandanalysisthepriornon-saturatingtotelemetrnon-vytoetoedground.counts(Figure<82.Me4V).showsthe
Thenumberofdetectorsinthearrayistoosmalltoachievegoodimagingperfor-

45

2DataanalysesofINTEGRAL/SPIatMPEFigure2.4:Distributionof19GeDsandthenon-saturating,non-vetoedGecountson
theSPIcamera(fromRoquesetal.2003,verticalaxisincounts/s,YsiandZsiaxis
=positionsofthedetectorsinmm,detectornumbersmarkedontheplot).The
outerdetectorsshowlesscounts,sinceComptonscatteredphotonsmayescape
intotheBGOcrystalshieldandactivateavetosignal,whichsuppressestheevent.

mancefromasinglepointing,sosuccessiveimagesofthesamefieldofviewaretaken
bothwithSPIslightlyanddifIBISferentaboardpointingINTEGRALangles.[VThisedrenneistheetal.,dithering2003].Atstrategyintervwhichalsofisusedtypicallyfor
30ditheringminutes,patterthen.spacecraftThisisstrategyreorientedallowstheusingsourinertialcestobewheelstoseparatedfollowfraomtheprescribedback-
ground,whichisparticularlydifficultduetothelowlevelofthesourcesignal(e.g.,
26AlemissionintheGalaxy,fewpercentofthebackground).Italsominimizesthe
effectsofthesmallvariationsofthebackgroundwhichoccurdespitethechoiceofan
eccentricorbitoutsidetheradiationbelts.
TheSPIdataarepre-processedusingtheISDCstandardanalysissoftware(OSA)
uptothelevelofeventhousekeepingintospectraandthecorrespondingassembly
ofpointingandlivetimeinformation.ThepresentversionisOSA6.0whichwaspub-
lishedinDecember2006.Theenergycalibrationisperformedusinginstrumental
backgroperationoundforγlines-raywithlineknostudieswnener(sub-kegiesV[Waccuracy),eidenspointnerforetcontinuumal.,2003].studies,Thisaisastandarcriticald

46

2.2Dataselectionandassemblycalibration(about1keVaccuracy)isquiteadequate.Variousenergybinningscanbe
useddependingontheavailablestatisticsasafunctionofenergy.Afterthefailureof
detector2attheINTEGRALJulianDate(startsat1Jan2000)1434,theremaining18
detectorswereused;andafterthefailureofdetector17attheJulianDate1659,the
remaining17detectorswereused(seeFigure2.4).

assemblyandselectionData2.2Wetestdataqualitypereach‘sciencewindow’(typicallyatimeintervalof∼30
minutescorrespondingtoonepointing),applyingselectionlimitsto‘sciencehouse-
keeping’parameterssuchasthecountratesinseveralonboardradiationdetectors,
instrumentstatuscodes,dataownership,andorbitphase(utilityspiselectscw).We
employtheINTEGRALRadiationEnvironmentMonitor(IREM;Hajdasetal.2003),
theSPIplasticscintillatoranticoincidencecounter(PSAC),andtherateofsaturating
evferredentstoinasSPI’sGEDSAGeTdetectorsrates)to(fromexcludeeventssolardepositing-flareev>ents8MeandVinotheradetector;erraticbackgrhereafteroundre-
increases.WeexcludedatawheretheGEDSATratesfallbelow50counts/sandexceed
105counts/storejectanomalies.Regularperigeebackgroundincreasesareaddition-
allyeliminatedthrougha0.05–0.95windowonorbitalphaseinordertoexcludeorbit
phasesneartheEarth’sradiationbelts.Fromtheselectedevents,spectraareaccu-
mulatedpereachdetectorandpointing,andtogetherwithdeadtimeandpointing
informationassembledintotheanalysisdatabase.
Instudies,weestablishdatabasesbothforSE(detectorIDs0–18)andME(detec-
torIDs19–84).Forγ-raylinestudies,e.g.,26Aland60Fe,theselecteddataarein
spectralrangeswith0.5keVor1keVbinsize(dependingonthescientificobjectives)
in∼20−30keVbandsaroundeachoftheγ-raylinesforscienceanalysis,andalso
includedataadjacenttothelinesfordeterminationofinstrumentalbackground(see
belostudies.w).ForTheseprexample,ocessesin26Alwillbestudiesrepeated(seeinChapterthe3),followewingusethechaptersSEdataforinthethedifspectralferent
rangemodellingof1785from–1826thekeVadjacentwith0.5kecontinuum,Vbinwsizeeformaketheuselineof1785analysis;–1802forketheVplusbackgr1815ound-
1826keV,whichhavebeencombinedintooneenergybin.Atpresent,SPIhasaccu-
mulated4yearsofdata.WewilluseallthesedatafromtheINTEGRALcoreprogram
(focusingonthesurveyoftheinnerGalaxy),datawithrights(PIownerships)residing
thewithinselectedtheMPEdatagr(foroup,aandsamplepublicof26Aldataavstudies)ailablehasinbeenMarchsho2007wn.inTheFigureexposure2.5.mapof

modellingBackground2.3SPIure2.3)spectraofarespacedominatedplatformsbyunderthegoingintensebackgrcosmic-rayound(CR)radiationbombardment.characteristicMuch(seeofthisFig-

47

2DataanalysesofINTEGRAL/SPI90135180

atMPE45

90

0

-90

60

30

315270225-30

-60

0

1.49 522.1.98e+034.37e+037.70e+031.20e+041.72e+04

Figure2.5:Exposuremapoftheallskyforthe26Aldataselectedfrom4-yearSPI
observations(fromorbits43–525).Thedatabasecoversthewholeskywith23701
pointings,equivalenttoatotaldeadtime-correctedexposuretimeof47Ms.The
observationsconcentrateontheGalacticPlane,speciallyintheGalacticcenter,the
innerGalaxyregion,CasA,Cygnus,andCarina-Velaregions.

48

2.3oundBackgrmodellingFigure2.6:Totalratesofsingleeventsfor19Gedetectors(SgleEvtsTotRate)versus
totalratesofGEDSATrates(GeSatTot).

oughtradiationtoisfolloprwompt,thatoftheresultingincidentdirectlyCRfrflux.omCROtherimpacts,componentswhosearisevfrariationomwithradioactivtimee
isotopesproducedbytheCRimpacts,whosedecaylifetimesarelongcomparedto
thetopeswillcoincidenceincreasewindoinwofabundancetheuntildetectorCRpr-anticoincidenceoductionbalanceselectrβonics.-decay.ARadioactivtimeeseriesiso-
oftheirγ-rayemissionwillbetheconvolutionofthetimebehaviorofthepr−tompt/τ0
CRsourcewithanexponentialdecreasefromdecay(anexponentialfunctione,
whereτ0istheisotope’sdecaytime.Localradioactivityinthespacecraftandinstru-
mentsthemselvesthuswillgeneratebothbroadcontinuumbackgroundemissionand
narrowgamma-raylinesfromlong-livedradioactiveisotopes.Varyingwithenergy,
backgroundcomponentsmayexhibitcomplextimevariabilityduetotheirorigins
frommorethanonephysicalsource.
inWeeachderivedetectormodelsfromforthecomparisonsbackgrtooundpresumedcontribution‘tracers’perofenerbackgrgybinound;andforpointingthese,
w2006ecuse,andseeindependentHalloinetal.contemporaneous2007inpreparation,measurementsalsoaboarbackgrdoundINTEGRALmodel(Diehlstudiesetbal.y

49

2DataanalysesofINTEGRAL/SPIatMPESizunetal.2007inpreparation).Comparingdifferentcandidatetracerstoobserved
backgroundvariations,wefindthebesttracer(s)ofbackgroundforeachenergyband.
CandidatetracerswithhighstatisticalprecisionaretheratesmeasuredinSPI’splastic
scintillatoranticoincidencedetector,therateofsaturatedeventsintheBGOshield,the
detector-by-detectorGEDSATrates(fromeventsdepositing>8MeVinaGedetector,
reflectingthevariabilityandeffectsofcosmicraysondetectors,alsoseeFigure2.6),
andGe-detectorratesintegratedoverratherwideenergybands(forbetterstatistical
precision).Acorrelationbetweentotalratesofsingleeventsfor19Gedetectorsandtotalrates
ofGEDSATratesisshowninFigure2.6.SotheGEDSATratesaregoodtracersto
variabilityofbackgrounddominatedinSPIrawspectra.Forouranalysis,wemake
useoftheGEDSATratestracingthepromptCRactivationtotheγ-raylines.
Eachsuchtracerisregardedasa‘template’fortimevariabilityofbackground.By
normalizingsuchatemplatetothesetofcountsperdetectorspectrum(i.e.perpoint-
ing),abackgroundmodelcanbeconstructed,whichisapplicabletotheanalysisdata
setofspectra.Weconstructourbackgroundmodelfromtwocomponents,onederived
fromadjacentenergybands,andanadditionalcomponentmodellinginstrumentalline
contributions.Thus,inafirststep,thedetector-by-detectorcountratesinthecontinuumbandsare
fittedbytheGEDSATtimeseries(atracerfortimevariabilityofbackgroundinthe
adjacentcontinuumbands,Figure2.6)toconstructan‘adjacent-energies’background
template.Otherbackgroundtemplateswhichareusedformodellinginstrumentallinecontri-
butionscomefromtheradioactivebackgroundlinecomponentsintheenergybands
intheindividualscienceanalysis.Generally,theGEDSATtraceristakenasanim-
portantbackgroundcomponenttracerfortheinstrumentallinefeatureswhichblend
intothecelestialγ-raylinesignals(seeabove).Thisadditionalbackgroundcomponent
tracerisenoughforthe26Alstudies(seeChapters3&4).Forstudiesofthe60Felines,
otherradioactivebackgroundlinecomponenttracersmaybeenneeded,whichwillbe
.5ChapterindiscussedThen,thesetofdetector-by-detectorspectraperpointingin0.5or1keVbinscover-
ingthe∼20keVintervalsaroundandincludingtheγ-raylinesarefittedtothesum
oftheadjacentcontinuumtemplate,plusaGEDSATtemplate,or/andplusadditional
radioactivitytemplatesforspecialimportantradioactivebackgroundlines,tocapture
anyadditionalpromptbackground.
Theadjacentcontinuumtemplateistakenastheprimebackgroundtracer.Andthen
weremovetheinformationofthisprimetracerfromothercandidatetracersthrough
orthogonalization.Sothatweorthogonalizethedifferentbackgroundmodelcompo-
nentsforanimprovedconvergenceofthefitting(utilityspiorthomodel,alsosee§2.2of
Diehletal.2006c).Thisstepensuresthateachhigher-orderbackgroundcomponent
carriesnewandindependentinformation.
Thuswegenerateabackgroundmodelforouractualsetofspectra,whichiscon-

50

2.3oundBackgrmodellingFigure2.7:Detectorratiosaroundthe26Alenergyband(1800–1820keV):relative
intensitiesof19detectors(seeFigure2.4)forthecontinuum(left)andforthe1810
(right).featurelineVke

strainedgamma-rainyitsevventsariabilityaresmall,bythetheirvvariousariabilitybackgrisoundintrinsicallytracers.differentContributionsduetoofthecelestialcoded-
maskexpectedonlymodulationtoafoffecttheglobalbackgrnoroundmalization,modelbutnotsupplementedthevbyariabilitydithering,withtime.andthusis

Asweknow,thecountsvarywithdetectors.Thisdetectorratioalsochangeswith
energiesandspectralfeatures.Figure2.7displaysthedetectorratiosaroundthe26Al
energyband.Theratiointhecontinuumband(1800–1820keV)isquitedifferent
fromtheratiointhelinefeature(1807–181326keV).This60behaviorshouldbetakeninto
accountinthebackgroundmodellingintheAlandFelinestudies,speciallyforthe
high-resolutionspectroscopyofstrongγ-raylines.

51

2Dataanalysesof1.21.00.8

Intensity0.6

0.4

0.2

0.0

INTEGRAL/SPIatMPEττ=0.0, f=0.1, fll=1.00, =1.02, ΔΔx=−0.00x=−0.01
ττ=0.3, f=0.2, fll=1.14, =1.08, ΔΔx=−0.01x=−0.01
ττ=0.4, f=0.5, fll=1.21, =1.28, ΔΔx=−0.02x=−0.02

−1−2−3E − E0

0

1

Figure2.8:Lineshapefunction(Gaussianplusatruncatedexponential)usedtomodel
thedegradationbehaviorofSPI’sinstrumentallinesaschangesofτvalues(from
Kretschmer2007),whichkeepstheshapeoftheunderlyingGaussianconstantata
fullwidthathalfmaximumof1energyunitandapeakintensityofoneintensity
unit.Linewidth(f1)increasesandtheshift(Δx)movestowardslowerenergies
withincreasingdegradationparameterτ.

Spectral2.4SPIofresponse

TheSPIinstrumentalresponseisbasedonextensiveMonteCarlosimulationsand
parameterization[Sturneretal.,2003].ThishasbeentestedontheCrabin-flightcal-
ibrationobservationsandshowntobereliabletobetterthan20%inabsolutefluxat
thecurrentstateoftheanalysis[Attiéetal.,2003].
Sinceouraimistoderivethehigh-resolutionspectroscopyofγ-raylines,e.g.,from
26Aland60Fe,weneedtostudypropertiesofthespectralresponseofSPIGedetectors,
whichmainlyrelatetotheinstrumentalbackgroundlinesinSPIrawdataspectra
(Weidenspointneretal.2003,seeFigure2.3).Anidealsemiconductordetectorwould
produceinstrumentallineswithGaussianshapes.However,theGeDperformance
degradeswiththedamagescreatedwithinthecrystalbyincidentradiation(mainly
protonsandneutrons).Inspace,thecosmicparticlefluxishighenoughtoproduce
substantialdegradationoftheGeDsoverafewmonths.Thisenergyresolutionof
theGeDsthusevolveswithtime(seeFigure2.2)andacontinuousmonitoringofthe

52

2.5Extractingspectraomfrthesky:diffuseemissionandpointcessourresolutionisrequiredtoanalyzeproperlytheSPIdata.
TheinstrumentallineshapevariationsduetodegradationoftheGeDsaredeter-
minedduringthemissionbyfittingaspecificspectralresponsetoinstrumentallines.
ThisspectralresponsefunctionconsistsofaGaussianshapewhichcharacterizeseach
particulardetector’sintrinsicspectralresolution,andaone-sideexponentialfunction
extendingfromthepeakoftheGaussiantowardslowerenergies(Kretschmer2007,
seeFigure2.8),whichcharacterizesthepulseheightlossesduetodetectordegrada-
tionthroughthedegradationwidthτ[Kretschmer,2007]:
si=s0∞(√1e−Ei+2σE2−E0∙1e−τE)dE,(2.1)
τσπ20wheresiistheamplitudeperenergybinEi,E0isthephotopeaklineenergy,andσ
theinstrumentalresolutionoftheGedetectors.Byfittingmanycalibrationlinesover
theSPIenergyrange,thetime-variabledegradationcanbedetermined.Figure2.9
presentsthetimeevolutionoffullwidthathalfmaximumofSPIinstrumentalback-
groundlinesinfourdifferentenergyrangescausedbythetime-variabledegradation
effect.Wewilltakethistime-variablespectralresponsedirectlyintoaccountinthe
instrumentalresponseinourmodelfittingandspectradetermination(nextsection).
Degradationbelongstothelonglifetimevariationofspectralresponse(months).
GainvariationswiththeGeDtemperatureinashorttimescales(<1orbit)have
beenalsofound[Lang,2006].InoneINTEGRALorbit,thechanges(small,∼0.2%)
ofEpeak(instrumentalbackgroundlines)correlatetothedetectortemperature.This
temperaturecorrectionhasnotyetbeenappliedtoenergycalibrationofSPI[Lang,
2006],andwouldimprovespectralresolution.

2.5Extractingspectrafromthesky:diffuseemissionand
sourcespoint

Weobtaintheskyinformationbyusingamodel-fittingprogramwhichfitsthedatato
alinearcombinationofinputastronomicalskymapsandpointsources,togetherwith
backgroundcomponents.Thefitisperformedbythemaximum-likelihoodfunction
basedonPoissonionstatisticswithoneparameterpercomponentandenergyrange
[Strong,2003,Strongetal.,2003].Hencethetime-dependenceofthebackgroundis
explicitlydeterminedfromthedatathemselvesontheassumptionofconstantdetector
ratios.Theearlyimplementationofsuchamodelfitting(spidiffit)hasbeenapplied
fordiffusecontinuumstudies[Strongetal.,2003],and26Allinestudies[Diehletal.,
2003b]fromtheinnerGalaxyregionusingtheearlySPIdata.
Inthepresentupdatedprogramspimodfit(formoredetailsofmethods,seeStrongetal.
2005),wesimilarlycombinebackgroundmodelswithspatialmodelsforskyemission
tofitourdata,allowingformorepowerfuladjustmentsoffitparametersforback-
groundandskyintensities.ThespimodfitoutputFITSfilescontain,perenergybin,

53

2Dataanalyses4.0

3.5

FWHM [keV]3.0

2.5

of2.00

INTEGRAL/SPI100

atMPE200time [rev]

300

400

Figure2.9:TimevariationsoffullwidthathalfmaximumofSPIinstrumentalback-
gr2007,oundfromlinesinbottomfourtodiftop):ferent415ener–gy460kerangesV,860due–to930keV,degradation1060–(fr1175omkeV,Kretschmer1740–
..Vke1820

54

2.5Extractingspectraomfrthesky:diffuseemissionandpointcessourthefittedparametervalueswithuncertainties,thecovariancematrices,andthefitted
modelcomponents.Thecountsperenergybin,perdetector,andperpointingare
fittedtothebackgroundmodeldescribedin§2.3andtheassumedskymapsofce-
lestialemissionplusasourcecatalogueasconvolvedintothedomainofspectraper
detector,energybin,andpointingthroughthepointingsequenceandtheinstrumental
response:

k1k2k3
De,d,p=∑∑Aej,,dm,,pnβsjIjm,n+∑Aeh,d,p(m,n)βchSh(m,n)+∑∑βbi,tBei,d,p+δe,d,p,(2.2)
m,nj=1h=1ti=1

indiceswherefore,d,ptheareskyindicesdimensionsfordata(galacticspacelongitude,dimensions:enerlatitude);gy,Aisdetectorthe,pointing;instrumentm,nre-
sponsematrix,Iistheintensityperpixelonthesky,S(m,n)istheinputsourcecata-
logue(pointsources).k1isthenumberofinputskymaps;k2isthenumberofpoint
sourcesinthecatalogue;k3isthenumberofbackgroundcomponents.Coefficients
βsfortheskymapintensity(constantintime),βcfortheinputpointsources(time
entdependentbackgroundwhereintensitiesrequired,aree.g.,derivtransiented(seesour§2.3ces);difandferentβb,tnor(timemalizationsdependent)alloforweddifforfer-
eachcameraconfigurationof19/18/17functionaldetectorelementsduetofailureof
frtwomoGetheskydetectors).(diffuseAndtheemission),βamplitudectheβsspectracomprisesofthetheinputpointresultantsourspectraces.δofisthethesignalcount
residueafterthefitting.Generally,agoodfitwillleadtoresidualsbeingstatistically
60distributedaroundzero(seeexamplesinChapters3&5aftermodelfittingsof26Al
data).Feand

55

2

56

Data

analyses

of

INTEGRAL/SPI

at

MPE

326Alemissionandlineshapesinthe
Galaxy

26Alisthefirstcosmicradioactivityeverdetectedwithgamma-raydetectors[Mahoneyetal.,
1982electr,onPrantzoscaptureand(15%),Diehl,with1996a].meanItislifetimeunstableof1.04isotopeMyrto.26Alpositrcanonfirstemissiondecayinto(82%)anorex-to
citedstateof26Mg,whichde-excitesintotheMggroundstatebyemittinggamma-ray
photonswiththecharacteristicenergyof1808.66keV(seeFigure3.1).
Thedetailedstudiesof1809keVlineemissionfromtheGalactic26Alistheoneof
themaindesigngoals26oftheINTEGRALmission.As26describedin§1.3,theCOMP-
TELimagingofthe26AllinefromallskysuggestedAlemissionextendsalongthe
Galacticplane,thenAlnucleosynthesisisacommonGalacticphenomenonrather
thanalignmentslocaltooftheemissionsolarsystemmaxima[withPlüschkeetspiral-aral.,m2001].tangent,Theandstructurecomparisonsofthiswithemission,trac-
26ersdominateof26candidateAlAlnucleosynthesissources,[allChenhaveetal.,pointed1995,toDiehltheetal.,conclusion1995a,thatKnödlsedermassiveetstarsal.,
1999b,Knödlseder,1999].
1809SPIkeisV,ahighwhichisresolutionexpectedspectrtorevometerealmorewithinforenergymationresolutionaboutofthe326keAlVsour(FWHM)cesandat
26theirGalacticrlocationotationthrandoughdynamicsDopplerbroftheoadeningejectedand26Alshiftas(itAlprlineopagatesshapes),intheinducedinterstellarfrom
tantmediumastrarophysicsoundonthe26Alsources.studies,Sobetheyondspectralwhatcouldanalysisbewlearouldnedproonlyvidefrneomwtheandimagingimpor-
oftheCOMPTELall-skysurvey.
isticsInofthis26AlChapteremission,wewillalongtheconcentrateGalacticonPlane.studyingThethemainlargoalge-scaleistospectralmeasurethecharacter26Al-
lineshapesintheinnerGalaxyanddifferentregionsalongtheGalacticplane,and
togloballyprobetheeffectofGalacticrotationonthe26Allinebroadeningandshift
(adetailedstudyoftheinnerGalaxyregionisalsopresentedinKretschmer2007).
The26Allinewidthwouldrelatetotheenvironmentoftheinterstellarmedium.Then
sources:measurementsjustnorofmalthelineinterstellarwidthcouldturbulenceprovide(narrothewinforline),ormationexistenceofISMofarsuperoundno26vAla,
stellarwindbubbles(broadline).
Inspectralstudies,the26AlspectrumfromthemodelfittingofSPIdatacouldde-
pendtraceronmapstheofinput26Alskysourmapce(see§candidates,2.5).Forandatheirconsistenteffectscheck,onthewederivwilledcomparespectrum,differentspe-

57

326AlemissionandlineshapesintheGalaxyFigure3.1:Thedecaychainoftheradioactiveisotope26Al.26Aldecayswithachar-
26theacteristicMggroundlifetimestateof1b.y04Myremittingintoγ-raanylinesexcitedwithstatetheofenerMg,giesofthen1809kede-excitesV(99.7into%)
and2.936keV(0.3%).

58

ciallyonthe26Alflux.

analysisandpreparationData3.1

3.1DataeparationprandanalysisIn26Allinestudies(Chapters3&4),weestablishaSEdatabaseusingthe4-yearSPI
observations,fromtheINTEGRALorbit43–525,whichcoversthewholeskywith
23701pointings,equivalenttoatotaldeadtime-correctedexposuretimeof47Ms(see
theexposuremapinFigure2.4).Theselecteddataincludeadatabaseinthespectral
rangeof1785–1826keVwith0.5keVbinsizeforthelineanalysis,andadatabase
forthebackgroundmodellingfromtheadjacentcontinuum.Arawdataspectrum
aroundthe26Alline(1750–1830keV)ispresentedinFigure3.2.Threeinstrumental
linefeaturesdominatetherawspectrum:1764keV(205Bi),1779keV(28Al),and1809.4
keV,acomplexfeatureincluding22Na(1808.7keV)with56Mn(1810.9keV).
Forthebackgroundmodellingfromtheadjacentcontinuum,wemakeuseof1785
–1802keVplus1815-1826keV,whichhavebeencombinedintooneenergybin.
Aspresentedin§2.3,thedetector-by-detectorcountratesinthethecontinuumband
arefittedbytheGEDSATtimeseriestoconstructan‘adjacent-energies’background
templateforthe26Alline.AndtheGEDSATratescanbeusedtotracethepromptCR
activationtotheinstrumentallines.Thenthesetofdetector-by-detectorspectraper
pointingin0.5binscoveringthe∼10keVintervalsaroundandincludingthe26Alline
(1785–1826keV)arefittedtothesumofadjacentcontinuumtemplate,plusaGEDSAT
template.Thus,wegenerateabackgroundmodelforouractualsetofspectraforthe
analysis.scienceThenthemodelfittingsaremadefortheSEdatabaseof26Alline.Sinceinthesky
regionwestudied,i.e.,−120◦<l<120◦,−30◦<b<30◦,SPIcannotdetectand
resolveanypointsourcesinournarrowenergybandsforγ-raystudies(including
26Aland60Felines),wedonotneedtoinputthepoint-sourcecatalogueforthemodel
fittings.Soonlytwocomponentsareusedtofitthedata:diffuseemissionwithinput
skymapsplusthebackgroundmodel,thenEquation2.2canbesimplifiedasfollows:

kk21De,d,p=∑∑Aje,,md,,pnβsjIjm,n+∑∑βbi,tBei,d,p+δe,d,p,(3.1)
m,nj=1ti=1
wherek1isthenumberofinputskymaps;k2isthenumberofbackgroundcompo-
nents.Weusetheskyintensitydistributionof26Alfrom9-yearCOMPTELobservations
(maximumentropymapasastandardskymodel,Plüschkeetal.2001)asthestandard
skymodelinthemodelfittings.Anddifferent26Alemissiontracermodelsormaps
arealsoappliedinfittings.Wewillcomparetheeffectofthesedifferentskymodels
onthe26Allineshapeandintensityin§3.4.Theresidualsafterthemodelfittingare
showninFigure3.3(residualswiththetime)andFigure3.4(withtheenergies).The
reducedχ2valuesaround1.0confirmthatthepresentbackgroundmodelisadequate.

59

326Alemissionandline6000

5000

4000Counts3000

2000

1000

shapesin1760

theGalaxy18001780Energy (keV)

det=0,18pnt=(88.0,-2.1)

1820

Figure3.2:Rawdataspectraaroundtheenergiesofthe26Allineinone-INTEGRAL-
orbitobservations(3days),representingtheinstrumentallinesandcontin-
uumbackgr205ound.Lineorigins28areradioactivities22excitedbycosmicbombar56d-
(ment,1810.9e.g.,keV)Bifor(the1764keV),compositeAl(feature1779ke(V),1809.4andkeV)Na(1808underlying.7keV)thewith26AllineMn
[Weidenspointneretal.,2003].Fittingtheseinstrumentallines,wetypicallyfind
instrumentalwidthsofthreelines(keV):1764/3.1,1779/3.2,1809/4.2,respectively.
Thementalintensitylineatof1779keVinstrumental.featureat1809keVistypically∼12%oftheinstru-

60

0.30.20.1Residual Counts0.0-0.1-0.2-0.3

3.1DataeparationprandDet=0-18E =1800.2 - 1819.8E =1800.2 - 1819.8

5.0•1031.0•1041.5•1042.0•104
Pointing No.

analysisFigure3.3:Residualsversustime−(pointings,1per30min)aftermodelfittingforthe
SEbackgrdatabaseound(inmodelsunitsareofadequate,countsswith).χ2/dResiduals.o.f.=ar0.958ound(zer450315oconfird.o.f.m)inthatmodelour
fittings.

0.40.2Residual Counts0.0-0.2-0.4

Det=0-18MJD=1145.7-2588.3MJD=1684.9-2588.3

181518101805Energy (keV)

Figure3.4:ResidualsofcountsversusenergiesaftermodelfittingfortheSEdatabase
(inunitsofcountss−1keV−1).Reducedχ2changeswithenergies:from0.82
aroundoff-lineenergiesat1800and1820keV,to0.95–1.05inthelineenergy
band.Thesolidlinerepresentsresidualsforthewholedata(e.g.,JulianDate1146
–2588),andthedashedlineforthelate-timedata(JulianDate1685–2588).

61

326AlemissionandlineshapesGLAT (deg)1050-5-10

intheGalaxy020GLON (deg)

340

Figure3.5:TheCOMPTELmaximumentropy26Almap[Plüschkeetal.,2001](here-
afterCOMPTELMaxEnt)fortheinnerGalaxyastheinputskymodelinthemodel
fitting.

3.226AlemissionandlineshapesintheinnerGalaxy
Thelarge-scalespectralfeatureof26AlemissionintheinnerGalaxy(e.g.,−30◦<
l<30◦,−10◦<b<10◦)isimportanttostudytherecentstar-formationactivityin
theGalaxyandthepropertiesoftheinterstellarmediumnearthe26Alsources.By
fittingoursetofobservationsat0.5keVwideenergybinningwiththeskyintensity
distributionof26Al(maximumentropyimagebyCOMPTEL,seeFigure3.5)together
withthebackgroundmodel,weobtainthespectrumfor26Alemissionfromtheinner
GalaxyshowninFigure3.6.The26Allineisdetectedat∼30σsignificance.Theline
componentabovethelinearcomponentofourspectralfitdeterminestheintensityof
observed26Al.TheupperoneinFigure3.6showtheGaussianfitofthespectrum,the
obtainedwidth3.09±0.15isconsistentwiththeinstrumentalwidth∼3.1keVnear
1.8MeV.Thisfitimpliesthe26AllineintheinnerGalaxywouldbeintrinsicallynarrow.
Wealsofittheadjacentcontinuumwiththeflatindex,andthiscontinuumisnearzero,
whichisexpectedfromtheadjacentbackgroundmodelling.Andthediffusegamma-
raybackgroundcontinuumintheinnerGalaxyis∼2×10−6phcm−2s−1rad−1keV−1
intheenergybandof1–2MeV[Strongetal.,1999],whichisalsoconsistentwiththe
fitting.continuumThederived26Alintensityhasbeennormalizedtotheregionof−30◦<l<30◦,−10◦<
b<10◦(i.e.thefluxintheinnerradian).Thenthetotal26Algamma-rayfluxofthe
GaussianfitasdeterminedfortheinnerGalaxyregionis(2.96±0.19)×10−4phcm−2s−1rad−1.
ThisisconsistentwiththeCOMPTELimaging-analysisvalueof(2.8±0.4)×10−4ph
cm−2s−1rad−1[Oberlack,1997,Plüschke,2001]andthevaluefromthepreviousanal-
ysisoftheSPIdata,(3.3±0.4)×10−4phcm−2s−1rad−1[Diehletal.,2006c].
Duetothehighsignificance(∼30σ)ofthedetectionof26Allineinthewholeinner
Galaxy,wealsotrytoderivetheastrophysicalwidthbyfittingtheinstrumental-line-

62

3.22.0-1]E=1808.98 FWHM=3.08 (±(±0.06)0.15)
0.19)±(I=2.96 keV-11.5rad-1s-21.0 ph cm-40.5Intensity [100.0

26Alemission

andline1815181018051800Energy [keV] 2.00.08)±(E=1809.07 -1]FWHM=0.46 (±0.32)
I=2.97 0.17)±(keV-11.5rad-1s-21.0 ph cm-40.5Intensity [100.0

1815181018051800Energy [keV]

shapesin1820

1820

theinnerGalaxyFigure3.6:SpectrumderivedfromskymodelfittingusingtheCOMPTEL26AlMaxi-
mumEntropyimage.Theupperfigureshowsthe26AlspectrumwithaGaussian
fit,obtainingaGaussianwidthof∼3.08keVwhichisconsistentwiththeinstru-
mentallinewidtharound1.8MeV.Thelowerfigureshowsthespectrumwithafit
usingtheshapeoftheinstrumentalresolutionasitresultsfromcosmic-raydegra-
dationandannealingsduringthetimeofourmeasurement(seeFigure2.9),con-
volvedwithaGaussian,toobtaintheintrinsic26Allinewidth(∼0.46keV).Both
twofitsdemonstratethatthelineisintrinsicallynarrow.Derivedfluxesarecon-
sistentwitheachother.Fluxesarequotedinunitsof10−4phcm−2s−1rad−1.The
26Allinecentroidenergyisdeterminedat1808.98±0.06keVfromtheGaussian
fitand1809.07±0.08keVfromtheinstrumental-response-convolvedfit,which
isalittlehigherthanthelaboratoryvalueforthe26Allineof1808.65±0.07keV
[FirestoneandEkstrüm,2004].

