Identifying XMM-Newton observations affected by solar wind charge exchange - Part II

0.44 MB·English

by  J. A. Carter

Astronomy&Astrophysicsmanuscriptno.15817main (cid:13)c ESO2011
January11,2011
Identifying XMM-Newton observations affected by solar wind
⋆
charge exchange - Part II
J.A.Carter1,S.Sembay1,andA.M.Read1
DepartmentofPhysicsandAstronomy,UniversityofLeicester,Leicester,LE17RH,UK
e-mail:[email protected]
e-mail:[email protected]
e-mail:[email protected]
1
1 Received;accepted
0
2 ABSTRACT

n
Aims.WewishedtoanalyseasampleofobservationsfromtheXMM-NewtonScienceArchivetosearchforevidenceofexospheric
a
J solarwindchargeexchange(SWCX)emission.
Methods.Weanalysed3012observationsuptoandincludingrevolution1773.Themethodemployedextendsfromthatoftheprevi-
0
ouslypublishedpaperbytheseauthorsonthistopic.WedetecttemporalvariabilityinthediffuseX-raybackgroundwithinanarrow
1
low-energybandandcontrastthistoacontinuum.Thelow-energybandwaschosentorepresentthekeyindicatorsofchargeexchange

emissionandthecontinuumwasexpectedtobefreeofSWCX.
]
M Results.Approximately3.4%ofobservationsstudiedareaffected.WediscussourresultswithreferencetotheXMM-Newtonmis-
sion.WefurtherinvestigateremarkablecasesbyconsideringthestateofthesolarwindandtheorientationofXMM-Newtonatthe
I timeoftheseobservations.Wepresentamethodtoapproximatetheexpectedemissionfromobservations,basedongivensolarwind
h. parameterstakenfromanupstreamsolarwindmonitor.WealsocomparetheincidenceofSWCXcaseswithsolaractivity.
p Conclusions.We present a comprehensive study of the majority of the suitable and publically available XMM-Newton Science
- Archivetodate,withrespecttotheoccurrenceofSWCXenhancements.WepresentourSWCX-affectedsubsetofthisdataset.The
o mean exospheric-SWCX flux observed within this SWCX-affected subset was 15.4 keVcm−2s−1sr−1 in the energy band 0.25 to
r 2.5keV.ExosphericSWCXispreferentiallydetectedwhenXMM-NewtonobservesthroughthesubsolarregionoftheEarth’smag-
t
s netosheath.Themodeldevelopedtoestimatetheexpectedemissionreturnsfluxeswithinafactorofafewoftheobservedvaluesin
a themajorityofcases,withameanvalueat83%.
[
Keywords.X-rays-diffusebackground-Sun:solar-terrestrialrelation-methods:dataanalysis

1
v
8
1. Introduction comets, within the heliosphere and at the heliospheric bound-
4
8 arywheretheouterreachesoftheinterplanetarymagneticfield
This paper follows that of Carter&Sembay (2008) (hereafter
1 encounter that of the surrounding interstellar medium. Those
PaperI),wherebyasetofapproximately180XMM-Newtonob-
. charge exchange processes that result in the emission of X-
1 servations,takenbetweenrevolutions52and1104(March2000
raysoccurforinteractionsinvolvingahighly-chargedsolarwind
0 untilDecember2005),wereanalysedtosearchforcasesofsolar ion, for example the bare oxygen ion O8+, and a donor neu-
1
wind charge exchange emission (SWCX) occurring within the
1 tralspecies,suchashydrogeninthecaseofgeocoronalSWCX
Earth’smagnetosheathorinnearinterplanetaryspace.InPaper
: emission.
v I we searched for time-variable SWCX signatures and found
ThepossibilitiesforviewingexosphericSWCXemissionin
i that approximately 6.5% of the observations studied were af-
X thevicinityoftheEarthdependontheorbitalandviewingcon-
fected.One case ofSWCX showedanextremelyrichemission
r line spectrum,which was attributed to a passing CoronalMass straints of the observatoryin use, however one may be able to
a detect enhancements due to charge exchange occurring within
Ejection(CME)andhasbeendiscussedindetailbyCarteretal.
theheliosphere.Thiswillshowfluctuationsonlongertimescales
(2010). This paper extends the work of Paper I, to provide a
toSWCXoccurringwithintheEarth’sexosphere(Cravensetal.
complete sample, covering3012suitable XMM-Newton obser-
2001).SWCX emissionfromwithintheheliosheath(forexam-
vations downloaded from the XMM-Newton Science Archive
ple resulting from the helium-focusing cone) is a contributor
(XSA)1.Weincludedatafrombetweenrevolutions28and1773
to the X-ray emission from the supposed Local Hot Bubble in
(February2000untilAugust2009).
which the Sun resides (Koutroumpaetal. 2007, 2008), and to
SWCX processes occur at many locations within the so-
whatlevelitcontributesis undergreatdebate.We concentrate,
lar system including planetary magnetosheaths, the corona of
however, on SWCX emission occurring in the near vicinity of
theEarthandusethetimevariablenatureofthisemissionasour
Sendoffprintrequeststo:J.A.Carter
markerforselectingaffectedXMM-Newtonobservations.
⋆ Based on observations with XMM-Newton, an ESA Science
Mission with instruments and contributions directly funded by ESA In addition to the pioneering work noting the so-called
MemberStatesandtheUSA(NASA). Long Term Enhancements (subsequently known to be due to
1 http://xmm.esac.esa.int/xsa/ SWCX emission) within ROSAT observations (Snowdenetal.
2 J.A.Carteretal.:IdentifyingXMM-Newtonobservationsaffectedbysolarwindchargeexchange-PartII
1995),theassignmentofsuchenhancementstoSWCXemission combined if the same filter was used for both. We considered
has also been observed in studies undertaken using data from observationsuptoandincludingrevolution1773(August2009).
