Table Of Content(DOI:willbeinsertedbyhandlater)
Coronal properties of G-type stars
in different evolutionary phases
L.Scelsi1,A.Maggio2,G.Peres1,andR.Pallavicini2
1 DipartimentodiScienzeFisicheedAstronomiche,SezionediAstronomia,Universita` diPalermo,PiazzadelParlamento1,
90134Palermo,Italy
2 INAF-OsservatorioAstronomicodiPalermo,PiazzadelParlamento1,90134Palermo,Italy
Received,accepted
5
0 AbstractWereportontheanalysisofXMM-NewtonobservationsofthreeG-typestarsinverydifferentevolutionaryphases:
0 the weak-lined T Tauri star HD 283572, the Zero Age Main Sequence star EK Dra and the Hertzsprung-gap giant star 31
2 Com. Theyallhave highX-rayluminosity (∼ 1031ergs−1 for HD283572 and 31Comand ∼ 1030ergs−1 forEK Dra).We
n compare the Emission Measure Distributions(EMDs) of these active coronal sources, derived from high-resolution XMM-
a Newtongratingspectra,aswellasthepatternofelementalabundances vs. FirstIonizationPotential(FIP).Wealsoperform
J time-resolvedspectroscopy ofaflaredetectedbyXMMfromEKDra.Weinterprettheobserved EMDsastheresultofthe
8 emissionofensemblesofmagneticallyconfinedloop-likestructureswithdifferentapextemperatures.Ouranalysisindicates
2 that the coronae of HD 283572 and 31 Com are very similar in terms of dominant coronal magnetic structures, in spite of
differences in the evolutionary phase, surface gravity and metallicity. In the case of EK Dra the distribution appears to be
1 slightlyflatterthanintheprevioustwocases,althoughthepeaktemperatureissimilar.
v
1
Key words. X-rays: stars – stars: activity – stars: coronae – stars: individual: 31 Com – stars: individual: EK Dra – stars:
3
individual:HD283572
6
1
0
5 1. Introduction result is consistent with the one previously obtainedfrom the
0 analysesofEinsteinandROSATdata,i.e.thatthereisagood
/ During the last decade, the analysis of high-resolution X-ray correlation between the effective coronaltemperature and the
h
spectra of late-type stars, obtained with EUVE, XMM-
p X-ray emission level (see, for example, Schmittetal. 1990;
- Newton and Chandra (e.g. MonsignoriFossietal. 1995; Preibisch1997).
o Schmittetal. 1996; Gu¨deletal. 1997; Griffiths&Jordan
r Theobservationthatinactivestarsaconsiderableamount
t 1998; Laming&Drake 1999; Sanz-Forcadaetal. 2002;
s ofplasmasteadilyresidesatveryhightemperatures,whichare
Argiroffietal. 2003), revealed that the thermal structure of
a
achievedon the Sun only duringflaring events,led to the hy-
: coronal plasmas is better described by a continous Emission
v pothesisthatasuperpositionofunresolvedflaresmayheatthe
Measure Distribution, EMD, rather than by the combination
i
X of a few isothermal componentsusually employed to fit low- plasmacausinganenhancedquasi-quiescentcoronalemission
level.Followingthisidea,Gu¨del(1997)showedthatthetime-
r and medium-resolution spectra. Since the coronal plasma is
a averagedEMDresultingfromhydrodynamicsimulationsofa
optically thin, the EMD of the whole stellar corona can be
statistical set of flares, distributed in total energy as a power
viewed as the sum of the emission measure distributions of
law, could be made quite similar to the EMD of stars of dif-
all the loop-like structures where the plasma is magnetically
ferent activity level (and age). In particular, he obtained dis-
confined;therefore,itcanbeusedtoderiveinformationabout
tributionswith two peaksand a minimum around10MK; the
the properties of the coronal structures and the loop popula-
amountofthehottestplasma(at∼12−30MK)decreaseswith
tions (Peresetal. 2001). In particular, the studies mentioned
decreasingL (or,equivalently,withincreasingage)and,atthe
above have indicated that the coronae of intermediate and X
sametime,thefirstpeakmovestowardslowertemperatures.
high activity stars appear to be more isothermalthan coronae
of solar-type stars, and that the bulk of the plasma emission A qualitatively different scenario for the evolution of the
measure is around logT ∼ 6.6 for stars of intermediate EMD with activityhasbeenproposedbyJ. Drake(seeFig.2
activity and up to logT ∼ 7 for very active stars. The latter inthereviewby Bowyeretal.2000):thedistributionincreases
monotonically from the minimum, which occurs at logT be-
Send offprint requests to: L. Scelsi, e-mail: tween ∼ 5 and 6, up to the peak at coronaltemperatures;the
[email protected] locationofthepeakshiftstowardshigherandhighertempera-
tures (up to logT ∼ 7 in the most active stars) for increasing
X-rayactivitylevel.Alongwiththeshiftofthepeak,thesteep-
nessoftheascendingpartofthedistributionincreases.