63

326AlemissionandlineshapesGLAT (deg)1050-5-10

in20

theGalaxy0GLON (deg)

-20

Figure3.7:TheCOMPTELMREM26Almap[Plüschkeetal.,2001]fortheinnerGalaxy
astheinputskymodelinthemodelfitting.

withshapethe26responseAlspectrum.functionInwithfittingtimeduenon-analyticaltodegradationspectral(e.g.,shapestoFigureour2.9)spectra,convolvsuched
asourtime-integratedinstrumental-responselineshapeconvolvedwithintrinsically-
broadened26Alemission,gradient-drivenfitalgorithmsareinadequate,inparticular
whenweaimtodetermineaparameterliketheintrinsic26Alwidth,whichiscon-
venerolvedgybin.withThetheprinstrumentalobabilitylinedistributionprofileinbeforethiscasebeingisverycomparedasymmetric,toourbutfluxwvealueswishperto
performquantitativestatisticalanalysisofourfindings.Thisisnotpossiblewithtools
thatinherentlyassumesymmetricandsmoothprobabilitydistributionsforparameter
fluctuations.WemakeuseoftheMonteCarloMarkovChain(MCMC)method,pre-
processedbysimulatedannealing(alsoseediscussionsinKretschmer2007,Diehletal.
).c2006Wefittheeffectiv26eaccumulatedinstrumentallineshape,convolvedwithaGaussian
forthecelestialAllinebroadening,toourspectrum.Thefittedparametersarethe
linecentroid,theintrinsicwidthofcelestial26Al,theintensityoftheline,andtwo
parametersfortheunderlyingcontinuum.Theintrinsiclinewidthalsoappearstobe
ratherconsistentsmall,with0.46the±0.32earlykeVresults(thebybeloSPIw[oneDiehlinetal.Figure,20043.,6).2006Thec].derivAndedthewidthlinevcentralueoidis
energyand26AllinefluxarestillconsistentwiththeresultsfromtheGaussianfit.
The26AllinecentroidenergyfromtheinnerGalaxy(|l|<30◦,|b|<10◦)isde-
terminedat1808.98±0.06keVfromtheGaussianfitand1809.07±0.08keVfrom
theinstrumental-response-convolvedfit.Evenafterconsideringthesystemerrorof
0.1keV,themeasuredcentroidenergyisalittlehigherthanthelaboratoryvaluefor
the26Allineof1808.65±0.07keV[FirestoneandEkstrüm,2004].Theeffectofthis
blueshiftwilldiscussedinmoredetailsinfollowingsections.
AccordingtoEq.3.1,ourmodelfittingdependsontheinputskymodel.26Fora
test,wealsoderivethespectrumafterthefitusingtheCOMPTELMREMAlmap

64

3.22.0-1]E=1809.03 FWHM=3.06 (±(±0.06)0.15)
0.19)±(I=2.98 keV-11.5rad-1s-21.0 ph cm-40.5Intensity [100.0

26Alemission

andline1815181018051800Energy [keV] 2.00.09)±(E=1809.12 -1]FWHM=0.47 (±0.33)
0.17)±(I=3.01 keV1.5-1rad-1s-21.0 ph cm-40.5Intensity [100.0

1815181018051800Energy [keV]

shapesin1820

1820

theinnerGalaxyFigure3.8:SpectrumderivedfromskymodelfittingusingtheCOMPTEL26AlMREM
image.SimilartoFigure3.6,theupperfigureshowsthe26Alspectrumwith
aGaussianfit,andthebelowoneshowsthespectrumwithafittoderivethe
intrinsicwidth.Theresultsoftwofitsareconsistentwitheachother,narrow26Al
lines,energycentroids,asamelevelof26Allinefluxes,andalsoconsistentwith
theresultsinFigure3.6.

65

326AlemissionandlineshapesintheGalaxyFigure3.9:Reported26AllinefluxvaluesfromtheinnerGalaxyfortendifferentex-
periments:(1)HEAO-C[Mahoneyetal.,1984];(2)SMM[Shareetal.,1985];(3)
MPEballoon[vonBallmoosetal.,1987];(4)Bell-Sandia[MacCallumetal.,1987];
(5)GRIS[Teegardenetal.,1991];(6)HEXAGONE[Durouchouxetal.,1993];(7)
COMPTEL[Plüschke,2001];(8)OSSE[Harrisetal.,1997];(9)RHESSI[Smith,
2004b];(10)SPI(thiswork).

66

3.3ophysicalAstroriginsof26AllinewidthFigure3.10:Reportsontheconstraintsofthe26Allinewidthfromdifferentexperi-
ments(1σerrorbars):HEAO-C[Mahoneyetal.,1984],GRIS[Nayaetal.,1996],
RHESSI[Smith,2003]andSPI.

(Figure3.7).The26AlspectrumisshowninFigure3.8.Sincethesetwomapsare
quiteamongthesimilardifwithferenteachtracerother,mapstheof26Alconsistentsourcesresultsfortheareinnerexpected.GalaxyFurtherwillbecomparisonpresented
anddiscussedin§3.3.
Sincethefirstdetectionof26AlemissionfromtheinnerGalaxybyHEAO-C[Mahoneyetal.,
26v1982alues],frupomtonothew,innerthereareGalaxyten.WdifeferentcollectedtheseexperimentsvalueswhichwhichhavareereportedplottedintheFigureAl3.flux9.
TheresultsSPIbyresultCOMPTELhas[obtainedOberlackthe,most1997,significantPlüschke,2001report,]andwhichOSSEis[Harrisconsistentetal.,with1997the],
26andinneralittleGalaxylowiser∼3than×10the−4phrecentcm−2results−1byrad−1.RHESSI[Smith,2004b].TheAlfluxinthe
26ureSe3v.10eral).TheexperimentsresultshaveobtainedalsobyreportedHEAO-Cthe[widthMahonevyaluesetal.of,the1984],AlRHESSIline(see[SmithFig-,
[Na2003y]aetandal.,SPI1996]suggestedreportedtheaveryintrinsicbroadnarrolinewwithlineafeaturewidthfor∼5.4thekeinnerV,whichGalaxyis.incon-GRIS
sistentwiththeothermeasurements,andisclearlyruledoutbyourpresentresult.

3.3Astrophysicaloriginsof26Allinewidth

AlineDopplerbrvoadeningelocitiesofof∼∼0.605kekmVsfr−1.omAstrourophysicalline-shapeoriginsconstraintsof26Allinecorrespondswidthtomaythercomemal
fromtwoeffects:randommotionsintheinterstellarmedium[Chenetal.,1997]and
Galacticdifferentialrotation[Kretschmeretal.,2003].

67

326AlemissionandlineshapesintheGalaxyThetypicalturbulentvelocitiesinISMarearoundtensofkms−1.Ifejected26Alis
mixedwiththenormalISMinthetimescaleofMyrintoanequilibriumISMstate,
the26Allineisexpectedtobeintrinsicallynarrow.Inwindbubblesnearbymassive
starsandsupernovaremnants,ISMvelocitieswouldreachahighvaluerange.Aline
broadeningeffectcouldbeobservedifalargefractionofISMresidesinconnected
bubblesorotherunusualcircumstances[Chenetal.,1997].
HowmuchlinebroadeningfromISMisplausible?Significantbroadeningisex-
pectedfromtheapparentexistenceofmajorinterstellarcavitiesintheregionsof
massive-starclusters[Oey,1996],andfromtheejectionkinematicsofthesepresumed
26Alsources,i.e.windsandsupernovae.WRwindvelocitiesare1200kms-1[VinkanddeKoter,
2005]orhigher(Plüschke2001),andmodelsforejectionof26Albycore-collapsesuper-
novaepredictvelocitiesinasimilarrange[HerantandWoosley,1994].Heresources
maynotyethaveobtainedalong-termequilibriumwiththeirsurroundingISM,and
26Almaythusbepreferentiallydecayingfromitsinitialandfastphase,ratherthan
alreadybeingsloweddowntonormalISMvelocitiesinthe10kms−1range.Chenet
al.(1997)alsosuggestedthatahighfractionof26Albringdepositedintohigh-speed
grainsnearthe26Alsourcescouldproducesignificantlybroadlinefeatures.Thebroad-
eningcouldalsobearesultofre-accelerationofdustgrainsbyinterstellarshocksin
theneighbourhoodofthe26Alsource,allowingthemtomaintainahighvelocityover
the26Aldecaytimescale[SturnerandNaya,1999].Therefore,itwillbeinterestingto
testwithSPIonINTEGRALwhetherwecanobservesucheffectsforlocalizedregions
26emission.AlofIfourconstraintsonlarge-scaleintegratedbroadeningofthe26Allinefromtheinner
Galaxycanbeinterpretedintermsofinterstellar-mediumcharacteristics,theintrinsic
widthfromISMturbulencewouldbe∼0.5keV,evenconsideringa2σupperlimitof
1.2keV.Thiscorrespondsto150kms−1fora2σlimitonISMvelocities,wellwithin
theacceptablerange.Thereforeweconcludethat,withinuncertainties,theaverage
velocitiesofdecaying26AlintheGalaxyareprobablynotinexcessoftypicalvalues
fortheISMnearmassivestars.
IntheinnerGalaxy,GalacticdifferentialrotationalonecanleadtoDopplershifts
andlinebroadening.Kretschmeretal.(2003)havesimulatedthe26Allineshape
diagnosticsintheinnerGalaxyduetotheGalacticrotationeffect.Theyadopteda
three-dimensionalmodelforthespacedensityoffreeelectronsasourparentdistri-
butionfor26AlsourcesintheGalaxyderivedbyTaylor&Cordes(1993)frompulsar
dispersionmeasureobservations.TheDopplershiftsduetoGalacticrotationcanthen
bedeterminedfromtheGalacticrotationcurve.Insimulations,theyusedtheresults
obtainedbyOllingandMerrifield[2000]fromfittingradialvelocitymeasurements
withafive-componentmassmodeloftheGalaxy,consistingofastellarbulge,astel-
lardisc,twogasdisks(HI,H2)andadark-matterhalo.SuperimposedonGalactic
rotationisthemotionoffreshlysynthesizedradioactivematerialduetotheparental
supernovaexplosionortheejectingWolf-Rayetstarwindanditsslowed-downmotion
intheISMbeforedecay,∼106yr.Finallytheirresultspronouncedthelinecentroid

68

3.4Comparisonoftheentdiffertracermapsof26AlcessourintheGalaxyshiftsof∼0.25keVtowardslongitudes±30◦[Kretschmeretal.,2003];apossibleline
broadeningofupto1keVhadbeenestimatedifintegratedoverthisinnerregionof
theGalaxy.Arealmeasurementwouldyieldalinewidththatislowerthanthewidth
obtainedfortheentireinnerGalaxy(Kretschmeretal.2003).Ourpresentline-shape
constraintsarealsoconsistentwithGalactic-rotationeffects.
Sowebelievethatthemeasuredlinewidthof26AlfromtheinnerGalaxyiscon-
sistentwithGalacticrotationandmodestinterstellar-mediumturbulencearoundthe
sourcesof26Al.WethusconfirmearlierresultsobtainedbyHEAO-C(Smith2003),
RHESSI(Mahoneyetal.1982),andSPIonINTEGRAL(Diehletal.2006c).Thisis
reassuringthatneitherthesuppressionofdecelerationof26Albeforedecayinlarge,
kpc-sizedcavitiesarounditssourcesnorotherexoticexplanations[Chenetal.,1997]
arerequiredtoaccountforalargebroadeningofthe26Alline.

3.4Comparisonofthedifferenttracermapsof26Alsourcesin
GalaxytheTheall-skyimaginginthe1.8MeVgamma-raylinefrom2626Alobtainedbythe9-year
alongCOMPTELtheplaneobserofvtheations[GalaxyPlüschke.Theetal.,irregular2001]shostructurewsthatoftheAlemission,emissionextendsalignmentsall
ofemissionmaximum26withspiral-armtangentpointstotheconclusionthat26mas-
sivestarsdominateAlnucleosynthesis.ComparisonsoftracersofcandidateAl
sourceshavealsobeenagoodprobeoftheoriginof26Alproduction[Chenetal.,1995,
Knödlsederetal.,1999b].Knödlseder[1999]revealedaclosecorrelationbetween53
GHzmicrowavefree-freeemissionand26Alγ-raylineemission.Whilemicrowave
starsfree-free(>20Memission)arearisesatthefromoriginionizedofGalacticinterstellar26Al.medium,theyarguedthatmassive
Inthissection,wewillcomparethedifferenttracersof26AlsourcesintheGalaxy.
AllthetracermapsexceptthemaximumentropyandMREM26AlmapsbyCOMPTEL
arepresentedinFigure3.11.Briefdescriptionsonthesetracermapsareshownhere
ed:wfolloasAnexponentialdiskmodel:scaleradius4kpc,scaleheight180pc(hereafterExp
Disk);tudes,Anartificiallyexponential-likecreatedinlatitudeshomogenouswithdiskscalemodelheight:200constantpc(hereafterbrightnessHomoalongDisk);longi-
TheHIsurveymap(21cm,DickeyandLockman1990)fromtheLeiden/Dwingeloo
y;evSur5-yTheearCOdatasurofvtheeymapColumbia(Dame1.2etmal.1987telescopes,StrinongNeetwal.York1988),CityfromandaonCerrcompositeoTololoofthein
Chile,probingthedistributionofmolecularcloudsintheGalaxy;
TheIRAS12µmsurveymap(Wheelocketal.1991);
Theradio408MHzsurveymap(Haslametal.1995)fromtheParkessurvey;

69

370

26AlemissionandlineshapesintheGalaxyExp DiskHomo DiskGLAT (deg)GLAT (deg)10105500-5-5-10-10200340200340
GLON (deg)GLON (deg)HICOGLAT (deg)GLAT (deg)10105500-5-5-10-10200340200340
GLON (deg)GLON (deg)DIRBE 240um408MHzGLAT (deg)GLAT (deg)10105500-5-5-10-10200340200340
GLON (deg)GLON (deg)IRAS 12umEGRET > 100 MeVGLAT (deg)GLAT (deg)10105500-5-5-10-10200340200340
GLON (deg)GLON (deg)Robin YoungDiskTC93 Arm 150pcGLAT (deg)GLAT (deg)10105500-5-5-10-10200340200340
GLON (deg)GLON (deg)TC93 Arm 300pcNE2001 140pcGLAT (deg)GLAT (deg)10105500-5-5-10-10200340200340
GLON (deg)GLON (deg)NE2001 Arm 330pcGLAT (deg)1050-5-10200340GLON (deg)

Figure3.11:Thedifferenttracermapsof26AlsourcesintheinnerGalaxyusedinour
fittings.model

3.526AlmassTheDIRBE/COBE240µmsurveymapfromthe4-yearobservations[Bennettetal.,
];1996TheEGRET(>100MeV)surveymap(fromHunteretal.1997);
Ayoungdiskmodel(fromRobinetal.2003,scaleheight125pc);
Thefreeelectronspiralarmmodelfrompulsardispersionmeasurements(from
TaylorandCordes1993,scaleheight150pc,hereafterTC93150pc);
Thefreeelectronspiralarmmodel(fromTaylorandCordes1993,scaleheight300
pc,hereafterTC93300pc);
Therecentfreeelectronmodelfrompulsardispersionmeasurements(fromCordesandLazio
2002,nothickdisk,scaleheight140pc,hereafterNE2001140pc);
Therecentfreeelectronspiralarmmodelfrompulsardispersionmeasurements
(fromCordesandLazio2002,scaleheight330pc,hereafterNE2001330pc);
Inourmodelfittings,theoutputresultsshoulddependontheinputskymodels.
Since26Alisdominatedbymassivestarorigin,allselectedtracermodelsmayreflect
thedistributionofthepresentstarformationandmassivestarintheGalaxy.Ifsome
scientificallyimplausibleskydistributionmodels(e.g.,apointsourceatl=20◦,a
bulgemodeloraCOMPTEL26AlmapzeroedintheGalacticPlane,|b|<10◦)are
takeninmodelfittings,thennosignificant26Alsignalcouldbedetected.Wecarried
outtheseparatemodelfittingsusingaskymodelofeachtracermappresentedin
Figure3.11,andobtainedthe26Alspectrawiththesedifferenttracermaps.These
spectrahavebeendisplayedinFigures3.12and3.13.Foracompletecomparison
amongthesetracermapsfor26Alsources,wehavepresentedthe26Allinecentroids,
fluxesandwidthvaluesobtainedfromthedifferenttracersincludingtheCOMPTEL
MaxEntandMREMmapsinTable3.1.
Allthefittedparametersofthe26Alline:linecentroid,fluxandwidth,usingdiffer-
enttracermaps,areconsistentwitheachotherwithinerrorbars.Thelinecentroidis
at∼1809keV,withalittleblueshiftrelativetothelaboratoryvalue.Thelinewidth
valuesare∼0.5keVforallfittings.26AllineintheinnerGalaxyisintrinsicallynar-
row,withtheoriginofGalacticrotationandnormalinterstellarmediumvelocities.
The26AllinefluxintheinnerGalaxyvariesalittlewiththedifferentskymodels:a
littlelowerfluxobtainedusingtheexponentialdiskmodel;forthefree-electrondis-
tributionmodelsincludingbothTC93andNE2001,higherfluxesobtainedusingthe
modelswithlargerscaleheightvalues.Thederived26AlfluxintheinnerGalaxyis
(3.0±0.2)×10−4phcm−2s−1rad−1.Thisdemonstratesthatthederived26Alspectrum
doesnotdependondetailsoftheinputskyaslongasthemaincharacteristicsofthe
26Alskyarerepresentedbythesetracers.

26massAl3.526alloFluxwvforaluesaindirecttheconvAlersionlineasinthedeterGalaxymined,forbecausegeometricalthe3Dspacesourcedistributiondistributionisknomodelswn

71

326AlemissionandlineshapesintheGalaxy
0.2]E=1809.09(±0.09)]E=1809.11(±0.09)
1-V1.5FI=W2.H6M5(=±00..4166)(±0.35)11-VIF=2W.9H8M(=±00..4167)(±0.35)2
1-ek1-ek1.5
dd-1ar1-ar
2-s1.02-s1.0
mc mc
hhpp4- 00.54-0 0.5
1[ y1[ y
tissti
etn0.0net0.0
nnII1800180518101815182018001805181018151820
Energy [keV]Energy [keV]
2.0E=1809.14(±0.09)E=1809.09(±0.09)
1-]VFWHM=0.48(±0.35)31-]VFWHM=0.49(±0.38)4
ekI=3.23(±0.17)ek1.5I=3.08(±0.17)
1-d1.51-d
1-ar1-ar
ss-2m1.02-m1.0
cc hp hp
4-010.54-010.5
yt[ [ yt
iissnet0.0net0.0
nnII18001805Ener1g8y1 [0keV]1815182018001805Ener1g8y1 [0keV]18151820
]2.0E=1809.09(±0.08)]2.0E=1809.10(±0.09)
1-VIF=2W.9H5M(=±00..4166)(±0.33)51-VIF=3W.1H8M(=±00..4197()±0.36)6
ee1-kd1.51-dk1.5
1-ar1-ar
2-s2-s
cm1.0mc1.0
hh4-p 4-p
01[ 0.501[ 0.5
iytytis
snnetnI0.0etnI0.0
1800180518101815182018001805181018151820
Energy [keV]2.0E=1809.07(±0.08)Energy [keV]
1-]VFI=W3.H1M1=(±00..4187)(±0.35)7
1-ekd1.5
a2-1-sr
mc 1.0
hp 4-01[ 0.5
ytisnnIet0.0
18001805Ene1rg8y1 0[keV]18151820
Figure3.12:26Alspectraderivedfromskymodelfittingsusingdifferenttracermaps:
1.ExpDisk;2.HomoDisk;3.HI;4.CO;5.408MHz;6.DIRBE240µm;7.IRAS
m.µ1272

3.526Almass
2.02.0
1-]EF=W18HM09=.00.650(±(0±.00.93)5)1-]FEW=1H8M09=.00.951(±(0±.00.83)6)
VeI=3.10(±0.17)1VeI=3.14(±0.16)2
1-k1.51-k1.5
dd1-ra1-ar
ss22-m1.0-m1.0
c h hc
4- p4-p
01[ 0.501[ 0.5
ytitiy
ssnnetnI0.0etnI0.0
18001805Ener1g8y1 [0keV]1815182018001805Ener1g8y1 [0keV]18151820
1-]FEW=1H8M09=.00.454(±(0±.00.94)0)1-]2.0FEW=1H8M09=.00.652(±(0±.01.03)8)
Ve1.5I=3.07(±0.16)3VeI=3.37(±0.19)4
1-k1-k
dardar1.5
11--2-s1.02-s
mc mc 1.0
hh-4p 4-p
01[0.5[010.5
ytissyti
nneenIt0.0tnI0.0
1800180518101815182018001805181018151820
Energy [keV]Energy [keV]
0.21-]FEW=1H8M09=.01.149(±(0±.00.93)6)51-]2.0FEW=1H8M09=.00.65(0±(0±.00.93)7)6
VeI=2.99(±0.16)VeI=3.41(±0.19)
1-kd1.51-dk
1-ar1-ar1.5
2-s2-s
c m1.0mc 1.0
hh4-p 4-p
001[ y0.51[ y0.5
tistis
nnetnI0.0etnI0.0
1800180518101815182018001805181018151820
Energy [keV]Energy [keV]
Figure3.13:26Alspectraderivedfromskymodelfittingsusingdifferenttracermaps
(continued):1.EGRET(>100MeV);2.Robinyoungdisk;3.TC93150pc;4.
TC93300pc;5.NE2001140pc;6.NE2001330pc.
73

326AlemissionandlineshapesintheGalaxyTable3.1:26Allineresultsfordifferentskydistributionmodels

SkymodelLinecentroid(keV)26Alflux(10−4phcm−2s−1rad−1)Width(keV)
COMPTELCOMPTELMREMMaxEnt1809.121809.07±±0.090.083.012.97±±0.170.170.470.46±±0.330.32
ExpHomoDiskDisk1809.111809.09±±0.090.092.982.65±±0.170.160.460.46±±0.350.35
HI1809.14±0.093.23±0.170.48±0.35
CO1809.09±0.093.08±0.170.49±0.38
408MHz1809.09±0.082.95±0.160.46±0.33
DIRBE240µm1809.10±0.093.18±0.170.49±0.36
IRAS12µm1809.07±0.083.11±0.170.48±0.35
RobinEGRETy(>oung100MediskV)1809.091809.06±±0.080.093.143.10±±0.160.170.510.50±±0.360.35
TCTC9393150300pcpc1809.041809.06±±0.090.103.073.37±±0.160.190.540.52±±0.400.38
NENE20012001330140pcpc1809.061809.11±±0.090.093.412.99±±0.190.160.500.49±±0.370.36

andcanbeintegratedintermsof26Almass.Normalizationisobtainedbystandardiz-
ingthemodelprojectedskybrightness(fortheinnerGalaxy:−30◦<l<30◦,−10◦<
b<10◦),andthebrightnessfitofrespectivemodeltoourdata(e.g.Table3.1).
Fromcomparisonsofalargefamilyofmodelsacrossallwavelengthregions(from
radiotogamma-rays)andplausibleparameterrangesin§3.3,modelsforthegas
diskoftheGalaxy[Robinetal.,2003],ofwarmdustemissionbasedontheIRASmea-
surements[Wheelocketal.,1991]andtheDIRBE/COBEobservations[Bennettetal.,
1996],andfreeelectrondistributioninferredfrompulsardispersionmeasurements
[TaylorandCordes,1993,CordesandLazio,2002],havebeenidentifiedasthemost
plausibleones(seeKnödlsederetal.1999b).
Forconversionto26Almass,wenormalizethetotalintegrated26Alluminosityofthe
geometricalmodelstotheline-of-sightintegratedfluxfittedtoourdata.Duetovery
differentdistributionof26Alinthesemodels,thedispersionofderived26Alamounts
fortheentireGalaxyislargerthantheinner-Galaxyfluxvariations.Weusedthesky
models(e.g.,theexponentialdiskmodel,theRobinyoungdiskmodel,theIRdust
geometricalrepresentation,andthefreeelectronbasedmodels),andtheaverage26Al
massis(2.7±0.6)Mincludingthe26Alfluxandsourcedistributionuncertainties.
WealsousethemassofinterstellargasintheGalaxy(4.95×109M,Robinetal.
2003)andrecently-updatedstandardabundances(logNAl=6.40bynumber,normal-
izedtologNH=12,Asplundetal.2005)todeterminetheinterstellarmassofAl,and

74

3.526Almassthusobtainanisotopicratio26Al/27Alof∼8.1×10−6.This26Al/27Alratioisabout
oneorderofmagnitudelowerthanthevalueinthesolarnebula∼5×10−5.Itseems
plausiblefromcontinuedproductionofAlsincethebirthofthesolarsystems4.6Gyr
ago,andfromseveralotherhintsforspecialenrichmentofthesolarnebulaatbirth
fromanearbynucleosynthesissource[Wasserburgetal.,1998,Goswamietal.,2001],
orfromlocally-produced26Alenrichmentsmadepossiblybythejetoutflowsinthe
earlystageofsolar-nebulaaccretionontothecentralSun[Gounelleetal.,2006].
Radioactive26AlisproducedanddispersedthroughouttheGalaxyatvarioussitesof
nucleosynthesis,anddecayswithameanlifeofτ∼106yrs.Thisprocessisdominated
byejectedmatterfromcorecollapsesupernovaeandtheirprecedingwindsduringthe
Wolf-Rayetstage(see§1.6).Then26Alfluxandderived26Almasscanbeusedto
estimatethestarformationrate(SFR)intheGalaxy.Therelativelylongmeanlifein
comparisontothetimebetweensupernovae(∼30–50yrs)impliesthataverylarge
numberofeventscontributestoasteadystateabundance,whichinturnresultsina
diffuseglowoftheGalaxyinthe1.8MeVgamma-raylinefromthedecayof26Al.
Wecanconverseobservedtotalgamma-raylinefluxfor26Altothecorresponding
supernovarate(SNRate),andconsequentlytotheaveragestarformationrate.The
mainadvantageofthismethodisthelackofextinctioncorrectionsinthegamma-ray
band,sothatwecanseethefulldiskoftheGalaxyandarenotlimitedtoasmall
samplingvolumearoundtheSun,e.g.,methodsbasedonstarcounts.Itisalsoworthy
tonotethataveragingoverafewmillionyearsimpliesalargenumberofevents,which
providesabetterstatisticalensembleincomparisontoothermethodsinwhichthe
tracershaveamuchshorterobservationallifetimeandoftenverysignificantselection
effectsandhighlyuncertainevolutioncorrections(pulsars,radiosupernovaremnants,
historicSNrecord,HIIregions,seeDiehletal.2006a).
The26Almethodrelatestheobservedlinefluxtothesteadystateequilibriummassof
26AlintheGalaxyviaaspatialdistributionmodel.Sincetheobservedmapat1.8MeV
indicatesthatmassivestarsarethedominantcontributorstothisradioactivespecies
intheISM,wegobeyondasimpleaxis-symmetricmodel,andincludespiralstructure
ofourSbc-typeMilkyWay.Thedynamicevolutionoftheinjected26Alshouldalsobe
accountedfor,asstellarwindsandexplosiveoutflowscausethedensitydistribution
tobemoreextendedthanwhatonewouldinferfromtheknownspatialdistribution
ofmassivestars.Theresultingdensitymapissomewhatuncertain;butperhapsthe
largestsourceofpotentialerroristheoverallscaleoftheGalaxy:TheofficialIAUdis-
tanceoftheSuntotheGalacticcenteris8.5kpc,butarecentreviewofmeasurements
doesindicatevaluesassmallas7kpc[Reid,1993].Fluxscalesastheinversesquare
ofdistance,therefore26suchaglobalreductioninsizeoftheGalaxywoulddecreasethe
inferredmassofAlby∼34%.
Theamountof26Alismaintainedinsteadystatebyacore-collapsesupernovarate
viaMeq=SNRate∙τ∙Y,wheretherateismeasuredineventsperyear,τisthemean
lifeof26Al,andYistheIMF-averaged26AlyieldinunitsofMpersupernova.The
yieldinthiscontextmustincludetheexplosiveyieldsfromthesupernovamodelas

75

326AlemissionandlineshapesintheGalaxywellasany26AlejectedintheWolf-Rayetwindphase.Yieldsaremoderatedbythe
steepinitialmassfunction(IMF),ξ∝m−α,inourrelevantmassrange10–120M.
WeusetheMiller-ScaloIMF(ξ∝m−2.7)forthishigher-massrange[MillerandScalo,
1979],supportedbyawiderangeofastronomicalconstraints.Yisobtainedfromthe
high-massinitialmassfunction(IMF)andthenucleosynthesisyieldsofmodels(see
§1.6).Theresulting26Alyieldpermassivestaris(1.4±0.7)×10−4Mbasedonvar-
iouspublishedyieldsasafunctionofprogenitormass(see§1.6).Thecorresponding
supernovarateisSNRate=1.90±0.95eventspercentury(thisdoesnotincludetype
Iasupernovae,whichhavebeenfoundtobenegligiblesourcesof26Al).Theresulting
rangeofonetothreecorecollapsespercenturycoincideswiththerecentvaluesob-
tainedfromonasurveyoflocalO3-B2dwarfs[Reed,2005],extrapolatedtotheGalaxy
aswholewithspatialdistributionmodels,andthestudyoftheluminosityfunctionof
OBassociations[McKeeandWilliams,1997].
Theconversionfromcore-collapsesupernovarate,whichassumesthatallstarsmore
massivethantensolarmassesendtheirlivesassupernovae[Hegeretal.,2003],is
givenbySFR=SNRate∙<m>∙fSN−1,where<m>istheaveragestellarmassina
starformationevent,andisthefractionofallstarsthatbecomesupernovae.Starsare
predominantlyformedinclusterswithaquasi-universalIMF(narrowlydistributed
aboutthecanonicalSalpeterpowerlaw,Kroupa2002,WeidnerandKroupa2005),but
thattheintegratedgalaxialinitialmassfunction(IGIMF)mustbesteeperthanthe
canonicalIMF.AsteeperIGIMFimpliesareducedsupernovafractioningeneral,and
impliesadependenceongalaxymassandthemuchlessestablishedclustermass
distributionfunction.ItisthuspossiblethatthetrueSFRcouldbesignificantlylarger
thantheratederivedfromthecanonicalIMF.WereportaSFRbasedontheconversion
choicemadebyMcKeeandWilliams(1997):SFR=1.96SNRate.Withanaverage
stellarmassof<m>=0.51MandfSN−1=2.6×10−3,wefindSFR=(3.8±1.9)
Myr−1,whichagreeswiththevaluederivedfromtheluminosityfunctionsofOB
associations[McKeeandWilliams,1997].

3.626Allineshapesforthedifferentlongitudesalongthe
PlaneGalactic

The9-yearCOMPTELimagingof26Allineemissionhassuggestedthestructuresordif-
ferentbrightnessalongtheplaneoftheinnerGalaxy(bothforthemaximumentropy
andMREMimages,seeFigures3.5and3.7).COMPTELcouldprovidetheimagede-
26tailsresolutionofAlarinoundthe1.8GalaxyMeV,,butwhichspectralprovideinforusamationchanceistopoorpr.obeSPIthehas26Althelinehighshapesspectralin
thedifferentregionsalongtheGalacticplane.
26Almainlyoriginatesinmassivestars,whichshouldcorrelatetostar-formation
spiral-arcomplexesminthestructures.GalaxySofrom(Figurethe3.14sight).atStarthe-forSun,mationthe26Alcomplexesemissionarewoulddistributedappearin

76

3.626AllineshapesfortheentdifferlongitudesalongtheGalacticPlane26Figure3Galaxy.14:[RusseilDistribution,2003of].theTheAlSuncandidatepositionissourgivces:enbstary-forthelarmationgestarcomplexessymbol.inThethe
four-spiral-armmodelisadoptedtofittheallcomplexes:1:Sagittarius-Carina
arm,picture2:Salsocutum-Cruxsketchesthearm,local1’:armNorfeaturema-Cygnus(longarmdashedand2line),’:thePerseusbararm.orientationThe
andlength(dashed-dot-dotline)fromEnglmaierandGerhard1999,theexpected
departurefromalogarithmicspiralarmobservedfortheSagittarius-Carinaarm
(shortdashedline)andfinallyafeaturecertainlylinkedtothethree-kpcarm
line).(solid

77

326AlemissionandlineshapesintheGalaxyFigure3.15:Theall-skyimagingof26Alemission(1805–1811keV)with3-yearSPI
data(up)andthe26Alintensitydistributionalongthelongitudes(down)and
intensityprofilealongthelatitudes(smallbox).Thisimageconfirmsextended
26Alemissionalongtheinner-Galaxyridge,thebright26Alemissionfromthe
Cygnusregion.Andthe4thquadrantisslightlybrighterthanthe1stone(from
Halloinetal.2007,inpreparation).

brighterinthedirectionofthespiralarms.FortheGalacticcenterregion,starforma-
tionshouldbeactiveinthelastmillionyears,molecularclouds,andmanySNRshave
beendiscoveredintheGalacticcenter[LaRosaetal.,2000].IntheGalacticcenter,a
barstructurelyinginsidehasbeensuggested[EnglmaierandGerhard,1999,Gerhard,
2002].Figure3.14alsoshowsahintofthepresenceofthelocalarm,possiblybelonging
tothenearestpartsoftheSagittarius-Carinaarm.Alargenumberof26star-formation
complexesexistneartheSun,whichwouldcontributetoobservedAlemissionto-
wardthesedirections.Butthenumberandspacedensityof26AlneartheSunare
26prAlobablysourcesloweronlythanfromthatthein26Alspirallinearms.brightnessWedoinnottheknoskyw,thebutdistancedetailedinforstudiesmationoftheof
lineshapes,e.g.linecentroidsandwidths,possiblyprobethegeometricaldistribution
ofGalactic26AlwhichmaylocateindistantregionstowardtheGalacticcenterregion
andlarge-scalespiralarms,orthelocalstar-formationregions(<1kpc).

