Chandra X-ray spectroscopy of focused wind in the Cygnus X-1 system

24 Pages·2016·2.72 MB·English

A&A590,A114(2016) Astronomy
DOI:10.1051/0004-6361/201322490
&
(cid:13)c ESO2016 Astrophysics
Chandra X-ray spectroscopy of focused wind
in the Cygnus X-1 system
II. The non-dip spectrum in the low/hard state – modulations with orbital phase
IvicaMiškovicˇová1,NatalieHell1,2,ManfredHanke1,MichaelA.Nowak3,KatjaPottschmidt4,5,NorbertS.Schulz3,
VictoriaGrinberg1,3,RefizDuro1,6,OliwiaK.Madej7,8,AnneM.Lohfink9,JérômeRodriguez10,
MarionCadolleBel11,ArashBodaghee12,JohnA.Tomsick13,JuliaC.Lee14,GregoryV.Brown2,andJörnWilms1
1 Dr.KarlRemeis-SternwarteandErlangenCentreforAstroparticlePhysics,UniversitätErlangen-Nürnberg,Sternwartstr.7,
96049Bamberg,Germany
e-mail:[email protected]
2 LawrenceLivermoreNationalLaboratory,7000EastAve.,Livermore,CA94550,USA
3 MITKavliInstituteforAstrophysicsandSpaceResearch,NE80,77Mass.Ave.,Cambridge,MA02139,USA
4 CRESST,UniversityofMarylandBaltimoreCounty,1000HilltopCircle,Baltimore,MD21250,USA
5 NASAGoddardSpaceFlightCenter,AstrophysicsScienceDivision,Code661,Greenbelt,MD20771,USA
6 AITAustrianInstituteofTechnologyGmbH,Donau-City-Str.1,1220Vienna,Austria
7 DepartmentofAstrophysics/IMAPP,RadboudUniversityNijmegen,POBox9010,6500GLNijmegen,TheNetherlands
8 SRONNetherlandsInstituteforSpaceResearch,Sorbonnelaan2,3584CAUtrecht,TheNetherlands
9 InstituteofAstronomy,UniversityofCambridge,MadingleyRoad,CambridgeCB30HA,UK
10 LaboratoireAIM,UMR7158,CEA/DSM-CNRS-UniversitéParisDiderot,IRFU/SAp,91191Gif-sur-Yvette,France
11 Max-PlanckComputingandDataFacility,Gießenbachstr.2,85748Garching,Germany
12 DepartmentofChemistry,Physics,andAstronomy,GeorgiaCollege&StateUniversity,Milledgeville,GA31061,USA
13 SpaceSciencesLaboratory,7GaussWay,UniversityofCalifornia,Berkeley,CA94720-7450,USA
14 Harvard John A. Paulson School of Engineering and Applied Sciences, and Harvard-Smithsonian Center for Astrophysics,
60GardenStreetMS-6,Cambridge,MA02138,USA
Received14August2013/Accepted1April2016
ABSTRACT
AccretionontotheblackholeinthesystemHDE226868/CygnusX-1ispoweredbythestrongline-drivenstellarwindoftheO-type
donorstar.WestudytheX-raypropertiesofthestellarwindinthehardstateofCygX-1,asdeterminedusingdatafromtheChandra
HighEnergyTransmissionGratings.Largedensityandtemperatureinhomogeneitiesarepresentinthewind,withafractionofthe
windconsistingofclumpsofmatterwithhigherdensityandlowertemperatureembeddedinaphotoionizedgas.Absorptiondips
observedinthelightcurvearebelievedtobecausedbytheseclumps.Thisworkconcentratesonthenon-dipspectraasafunctionof
xvii xxiv
orbitalphase.ThespectrashowlinesofH-likeandHe-likeionsofS,Si,Na,Mg,Al,andhighlyionizedFe(Fe –Fe ).We
measurevelocityshifts,columndensities,andthermalbroadeningofthelineseries.Theexcellentqualityofthesefiveobservations
allowsustoinvestigatetheorbitalphase-dependenceoftheseparameters.Weshowthattheabsorberislocatedclosetotheblack
hole.Dopplershiftedlinespointatacomplexwindstructureinthisregion,whileemissionlinesseeninsomeobservationsarefroma
densermediumthantheabsorber.Theobservedlineprofilesarephase-dependent.Theirshapesvaryfrompure,symmetricabsorption
atthesuperiorconjunctiontoPCygniprofilesattheinferiorconjunctionoftheblackhole.