In the pictures sketched above, the shape of the EMD
changes with the stellar activity level; however, it is not yet
understoodwhich stellar parameters(luminosity,surface flux,
surface gravity, evolutionary phase, or others) have a major
role in determining the physical characteristics of the domi-
nantcoronalstructuresand,hence,thepropertiesofthewhole
EmissionMeasureDistribution.
In order to investigate this issue, we have examined the
cases of three G-type stars, in different evolutionary phases:
the Pre-Main Sequence star HD 283572, the Zero-Age Main
SequencestarEKDraconis(HD129333)andtheHertzsprung-
gap giant star 31 Com (HD 111812). Here we report on the
analysesof recentXMM-Newton observationsof these bright
targets,characterizedbysimilarandrelativelyhighX-raylumi-
nosities(L ∼ 1030ergs−1 forEKDra,and L ∼ 1031ergs−1
X X
forHD283572and31Com)withrespecttotheSun.Previous
analyses(e.g. Gu¨deletal.1997;Ayresetal.1998;Favataetal.
1998) showed that the characteristic coronal temperatures of
the stars of our sample lie around 107K; the EPIC and RGS
detectorsonboardXMMareverysensitivetothistemperature Figure1. Positions of HD 283572, EK Dra and 31 Com in
regime,allowingustogetratheraccurateandreliableinforma- the H-R diagram. We have superimposed pre-main-sequence
tionabouttheplasmaEmissionMeasureDistributionsofthese (dashed lines) and post-main-sequence(solid lines) tracksfor
stars. themassvaluesreportedinthe plot.Theevolutionarymodels
are those of Venturaetal. (1998a,b), exceptfor the 2M pre-
The analysisof the XMM observationof 31 Com was re- ⊙
M.S. track, for which we have used the model of Siessetal.
ported in Scelsietal. (2004), while the reconstruction of the
(2000). All tracks are calculated for solar photospheric abun-
EMD ofHD 283572usinga high-resolutionspectrumis pre-
dances.
sentedhereforthefirsttime.Foreaseofcomparisonwiththese
twosources,wehavealsore-analyzedtheXMMobservationof
EKDrausingthesamemethodemployedfor31ComandHD
283572, thus ensuring homogeneity of the results; note how-
The latter star is a member of the Taurus-Auriga star
everthatindependentanalysesofthesameXMMobservation
forming region and its age is estimated to be ∼ 2 × 106yr
of EK Dra have been published since 2002 (e.g. Gu¨deletal.
(Walteretal. 1988). HD 283572 shows no sign of accretion
2002;Telleschietal.2003)andmorerecentlyandcomprehen-
fromacircumstellardisk,whichcharacterizestheearlierstage
sivelybyTelleschietal.(2004)inthecontextofastudyofso-
of classical T Tauri stars; the decoupling from the disk al-
laranalogsatdifferentages.Thelatterworkiscomplementary
lowed this star to spin up cosiderably, due to its contraction,
to our present study because it considers stars having similar
upto severaltens ofkms−1 (vsini = 78kms−1, see Table 1),
mass,sizeandgravity,butlargelydifferentL andcoronaltem-
X
probably with a consequently enhanced dynamo action and a
perature.
very high X-ray luminosity (L ∼ 1031ergs−1). From Fig. 1
This paper is organized as follows: we describe the three X
we deduce that HD 283572 will be an A-type star during its
targetsinSect.2andwepresenttheobservationsinSect.3.In
main sequence phase, and we estimate a mass between ∼ 1.5
Sect.4wedescribethedatareductionandthemethodsusedfor
and ∼ 2M , in agreement with the estimate of 1.8±0.2M
theanalysesofEPICandRGSspectra.Theresultsareshown ⊙ ⊙
byStrassmeier&Rice(1998a).TheradiusofHD283572has
inSect.5anddiscussedinSect.6.
beenderivedbyWalteretal.(1987)throughtheBarnes-Evans
relation, R ∼ 3.3R at an assumed distance of 160pc, which
⊙
becomes 2.7R at the new distance of 128pc measured by
2. Thesample ⊙
Hipparcos; more recently, Strassmeier&Rice (1998a) com-
In Fig. 1 we plot the positionsof the sample stars in the H-R bined photometric measurements, rotational broadening and
diagram,toshowtheirrespectiveevolutionaryphases.Weused Doppler imaging technique to determine the radius of HD
visualmagnitudes,B−Vcolorindexesanddistancesmeasured 283572intherange3.1−4.7R ,witha bestvalueof4.1R .