78

3.626AllineshapesfortheentdifferlongitudesalongtheGalacticPlane3.6.1Differencesinthe1stand4thquadrants
Thethree-yearSPIobservationsalsohaveobtainedanall-sky26Almap(Figure3.15,
Halloinetal.2007,inpreparation),usingSPI’scoded-maskimagingatalimiting
spatialresolutionof∼2.7◦atenergiesof1.8MeV(comparedwithCOMPTEL,∼
3.8◦).Thebasiccharacteristic26Alemissionfeaturesareconfirmed:26Alemission
extendsalongtheGalacticplane,prominentemissionfromtheCygnusregion,and
brightnessasymmetriesfortheleftandrightpartsoftheGalacticplane,thefourth
Galacticquadrantbeingbrighterthanthefirstone(lowerpartinFigure3.15).
Inthissection,wewillprobetheasymmetriesinadifferentway.Wewilldirectly
derivethe26Allinespectraintheregionsofboththe1stand4thGalacticquadrants.
Sothe26Allinefluxes,linecentroidsandwidthsinthetwoquadrantscanbestudies
together.Themodelfittingsarecarriedoutusingseparateskymapscoveringtheeach
skyregion.Ahomogenousdiskmodel(seeFigure3.11,−60◦<l<60◦,−10◦<b<
10◦,scaleheight200pc)isusedheretoavoidthebiasofmapsinmodelfittingsalong
plane.GalactictheWefirstsplitthisskymodelintotwosub-maps(the1stquadrant0◦<l<60◦,and
the4thquadrant−60◦<l<0◦),andapplythemodelfittingusingthesetwomapsto
determinethe26Alspectrafromboththe1stand4thquadrantssimultaneously.Two
outputspectraaredisplayedinFigure3.16,andaGaussianfitisappliedtodetermine
the26Allineshapesandfluxes.Bothspectrashowenergyshiftsrelativetothecentroid
energyof26Allineinthelaboratory,1808.66keV:aminorredshift0.12±0.10inthe1st
quadrantandasignificantblueshift0.54±0.07inthe4thquadrant.Boththespectra
havewidthvaluesnearthevalueoftheinstrumentallinewidth,implyingthat26Al
emissionsinthesetworegionsarealsothenarrowlines.The26Alfluxfromthe4th
quadrantishigherthanthatfromthe1stquadrant,withthefluxratioof∼1.3.
Wecanfurthersplittheskymodelintosmalllongitudedegreebinsalongtheplane.
Figure3.17shows26AlspectrainfoursegmentsalongtheGalacticplane:(1)0◦<l<
30◦,(2)30◦<l<60◦,(3)−30◦<l<0◦,and(4)−60◦<l<−30◦.Theenergyshifts
ofthe26Allinearedetectedintheregions(1)and(3)alongtheplane.Theinnerregion
(−30◦<l<30◦)ismuchbrighterthantwooutsideregions.The1stquadrantbrighter
thanthe4thquadrantisalsoconfirmed:theregion(3)isbrighterthantheregion(1),
withafluxratioof∼1.2,and(4)isbrighterthan(2).
Wealsocarriedoutmodelfittingsusingsixsub-mapswith20degreebinalong
theGalacticlongitude,andobtainedsix26Alspectrafromtheseregions(Figure3.18):
(1)0◦<l<20◦,(2)20◦<l<40◦,(3)40◦<l<60◦,(4)−20◦<l<0◦,(5)
−40◦<l<−20◦,(6)−60◦<l<−40◦.The26Allinespectracanbestilldetected
significantlyfortheinnerGalaxypart(−40◦<l<40◦,>6σforeach20◦binregion).
Forthetwooutsideregions(40◦<l<60◦and−60◦<l<−40◦),26Alisdetectedwith
alowersignificancelevelof<4σ.Thismaybeduetothatthe26Alfluxesintwooutside
regionsareintrinsicallylowandtheexposuretimefortheregionislimitedrelativeto
thecenterregion.ForthefourregionsintheinnerGalaxy,26Allineenergyshiftsare

79

326AlemissionandlineshapesintheGalaxy1.5-1]E=1808.54 FWHM=3.35 (±(±0.10)0.26)
I=2.18 0.22)±(keV-11.0rad-1s-2 ph cm0.5-4Intensity [100.0

1815181018051800Energy [keV] 0.07)±(E=1809.20 -1]FWHM=3.13 (±0.18)
keV1.5I=2.83 (±0.21)
-1rad-1s-21.0 ph cm-40.5Intensity [100.0

1815181018051800Energy [keV]

1820

1820

Figure3.16:26Alspectraintwosegments(upper0◦<l<60◦,anddown−60◦<l<
0◦oftheGalacticPlanederivedfromthemodelfittingusingthetwohomogenous
diskmodels.ThespectraarefittedwithGaussianprofiles.Twospectrashowthe
energyshiftsrelativetothecentroidenergyof26Allineinthelaboratory,1808.66
keV:aminorredshiftinthe1stquadrantandasignificantblueshiftinthe4th
quadrant.Boththespectrahavewidthvaluesnearthevalueoftheinstrumental
linewidth,implyingthat26Alemissionsinthesetworegionsarealsothenarrow
lines.Andthe26Alfluxfromthe4thquadrantalittlehigherthanthatfromthe
1stquadrant,thefluxratiois∼1.3.

80

3.626AllineshapesforthedifferentlongitudesalongtheGalacticPlane
0.21]FEW=1H8M08=.35.715(±(0±.01.23)0)11]1.0FEW=1H8M08=.39.805(±(0±.02.66)4)2
--Ve1.5I=2.71(±0.34)VeI=1.38(±0.38)
1-kd1-kd
aarr1-2-s1.01-2-s0.5
mc mc
hh4-p 0.54-p
01[ 01[ 0.0
yytis0.0sti
netnet
InnI-0.5
5.0-18001805Ener1g8y1 [0keV]1815182018001805Ener1g8y1 [0keV]18151820
5.11-]2.0FEW=1H8M09=.33.401(±(0±.00.92)2)31-]FEW=1H8M09=.32.265(±(0±.02.90)0)4
VeI=3.27(±0.31)VeI=1.72(±0.27)
1-kd1.51-kd1.0
1-ra1-ar
2-s1.02-s
mc mc0.5
hh4-p 0.54-p
[ 0101[ 0.0
ytis0.0ytis
netnet
nI-0.5nI-0.5
1800180518101815182018001805181018151820
Energy [keV]Energy [keV]
Figure3.17:26AlspectrainfoursegmentsalongtheGalacticplane:1.0◦<l<30◦,2.
30◦<l<60◦,3.−30◦<l<0◦,4.−60◦<l<−30◦.
81

326AlemissionandlineshapesintheGalaxy26thealsoregionobserv(2ed.)shoThewsaregionbroad(5)feature,appearswithbrightestainGaussianthewidthGalaxy.ofWhileFWHMthe∼3.87Al±line0.91in
keV.Thispossiblebroadfeatureisquiteinterestingandimportanttotheoriginof26Al
towardthisdirection,i.e.,theSagittarius−1arm.Thebroad26Allinecannotoriginate
frlineomtheshouldbeGalacticduertootationvery(<high100kmsturbulent,seevelocitiesKretschmer(∼300etal.±2002003).kmsThen−1)ofthisbrISMoadin
thestar-formationregionstowardSagittarius(alsoseecomplexstructurestowardthis
directioninFigure3.14).Uncertaintiesarestilllarge,thusthebroad26Allinetoward
the26Sagittariusarmneedfurther◦studies.
AsaAlcheck,emissionwefrapplyomthethemodelfour20fitting-binusingregionseightof10the◦-bininnersub-mapsGalaxyisinthequiteinnersignificant.region
of−40◦<l<40◦,andobtaintheeight26AlspectraalongtheinnerGalacticplane
26iswhichnotaresignificantshown(in<3σFigure)3except.19andthe3.20region.Theof−40◦detections<lof<−the30◦.AlAlineinGaussianeachprregionofile
awithaGaussianfreelineparameterwithanforthelineinstrumentalwidthlinecannotwidthgiv(e∼a3.06goodkefitVartotheound1.8spectra.MeV)Whilecan
fitthederivedspectrarelativelywellwithareducedχ2around1.26Allineemission
intheregionof−40◦<l<−30◦isstillsignificant(∼6σ)andappears◦◦brightest
twalongotheadjacentGalacticsub-mapsplane.(scalesBecausesmallerofthethanwidethefieldfieldvievieww)ofareSPIused(∼to16fit×SPI16),data,if
gamma-raycountsinonepointing(thetimeunitinfitting)canaffectboththeregions
oftwosub-mapssimultaneously.Therefore,themodelfittingwithtoomanysub-
mapscoveringrelativelysmallerregions(<16◦)wouldcontainsomecontaminations
offittingsinadjacentsub-maps:apossibleanti-correlationofthefittedcoefficientsin
ofmodelf-linefittingspectramaforyhaveadjacentdifficultymapsto(seedeterexamplesminethein26AlFigurespectra3.19andwith3.20high).Hence,significancethe
usingseveraladjacentsmallsub-maps(scales<16◦).Tostudy26Alemissioninsmall
intervalswithSPIdatamoreaccurately,otherapproachesarestillneeded(seeone
approachforourtestin§3.6.3).

effectrotationGalacticRevisiting3.6.2Galacticdifferentialrotationcanresultinnotonlythebroadeningof26Allinebutalso
theshiftsofthelinecentroidenergyduetotheDopplereffect[GehrelsandChen,1996,
Kretschmeretal.,2003].AssumingapeoplelooksthroughtheGalaxyabovetheplane
(e.g.thepictureofspiral-armstructuresinFigure3.14),theGalaxyrotatesintheclock
direction,sothatGalacticrotationwouldinducethe26Allinecentroidenergyredshifts
inthe1stquadrantandblueshiftsinthe4thquadrant.TheGalacticrotationcurvecan
bebeendeterminedfromobservationsofvariousobjects(HI,CO,HII,BrandandBlitz
).1993Theangulardistributionof26Alemissionontheskycorrelateswellwithtracersof
ionization,suchHαandfree-freeemission.Adoptingathree-dimensionalmodelfor

82

3.626AllineshapesforthedifferentlongitudesalongtheGalacticPlane
2.02.0
1-]FEW=1H8M0=8.35.410(±(0±.01.64)0)1-]EF=W18H0M9=.31.407(±(0±.01.53)8)
VekI=2.58(±0.44)1Vek1.5I=2.72(±0.44)4
1-d1.51-d
aarr1-2-s1-2-s1.0
0.1mmc hc h0.5
4-p 00.54-p 0
11[ yt yt[0.0
iinets0.0snet-0.5
nnII-0.5-1.0
18001805Ener1g8y1 [0keV]1815182018001805Ener1g8y1 [0keV]18151820
1-]1.5FEW=1H8M0=8.36.387(±(0±.03.59)1)1-]EF=W18H0M9=.537.23(±(0±.01.43)5)
VekI=2.36(±0.73)2Vek2I=3.57(±0.52)5
11--1dar1.0dar
2--s1-2-s
mc 0.5mc 1
hhp 4-010.04-p 0
[ y1[
tis-0.5syti0
nnetnet
I-1.0nI
1-18001805Ener1g8y1 0[keV]1815182018001805181018151820
Energy [keV]
1-]1.5FEW=1H8M0=9.20.747(±(0±.04.60)0)1-]1.5FEW=1H8M08=.23.124(±(0±.04.10)0)
VeI=1.03(±0.35)3VeI=0.86(±0.28)6
1-kd1.01-kd1.0
1-ars1-ars
2-m0.52-m0.5
c hc h
4-p 00.04-p 0
1[1[0.0
yti-0.5 yti
ssnetnet-0.5
nI-1.0nI
0.1-18001805Ener1g8y1 [0keV]1815182018001805Ener1g8y1 [0keV]18151820
Figure3.18:26AlspectrainsixsegmentsalongtheGalacticplane:1.0◦<l<20◦,2.
20◦<l<40◦,3.40◦<l<60◦,4.−20◦<l<0◦,5.−40◦<l<−20◦,6.
−60◦<l<−40◦.Smalllongitudedegreebin(20◦)makesthedetectionsof26Al
notsignificantintheregionsof40◦<l<60◦and−60◦<l<−40◦.
83

326AlemissionandlineshapesintheGalaxy
41-]FEW=1H8M08=.19.280(±(0±.03.38)0)11]FEW=1H8M08=.39.806(±(0±.05.10)0)2
-VeI=2.16(±1.27)Ve3I=3.10(±0.97)
1-k21-k
dd1-ar1-ar2
2-s2-s
mc 0mc 1
hh-4p 4-p
0001[ y-21[ y
tistis-1
tnenet
nI-4In-2
18001805Ener1g8y1 [0keV]1815182018001805Ener1g8y1 0[keV]18151820
E=1808.86(±0.38)
1-]4FWHM=3.06(±0.00)31-]FEW=1H8M09=.34.806(±(0±.07.70)0)4
VekI=3.98(±0.94)Ve4I=2.01(±0.96)
1-d1-kd
a-1r1ra
2-s2-2-s2
mc hmc
hp-40 4- p
001[ 1[ 0
ytisytis
nneetnI-2tnI
2-1800180518101815182018001805181018151820
Energy [keV]Energy [keV]
Figure3.19:26AlspectrainfourregionsoftheinnerGalaxy:1.0◦<l<10◦,2.10◦<
l<20◦,3.−10◦<l<0◦,4.−20◦<l<−10◦.
84

3.626AllineshapesforthedifferentlongitudesalongtheGalacticPlane
31]4FEW=1H8M08=.32.106(±(0±.05.70)0)11]FEW=1H8M08=.39.806(±(0±.04.80)0)
--VeI=2.83(±1.00)VeI=2.35(±0.71)2
1-kd1-kd2
1-ar21-ar
2-s2-s1
mc mc
hh4-p 04- p0
01[ 01[
ytiyti
sne-2sne-1
tnItnI
2-18001805Ener1g8y1 [0keV]1815182018001805Ener1g8y1 [0keV]18151820
41]FEW=1H8M09=.37.206(±(0±.60.20)0)31]FEW=1H8M09=.32.606(±(0±.02.90)0)4
--Vek2I=2.71(±1.05)Vek3I=4.49(±0.80)
11--dardar
1-2s1-2s2
--mc 0mc
1hh4-p 4-p
00[ 1-21[ 0
ytiyti
ssetnnet-1
nI-4nI
2-18001805Ener1g8y1 [0keV]1815182018001805Ener1g8y1 [0keV]18151820
Figure3.20:26AlspectrainfourregionsoftheinnerGalaxy(continued):1.20◦<l<
30◦,2.30◦<l<40◦,3.−30◦<l<−20◦,4.−40◦<l<−30◦.
85

326Alemissionandline1.00.5Al line energy shift [keV]0.0−0.526−1.0

shapesinthe50

Galaxy0Galactic longitude [deg]

−50

Figure3.21:26Allinepositionshiftswithdifferentviewingdirectionsalongtheinner
Galacticplanederivedfromthe1.5-yearSPIdata[Diehletal.,2006a].Galactic
rotationwillshifttheobserved26AllineenergyduetotheDopplereffect,to
appearblueshiftedatnegativelongitudesandredshiftedatpositivelongitudes.
Colourscalesareexpectationsof26Allinepositionsandintensitymodelledfrom
theGalacticrotationcurveandathree-dimensionaldistributionof26Alsources
[Kretschmeretal.,2003].TheyusedthecentralGalacticregion(−10◦<l<10◦)
asareference(fittedlineenergy1808.72keV)andfoundcentroidenergyshifts
of-0.14keV(10◦<l<40◦)and+0.44keV(−40◦<l<−10◦),respectively.So
theyconcludedthatthemeasurementsareconsistentwiththeGalacticrotation
expectationsof±0.3keVinmodelsthoughtheerrorbarsareverylarge.

86

3.626AllineshapesfortheentdifferlongitudesalongtheGalacticPlanethe26spacedensityoffreeelectrons[TaylorandCordes26,1993]stheparentdistribution
forAlsourcesintheGalaxy,theDopplershiftsofAllineduetoGalacticrotation
havebeenmodelledusingthefittedGalacticrotationcurve(Kretschmeretal.2003,
alsosee§3.2).Typically,themodelsexpectDopplershiftsof0.25keV,varyingby∼
0.05keVwithassumptionsaboutinner-Galaxyrotationandspatialsourcedistribution,
fortheintegratedlongituderanges10◦<l<40◦and−40◦<l<−10◦(seesimulated
profilesinFigure3.21).
WefirststudiedtheDopplershiftsof26AllineduetoGalacticrotationwiththe
1.5-yearSPIdata(Diehletal.2006).ForextractionofDopplershiftsfromSPIobser-
vations,weadoptedasymmetricsourcedistributionintheinnerGalaxyintheform
ofanexponentialdisk(inFigure3.11,scaleradius4kpc,scaleheight180pc).Asym-
metricmodelavoidsbiasfromuncertaintiesofspatialdistributiondetails,andthis
exponentialdiskmodelisalsoplausiblebecausespiralstructureisinsignificantinthe
innerGalaxybelow∼35◦.Thenwesplitthismodelintothreelongitudeintervals
(−40◦<l<−10◦,−10◦<l<10◦,10◦<l<40◦),andsimultaneouslyderivedthe
spectraforthreesegmentsusingthemodelfitting.Theresultsareplottedascrosses
inFigure3.21.UsingthecentralGalacticregion(−10◦<l<10◦)asareference(fitted
lineenergy1808.72keV),wefoundcentroidenergyshiftsof-0.14keV(10◦<l<40◦)
and+0.44keV(−40◦<l<−10◦),respectively.Thoughtheerrorbarsfortheline
positionsareverylarge,weconcludedthattheresultsofthelineenergyshiftsare
consistentwiththeGalacticrotationeffect[Diehletal.,2006a].
Here,werevisitthe26AllineenergyshiftsduetoGalacticrotationwithfour-year
SPIdata.Anewskymodelareappliedhereinthemodelfitting.Wehavecreateda
homogenousdiskmodel(Figure3.11,scaleheight200pc)toavoidthestructuresalong
theGalacticplane.Theexponentialdiskmodelmaybenotgoodenough,becausewe
havefoundthat26Alemissionisnotweakfromlongitudes|l|>30◦(see26Alspectra
inFigure3.20).
Wesplitahomogenousdiskmodelmapintothreeintervals(−10◦<b<10◦):
−40◦<l<−10◦,−10◦<l<10◦,10◦<l<40◦.Thenwederivethespectraforthree
segmentssimultaneouslyusingmodelfitting,whichcanbecomparedwiththeearly
SPIresults[Diehletal.,2006a].ThreederivedspectraarepresentedinFigure3.22.The
linecentroidpositionshiftsareclearlyobservedalongthethreelongitudeintervals.
The26Allinecentroidenergyfromthecentralinterval−10◦<l<10◦is1808.82±
0.11(stat)±0.1(sys)keVwithaminorblueshift,thoughitwouldbeconsistentwiththe
laboratoryvalueconsideringtheerrorbars.Foracomparison,similartotheprevious
SPIwork[Diehletal.,2006a],weusethecentralGalacticregion(−10◦<l<10◦)as
areference(fittedlineenergy1808.82keV)andfindcentroidenergyshiftsof−0.16±
0.13keV(10◦<l<40◦)and+0.62±0.09keV(−40◦<l<−10◦),respectively.
ThelineenergyshiftshavebeenplottedinFigure3.23.Theredshiftof∼0.16keV
inthepositivelongitudeintervalislowerthanthepredictionof∼0.3keV,butis
stillconsistentwiththeGalacticrotationexplanation.Theblueshiftof∼0.62keV
innegativeintervalisquitesignificant,evenconcludingtheerrorbars,thisblueshift

87

326AlemissionandlineshapesintheGalaxy

]2.0E=1808.66 (±0.13)
-1I=3.21 FWHM=3.58 (±0.40)(±0.34)
keV1.5-1rad-1s-21.0 ph cm-40.5Intensity [100.0-0.51815181018051800Energy [keV] 2.0-1]E=1808.82 FWHM=2.61 (±(±0.11)0.26)
keV1.5I=2.27 (±0.30)
-1rad-1s-21.0 ph cm-40.5Intensity [100.0-0.51815181018051800Energy [keV] 0.09)±(E=1809.44 -1]2.0I=3.67 FWHM=3.23 (±0.34)(±0.23)
keV-1rad1.5-1s-21.0 ph cm-40.5Intensity [100.0-0.51815181018051800Energy [keV]

1820

1820

1820

Figure3.22:26AlspectraofthreesegmentsalongtheGalacticplanederivedfromthe
◦◦◦◦−10model◦<lfitting<10◦using(middle),three−40◦sub-maps<l<(−−1010◦<(dobwn).<10Ener):gy10shifts<lof<the4026Al(top),line
centroidareobservedalongtheplane.

88

3.626AllineshapesfortheentdifferlongitudesalongtheGalacticPlaneFigure3.23:26AllineenergyshiftsalongtheGalacticplane.Similartotheprevious
SPIwork[Diehl◦etal.,◦2006a],◦heretheline◦centroidofthe26Alspectrumfromthe
regionof−10<l<10,−10<b<10hasbeensettobezero.

0.23)±(E=1808.66 ]-14I=4.82 FWHM=2.58 (±1.41)(±0.57)
keV-1rad-1s-22 ph cm-40Intensity [10-218001805Energy [keV]181018151820

Figure3.24:26AlspectrumfromtheGalacticcenter(−5◦<l<5◦,−10◦<b<10◦)
derivedfrommodelfitting.Thedeterminedlinecentroidenergyisconsistent
withthelaboratoryvalue.

89

326AlemissionandlineshapesintheGalaxy◦◦26Figuregitude3.25:binAlofline20◦,enerfittedgy26AlshiftsspectraalongfrtheomGalacticFigure3.plane18)(−relativ60e<tolthe<60line,withcentroidlon-
ofthe26Alspectrum(fittedenergy1808.66keV,seeFigure3.24)fromtheGalactic
centerregionof−5◦<l<5◦,−10◦<b<10◦whichissettobezero.
cannotbeexplainedbythepresentGalacticrotationmodels[Kretschmeretal.,2003,
].2007,KretschmerWehaveusedthe26Allinecentroidenergyfortheregion−10◦<l<10◦,−10◦<
b<10◦asareferencetostudytheGalacticrotationeffect.Butwehavefoundthe
thecentroidGalacticalsorhasotationaefminorfectonblueshiftthe26Alrelativlineetoenerthegyshifts,laboratorwyevarealue.unhappForystudiesforthisof
reference.aasshiftsystematicWithmoreSPIdataanalyzedhere,wecanobtainthe26AlspectrumtheGalactic
center26region(−5◦<l<5◦,−10◦<b<10◦)usingmodelfitting.Thedetection
ofdeterAlminedintheat1808Galactic.66keVcenterisiswellstillconsistentsignificantwith(>4theσ).Thelaboratorlineyvcentralue.oidSoenerwegyuseis
the26AllinespectrumfortheGalacticcenter(centroidenergyof1808.66keV)asa
reference26tostudylineenergyshiftsforthe1stand4◦thquadrants.Furthermore,we
used(Figurethe3.18Al).26Alspectralinederivenergyedfrshiftsom6alongsegmentslongitudesin20arelongitudeplottedinbinsFigurealong3.25the.plane
Inpositivelongitudesof∼0◦−40◦,theredshiftinenergyis∼0.1±0.2keV,lower
thantheexpectationof0.2–0.5keVforthislongitudeintervalfromGalacticrota-
tion.shiftsintheConsideringpositivtheelongitudes.uncertaintiesofWhile,themeasuredblueshiftlineforpositions,thetherenegativearenolongitudesignificantcase
islarge,∼0.48keVfor−20◦<l<0◦andupto0.9keVfor−40◦<l<−20◦.This
asymmetryofthelineenergyshiftmayimplytheBarstructureexistsintheGalactic
centerregion(Figure3.14,EnglmaierandGerhard1999).IntheinnerGalacticregion
|l|<30◦,arapidlyrotationbarwillaffectthedynamicsofgasesandstars.Thelarge

90

3.626AllineshapesfortheentdifferlongitudesalongtheGalacticPlanenon-circularmotionshavebeenseeninHIandCOobservations[MulderandLiem,
19971986,],GersourharcedandcountVietri,asymmetries1986],and[alsoNikolaefrvomandtheWNIReinberlightg,1997]distributionand[gasBinneydynamicsetal.,
[tionEnglmaierangle∼and25◦ofGerthehard,bar(1999].EnglmaierThenon-cirandGercularhardmotions1999,inalsotheseeinnerFigurering3.14and)awouldposi-
leadtotheasymmetryofthe26AllineenergylineshiftsintheinnerGalaxy.
Blueshiftsatpositivelongitudesarehigherthanthe◦expectationof◦Galacticrotation
[0.9keVKretschmercannotetbeal.,2003explained].byEspeciallythepresentfortheGalacticcaseofr−otation40<lmodels<−[20,aKretschmerblueshiftetal.of,
2003,Kretschmer,2007].IfGalacticrotationleadstoablueshift26of0.3keV,anad-
ditionalshiftof∼0.6keVcomesfromotherDopplershiftsofAlsourceswitha
inatebulk-motioninlocalvstarelocity-forof∼mation100kmregions,s−1toe.g.,wardtheus.nearestThisbulkpartofmotionstheof26AlSiguttarius-Carinacouldorig-
inarmabout(Figurelast330.14),Myror[theGrenierGould,2000Belt]andwhichincludeisasuggestedlargetoamountbeactivofelocalinstarstarfor-formationsmation
complexes[PerrotandGrenier,2003].Contributionsoftheselocal26Alsourceswould
afbefectstudiedthe26Alandlinediscussedshapesinto§w3.ar7.dthisdirectionoftheGalacticplane.Thisissuewill

3.6.326Alemissioninsmalllongitudeintervals
Detectionsof26Allineshapesforthe∼10◦-longitude-binintervalsoftheinnerGalaxy
arenotsignificantyetbythepresentSPIobservations(seeFigures3.19and3.20).So
directprobeof26Allineshapesinsmallerregionsisnotpossible.However,SPIhasan
angularresolutionof∼2.7◦forimagingat1.8MeV(Figure3.15).Thenitispossible
tostudythechangesof26Alspectraforlongitudeintervalswithincreaseorshiftof
longitudebinsofinputskymapsbyastepassmallas∼2.7◦usingmodelfitting.The
significantdifferencesof26Alspectramaynotbefoundduetothechangesofsmall
scales(largeerrorbars,orotherreasons),butthiswouldbeagoodtestfortheSPI
26imaging.AlofabilityMultiplemodelfittingjobsshouldbecarriedouttoprobethedifferencesof26Al
spectrawithchangesoflongitudeintervalsbyastepofseveraldegrees.Statistical
uncertaintiesmaybealittledifferentforvariousjobsofmodelfittings.Butthesame
database,samebackgroundmodelsandsameskymodels(sub-mapsofahomogenous
diskmodel)areappliedinmodelfittings,sothechangesof26Alspectrashouldnot
resultfrombiasesexceptforstatisticaluncertainties.
TheGalacticcenterregion|l|<15◦
Figure3.26showschangesof26Alspectraforthedifferentlongitudeintervalsin
theGalacticcenterregion(|b|<10◦):from|l|<2.5◦to|l|<15◦withastepof
2.5◦.26AlemissionisbrightestintheGalacticcenter(|l|<2.5◦and|l|<5◦),andthe
linecentroidenergyiswellconsistentwiththelaboratoryvaluewithinuncertainties.

91

326AlemissionandlineshapesintheGalaxy
8E=1808.58(±0.33)E=1808.66(±0.23)
1-]VIF=7W.3H1M(=±22..7893)(±0.82)11-]V4IF=4W.8H2M(=±21..5481)(±0.57)2
1-ek6-1ek
1dar1dar
-2-s4-2-s2
mmc h2c h
4-p 4-p
[ 01001[ 0
ytisytis
net-2net
nInI-2
4-18001805Ener1g8y1 [0keV]1815182018001805Ener1g8y1 [0keV]18151820
0.231-]FEW=1H8M08=.28.115(±(0±.01.74)2)31-]FEW=1H8M08=.28.261(±(0±.01.12)6)4
VeI=2.90(±0.75)VeI=2.27(±0.30)
1-k2-1k1.5
dd1-ar1-ar
2-s2-s1.0
mc 1mc
hh4-p 4-p 0.5
0011[ yt0[ yt
iisnesne0.0
ttIn-1nI
5.0-1800180518101815182018001805181018151820
Energy [keV]Energy [keV]
1-]2.0FEW=1H8M08=.28.668(±(0±.01.02)4)1-]2.0FEW=1H8M08=.29.279(±(0±.00.92)1)
VeI=2.48(±0.29)5VeI=2.55(±0.26)6
k1-kd1.51-d1.5
1-ar1-ar
ss2-m1.02-m1.0
c hc h
pp 4-00.54-00.5
11[[ ytiyti
sne0.0sne0.0
tnItnI
-0.5-0.5
1800180518101815182018001805181018151820
Energy [keV]Energy [keV]
Figure3.26:Comparisonof26AlspectraintheGalacticcenterregionforsixdifferent
longitudeintervals(−10◦<b<10◦):1.−2.5◦<l<2.5◦,2.−5◦<l<5◦,3.
−7.5◦<l<7.5◦,4.−10◦<l<10◦,5.−12.5◦<l<12.5◦,6.−15◦<l<15◦.
92

3.626AllineshapesfortheentdifferlongitudesalongtheGalacticPlane 34E=1808.93 (±0.59)3E=1808.88 (±0.36)E=1808.52 (±0.39)
-1]FWHM=3.06 (±0.00)-1]FWHM=3.06 (±0.00)-1]FWHM=3.18 (±0.00)
keVI=2.67 (±0.98)keVI=2.39 (±0.53)keV2I=2.25 (±0.53)
2-1-12-1radradrad-1s-1s-1s
-20-21-21
-4 ph cm-4 ph cm-4 ph cm
00-2Intensity [10Intensity [10Intensity [10-1-1-4180018051810181518201800180518101815182018001805181018151820
Energy [keV]Energy [keV]Energy [keV] E=1808.86 (±0.38)3E=1808.48 (±0.42)E=1809.28 (±0.32)
-1]4FWHM=3.06 (±0.00)-1]FWHM=3.18 (±0.00)-1]FWHM=3.57 (±0.80)
I=3.98 (±0.94)I=2.49 (±0.63)2I=3.36 (±0.99)
-1keV-1keV2-1keV
-1rad-1rad-1rad
-2s2-2s1-2s1
-4 ph cm-4 ph cm0-4 ph cm
00Intensity [10Intensity [10Intensity [10-1-1-2-2180018051810181518201800180518101815182018001805181018151820
Energy [keV]Energy [keV]Energy [keV]

Figure3.27:Comparisonof26Alspectrawithchangesoflongitudeintervalsbyastep
of2.5◦fortheleftpart(upper,l>0◦)andrightpart(below,l<0◦)oftheGalactic
center:0◦<|l|<10◦(left),0◦<|l|<12.5◦(middle),0◦<|l|<15◦(right).

2.5E=1808.67 (±0.20)E=1808.62 (±0.21)E=1808.70 (±0.31)
-1]2.0FWHM=3.61 (±0.50)-1]FWHM=3.54 (±0.53)-1]3FWHM=3.18 (±0.00)
keVI=3.29 (±0.60)keV2I=3.47 (±0.68)keVI=3.27 (±0.61)
-1-1-11.5-1rad-1rad-1rad2
-2s-2s-2s
1.011-4 ph cm0.5-4 ph cm-4 ph cm
0Intensity [100.0Intensity [10Intensity [100-1-0.5180018051810181518201800180518101815182018001805181018151820
Energy [keV]Energy [keV]Energy [keV] E=1809.55 (±0.16)E=1809.61 (±0.17)E=1809.17 (±0.40)
-1]2FWHM=3.22 (±0.41)-1]2FWHM=3.13 (±0.42)-1]FWHM=3.18 (±0.00)
I=3.21 (±0.54)I=3.34 (±0.60)2I=2.70 (±0.66)
-1keV-1keV-1keV
-1rad-1rad-1rad
-2s1-2s1-2s1
-4 ph cm-4 ph cm-4 ph cm0
00Intensity [10Intensity [10Intensity [10-1-1-1-2180018051810181518201800180518101815182018001805181018151820
Energy [keV]Energy [keV]Energy [keV]

Figure3.28:Comparisonof26Alspectrawithchangesoflongitudeintervalsbyastep
of2.5◦for◦thepositive◦(upper)and◦negative◦(below)longitudes:10◦<|l|<30◦
(left),12.5<|l|<30(middle),15<|l|<30(right).