Keywords. accretion,accretiondisks–stars:individual:CygX-1–stars:individual:HDE226868–X-rays:binaries–
stars:winds,outflows
1. Introduction M = 19.2±1.9M for the companion star and M = 14.8±
2 (cid:12) 1
1.0M for the compact object have been deduced (Oroszetal.
(cid:12)
In the 50 years of persistent X-ray activity since its discov- 2011).
ery in 1964 (Bowyeretal. 1965), the high-mass X-ray binary
CygnusX-1(CygX-1)hasbecomeoneofthebestknownX-ray Withamasslossrateof∼10−6M yr−1(Herreroetal.1995),
(cid:12)
sources,buttherearestillmanyopenquestions,evenwithregard HDE 226868 shows a strong wind. These winds are driven
to one of its most basic properties, the nature of the accretion by radiation pressure which, owing to copious absorption lines
process.ThesystemconsistsofthesupergiantO9.7Iab-typestar present in the ultraviolet part of the spectrum on material in
HDE226868(Walborn1973)andacompactobjectina5.6dor- thestellaratmosphere(line-drivenorCastoretal.1975,[CAK]
bit (Webster&Murdin 1972; Brocksoppetal. 1999; Giesetal. windmodel),canreachveryhighvelocities(v (cid:29)2000kms−1;
∞
2003)withaninclinationofi = 27◦.1±0◦.8(Oroszetal.2011). Muijresetal.2012)andareonlyproducedbyhot,earlytypeO
Thesystemhasadistanced = 1.86+0.12kpc(Xiangetal.2011; orBstars.Simulationsshowthatasteadysolutionofline-driven
−0.11
Reidetal. 2011). Based on these measurements, masses of windsisnotpossible,i.e.,perturbationsarepresentinthewind
ArticlepublishedbyEDPSciences A114,page1of24
A&A590,A114(2016)
(Feldmeieretal. 1997), causing variations of density, velocity,
and temperature, which compress the gas into small, cold, and 0 . 8 0.7
overdensestructures,oftenreferredtoasclumps(Oskinovaetal.
3815
2012; Sundqvist&Owocki 2013, and references therein). The
existence of clumps is supported by observations of transient 9
X-rayabsorptiondips(lowerflux)inthesoftX-raylightcurves 0. 0.6
ofsuchsystems.Sakoetal.(1999)estimatethatmorethan90%
of the total wind mass in VelaX-1 is concentrated in clumps,
whiletheionizedgascoversmorethan95%ofthewindvolume.
4
Clumpinesswithafillingfactorof0.09–0.10hasbeenconfirmed
1 1
inCygX-1byRahouietal.(2011). 8 1
0 3 x 0 0
Strong tidal interactions between the star and the black . 4 .
0 4 5
hole and centrifugal forces make the wind distorted and asym-
metric. Both the wind density and the mass loss rate are en- 5
2
hanced close to the binary axis, creating a so-called focused 5
8
wind(Friend&Castor1982).Evidenceforthepresenceofsuch
a wind around HDE 226868, which fills more than ∼90% of
0
its Roche volume (Gies&Bolton 1986), is the strong mod- 1. 4.