⊙ ⊙
byHipparcos;weassumednegligibleopticalextinctioninthe Duetotheuncertaintiesoftheseestimates,wedecidedtocon-
casesofEKDraand31Com,coherentwiththelowinterstellar siderbothofthem.Weanticipatethatourmainresultsareonly
absorptionusedintheanalysisoftheirX-rayspectra(Sect.5), weaklyaffectedbythechoiceofoneofthesevalues.
whileweusedavisualextinctionAV =0.57(Strometal.1989) EK Dra is a G1.5-type star with mass and radius about
andEB−V ∼AV/3forHD283572. equal to the solar values. It has just arrived on the main se-
Table1.Stellarparameters.DistancesaremeasuredbyHipparcos;L (0.3−8keV)arederivedinthisworkfrom3−T models;
x
gravitiesaredeterminedfromthecorrespondingMandR,andsurfacefluxesfromthecorrespondingL andR.Inthelastcolumn,
X
thereferencesforL areindicatedintheentries.
bol
M/M R/R Spectral P vsini d L g/g F L /L
⊙ ⊙ rot x ⊙ x x bol
type [d] [Kms−1] [pc] [1030ergs−1] [106ergs−1cm−2]
HD283572 1.8a 2.7b;4.1a G2 1.55a 78a 128 ∼9 0.25;0.12 20;9 5× 10−4c
EKDra 1.1d 0.95d G1.5V 2.75d 17.3e 34 ∼1 1.2f 18 3× 10−4g
31Com 3h 9.3h G0III <7.2i 66.5j 94 ∼7 0.035 1.3 3× 10−5h
aStrassmeier&Rice(1998a).
bWalteretal.(1987)andtheHipparcosmeasurementofd.
cWalteretal.(1988)andtheHipparcosmeasurementofd.
dGuinanetal.(2003).
eStrassmeier&Rice(1998b).
f AlsoconsistentwiththeestimatebyStrassmeier&Rice(1998b).
gRedfieldetal.(2003).
hPizzolatoetal.(2000).
iFromP andvsini.
rot
jdeMedeiros&Mayor(1999).
quence,thusrepresentingananalogoftheyoungSun.Because panion(Duquennoyetal.1991),whosemassislikelybetween
of its age (∼ 7 × 107yr, Soderblom&Clements 1987), it 0.37M and0.45M (Gu¨deletal.1995a).Gu¨deletal.(1995b)
⊙ ⊙
sufferedlittle magneticbrakingand its shortrotationalperiod found that the X-ray and radio emissions are modulated with
(∼2.7days, Guinanetal.2003)makesitabrightX-raysource therotationalperiod,stronglysuggestingthatthecoronalemis-
(L ∼1030ergs−1). sion comespredominantlyfrom the G star. If we assume that
X
The more massive (M ∼ 3M⊙) giant star 31 Com (age thesecondarystarhasM ∼ 0.4M⊙ andage∼ 70Myr,andhas
∼4×108yr, Friel&Boesgaard1992)hasalreadyevolvedout asaturatedcorona(theworstcase),itsX-rayluminositywould
of the main sequence andnow it is crossingthe Hertzsprung- be∼ 1029ergs−1,sowemightexpectcontaminationoftheX-
gap.Thepositionof31ComintheH-Rdiagramandtheevo- rayemissionoftheGstarfromthecompanionatmostat∼10
lutionarymodelsindicateaspectraltypelate-B/early-Aonthe %level.
mainsequence;therefore,thisstarhasdevelopedaconvective
subphotosphericlayer and a dynamo only in its current post- 3. Observations
mainsequenceevolutionaryphase(Pizzolatoetal.2000).The
X-rayluminosityis∼7×1030ergs−1. TheobservationsofHD283572,EKDraand31Comwereper-
formedwithXMM-NewtonrespectivelyonSeptember,5,2000
Thestellarparametersofthethreetargets,withtherelevant
(PI:R.Pallavicini),onDecember,30,2000(PI:A.Brinkman)
references,aresummarizedinTable1.ForHD283572were-
andonJanuary9,2001(PI:Ph.Gondoin).Thenon-dispersive
portbothestimates,mentionedabove,ofthestellarradiusand
CCD cameras (EPIC MOS and EPIC , Turneretal. 2001;
thecorrespondingvaluesofgravityandsurfaceX-rayflux.
Stru¨deretal.2001),lyinginthefocalplaneoftheX-raytele-
HD 283572, EK Dra and 31 Com were chosen because
scopes,havespectralresolutionR= E/∆E ∼5−50intherange
their stellar parameters offer the possibility to get useful in- 0.1 − 10 keV, while the two reflection grating spectrometers
sightintotheircoronalpropertiesfromthecomparisonoftheir (RGS, denHerderetal.2001)provideresolutionR∼70−500
EMD. Note, in particular, that while the X-ray luminosity of
inthewavelengthrange5−38Å(0.32−2.5keV).