Suzaku (Fujimotoetal. 2007; Bautzetal. 2009) and Chandra Individual observation event lists were created from the
(Smithetal. 2005). Enhancements in soft X-ray band XMM- OriginalDataFilesforeachEPIC-MOScameraandfilteredfor
Newton spectra have been attributed to SWCX contamination soft-protoncontaminationusingthepubliclyavailableExtended
intheliterature(Snowdenetal.2004;Kuntz&Snowden2008; SourceAnalysisSoftware(ESAS)2package(mos-filtertool).At
Snowdenetal. 2009;Henley&Shelton2008). Thishasalmost the time of data processing, ESAS was only available for the
exclusivelyinvolvedthecomparisonofmultiplepointingsofthe EPIC-MOScameras.ThistoolfitsaGaussiantoahistogramof
samefieldwhichhasenabledtheserendipitousdetectionofthe in-field-of-view count rates in a high energy band (2.5keV to
low-energyenhancement,mostnotablyaroundtheOviihelium- 12.0keV).Timeperiodswithcountratesbeyondathresholdof
liketripletatapproximately0.56keV.ObservationsoftheGroth- ±1.5σ away from the peak of this Gaussian were removedby
Westfall Strip, Polaris Flare region and the Hubble Deep Field applyingaGoodTimeInterval(GTI)filetotheeventlists.The
as discussed in Kuntz&Snowden (2008) were included in the GTIfilescreatedforeachEPIC-MOSeventlistwerecombined
analysisofPaperIascontrolcasesforourmethod. togethertoformoneEPIC-MOSGTIfileandthisfilewasreap-
The XMM-Newton observatory(Jansenetal. 2001) has the pliedtoboththeMOS1andMOS2eventliststoprovidesimulta-
largest collecting area in the band 0.2 to 10keV of all X- neouscoverageduringeachobservation.Resolvedpointsources
ray telescopes currently in orbit and the European Photon (fromlistscreatedforthe2XMMcatalogue(Watsonetal.2009)
Imaging Camera (EPIC) suite of instruments on board (two usingaminimumlikelihoodthreshold≥6),wereremovedfrom
MOS(Turneretal.2001)andonepn(Stru¨deretal.2001)CCD the field of viewby extractingeventsin a circularregionof35
cameras)providemediumspectralresolution( E ∼17).During arcsecondsradiusaboutthesourceposition.Thoseobservations
∆E
variousperiodsofitsorbitanddependingonobservationalcon- judged, after a visual inspection (and prior to creation of any
straints,XMM-NewtonmayviewregionsoftheEarth’smagne- spectra, see Section 4), to show residual source contamination
tosheathwhicharepredictedtoexhibitthehighestX-rayemis- (thewingsofthepointspreadfunctionoftheEPIC-MOScam-
sivityduetoSWCXbetweenhighlychargedsolarwindionsand era beingevidentin an image of the eventfile) passed through
hydrogen in the Earth’s exosphere (Robertsonetal. 2006, and an additionalspatialfiltering stage using a largerextractionre-
references therein). We use data from the EPIC-MOS cameras giontofurthercleanthedataset.Amoredetaileddescriptionof
to identify times of variability in a low energyband thatis not thefilteringstepsandnatureofsoftprotoncontaminationcanbe
mirroredinahigherenergyband,toselectthoseperiodswesus- foundinPaperIandCarteretal.(2010).
pecthavebeenaffectedbyvariableSWCX. The method employed here followed the first steps as de-
SWCX emission can be considered as a contaminant scribed in detail within Paper I. In summary, two lightcurves
by sections of the community wishing to study Galactic or with bin size of 1ks were created for each observation from
Extragalacticsourcesbeyondtheheliosheath.Understandingthe eventswithinthefullfield-of-view(radiusof13.3arcminutes).
occurrence and nature of the emission is important therefore When both EPIC-MOS cameras were used during an observa-
for the process of elimination from astronomical observations. tion and employed the same filter, events from both cameras
However, SWCX emission can be used to provide diagnostics were usedto constructthe lightcurves.The firstlightcurvewas
of the solar wind. X-rays emitted from the comas of comets chosentorepresentthecontinuum,coveringeventswithenergies
have been suggested and used for this purpose (Cravens 1997; in the range2.5keV to 5.0keV. The second lightcurvewas ex-
Dennerletal. 1997; Lisseetal. 2001; Bodewitsetal. 2007). tractedusingeventswithenergiesintherange0.5to0.7keVto
A detailed study of the heavy ion constituents of a CME coverthestrongSWCXemissionfromOviiandOviii(theline-
was discussed in Carteretal. (2010). In addition, exospheric- bandlightcurve).Thisenergyrangeincorporatesemissionener-
SWCX may be used to test models of dynamical processes gies from the Ovii triplet and resonance lines which are dom-
within the Earth’s magnetosheath and the solar-terrestrial con- inated by the forbiddenline transition at 0.56keV. Lightcurves
nection,suchasflux-transfereventsorboundarylayerphenom- werethenexposure-correctedforperiodsremovedduringthefil-
ena(Collieretal.2010). teringsteps.AstheMOS1andMOS2eventfilesforeachobser-
The layout of the paper is as follows. In Section 2 we de- vation,priortothelightcurvecreation,werefilteredusingasin-
scribe the method employed to identify those observations of gleGTIfilewhichisnotenergyspecific,theexposure-coverage
interest.InSection3wepresenttheoverallresultsforthewhole for each bin was the same for the line-band and continuum
samplestudiedandcomparetheoccurrenceofSWCX withthe lightcurves. Therefore the same exposure-correction factor for
solar cycle. In Section 4 we describe fitting spectral models to anindividualbinwasappliedtobothlightcurvesinthisstep.We
cases of SWCX enhancement data. In Section 5 we discuss in keepbinsofthelightcurvewithatleast60%ofthefullexposure
moredetailseveralofthoseobservationsaffectedbySWCX.In forthatbinandrejecttheremainingbins(incontrasttoPaperI
Section 6 we describe a modelof the expectedX-ray emission whenwekeptallbinswithatleast40%exposure).Lightcurves
using the orbitof XMM-Newton and the solar wind conditions wererejectedfromfurtheranalysisiftheywerelessthan5ksin
atthetimeoftheobservation.Wefinishwithourdiscussionand length. By increasing the strictness of the bin coveragethresh-
conclusionsinSection7. olds,wereducedtheincidenceoftypeIerrors(thoseincorrectly
labelled detections), but ran the risk of increasing the number
ofcasesthatwereincorrectlylabelledasnon-detections(i.e.not
2. Dataanalysis
showing a deviance from the null hypothesisof a linear fit be-
Allobservationaldatausedinthisstudyareavailablethroughthe tween the line-band and continuum (type II errors). We scaled
XSAwhereweextracttheOriginalDataFiles(ODF).We con- eachlightcurvebyitsmeantoproduceadjustedlightcurves.The
siderobservationstakenbytheEPIC-MOS(Turneretal.2001) countratesfortheline-bandandthecontinuumbandwerethen
cameras (MOS1 and MOS2) using the full-frame mode only. always of the same orderwhich facilitated the identificationof
Thereforetheobservationsusedprovidedanevensampleacross
the mission and the event list data from the cameras could be 2 http://heasarc.gsfc.nasa.gov/docs/xmm/xmmhp xmmesas.html
J.A.Carteretal.:IdentifyingXMM-Newtonobservationsaffectedbysolarwindchargeexchange-PartII 3
periods of enhancementin the line-band lightcurve. An exam- Table1.Highestrankedobservationbyχ2.Thoseobservations
µ
pleofthecombined-MOSlightcurveadjustmentprocesscanbe with severe residual soft proton contamination have been ex-
foundinFigure1(panelstop-left,top-rightandbottom-left). cluded.