ulation of the Heii λ4686 emission line with orbital phase 0
(Gies&Bolton1986)andthestrongphase-dependenceofother 7489
opticalabsorptionlines,whicharedeepestaroundorbitalphase
φ = 0(Giesetal.2003),i.e.,duringthesuperiorconjunction 2.0 3 . 0
orb
oftheblackhole.Furtherevidencecomesfromthemodelingof
the IR continuum emission (Rahouietal. 2011). Finally, in the Fig.1.OrbitalphasecoverageoftheChandraobservationsofCygX-1
X-rays,theoverallabsorptioncolumndensityanddippingvary inthehardstateanalyzedinthiswork.Fullarcs(dashedarcs)display
strongly with orbital phase and are largest at φorb ∼ 0.0 (e.g., TE mode (CC mode; for explanation see Sect. 2.3) observations. The
Li&Clark 1974; Masonetal. 1974; Parsignaultetal. 1976; labelscorrespondtotheChandraObsIDs.Phaseφ = 0corresponds
orb
Pravdoetal.1980;Remillard&Canizares1984;Kitamotoetal. tothesuperiorconjunctionoftheblackhole.
1984, 1989; Wenetal. 1999; Bałucin´ska-Churchetal. 2000;
Feng&Cui 2002; Lachowiczetal. 2006; Poutanenetal. 2008;
a study it is essential that the source is in the hard state or the
Hankeetal. 2009), which is consistent with the focused wind
hard-intermediate state. The strong X-ray emission during the
picture (e.g., Li&Clark 1974; Remillard&Canizares 1984;
softstateissufficienttocompletelyphotoionizethestellarwind.
Bałucin´ska-Churchetal. 2000; Poutanenetal. 2008). Owing
Asaconsequence,thewindissuppressedbecausetheradiative
to the presence of the focused wind, the accretion pro-
drivingforceoftheUVphotonsfromthedonorstarisreduced
cess in Cyg X-1 is not primarily wind accretion like in
other HMXBs, such as Vela X-1, SMC X-1, or 4U1700−37 sincetheionizedgasistransparentforUVradiation.
(Bondi&Hoyle 1944; Blondin&Woo 1995; Blondin 1994; In this paper, we extend earlier work on Chandra high-
Blondinetal. 1991), but a small accretion disk is present. Ev- resolutiongratingspectrafromCygnusX-1,studyingtheionized
idence for this disk comes from detections of the disk’s ther- materialofthestellarwindofHDE226868duringthelow/hard
malspectrum(Bałucin´ska-Churchetal.1995;Priedhorskyetal. state (Hankeetal. 2009, hereafter Paper I). We perform a de-
1979), an X-ray disk reflection component, and a strong and tailed study of four observations at phases φorb ∼ 0.05, ∼0.2,
broadFeKαline(Tomsicketal.2014;Duroetal.2011,andref- ∼0.5,and∼0.75,andcombineitwithpreviousresultsoftheob-
erencestherein). servationat∼0.95(PaperI).Thesedataprovideauniquesetof
observations that allows us to probe all prominent parts of the
Blackholebinariesshowtwocharacteristicbehaviorscalled
the low/hard and the high/soft state. These states differ in the wind of CygX-1 (Fig. 1). Our aim is to describe the complex
structure and the dynamics of the stellar wind. Section 2 sum-
shapeoftheX-rayspectrum,thetimingproperties,andtheradio
marizes observations and data used in the analysis. In Sect. 3
emission (Fenderetal. 1999; Belloni 2004; Wilmsetal. 2006;
and Sect. 4, we present the data analysis and results related to
Belloni2010,andreferencestherein).ThespectrumofCygX-1
thecontinuumfittingandtheH-likeandHe-likeabsorptionlines
in the hard state is well described by a hard, exponentially cut-
off broken power law with photon index Γ ∼ 1.7 (Wilmsetal. observedinthenon-dipspectra.InSect.5,wediscusstheirmod-
2006).Duringthesoftstate,thepowerlawissteeper(Γ ∼ 2.5) ulation with orbital phase. Section 6 discusses line profile vari-
ationsandresultsfromplasmadensitydiagnostics.Wesumma-
and a luminous and less variable thermal disk component ap-
rizeourconclusionsinSect.7.Overviewtablesandplotsofall
pears(Wilmsetal.2006).
fullrangespectraaregiveninAppendixA.Technicalissuesre-
CygX-1isoftenconsideredtobeahardstatesource,since
latedtotheanalysisarediscussedinAppendicesBandC.
itspentmostofthetimeinthehardstate(Grinbergetal.2013).