93

326AlemissionandlineshapesintheGalaxy 3E=1808.54 (±0.48)E=1808.98 (±0.48)E=1808.60 (±0.41)
-1]2FWHM=4.62 (±1.27)-1]FWHM=3.06 (±0.00)-1]1.5FWHM=3.57 (±1.05)
keVI=2.66 (±0.95)keV2I=2.35 (±0.71)keVI=1.76 (±0.68)
-1-1-11.0radradrad-1-2s1-1-2s1-2-1s
0.5-4 ph cm-4 ph cm0-4 ph cm
0.00Intensity [10Intensity [10Intensity [10-1-0.5-1-1.0-2180018051810181518201800180518101815182018001805181018151820
Energy [keV]Energy [keV]Energy [keV] 3E=1809.81 (±0.25)4E=1809.23 (±0.26)E=1809.21 (±0.21)
-1]FWHM=3.41 (±0.62)-1]FWHM=2.79 (±0.65)-1]FWHM=3.01 (±0.52)
I=4.44 (±1.06)3I=4.26 (±1.31)2I=3.04 (±0.69)
-1keV2-1keV-1keV
-1rad-1rad2-1rad
-2s-2s-2s
11-4 ph cm-4 ph cm-4 ph cm
1000Intensity [10Intensity [10Intensity [10-1-1-1-2180018051810181518201800180518101815182018001805181018151820
Energy [keV]Energy [keV]Energy [keV]

Figure3.29:Comparisonof26Alspectrawithchangesoflongitudeintervalsbyastep
of5◦for◦thepositiv◦e(upper)and◦negative◦(below)longitudes:25◦<|l|<40◦
(left),30<|l|<40(middle),30<|l|<45(right).

Intheregionsof|l|<7.5◦and|l|<10◦,thebrightnessdecreasesbyafactorof2
relativetotheGalacticcenter,whichsuggeststhat26Alemissioncouldbeweakin
thedirectionsof5◦<|l|<10◦.Extendingtothelargerregionsof|l|<12.5◦and
|l|<15◦,thedetectionsof26Alarequitesignificant(>10σ),andthe26Albrightness
doesnotchange(brighterthantheaveragebrightnessfor|l|<10◦).Sosignificant
26Alemissionexistsinthedirectionsof10◦<|l|<15◦,withanaveragebrightness
of∼2.5×10−4phcm−2s−1rad−1.The26Allinecentroidswithaminorblueshiftare
detectedfrom|l|≥7.5◦.Onepossiblereasonmaybetheasymmetriesof26Alemission
inthepositiveandnegativelongitudes.Thenegativelongitudeisbrighterthanthe
positiveone,thenduetoGalacticrotation,theblueshiftcouldbeobservedasawhole
behavior.Anotherpossibilitymaybeduetothebulkvelocitiesof26Alsourcesin
localstar-formationsystemstowardus.Wewillfurtherdiscussthisprobleminthe
sections.wingfollo26Alspectrawithchangesoflongitudeintervalsfortheleftandrightpartsofthe
Galacticcenterregionarealsostudiedseparately.Thelongitudebinincreasesfrom
l=10◦tol=15◦byastepof2.5◦forbothsides.WecomparethesixspectrainFigure
3.27.Thedetectionsof26Alfortheselongitudeintervalsarenotsignificantexceptthe
regionof−15◦<l<0◦.Thoughlargeerrorbarsexists,thebrightnessof26Alforthe
negativelongitudesissystematicallyhigherthanthatforthepositiveones(|l|<15◦).
10◦<|l|<30◦

94

3.7Latitudestudiesof26Alemission E=1809.35 (±0.36)E=1808.78 (±0.08)E=1807.61 (±****)
-1]FWHM=1.34 (±0.87)-1]2.0FWHM=3.19 (±0.19)-1]3FWHM=12.39 (±****)
2I=1.08 (±0.92)I=3.38 (±0.27)I=0.99 (±7.45)
-1keV-1keV1.5-1keV
2-1rad1-1rad-1rad
-2s-2s-2s
1.010-4 ph cm-4 ph cm-4 ph cm0
0.5-1Intensity [10Intensity [10Intensity [10-10.0-2-2-0.5180018051810181518201800180518101815182018001805181018151820
Energy [keV]Energy [keV]Energy [keV]

Figure3.30:26AlspectraforthreeGalacticlatitudeintervals(−60◦<l<60◦):5◦<
b<20◦(left),−5◦<b<5◦(middle),−20◦<b<−5◦(right).

Furthermore,westudy26Alspectrawithsmalllongitudechangesfortheouterparts
oftheinnerGalaxy.InFigure3.28,wepresentthecomparisonof26Alspectrawith
changesoflongitudeintervalsbyastepof2.5◦forboththepositiveandnegative
longitudesof10◦<|l|<30◦.Significantblueshiftsofthe26Allinearedetectedfor
thenegativeparts,butsmallredshiftsforthepositiveregions.For10◦<l<30◦,the
26Allinebrightnessisnearlyconstantwithalevelof∼3.3×10−4phcm−2s−1rad−1.
Thenegativepartisnotbrighterthanthepositiveone.Andtheregionof−30◦<
l<−15◦issignificantlydarkerthan−30◦<l<−12.5◦.Thisimpliesthat26Al
emissionfor−15◦<l<−12.5◦shouldbehigherthantheaveragelevelof∼3.4×
10−4phcm−2s−1rad−1.

25◦<|l|<45◦
Figure3.29displaysthecomparisonof26Alspectrawithchangesoflongitudeinter-
valsbyastepof5◦forboththepositiveandnegativelongitudesof25◦<|l|<45◦.
Thesignificantenergyblueshiftsofthe26Allinecentroidarealsodetectedfortheneg-
ativelongitudes.The26Allinecentroidfor−40◦<l<−25◦hasalargeblueshiftof
◦◦∼This1.1difkeVferentrelativeblueshiftstotheimplylaboratorthatyinvthealue,anddirectiononlyof∼−0.630◦ke<Vlfor<−−2540◦,26<All<sour−30ces.
Thelocatedinbrightnessnearbofy26starAl-forforthemationnegativcomplexesehalongitudesveaishighmeansystematicallybulkvelocityhighertowthanardthatus.
◦◦isforthethebrightestpositiveoneones,withwithanaavfluxerageratiofluxoflevnearelof1.7.∼4.4The×10region−4phof−cm40−2s<−1lrad<−−1.25It
isinterestingthatwecompare26Alspectrafortheregionsbetween30◦<|l|<40◦
◦◦◦◦30and◦<30|l|<<|l40|◦<by45a.factorTheof∼brightness1.35,of30suggesting<|l|that<2645Aldropsemissionforsignificantly|l|>40◦relativisveertoy
weak,whichisconsistentwiththeresultsinFigure3.18.

95

326AlemissionandlineshapesintheGalaxy E=1809.37 (±0.97)3E=1808.45 (±0.27)4E=1808.48 (±2.43)
-1]4FWHM=1.11 (±2.32)-1]FWHM=3.49 (±0.70)-1]FWHM=2.49 (±5.96)
keVI=0.47 (±1.30)keVI=3.78 (±0.99)keVI=0.63 (±2.00)
-1-122-1radradrad2-1s-1s-1s
-2-2-2100-4 ph cm-4 ph cm-4 ph cm
0-2-2Intensity [10Intensity [10Intensity [10-1-4-4180018051810181518201800180518101815182018001805181018151820
Energy [keV]Energy [keV]Energy [keV] E=1809.44 (±0.46)E=1809.20 (±0.23)E=1809.56 (±1.45)
-1]4FWHM=2.06 (±1.12)-1]3FWHM=3.33 (±0.59)-1]4FWHM=1.30 (±3.47)
I=1.95 (±1.40)I=4.23 (±0.98)I=0.39 (±1.37)
-1keV-1keV-1keV
2-1rad2-1rad-1rad2
-2s-2s-2s
1-4 ph cm-4 ph cm0-4 ph cm0
0Intensity [10Intensity [10-1Intensity [10-2-2-2-4180018051810181518201800180518101815182018001805181018151820
Energy [keV]Energy [keV]Energy [keV]

Figure3.31:26Alspectrafordifferentlatitudeintervalsofthe1st(upper)and4th
(below)quadrants:5◦<b<20◦(left),−5◦<b<5◦(middle),−20◦<b<−5◦
(right).

3.7Latitudestudiesof26Alemission
Boththe26AlimagingbyCOMPTELorSPI,and26Alspectraindifferentlongitudes
obtainedbySPIcannotdirectlyprobethedistanceinformationof26Alsources.The
detected26Alsignalsintheskycouldoriginatefromthelocalstar-formationcomplexes
(∼100pc),orfromthenearestpartoftheSiguttarius-Carinaarm(1−2kpc),orfrom
theGalacticcenterregion(∼8kpc),orevenfromtheothersideoftheGalaxy(>10
kpc).Inthissection,wetryapossiblewaytoresolvethe26Alsignalsforthelocalcom-
plexesfromthelargescalesoftheGalacticplane.Wesplitthehomogenousdisk
modelintosub-mapsalongGalacticlatitudes,andderive26Alspectraforthedifferent
latitudesusingmodelfitting.Wehavesplittedtheskymapintothreepartsalonglow
latitudes,−5◦<b<5◦,andintermediatelatitudes5◦<b<20◦and−20◦<b<−5◦.
26Alemissionforlowlatitudes(|b|<5◦)wouldbedominatedbythelarge-scaleorigin
intheGalacticdisk.While26Alsourcesforhighlatitudes(|b|>5◦)shouldoriginate
fromlocalstar-formationsystemsintheGouldBelt.Thisdefinitionisalsosimilar
topulsarpopulationstudies[Wangetal.,2005b,Chengetal.,2004].TheGouldBelt
appearsasanellipsoidalshapedringwithsemi-majorandminoraxesequalto∼500
pcand340pc,respectively[PerrotandGrenier,2003].TheSunisdisplacedfromthe
centeroftheGouldBeltabout200pctowardsl=130◦[Guilloutetal.,1998].The
VelaregionislocatedneartheboundaryoftheGouldBelttowardsl∼−90◦.The

96

3.7Latitudestudiesof26AlemissionnearbySco-CenregionalsobelongstotheGouldBelt,extendingfrom(l,b)=(0◦,20◦)
towards(l,b)=(−30◦,0◦)[Sartorietal.,2003].
Figure3.30displaysthree26AlspectrawithGalacticlatitudes(|l|<60◦).No26Al
signalisdetectedinthenegativelatituderegion(−20◦<b<−5◦).Andweak26Al
emissionisdetectedinthepositivelatituderegion(∼2σ,5◦<b<20◦).Itisquite
interestingthatthelinecentroidhasasignificantblueshiftof0.7±0.4keV.The26Al
emissionforthelowlatitudes(|b|<5◦)isbrightestasexpected,withanarrowline
feature,andlinecentroidenergyof1808.78±0.08keV.Theblueshiftrelativetothe
laboratoryvalueisnotsignificantyet(comparedtotheresultsinTable3.1).Thenwe
haveexplanationofthelargeblueshiftofthe26AllinecentroidinTable3.1:thelarge
blueshiftcannotoriginateinspiralarmsandthepossibleasymmetryof26Alemission
intheGalacticplane,butmainlycomesfromintermediatelatituderegionswherethere
existnearby26Alsourceswithbulkvelocitiestowardus.
Wefurtherspliteachlatitudeintervalintotwoparts:the1stand4thquadrants.
26ThenwecarryoutthemodelfittingandderivesixspectrashowninFigure3.31.No
Alsignalsaredetectedfortwointermediatelatituderegionsofthe1stquadrant.
Whileinthe4thquadrant,26Alemissionisstillclearlydetectedinthelatituderegion
of5◦<b<20◦,withthesignificantblueshiftofthelinecentroid,0.8±0.5keV.
Additionally,averyweak26Alsignalmaybehintedtowardthelatituderegionof
−60◦<l<0◦,−20◦<b<−5◦,whichneedafurthercheckinfuture.Inthelow
latitudecases,asanexpectation,relativetothelaboratoryvalue,the26Allinecentroid
showsaredshiftof0.21±0.27keVforthe1stquadrant,andablueshiftof0.54±0.23
keVforthe4thquadrant,whichisconsistentwiththeexplanationofGalacticrotation.
The26Alspectraforthe1stand4thquadrantsalsoshowthenarrowlinefeatures,and
theasymmetryof26Albrightnessfortwopartsisalsodetected(afluxratioof1.12),
thoughtheerrorbarsarelargeforbothspectra.
In§3.6.2,wehavestudiedtheGalacticrotationeffecton26Alspectraforthreelon-
gitudeintervals:10◦<l<40◦,−10◦<l<10◦,−40◦<l<−10◦.Butthespectra
obtainedin§3.6.2possiblyincludesthecontributionsfromhighlatitudes.Toavoid
thecontaminationsfromhighlatitudes,were-studythe26Alspectraforthesethree
longitudeintervalsonlyinthelowlatitudes(|b|<5◦)usingthemodelfitting.
Figure3.32presentsthe26AlspectraofthreesegmentsalongtheGalacticplane.
Forthelowlatitudecases,energyshiftsofthe26Allineareobservedalongtheplane.
SimilartoFigure3.22,aredshiftof0.11±0.35keVisnotsignificant,whilealarge
blueshiftof0.75±0.27keVisstillinconsistentwiththeexpectofGalacticrotation.So
thedynamicsofthebarstructureinthebulgemayaffectthe26Allineshiftsinpositive
longitudes,andinnegativelongitudes,motionsofnearby26Alsourcesintheplane
direction(|b|<5◦)towardsuswouldpartlycontributetothislineblueshift.
Thelinecentroidenergyforthecenterregion(|b|<5◦,|l|<10◦)isdetermined
at1808.65±0.27keV,wellwithinthelaboratoryvalue.Comparedwiththe26Alspec-
trumfortheGalacticcenter(|b|<10◦,|l|<5◦)inFigure3.24andspectrainhigh
latitudesinFigure3.31,weconcludethatnearby26Alsourceslocatedinthedirection

97

326AlemissionandlineshapesintheGalaxy 3-1]E=1808.54 FWHM=3.78 (±(±0.35)0.91)
1.18)±(I=3.76 keV2-1rad-1s-21 ph cm-40Intensity [10-11815181018051800Energy [keV] 0.27)±(E=1808.65 ]-13I=3.75 FWHM=3.30 (±1.03)(±0.69)
keV-12rad-1s-21 ph cm-40-1Intensity [10-21815181018051800Energy [keV] -1]3E=1809.40 FWHM=3.74 (±(±0.27)0.68)
1.05)±(I=4.38 keV-12rad-1s-21 ph cm-40Intensity [10-1-21815181018051800Energy [keV]

1820

1820

1820

Figure3.32:26AlspectraofthreesegmentsalongtheGalacticplane(−5◦<b<5◦):
10◦<l<40◦(top),−10◦<l<10◦(middle),−40◦<l<−10◦(down).Similar
toFigure3.22,energyshiftsofthe26Allinecentroidareobservedalongtheplane.
Andthelinecentroidenergyforthecenterregion(|l|<10◦)isdeterminedat
1808.65±0.27keV,consistentwiththelaboratoryvalue.

98

3.81809VkeemissionintheinnerGalaxy:simulationsversusobservations◦◦oftiesbto>w5ar,dsl<us,−5,inducingpossiblethetheobserSvedco-Cen26AllineregioninblueshiftstheGouldfortheBelt,innerhaveGalaxyhighvregioneloci-
(e.g.,the26AlspectrainFigure3.26).Wewillstudythepossible26Alemissionforthe
Sco-CenregioninthenextChapter.

3.81809keVemissionintheinnerGalaxy:simulationsversus
observationsWehavedoneanalysesof26AllineshapesforthedifferentlongitudesalongtheGalac-
ticplanewiththeSPIdata.Thedifferentappearancesof26Alemissioninthe1stand
4thquadrants,andthelinecentroidenergyshifts26alongthelongitudesinlargescales
aredetected.Theselarge-scalecharacteristicsofAlemissionshouldbecorrelatedto
thespiral-armstructuresoftheGalaxy,where26Alsourcecandidatesaremostproba-
blylocated.Therefore,ifweknowthedistributionof26AlsourcetracersintheGalaxy,
wecansimulatethe26Alintensityandlineshapediagnosticsalongtheplane.This
simulatedresultscanbedirectlycomparedwiththeSPIresultson26Allineshapes
e).vabo(seeKretschmeretal.[2003]havesimulatedthelineshapediagnosticsofGalactic26Al
.TheirresultsshowedtheGalacticrotationeffecton26Allineshapesinlargescales
alongtheinnerGalacticplane.Herewewillupdatethesimulationsof26Alemission
featuresintheinnerGalaxyusingsomeupdatedobservationsrecently[Kretschmer,
2007],e.g.,athree-dimensionalmodelforthespacedensityoffreeelectronsbyCordesandLazio
[2002]insteadoftheolderonebyTaylorandCordes[1993],alsonewobservationalre-
sultsontheGalacticrotationcurve[Avedisova,2005].

3.8.1ModelsofGalactic26Al
ThespacedistributionofGalactic26Alshouldcorrelatewiththedistributionofstar-
formationcomplexesintheGalaxy.Thisdistributioncanbewelltracedbyobserva-
tionsonionization(Horfree-freeemission),molecularclouds,freeelectrons,andso
on(seecomparisonαamongthedifferenttracersof26Alsources§3.4).Therefore,insim-
ulations,weadoptathree-dimensionalmodelforthespacedensityoffreeelectrons
basedonpulsardispersion26measurements[CordesandLazio,2002]astheparentdis-
tributionofGalacticAl.AcutthroughthismodelalongtheGalacticplanecan
befoundinFigure3.33[Kretschmer,2007].Wehaveusedthreecombinationsofsev-
eraldifferentcomponentsoffreeelectrondensity:(1)allcomponents;(2)allexcept
thethickdiskandthe26Galacticcenter;(3)onlythethindiskandspiralarms.These
threecombinationsasAlsourcedistributionmodelswillbeappliedinsimulations
comparison.aforseparatelydiskWe[takeOllinganandapprMerrifieldoximate,for2000mula]forofthetheIAUGalacticstandarrdotationvaluecurRv0e=in8.5thekpc,innerandΘGalactic0=

99

326Al10

Y [kpc]0

-10

emission-10

andlineshapesinthe100X [kpc]

10

Y [kpc]0

-10

Galaxy-10

10

0Y [kpc]

-10

100X [kpc]

-10

0X [kpc]

10

Figure3.33:Areadensitymodelsofsupernovaeobtainedfromthefree-electrondis-
tributionfrompulsardispersionmeasurements(NE2001,CordesandLazio2002).
Threecombinationsofseveraldifferentcomponentsofdensityarepresentedhere
(fromKretschmer2007):allcomponents(upper-left);allexceptthethickdiskand
theGalacticcenter(upper-right);onlythethindiskandspiralarms(down).

100

3.81809VkeemissionintheinnerGalaxy:220kms−1withtheradialdependenceisgivenby:

simulations|v|(R)=220[1−exp(−R/1kpc)]kms−1.

versusobservations)2.3(

Thevelocityvectorisparalleltotheplaneandperpendiculartothevectorpointing
fromtheGalacticcentertothesourcelocation.
Thekinematicsof26AlsourcesinthelocalinterstellarmediumandthewholeGalac-
ticdiskmustbeconsideredinourmodel.DominatedGalactic26Alisejectedbystellar
windsofmassivestarsandsupernovaexplosions.SotheGalacticrotationinlarge
26thescalessuperisnovasuperimposedexplosiononortheejectingmotionWofolf-Rayfreshlyetstarwindsynthesizedanditssloradioactivwed-doewnAlduemotionto
intheISMbeforedecay,i.e.,within∼106years[Kretschmeretal.,2003].
thatRecentthehydrexpansionodynamicofbulkofsimulationsofnucleosyntheticTypeIISNsuperprnovoductsae[suchKifonidisaset26Alal.,ma2003y]befindat
avvelocitieselocityofof∼15001200kmkms−1s−1until.itThus,reachesinourthemodelradiusweofallothewSN26reAlvtoerseexpandshockforfreelymedwithby
26wciravecumstellarshock.Thisprinteraction.ovidesusAfterathisconservpoint,ativeweestimateexpandbecauseAlat26theAlvisprelocityobablyofthetomoblastve
slowerthantheforwardshock.ForourmodelforSNRdynamicsfromcircumstellar
velocitiesinteraction,givwenebyadoptMcKeethevandaluesTofrueloveKepler[1995’s].supernovaremnantshockpositionsand
TypeIb/IcsupernovaeandWRstarswouldejectmatteratsimilarorevenhigher
2002speeds].InthanTaddition,ypeIISNethe[Prinjainteractionetal.,of1990,ejectedGarcia-Smatteregurawithettheal.,surr1996,oundingMellemaandmediumLundqvist,
dependsonthestarformationhistoryofthesourceregion,wherebubblesforming
aroundgroupsofyoungmassivestarsplayapotentiallylargerole.Cavitiesextending
otheversevEridanuseralcahundredvitypcpresentshaveusbeenwithobseravnearbedinygalaxiesexample[ofOeysuchandacaClarkevity,,1996],extendingand
fromtheOrionstarformationregiontoveryneartheSun[Burrowsetal.,1993].Mat-
terejectedintosuchalow-densitybubblecouldexpandalmostfreelyuntilreaching
theboundarywhereastypicallyassumedISMdensities,e.g.,nH∼1cm−3wouldslow
.rapidlywndoitWiththeaboveassumptions,theejected26Almovesfreelyfornearly2kyr,and
thethenISMatdeceleratesanagewiththecomparableSNR’stoshell.theThelifetimeshellof26Alreaches,awhenradiusawheresignificantitdissolvfractioneshasin
rthereforeotationalvalreadyelocityofdecay220ed.kmSinces−1theataboutexpansion40vkyr,elocitywhendrops96%ofbelo26wAltheisstillcharacteristicleft,the
modelcontribution[Kretschmerfrometal.expansion,2003to].theoverall26Allinewidthwillberathersmallinthe

101

326Alemissionandlineshapes410•2.0

410•1.5

410•1.0

35.010•

0-38000

6000

4000

2000

0-38000

6000

4000

2000

0-3

inthe-2

-2

-2

Galaxy-1ΔE0 [keV]1

-1ΔE0 [keV]1

-1ΔE0 [keV]1

2

2

2

3

3

3

Figure3.34:IllustrationsoflinesfromdifferentGalacticregionsforthreemodelssepa-
therately(frGalacticomcenterKretschmer(middle);2007):onlyallthecomponentsthindisk(top);andallspiralexceptarmsthe(dothickwn).diskTheandhis-
region:togramstheineachGalacticfigurecentersho(w|l|the<5◦,spectragreen),ofthethedifeasterferentnpartparts(5of◦<thel<inner40◦,galacticblue),
thewesternpart(−40◦<l<−5◦,red).Thedottedlinesrepresentthe26Alsource
spectrum,thedashedlinesarethebest-fittingGaussian,andthedash-dottedlines
102talresultfrresolutionomaofconvtheolutionmeasuringwitha3.0detectorkeV.FWHMGaussianadoptedforinstrumen-

3.81809VkeemissionintheinnerGalaxy:simulationsversusobservations3.8.226Alintensityandlineshapes:simulationsversusobservations
Weobtainsimulatedsky26AlintensityandspectrafromaMonteCarloscheme:26Al
sourcelocationsarechosenrandomlyfromaspatialdistributionproportionaltothe
freeelectrondensity(Figure3.33,CordesandLazio2002,Kretschmer2007),distribut-
ingcandidatesourcepositionswithinavolumecenteredontheGalaxyandextending
25kpcintheplaneand7.5kpcperpendiculartotheplane.Foreachnucleosynthesis
event,arandomageischosenwithintheinterval[0,2×107yr].Fromthisweevalu-
atetheintrinsicvelocitydistributionandsizeofthe26Alsource,followingtheabove
expansionmodel.Theageofthenucleosynthesiseventthusdeterminesextentand
intrinsicvelocityofitsejecta,aswellastheir1.809MeVluminosity.Werepresenteach
eventby210masselementstoreflectitsspatialextent[Kretschmer,2007].
AssumingR0=8.5kpc,andΘ0=220kms−1,weobtainviewingdirectionand
radialvelocityof26Alsources.Directionandradialvelocitygiveusthecoordinatesof
the26Alsourcemasselementinadataspaceof26Aldecayluminosityasafunction
oflongitude,latitudeandphotonenergy[Kretschmeretal.,2003,Kretschmer,2007].
Simulationshavebeencarriedoutusingthreecombinationsofthefreeelectrondensity
modelrespectively(Figure3.33).
InordertocomparewiththeSPIresults,weintegrateoursimulatedskymapof
lineenergiesandintensitiesoveraregionofinterestandobtainaresultingspectrum
correspondingtoanobservationofthisregionwithperfectspectralresolution.The
integrated26AlspectrafromourmodelsforthreepartsoftheinnerGalaxyregion
(−5◦<l<5◦,5◦<l<40◦,−40◦<l<−5◦,respectively)arepresentedinFigure
3.34.Ourmodelspectra(dottedlines)showthedeviationfromafittedGaussianshape
(dashedlines).FortherealisticSPImeasurements,Gedetectorshaveinstrumentalres-
olution3.0keVFWHMaround1.8MeV(seeFigure2.9).Thisinstrumentalresolution
convolutionwiththeinstrumentresponsesuppressesthedifferencebetweenthemea-
suredfluxandaGaussianbyafactorof≈2000(seedash-dottedlinesinFigure3.34).
Thesimulated26AllineshapescanbedeterminedandcomparedwiththeSPImea-
surements.The26AllinesforthreepartsoftheinnerGalaxyshownarrowlinefeatures,
whichiswellconsistentwithourSPIconstraints.WhenwefitGaussianstothesespec-
tra,weobtainalinewidthof∼0.3keVfortheGalacticcenter(greenlines,|l|<5◦),
and∼0.8−0.9keVfortheleftandrightpartsoftheinnerGalaxy(blueandredlines,
5◦<|l|<40◦).
Thereareno26AllinecentroidshiftsfortheGalacticcenter(|l|<5◦)inallthree
distributionmodels,whichisalsoconsistentwiththeobservedspectrumbySPI(see
Figure3.24).Butthe26Allinecentroidenergiesfortheleftandrightpartsoftheinner
Galaxy(5◦<|l|<40◦)areshiftedbyasignificantlyoffsetrelativetothecentralregion
(|l|<5◦).Andtheshiftsaresymmetricalforthepositiveandnegativelongitudes.The
shiftedvaluesarealittledifferentforthreemodels.Forthemodelofallcomponents,
the26AllineshiftduetoGalacticrotationis∼0.2keV,while∼0.3keVforthemodel
ofallexceptthethickdiskandtheGalacticcenterand∼0.4keVforthemodelofonly

103

326AlemissionandlineshapesintheGalaxythethindiskandspiralarms.ComparedtoourSPIresultson26Allinecentroidshifts
alongtheGalacticplane(seeFigure3.25),aredshiftforpositivelongitudesof∼0.1
keVislowerthanthepredictionofmodels,whileablueshiftfornegativelongitudesof
≤0.5keVishigherthantheGalacticrotationexpectationin26simulationsevenforthe
modelofonlythethindiskandspiralarms.AsymmetryofAllinecentroidshifts
forpositiveandnegativelongitudesisinconsistentwiththeexplanationofGalactic
otation.rWealsocomparethe26Alintensitiesfortheleft(5◦<l<40◦)andright(−40◦<l<
−5◦)partsoftheinnerGalaxy.Forthemodelofallcomponents,the26Alintensityfor
theleftpartisalittlehigherthanthatfortherightpart,whichisnotconsistentwith
ourSPIresultsin§3.6.1.ForthemodelofallexceptthethickdiskandtheGalactic
center,the26Alintensityofeachpartissimilar,thesignificantasymmetrydoesnot
appear.Butforthemodelofonlythethindiskandspiralarms,the26Alintensityfor
−40◦<l<−5◦issignificantlyhigherthanthatfor5◦<l<40◦,withafluxratio
of∼1.15.Thisasymmetryhasbeendetectedin26Alspectroscopyforthe1stand4th
quadrants(seeFigure3.16,3.17and3.22).InourSPIobservations,wehavereported
afluxratioof∼1.3for|l|<60◦,∼1.2for|l|<30◦and∼1.14for10◦<|l|<40◦.
Therefore,weconcludethatthefree-electrondensitymodelofonlythethindiskand
spiralarmsreflectstherealdistributionofGalactic26Al,whichisalsoconsistentwith
theexpectationofthemassivestardominatedoriginof26AlsourcesintheGalaxy.

104

426Alsourcesandlineshapesinstar
formationregions

TheFigures1.COMPTEL6and1.71809,kePlüschkeVallsky2001)mapsattributed(i.e.,tomaximumradioactiventreopdecayyandofMREMGalactic26maps,Alhavseee
confirmedthediffuseemissionalongtheinnerGalaxy.Severalsignificantfeaturesof
thereconstructedintensitypatternarefluxenhancementsinthedirectionsofnearby
[starDiehl-foretal.mation,1995b],regions,thee.g.,CarinatheCyregiongnus[Plüschkeregion,[del2001Rio].Aetal.,possible1996],featuretheVelaaboveregionthe
Galacticcenter(inhighlatitudes)hasbeenalsoreported,which26isattributedtothe
thenearbythree-ySearco-CenSPIregionobser[vationsPlüschke,(Figure20013]in.15,theHalloinGouldetBelt.al.The2007,Alinmappreparation)obtainedalsoby
confirmedthebrightemissionfromtheCygnusregion.
Thesenearbystar-formationregionsareveryyoung(severalMyr),and26containa
wealthofmassivestars,andrecentsupernovaremnants.DetectionsofAlinthese
regionsstronglysupportthehypothesisofmassivestarsandtheirdescendentsuper-
novaebeingthedominantsourcesofinterstellar26Al.26Alneartheseyoungpopula-
tionsmaybelocatedthedifferentmediumenvironmentfromthelargescalesininner
Galaxy,e.g.highturbulentvelocitiesofinterstellarmediuminthesestar-formation
regionsduetostellarwindsandsupernovaexplosions.
26spectralThestudiesINTEGRAL/SPIof26Alislineainpowtheseerfulstarspectr-formationometertoregionsprobecantheproAlvidelinenewshapes.inforSma-o
tionof26Alsourcesandtheirenvironments.
26CyIngnus,thisVelaChapterand,wSewillco-Cen,studywithAlthelinefour-yshapesearofSPIthreedatabasenearbystar(same-forwithmationthatinregions:§3).
Modelfittingsarealsocarriedouttoderivethe26Alspectrausingthespecifiedsky
region:eachformaps(1)splittheskydistributionmodelmapintotwoindependentsubmaps,onemap
coveringthedefinedstar-formationregion(Cygnus,VelaorSco-Cen)andtherestof
sky;the(2)modelfittingisappliedwithtwoindependentsubmapstofitthedata;
(3)twospectraforthedefinedstar-formationregionandtherestpartrespectively
arederivedsimultaneouslyusingmodelfitting.