We plotted a scatter plot between the two bands (using the
adjustedline-bandasthedependentvariable),showninFigure1 Revn. Obsn. χ2 Comment
µ
(bottom-right).Alinearmodelfittoeachscatterplotwascom- 0369 0103461101 856.7 CometC2000WM1(LINEAR)
puted using the IDL procedure, linfit, which minimises the χ2 0808 0164960101 226.6 CometC2001Q4(Neat)
statistic. 0342 0085150301 27.2 PaperI&Carteretal.(2010)
Wecomputedthereduced-χ2forthefit,hereafterreferredto 0209 0093552701 23.0 PaperI
1014 0305920601 15.0 PaperI
as χ2, by dividing the χ2 by the number of bins minus one, to
µ 0690 0149630301 14.1 PaperI
accountfor the reduction in the numberof degreesof freedom
0623 0150610101 13.5 Newcase
made by fitting a linear modelto the data. A high χ2 indicates
µ 1177 0406950201 13.3 Comet73p
thatasignificantfractionofthepointsdeviatesignificantlyfrom 0339 0054540501 13.2 Newcase
the bestfit line. We expectedthese cases wouldbe morelikely 0422 0113050401 12.7 Newcase
toshowvariableSWCX-enhancement.Inadditionwecomputed
the χ2 valuesfor each individuallightcurvein termsof the de-
viationfromthemeanofthatlightcurve.Wecalculatedtheratio method described in this paper tested only for variable SWCX
between the line-band and continuum χ2 values to add to our onshorttimescales.SpectralanalysisofsuspectedSWCXcases
diagnostic(hereafterdenotedasR ). wasthereforeonlypossiblewhenaclearline-bandenhancement
χ
3012 observations made up the final sample to be used periodcouldbeidentifiedduringthedurationoftheobservation.
for further analysis. The results from Paper I showed us that We find 103 observations in our sample that show indica-
those observations exhibiting both high χ2 and high R were tions of a time-variable exospheric SWCX enhancement, that
µ χ
mostlikelytoshownear-Earthtime-variableSWCXsignatures. are not excluded from consideration based on the reasons de-
Observationsthatfulfilledthesecriteriawereconsideredforfur- scribed above. All of these cases had a χ2 value greater than
µ
theranalysis,spectrally,temporallyandwithregardtotheorien- or equal to 1.2 and a R value of greater than or equal to 1.0.
χ
tationofXMM-Newton. These cases make up only ∼20% of all observations that have
values of χ2 and R above these thresholds, indicating that al-
µ χ
though the values of χ2 and R are indicators of a SWCX-
3. Globalresults µ χ
enhancement, considerable inspection of an observation on a
Allobservationsin ourfinalsample,afteranyrejectionsasde- case-by-casebasisisstillrequired.Themajorityofcasesinthe
scribed below, were ranked by χ2. The two highest ranked ob- wholesample havea R valueless than1,indicatingthatthere
µ χ
servationswereobservationsofcomets.Althoughresolvedpoint is more variation seen in the continuum compared to the line-
sourceshavebeenremoved,cometaryX-raysarediffuseandwill band.Thecontinuumincorporatesthebreakenergyat∼3.2keV
likely be spread over a large fraction, if not all of the field of inthetwo-powerlawmodelsofresidualsoftprotoncontamina-
view. We assume that the dominant variations in the line-band tion (Kuntz&Snowden2008). For higherintensity soft-proton
lightcurve that result in such high ranking are due to SWCX flares,theslopeofthepower-lawbecomesflatter.Thereforethe
emission occurring within the cometary coma and not to any higherenergyandwidercontinuumwillhaveagreatervariance
emissionoccurringwithinthevicinityoftheEarth.We discuss thanthesofter,narrowerline-bandduetothepresenceofunfil-
thecometarycasesinSection3.3. teredresidualsoftprotons.
Afterrankingtheobservations,wewereabletostudythose Thetop10observationsasrankedbyχ2 aregiveninTable1
µ
thatexhibitedthehighestχ2µ andRχ inmoredetailonacaseby (excludingthose observationsthathad beenrejectedfromcon-
casebasis. sideration). Three of these top ten result from cometary obser-
Observationswere examinedfor residualpointor extended vations, four were identified in the work presented in Paper I
sources that may contribute to the high variations in the line- andthreearenew observationsdiscoveredduringthisanalysis.
band. Extended or diffuse residual sources that remain in the Manyofthehighestrankedcaseshadpreviouslybeenidentified
field ofviewwillnotaffectthisdetectionmethodprovidingno in the literature. It is our intention to make the full ranked list
inherent variation occurs within these sources, as expected for ofobservationsusedinthisanalysispublicallyavailableonline
sources outside the solar system, in either one of the bands. in the future, most probably through the XMM-Newton EPIC
Cases were also examined for residual soft proton contamina- BackgroundWorkingGroup(BGWG)3.Weincludeasummary
tion. Although the files used in this analysis have been filtered tablelistingtheSWCXcasesinTableA.1.
for periods of soft proton flaring, residual contamination may Ascatterplotofχ2versusR isgiveninFigure2.Thoseob-
µ χ
remain.Excessivescatterduetoresidualsoftprotoncontamina- servationsexhibitingbothhighχ2 andhighR butwhichdonot
tionimpedestheabilitytoidentifyperiodsofexosphericSWCX µ χ
showclearSWCXsignatures(i.e.anenhancementinthemean-
andwillresultinsometypeIIerrorsinoursample.Also,ifexo-
adjustedline-bandlightcurvecomparedtothatofthecontinuum)
sphericSWCXoccursthroughouttheentiretyoftheobservation
havebeenrejectedfromfurtheranalysis.We discussthe detec-
little ornovariationcanbeseeninthelineband,resultingina
tion methods of SWCX cases in the literature, other than by a
rejectionofthis case. Excessiveandsimultaneousvariationsin
searchforatime-variablelow-energycomponentinSection3.2.
boththe line-bandand continuumwillresultin a highχ2 yeta
µ CometaryX-rayemissionisdiscussedinSection3.3.
lowvalueofR .Weconcentrateouranalysisthereforeoncases
χ InFigure3weplotthetotalnumberofobservationsandthe
thatexhibitbothhighχ2µandhighRχ. fraction of observations that show SWCX enhancements, ver-
Forsomeobservationswithshortlightcurves,itwasimpos- sus the GSE-X position of XMM-Newton at the mid-point of
sible to identify any time-periods of boosted line-band emis- eachobservation.WecanseefromthisfigurethatXMM-Newton
sion(theputativeSWCXenhancementperiods).Theseobserva-
tionsweredisregardedasSWCX-enhancementcases.Also,the 3 http://xmm.vilspa.esa.es/external/xmm sw cal/background
4 J.A.Carteretal.:IdentifyingXMM-Newtonobservationsaffectedbysolarwindchargeexchange-PartII
0.4
1.0
-1ct s) 0.3 0.8
e (
unt rat e (ks) 0.6
d co 0.2 osur
e p
ust Ex 0.4
dj
a
n- 0.1
o 0.2
N
0.0 0.0
0 2 4 6 8 10 0 2 4 6 8 10
Time (hrs) Time (hrs)
1.8 χ2 : 10.6
µ
-1s) Rχ : 4.8
-1s) 1.5 d (ct 1.6
ct an
ate ( ne b 1.4
nt r 1.0 e, li
d cou nt rat 1.2
e u
Adjust 0.5 sted co 1.0
u
dj 0.8
A
0.0
0 2 4 6 8 10 0.8 0.9 1.0 1.1 1.2 1.3
Time (hrs) Adjusted count rate, continuum (ct s-1)
Fig.1.Lightcurvecorrectionprocedureexampleforobservationwithidentifier0150680101(line-band(black),continuum(red)for
panelstop-leftandbottom-left).Topleft:examplelightcurvesshowingapeakintheline-bandthatisnotreflectedinthecontinuum.