Transitions into the soft state are observed every few years
and “failed state transitions”, where the soft state is not en-
2. Observationsanddatareduction
tirelyreached,arealsopossible(Pottschmidtetal.2003).Since
2010 the source’s behavior has been rather unusual and it has
2.1. Selectionofobservations
spentsignificantlymoretimeinthesoftstate.SeeGrinbergetal.
(2013) for a discussion of the long-term changes in Cyg X-1 The main purpose of this paper is to study the variation of the
from1996untiltheendof2012. X-rayspectrumwithorbitalphaseduringthehardstate.Onlysix
SinceX-raysfromtheblackholepropagatethroughthestel- ofthe17availableChandra-HETGobservationsmeetthiscon-
larwind,wecanusetheseX-raystoprobeitsstructure.Forsuch dition: ObsIDs 2415, 3814, 3815, 8525, 9847, and 11044 (see
A114,page2of24
I.Miškovicˇováetal.:ChandraspectroscopyofthefocusedwindofCygnusX-1.II.
Table1.Logofhard-stateChandra-HETGSobservationsofCygX-1usedinthispaper.
Startdate Countrates Non-dipspectrum
ObsID Date MJD Mode T φ r r h T L
exp orb ASM Chandra exp,non−dip 0.5−10keV
yy-mm-dd [ks] [cps] [cps] [ks] [1037erg/s]
3815 03-03-04 52702 CC/g 58.4 0.70–0.82 25.8 210.5 1.05–1.8 45 0.67
3814 03-04-19 52748 TE/g 48.3 0.93–0.03 17.5 88.6 – 16.1 0.39a
8525 08-04-18 54574 TE/g 30.1 0.02–0.08 14.2 87.5 0.667–0.94 4.4 0.43
9847 08-04-19 54575 TE/g 19.3 0.17–0.21 18.0 100.9 0.692–1.0 4.4 0.49
11044 10-01-14 55210 TE/g 30.1 0.48–0.54 18.4 91.1 – 30.1 0.41
Notes. Mode: CC/g – continuous clocking, graded; TE/g – timed exposure, graded; T : exposure time; φ : orbital phase according to the
exp orb
ephemerisofGiesetal.(2003);r :RXTE-ASM(1.5–12keV)countrate,averagedovertheobservation;r :Chandraspectral(non-dip)
ASM Chandra
count rate; h is the hardness ratio of the Chandra first order count rates in the 0.5–1.5keV band to those in the 3–10keV band; T –
exp,non−dip
exposuretimeafterexclusionofdips.NotethatObsID8525and9847werescheduledaspartofamulti-satellitecampaignontothesystem.For
technicalreasonsthis50ksecobservationwassplitintotwoparts.L –sourceluminosityinthe0.5–10keVenergyrange,derivedfromthe
0.5−10keV
non-dipspectraassumingasourcedistanceof1.86kpc(Xiangetal.2011;Reidetal.2011).(a)Hankeetal.(2009),correctedforthenewersource
distance.
60 15 14 2547 D4 4M
[cps]50 D38 D38 ∼90cps D85D98 ObsI1104 AXI
rate40 ObsI ObsI ObsIObsI 22–20k
nt30 0eV
u
o
c20
M
S
A10
Nov Dec Jan FebMar AprMay JunJuly Dec Jan FebMar AprMay JunJulyAug Sep Oct Nov Dec Jan FebMar AprMay
2002/2003 2007/2008 2009/2010
Fig.2.1.5–12keVRXTE-All-SkyMonitorcountrateofCygX-1withonedaybinningintheintervalsofninemonthsduring2003,2008,and
2010centeredatthetimesofChandraobservations:ObsIDs3815and3814(left),8525and9847(middle),and11044(right),indicatedbyvertical
lines.TherightpanelalsoshowstheMAXIlightcurve.