105

426AlcessourandlineshapesinstarformationegionsrOB 8

2 kpc1 kpc0.5 kpc8.48.2OB 7OB 4OB 28.0OB 87.8y [kpc]7.67.47.2OB 9OB 3OB 17.00.00.51.01.52.02.53.0
x [kpc]

Figure4.1:DistributionofOandWRstars(redandblackstars)aswellasSNR(blue
frciromcles)aboandvetheOBgalacticassociationsplane(green(fromellipses)Plüschkeinthe2001).directionThesunofisCygnusallocatedasviewalonged
they-axisat8.0kpc.

regionCygnus4.1TheCygnusregionisoneofthemostactivenearbystarformingregionsintheGalaxy
withameanageof∼3Myr[Plüschke,2001].TheregionasdefinedbytheCygnus
126.8MeVemissionfeaturecontainsnumerousmassivestars.TheCygnusregionandits
AlemissionhavebeenstudiedanddiscusseddetailedlyinPlüschke’sthesis(2001).
Here,wewillsummarizedsomemainfeatureswhicharerelatedto26Alsourcesin
gnus.CyThegalacticOstarcataloguelists96Ostarsinthisfield[Garmanyetal.,1982].In
addition,onefinds23Wolf-Rayetstarsinthisregionofwhich14areofWN-type,8of
WC-typeandoneisclassifiedasWO-star[vanderHucht,2001].Between5and10of
theseWRstarsarebelievedtobemembersofOBassociationsinthisareaofthesky.
Inthepastmillionyears,supernovaexplosionswerealsoactiveintheCygnusregion.
TheGalacticSNRCataloguelists19remnantsinthisregion,for9ofthoseageand
distancehavebeenestimatedwithsufficientaccuracy[Green,2006].
Besidenumerousopenclusters,theregioncontainsnineOBassociations[Alteretal.,
1970,Plüschke,2001].Meanwhiletwoofthose(OB5&6)arethoughttobeartificial
associationduetoprojectionandselectioneffects[GarmanyandStencel,1992].
Figure4.1showsthedistributionofOandWRstars,SNRsandOBassociations
inthedirectionofCygnusasviewedfromabovethegalacticplane[Plüschke,2001].
ForSNRsthegivensizeofthecirclerepresentstheappearanceatthesky,whereas
theellipsesmarkingtheOBassociationhaveatwofoldmeaning.Theminoraxis
representstheappearanceinalongitude-latitudeprojectionandmajoraxisgivesthe
association.eachforuncertaintydistanceOpticalpopulationstudiestowardstheCygnusregionarelargelyhamperedbya

106

105II (deg)0b−5−10100

4.164.11−10 km/s < v <20 km/s−10 km/s < v <20 km/s55.9647.81OB839.66OB3OB7OB1OB9OB231.52115 GHz intensity23.3715.22OB47.08−1.0790lII (deg)8070

Cygnus01egionrFigure4.2:The115GHzCOemissiontowardsCygnus[Dameetal.,2001].Theellipses
markthepositionsoftheCygnusOBassociations.

giantmolecularcloudcomplexlyingatadistanceof500–1000pc[Dameetal.,2001].
Figure4.2showsavelocityintegratedintensityimageofCO115GHzemissionto-
wardstheCygnusregionreflectingthelargeangularextentoftheCygnusMolecular
Rift[Dameetal.,2001].Byanalyzingthe2MASSNIRdataofafieldcenteredon
CygnusOB2,Knödlseder[2000]hasfoundafactorofthreemoreOstarsthande-
terminedfromopticalobservations.Evenmore,CygOB2nowisthemostmassive
conglomerateofthestarsknownintheMilkyWayandresemblesmorelikeayoung
globularclusterthanatypicalOBassociation.Knödlseder[2000]hasfound120±20
Ostarand2600±400OBstarmembers.Theslopeofthemassdistributionwasfound
tobeΓ=1.6±0.1andtotalmassofthestellargroupisoftheorderof105M.The
overalllargedifferencesoftheinferredpropertiesofCygnusOB2dependingmoreor
lessontheimpactofthemolecularcloudsinfrontoftheassociationletoneexpect
similareffectsfortheotherOBassociationsfoundinCygnus.
The26AlfluxfromtheCygnusregionhasbeenreportedfromtheCOMPTELand
INTEGRAL/SPIobservations.FromtwoyearsofCOMPTELdata,delRioetal.[1996]
derivedafluxof(7.0±1.4)×10−5phcm−2s−1fromaskyregiondefinedas73◦<
l<93◦,−9◦<b<9◦,correspondingtoadetectionsignificanceof5σ.Fromthe
analysisof9yearsofCOMPTELdata,Plüschke[2001]foundafluxof(10.3±2.0)×
10−5phcm−2s−1fromaslightlylargerregioncovering70◦<l<96◦,−9◦<b<25◦.
WithearlySPIdata,wepresentedfirstresultsof26AllineemissionfromtheCygnus
region,andobserveda1809keVlinefluxof(7.3±1.8)×10−5phcm−2s−1fromasky
regiondefinedas73◦<l<93◦,|b|<9◦atasignificancelevelof4σ[Knödlsederetal.,
2004].The26AllinefromCygnusappearsmoderatelybroadening,withanintrinsic
FWHMof(3.3±1.3)keV.Butthisbroadfeatureisstillindispute,andneedsfurther

107

426AlcessourandlineshapesinstarGLAT (deg)

10

formationegionsrGLAT (deg)

10

GLAT (deg)GLAT (deg)GLAT (deg)1010105000-5-10-10-10908070708090708090
GLON (deg)GLON (deg)GLON (deg)

0

0

Figure4.3:ThreedifferentskymodelsfortheCygnusregion(65◦<l<95◦,−13◦<
b<17◦)usedinmodelfitting:theCOMPTELmaximumentropymap(left),the
COMPTELMREMmap(middle),thehomogenousdiskmodel(right).

studies.Herewere-analyzedthe26Allinelineshape26fromtheCygnusregionwithalargeSPI
database.Intherealanalysis,wederivetwoAlspectrasimultaneouslyusingmodel
fittingwithtwoskymaps,onecoveringtheCygnusregion,andtheothercovering
therestofthesky(includingtheinnerGalaxy).Threedifferentskymodelsastracers
of26AlinCygnusareusedinmodelfittingstoobtainthe26Alspectra.Comparison
amongthe26Alspectrafromdifferentskymodelsisaconsistentcheckforthe26Alline
shapesandfluxintheCygnusregion.◦
95◦,Figure−13◦4.3<bdispla<ys17◦)threetodifobtainferent26AlskyspectramapsusingforthemodelCygnusfittings:regionthe(65<COMPTELl<
maximumentropymap,theCOMPTELMREMmap,andahomogenousdiskmodel
employedfortheCygnusregion.Three26AlspectraforCygnusderivedfrommodel
26linefittingsemissionwithfrtheseomCythreegnusdifferentwithaskysignificancemodelsarelevelshoofwn∼in6σ.FigureThe4.4.reportedWe26detectAlfluxAl
doesnotvarysignificantlywiththedifferent−5models:(−27.3−±11.1)×10−5phcm−2s−1for
theCOMPTElMaxEntmap;(7.2±1.2)×10phcmsfortheCOMPTELMREM
map;(6.6±1.2)×10−5phcm−2s−1forthehomogenousdiskmodel.Thesefluxvalues
arestillconsistentwithpreviousresultsreportedbyCOMPTELandtheearlySPIdata
analysis.26Allineshapesobtainedfromthreedifferentmodelsarealsosimilar.Thelinewidth
valueisnear26theinstrumentallinewidthvalue(FWHM,∼3.1keV).Wealsoderive
theintrinsicAllinewidthbyfittingthespectradeconvolvedwiththeSPIspectral
response(see§3.2),andfindanintrinsicFWHM0.93±0.8226keVfortheCOMPTEL
alsoMaxEntintrinsicallymap,0.59narr±ow0.53,withkeVafortypicaltheISMMREMturbulentmap.SvotheelocityAlof≤line200frkmomsCy−1ingnusthisis
region.Fast-expandingbubblesarenotthedominatingISMcomponentsinCygnus.

108

2•10-5E=1808.81 FWHM=3.39 (±(±0.31)0.78)
)-1-5 keV1•10-1 s-20Flux (cm-5-1•101815181018051800Energy [keV] E=1808.93 FWHM=3.10 (±(±0.27)0.67)
-52•10)-1 keV-5-11•10 s-2Flux (cm0-5-1•101815181018051800Energy [keV] 0.34)±(E=1808.76 2•10-5FWHM=3.27 (±0.85)
)-1 keV-51•10-1 s-20Flux (cm-5-1•101815181018051800Energy [keV]

1820

1820

1820

4.1CygnusegionrFigure4.4:26AlspectraforCygnusderivedfrommodelfittingsusingthreedifferent
skymodels:theCOMPTELMaxEntmodel(top),withthederived26AlfluxF∼
(7.3±1.1)×10−5phcm−2s−1fortheregionof65◦<l<95◦,−13◦<b<17◦;
theCOMPTELMREMmap(middle),F∼(7.2±1.2)×10−5phcm−2s−1;the
homogenousdiskmodel(down),F∼(6.6±1.2)×10−5phcm−2s−1.

109

426AlcessourandGLAT (deg)10

5

0

-5

lineshapesinstarformationegionsrGLAT (deg)10

5

0

-5

-10-10255260265270275260265270275
GLON (deg)GLON (deg)

Figure4.5:TwodifferentskymodelsfortheVelaregion(−104◦<l<−82◦,−10◦<
b<10◦)usedinmodelfitting:theCOMPTELMREMmap(left),thehomogenous
(right).modeldisk

EvenFWHMofconsidering∼3.3keVuncertaintiesreportedbyofthetheearlyFWHM,SPIthedatabroadanalysislineisnotfeatureheldwithupanbyintrinsicpresent
results.Thevalueofthe26AllinecentroidenergyinCygnusisconsistentwiththelaboratory
value.indicated,ForbutthethecaseenerofgytheshiftMREMisstillwmodel,ellawithinsmalltheblueshiftuncertaintiesof∼(00.27.27kekeVV)maofythebe
bulkmeasurements.motionsof26AlTherefore,sourcesweinCyconcludegnustofrwomardorpresentforwarobserdus.vationsthatthereareno

regionelaV4.2Inemissionthedirectioncenteredtowatarld∼the−100V◦ela,b∼region,−2◦,theretheGumexistsaNebula,hugewithextendedanangularstructureofdiameterHα
◦of300∼–36450[pc[ChanotBrandtandetSival.,an,19711983,].FrancoThe,1990mean].ThedistanceGumoftheNebulanebulamayisbeanapprinterstellaroximately
2systembubblebloinsidewnbtheystellarnebula[Wwindseaverofettheal.,bright1977],O5orIaanstarζexpandingPuppisSNRandwiththeγanVageelorumof1
Myr[Reynolds,1976].
(GThe263.9-3most.3).prThisominentSNRisastructureremnanttowarofdatheskycore-collapseoftheGumsupernovNebulaaisexplosiontheVelaaboutSNR∼
4101981y],earswithagoadistanceestimatedofby∼the200−spin-do400pcwn[rateChaofetal.the,V1999ela,pulsarGreen[,2006Manchester].ThesoftandTX-raaylory,
7.3obser◦[vationsAschenbachby,ROSA1993T].revThisealednearbthatytheangularcore-collapsediameterSNRisofthetheVmostelaprSNRomisingisat26leastAl
sourcetowardtheVelaregion.

110

E=1809.18 FWHM=4.49 (±(±0.60)1.60)

4.2elaVr-52.5•10E=1808.92 (±0.88)E=1809.18 (±0.60)
FWHM=5.47 (±2.87)FWHM=4.49 (±1.60)
-52.0•10-52•10))-1-1-51.5•10 keV keV-5-1 s1.0•10-5-1 s1•10
-2-2-65.0•10Flux (cmFlux (cm00-5-1•10-6-5.0•10180018051810181518001805181018151820
Energy [keV]Energy [keV]

egionFigure4.6:26AlspectraforVeladerivedfrommodelfittingsusingtwodifferentsky
models:theCOMPTELMREMmap(left),F∼(4.1±2.7)×10−5phcm−2s−1for
theregionof−104◦<l<−82◦,−10◦<b<10◦;thehomogenousdiskmodel
(right),F∼(3.9±2.2)×10−5phcm−2s−1.

E=1808.81 FWHM=3.77 (±(±0.46)0.00)

-52.5•10E=1808.62 (±0.57)E=1808.81 (±0.46)
FWHM=3.77 (±0.00)FWHM=3.77 (±0.00)
-52.0•10-52•10))-1-1-51.5•10 keV keV-5-1 s1.0•10-5-1 s1•10
-2-2-65.0•10Flux (cmFlux (cm00-5-1•10-6-5.0•10180018051810181518001805181018151820
Energy [keV]Energy [keV]

Figure4.7:26AlspectrafittedwithfixedGaussianwidthforVeladerivedfrommodel
fittingsusingtwodifferentskymodels:theCOMPTELMREMmap(left),F∼
(3.7±1.2)×10−5phcm−2s−1fortheregionof−104◦<l<−82◦,−10◦<b<
10◦;thehomogenousdiskmodel(right),F∼(4.1±1.2)×10−5phcm−2s−1.

111

426AlcessourandlineshapesinstarformationegionsrBesidesthelargeVelaSNR,severalsmallSNRsarealsodetectedintheVelaregion,
e.g.PuppisA(G260.4-3.4),G261.9+5.5,andG272.2-3.2.ThedistancetoPuppisAis
about2kpcbasedonobservationsofradialvelocitiesandpropermotionsofopticalfil-
aments[Winkleretal.,1988],whichmaybealsoacore-collapseSNR[HuiandBecker,
2006].G261.9+5.5isquitefaintonewithanestimateddistanceof∼2.9−4.8kpc
[ColombandDubner,1980].G272.2-3.2discoveredbyROSATisapossibleveryyoung
SNR(anestimatedage∼1800years)withadistanceofabout1.8kpc.Thediscovery
ofthenearestyoungsupernovaremnantRXJ0852.0-4622/GROJ0852-4642inthe
GalaxytowardthedirectionoftheVelaregionbyROSATandCOMPTELhasalso
beenreported[Iyudinetal.,1994,Aschenbachetal.,1999].Ageanddistanceareesti-
matedtobe∼680yearsand∼0.2kpcbytheX-raydiameterandthegamma-rayline
fluxofradioactive44Ti[Aschenbachetal.,1999].Theserecentcore-collapsesupernova
arealsogoodcandidate26AlsourcesintheVelaregion.Inaddition,therearepossible
unresolvedSNRsinsidethegiantmolecularcloudstowardthedirection,so-calledthe
VelaMolecularRidge[MurphyandMay,1991]whichisthemostmassivestructurein
thisdirectionatadistanceof∼1−1.5kpc,alsowithalargenumberofOstars,IR
sources,dustemission,andthebrightestHIIregionofthesouthernsky.
AnOBassociationVelOB2existsinthisregion,includingmanyyoungstarswitha
meandistanceof∼410pc[deZeeuwetal.,1999].Themassivestarbinarysystemγ2
VeloruminthisgroupcontainsWR11,thenearestknownWolf-Rayetstaratd∼260
pc[vanderHuchtetal.,1997].
Inaword,theVelaregionincludessomeactivenearbystarformationregions.26Al
emissioncanbedetectedtowardthisregion[Oberlacketal.,1994].COMPTELhas
detectedthe1809keVgamma-raylineemissionfromtheVelaregion,whichmaybe
associatedwiththeVelaSNR[Diehletal.,1995b]orγ2Velorum[Oberlacketal.,2000].
Themeasured26Alfluxis(3.6±1.2)×10−5phcm−2s−1.
Herewestudythespectralinformationof26AllineintheVelaregionwithSPI.The
26Alsourcecandidatesdescribedabove,e.g.,theVelaSNR,WR11,otherSNRs,are
locatedinthedirectionof−104◦<l<−82◦,−10◦<b<10◦.Sowesplitthesky
modelintotwoparts,onecoveringtheVelaregion(−104◦<l<−82◦,|b|<10◦)
andtherest.Thenwecanderivethetwo26AlspectraofVelaandtheinnerGalaxyby
modelfittingusingtwoskymapssimultaneously.
TwodifferentskymodelsfortheVelaregionhavebeenusedinmodelfittingfora
comparison(seeFigure4.5):theCOMPTELMREMmapandthehomogenousdisk
model.26AlspectrafortheVelaregionderivedfrommodelfittingusingtwodifferent
skymodelsareshowninFigures4.6and4.7.
InFigure4.6,wefitthe26AlspectrawithaflatcontinuumbackgroundplusaGaus-
sianlineprofilewithfreeparametersoflinepositionandwidth.Thedetectionisnot
significant(∼2σ),andthemeasured26AlfluxforVelais(4.1±2.7)×10−5phcm−2s−1
fortheCOMPTELMREMmap,andis(3.9±2.2)×10−5phcm−2s−1forthehomoge-
nousdiskmodel.Thederived26AlfluxforVelabySPIisstillconsistentwiththat
obtainedbyCOMPTEL[Diehletal.,1995b]thoughthe26AlsignalforVelaisweak

112

4.3Sco-Cenegionrwiththepresentobservations.TheGaussianFWHMisdeterminedat∼4.5−5.4
keV,iftakingtheinstrumentallinewidthFWHM=3.18keV[Kretschmer,2007],the
derivedintrinsic26AllinewidthisFWHM>3keV.Itcouldbeabroadlinefeature
thoughthesignalisweakwithlargeerrorbarsforwidthinfit.
Toreducetheeffectofthelargeerrorbarsforthelinewidthonthe26AlfluxinGaus-
sianfit,wecanletthelinewidthbeafixedparameterinGaussianfittolineprofiles.
InFigure4.7,wepresentthe26AlspectrafittedwithfixedGaussianwidth(FWHM=
3.77keV)plustheflatcontinuum,whereweassumetheintrinsic26AlwidthinVela
asFWHM∼2keV.Thederived26Allinefluxis(3.7±1.2)×10−5phcm−2s−1forthe
COMPTELMREMmap,andis(4.1±1.2)×10−5phcm−2s−1forthehomogenousdisk
model,withasignificancelevelof∼3σ.ThefluxisconsistentwithCOMPTELresults,
butislowerthanthereportedvaluebytheearlyanalysisofSPIdata[Schanneetal.,
2007].The26Allinecentroidenergyiswellwithinthelaboratoryvalue.
Wehaveassumedanintrinsic26Allinewidthof2keVfortheVelaregionintheline
fit.Thebroad26AllinefeaturefortheVelaregionindicatesthattheenvironmentand
kinematicsof26AlsourcesinVelacanbequitedifferentfromthoseinlargescalesof
theinnerGalaxy.IntheVelaregion,severalrecentsupernovaexplosions,andstellar
windsfromWR11couldleadtoexpandingbubbleswhichprobablydominatethe
interstellarmediumof26Alsourcesfrompresentobservationlimitsonthelineprofile.
Thenaverageturbulentvelocitiesof26Alsourcesmayreach∼200−300kms−1inthe
Velaregion.Anyway,thisissueneedthefurtherstudieswithmoreSPIdata.

Sco-Cen4.3region

Thenearestsiteofmassivestarformation,theScorpius-Centaurus-Lupusregion(hereafter
theSco-Cenregion),isabout100–150pcfromtheSun.Nearbyyoungstarsareseen
inSmostlyco-CenintheOBSoutherassociationnthatHemisphere,consistswhichofisthreerelatedsub-regions:torecentUppermassivSecorstarpiusfor(US),mationUp-
byperdifferentCentaurusskyLupuspositions,(UCL),ageand[deLowGeuseretal.,Centaurus1989],Cruxand(LCC),kinematicseach(deZeeuwdistinguishableetal.
1999,seeFigure4.8).
Theclosestofthethreesub-regionsisLowerCentaurus-Crux(LCC)withanesti-
10–mated20Myrdistance[deofGeus∼et120al.pc,[1989de].ZeeuwLCCetisal.,located1999].towTheardagetheofLCCdirectionisofthoughtthetoGalac-be
ticplane(−70◦<l<−40◦),andmaycontributeto26Allineemissionobservedby
COMPTELandSPIintheplane(see§3.6).
USandUCLarelocatedatlatitudesabovetheGalacticplane.Themeandistance
of1989].UCLUCisis∼the140ypc,oungestwithanoneagewithof∼an10ageMyrof∼[de5−10ZeeuwMyret.al.Its,1999mean,dedistanceGeusetisal.∼,
145regionpc[isdelocatedZeeuwetnearal.,the1999center,deofGeusthisetgral.,oup1989].(FigureThe4.8r).hoUCisOphiuchusjuststarlocatedforinmingthe

113

426AlcessourandlineshapesinstarformationegionsrFigure4.8:Distributionofyoungstarsincludingpre-mainsequencestars(PMS)
andyoungearly-typestarsinSco-Cenregion(fromSartorietal.2003).Three
sub-groups,UC,UCLandLCCarenotifiedwiththedashed-lineboxes.The
ChamaeleonOBassociationcouldbeanextensionoftheSco-Cenone.Thepo-
sitionsoftherhoOphiuchusactivestarformationregionandtheLupuscloud
marked.alsoarecomplex

GLAT (deg)30

25

20

15

10

5

GLAT (deg)30

25

20

15

10

5330340350330340350
GLON (deg)GLON (deg)

Figure4.9:TwodifferentskymodelsfortheSco-Cenregion(−35◦<l<−5◦,5◦<
b<30◦)usedinmodelfitting:theCOMPTELMREMmap(left),thehomogenous
right).(modeldisk

114

4.3Sco-Cenr -54•10E=1810.19 (±0.21)3•10-5E=1810.44 (±0.23)
FWHM=2.46 (±0.50)FWHM=2.37 (±0.55)
)3•10-5)2•10-5
-1-1 keV2•10-5 keV1•10-5
-1-1 s s-2-20-51•10Flux (cmFlux (cm-5-1•100-5-2•101800180518101815182018001805181018151820
Energy [keV]Energy [keV]

egion26Figuresky4.10:models:AlthespectraCOMPTELforSco-CenMREMderivmapedfr(left),omF∼model(6.7±fittings2.0)×using10−5twphocmdif−2ferents−1
◦◦◦◦for(right),theFregion∼(6.0of±−1.935)×<10l−<5ph−5cm,−52s−<1b.<Both30;spectrathehavehomogenousshowndiskthesignifi-model
cantblueshiftsof∼1.5keVrelativetothe26Allinecentroidinlaboratory.
directionabovetheGalacticcenter(l∼−5◦,b∼18◦),sothe26Alemissionstructurein
thisdirectionobservedbyCOMPTEL(seeFigure1.6,Plüschke2001)ismostprobably
attributedtotheUCOBassociation.
2inT12CO,achiharaandetreal.v[ealed2001]thehavesurmolecularveyedgasmorethandistribution500indegthisaroundregion.theAStotalco-Cenmassregionof
4In∼10addition,Mandthere∼is100anofsmallexpandingcloudsHIhashellvebeencenteredindetectedUpperas-Stheco,LupusandthecloudLupuscomplex.cloud
complexexistsbetweentheshellandtheUpper-Cen-Lup.Theyoungstarclustersare
locatedattheedgeoftheshell,whichsuggeststriggeredstarformation.
Here,wewillstudythespectralinformationof26AlsourcestowardtheSco-Cen
inregionlatitudeswithofSPI.b>The5◦,studiedwhichregioncouldbeconcentratesresolvedonfrtwomogrthe26oups,AlUSandcontributionsUCL,oflocatedthe
26Galacticsubmaps:disk.oneforToSderivco-CenethedefinedAlasspectra−35◦of<Sl<co-Cen−5◦,in5◦<modelb<30fittings,◦andwetheuseresttwofo
themapskyandtothefitthedata.homogenousTwoskydiskmodeldistribution(Figuremodels4.9).areThe26applied:AlspectratheforSCOMPTELco-CenfrMREMom
atwoskysignificancemodelslevareelof∼presented4σ.inTheFiguremeasured4.10.Wfluxehaisv(e6.7±detected2.0)×26Al10−fr5omphScm−co-Cen2s−1withfor
theCOMPTELMREMmapand(6.0±1.9)×10−5phcm−2s−1forthehomogenous
diskmodel.The26Allinedetectedtowardthelatituderegionof−60◦<l<0◦,5◦<
b<20◦in26Figure3.31(§3.7)wouldbeattributedtoSco-Cen.
BoththeAlspectrafromtwoskymodelsshowthesignificantblueshifts.Theline
centrreflectsoidthehasquiteablueshiftinterestingof∼1.5dynamics−1.7ofkeVthe26relativAleejectatointhetheSlaboratorco-Cenyvalue.region.ThisThe26shiftAl

115

426Alsourcesandlineshapesinstarformationregions
26ofsour200ces±in120Skmco-Cens−1aretowarddominatedus.OurbythestudiescomponentimplyofthatAlthevejectaerywithnearbay26bulkAlvsourelocityces
haveaquitedifferent26behaviorindynamicsfromthelargescalefeatureintheinner
Galaxy.ThesenearbyAlsourcesarestronglyaffectedbystellarwindsofmassive
starsandejectaofsupernovaexplosions,andarealsorelativelyeasilybeendetected
andresolvedbythepresentobservations.
116

5Diffuse60FeemissionintheGalaxybySPI

660Felocatedisanattheunstableneutronnucleusrichsidewhoseofthe“vterrestrialalleyofhalf-lifestableisisotopes”1.5×(also10yseeears,Figureand1.it11is).
The60Feisotopeissynthesizedinneutroncapturereactionsfromthe56Feisotope
whichisabundantfromformerequilibriumnucleosynthesisof56Nianditsdecay,
Suchreactions,s-pre.g.,ocess13isCandexpected20Netoαoccurcaptures,instellarhenceinregionstheO/Newithburefningficientshellneutrandon-liberatingbottomof
Heburningshellofcore-collapsesupernovae,andtheHeburningshellinsidemassive
stars(moredetailsin§1.7).
Thedecaychainsof60FeareshowninFigure5.1.60Fefirstlydecaysto60Co,with
emittingγ-rayphotonsat59keV,andthendecaysto60Ni,withemittingγ-rayphotons
ofat1173thoseandat11731333keandV.1333ThekeV,gamma-rasotheyefgamma-raficiencyyoffluxtheat5959keVkeVistransitionmuchisloonlywer∼than2%
thefluxesofthehighenergylines.The59keVgamma-raylineisverydifficulttobe
detectedwithpresentmissions.Measurementsof60thetwohighenergylineshavebeen
themainscientifictargettostudytheradioactiveFeisotopeintheGalaxy.
[60FeShukolyukohasvbeenandfoundLugmairto,be1993part].ofTheinferredmeteorites60Fefor/med56Feinratiotheforearlythesesolarmeteoritessystem
exceededtheinterstellar-mediumestimatesfrom60nucleosynthesismodels,whichled
tosolarsuggestionsystem[TthatachibanathelateandsuperHussno,va2003,ejectionTachibanaofFeetal.,occurred2006].Ybeforeet,thisforisamationproofofforthe
cosmic60Feproduction,accelerator-massspectroscopyofseafloorcrustmaterialfrom
thesouthernPacificoceanhasrevealedan60Feexcessinacrustdepthcorrespond-
ingtoconcludedanagethatofa2.8superMyrno[vaKnieetexplosional.,2004ev].entFrnearomthethissolarinterestingsystemoccurredmeasurement,aboutitis3
Myrago,depositingsomeofitsdebrisdirectlyintheearth’s60atmosphere.Allthese
measurementsbasedonmaterialsamplesdemonstratethatFenecleosynthesisdoes
throccuroughinnature.detectingItisnoradioactivwinterestinge-decaytoγ-rasearychlines.forcurrent60FeproductionintheGalaxy
Manyexperimentsandeffortsweremade(alsoseeTable5.2)tomeasurethe60Fe
60thegamma-ra60Feyfluxfromemission.theTheinnerfirstGalaxydetectionregionof(−30Fe◦<lineslw<as30pr◦)oisvided∼(b5.3y±HEAO-4.3)×3,
10−5phcm−260s−1rad−1[Mahoneyet−5al.,1982−].2−The1SMM−1Gamma-RaySpectrome-
terreportedaFefluxof2.9±2.510phcms60rad[Leisingand−5Share,1994−2].
s−1OSSErad−aboar1[dHarristheetal.COMPT,1994ON,1997Obser].vatorAndygaveCOMPTELaFefluxaboarofd6.3the±4.5COMPT10ONphcmObser-

117

5Diffuse60FeemissionintheGalaxybySPIFigure5.1:Thedecayschemeof60Fe.Themeanlifetimeis2×106years.Thegamma-
rayfluxat59kevlineis∼2%ofthoseat1173and1333keV.

vatoryalsoreporteda60Fefluxlimitof1.2×10−4phcm−2s−1rad−1(2σ,Diehletal.
1997).TheGamma-RayImagingSpectrometer(GRIS)reportedanupperlimitforthe
60Fefluxof6.8×10−5phcm−2s−1rad−1(2σ,Nayaetal.1998).Recently,RHESSI
reportedobservationsofthegamma-raylinesfrom60Fewithasignalof2.6σsig-
nificance,andanaveragefluxof(3.6±1.4)×10−5phcm−2s−1[Smith,2004b]from
theinnerGalaxy.TheanalysisofthefirstyearofdatafromtheSPIspectrometer
onINTEGRALspacecraftresultedinasimilarlymarginally-significantdetectionof
theseγ-raylinesfrom60Fe(∼3σ,Harrisetal.2005),withanaveragelinefluxof
(3.7±1.1)×10−5phcm−2s−1fromtheinnerGalaxy.
SPIhasaccumulatedmoredatasincethen,hereweanalyze3y60earsofdata,aiming
ataconsolidationoftheINTEGRAL/SPImeasurementoftwoFegamma-raysat
1173and1333keV.

analysisandpreparationData5.1

AsdescribedinChapter2,weestablishdatabasesforsingleandmultipleevents(SE
andME),inspectralrangeswith1keVbinsizein∼20keVbandsaroundeachofthe
60Felines,andinadjacentenergybandsfordeterminationofinstrumentalbackground.
Forthescienceanalysesoftwo60Felines,weusethedatawiththeenergyranges
aroundthetwoγ-raylinesextendfrom1153-1193keVand1313-1359keV.Theraw

118

5.1DataeparationprandanalysisdataspectraaroundtwolinesarepresentedinFigure5.2.Stronginstrumentallines
spectra.wrathedominateForthebackgroundmodellingfromtheadjacentcontinuum,weuse1163-1169
keVplus1177-1184keV,and1318-1328keVplus1336-1349keVenergybands,
respectively.Forthecaseof1333keVline,sincethestronginstrumentallineat1337
keVblendswiththe60Feline,wehaveincludetwostronginstrumentallines(at1337
and1347keV,bothproducedby69Geelectroncaptures,seeFigure5.2)intheselected
adjacentenergybands.Ouraimsaretoreducethecontaminationoftheinstrumental
lineat1337keV.Fromorbits17to359,atotalobservationtimeof∼24Msisthus
obtained,withadatasetconsistingof14623pointingswithonespectrumpereach
detectorandeventtype.
Inbackgroundmodelling,firstthedetector-by-detectorcountratesinthefourcon-
tinuumbandsarefittedbytheGEDSATtimeseriestoconstructan‘adjacent-energies’
backgroundtemplateforeach60Feline.Othercomponentsmodellingtheinstrumental
required.alsoarecontributionlineThemostimportantradioactivebackgroundlinecomponentforouranalysisisthe
emissionfrom60Codecayinsidetheinstrument,whosetwolinesat1173and1333keV
areactuallypartofthecelestial60Fedecaychainwewanttomeasure.Theconvolution
oftheGEDSATCRsourcetracerwithanexponentialfunctione−t/τ0(see§2.3),in
whichthedecaytimeτ0isthatof60Co(7.6yr),isexpectedtogiveagoodbackground
modelforenergybinscontainingcontributionsfromtheselines,i.e.around1173and
1333keV.Asecondpossibleradioactivecontaminantmaybethestrong69GeK-shell
electroncaptureline(1337keV)whichblendsintothe1333keV60Feline;itslifetime
is2.35days.Becausethedecaytimeisshort,theconvolutionoftheGEDSATtracer
withtheexponentialfunction(τ0=2.35days)issimilartotheGEDSATtraceritself,
thereforewejusttaketheGEDSATratesasabackgroundcomponenttracerforthis
feature.lineinstrumentalongstrThenthesetofdetector-by-detectorspectraperpointingin1keVbinscoveringthe
∼20keVintervalsaroundandincludingthe60Felinesarefittedtothesumofadjacent
continuumtemplate,plustheradioactivitytemplatefor60Co,plusanadditionalGED-
SATtemplatetocaptureanadditionalpromptbackground.Finally,weorthogonalize
thethreedifferentbackgroundmodelcomponentsforanimprovedconvergenceof
thefitting(see§2.3).Wedisplayexamplesofthreebackground-componentspectrafor
boththe1173keVlineandthe1333keVcases(Figure5.3).Allthesethreebackground
spectrashowsignificantlinefeaturesaroundthe60Felineenergyranges,confirming
newandindependentcontributionofeachbackgroundcomponenttothemodelfit-
ting.SeparatemodelfittingsaremadeforeachoftheFelinesandforSEandME,respec-
tively.TheresidualsafterthemodelfittingareshowninFigure5.4(residualswiththe
time)andFigure5.5(withtheenergies)foreachofthetwolines.Representingour
datasetswithbackground-modelsonly,wethenobtainreducedχ2valuesof1.185and
1.194forthe1173and1332keVlinebands(SE,277824d.o.f),and0.665and0.663for

119

5Diffuse60Feemissioninthe36003400Counts32003000

2800

450040003500Counts3000

GalaxybySPIdet=0,18pnt=(88.0,-2.1)

11601170Energy (keV)118011901200

det=0,18pnt=(88.0,-2.1)

2500132013301340Energy (keV)13501360

Figure5.2:Rawdataspectraaroundtheenergiesoftwo60Felinesinone-INTEGRAL-
orbitobservations(3days),representingtheinstrumentallinesandcontinuum
obbackgrvious:ound.44Sc(For1157thekeV),117360keCoV(case1172.9(abokevV),e),182threeTa(str189.4ongkeV),instrumentalandthe60linesCoareline
blendswiththe60Feline[Weidenspointneretal.,2003].Forthe1333keVcase
60line(beloofw),69Gethe(1336Co.8kebackgrV)oundblendlinewith(1332the.560FekeV)line,andatheinstrumentalotherstronglineat1347instrumentalkeV
alsocomesfrom69Geelectroncaptures[Weidenspointneretal.,2003].