31 Com and HD 283572 are about equal and larger than that Forthepresentstudy,weconsideredonlytheEPICand
of EK Dra by about an order of magnitude, EK Dra and HD
RGSdata;inTable2wereportdetailsontheinstrumentcon-
283572arethestarswiththehighestsurfacefluxes,whoseval-
figurationsandontheobservations.
uesexceedsignificantlythatof31Com,byaboutoneorderof
Atthetimeoftheseobservations,bothCCD7ofRGS1and
magnitude.Notealsothatthedifferentevolutionaryphasesim-
CCD4ofRGS2werenotoperating.TheseCCDscorrespondto
plydifferentstellarinternalstructures;moreover,thesetargets
thespectralregionscontainingtheHe–liketripletsofneonand
havequitedifferentgravities,implyingdifferentpressurescale
oxygen,respectively.NotealsothattheRGS1spectrumofHD
heightsandpossiblechangesinthepropertiesofthedominant
283572isentirelymissing,duetoinstrumentsetupproblemsin
coronalloops.
theearlyphaseofXMM-Newtonobservations;hencewehave
Finally,therapidlyrotatingstarsHD 283572and31Com noinformationontheOtripletforthissource.
areputativesinglesources:thisavoidsdifficultiesintheinter-
pretationoftheresults,duebothtotheuncertainoriginofthe
4. Dataanalysis
emission,incaseofmultiplecomponents,andtothepossibility
ofanenhancedactivityasfound,forexample,intidally-locked We usedSASversion5.3.3,togetherwiththecalibrationfiles
RSCVnsystems.Onthecontrary,EKDrahasadistantcom- availableatthetimeoftheanalysis(June2002),toreducethe
Table2.LogoftheXMM-Newtonobservations.
Exposuretime(ks) EPIC Q.E.Exposurea(ks) Count-rateb(s−1)
RGS1 RGS2 Mode/Filter RGS1 RGS2 RGS1 RGS2
HD283572 41.1 0 48.7 FullFrame/Medium 41.1 0 47.4 2.20 0 0.15
EKDra 46.9 51.7 50.2 LargeWindow/Thick 38.5 44.9 43.6 2.20 0.16 0.22
31Com 33.5 41.7 40.5 FullFrame/Thick 32.2 39.6 38.5 1.45 0.11 0.16
aExposuretimefortheanalysisofthequiescentemission(Q.E.),i.e.excludingthetimeintervalsaffectedbyprotonflares,
occurredinthecasesofHD283572and31Com,andbythesourceflareinthecaseofEKDra(seeSect.4.1).
bMeancount-rateinthe1.2−62Å(0.2−10keV)bandforandinthe5−38Å(0.32−2.5keV)bandforRGS(1storder
spectrum)relevanttotheQ.E.Exposure.
dataofHD 283572and31Com;thedataofEKDrawerere- ceedsbymorethan1σtheaverageHRvaluecalculatedfrom
ducedwithSASversion5.4andtheanalysiswasperformedin timeintervalsbeforeandaftertheflare.Weexcludedthetime
September2003.We generatedall responseswiththeSAS intervaloftheflarefromtheemissionmeasureanalysis(Sect.
andtasks. 4.3),sincewewanttostudythethermalpropertiesofthequi-
GoodTimeIntervalswereselectedbyexcludingthosetime escent corona, and we analyzed the flare separately. The qui-
intervalsshowing the presenceof presumableprotonflares in escentemissionofEKDraisstillvariable,yieldingareduced
thebackgroundlightcurveextractedfromCCD9oftheRGS, χ2 = 6.2(190d.o.f.)againstthenullhypothesisofaconstant
r
following denHerder (2002): we cut the intervals where the source;thevariabilityisonatime-scaleof∼15ksanditsam-
count-rate exceeds 0.1 ctss−1 for 31 Com and 1.6 ctss−1 in plitude(calculatedasabove)is∼16%.
thecaseofHD283572(whoseobservationiscontaminatedby
highlevelofbackground),whilewedidnotexcludeanyinter-
4.2.EPICspectra
valinthecaseofEKDra.
In order to obtain X-ray light curves and spectra, we ex- WehaveperformedglobalfittingoftheEPICspectra(Fig.3)
tractedtheeventsfromacircularregion(∼ 50′′ radius)within withtheaimofderiving,frommulti-componentthermalmod-
CCD 4 for HD 283572, while we used annular regions (∼ els,theinitialguessofthecontinuumlevelforthelinemeasure-
7.5′′−50′′radii)for31ComandEKDra,becausetherelevant mentsin theRGSspectra.Moreover,theabundancesofsome
datawereaffectedbypile-up.Inallcases,backgroundphotons elements(Si,S,Ar,Ca)canbebetterdeterminedfromspec-
were extractedfromtherestof CCD 4,excludingthe sources tra,ratherthanfromRGSspectra,thankstothewiderspectral
andtheirout-of-timeevents. rangeoftheformer–whichincludesthestrongK-shelllinesof
therelevantH–likeandHe–likeions–andtothehigherphoton
countingstatistics2.