Topright:exposurecoverageforeachbin,thethresholdat60%ismarkedbythereddashedline.Bottomleft:lightcurvesafterthe
adjustmentsforexposurecorrectionandscalingbythemean.Bottomright:examplescatterplotforthisobservation.
off distance, at approximately 10R . The XMM-Newton line-
E
of-sight for these cases traversed the subsolar region (sunward
800
n 600 side) of the magnetosheath,as expectedaccording to the mod-
s
ob 400 ellingworkofRobertsonetal.(2006).XMM-Newtonviewingis
No. 200 constrainedbythefixedsolarpanelsandlimitsimposedtoavoid
0 directlyobservingtheSun,MoonandEarthandisonlyableto

100.0 viewthesunwardsideofthemagnetosheathatcertaintimesof
theyear(Carter&Sembay2008).
We also consider the seasonal variation of the occurrence
10.0
of SWCX in this sample. 64 SWCX cases occurredduring the
χ2µ summermonths(ApriluntilSeptemberinclusive)and39during
thewinter(OctoberuntilMarch).MoreexosphericSWCXcases
1.0
areexpectedinthesummermonths,asviewedbyXMM-Newton
(Carter&Sembay2008).
0.1
0.1 1.0 10.0 0 200 400 600 800
Rχ No. obsn 3.1.Relationshipwiththesolarcycleandsolarwind
Fig.2. χ2 versus R . Red filled circles indicate those observa-
µ χ InFigure4weplotthesolarsunspotnumber4fromthelatterhalf
tions with time-variable SWCX signatures. Histograms of the
of Solar Cycle 23 and mark the times at which XMM-Newton
χ2 andR valuesareshowninthesidepanels(usinga binsize
µ χ observationswithknownSWCX(thispaper,paperIandXMM-
of0.25).
Newtonexospheric-SWCXcasesintheliterature)occurred.We
plot a histogram of the number of SWCX cases per half year,
to remove any bias resulting from the seasonal constraints on
is preferentially found on the subsolar side of the Earth when pointing angle experienced by XMM-Newton. Each histogram
SWCX enhancementsoccur. In addition, the largest bin of the
fractional plot occurs around the nominal magnetopause stand 4 http://www.sidc.be/index.php
J.A.Carteretal.:IdentifyingXMM-Newtonobservationsaffectedbysolarwindchargeexchange-PartII 5
500 0.10 portionoftheirsetaffectedtodifferentdegreesbySWCXemis-
sion.Bautzetal.(2009)inferredSWCX-enhancementsinobser-
et vationsbySuzakutowardstheclusterAbell1795afterexamina-
400 0.08CX s tionofalow-energylightcurverevealedpeakscoincidentintime
W with enhancements in the solar wind proton flux as measured
S
300 0.06n by ACE. Henley&Shelton (2010) used a large set of XMM-
bsn sn i NewtonobservationstocompareintensitiesfromOviiandOviii
o b
No. 200 0.04no. o lTinheesyffironmd nseotsunoifvoerbssaelrvaastsioocnisatwiointhbtehtewseaemneentahragnectepdoiSnWtinCgXs.
of
n emission and the closeness of the the line-of-sight to the sub-
o
100 0.02acti solar region of the magnetosheath. In this paper we do see a
Fr tendencyforXMM-Newtontobeclusteredaroundthesub-solar
regionfortheSWCXcases,asdiscussedinSection3.
0 0.00
OneoftheHenley&Shelton(2010)SWCX-enhancedcases
-15 -10 -5 0 5 10 15
GSE-X (R ) is part of the data set used in this paper, however this was not
E
detected by our method as no discernable variability occurred
Fig.3.Totalnumberofobservations(black)versusGSE-X po-
duringtheobservation.CasesofSWCXoccurringintheEarth’s
sitionandthefractionofobservationsdetectedwithexospheric
exosphere,identified by detectingtime-variableemission, have
SWCXenhancements(red).
alsobeenobservedbySuzaku(Fujimotoetal.2007;Ezoeetal.
2010).Carteretal.(2010)useda previousobservationofa tar-
er 200 getfieldtoconstrainthediffuseX-rayemissioninherenttothat
b
m 150 lookdirectiontocalculatethestrengthofSWCXemissionlines
u
pot n 100 adsastoacsieattepdrewseinthteadCinMthEepparsessienngtipnaptheercvoicnitnaiitnysoafnthuemEbearrtohf.pTrhee-
s 50
n viously knowncases of SWCX, however,not all of these were
u
S 0 timevariableandthereforewerenotpickedupbythetechnique
2000 2002 2004 2006 2008 2010
usedhere.Theydidhoweverbecometests oftheabilityofthis
ear 0.12 Year methodtoidentifySWCX-affectedobservations.Onelimitation
alf y 0.10 12 136 SWuinmtemre r of this technique is that SWCX emission occurring at an ap-
per h 0.08 188 184 143 praryoxeimmiastseiloynstweaildlynsottatbeeasidpeanrttifioefda.nyInqaudiedsictieonnt,gaenoycoorbosnearlvXed-
CX 0.06 128 159238 169 quiescent heliospheric SWCX will in general be several times
W 0.04 198
Fract. S 00..00202000 2002 2107034 156 2108006189157169210806817272 3 2010 tAshtrleothinnogcuergrehathsaeacdnoimanntbyeignqraauttiiioeosnncoelfenntteggctehhonsciioqnruvoeonslavwleodXu(-lCrdarybaeveeimdneissaeslit,otanhl.,is2di0us0en1to)ot.
Year possibleforthevastmajorityofsingleXMM-Newtonpointings.
The greatest advantage of this method is that observations are
Fig.4. Top panel: sunspot number versus time. Bottom panel:
considered on an individual basis. An XMM-Newton user can
the coloured histogram of the fraction of observationsaffected
makeajudgementastowhetherextracautionisrequiredwhen
by exospheric SWCX is binned into six month periods (blue -
analysingtheirresultsforpossibleSWCX-contamination.
summer, black - winter). The total number of all observations
foreachperiodisnotedabovethebin.