120

5.1Datapreparationandanalysis
1.05E=1173.11(0.56)FWHM=-1.27(2.50) 2.52.0E=1172.30(0.16)FWHM=2.42(0.39) 0.2E=1172.22(0.08)FWHM=2.40(0.20)
1.040.11.51.031.00.0Intensity (ref units)1.021.01Intensity (ref units)0.50.0Intensity (ref units)-0.1
-0.51.0011651170Energy (keV)1175118011651170Energy (keV)11751180-0.211651170Energy (keV)11751180
1.05E=1173.00(Inf)FWHM=2.76(Inf)3E=1332.00(0.08)FWHM=1.35(0.31)0.12E=1333.21(0.21)FWHM=2.36(0.51)
0.101.0420.081.0310.061.02Intensity (ref units)1.011.00Intensity (ref units)0-1Intensity (ref units)0.040.02
0.9913251330Energy (keV)1335134013251330Energy (keV)133513400.0013251330Energy (keV)13351340
Figure5.3:Spectraofthreebackgroundcomponentsforthe1173keVlineandthe
601333CokeVradioactivitycases:adjacent-build-upener(right).gy-bandThelinetemplateintensity(left);isgivGEDSAenTinrelativtracer(middle):eunits
fittings).inficients(coefthe1173and1333keVlinebands(ME,2615038d.o.f.),respectively.Notethatforthelow
numberofcountsinMEspectra,χstatisticsdoesnotapplyhere.Therefore,inour
modelfittingapproach(bothforSEandMEcases,see§2.5),wehaveusedPoissonion
function.likelihoodtheandstatisticselsareThesefitsadequate,producetheskycontinuumbrightnessenergyamplitudesbinsoutsideperenerthegy60Febin.linesIfbackgrshouldoundshowmod-sky
amplitudesstatisticallyconsistentwithzeroskybrightness.Non-zeroskybrightness
width,shouldshopossiblywupbrinoadertheforifmtheoflinescelestialwithlineaenershapegyisconforbrmingoadenedtothealready(asinstrumentalhadbeenline
discussedfor26Al,seeChapter3andDiehletal.2006c).Duetothelowintensity
ofexpected60Feemission,60wedonotattempttoderivelineshapeinformationatall,
talandwidthrathertothederivethespimodfitFeresultlinespectra,amplitudedeterbyfittingminingaamplitudesGaussianwithandlinefixedcentrinstrumen-oidsin
thisway.Fromstrongnearbyinstrumentallines(e.g.theinstrumentallineat1107
kenarrV),owwelinesdeterinminethethatregimeaofGaussianthe60Fewithlines;awwidtheuseof2.this76keVshapeistoandeteradequatemine60modelFelinefor
parametersfromthespimodfitresultspectra.Foraconsistentcelestialsignalfrom60Fe,
thelinesshouldbearoundthe60Fedecayenergiesof1173and1333keVandequal
(withinerrors)inamplitudesforallfourdatasets.
121

5Diffuse60FeemissionintheGalaxybySPIDet 00 - 180.050.040.030.020.010-0.01-0.02-0.03-0.04160014001200Det 00 - 180.060.040.020-0.02-0.04-0.06

160014001200

1163.00 - 1182.00 keV

200018001323.00 - 1342.00 keV

20001800

Figureunits5.4:ofcountsResidualss−of1).Thecountsupperafter(1173fittingkeinV)difandferentlowprer(1333ojectionskeV)forSEfiguresdatabasesshowthe(in
residualsversustime(daysoftheJulianDate,startsat1Jan2000),wherecounts
zerperoconfirpointingmhathatveourbeenbackgrre-binnedoundinto3models-dayareintervalsadequate,forwithclarity.χ2/d.oResiduals.f.inarmodelound
fittings:1.182(SE)and0.670(ME)forthe1173keVline;1.191(SE)and0.667(ME)
forthe1333keVline.

122

0.20.0Residual Counts-0.2-0.4-0.6

0.60.40.2Residual Counts0.0-0.2-0.4-0.6

1165

1325

5.1DataprDet=0-18MJD=1073.4-1433.4MJD=1433.4-2089.0

118011751170Energy (keV)

Det=0-18MJD=1073.4-1438.4MJD=1433.4-2089.0

134013351330Energy (keV)

eparationandanalysisFigure5.5:Theupper(1173keV)andlower(1333keV)figurespresenttheresiduals
(inunitsofcountss−1keV−1)projectedwiththeenergyforSEcases.Forthe1333
keVlinecase,becauseofthestronginstrumental69Gelineat1337keV,therestill
existthelargeresidualsaround1337keV.Thesolidlinerepresentsresidualsfor
theearlydata(e.g.,fromtheJulianDate1074–1434),andthedashedlineforthe
recentdata(fromtheJulianDate1434–2089).

123

5Diffuse60FeemissionintheGalaxybySPI5.2Observationsoftwo60Felinesat1173and1333keVin
Galaxythe5.2.1Detectionsof60FefromtheinnerGalaxy
26Aland60Fewouldshareatleastsomeofthesameproductionsites,i.e.massivestars
andsupernovae(alsosee§1.6and§1.7).Inadditionbotharelong-livedradioactive
isotopes,sowehavegoodreasonstobelievetheirgamma-raydistributionsaresimilar
aswell.Thereforewecanadopttheskydistributionof26Algamma-raysasourbest
modelforcelestial60Fegamma-raydistribution.Thus,ourfittedmodelconsistsofthe
skyintensitydistributionof26Alfrom9-yearCOMPTELobservationsasastandard
model(maximumentropyimage,Plüschkeetal.2001),togetherwiththebackground
modelbasedonmeasurementsinenergybandsadjacenttothe60Felinesandon
backgroundtracers(seeabove).FromindependentlyfittingtheSEandMEdatasets
at1keV-wideenergybinsintheenergyrangesofthetwolinesof60Feemission(1173
keVand1333keV),weobtainthefourspectrafromspimodfit(§2.5)showninFigure
5.6.Wefindexcessamplitudesforthecelestialcomponentineachofthefourspectra.
Residualbackgroundimperfectionsaresmallexceptfortheinstrumentallineat1337
keV.Thecontinuumfluxlevelsareclosetozerowithin10−5phcm−2s−1rad−1keV−1,
withoutapparentenergydependence.Addingalinearslopetoourspectralfitfunction
doesnotimprovethefit.Furthermore,thediffusegamma-raycontinuumemission
around60FelineenergiesintheinnerGalaxyis∼2×10−6phcm−2s−1rad−1keV−1
fromCOMPTELmeasurements[Strongetal.,1999],whichiswellbelowtheerrorbars
ofspectralfluxesinFigure5.6.
The69Gelineat1337keVhasalsobeeneliminatedratherwell,thoughnotcom-
pletely(aboutafactof4comparedwithHarrisetal.2005).However,thisstrong
instrumentallinestillleadstoaspectralsignalwhichisrecognizedclearlyasanarti-
factandthusnotconfusedwiththecelestialline.
Wethendeterminethe60Felineparametersinthesespectrafittingacontinuumand
GaussiansrepresentingtheSPIspectralresponse(§2.4).ThewidthoftheGaussian
profileshasbeenfixedat2.76keV.Wethusobtain60Felinepositionsandintensities
fromfittingthefourspectraofFigure5.6,wherethetotallinefluxforeachlineinthe
innerGalaxyregion(e.g.−30◦<l<30◦and−10◦<b<10◦)isdeterminedfromthe
intensityofourfittedGaussiansandfromthenormalizationoftheinputskymap.We
presentlinefluxesoftwo60FelinesfortheSEandMEdatainTable5.1,respectively.
Alllinefluxvaluesareconsistentwithinuncertainties.Anaverage60Fefluxofthese
fourspectraisabout4.6×10−5phcm−2s−1rad−1.
AsuperpositionofthefourspectraofFigure5.6isshowninFigure5.7.Lineenergies
ofthe60Felinesinthelaboratoryare1173.23and1332.49keV.Sincetheenergybinfor
the60Fedatais1keV,wecannotsuperposetwolineswiththelinecentroidenergies
exactlyatthelaboratoryvalues.Forthissuperposition,wethereforedefinethezero
oftherelativeenergyaxisat1173and1333keV,toderivethesummedspectrumof

124

5.2Observationsoftwo60Felinesat1173and1333keVintheGalaxy
1-]0.2FEW=1H1M73=.25.676(±(0±.05.70)0)SE1-]EF=W13H3M3=.02.97(6±(0±.60.50)0)SE
VeI=0.42(±0.16)Ve0.1I=0.35(±0.15)
1-kd1-kd
aa1-rs0.11-rs0.0
22--mc mc
hh4-p 0.04-p -0.1
001[ y1[ y
ttiisn-0.1sn-0.2
nIetIetn
3.0-11651170117511801325133013351340
Energy [keV]Energy [keV]
4.01-]0.3FEW=1H1M73=.23.076(±(0±.04.70)0)ME1-]FEW=1H3M32=.23.876(±(0±.05.90)0)ME
VekI=0.58(±0.19)Vek0.3I=0.52(±0.21)
11--dar0.2dar
1-2-s1-2-s0.2
mc 0.1 mc
hh4-p 4-p 0.1
001[ y0.01[ y0.0
tisist
nnetnI-0.1etnI-0.1
11651170117511801325133013351340
Energy [keV]Energy [keV]
Figure5.6:Thespectraoftwogamma-raylinesof60FefromtheinnerGalaxy:1173
keVand1333keV(fromWangetal.2007).Wehaveshowntheresultsboth
fromSEandMEdatabases.ThedatapointsarefittedwithGaussianpro-
filesoffixedinstrumentalwidth(2.76keV),andfixedcontinuumslope(flat),
withχ2/d.o.f.offits:1.06(SE)and1.09(ME)forthe1173keVlineand
1.16(SE)and1.11(ME)forthe1333keVline.FortheSEdatabase,wefind
alinefluxof(4.2±1.6)×10−5phcm−2s−1rad−1forthe1173keVlineand
(3.5±1.5)×10−5phcm−2s−1rad−1forthe1333keVline.FortheMEdatabase,
thelinefluxis(5.8±1.9)×10−5phcm−2s−1rad−1forthe1173keVlineand
(5.2±2.1)×10−5phcm−2s−1rad−1forthe1333keVline(alsoseeTable5.1).
125

5Diffuse60FeemissionintheGalaxybySPI-1]E=0.07 (±0.30)
-1keV0.20I=0.44 FWHM=2.76 (±0.09)(±0.00)
rad-10.15s-2 ph cm0.10-40.05Mean Intensity [100.00-0.05

-5Energy offset from 060Fe line [keV]5

Figure5.7:Thecombinedspectrumofthe60FesignalintheinnerGalaxy,superimpos-
ingthefourspectraofFigure5.7(fromWangetal.2007.Inthelaboratory,theline
energiesare1173.23and1332.49keV;heresuperimposedbinsarezeroat1173and
1333keV.Wefindadetectionsignificanceof4.9σ.Thesolidlinerepresentsafitted
Gaussianprofileoffixedinstrumentalwidth(2.76keV),andaflatcontinuum.The
averagelinefluxisestimatedas(4.4±0.9)×10−5phcm−2s−1rad−1.

126

Table5.1:60FeintensityintheinnerGalaxy

Flux(10−5phcm−2s−1rad−1)
1173keV(SE)4.2±1.6
1173keV(ME)5.8±1.9
1333keV(SE)3.5±1.5
1333keV(ME)5.2±2.1

5.2Observationsoftwo60Felinesat1173and1333Vkeinthe 0.30.3-1]E=1170.05 FWHM=24.59 (±(±9.00)****)-1]E=1345.92 FWHM=5.85 (±(±****)****)
keVkeV0.2I=2.72 (±****)0.2I=0.02 (±****)
-1-1radrad-1-1-2s0.1-2s0.1
-4 ph cm0.0-4 ph cm0.0
Intensity [10-0.1Intensity [10-0.1-0.2-0.211651170117511801325133013351340
Energy [keV]Energy [keV]

GalaxyFigure5.8COMPTEL:Spectra26Alderivmapedfrwithomzeromodelinthefittinginnerusingregiontheofsky−40◦distribution<l<40◦model,−10of◦the<
b<10◦.Nofluxexcessesarefoundnearthe60Felineenergies(1173and1333
V).ke

all60Fesignals.Theenergydeviationof0.23/0.5keVdoesnotaffectthesuperposed
spectrumandderived60Feflux.Thelinefluxestimatedfromthecombinedspectrum
is(4.4±0.9)×10−5phcm−2s−1rad−1.Oursignificanceestimateforthecombined
spectrumis∼4.9σ,addinguncertaintiesoftheindividualspectrainquadrature.This
improvesuponearlier60FesignalreportsfromRHESSI[Smith,2004a,b]andthefirst
yearofSPIdata[Harrisetal.,2005].
Thesignalofcelestial60Feisveryweak,andmarginallysignificant(<3σ)ineach
ofourfourspectra.Thereforewecannotevaluatelineshapeinformation.Thelinesin
ourspectraappearwellrepresentedbyGaussianswithinstrumentalwidths(Figures
5.6and5.7),suggestingthatthecelestial60Felinesareintrinsically-narrowlines.This
wouldimplythatbroadeningof60Felinesfromastrophysicalprocessesisnotsignifi-
cant;mostofthe60Femaybedistributedinarathernormalinterstellarmediumwith
turbulentvelocitiesbelow∼300kms−1.Ifweassumealinebroadeningof1keV,line
fluxeswouldincreaseby∼6%.
Fromourmodel-fittingapproach(§2.5&§5.1),inprinciple,thespectralresultde-
pendsontheinputskymap(alsoseeEq.3.1).Wedonotknowtherealdistribution
of60FesourcesintheGalaxy,butweneedarealisticmodelforthesesourcestode-
rivecorrectspectraandlineintensity.WeusetheCOMPTEL26Almaximumentropy
mapasthestandardskymodelappliedinthemodelfittings.Whilethemaximum
entropymap26includessomestructuresandpatchesintheGalacticplane,wealsoused
theCOMPTELAlMREMmap(Plüschkeetal.2001),asmootherone,inmodel
fittings.AsanalternativetotheCOMPTEL26AlmaximumentropyandMREMsky
maps,wealsotriedadifferentskymap:anexponential-diskmodelwithscaleradius
4kpc,andscaleheight180pc.Thisshouldalsobeagoodfirst-orderrepresentationof

127

5Diffuse60FeemissionintheGalaxybySPI26AlemissionfromtheinnerGalaxy,andavoidsthefinestructureoftheCOMPTEL
imagewhichmaypartlyarisefrominstrumentaloranalysisimperfections.Wefind
thatcantlydif(vferentariationsinputofskythelinemapsdofluxesnotarechangewithinthe10%lineofprtheofilesquotedandvalues,intensitiesi.e.belosignifi-w
thebutionuncertainties).models,e.g.,Ifawepointusesourquiteceatdif±lferent=20and◦,abulgescientificallymodel(aimplausibleGaussianskyprdistri-ofile
intheGalacticcenter),ortheCOMPTEL26Almapwithzerointheinnerregionof
−near60◦the<l60<Fe60line◦,−ener10◦gies<b(see<10◦)examplesinmodelinFigurefittings,5.8).andnofluxexcessesarefound

5.2.2Searchingfor60FesignalfromCygnusandVela
TheCygnusregionisoneofthemostactivenearbystarformationregionsinour
Galaxy(see§4.1).TheVelaregioninthesouthernskyincludesevenmorenearby
massivestars,andseveralrecentcore-collapsesupernovaremnant(§4.2).The1809
keVlineemissionof26AlfromtheCygnusandVelaregionshavebeendetectedwith
SPIdata(seedetailsinChapter4).Hence,thesetwostar-formationregionsshould
bealsogoodcandidatesforgamma-raylineemissionfrom60Fe.Sincethemajorityof
Cygnusregionstar-clustersareyoung(∼3Myr),frompopulationsynthesisstudiesof
themassivestarsintheCygnusregionithasbeensuggestedthatthe60Feproduction
islow,consistentwiththesmallnumberofrecentsupernovaeventsinferredforthis
region(F(1173keV)∼2×10−6phcm−2s−1,seeKnödlsederetal.2002).
Wesearchfor60Fesignalfromthesetworegionsusingthesimilarmethodsdescribed
inChapter4.Splittingthemodelskymap(theMREMimageoftheGalactic1809keV
emission,seeFigure1.7)intoindependenttwocomponents(theCygnusregionandthe
restmap,ortheVelaregionandtherestmap)tofitthedata,wedonotseesignificant
contributionsfromtworegions.Ourestimatedupperlimitsfor60Fegamma-raysfrom
theCygnus(65◦<l<95◦,−13◦<b<17◦)andVela(256◦<l<278◦,−10◦<b<
10◦)regionsare∼1.1×10−5phcm−2s−1(2σ).

5.3Theratioof60Fe/26Al
26Aland60Fewouldshareatleastsomeofthesameproductionsites,i.e.massive
starsandsupernovae(Timmesetal.1995;LimongiandChieffi2006,alsoseeChapter
1).Inadditionbotharelong-livedradioactiveisotopes,sowehavegoodreasonsto
believetheirgamma-ra26ydistributionsaresimilaraswell.Therefore60weadoptthe
skydistributionofAlgamma-raysasourbestmodelforcelestialFegamma-ray
distribution.Andweusean26Aldistributionobtainedindirectobservations,from
the9-yearCOMPTELdata[Plüschkeetal.,2001].
Differenttheoreticalmodelshavepredictedtheratioof60Fe/26Al[Timmesetal.,
1995,Prantzos,2004,LimongiandChieffi,2006].Gamma-rayobservationscouldde-

128

5.3Theratioof60Fe/26AlTable5.2:Differentmeasurementsof60FefluxfromtheinnerGalaxyand60Fe/26Al
ratioflux

Experiments60Feflux(10−5phcm−2s−1rad−1)F(60Fe)/F(26Al)references
HEAO-35.3±4.30.09±0.08Mahoneyetal.1982
SMM2.9±2.50.1±0.08Leising&Share1994
OSSE6.3±4.50.21±0.15Harrisetal.1997
COMPTEL<12(2σ)0.17±0.135Diehletal.1997
GRIS<6.8(2σ)<0.14(2σ)Nayaetal.1998
RHESSI6.3±5.00.16±0.13Smith2004a
RHESSI3.6±1.40.10±0.04Smith2004b
SPI3.7±1.10.11±0.07Harrisetal.2005
SPI4.4±0.90.148±0.06thiswork

tectthesetwoisotopesandreportthefluxratioof60Fe/26Alwhichcanbedirectly
comparedwiththeories.Therefore,themeasurementofthegamma-rayfluxratio
60Fe/26Alisimportantfordiscussionsoftheastrophysicaloriginsofthetworadioac-
tiveisotopes,andthenuclearphysicsinvolvedinmodelsfortheirproduction(ad-
dressingtheuncertainnuclearreactioncrosssectionsandhalf-lives).Forthispur-
pose,weapplythesameanalysismethodonthe26Aldatafromthesameobserva-
tions,analyzingthe1785—1826keVenergyband(1keVbins).Weagaingenerate
aspecificbackgroundmodel(fromadjacentenergybandsandtheGEDSATback-
groundintensitytracer),andapplythesameinputskymap(the9-yearCOMPTEL
26Almaximumentropymap)inmodelfittingtheSEdata.Thisproducesafluxra-
tioof60Fe/26Alinaself-consistentway,andyieldsF(60Fe)/F(26Al)=(14.8±6.0)%.
Here,theuncertaintyhasbeenestimatedfromtherespectivemodel-fittinguncertain-
tiesofthetwoSEdatabases.Alternatively,fromthecombinedspectrumof60Felines
(Figure5.7),andadoptingthe26AlintensitymeasuredwithSPIbefore(F(26Al)=
(3.04±0.31)×10−4phcm−2s−1rad−1fortheinnerGalaxy,seeDiehletal.2006c),we
obtainF(60Fe)/F(26Al)=(14.5±4.0)%.
Manyexperimentsandeffortsweremade(seeTable5.2andFigure5.9)tomeasure
the60Fe/26Alfluxratiothatwaspredictedbytheory–wenowprovidethemostsig-
nificant60detection26todate.InFigure5.9,weshowthepreviousconstraintsontheflux
ratioofFe/Altogetherwiththeresultofthiswork,andcomparetheobserva-
tionalresultswithdifferenttheoreticalpredictions.Theearliestobservationallimit
wasgivenfromHEAO-3,F(60Fe)/F(26Al)=0.09±0.08,anupperlimitbeing0.27
[Mahoneyetal.,1982](inFigure5.8,wechosetogivelimitsat2σforallreportedval-
uesbelowasignificanceof3σ).AnotherlimitwasobtainedwiththeSMMGamma-Ray
Spectrometer,afluxratioof0.1±0.08,theupperlimitbeing∼0.27[LeisingandShare,

129

5Diffuse60FeemissionintheGalaxybySPIFigure5.9:60Flux26ratioofthegamma-raylinesfromthetwolong-livedradioactiveiso-
5.2topes,fromFe/WangAletfral.omse2007v),eralwithobservupperations,limitsshoincludingwnatour2σSPIforallresult(alsoreportedseevTalues,able
andcomparisonwiththerecenttheoreticalestimates(theupperhatchedregion
fromPrantzos2004;thestraightlinetakenfromTimmesetal.1995;thelower
hatchedregion,seeLimongi&Chieffi2006).Ourpresentworkfindsthelineflux
ratiotobe(14.8±6.0)%.Seemoredetailsinthetext.

1994].OSSEaboardtheCOMPTONObservatorygaveafluxratioof0.21±0.15,and
theupperlimitis∼0.51[60Harrisetal.,1994,1997].COMPTELaboardtheCOMPT60ON
ObservatoryalsofoundFegamma-rays,andreportedafluxratiovalueofFe/
26TheAlofGamma-Ra0.17±y0.135,ImagingwhichSpectrtranslatesometerinto(GRIS)anupperreportedlimitan∼upper0.44[limitDiehlforettheal.,ratio1997of].
<0.14(2σ,Nayaetal.1998).RHESSIhasreportedthefirstdetectionof60Fegamma-
yraearylines,dataofandSPIgavgaeveaafluxfluxratioratio0.160.11±±0.130.07for[twHarriso-yetearal.,data2005[],Smithand,the2004a].presentThefirstanal-
ysisofthe3-yearSPIdatafindsafluxratio600.14826±0.06(alsoseeWangetal.2007).
sinceTTheoreticalimmesetal.predictions[1995]ofthepublishedratiotheoffirstFe/detailedAlhavetheoreticalundergoneprediction.someInchangestheir
paper,theycombineamodelfor26Aland60Fenucleosynthesisinsupernovaexplo-
sionswithamodelofchemicalevolution,topredictthatthesteadyproductionrates
are(2.0±1.0)MMyr−1for26Al,and(0.75±0.4)MMyr−1for60Fe,whichcorre-
wspondsouldbetoaconsistentgamma-rawithyfluxourratiopresentF(60Fe)/Fmeasurements.(26Al)=Since0.16±20020.12.,Thistheoreticianspredictionhave
improvedvariousaspectsofthestellar-evolutionmodels,includingimprovedstellar

130

5.3Theratioof60Fe/26Alwindmodelsandthecorrespondingmasslosseffectsonstellarstructureandevolu-
tion,ofmixingeffectsfromrotation,andalsoupdatednuclearcrosssectionsinthe
nucleosynthesispartsofthemodels.Asaresult,predictedfluxratios60Fe/26Al
ratherfellintotherange0.8±0.4(seePrantzos2004,basedon,e.g.Rauscheret
al.2002,Limongi&Chieffi2003)–suchhighvalueswouldbeinconsistentwith
severalobservationallimitsandourSPIresult(butseeWoosleyandHeger2007for
commentsonnuclearreactionrateupdates).Recently,new26Aland60Feyieldmod-
elsarepresented[LimongiandChieffi,2006,WoosleyandHeger,2007],forstarsof
solarmetallicityranginginmassfrom(11−120)M.Limongi&Chieffi(2006)then
combinedtheirmodelsoffullstellarevolutionuptoandincludingthesupernova
withanadoptedmassfunctiontoobtainanewpredictionwithlatestnuclearreaction
rateinputs.Theircalculationsyieldalowerpredictionforthe60Fe/26Alfluxratioof
0.185±0.0625,whichisagainconsistentwiththeobservationalconstraints(seeFigure
).9.5Insummary,uncertaintiesstillexist,bothinmodelsandmeasurementsof60Fe.On
thetheoryside,stellarevolutioninlatestagesiscomplex,nuclearreactionsinclude
neutroncaptureonunstableFeisotopes,andexplosivenucleosynthesisaddsyetan-
othercomplexingredient.Ontheexperimentalside,cosmicrayinduced60Coradioac-
tivityintheinstrumentandspacecraftandthelimitationsofspatialresolutionsand
sensitivityareissuesreflectedinthesubstantialuncertaintiesinexperimentalvalues.
WithmoreINTEGRAL/SPIdatatocome,andalsowiththedevelopmentofnext-
generationgamma-rayspectrometers/telescopes,gamma-rayobservationshopefully
canhelpwithanindependentviewontheastrophysicalmodelcomponents.

131

5

Diffuse

132

60Feemission

in

the

Galaxy

by

SPI

perspectivesandSummary6

Inthisthesis,wehavestudiedthelong-livedradioactivesources,26Aland60Feinthe
GalaxywiththeINTEGRALspectrometer(SPI).SPIhasahighspectralresolutionof
2.5keVat1MeV,suitable26for60detectionsandstudiesofindividualgamma-raylinesof
radioactiveisotopes,AlandFe,andtheirlineshapes.Detailsoflineshapeshelp
usnearbtoystarunderstandformationtheregions,dynamicsoftheKnödlsederejectedetal.isotopes2004)inandtheinlargeinterstellarscalesofmediumtheGalaxy(e.g.,
[Diehletal.,2006a,c].
6026andAlejectionandFeintohavtheesimilarinterstellarastrmediumophysicalareoriginsindominatedthebyGalaxy.massivTheirestars,theirnucleosynthesisevolu-
tion,winds,andtheirsubsequentcore-collapsesupernovaexplosions.Withitsmuch
longerhalf-life(∼106years),26Almaypropagateoversignificantdistancesoffew
novhundredaeuntilpc,andinjectionandaccumulatesβ-decainytheareininterstellarbalanceinmediumtheISM.from26Almanyemissionstarsandappearssuperas-
aofdif60FefusewouldGalaxy-widefollowthatglowof[26Al,Mahonesinceyetitsal.,half-life1982,isPlüschkesimilar,et∼al.2.2,×200110].6yTheears.behavior
Large-scalecharacteristicsof26AlemissionintheinnerGalaxy
Studying26AllineshapedetailsintheGalaxyisoneofthemaingoalsintheSPI
scientificanalyses.With4yearsofSPIdata,wedetect26AlintheinnerGalaxy(|l|<
30flux◦,|forb|<the10◦inner)withGalaxyahighis(3.0±significance0.2)×10le−v4elphofcm∼−302σs−1(seerad§−3.12.).TheThe26Allinemeasuredcentr26oidAl
enersomegywhatfromhigherthethanwholetheinnerlaboratorGalaxyyvisaluedeterforminedthe26atAlline1809.0of±0.21808.65keV±,0.07whichkeVis.Thestill
26AllinefortheinnerGalaxyappearsanarrowfeature,withanintrinsicwidthvalueof
0.5±0.4keV.Thismeasuredlinewidthof26AlfromtheinnerGalaxyisconsistentwith
sourexpectationcesof26ofAlGalactic(turbulentrvotationelocitiesand<modest200kms−1interstellar).Our-mediumresultsareturbulenceconsistentaroundwiththethe
brpreoadviouslinereportswithabywidthHEAO-C∼5.4[keVMahoneasyetreportedal.,1982by],GRISRHESSI[Na[yaSmithet,al.,20031996].]Butisavclearlyery
ruledoutbyourSPImeasurement.
Forourmodel-fittingapproachtoderivethespectra26fromSPIdata(Eq.3.1),the
pre-assumedskydistributionmodelsmayaffecttheAlfluxandlineshapes.As
aconsistent26checkandinordertoestimatesystematicuncertainties,26wehavecom-
paredGalaxyinAllinemodelspectrafittingsdeter(seeTminedable3fr.1om).theFifteendifdifferentferenttracersskyofmapsAlsour(Figureces3.in11)theas

133

6SummaryandperspectivesthedistributionmodelsofGalactic26Alhavebeenusedtoderive26Alspectra.Using
thesedifferenttracermaps,thefittedparametersofderived26Alspectrafortheinner
Galaxy:linecentroid,fluxandwidth,areallconsistentwitheachotherwithinerror
bars.Therefore,differentskydistributionmodelsdonotaffect26Alintensityandline
shapesfortheinnerGalaxysignificantly.
Wecanconversethedetermined26Alintensityinto26AlmassintheGalaxyusing
assumedgeometricalsourcedistributionmodels.Withthemeasured26Alfluxof∼
3×10−4phcm−2s−1rad−1fortheinnerGalaxy,weobtainaGalactic26Almassof
(2.7±0.6)M,where26wehavetakenthedistanceoftheSun26tothe27GalacticcenterR0−=6
8.5kpc.ThisderivedAlmassleadsto26anisotopicratioAl/Alof∼8.1×10
intheinterstellarmedium.ThederivedAlmasscanbeusedtoestimatethecore-
collapsesupernovarate(SNRate)orthestarformationrate(SFR)intheGalaxy.Sowe
obtainacore-collapsesupernovarateofSNRate=1.90±0.95,correspondingtoaSFR
rateofSFR=(3.8±1.9)Myr−1,whichiswellconsistentwiththevaluederivedby
othermethods(McKeeandWilliams1997,alsoseeDiehletal.2006aandreferences
therein).26longitudesalongemissionAlSPIcanprobethespectralinformationtowardpartsofskyusingmultiplesub-maps
coveringthesespecifiedregions,whichisthenewscientifictargetfor26Alstudies.
Studyingthe26Allineshapesfordifferentpartsoftheskyisthemainprojectinmy
thesis.Hence,wehavecarriedoutmodelfittingstoobtain26AlspectraalongtheGalac-
ticlongitudes(§3.6)andalonglatitudes(§3.7),andfordifferentnearbystar-formation
).4§(seeregionsWefirststudiedthe26Alspectraforthe1st(0◦<l<60◦)and4th(−60◦<l<0◦)
quadrants.Brightnessasymmetriesforthetwoquadrantsarefound,andthefluxratio
ofthe4thquadranttothe1stoneis∼1.3forthepartsof|l|<60◦(Figure3.16),and
∼1.2forthepartsof|l|<30◦(Figure3.17).The26Allinecentroidenergiesshow
theshiftsrelativetothecentroidenergyof26Allineinthelaboratory,1808.66keV:a
minorredshift(∼0.1keV)inthe1stquadrantandasignificantblueshift(∼0.6keV)
inthe4thquadrant.WesplittedtheinnerGalaxy(|l|<60◦)intosix20◦-binparts
alonglongitudes,andderivethe26Alspectrumforeachpart.26Allinespectracanbe
stilldetectedsignificantlyforfour20◦binregionsof−40◦<l<40◦,andlineenergy
shiftsarealsodetectedalongtheplane.The26Allinefortheregion20◦<l<40◦
(theSagittariusarm)showsapossiblebroadlinefeature.26Alfluxesfortwooutside
regions(40◦<|l|<60◦)aremuchlowerthanthoseoftheinnerregions.
LineenergyshiftsalongtheGalacticplanewouldbeattributedtoGalacticrotation.
Wehaveobtained26AlspectraalongthelongitudesincludingtheGalacticcenterregion
todeterminethelineenergyshiftsoftheleftandrightpartsrelativetotheGalactic
centerforprobingtheGalacticrotationeffecton26Allinecentroid.26Alspectrafor
threeregionsaresimultaneouslydetermined(|b|<10◦):−40◦<l<−10◦,−10◦<
l<10◦,10◦<l<40◦.The26Allinecentroidenergyfromthecentralinterval