4.1.Lightcurves
We analyzed these spectra with XSPEC v11.2 and we
Figure 2 shows the background-subtractedlight curves of found that an absorbed, optically-thin plasma with three
the sources, with a 200s time binning. The light curve of 31 isothermal components provides an acceptable description
Comistheonlyonethatisconsistentwiththehypothesisofa of each of them (see results in Sect. 5.1). The models
constantemission(seeSect.3.1in Scelsietal.2004). are based on the Astrophysical Plasma Emission Database
In the case of HD 283572,there is evidenceof variability (APED/ATOMDB V1.2) and have variable abundances; we
of the emission, on a time-scale of the order of 30ks, which adopted the criterion of leaving free to vary only the abun-
is nota typicalflare event.The reducedχ2 is 5.9(229d.o.f.) dances of those elements (O, Ne, Mg, Si, S, Fe, Ni, in some
r
in the null hypothesis of a constant emission; the variability casesCaandAr)withstrongandclearlydetectablelinecom-
amplitude,calculatedas0.5[max(rate)-min(rate)]/mean(rate), plexesinEPICspectra.Theabundancesoftheotherelements
is∼20%.Thereisalesspronouncedvariabilityonatime-scale weretiedtothatofiron,theirbest-fitvaluesbeingpoorlycon-
of∼10ks.FromTables1and2,wenotethatthedurationofthe strainedwhenleftfreetovary.
observation is about one third of the stellar rotational period, Weeventuallyusedthehigh-energytailofthespectrum
hencea largefractionofthestellarsurfacewasvisibleduring alsotocheckthehigh-temperaturetailoftheemissionmeasure
thepointing;thissuggeststhatatleastpartofthevariabilityis distributions,asdescribedinthenextsectionandinAppendix
due to an inhomogeneousdistributiuon of active regionsover A.
thestellarsurface. Finally,weperformedtime-resolvedspectroscopyofthe
ThelightcurveofEKDraclearlyshowsthepresenceofa dataofEKDraduringtheflare,togetinformationontheprop-
flare;theverticallinesinthefiguremarkthestartandtheendof
theflare,obtainedastheminimumandmaximumtimeswhere 2 The Si- lines fall also in the RGS spectral range, but the
thehardness-ratio,HR=(H−S)/(H+S)1,systematicallyex- statisticsareusuallyverylowandthecalibrationoftheeffectiveareais
lesspreciseatthesewavelengths;nonetheless,theresultsofouranal-
1 We have evaluated the soft emission count-rate, S, in the 0.3− ysisshowthattheSiabundancesderivedfromdataareconsistent
1keV band and the hard emission count-rate, H, in the 1−10keV withthoseobtainedfromRGSspectrawithinstatisticaluncertainties
band. (seeTable4).
Figure3.EPICspectraofHD283572,EKDraand31Comwith theirbest-fitmodelspectra(theparametersofthemodels
arelistedinTable3).Thespectraof31ComandHD283572,withtheirrelevantbest-fitmodels,havebeenshiftedby-0.5and
+0.5dexforclarity.
erties(inparticularthesize)oftheflaringloop,employingthe selectedasetoflines,amongtheidentifiedones,withreliable
methodbyRealeetal.(1997).Thisanalysisanditsresultsare flux measurementsand theoreticalemissivities. Most of them
reportedinAppendixB. are blendedwith other lines, so the measuredspectralfeature
isactuallythesumofthecontributionsofanumberofatomic
transitions; accordingly, we evaluated the ”effective emissiv-
4.3.EmissionMeasureReconstruction
ity”ofeachlineblendasthesumoftheemissivitiesofthelines
whichmostlycontributetothatspectralfeature.Moreover,we
The approach we adopted for the line-based analysis of the
carefullyselectedonlyironlinesnotblendedwithlinesofother
RGS spectra of each star is discussed in detail in Scelsietal.
elements,becausetheprocedureweemployed(seebelow)uses
(2004)togetherwithastudyofitsaccuracy;herewelimitour-
theseironlinesinthefirststepofthe EMDanalysis,andesti-
selvestoreportthemainpointsofouriterativemethod.
matesoftheabundancesoftheotherelementsarenotyetavail-
We employed the software package PINTofALE
ableatthisstep.
(Kashyap&Drake 2000) and, in part, also XSPEC, and
used the APED/ATOMDB V1.2 database which includes the We performedthe EMD reconstructionwith the Markov-
Mazzottaetal.(1998)ionizationequilibrium. Chain Monte Carlo (MCMC) method by Kashyap&Drake
Wefirstrebinnedandco-addedthebackground-subtracted (1998). This method yields a volume emission measure dis-
RGS1andRGS2spectrafortheidentificationofthestrongest tribution, EM(T ) = dem(T )∆logT, and related statistical
k k
emission lines and the measurement of their fluxes. In this uncertainties∆EM(T ), where dem(T) = n2dV/dlogT is the
k e
latter step, we adopted a Lorentzian line profile and we as- differential emission measure of an optically thin plasma and
sumed initially the continuum level evaluated from the 3-T ∆logT = 0.1isa constantbinsize;themethodalsoprovides
modelbestfittingthespectrum,becausethewidelinewings estimatesofelementabundances,relativetoiron,withtheirsta-
makeitimpossibletodeterminethetruesourcecontinuumbe- tisticaluncertainties.Theironabundanceisestimatedbyscal-
low ∼ 17Å directly from the RGS data, in particular in the ingtheemissionmeasuredistributionassumingdifferentmetal-
∼ 10−17Årange,wherethespectrumisdominatedbymany licities andby comparingthe synthetic spectrumwith the ob-
strongoverlappinglines.Then,withtheaimtoreconstructthe servedoneatλ > 20ÅintheRGSspectrum(thisisaspectral
Emission Measure Distribution (EMD) vs. Temperature, we region free of strong overlappingemission lines). Finally, we
are importanttests for the reliability of the amountof plasma
inthehigh-temperaturetailoftheEMD.