3.3.Cometaryemission
bin starts at the start of summer (1st April) or start of winter CometHyakutakewasthefirstcometwhoseX-rayemission,as
(1stOctober).AsexpectedforcasesofexosphericSWCX,there observedbyROSATandRXTE(Lisseetal.1996),wasassigned
aremorecasesattimesofhighsolaractivityaroundsolarmaxi- to the SWCX emission process (Cravens 1997). This SWCX
mumthanwhenapproachingsolarminimum.Also,forthecases emissionoccursfromtheinteractionofthesolarwindwithneu-
from 2002 onwardsand approachingsolar minimum,we see a tralspeciesthatoutgasfromthecometasitenterstheinnerso-
higherproportionofSWCXcasesinthesummersix-monthpe- lar system, and the amount of out-gassing is dependent on the
riodcomparedtothewinterperiod,forthesame year.Thesig- comet’sdistancefromtheSun.Theseneutralspeciesaremainly
nificanceof thistrendhowevershouldnotbe overstateddue to water and its dissociation products.The SWCX emission must
thelownumberofcasesineachbin. occurincometaryregionswherephotoionisationanddestruction
of the neutralspecies can occur.Water and hydroxylions have
shortlifetimeswhenexposedtosolarUVphotonsandtherefore
3.2.Multiplepointingsoftargetfields survive the longest in the coma interior, whereas the dissocia-
tionproductsofwater(alongwithCOprovidingthecomethasa
Multiplepointingstowardsthesametargetallowonetocompare
sufficientlyhighcarbonabundance)cansurvivefurtherintothe
diffuseandextendedemissionspectralmodelsthatmayexhibit
outer coma regions. A detailed description of X-ray emission
spectralvariationsindicativeofSWCXcontamination.Thelong
from comets, primarily using data from the Chandra observa-
termenhancementsoftheROSATall-skymaps,whichweresub-
tory,canbefoundinBodewitsetal.(2007);Bodewits(2007).
sequentlyattributedto SWCX, werefirstidentifiedbycompar-
isons between fields (Snowdenetal. 1995). Kuntz&Snowden SeveralXMM-Newtonobservationsofcometswereincluded
(2008) examined multiple XMM-Newton observations of the inthesampleinthispaper.Theχ2µ andRχvaluesforthecomets
HubbleDeepField,amongstothertargets,andidentifiedapro- are given in Table 2. The highest overall values of χ2 and R
µ χ
6 J.A.Carteretal.:IdentifyingXMM-Newtonobservationsaffectedbysolarwindchargeexchange-PartII
Table 2. XMM-Newton observation of comets within the data 25
set. 1200 -20)
1
x
R02e0v9n. O01b0s3n4.60901 NMacmNeaught-Hartley χ1.2µ5 R1.χ0 1000 206+O (
00376199 00110631476610110011 CC2o0m0e0tW2pM(E1n(cLkIeN)EAR) 28.506.7 12.959.9 -1s s) 800 157+nd O/
0720 0113041301 C2001Q4(Neat) 1.2 0.9 e (ct 70) a
01810787 00146046996500120011 CC2o0m0e1tQ734p(Neat) 1232.63.6 1374.92.77 unt rat 600 10-1s x 1
Co -2m
400 c
x (
u
occurredduringobservationsofcomets.Exampleline-bandand 5 n fl
200 o
continuumlightcurvesfrom comet C2001 Q4 (Neat) (observa- Prot
tion 0164960101) are shown in Figure 5, where the line-band
0 0
lightcurve clearly dominates and is highly variable. Also, an 0 2 4 6 8 10 12
earlier observation of the same comet exhibited very different Time (hr)
values of χ2 and R . Although the upwind solar wind moni-
µ χ Fig.5. Lightcurve from comet C2001 Q4 (Neat) that resulted
tor Advanced Composition Explorer (ACE level 2, combined in the highest χ2 values (black - SWCX band, red - contin-
instrumentdata, Stoneetal. (1998)) protonflux, also shown in µ
uum band). The time axis is given in hours since the start of
this graph,is steady andnotremarkablein intensity,the comet
the XMM-Newton lightcurves. The solar wind proton flux as
will more likely be sampling solar wind that originates from a recordedbyACEisgiveninblue.TheO7+toO6+ratioisplotted
differentlocationin the solar corona.In Figure 5 we have also
usingtheprotonaxisandisshownbythebluedashedline.
plotted the O7+ to O6+ ratio, using data from the ACE SWICS
instrument.Thisratiovariesslightlyoverthelengthoftheobser-
vationbutanyfluctuationsarenotreflectedintheXMM-Newton evidence of SWCX enhancement. Therefore, we have no rea-
line-bandlightcurve,furtherimplyingthatthecometissampling sontorejectthisobservationfromtheSWCXset.Wewerecon-
a different solar wind to that seen by ACE. The solar wind is cerned that the planet’s movementthroughand possibly outof
a collisionless plasma and so its ion composition remains un- thefieldofviewmayhavecausedthiseffect.However,thedom-
changedasitflowsawayfromtheSun.SignaturesofSWCXoc- inantmovementisinrightascensionwithamaximumspeedof
curringthroughoutthesolarsystemcanthereforebeusedtoin- 8arcsecondsperhour(Branduardi-Raymontetal.2010),result-
ferthecompositionofthesolarwind,whichvariesconsiderably inginashiftofonly0.78arcminutesoverthecourseofthe21ks
throughoutthesolarcycleandwithsolarlatitude.Ascometary observation. In addition, resolved sources have been removed
orbitsarenotrestrictedtotheeclipticplanetheyareidealloca- fromthefieldofviewasforallotherobservations.
tionstostudycompositionalsignaturesfromsolarwindoriginat-
ingfromavarietyofsolarwindlatitudes(Dennerletal.1997).
4. Spectralanalysis
Bodewitsetal. (2007) was able to use cometaryX-rays to dis-
tinguishemissionresultingfromthreesolarwindtypes:thecold
We continue our study by observing the spectral signatures of
and fast wind, the warm and slow wind and the warm and dis-
new SWCX cases identified in this paper and in Paper I. We
turbed wind. A more complete discussion of cometary X-ray
omitthoseSWCXobservationsthathavebeencomprehensively
emissionisbeyondthescopeofthispaper.
investigatedintheliteratureyetshownotemporalvariabilityin
our tests. We also omitobservationsof comets. This set of ob-
servationswhichwehaveusedforfurtherstudythroughoutthis
3.4.Planetaryemission
paper is hereafter known as the SWCX set and comprises 103
We include within the SWCX set an observation of the planet observations.
Saturn, taken on 1st October 2002. This observation has been TheScienceAnalysisSystem(SAS)software(version9.0.0;
comprehensivelyanalysedbyBranduardi-Raymontetal.(2010) http://xmm.esac.esa.int/external/xmm data analysis/) was used
whoattributetheX-rayfluxtoemissionfromtheplanetarydisk, to produce spectral products and instrument response files for
producedbythescatteringofsolarX-raysandanadditionalflu- all observations of the SWCX set. The SAS accesses instru-
orescentemissionlineofoxygenat∼0.53keVoriginatingfrom mentcalibrationdatainso-calledcurrentcalibrationfiles(CCFs)
therings.ThisobservationwaspreviouslystudiedbyNessetal. which are generally updated separately from SAS release ver-
(2004) but they did not investigate any low-energy variability. sions.InthispaperweusedthelatestpublicCCFsreleasedasof
We find a high χ2 (2.9) and R (2.3) in our time variability February2010.