134

−10◦<l<10◦isnotexactlyatthelaboratoryvalue.Takingthe26Alspectrumfrom
−10◦<l<10◦asareference,thelinecentroidenergyshiftsof∼−0.16keV(10◦<
l<40◦)and∼+0.62keV(−40◦<l<−10◦)areclearlyobserved(Figures3.22and
3.23).Inaddition,weobtainedthe26AlspectrumfortheGalacticcenter(−5◦<l<
5◦,−10◦<b<10◦,Figure3.24),withthelinecentroiddeterminedat1808.66±0.23
keV.Thenweusedthe26Alspectraderivedfrom6segmentsin20◦longitudebins
alongtheplane(Figure3.18)toshowthelineenergyshiftsrelativetothereference
spectrumfortheGalacticcenter(Figure3.25).Wehavefoundalowredshiftof∼0.1
keVforpositivelongitudes(0◦<l<40◦)andahighblueshiftof∼0.48keVfor
−20◦<l<0◦and0.9keVfor−40◦<l<−20◦.Thisasymmetryofenergyshiftsis
notconsistentwiththepresentGalacticrotationmodels(Kretschmeretal.2003,also
see§3.8).
Significant26Alspectraaredetectedforthe20◦-binlongitudeintervalsoftheinner
Galaxy(|l|<40◦),andourmodelfittingcouldalsoallowforprobingthespectrainthe
smallerregions.Sowehavetriedtoderivethe26Alspectraforeight10◦-binlongitude
intervalsof10◦oftheinnerGalaxysimultaneously(Figures3.19and3.20),however,
thedetectionofthespectraforeachregionisnotsignificant.Becausethefieldview
ofSPIiswiderthanthescalesofthesesub-maps,contaminationsbetweenadjacent
sub-mapscouldaffectthespectrainmodelfittings,whichleadstofluctuationsofthe
fittedcoefficientsinoff-linespectraanddetectionsof26Alforeachsub-mapinlow
significancelevels.Onepossiblewaytoprobethe26Allinepropertiesinsmallerscales
iscomparingthespectrawithchangesoflongitudeintervalsbyastepof2.5◦or5◦.The
preliminaryresultsoncomparisonof26Allinepropertieswithchangesoflongitude
sizesarepresented(Figures3.26–3.29)anddiscussedin§3.6.3.Theimplicationsare
interestingandrequirefurtherstudieswithmoreSPIdata.
26latitudesalongemissionAlWehavealsoderived26AlspectraalongGalacticlatitudeswithSPI(|l|<60◦).Possi-
ble26Alemissionatlatitudesof|b|>5◦shouldoriginatefromlocalstar-formationsys-
temsintheGouldBelt,while26Alatlatitudesof|b|<5◦wouldbedominatedbythe
large-scaleoriginintheGalacticdisk/spiralarms.Weak26Alemissionisdetectedin
thepositivelatituderegionof5◦<b<20◦withthelinecentroidblueshiftof0.7±0.4
keV,butno26Alsignalisdetectedinthenegativelatituderegionof−20◦<b<−5◦
(seeFigure3.30).Furtheranalysisofthelatitudeinterval5◦<b<20◦foundthat
no26Alisdetectedforthe1stquadrant,and26Allineemissionforthe4thquadrant
issignificantlydetected,withtheblueshiftof0.8±0.5keV(Figure3.31).This26Al
emissioncouldbeattributedtothenearbySco-CenOBassociation.26Alemissionfor
lowlatitudes(|b|<5◦)issignificantlydetectedasexpected.The26Allineforthe
wholeplaneshowsanarrowfeature,withthelinecentroidenergyof1808.78±0.08
keV.Theasymmetryof26Albrightnessforthe1stand4thquadrantsisalsodetected,
andthe26Allinecentroidshowsaredshiftof0.21±0.27keVforthe1stquadrant,and
ablueshiftof0.54±0.23keVforthe4thquadrant,whichisconsistentwiththeexpla-

135

6SummaryandperspectivesnationofGalacticrotation.ToprobetheGalacticrotationeffecton26Alspectraalong
theplanewithoutcontaminationsof26Alemissionfromhighlatitudes,wederived
26Alspectraforthreelongitudeintervalsofonlylowlatitudes(|b|<5◦):10◦<l<40◦
,−10◦<l<10◦,−40◦<l<−10◦.Thelinecentroidenergyforthecenterregion
(|l|<10◦)isdeterminedat1808.65±0.27keV,wellwithinthelaboratoryvalue.Energy
shiftsof26Allinecentroidsarealsoobserved:asmallredshiftof0.11±0.35keVplusa
largeblueshiftof0.75±0.27keV.Theasymmetryofenergyshiftsalongthediskplane
isstillinconsistentwiththeexpectationofGalacticrotation(Kretschmeretal.2003,
Kretschmer2007,see§3.8),maybeinducedbyothereffects,e.g.,theBarstructureor
thecontributionofnearby26Alsources.
26AlemissionoftheGalaxy:simulationsversusobservations
SPIobservationsof26AlspectraintheGalaxyfirstlyconfirmthedominant26Al
origininmassivestarsandsupernovae,andalsoprobethedynamicsof26Alejecta
fromstellarwindsandsupernovaexplosions,andGalacticrotationeffecton26Alline
shapes.26Alintensityandlineshapescanbesimulatedwiththeknowndistribution
modelsof26AlsourcesintheGalaxy[Kretschmer,2007].Comparisonbetweenthe
simulatedresultsandSPIobservationswillbeagoodconstrainton26Aloriginand
disk.GalactictheindistributionWehaveadoptedathree-dimensionalmodelforthespacedensityoffreeelectrons
basedonpulsardispersionmeasurements[CordesandLazio,2002]astheparentdis-
tributionofGalactic26Al,whichalsoincludesthreedifferentdistributionmodels
[Kretschmer,2007]:(1)allcomponents;(2)allexceptthethickdiskandtheGalactic
center;(3)onlythethindiskandspiralarms.Thekinematicsof26Alsourcesfrom
stellarwindsandsupernovaexplosionsinthelocalinterstellarmediumsuperimposed
onGalacticrotationisconsideredinoursimulations.
ForcomparisonbetweensimulationsandSPIresults,weintegrateoursimulated
skymapoflineenergiesandintensitiesoveraregionofinterest,andthenobtain
resultingspectraforthreelongitudeintervalsoftheinnerGalaxy:−5◦<l<5◦,5◦<
l<40◦,−40◦<l<−5◦(seeFigure3.34,Kretschmer2007).Simulated26Alspectra
fortheinnerGalaxyshowthenarrowlinefeatures,withanintrinsicwidthvalue
<0.9keV,whichisconsistentwiththeSPIconstraints.Nolinecentroidshiftsfor
theGalacticcenter(|l|<5◦)inallthreedistributionmodelsareconsistentwiththe
SPIresult.26AllineenergyshiftsduetoGalacticrotationaresymmetricalforpositive
andnegativelongitudes,theexpectedshiftvalueis∼0.2keVforthemodelofall
components,∼0.3keVforthemodelofallexceptthethickdiskandtheGalactic
centerand∼0.4keVforthemodelofonlythethindiskandspiralarms.However,
theobservedlinecentroidshiftsofpositiveandnegativelongitudesisasymmetrical,
∼0.1keVforpositivelongitudes,and∼0.5−0.9keVfornegativeones.Inconsistence
betweentheGalacticrotationexpectandSPIobservationssuggestedthatsomeeffects
notconsideredinoursimulationsleadtothe26Allineshiftasymmetry,e.g.,theBar
structure,anddynamicsofnearby26Alsources.The26Alintensityratioofthe4th

136

quadranttothe1stoneoftheinnerGalaxycanbeobtainedinsimulationstocompare
withtheSPIobservation.Forthemodelofallcomponentsandthatofallexceptthe
thickdiskandtheGalacticcenter,26Alintensitiesforthe1stand4thquadrantsare
nearlysymmetrical,whichisnotconsistentwithourSPIresultsin§3.6.1.Butforthe
modelofonlythethindiskandspiralarms,significantasymmetryappears,witha
26Alintensityratioof∼1.15whichmeetsthepresentobservedlimits.Sothefree-
electrondensitymodel26ofonlythethindiskandspiralarmswouldreflectthereal
.AlGalacticofdistribution26Alemissionfromstar-formationregions
Detectionsof26Alfromnearbystar-formationregionsarethegoodprobeforthe
massivestaroriginofGalactic26Alandkinematicsof26AlejectainISM.Westudied
26Alspectrafromtheseregionsusingmodelfittingswithtwoskymapscoveringthe
star-formationregionandtherest.Thenwehaveobtained26Alspectrafromthree
nearbystar-formationregionswiththepresentSPIdata(§4):Cygnus,VelaandSco-
Cen.Wederivedthe26AlspectrumforCygnus(65◦<l<95◦,−13◦<b<17◦)with
asignificancelevelof∼6σ.Areported26Alfluxis∼(7.0±1.2)×10−5phcm−2s−1,
whichisconsistentwiththeCOMPTELresult.Noshiftofthelinecentroidisobserved,
andthe26Allineshowsthenarrowfeature,withanintrinsiclinewidth<2keV.A
typicalISMturbulentvelocityis≤200kms−1intheCygnusregion.
26AldetectionfortheVelaregion(−104◦<l<−82◦,−10◦<b<10◦isnot
significant(<3σ).The26Alfluxisdeterminedat(4.0±1.2)×10−5phcm−2s−1which
isconsistentwiththeCOMPTELresult.The26AllineinVelaappearsabroadfeature,
withoutalinecentroidshift.AGaussianwiththeinstrumentallinewidthcannotfit
thespectrumwell,sowefitthespectrumwithfixedGaussianwidth(FWHM=3.77
keV)afterassumingtheintrinsic26AlwidthinVelaasFWHM∼2keV.Apossible
broad26AllinefeatureofVelashouldrequirethefurtherstudieswithmoreSPIdata.
Sco-Cenistheneareststar-formationregionfromthesun.Weobtainthe26Alspec-
trumforSco-Cen−35◦<l<−5◦,5◦<b<30◦withasignificancelevelof∼4σ.A
reported26AlfluxforSco-Cenis(6.3±2.0)×10−5phcm−2s−1.Importantly,The26Al
spectrumshowsasignificant26blueshiftof∼1.5keVrelativetothelaboratoryvalue.
ThisshiftimpliesthatAlsourcesinSco-Cenaredominatedbythecomponentof
26Alejectawithabulkvelocityof∼200kms−1towardus.
Diffuse60FeemissionintheGalaxy
Finally,Wehavereportedthedetectionof60Fedecaygamma-raylinesintheGalaxy
from3yearsofSPIobservations.Withournewmeasurementswedetectboththe1173
keVand1333keVlineof60FefromSPIsingleandmultiple-detectorevents.Combin-
ingourfourspectrafromindependentmodelfitsweobtainan60Fesignalfromthe
Galaxywithasignificanceof4.9σ.Thisimprovesuponourpreviousmeasurements
(seeHarrisetal.2005).Theaverage60FelinefluxfromtheinnerGalaxyregionis
(4.4±0.9)×10−5phcm−2s−1rad−1,assumingintrinsicallynarrowlinesandasky

137

6Summaryandperspectivesdistributionequaltothatof26AlasmeasuredbyCOMPTEL(Plüschkeetal.2001).
Fromthesameobservationsandanalysisprocedureappliedto26Al,wefindaflux
ratioof60Fe/26Alof(14.8±6.0)%.The60Fesignalsaretooweaktodetermineline
shapedetails;itappearsthataGaussianwiththewidthoftheinstrumentalresolution
canfitthedatawell.Thiswouldimplythatbroadeningof60Felinesfromastrophysi-
calprocessesisnotsignificant;mostofthe60Femaybedistributedinarathernormal
interstellarmediumwithturbulentvelocitiesbelow∼300kms−1.Inordertoinvesti-
gatethevariationsovertheGalaxy,wesearchfor60FeemissionfromtheCygnusand
Velaregions,anddonotfind60Fesignals.
Inthisthesis,wehaveusedthefour-yearSPIdataforspectralanalysesof26Aland
60FeintheGalaxy.INTEGRAL/SPIwillcontinuetoaccumulatemoredataforseveral
years,whichwouldcovermoreskyregions.WithmoreSPIdata,wecouldimprove
spectralresolutionof26AlintheinnerGalaxyand26Alspectraalongtheplane,and
thenconfirm26AllineshapesforCygnus,VelaandSco-Cen,furtherobtain26Alspectra
fromsomeothernearbystar-formationregions,likeOrionandCarina.For60Festudies,
thesignificancelevelofdetectionsfortheinnerGalaxywillbepushedforwardsagain,
andpossibledifferencesof60Fesignalsforthe1stand4thquadrantscanbeprobed
withmoreSPIdata.Finally,wecansearchforthe60Fegamma-raysignalfromthe
59keVlineusing60INTEGRAL/IBISdataandSWIFTdata,whichmightbeaself-
studies.FeforcheckconsistentSPIaboardINTEGRALhasahighspectralresolutionbutaverylimitedsensitivity
over0.1–2MeV(seeFigure6.1).PresentlytheGLASTmission[Michelson,2001],
athirdgeneration(afterSAS-2/COS-BBignamietal.1975andEGRETKanbachetal.
1988)pair-creationtelescopesensitiveabove20MeV,isscheduledtobelaunchedin
October2007.InFigure6.1,thepointsourcesensitivity,i.e.theminimumdetectable
flux,forpast,present(e.g.,instrumentsaboardINTEGRAL),andfutureinstruments
inthelowthroughhigh-energyregimesisshown.Thelackofasensitiveinstrument
from0.3–10MeVisobvious.Therefore,itistimelytodevelopasecond-generation
γ-raytelescopeintheComptonenergyrangetomaintaintheoverallmulti-wavelength
sensitivity,whichissoimportantforafullunderstandingofhigh-energyastrophysics.
Astronomyinthelow-tomedium-energyγ-rayrangehasalwayschallengedthe
mostadvancedexperimentaltechniquesfortwomainreasons:(1)photoninteraction
cross-sectionsintheMeVrangegothroughaminimumintheirtransitionfromthe
photoelectriceffect(∼100keV)topaircreation(∼10MeV).Inthisso-calledComp-
tonrange,theinteractionsarecharacterizedbysmallenergydepositsandlong-range
secondaryradiation.Itisthereforenecessarytobuildadeepdetectortoachievea
reasonableefficiencyandtofinelysegmentthedetectorinordertorecordandtrigger
widelyseparatedinteractionsofthescatteredphotons.(2)Thenuclearenergylevels
ofalldetectorandstructuralmaterialslieintheMeVrangeandarethereforeeasily
excitedbyenergeticparticles.Theresultofthissystematicradio-activationinorbitis
aprolificγ-raybackgroundoflocaloriginwhichmustbeeffectivelydiscriminated
againsttoachieveausefulsensitivityforastronomicaltargets.

138

Figure6.1:SensitivitylevelsforselectedhardX-andgamma-raytelescopes(10keV–
GeV).Thelackofapresentsensitiveinstrumentfrom300keV–10MeVisevident.
ThegoalforMEGAissettobeaboutoneorderofmagnitudeimprovementover
COMPTEL(dashedline).ThehistogramforMEGAshowspreliminaryresultsof
asimulationforaMEGAspacemission[Kanbachetal.,2005].

139

6SummaryFigure140

.6:2matics

andperspectivesSchematicoftheMEGAinstrument(fromKanbachetal.2005)and
ofgamma-rayinteractions:Comptonscatterorpairconversionof

photons.

kine-γy-ra

TelescopesbasedonthedetectionofCompton-andpair-creationinteractionshave
intrinsicallylargefields-of-viewandarewellsuitedasskysurveymonitorsandtomap
large-scaleextendedemissions.Theseextendedsourcesofhigh-energyphotonsarise
partlyfrominteractionsofcosmic-rayswithinterstellargas,fromdispersedradioactive
isotopes(mainly26Al,60Fe,β+-decaywithfollowingannihilationradiation),and
possiblyfromthedecayofdarkmatterinagalactichalo.
ThesensitivitygoalforthenextMeVspacecraftmissionshouldfallintoaninter-
mediatelevel,betweentheCOMPTEL/INTEGRALsensitivitiesandtheextensionof
theGLASTsensitivityintotheMeVrange.ItwilllikelyrevolutionizeMeVastronomy
inthemannerthatEGRETrevolutionizedastronomyat100MeV.Toaccomplishthese
goalswithamedium-sizedmissionrequiresnewdesignsbeyondthatofCOMPTEL
andEGRET.AnewCompton/low-energypairtelescopeforMediumEnergyGamma-
rayAstronomy(MEGA)intheenergyband0.4–50MeV,hasbeenindevelopmentand
test.MEGAfunctionsinsomewayssimilartoCOMPTELandEGRET,butalsodiffers
inmanyaspects.MEGAusesastackofdouble-sidedsiliconstripdetectors(DSSD)as
ascatteringandconversiontrackingdetectorandfinelypixelatedCsI/PINdiodescin-
tillationdetectorsfortheabsorptionofthescatteredradiation[Kanbachetal.,2005].
AschematicofthebasicMEGAdesignisshowninFigure6.2.SinceMEGAhasa
smallsizeandmaystillbeaninsufficientstepforthefutureγ-raylineastronomy,
othervisionaryMeVtelescopeshavealsobeenproposed.
AvisionaryAdvancedComptonTelescope(ACT)coveringtheenergyrangeof
0.4–50MeVhasagoalsensitivityof∼10−6MeVcm−2s−1in106sforcontinuum,
correspondingtoGLASTabove100MeV,andasensitivityfornarrowlinesof∼
10−7phcm−2s−1in106s(Figure6.3).ACThasadesignprinciplesimilartoMEGA
showninFigure6.2withalargersize,thenwouldhavehigherenergyresolution
[Boggsetal.,2003].AnothervisionaryEuropeanGamma-RayImager(GRI)mission
basedonLauelenses(theBraggdiffractionfromcrystalsintransmissionconfigura-
tion)from50keV–2MeVisnowunderstudyfortheESAplanCosmicVision2015-
2025[Knödlseder,2006].GRIhasanangularresolutionof∼1arcmin,andahigh
energy−7resolution−2−1ofE/Δ6E∼500at1MeV.Thegoalsensitivityfornarrowlinesis
∼10phcmsin10s(alsoseeFigure6.3).
Thesenext-generationMeVgamma-raytelescopeswillgreatlyimproveγ-rayline
astronomy.Presentandpastinstruments,e.g.SPIandCOMPTEL,candetectthe
511keVannihilationline,60FelinesfromtheinnerGalaxy,56Colineat847keVfrom
extragalacticTypeIaSNewithin20Mpc(broadline,SN1991TinNGC4527,∼17
Mpc,Morrisetal.1995),44TiinCasA(broadline),26AlintheinnerGalaxyandnearby
activestar-formationregions.Whilemanypotentialγ-raylinesources:the511keV
linefromGalacticnovaeandnearbysupernovaremnants;60FefromtheVelaSNR;
56CofromSNIaatadistanceupto80Mpc;44TifrommoreyoungSNRs,e.g.Tycho,
Kepler,SN1987A;26Alfromindividualsources,liketheVelaSNR,theclosestWolf-
RayetstarWR11intheVelastar-formationregion,evenfromCrabandCasA;and
thenuclearexcitinglinesinISMfromCandO(seeFigure6.3),arewellbelowthe

141

6SummaryFigure142

.6:3andperspectivesFuture

goals

for

-raγy

line

onomyastr

2.0(

10

MeV):

next-generation

in-

struments(e.g.,Advanced-ComptonTelescope,Gamma-RayImager)andpoten-
tialcandidateγ-raylinesourcesfore−e+annihilationline,56Co,44Ti,60Feand
26Al(fromBoggsetal.2003).SensitivitiesofSPIandCOMPTELarealsoshown

comparison.forhere

SensitivitiesofSPIandCOMPTELarealsoshown

γ-raysensitivitytelescopes,limitsofi.e.,ACTpresentandmissions,GRI(Figurebut6could.3).beDetectionsdetectedofbythethe511futurekeVadvline,44ancedTi,

26Aland60FefromsomeindividualSNRswillputtheconnectionwithyieldsofthese

isotopesradioactivity

-raγliney

missions.

onomyastr

omfr

yma

stellarnucleosynthesisonsolidground.Eraforaccurate

come

near

with

the

launch

of

next-generation

VMe

y-raγ

143

6

Summary

144

and

perspectives

Bibliography

S.J.Allen.Phys.Rev.,34:296,1911.
G.Alter,B.Balazs,J.Ruprecht,andJ.Vanysek.Catalogueofstarclustersandassociations.
Budapest:AkademiaiKiado,1970,2nded.,editedbyG.Alter,B.Balazs,J.Ruprecht,
.1970W.D.physicsAr,nett.pagesIn90M.–+,D.1977.Papagiannis,editor,EightTexasSymposiumonRelativisticAstro-
W.D.Arnett.SupernovaeandNucleosynthesis.PrincetonUniversity,Princeton,NJ,
.1996Z.Arzoumanian,D.F.Chernoff,andJ.M.Cordes.ApJ,568:289–301,March2002.
B.Aschenbach.X-rayemissionfromsupernovaremnantsobservedwithROSAT.Ad-
vancesinSpaceResearch,13:45,December1993.
B.Aschenbach,A.F.Iyudin,andV.Schönfelder.A&A,350:997–1006,October1999.
M.Asplund,N.Grevesse,andA.J.Sauval.InT.G.Barnes,IIIandF.N.Bash,editors,
sisASP,pagesConf.25Ser.–+,336S:eptemberCosmic2005Abundances.asRecordsofStellarEvolutionandNucleosynthe-
D.Attié,B.Cordier,M.Gros,P.Laurent,S.Schanne,G.Tauzin,P.vonBallmoos,
L.drenne,Bouchet,R.P.GeorJean,gii,J.A.vonKnödlseder,Kienlin,P.G.MandrLichti,ou,V.P.SPaul,chönfelderJ.-P.,A.Roques,Strong,G.C.SkinnerW,G.undererVe-,
C.Shrader,S.Sturner,B.Teegarden,G.Weidenspointner,J.Kiener,M.-G.Porquet,
V.Tatischeff,S.Crespin,S.Joly,Y.André,F.Sanchez,andP.Leleux.A&A,411:
L71–L79,November2003.
V.S.Avedisova.AstronomyReports,49:435–445,June2005.
H.Becquerel.Compt.Rend.,130:809,1900.
H.Becquerel.Compt.Rend.,122:420,1896.
K.BelczynskiandR.E.Taam.ApJ,616:1159–1166,December2004.
C.G.L.F.Bennett,Smoot,A.D.J.T.WBanday,ilkinson,K.M.andGorski,E.L.G.Wright.HinshaApJw,,P.464:LJackson,1+,P.June1996Keegstra,.A.Kogut,

145

BibliographyG.F.Bignami,G.Boella,J.J.Burger,B.G.Taylor,P.Keirle,J.A.Paul,H.A.Mayer-
Hasselwander,E.Pfeffermann,L.Scarsi,andB.N.Swanenburg.SpaceScienceIn-
strumentation,1:245–268,August1975.
J.Binney,O.Gerhard,andD.Spergel.MNRAS,288:365–374,June1997.
H.Bloemen,W.Hermsen,B.N.Swanenburg,C.P.deVries,R.Diehl,V.Schoenfelder,
H.Steinle,A.W.Strong,A.Connors,M.McConnell,D.Morris,G.Stacy,K.Bennett,
andC.Winkler.ApJS,92:419–423,June1994.
C.Boehm,D.Hooper,J.Silk,M.Casse,andJ.Paul.PhysicalReviewLetters,92(10):
101301,March2004.
S.E.Boggs,M.S.Amman,P.Jean,S.Amrose,W.Coburn,R.P.Lin,D.M.Smith,
P.N.Luke,N.W.Madden,M.Burks,E.L.Hull,P.vonBallmoos,W.W.Craig,
andK.Ziock.InJ.E.TruemperandH.D.Tananbaum,editors,X-RayandGamma-
RayTelescopesandInstrumentsforAstronomy.EditedbyJoachimE.Truemper,HarveyD.
Tananbaum.ProceedingsoftheSPIE,Volume4851,pp.1221-1227(2003).,volume4851
ofPresentedattheSocietyofPhoto-OpticalInstrumentationEngineers(SPIE)Conference,
pages1221–1227,March2003.
J.BrandandL.Blitz.A&A,275:67–+,August1993.
J.C.Brandt.Sol.Phys.,6:171–+,1969.
J.C.Brandt,T.P.Stecher,D.L.Crawford,andS.P.Maran.ApJ,163:L99+,February
.1971E.M.Burbidge,G.R.Burbidge,W.A.Fowler,andF.Hoyle.ReviewsofModernPhysics,
29:547–650,1957.
D.N.Burrows,K.P.Singh,J.A.Nousek,G.P.Garmire,andJ.Good.ApJ,406:97–111,
.1993chMarR.W.Bussard,R.Ramaty,andR.J.Drachman.ApJ,228:928–934,March1979.
M.Cassé,B.Cordier,J.Paul,andS.Schanne.ApJ,602:L17–L20,February2004.
G.R.CaughlanandW.A.Fowler.AtomicDataandNuclearDataTables,40:283–+,1988.
A.N.Cha,K.R.Sembach,andA.C.Danks.ApJ,515:L25–L28,April1999.
A.ChanotandJ.P.Sivan.A&A,121:19–25,May1983.
W.Chen,N.Gehrels,andR.Diehl.ApJ,440:L57–L60,February1995.
W.Chen,R.Diehl,N.Gehrels,D.Hartmann,M.Leising,J.E.Naya,N.Prantzos,
J.Tueller,andP.vonBallmoos.InC.Winkler,T.J.-L.Courvoisier,andP.Durou-
choux,editors,ESASP-382:TheTransparentUniverse,pages105–+,1997.

146

BibliographyK.S.Cheng,L.Zhang,P.Leung,andZ.J.Jiang.ApJ,608:418–426,June2004.
K.S.Cheng,D.O.Chernyshov,andV.A.Dogiel.ApJ,645:1138–1151,July2006.
L.X.Cheng,M.Leventhal,D.M.Smith,W.R.Purcell,J.Tueller,A.Connors,D.Dixon,
R.L.Kinzer,andJ.G.Skibo.ApJ,481:L43+,May1997.
X.Chi,K.S.Cheng,andE.C.M.Young.ApJ,459:L83+,March1996.
E.Churazov,R.Sunyaev,S.Sazonov,M.Revnivtsev,andD.Varshalovich.MNRAS,
357:1377–1386,March2005.
D.D.Clayton.Nature,368:222–+,March1994.
F.R.ColombandG.M.Dubner.A&A,82:244–248,February1980.
A.H.Compton.Naturwissenschaften,17:507,1929.
J.M.CordesandT.J.W.Lazio.NE2001.ArXivAstrophysicse-prints(astro-ph/0207156),
.2002JulyT.J.-L.Courvoisier,R.Walter,V.Beckmann,A.J.Dean,P.Dubath,R.Hudec,
P.Kretschmar,S.Mereghetti,T.Montmerle,N.Mowlavi,S.Paltani,A.PreiteMar-
tinez,N.Produit,R.Staubert,A.W.Strong,J.-P.Swings,N.J.Westergaard,N.White,
C.Winkler,andA.A.Zdziarski.A&A,411:L53–L57,November2003.
T.M.Dame,H.Ungerechts,R.S.Cohen,E.J.deGeus,I.A.Grenier,J.May,D.C.
Murphy,L.-A.Nyman,andP.Thaddeus.ApJ,322:706–720,November1987.
T.M.Dame,D.Hartmann,andP.Thaddeus.ApJ,547:792–813,February2001.
H.deBoerandet.al.InV.DiGesuandet.al.,editors,DataAnalysisinAstronomyIV,
volume59,page241,1991.
E.J.deGeus,P.T.deZeeuw,andJ.Lub.A&A,216:44–61,June1989.
P.T.deZeeuw,R.Hoogerwerf,J.H.J.deBruijne,A.G.A.Brown,andA.Blaauw.AJ,
117:354–399,January1999.
E.delRio,P.vonBallmoos,K.Bennett,H.Bloemen,R.Diehl,W.Hermsen,
J.Knoedlseder,U.Oberlack,J.Ryan,V.Schoenfelder,andC.Winkler.A&A,315:
237–242,November1996.
J.M.DickeyandF.J.Lockman.ARA&A,28:215–261,1990.
R.DiehlandF.X.Timmes.PASP,110:637–659,June1998.

147

BibliographyR.Diehl,C.Dupraz,K.Bennett,H.Bloemen,W.Hermsen,J.Knoedlseder,G.Lichti,
D.Morris,J.Ryan,V.Schoenfelder,H.Steinle,A.Strong,B.Swanenburg,M.Varen-
dorff,andC.Winkler.A&A,298:445–+,June1995a.
R.Diehl,K.Bennett,H.Bloemen,C.Dupraz,W.Hermsen,J.Knoedlseder,G.Lichti,
D.Morris,U.Oberlack,J.Ryan,V.Schoenfelder,H.Steinle,M.Varendorff,and
C.Winkler.A&A,298:L25+,June1995b.
R.Diehl,U.Wessolowski,U.Oberlack,H.Bloemen,R.Georgii,A.Iyudin,J.Kn-
odlseder,G.Lichti,W.Hermsen,D.Morris,J.Ryan,V.Schonfelder,A.Strong,P.von
Ballmoos,andC.Winkler.InC.D.Dermer,M.S.Strickman,andJ.D.Kurfess,edi-
tors,AIPConf.Proc.410:ProceedingsoftheFourthComptonSymposium,pages1109–+,
.1997embervNoR.Diehl,N.Baby,V.Beckmann,P.Connell,P.Dubath,P.Jean,J.Knödlseder,J.-P.
Roques,S.Schanne,C.Shrader,G.Skinner,A.Strong,S.Sturner,B.Teegarden,
A.vonKienlin,andG.Weidenspointner.A&A,411:L117–L121,November2003a.
R.Diehl,J.Knödlseder,G.G.Lichti,K.Kretschmer,S.Schanne,V.Schönfelder,A.W.
Strong,A.vonKienlin,G.Weidenspointner,C.Winkler,andC.Wunderer.A&A,
411:L451–L455,November2003b.
R.Diehl,K.Kretschmer,G.Lichti,V.Schönfelder,A.W.Strong,A.vonKienlin,
J.Knödlseder,P.Jean,V.Lonjou,G.Weidenspointner,J.-P.Roques,G.Vedrenne,
S.Schanne,N.Mowlavi,C.Winkler,andC.Wunderer.InV.Schoenfelder,G.Lichti,
andC.Winkler,editors,ESASP-552:5thINTEGRALWorkshopontheINTEGRAL
Universe,pages27–+,October2004.
R.Diehl,H.Halloin,K.Kretschmer,G.G.Lichti,V.Schönfelder,A.W.Strong,
A.vonKienlin,W.Wang,P.Jean,J.Knödlseder,J.-P.Roques,G.Weidenspointner,
S.Schanne,D.H.Hartmann,C.Winkler,andC.Wunderer.Nature,439:45–47,2006a.
R.Diehl,N.Prantzos,andP.vonBallmoos.NuclearPhysicsA,777:70–97,October
b.2006R.Diehl,H.Halloin,K.Kretschmer,A.W.Strong,W.Wang,P.Jean,G.G.Lichti,
J.Knödlseder,J.-P.Roques,S.Schanne,V.Schönfelder,A.vonKienlin,G.Weidens-
pointner,C.Winkler,andC.Wunderer.A&A,449:1025–1031,April2006c.
C.Dupraz,H.Bloemen,K.Bennett,R.Diehl,W.Hermsen,A.F.Iyudin,J.Ryan,and
V.Schoenfelder.A&A,324:683–689,August1997.
P.Durouchoux,P.Wallyn,C.Chapuis,J.Matteson,B.Bowman,M.Pelling,L.Peterson,
G.Vedrenne,P.vonBallmoos,I.Malet,M.Niel,R.Lin,P.Feffer,D.Smith,and
K.Hurley.A&AS,97:185–187,January1993.

148

BibliographyC.D.Ellison,C.F.Jones,andR.Ramaty.InInternationalCosmicRayConference,pages
.1990–+,68P.EnglmaierandO.Gerhard.MNRAS,304:512–534,April1999.
R.B.FirestoneandL.B.Ekstrüm.http://ie.lbl.gov/toi/.2004.
G.A.P.Franco.A&A,227:499–514,January1990.
C.FranssonandC.Kozma.NewAstronomyReview,46:487–492,July2002.
G.Gamow.PhysicalReview,70:572–573,October1946.
G.Garcia-Segura,N.Langer,andM.-M.MacLow.A&A,316:133–146,December1996.
C.D.GarmanyandR.E.Stencel.A&AS,94:211–244,August1992.
C.D.Garmany,P.S.Conti,andC.Chiosi.ApJ,263:777–790,December1982.
N.GehrelsandW.Chen.A&AS,120:C331+,December1996.
O.Gerhard.InASPConf.Ser.273:TheDynamics,Structure&HistoryofGalaxies:A
WorkshopinHonourofProfessorKenFreeman,pages73–+,2002.
O.E.GerhardandM.Vietri.MNRAS,223:377–+,November1986.
P.GoldreichandW.H.Julian.ApJ,157:869–+,August1969.
J.N.Goswami,K.K.Marhas,andS.Sahijpal.ApJ,549:1151–1159,March2001.
M.Gounelle,F.H.Shu,H.Shang,A.E.Glassgold,K.E.Rehm,andT.Lee.ApJ,640:
1163–1170,April2006.
D.A.Green.CatalogueofGalacticSupernovaRemnants,URL:
http://www.mrao.cam.ac.uk/surveys/snrs/.2006.
I.A.Grenier.A&A,364:L93–L96,December2000.
P.Guillout,M.F.Sterzik,J.H.M.M.Schmitt,C.Motch,andR.Neuhaeuser.A&A,
337:113–124,September1998.
M.J.Harris,W.R.Purcell,D.A.Grabelsky,M.P.Ulmer,W.N.Johnson,R.L.Kinzer,
J.D.Kurfess,M.S.Strickman,M.D.Leising,andG.V.Jung.InBulletinofthe
AmericanAstronomicalSociety,pages968–+,May1994.
M.J.Harris,W.R.Purcell,K.McNaron-Brown,R.J.Murphy,J.E.Grove,W.N.John-
son,R.L.Kinzer,J.D.Kurfess,G.H.Share,andG.V.Jung.InC.D.Dermer,M.S.
Strickman,andJ.D.Kurfess,editors,AIPConf.Proc.410:ProceedingsoftheFourth
ComptonSymposium,pages1079–+,November1997.