We also checked the consistency between the continuum
level assumed for flux measurements and the continuum pre-
dicted by the EMD. In fact, since our methodis iterative,the
continuum assumed for flux measurements in the RGS range
is adjusted at each iteration for consistency with the EMD,
and it maybecomedifferentfrom thatpredictedby the 3−T
modelbest-fittingthe spectrum,whichis adoptedasinitial
guess. Therefore, this procedure ensures that possible cross-
calibrationoffsetsbetween andRGS donotaffectthefinal
EMD.
5. Results
5.1.3-Tmodels
Figure3showsthespectrawiththeirrelevant3-Tmod-
els, obtainedby fitting the data in the 0.3−8keV range. The
best-fitparametersofthemodelsarelistedinTable3.
The presence of the Fe 6.7 keV emission line in all
thesespectraisindicativeofhotcoronae,asexpectedfromear-
lierworksandconfirmedbyouranalysis.Notethelargebest-fit
EMvaluesofthehottestcomponentsforallstars,comparable
totheEMsofthecoolercomponents;inparticular,thehottest
plasmadominatesthecoronaofHD283572,asalsoconfirmed
bytheanalysisofChandraspectra(Audardetal.2004).
Thelinecomplexesof Mg-(∼ 1.3−1.5keV),Si-
(∼1.8−2.1keV)andS(∼2.5keV),aswellasthelarge
bumpbetween0.6and1keVduetotheFe-,Ni-
andNe-linesallowedustoconstraintheabundancesofMg,
Si, S, Fe, Ne and Ni. These complexesare less evidentin the
spectrumofHD283572,asaconsequenceofthesignificantly
lowermetallicitywithrespecttotheothertwostars;instead,the
Figure2.Background-subtractedlightcurvesofHD283572 Calinecomplex(∼3.9keV)ismostprominentinthespec-
(upper), EK Dra (middle) and 31 Com (lower), in the 0.2 − trumofthisstar(Fig.3)andtheestimatedabundanceofthisel-
10keVbandandwith timebinsof200s. Theverticallinesin ementishigherthanfortheothertwocases.Notealsothatwe
the light curve of EK Dra mark the time interval of the flare, wereabletoconstraintheArabundanceforEKDra,thanksto
excludedfromtheanalysisofthequiescentemission. theclearlyvisiblelinesofArat∼3.1keV.Intheothertwo
cases we linked the abundancesof Ca and/orAr to that of Fe
assumingthesameratiosasinthesolarcorona(Grevesseetal.
1992).Moreover,we usedtheresultsofthe EMD analysesto
fixthecoronalC/FeabundanceratioforEKDraand31Comto
checkedthe solution obtainedwith the MCMC by comparing
respectively0.7and0.25solar,andtheN/FeratioforEKDra
(i)thelinefluxespredictedfromoursolutionwiththemeasured
onesand(ii)themodelspectrum,basedonthereconstructed to0.5solar.
EMD, with the observed spectrum at E > 2keV. These Finally,wecouldnotconstraintheinterstellarabsorptionin
checks are illustrated respectively in Fig. 6 and in Appendix thedirectionsofEKDraand31Comwiththefittingprocedure,
A, taking the case of EK Dra as an example (similar results sowefixedthemattherelativelylowvaluesof3×1018cm−2
wereobtainedfortheothertwostars).Inparticular,thecorrect and 1018cm−2 measured, respectively, by Gu¨deletal. (1997)
prediction of the O- line fluxes allowed us to check the and Piskunovetal. (1997). On the contrary, the spectrum of
reliabilityoftheamountofplasmainthelow-temperaturetail HD283572issignificantlyabsorbed,asexpectedfromthelo-
oftheEMD;analogously,thecorrectpredictionoftheFe- cationofthisstarintheTaurus-Aurigastarformingregion.The
linefluxesandofthehigh-energytailofthespectrum hydrogencolumndensitywederivedfromthefitiscompatible
Table3.Best-fitmodelsoftheEPICdata(inthe0.3−8keVband),with90%statisticalconfidencerangescomputedforone
interestingparameteratatime;nominalerrorsonT andEM areatthe10%level.Elementabundancesarerelativetothesolar
i i
ones (Grevesseetal. 1992). Mean temperaturesare calculated as < T >= P3 EM T/P3 EM. Abundancesand hydrogen
i=1 i i i=1 i
columndensitieswithouterrorswerefixedasexplainedinthetext.