µ χ
testandadistinctstepintheline-bandlightcurve,indicativeof Foreachexospheric-SWCXcaseweextractedspectraforthe
a SWCX enhancement, after the same filtering steps as to all EPIC-MOS cameras for the suspected SWCX-affected period
other data sets have been applied, including source exclusion and for the suspected SWCX-free period. The SWCX-affected
to remove emission from the planet and planetary exosphere. period was when the enhancement in the line-band lightcurve
The initial source exclusion radius equates to ∼ 3.7R . was judged (however not by a formal mathematical argument)
Saturn
Spectra from this observation were extracted from event files to have occurred.The enhancementcould have occurredat the
thathadbeenadditionallyfilteredwithalargerextractionregion beginning,middleorendoftheobservation.Theremainingtime
of ∼ 10.6R . Even after this additional source extraction periodsin the observationmadeuptheSWCX-free period.We
Saturn
stepthelightcurveproductionprocedureyieldsmetricsofχ2and usedeventsfromacircularextractionregion,centredondetector
µ
R of3.1and2.4respectively.TwolaterobservationsofSaturn coordinatepositions(DETX,DETY=-50,-180),withanextrac-
χ
taken in 2005 were included within our sample but showed no tion radiusof 16000detectorunitsor 13.3arcminutes.We also
J.A.Carteretal.:IdentifyingXMM-Newtonobservationsaffectedbysolarwindchargeexchange-PartII 7
Table3.Principleionspeciesemissionlinesusedinthemodel,
plusanyminoremissionlineenergiesused(seetext). 8
7
Ion Energy(keV) Minorenergies(keV)
Cv 0.299 0.304,0.308,0.354,0.379 6
Cvi 0.367 0.436,0.459,0.471 es 5
s
Nvi 0.420 0.426,0.431,0.523 ca
Nvii 0.500 0.593,0.625,0.640,0.650 o. 4
N
Ovii 0.561 0.569,0.574,0.713,0.666,0.698,0.723 3
Oviii 0.653 0.775,0.817,0.837,0.849
Nex 1.022 2
Mgxi 1.330 1
Sixiv 2.000
0
0 10 20 30 40 50
X-ray flux (keV cm-2 s-1 sr-1)
appliedtheflagandpatternselectionexpression’#XMMEA EM Fig.6.Histogramoftotalspectrallyfittedfluxbetween0.25and
&&PATTERN<=12&&FLAG==0’.Thispatternselectionse- 2.5keVfortheSWCXset.
lectseventswithinthewholevalidX-raypatternlibraryforthe
EPIC-MOS and the flag selection removes events from or ad-
WeusedVersion12.5.0oftheXSPEC5X-rayspectralfitting
jacenttonoisypixelsandknownbrightcolumns.We produced
packagetoperformthisanalysis.Wefittedthemodeltoeachdif-
instrumentalspectralresponsefilesforeachperiod.Theinstru-
ference spectrum by minimisingthe χ2 statistic. We calculated
ment effective area files were calculated assuming the source
the modelled flux and 1-sigma errorson the flux between 0.25
flux is extended, filling the field of view and with no intrinsic
and2.5keVforeachEPIC-MOSinstruments.Thekeywordgiv-
spatialstructure.
ing the area collected in the spectral extraction (BACKSCAL)
was converted into units of steradians and used to convert
4.1.Spectralmodelling the individual flux values to units of keVcm−2s−1sr−1. Fluxes
presented here onwards are for a combined error weighted-
We knew from the work of Carteretal. (2010) that exospheric
average EPIC-MOS flux. The error on the flux was calculated
SWCX can occur throughout the entirety of an observation,
fromthe individualflux errors,combinedin quadrature.A his-
although the line-band lightcurve may show an enhanced and
togramof the total spectrally fitted flux for each of the SWCX
a steady-state period. We used the spectra from the apparent
set can be seen in Figure 6. The minimum flux we observed
SWCX-affected period as the source spectra and that from the for a SWCX case was 2.2keVcm−2s−1sr−1 (observation id.
apparentSWCX-freeperiodasthebackgroundtoproduceadif- 0112490301)and the maximum 50.1keVcm−2s−1sr−1 (obser-
ferencespectrum.Thedifferencespectrumthereforeprovidesa
vationid.0085150301,Carteretal.(2010)).
lower limit to the SWCX enhancement that has occurred dur-
Therelativestrengthsofthecomponentlinestothe SWCX
ing an observation. Providing the particle-induced background
spectral model varied considerably within the SWCX set.
is reasonably constant over the duration of the observation (at
Individual line fluxes were calculated by finding the best fit
most±10%(DeLuca&Molendi2004))thisfactorwillbeelim-
modelusingalllines,asdescribedabove,thensettingotherline
inated.Aninspectionofthedifferencespectrawasmadeforen-
normalisationstozeroinXSPEC.Thefluxfortheindividualion
ergies above 2.5keV, to check for the presence of significant
speciescontributionwascalculatedintherange0.25to2.5keV.
variable residual soft-proton contamination. Each SWCX-case HighO7+ to O6+ andmagnesiumto oxygenratiosare used
showedacountratestatisticallyconsistentwithzeroabovethis
asindicatorsofthepresenceofCMEplasma(Zhaoetal.2009;
energy. Richardson&Cane2004;Zurbuchen&Richardson2006).Ovii
We modelled the resulting difference spectrum for each
is the dominantSWCX ion-speciesin the majority of cases. In
SWCXcasewithastandardmodelofemissionlines.Thespec- Figure 7 we plot the ratio of the fluxes of the lines Mgxi/Ovii
trum from MOS1 and MOS2 were fitted simultaneously, al- to Oviii/Ovii (using those oxygen transitions available to us
thoughaglobalnormalisationparameterfortheMOS2spectrum
within our X-rayspectral band).To look for plasma signatures
was allowed to vary.The relativeline strengthsfor a particular
with the highest charge states we only plot those cases where
ionspeciesbelow1keV,forexampleOvii(whichinvolvesseven
the normalisation of the numerator in the ratio is well con-
separatetransitionsincludingtheOviitriplet),weresetusingthe
strained. Three observationsare constrained to have both a ra-
velocitydependentcross-sectionsoflaboratorychargeexchange tio of Mgxi/Ovii > 0.6 and Oviii/Ovii > 1.0. One of these
collisions between highly charged ion and atomic hydrogen,
(withidentifier0085150301)wastheobservationpreviouslyas-
as found in Bodewits (2007). We assumed a solar wind speed
signed to a passing CME and described in Carteretal. (2010).