149

BibliographyM.J.Harris,J.Knödlseder,P.Jean,E.Cisana,R.Diehl,G.G.Lichti,J.-P.Roques,
S.Schanne,andG.Weidenspointner.A&A,433:L49–L52,April2005.
C.G.T.Haslam,C.J.Salter,H.Stoffel,andW.E.Wilson.AstronomyDataImageLibrary,
pages1–+,December1995.
A.Heger,C.L.Fryer,S.E.Woosley,N.Langer,andD.H.Hartmann.ApJ,591:288–300,
.2003JulyM.HerantandS.E.Woosley.ApJ,425:814–828,April1994.
W.HillebrandtandF.-K.Thielemann.ApJ,255:617–623,April1982.
C.Ho.MNRAS,221:523–536,August1986.
C.Y.HuiandW.Becker.A&A,454:543–552,August2006.
S.D.Hunter,D.L.Bertsch,J.R.Catelli,T.M.Dame,S.W.Digel,B.L.Dingus,J.A.
Esposito,C.E.Fichtel,R.C.Hartman,G.Kanbach,D.A.Kniffen,Y.C.Lin,H.A.
Mayer-Hasselwander,P.F.Michelson,C.vonMontigny,R.Mukherjee,P.L.Nolan,
E.Schneid,P.Sreekumar,P.Thaddeus,andD.J.Thompson.ApJ,481:205–+,May
.1997A.F.Iyudin,R.Diehl,H.Bloemen,W.Hermsen,G.G.Lichti,D.Morris,J.Ryan,
V.Schoenfelder,H.Steinle,M.Varendorff,C.deVries,andC.Winkler.A&A,284:
L1–L4,April1994.
A.F.Iyudin,R.Diehl,G.G.Lichti,V.Schönfelder,A.W.Strong,H.Bloemen,
W.Hermsen,J.Ryan,K.Bennett,andC.Winkler.InC.Winkler,T.J.-L.Courvoisier,
andP.Durouchoux,editors,ESASP-382:TheTransparentUniverse,pages37–+,1997.
A.F.Iyudin,V.Schonfelder,K.Bennett,H.Bloemen,R.Diehl,W.Hermsen,G.G.
Lichti,R.D.vanderMeulen,J.Ryan,andC.Winkler.Nature,396:142–144,1998.
F.A.Iyudin,V.Schönfelder,K.Bennett,H.Bloemen,R.Diehl,W.Hermsen,
J.Knödlseder,G.G.Lichti,U.Oberlack,J.Ryan,W.A.Strong,andC.Winkler.
AstrophysicalLettersCommunications,38:383–+,1999.
P.Jean,G.Vedrenne,J.P.Roques,V.Schönfelder,B.J.Teegarden,A.vonKienlin,
J.Knödlseder,C.Wunderer,G.K.Skinner,G.Weidenspointner,D.Attié,S.Boggs,
P.Caraveo,B.Cordier,R.Diehl,M.Gros,P.Leleux,G.G.Lichti,E.Kalemci,J.Kiener,
V.Lonjou,P.Mandrou,P.Paul,S.Schanne,andP.vonBallmoos.A&A,411:L107–
L112,November2003.
P.Jean,J.Knödlseder,W.Gillard,N.Guessoum,K.Ferrière,A.Marcowith,V.Lonjou,
andJ.P.Roques.A&A,445:579–589,January2006.

150

BibliographyP.L.Jensen,K.Clausen,C.Cassi,F.Ravera,G.Janin,C.Winkler,andR.Much.A&A,
411:L7–L17,November2003.

W.N.Johnson,IIIandR.C.Haymes.ApJ,184:103–126,August1973.

J.Jose,M.Hernanz,andA.Coc.ApJ,479:L55+,April1997.

G.Kanbach,D.L.Bertsch,C.E.Fichtel,R.C.Hartman,S.D.Hunter,D.A.Kniffen,
B.W.Hughlock,A.Favale,R.Hofstadter,andE.B.Hughes.SpaceScienceReviews,
49:69–84,1988.

G.Kanbach,R.Andritschke,A.Zoglauer,M.Ajello,M.L.McConnell,J.R.Macri,J.M.
Randyan,P.MethodsBloser,inS.PhysicsHunter,ResearG.chA,Dicocco,541:J.310–Kurfess,322,AprilandV.2005.Reglero.NuclearInstruments
A.I.KarakasandJ.C.Lattanzio.PublicationsoftheAstronomicalSocietyofAustralia,20:
279–293,2003.
K.Kifonidis,T.Plewa,H.-T.Janka,andE.Müller.A&A,408:621–649,September2003.

K.Knie,G.Korschinek,T.Faestermann,E.A.Dorfi,G.Rugel,andA.Wallner.Physical
ReviewLetters,93(17):171103,October2004.
J.Knödlseder.A&A,360:539–548,August2000.

J.Knödlseder.InM.J.L.TurnerandG.Hasinger,editors,SpaceTelescopesandInstru-
PrmentationoceedingsII:oftheUltravioletSPIE,toVolumeGamma6266Ray,.pp.Edited626623byT(urner2006,).,vMartinolumeJ.L.;6266ofHasingerPr,esentedGüntherat.
theSocietyofPhoto-OpticalInstrumentationEngineers(SPIE)Conference,July2006.

J.Knödlseder.ApJ,510:915–929,January1999.

J.J.RyKnödlseder,an,V.SD.chönfelder,Dixon,K.andP.Bennett,vonH.Ballmoos.Bloemen,A&AR.,345Diehl,:813W–.825Her,Mamsen,y1999U.a.Oberlack,

J.V.SKnödlseder,chönfelder,K.andBennett,P.vonH.Ballmoos.Bloemen,R.A&A,Diehl,344:68W–.82,HerAprilmsen,1999U.b.Oberlack,J.Ryan,

J.Knödlseder,M.Cerviño,J.-M.LeDuigou,G.Meynet,D.Schaerer,andP.vonBall-
moos.A&A,390:945–960,August2002.

J.Knödlseder,M.Valsesia,M.Allain,S.Boggs,R.Diehl,P.Jean,K.Kretschmer,J.-
PC..WRoques,inkler.V.InSV.Schönfelder,choenfelder,G.VG.edrenne,Lichti,P.vandonC.WBallmoos,inkler,G.Weditors,ESAeidenspointner,SP-552:5andth
INTEGRALWorkshopontheINTEGRALUniverse,pages33–+,October2004.

151

BibliographyJ.Knödlseder,P.Jean,V.Lonjou,G.Weidenspointner,N.Guessoum,W.Gillard,
G.Skinner,P.vonBallmoos,G.Vedrenne,J.-P.Roques,S.Schanne,B.Teegarden,
V.Schönfelder,andC.Winkler.A&A,441:513–532,October2005.
B.Kozlovsky,R.E.Lingenfelter,andR.Ramaty.ApJ,316:801–818,May1987.
K.Kretschmer.PhDthesis,TechnischeUniversitätMünchen,2007.
K.Kretschmer,R.Diehl,andD.H.Hartmann.A&A,412:L47–L51,December2003.
P.Kroupa.Science,295:82–91,January2002.
B.M.Kuzhevskij,S.N.Kuznetsov,andE.V.Troitskaia.AdvancesinSpaceResearch,22:
1141–1147,1998.
M.Lang.Dipl.Thesis,TechnischeUniversitätMünchen.2006.
N.Langer.A&A,220:135–143,August1989.
N.Langer,H.Braun,andJ.Fliegner.Ap&SS,224:275–278,February1995.
T.N.LaRosa,N.E.Kassim,T.J.W.Lazio,andS.D.Hyman.AJ,119:207–240,January
.2000T.N.LaRosa,C.L.Brogan,S.N.Shore,T.J.Lazio,N.E.Kassim,andM.E.Nord.ApJ,
626:L23–L27,June2005.
R.Launhardt,R.Zylka,andP.G.Mezger.A&A,384:112–139,March2002.
M.D.LeisingandG.H.Share.ApJ,424:200–207,March1994.
P.Leleux,F.Albernhe,V.Borrel,B.Cordier,R.Coszach,S.Crespin,J.M.Denis,
P.Duhamel,P.Frabel,W.Galster,J.-S.Graulich,P.Jean,B.Kandel,J.P.Meulders,
G.Tauzin,J.Vanhorenbeeck,G.Vedrenne,andP.vonBallmoos.A&A,411:L85–L90,
.2003embervNoM.Leventhal,C.J.MacCallum,andP.D.Stang.ApJ,225:L11–L14,October1978.
M.LimongiandA.Chieffi.ApJ,592:404–433,July2003.
M.LimongiandA.Chieffi.ApJ,647:483–500,August2006.
R.E.LingenfelterandG.J.Hueter.InS.E.Woosley,editor,AmericanInstituteofPhysics
ConferenceSeries,pages558–+,1984.
R.E.LingenfelterandR.Ramaty.InR.Ramaty,editor,AIPConf.Proc.101:Positron-
ElectronPairsinAstrophysics,pages267–272,1983.

152

BibliographyC.J.MacCallum,A.F.Huters,P.D.Stang,andM.Leventhal.ApJ,317:877–880,June
.1987K.MaedaandK.Nomoto.ApJ,598:1163–1200,December2003.
W.A.Mahoney,J.C.Ling,A.S.Jacobson,andR.E.Lingenfelter.ApJ,262:742–+,
.1982embervNoW.A.Mahoney,J.C.Ling,W.A.Wheaton,andA.S.Jacobson.ApJ,286:578–585,
.1984embervNoR.N.ManchesterandJ.H.Taylor.Observedandderivedparametersfor330pulsars.
AJ,86:1953–1973,December1981.
A.Mastichiadis,R.J.Protheroe,andS.A.Stephen.InD.O’Sullivan,editor,22dInt.
Cosmic-RayConf.,page2,1991.
H.A.Mayer-Hasselwander,D.L.Bertsch,B.L.Dingus,A.Eckart,J.A.Esposito,
R.Genzel,R.C.Hartman,S.D.Hunter,G.Kanbach,D.A.Kniffen,Y.C.Lin,
P.F.Michelson,A.Muecke,C.vonMontigny,R.Mukherjee,P.L.Nolan,M.Pohl,
O.Reimer,E.J.Schneid,P.Sreekumar,andD.J.Thompson.A&A,335:161–172,July
.1998M.McConnell,S.Fletcher,K.Bennett,H.Bloemen,R.Diehl,W.Hermsen,J.Ryan,
V.Schonfelder,A.Strong,andR.vanDijk.InC.D.Dermer,M.S.Strickman,andJ.D.
Kurfess,editors,AIPConf.Proc.410:ProceedingsoftheFourthComptonSymposium,
pages1099–+,November1997.
C.F.McKeeandJ.K.Truelove.Phys.Rep.,256:157–172,May1995.
C.F.McKeeandJ.P.Williams.ApJ,476:144–+,February1997.
G.MellemaandP.Lundqvist.A&A,394:901–909,November2002.
G.Meynet,M.Arnould,N.Prantzos,andG.Paulus.A&A,320:460–468,April1997.
P.F.Michelson.InS.Ritz,N.Gehrels,andC.R.Shrader,editors,Gamma2001:Gamma-
RayAstrophysics,volume587ofAmericanInstituteofPhysicsConferenceSeries,pages
.2001–+,713G.E.MillerandJ.M.Scalo.ApJS,41:513–547,November1979.
P.A.Milne,J.D.Kurfess,M.D.Leising,andD.D.Dixon.InBulletinoftheAmerican
AstronomicalSociety,pages1253–+,October2000.
P.A.Milne,J.D.Kurfess,R.L.Kinzer,M.D.Leising,andD.D.Dixon.InA.Gimenez,
V.Reglero,andC.Winkler,editors,ESASP-459:ExploringtheGamma-RayUniverse,
pages145–148,September2001.

153

BibliographyD.J.Morris,K.Bennett,H.Bloemen,W.Hermsen,G.G.Lichti,M.L.McConnell,J.M.
Ryan,andV.Schönfelder.InH.Böhringer,G.E.Morfill,andJ.E.Trümper,editors,
SeventeethTexasSymposiumonRelativisticAstrophysicsandCosmology,volume759of
NewYorkAcademySciencesAnnals,pages397–+,1995.
N.MowlaviandG.Meynet.A&A,361:959–976,September2000.
W.A.MulderandB.T.Liem.A&A,157:148–158,March1986.
M.P.Muno,F.K.Baganoff,M.W.Bautz,W.N.Brandt,P.S.Broos,E.D.Feigelson,
G.P.Garmire,M.R.Morris,G.R.Ricker,andL.K.Townsley.ApJ,589:225–241,May
.2003D.C.MurphyandJ.May.A&A,247:202–214,July1991.
S.Nagataki,M.-A.Hashimoto,K.Sato,andS.Yamada.ApJ,486:1026–+,September
.1997J.E.Naya,S.D.Barthelmy,L.M.Bartlett,N.Gehrels,M.Leventhal,A.Parsons,B.J.
Teegarden,andJ.Tueller.Nature,384:44–46,November1996.
J.E.Naya,S.D.Barthelmy,L.M.Bartlett,N.Gehrels,A.Parsons,B.J.Teegarden,
J.Tueller,andM.Leventhal.Gamma-RayLimitsonGalactic60FeNucleosynthesis
andImplicationsontheOriginofthe26AlEmission.ApJ,499:L169+,June1998.
S.NikolaevandM.D.Weinberg.ApJ,487:885–+,October1997.
I.Nofar,G.Shaviv,andS.Starrfield.ApJ,369:440–450,March1991.
T.NugisandH.J.G.L.M.Lamers.A&A,360:227–244,August2000.
U.Oberlack.PhDthesis,TechnischeUniversitätMünchen,1997.
U.Oberlack,R.Diehl,T.Montmerle,N.Prantzos,andP.vonBallmoos.ApJS,92:
433–439,June1994.
U.Oberlack,K.Bennett,H.Bloemen,R.Diehl,C.Dupraz,W.Hermsen,J.Knoedlseder,
D.Morris,V.Schoenfelder,A.Strong,andC.Winkler.A&AS,120:C311+,December
.1996U.Oberlack,U.Wessolowski,R.Diehl,K.Bennett,H.Bloemen,W.Hermsen,
J.Knödlseder,D.Morris,V.Schönfelder,andP.vonBallmoos.A&A,353:715–721,
.2000yJanuarM.S.Oey.ApJ,467:666–+,August1996.
M.S.OeyandC.J.Clarke.BulletinoftheAmericanAstronomicalSociety,28:857–+,May
.1996

154

BibliographyR.P.OllingandM.R.Merrifield.MNRAS,311:361–369,January2000.
A.Palacios,G.Meynet,C.Vuissoz,J.Knödlseder,D.Schaerer,M.Cerviño,and
N.Mowlavi.A&A,429:613–624,January2005.
C.A.PerrotandI.A.Grenier.A&A,404:519–531,June2003.
E.Pfahl,S.Rappaport,andP.Podsiadlowski.ApJ,571:L37–L40,May2002.
S.Plüschke.PhDthesis,TechnischeUniversitätMünchen,2001.
S.Plüschke,R.Diehl,V.Schönfelder,H.Bloemen,W.Hermsen,K.Bennett,C.Winkler,
M.McConnell,J.Ryan,U.Oberlack,andJ.Knödlseder.InA.Gimenez,V.Reglero,
andC.Winkler,editors,ESASP-459:ExploringtheGamma-RayUniverse,pages55–58,
.2001eptemberSM.Politano,S.Starrfield,J.W.Truran,A.Weiss,andW.M.Sparks.ApJ,448:807–+,
.1995ugustAN.Prantzos.A&A,420:1033–1037,June2004.
N.PrantzosandR.Diehl.Phys.Rep.,267:1–69,March1996.
R.K.Prinja,M.J.Barlow,andI.D.Howarth.ApJ,361:607–620,October1990.
W.R.Purcell,L.-X.Cheng,D.D.Dixon,R.L.Kinzer,J.D.Kurfess,M.Leventhal,M.A.
Saunders,J.G.Skibo,D.M.Smith,andJ.Tueller.ApJ,491:725–+,December1997.
R.Ramaty.A&AS,120:C373+,December1996.
R.RamatyandR.E.Lingenfelter.ApJ,213:L5–L7,April1977.
R.RamatyandR.E.Lingenfelter.Nature,278:127–132,March1979.
R.Ramaty,B.Kozlovsky,andR.E.Lingenfelter.ApJS,40:487–526,July1979.
T.Rauscher,A.Heger,R.D.Hoffman,andS.E.Woosley.ApJ,576:323–348,September
.2002B.C.Reed.AJ,130:1652–1657,October2005.
M.J.Reid.ARA&A,31:345–372,1993.
M.Renaud,F.Lebrun,J.Ballet,A.Decourchelle,R.Terrier,andN.Prantzos.In
V.Schoenfelder,G.Lichti,andC.Winkler,editors,ESASP-552:5thINTEGRALWork-
shopontheINTEGRALUniverse,pages81–+,October2004.
M.Renaud,J.Vink,A.Decourchelle,F.Lebrun,P.R.d.Hartog,R.Terrier,C.Couvreur,
J.Knödlseder,P.Martin,N.Prantzos,A.M.Bykov,andH.Bloemen.ApJ,647:L41–
L44,August2006.

155

BibliographyR.J.Reynolds.ApJ,206:679–684,June1976.
A.C.Robin,C.Reylé,S.Derrière,andS.Picaud.A&A,409:523–540,October2003.
C.E.RolfsandW.S.Rodney.CauldronsintheCosmos.UniversityofChicagoPress,
.1994J.P.Roques,S.Schanne,A.vonKienlin,J.Knödlseder,R.Briet,L.Bouchet,P.Paul,
S.Boggs,P.Caraveo,M.Cassé,B.Cordier,R.Diehl,P.Durouchoux,P.Jean,P.Leleux,
G.Lichti,P.Mandrou,J.Matteson,F.Sanchez,V.Schönfelder,G.Skinner,A.Strong,
B.Teegarden,G.Vedrenne,P.vonBallmoos,andC.Wunderer.A&A,411:L91–L100,
.2003embervNoM.Ruderman.ApJ,366:261–269,January1991.
M.A.RudermanandP.G.Sutherland.ApJ,196:51–72,February1975.
D.Russeil.A&A,397:133–146,January2003.
E.Rutherford.Radio-activity.Cambridge:CambridgeUniv.Press,1905.
E.Rutherford.Phil.Mag.,47:109,1899.
M.J.Sartori,J.R.D.Lépine,andW.S.Dias.A&A,404:913–926,June2003.
S.Schanne,P.Sizun,D.Maurin,B.Cordier,A.vonKienlin,andC.Hamadache.ArXiv
Astrophysicse-prints(astro-ph/0701640),January2007.
V.SchoenfelderandM.Varendorff.InP.DurouchouxandN.Prantzos,editors,AIP
Conf.Proc.232:Gamma-RayLineAstrophysics,pages101–115,1991.
V.Schoenfelder,H.Aarts,K.Bennett,H.deBoer,J.Clear,W.Collmar,A.Connors,
A.Deerenberg,R.Diehl,A.vonDordrecht,J.W.denHerder,W.Hermsen,M.Kip-
pen,L.Kuiper,G.Lichti,J.Lockwood,J.Macri,M.McConnell,D.Morris,R.Much,
J.Ryan,G.Simpson,M.Snelling,G.Stacy,H.Steinle,A.Strong,B.N.Swanenburg,
B.Taylor,C.deVries,andC.Winkler.ApJS,86:657–692,June1993.
V.Schönfelder,K.Bennett,J.J.Blom,H.Bloemen,W.Collmar,A.Connors,R.Diehl,
W.Hermsen,A.Iyudin,R.M.Kippen,J.Knödlseder,L.Kuiper,G.G.Lichti,M.Mc-
Connell,D.Morris,R.Much,U.Oberlack,J.Ryan,G.Stacy,H.Steinle,A.Strong,
R.Suleiman,R.vanDijk,M.Varendorff,C.Winkler,andO.R.Williams.A&AS,143:
145–179,April2000.
G.H.Share,R.L.Kinzer,J.D.Kurfess,D.J.Forrest,E.L.Chupp,andE.Rieger.ApJ,
292:L61–L65,May1985.
A.ShukolyukovandG.W.Lugmair.Science,259:1138–1142,February1993.

156

BibliographyG.SkinnerandP.Connell.A&A,411:L123–L126,November2003.
D.M.Smith.ApJ,589:L55–L58,May2003.
D.M.Smith.NewAstronomyReview,48:87–91,February2004a.
D.M.Smith.InV.Schoenfelder,G.Lichti,andC.Winkler,editors,ESASP-552:5th
INTEGRALWorkshopontheINTEGRALUniverse,pages45–+,October2004b.
S.Starrfield,J.W.Truran,M.Politano,W.M.Sparks,I.Nofar,andG.Shaviv.Phys.Rep.,
227:223–234,1993.
A.W.Strong.A&A,411:L127–L129,November2003.
A.W.Strongandet.al.InV.DiGesuandet.al.,editors,DataAnalysisinAstronomy
IV,volume59,page251,1991.
A.W.Strong,J.B.G.M.Bloemen,T.M.Dame,I.A.Grenier,W.Hermsen,F.Lebrun,
L.-A.Nyman,A.M.T.Pollock,andP.Thaddeus.A&A,207:1–15,December1988.
A.W.Strong,H.Bloemen,R.Diehl,W.Hermsen,andV.Schönfelder.Astrophysical
LettersCommunications,39:209–+,1999.
A.W.Strong,L.Bouchet,R.Diehl,P.Mandrou,V.Schönfelder,andB.J.Teegarden.
A&A,411:L447–L450,November2003.
A.W.Strong,R.Diehl,H.Halloin,V.Schönfelder,L.Bouchet,P.Mandrou,F.Lebrun,
andR.Terrier.A&A,444:495–503,December2005.
S.J.SturnerandJ.E.Naya.ApJ,526:200–206,November1999.
S.J.Sturner,C.R.Shrader,G.Weidenspointner,B.J.Teegarden,D.Attié,B.Cordier,
R.Diehl,C.Ferguson,P.Jean,A.vonKienlin,P.Paul,F.Sánchez,S.Schanne,
P.Sizun,G.Skinner,andC.B.Wunderer.A&A,411:L81–L84,November2003.
P.A.Sturrock.ApJ,164:529–+,March1971.
S.TachibanaandG.R.Huss.ApJ,588:L41–L44,May2003.
S.Tachibana,G.R.Huss,N.T.Kita,G.Shimoda,andY.Morishita.ApJ,639:L87–L90,
.2006chMarK.Tachihara,R.Neuhäuser,S.Toyoda,T.Onishi,A.Mizuno,andY.Fukui.InE.R.
Schielicke,editor,AstronomischeGesellschaftMeetingAbstracts,pages72–+,2001.
J.H.TaylorandJ.M.Cordes.ApJ,411:674–684,July1993.
B.J.Teegarden,S.D.Barthelmy,N.Gehrels,J.Tueller,M.Leventhal,andC.J.MacCal-
lum.ApJ,375:L9–L12,July1991.

157

BibliographyB.J.Teegarden,K.Watanabe,P.Jean,J.Knödlseder,V.Lonjou,J.P.Roques,G.K.
Skinner,P.vonBallmoos,G.Weidenspointner,A.Bazzano,Y.M.Butt,A.De-
courchelle,A.C.Fabian,A.Goldwurm,M.Güdel,D.C.Hannikainen,D.H.Hart-
mann,A.Hornstrup,W.H.G.Lewin,K.Makishima,A.Malzac,J.Miller,A.N.
Parmar,S.P.Reynolds,R.E.Rothschild,V.Schönfelder,J.A.Tomsick,andJ.Vink.
ApJ,621:296–300,March2005.
O.V.Terekhov,R.A.Syunyaev,A.V.Kuznetsov,C.Barat,R.Talon,G.Trottet,and
N.Vilmer.AstronomyLetters,19:65–68,March1993.
F.-K.Thielemann,K.Nomoto,andM.-A.Hashimoto.ApJ,460:408–+,March1996.
F.X.Timmes,S.E.Woosley,D.H.Hartmann,R.D.Hoffman,T.A.Weaver,and
F.Matteucci.ApJ,449:204–+,August1995.
P.Ubertini,F.Lebrun,G.DiCocco,A.Bazzano,A.J.Bird,K.Broenstad,A.Gold-
wurm,G.LaRosa,C.Labanti,P.Laurent,I.F.Mirabel,E.M.Quadrini,B.Ramsey,
V.Reglero,L.Sabau,B.Sacco,R.Staubert,L.Vigroux,M.C.Weisskopf,andA.A.
Zdziarski.A&A,411:L131–L139,November2003.
K.A.vanderHucht.TheVIIthcatalogueofgalacticWolf-Rayetstars.NewAstronomy
Review,45:135–232,February2001.
K.A.vanderHucht,B.Hidayat,A.G.Admiranto,K.R.Supelli,andC.Doom.A&A,
199:217–234,June1988.
K.A.vanderHucht,H.Schrijver,B.Stenholm,I.Lundstrom,A.F.J.Moffat,S.V.
Marchenko,W.Seggewiss,D.Y.A.SetiaGunawan,W.Sutantyo,E.P.J.vanden
Heuvel,J.-P.deCuyper,andA.E.Gomez.NewAstronomy,2:245–250,August1997.
G.Vedrenne,J.-P.Roques,V.Schönfelder,P.Mandrou,G.G.Lichti,A.vonKienlin,
B.Cordier,S.Schanne,J.Knödlseder,G.Skinner,P.Jean,F.Sanchez,P.Caraveo,
B.Teegarden,P.vonBallmoos,L.Bouchet,P.Paul,J.Matteson,S.Boggs,C.Wun-
derer,P.Leleux,G.Weidenspointner,P.Durouchoux,R.Diehl,A.Strong,M.Cassé,
M.A.Clair,andY.André.A&A,411:L63–L70,November2003.
J.Vink,J.M.Laming,J.S.Kaastra,J.A.M.Bleeker,H.Bloemen,andU.Oberlack.ApJ,
560:L79–L82,October2001.
J.S.VinkandA.deKoter.A&A,442:587–596,November2005.
P.vonBallmoos,R.Diehl,andV.Schoenfelder.ApJ,318:654–663,July1987.
G.Wallerstein,I.J.Iben,P.Parker,A.M.Boesgaard,G.M.Hale,A.E.Champagne,
C.A.Barnes,F.Käppeler,V.V.Smith,R.D.Hoffman,F.X.Timmes,C.Sneden,R.N.
Boyd,B.S.Meyer,andD.L.Lambert.ReviewsofModernPhysics,69:995–1084,1997.

158

BibliographyW.Wang.Chin.J.Astron.Astrophys.Supplement,ArXivAstrophysicse-prints(astro-
ph/0510461),6:268–272,December2006.
W.Wang,Z.J.Jiang,andK.S.Cheng.MNRAS,358:263–269,March2005a.
W.Wang,Z.J.Jiang,C.S.J.Pun,andK.S.Cheng.MNRAS,360:646–654,June2005b.
W.Wang,C.S.J.Pun,andK.S.Cheng.A&A,446:943–948,February2006.
W.Wang,M.J.Harris,R.Diehl,H.Halloin,B.Cordier,A.W.Strong,K.Kretschmer,
J.Knoedlseder,P.Jean,G.G.Lichti,J.P.Roques,S.Schanne,A.vonKienlin,G.Wei-
denspointner,andC.Wunderer.A&A,469:1005–1012,July2007.
R.A.WardandW.A.Fowler.ApJ,238:266–286,May1980.
G.J.Wasserburg,R.Gallino,andM.Busso.ApJ,500:L189+,June1998.
R.Weaver,R.McCray,J.Castor,P.Shapiro,andR.Moore.ApJ,218:377–395,December
.1977T.A.WeaverandS.E.Woosley.Phys.Rep.,227:65–96,1993.
G.Weidenspointner,J.Kiener,M.Gros,P.Jean,B.J.Teegarden,C.Wunderer,R.C.
Reedy,D.Attié,R.Diehl,C.Ferguson,M.J.Harris,J.Knödlseder,P.Leleux,V.Lon-
jou,J.-P.Roques,V.Schönfelder,C.Shrader,S.Sturner,V.Tatischeff,andG.Ve-
drenne.A&A,411:L113–L116,November2003.
C.WeidnerandP.Kroupa.ApJ,625:754–762,June2005.
A.WeissandJ.W.Truran.A&A,238:178–186,November1990.
S.Wheelock,J.Chillemi,N.Gautier,D.Gregorich,D.Kester,H.McCallon,C.Oken,
J.White,andT.Chester.InBulletinoftheAmericanAstronomicalSociety,pages908–+,
.1991chMarM.Wiescher,J.Gorres,F.-K.Thielemann,andH.Ritter.A&A,160:56–72,May1986.
C.Winkler,T.J.-L.Courvoisier,G.DiCocco,N.Gehrels,A.Giménez,S.Grebenev,
W.Hermsen,J.M.Mas-Hesse,F.Lebrun,N.Lund,G.G.C.Palumbo,J.Paul,J.-
P.Roques,H.Schnopper,V.Schönfelder,R.Sunyaev,B.Teegarden,P.Ubertini,
G.Vedrenne,andA.J.Dean.A&A,411:L1–L6,November2003.
P.F.Winkler,J.H.Tuttle,R.P.Kirshner,andM.J.Irwin.InR.S.RogerandT.L.
Landecker,editors,IAUColloq.101:SupernovaRemnantsandtheInterstellarMedium,
.1988–+,65pagesS.E.Woosley.InJ.Audouze,C.Chiosi,andS.E.Woosley,editors,Saas-FeeAdvanced
Course16:NucleosynthesisandChemicalEvolution,pages1–+,1986.

159

BibliographyS.E.Woosley.ApJ,476:801–+,February1997.

S.E.WoosleyandA.Heger.toappearinPhysicsReports,BetheCentennialVolume,ArXiv
Astrophysicse-print(astro-ph/0702176),2007.

S.E.WoosleyandT.A.Weaver.ApJS,101:181–+,November1995.

S.E.Woosley,W.D.Arnett,andD.D.Clayton.TheExplosiveBurningofOxygenand
Silicon.ApJS,26:231–+,November1973.

S.E.Woosley,D.Hartmann,andP.A.Pinto.ApJ,346:395–404,November1989.

S.E.JuneW1990oosle.y,D.H.Hartmann,R.D.Hoffman,andW.C.Haxton.

L.ZhangandK.S.Cheng.A&A,398:639–646,February2003.

160

ApJ,356:272–301,

Acknowledgements

IfirstlywouldliketothankmyadvisorDr.habil.RolandDiehlverymuchforhis
carefulandintensiveguidanceinthreeyearsofstudyatMPE.Hebringmetoacom-
pletelynewscientificfieldformebeforeIcamehere:gamma-raylineastronomy.The
progressofmyknowledgeandnowthefinishofmyPhDworkshouldbeattributed
help.Roland’stoIwanttothankProf.GüntherHasingerforrepresentingthisthesistotheTechnische
München.ersitätUnivwarImalsowthankelcomeKarstenmademefeelKretschmerthat,Ithecamefirstbackpeoplehome.ImetatThankMPEKarstenthreeyverearsyago.muchHisfor
hishelpinmystudyforthreeyears.
ThanksalottoZhangXiao-Lingforherfriendshipandassistanceinmystudyand
life.IGiselhershouldLichti,beveryAndreasgratefulvontoKienlin,JochenWernerGreiner,CollmarAndre,wHubertStrong,Halloin,GottfriedMichaelKanbach,Lang,
andtheotherpeopleintheGammaGroupfortheirhelpinmystudies.
Jean,IamJ.P.gratefulRoques,toS.someSchanne,colleaguesB.Corindier,otherG.Winstitutes,eidenspointnerM.J.,C.Harris,WJ.undererKnödlsederfortheir,P.
collaborationinmyresearchwork.
WangManyJie,WthanksangtoLan,myXiangChineseFei,friendsReginaintheHuang,GarDachingvidHui,campus,GuoFanQi,YLii,YZhangang-Fang,Yu-YGaoing,
Hong,ZhangChenandthepeopleintheGarchingCSSUfortheirfriendshipandhelp
duringthelastthreeyears.
IamalsogratefultotheMax-Planck-InstitutfürextraterrestrischePhysiktoprovide
methegoodchanceofstudyandresearchforthreeyearsinGermany.Theexcellent
academicenvironmentinMPEandthecampusofGarchingmakesmeveryenjoyable
studies.PhDmyduringtheThehardwareINTEGRALteams.projectTheisSPIprsupportedojectbhasygobeenvernmentcompletedgrantsunderinallmemberresponsibilitystatesandof
leadershipofCNES.IamgratefultoASI,CEA,CNES,DLR,ESA,INTA,NASA,and
support.forOSTCloveofFinallymy,specialmother,thanksmytosister,andencouragementmywife,IandcansupportsuccessfullyfrommyfinishfamilymyinresearChina.chworkFor
thesis.PhDtheand

161