HD283572 EKDra 31Com
logT (K) 6.64,7.04,7.43 6.58,6.94,7.33 6.44,6.92,7.28
1,2,3
logEM (cm−3) 53.5,53.5,53.7 52.5,52.4,52.3 52.6,53.1,53.0
1,2,3
log<T > (K) 7.21 6.99 7.06
C 0.37 0.57 0.38
N 0.37 0.42 1.54
O 0.236 ± 0.014 0.346 ± 0.015 0.58 ± 0.03
Ne 0.46 ± 0.03 0.83 ± 0.04 2.35 ± 0.14
Mg 0.32 ± 0.05 0.86 ± 0.06 1.95 ± 0.13
Si 0.25 ± 0.04 0.59 ± 0.06 1.23 ± 0.11
S 0.26 ± 0.09 0.15 ± 0.10 0.58 ± 0.20
Ar 0.37 0.82 ± 0.22 1.54
Ca 1.8 ± 0.3 0.83 1.54
Fe 0.37 ± 0.01 0.83 ± 0.01 1.54 ± 0.02
Ni 1.52 ± 0.11 1.80 ± 0.20 4.1 ± 0.3
N (cm−2) (8.7 ± 0.4)×1020 3×1018 1018
H
χ2/d.o.f. 1.1/688 1.26/411 1.1/367
ν
Figure4.Co-addedRGSspectraofHD283572(upper),EKDra(middle)and31Com(lower)withtheidentificationofthemost
prominentlines;thebinsizeis0.02Å.
withA andconsistentwithpreviousresultsobtainedbyfitting inallcases,whileO,Mg-andSi-emissionlines
V
ASCA,ROSAT,EinsteinandSAXdata(Favataetal.1998).
5.2.EmissionMeasureDistributionsandabundances
The rebinned and co-added RGS spectra (Fig. 4) show emis-
sionlinesfromFe-,Ne-,OandNi-ions
arevisible onlyforEK Draand31Com,andthe N linein
the case of EK Dra only. Actually, we could not identify any
lineoutsidethewavelengthrange10−20Åinthespectrumof
HD 283572,because of contaminationfromhigh background
andlackoftheRGS1spectrumaltogether.Thereconstruction
of the EMD of EK Dra and 31 Com was based on about 40
lines, while we used 25 lines in the case of HD 283572 as a
consequenceofthelowerqualityofitsspectrum.
ThederivedEMDsareplottedinFig.5;notethatthealgo-
rithmweusedisnotabletoconstrainstatisticallythevaluesof
the emissionmeasure in allthe temperaturebins.We show in
Fig.6theobserved-to-predictedfluxesforthecaseofEKDra,
which is representative of the spread of these ratios obtained
in ouranalyses, and in Fig. 7 we comparethe observedspec-
tra and the modelspectra generatedwith the solutions(EMD
andabundances)foundinthiswork.Theelementalabundances
areshowninTable4;weestimatedtheironabundances(rela-
tivetothesolarvalue)ofHD283572,EKDraand31Comat
0.7±0.2,1.2±0.2and1.4±0.2,respectively.Table5reports
theratios3R= f/iandG =(f+i)/rrelativetotheOtriplet,
andtheestimatesofelectrontemperatures,densitiesandpres-
sures, averagedoverthe regionwhere the triplet forms, using
thetheoreticalcurvesbySmithetal.(2001).
In Fig. 8 we show the element-to-iron abundance ratios
for the three stars, ordering the elements for increasing First
IonizationPotential(FIP).WheneverbothEPIC-andRGS-
based estimates were available, we always reportthe latter in
theplot,becauseweconsiderthevaluesderivedwiththeRGS
the most accurate. It is worth noting that, despite widely dif-
feringmethodsemployedto deriveelementalabundances,we
obtainedconsistency between the - and RGS-derivedabun-
dance ratios, except for Ne in the case of 31 Com (the value
indicatedbytheistwotimeslargerthantheRGSone)and
for Ni in the case of EK Dra and HD 283572(the values ob-
tainedwiththearelargerthantheRGSonesbyfactorsof2
and4respectively).
Thepatternsofabundancesvs.FIParesimilarinthecases
of 31Com andHD 283572,with an initialdecrease (with re-
specttosolarphotosphericvalues)downtoaminimumaround
carbon, followed by increasing abundancesfor elements with
higherFIP(> 11eV).Thispatternisalsosimilartowhatwas
foundfortheyoungactivestarABDorbySanz-Forcadaetal.
(2003),butitislessevidentinthecaseofEKDra.Notethat31
ComandEK Drahaveironabundancesdifferingfromthatof
HD 283572byabouta factorof2,hencethe patternofabun-
dancesvs.FIPappearstobealmostindependentoftheglobal
coronalmetallicity. Figure5.DistributionsofemissionmeasurederivedfromRGS
data.Valueswithouterrorbarsarenotstatisticallyconstrained
by the MCMC algorithm. Note the different ordinate scale in
theplotofEKDrawithrespecttothoseofHD283572and31
6. Discussion Com.
XMM-Newton data allowed us to derive the plasma emission
measuredistributionsforourthreetargetsandtheircoronalel-
ementalabundances;inparticular,theEMDofHD283572has
beenderivedhereforthefirsttimeusingahigh-resolutionspec-
3 r,iand f denotethefluxesoftheresonance,intercombinationand trum.Ourresultsaresufficientlywelldeterminedandhomoge-
forbiddenlines. neousforthe purposeofa detailedcomparisonofthecoronal
Figure7.ModelspectracomparedtotheoriginalRGSspectra.