of 400kms−1 forthese cross-sections.We also addedemission
These observations are therefore possible candidates for hav-
lines from Nex at 1.022keV, Mgxi at 1.330keV and Sixiv at
ing observed CME plasma with XMM-Newton and are listed
2keV.Theremaybeemissionfromotherionspeciespresentin
in Table 4. We quote the lower limit to the ratio in the case
thespectra,such asfromhighlychargedironor aluminium(as wheretheOviifluxisbadlyconstrained,i.e.veryweak.Inthis
seeninCarteretal.(2010)),butwewishedtosimplifythemodel tablewealsoquotethemeanvalueoftheO7+ toO6+ ratiodur-
appliedtoageneralcaseandthedominantSWCXemissionlines
ing the period of the observation, using values taken from the
are found below 1keV. We fixed the relative normalisationsof
ACE SWICS instrument. This data was only available in two
the minor transitionsto that of the principaltransitionfor each
of the cases in the table. As ACE is foundat Lagrangianpoint
ionspecies.Theprincipal,dominanttransitioninthecaseofCv,
L1,wehavetime shiftedthe solarwinddatatoaccountforthe
NviandOviiistheforbiddenlinetransition.Theprincipalion
transitionsusedinthismodellingcanbeseeninTable3. 5 http://heasarc.gsfc.nasa.gov/docs/xanadu/xspec/index.html
8 J.A.Carteretal.:IdentifyingXMM-Newtonobservationsaffectedbysolarwindchargeexchange-PartII
Table 4. SWCX set observations exhibiting the highest
Mgxi/OviiandOviii/Oviiratios,orthelowerlimit(95%con-
fidence) to this ratio when Ovii is badly constrained. We also
navoateilathbeleA, wCEithStWheICstSanmdeaardnvdaelvuieatoiofnthoefOth7i+stroatOio6+girvaetinoawshthene OVII 1.0
error. XI/
g
M
Revn Obsn Mgxi/ Oviii/ Mean s:
e
Ovii Ovii O7+/O6+ ux
0342 0085150301 2.5±1.9 8.3±6.4 0.58±0.54 o fl
0494 0109120101 ≥0.60 ≥2.26 1.53±0.83 ati
R
0747 0200730401 ≥1.09 ≥2.49 ...
0.1
traveltime to Earth, based on the mean speed of the solar pro- 0.1 1.0 10.0
tons during the XMM-Newton observation. Expected O7+/O6+ Ratio fluxes: OVIII/OVII
ratiosfortheslowandfastsolarwindare0.27and0.03respec- Fig.7.RatioofMgxi/OviitoOviii/Oviiwhereavailableforthe
tively (Schwadron&Cravens 2000). Both of the observations SWCXset.Whereappropriatewemarkthelowerlimit(red).
with ACE SWICS data surpass boththe nominalslow and fast
values by a considerable margin and would suggest that ACE
detected a CME plasma. Although the identification of CME
plasma generally involves many more criteria to be satisfied,
XMM-Newton could provide supplementary spectral evidence
to studies employing in-situ dedicated solar wind monitors in -1sr) 40
thefieldofsolarsystemspacescience. -1s
To test if any relationship exists between the flux of the -2m
c
SWCX linesandincreasedsolarwindflux,weplotinFigure8 eV 20
the observed flux versus the difference in the mean solar wind k
protonflux(asmeasuredbyACE)betweentheSWCX-affected ux (
and SWCX-free periods. We have again time-shifted the ACE y fl
a
data to account for the distance between L1 and the Earth. X-r 0
Although there is considerable scatter amongst these values,
there is a positive correlation between line flux and solar wind
proton flux. We include in the plot the linear fit as found by
the IDL procedure linfit. Proton flux, for our SWCX set, is a 0 50 100
Solar wind proton flux ( cm-2 s-1× 107)
good indicator of the presence of SWCX-enhancement, if not
thelevelofthisenhancement.Thisisincontrastwiththeresults Fig.8. Observed flux versus mean solar wind proton flux. The
of Henley&Shelton(2010), whose SWCX cases were consid- reddottedlineindicatesalinearfittothedata.
ered to be due to SWCX occurring within the heliosphere and
thereforenocorrelationwouldbeexpectedbetweenanupstream
solarwindmonitorandanySWCX enhancement.Heliospheric
tothedata,asdescribedinSection4.Individualfluxesforase-
SWCXisexpectedtovaryonlongertimescalesthanexospheric
lectionofprominentlinesaregivenforeachcaseinTable5.
SWCX and therefore will be harder to identify by the tech-
niqueinthispaper.However,atcertaintimesoftheyearXMM-
Newtonmayhavealine-of-sightthatpassesthroughthehelium – Observation0150610101(revolution0623)
focusing cone, that could potentially produce a variable signal Theline-bandlightcurveshowsaperiodof enhancedcount
in the line-bandthatmay bedetectableby thistechnique(vari- rateatthebeginningoftheobservation.TheACEsolarpro-
ationsovera fewhours,(Koutroumpaetal.2007)).We discuss ton flux is raised at the beginningof the lightcurve and re-
this possibility further in Section 6.1. The flux variations seen ducesasthelightcurveprogresses.Thedifferencespectrum
withinourSWCXsetarethereforeduetolocalX-rayemission exhibitsemissionatOvii,Oviii,alongwithevidenceofcar-
inthevicinityoftheEarth. bonemissionbelow0.5keV.Thefluxobservedbetween0.25
and2.5keVwas20.6keVcm−2s−1sr−1.
– Observation0054540501(revolution0339)
Theline-bandlightcurveshowsan enhancementduringthe
5. ExamplecasesofSWCXenhancement
latter partof the observation,which is also observedin the
In this section we comment on the three newly identified exo- ACE solar proton flux. The difference spectrum exhibits
spheric cases from Table 1. Lightcurves for each, over-plotted emissionintheoxygenband,alongwithevidenceofcarbon
with the solar proton flux as recorded by ACE, are given in emissionbelow0.5keV.Thefluxobservedbetween0.25and
Figure9.Thesolarprotonlightcurveshavebeenadjustedforthe 2.5keVwas7.9keVcm−2s−1sr−1.
distancebetweenACEandtheEarthbyaddingadelaybasedon – Observation0113050401(revolution0422)
thedistancetoACEandthemeansolarprotonspeed,assuming Theline-bandlightcurveshowsaperiodof enhancedcount
aplanarwavefronttravellingontheSun-Earthaxis.Wealsoplot rate at the beginning of the observation. A short enhance-
a difference spectrum for each observation by combining data ment period is seen in the ACE solar proton flux and the
frombothEPIC-MOScamerasandover-plotthebestfitmodel overallmagnitudeofthisfluxismuchhigherthantheother
J.A.Carteretal.:IdentifyingXMM-Newtonobservationsaffectedbysolarwindchargeexchange-PartII 9
two cases in this section. The flux observed between 0.25 – Using the magnetopause location as a guide, we approxi-
and2.5keVwas25.9keVcm−2s−1sr−1. mate the position of the bow shock. We base the shape on
asimpleparabolaandcalculatethebowshockstandoffdis-
tance using the solar wind pressure and the relationship in
6. Modellingtheexpectedemissivity
Khan&Cowley (1999). The magnetosheathand bowshock
TheexpectedX-rayemissivityofSWCXemissionfromtheso- togetherdefinethemagnetosheathregion.
lar wind interaction with the magnetosheath can be estimated – We create an Earth-centric square image for use in subse-
fromthe integratedemission alongthe line-of-sightfor the ob- quentsteps.Thesidelengthoftheimageis200RE.Thisim-
server.Wehavedevelopedamodel,applicabletolocalinterplan- ageisdividedintocells,withsidelength0.5RE.Themagne-
etary space, to calculate this emission. We have not attempted topauseshapeisprojectedontothisimage.Weareabletouse
toincludeanycontributionfromfurtherintotheheliosphere,as animageratherthanacubeduetotheassumptionmadepre-
to increase the integration length would lead to greater uncer- viouslyregardingthesymmetryofthemagnetosheathshape
tainty in the underlying parameters of the solar wind on large abouttheGSE-Xaxis.
spatial scales. We assume that the solar wind parameters used – Wefindtheneutraldensityofhydrogenatomsforeachcell.
inthemodelareapproximatelyconstant(excludingthemagne- WeusetheØstgaardetal.(2003)modelforneutralhydrogen
tosheathregion)alongtheline-of-sight.Theemissivityexpected densityprofilesaroundtheEarth,butlimitthistoaminimum
(Cravens2000)isgivenbytheexpression: densityof0.4cm−3(Fahr1971).
– We determine the line-of-sightof the XMM-Newton point-
P =αη η hgieVcm−3s−1 (1) ing through the grid by extracting the relevant information
χ SW n
fromtheODFandconvertingthepositionsandtargetpoint-
whereαistheefficiencyfactordependentonvariousaspects ingdirectiontotheGSEcoordinatesystem.
ofthechargeexchangesuchastheinteractioncross-sectionand – We find the velocityand density of the solar wind for each
theabundancesofthesolarwindions,ηSW isthedensityofthe cell, (Spreiteretal. (1966), K.D. Kuntz private communi-
solar wind protons, ηn is the density of the neutral species and cation). As the solar wind passes the bowshock and enters
hgiistheirrelativevelocity. themagnetopauseitsdensityincreases(byaboutafactorof
ForeachobservationunderstudyinpaperII,wewishtotest fourinthesubsolarregion,ascomparedtotheunperturbed
whether any relationship exists between the total SWCX flux value),andthevelocitydropstoaboutonetenth.
seenandthetheoreticalintegratedX-rayemissionalongtheline- – The value of α is dependent on the abundances of the ion
of-sight(basedonEquation2,Cravens(2000)). species contributing to the charge exchange process, along
with the cross-section and energy of each interaction with
1 ∞
F = P dseV cm−2s−1sr−1 (2) the neutral donor in the energy band of interest. The neu-
4πZ χ
0 tral donor is hydrogen in the geocoronalcase. The relative
abundancesfoundin the solar wind varyconsiderablywith
We take data describing the conditions in the solar wind
solar wind state. The composition of the solar wind gener-
from ACE (Level2 processed data, merged instrument data us-
ally follows abundances seen in the photosphere, but can
inghourlyaverages)atthetimeofeachobservation.Weneeded
vary by up to a factor of about 2 (fast wind) or 4 (slow
toapplyadelayto thesignalreceived,toaccountforthesepa-
wind) for elements with first ionisation potential (FIP) be-
rationbetweenACEandtheEarth.Thisdelaywillbetimevari-
lowtheLyman-αlimitof10.2eV(Richardson&Cane2004,
able and will depend on the speed and orientation of the solar
and references therein). However, we use the slow solar
wind.However,asafirstapproximation,wehavetakentheav-
wind abundances for an ion species with respect to oxy-
eragesolarprotonspeedofthedataandassumedaplanarwave-
gen, as listed in Schwadron&Cravens (2000). We use an
front travelling anti-sunward perpendicularto the GSE-X axis.
oxygen to hydrogen ratio of 1/1780 for solar wind speeds
Wecalculatethedelayrequiredforthewavefronttotravelfrom
of ≤ 650kms−1 or 1/1550 for speeds above this threshold.
ACEtotheEarth.
For this modelling we consider contributions to the emis-
Throughoutthisworkweassumeageocentricsolar-ecliptic
sionfromtheprincipalandminortransitionsasdescribedin
coordinatesystem(GSE),wherepositiveXisdirectedfromthe
Section 4.1. Cross-sections for charge exchange transitions
EarthtotheSun,positiveYopposesplanetarymotionandposi-
are dependenton solar wind speed. We calculate an α map
tiveZisparalleltothedirectiontowardsthenortheclipticpole.
with the same dimensions and binning as that of the Earth
Thenforeachtimebinofanobservation:
gridandpopulatethismapwithvaluesofαdependingonthe
– We extract the solar wind proton velocity and temperature speedof the solar wind,unperturbedoutside ofthe magne-
fromtheACEdata,andfortheseparameterswecalculatea tosheathorperturbedinsidethemagnetosheathasdescribed
thermal velocity and average speed, using Equations3 and above.
4. – Wemultiplythesolarwindvelocity,solarwinddensityand
neutralhydrogendensitytogetherforeachofthecellsinthe
ν = 3k T (3) line-of-sightandmultiplythisvaluebytheefficiencyfactor
th b
p αforeachcell.Thisistheemissivityofeachcell.
hgi= ν2 +u2 (4) – Wesumallcellsintheline-of-sight,accountingforthenum-
q th sw berofcellsincludedintheintegral,togivethevalueofthe
– Weestimatethepositionofthemagnetopause,basedonthe emissivitymetric,approximatingEquation2.
model of Shueetal. (1998). To do this we use information
Therearesomeknownlimitationstothismodel,suchas:
regarding the strength and direction of the interplanetary
magneticfield(B component).Wecurrentlyassumethatthe – There are no magnetosheath cusps (increased density or
z
magnetopause shape is symmetrical about the GSE-X axis modifications to the velocity of the solar wind specific to
andplacethemagnetopausestandoffdistancealongthisaxis. theseregions)includedintheSpreiterapproximation
10 J.A.Carteretal.:IdentifyingXMM-Newtonobservationsaffectedbysolarwindchargeexchange-PartII
0.40
140
0.35
120
70)
-1s s) 0.30 100-1s x 1
Count rate (ct 00..2205 468000 -2Proton flux (cm
0.15
20
0.10 0
0.0 0.5 1.0 1.5 2.0 2.5
Time (hr)
1.0
140
0.8 120
70)
-1s s) 0.6 100-1s x 1
Count rate (ct 0.4 468000 -2Proton flux (cm
0.2
20
0.0 0
0 1 2 3 4
Time (hr)
1.0
140
120
0.8 70)
-1s s) 100-1s x 1
Count rate (ct 00..46 468000 -2Proton flux (cm
20
0.2 0
0 1 2 3 4 5 6
Time (hr)
Fig.9.ExamplecasesofSWCXenhancement,for(toprow)observationid.0150610101,(middlerow)observationid.0054540501
and (bottom row) observationid. 0113050401.Each left-handpanelshows the line-band (black)and continuumband (red) non-
mean-adjustedlightcurvesalongwiththesolarprotonflux(blue,right-handy-axis).ThesplitbetweentheSWCX-freeandSWCX-
affected period is indicated by the dashed vertical line. In the right-hand panels we show the combined EPIC-MOS difference
spectrumandthemodelfittedtothedataforeachcase(solidline).
Table 5. Most prominent ion line fluxes for example cases. Fluxes are quoted in units of keVcm−2s−1sr−1. Upper limits (95%
confidence)aregivenforveryweaklines.
Revn Obsn Cvi Nvi Ovii Oviii
0623 0150610101 4.86±1.10 2.19±1.48 7.73±0.37 2.42±0.61
0339 0054540501 5.61±0.85 ≤3.50 11.73±0.27 4.74±0.41
0422 0113050401 9.05±0.41 ≤2.80 11.60±0.27 4.36±0.46