Table4.Ratiosbetweenelementalandironcoronalabundances,relativetothesolarphotosphericratios(Grevesseetal.1992),
derivedfromRGSdata;errorsareat68%confidencelevel.Forcompleteness,wealsoreporttheabsoluteironabundance.The
-derivedvaluesareshownforpurposeofcomparison.Foreachstar,thenumberoflinesusedforthe EMD reconstructionis
reported(thenumberoflinesofagivenelementisshowninparenthesisneartherelevant(RGS)abundancevalue).
HD283572 EKDra 31Com
RGS RGS RGS
C/Fe 0.69+0.28(1) 0.24+0.22(1)
−0.08 −0.07
N/Fe 0.52+0.4 (1)
−0.16
O/Fe 0.6+0.4(2) 0.64 ± 0.04 0.50+0.04(3) 0.42 ± 0.02 0.49+0.15(2) 0.38 ± 0.02
−0.2 −0.07 −0.08
Ne/Fe 1.2+0.23(2) 1.24 ± 0.09 1.00+0.21(3) 1.00 ± 0.05 0.78+0.13(2) 1.53 ± 0.09
−0.3 −0.23 −0.3
Mg/Fe 0.86 ± 0.14 0.88+0.6 (2) 1.04 ± 0.07 1.0+0.4(2) 1.27 ± 0.08
−0.13 −0.3
Si/Fe 0.68 ± 0.11 0.7+0.5(1) 0.71 ± 0.07 0.9+0.9(1) 0.80 ± 0.07
−0.4 −0.3
Ni/Fe 1.2+1.0(2) 4.1 ± 0.3 0.9+1.0(1) 2.17 ± 0.24 3.5+2.1(6) 2.7 ± 0.2
−0.6 −0.3 −0.7
Fe 0.7 ± 0.2(19) 0.37 ± 0.01 1.2 ± 0.2(24) 0.83 ± 0.01 1.4 ± 0.2(27) 1.54 ± 0.02
totallines 25 36 41
Figure6.Comparisonbetweenobservedfluxesandthe fluxes
Figure8.Element-to-ironabundanceratios,relativetotheso-
predicted with the EMD model, for lines used in the EM re-
larphotosphericvalues(Grevesseetal.1992),forHD283572
construction of EK Dra; Fe: open diamonds, Ne: triangles,
(squares),EKDra(triangles)and31Com(diamonds).Theel-
Mg:opensquares,Si:filleddiamond,Ni:filledcircle,O:filled
ementsareorderedbyincreasingFIP.
squares,N:asterisk,C:opencircle.
theyareproportionalto∼T5,wheretheexponentofthepower
properties of the selected stars. We recall that these stars are
lawhasaformalconfidenceintervalbetween∼ 3.4and∼ 6.6;
in differentevolutionarystages, butsharethe characteristicof
therearealso indicationsofasignificantamountofplasmaat
beingactive(highX-rayluminosity)G-typestars.Ouranalysis
temperatures hotter than T (up to logT ∼ 7.6) and, at least
p
hasconfirmedthatthethreestarshaveveryhotcoronae,with
in the case of 31 Com, in the range logT ∼ 6−6.2. We re-
similar average temperatures(∼ 11−12MK for EK Dra and
callthatwe arenotabletostatistically constrainthe emission
31Com,and∼16MKforHD283572).
measureinallthetemperaturebins,andhencetogetinforma-
A remarkableresult of this work is the close similarity of
tionontheexactshapeofthedistributionsbelowlogT ∼ 6.5
theemissionmeasuredistributionsofHD283572and31Com,
andabovelogT ∼ 7.3,yetthepresenceinbothstarsofanon-
whichhavesimilar L aswell.Bothdistributionshaveawell-
X negligibleamountofplasma upto logT ∼ 7.6hasbeenveri-
definedpeakatT = 107Kand,intherangelogT ∼ 6.5−7,
p fied,asdescribedinSect.4.3,throughthecorrectpredictionof
the Fe-linefluxesandbycomparisonwith the high-
Table5.PressureestimateswithO. energy tail of the observed EPIC spectra (App. A), while the
correctpredictionoftheO-linesallowedustoverifythe
presence of cool plasma down to logT ∼ 6 in the EMD of
R ne(range) G T P(range) 31 Com (the O line is notavailable in the spectrumof HD
(1010cm−3) (106K) (dyncm−2)
EKDra 3.0±1.7 1(<7) 0.93±0.25 1.5+2.0 4(<70) 283572).Note,also,thattheshapeoftheconstrainedpartofthe
−0.5
31Com 2.0±1.4 3(0.6−20) 0.94±0.36 1.5+2.5 13(1.5−220) distributionof31Comandthepresenceofsignificantemission
−0.7