Sect. 2.2 for a discussion of the state classification). All other simultaneouswithObsID11044:theyallshowthesourceinthe
Chandra observations of CygX-1 either caught the source in hardstate,asdefinedbyGrinbergetal.(2013).
the high/soft state or were too short to obtain spectra of a suf-
ficientS/N.ObsID2415,takenduringtheintermediatestateat
φ ∼ 0.76 has already been analyzed by Milleretal. (2005). 2.3. Datareduction
orb
Table1givesalogoftheremainingobservationsstudiedinthis
For the spectral analysis we used the first order spectra of
paper and Fig. 1 depicts their orbital coverage. Results on Ob-
Chandra’s high and medium energy gratings (HEG, MEG;
sID3814fromPaperIwillbecombinedwiththesenewresults
Canizaresetal. 2005). Most HETG observations of CygX-1 in
inSects.4and5.
the analysis were performed in timed exposure (TE) mode of
Togaugethequalityofthecontinuummodeling,weusesi- theadvancedCCDimagingspectrometer(ACIS;Garmireetal.
multaneouspointedRXTEobservations(Sect.3.2),whichwere 2003). In this mode data are nominally accumulated for a 3.2s
reduced using our standard procedures as described, e.g., by frame time before being transferred into a framestore and read
Wilmsetal.(2006)orGrinbergetal.(2015). out, increasing the probability of pile up. In the cases of Ob-
sIDs3814,8525,9847,and11044,theframetimewasreduced
to1.7sbyusinga512rowCCDsubarrayreadout.
2.2. Sourcestate
Inevenbrightercases,asinObsID3815,whereCygX-1was
inthehard-intermediatestate,acontinuousclocking(CC)mode
Figure2showstheRXTEAll-Sky-Monitor(ASM;Levineetal.
wasappliedtoavoidpileupandconservelinefeatures.Herethe
1996) lightcurve of CygX-1 around the times of the Chandra
CCDrowsarereadoutcontinously,whichreducestheexposure
observations. The average daily ASM data point towards sim- perrowto2.85ms(Garmireetal.2003)andnoeffectsfrompile
ilar conditions during the first four observations. According to
up are expected. The application of CC-mode comes at the ex-
the scheme of Grinbergetal. (2013), all four observations are
pense of one spatial dimension as the image is then reduced to
found in the hard state, with low count rate (14–18cps) in Ob-
1024×1 pixel frames and the y-image dimension is lost. As a
sIDs 3814, 8525 and 9847, and a higher countrate of 26cps in
consequence,allimagephotons,includingnon-sourcedispersed
ObsID3815.
photons, such as from the source scattering halo, get collapsed
ObsID 11044 was taken during the time when ASM was intothespectrum,ordersortingbecomesmoredifficult,andcali-
deteriorating (Vrtilek&Boroson 2013; Grinbergetal. 2013). brationuncertaintiesarehigher,whichleadstolesserdetermined
We therefore used MAXI data (Matsuokaetal. 2009) for the continua.Someofthesetechnicalissuesarediscussedindetail
assessment of this observation. Six MAXI measurements are inAppendixB.
A114,page3of24
A&A590,A114(2016)
) 500 500
s
c/ ObsID3814 ObsID8525 ObsID9847 ObsID11044
(
te 200 2003-04-19/20 2008-04-18/19 2008-04-19 2010-01-14 200
a
r
nt 100 100
u
o
c 50 50
V
e
k
0
1 20 ObsID3815 20
–
0.5 10 2003-03-04/05 10
)
V 58ks 48ks 30ks 19ks 30ks
e
k
10 1 1
–
3
(
/
)
V
e 0.1 0.1
k
5
1.
–
5
(0. 0.01 0.01
0.70 0.75 0.80 0.95 0.00 0.05 0.20 0.50 0.55
Orbitalphaseφ
Fig.3.Upperpanel:lightcurvesofallfiveobservationsasafunctionoforbitalphase.Lowerpanel:variationofthehardnessratio.Exceptforthe
somewhatsofterhard-intermediatestateobservationObsId3815,allobservationswereinthelow/hardstate.Dipsarestrongestatφ ∼0.0,still
orb
presentatφ ∼0.2,andφ ∼0.75,andtheyvanishatφ ∼0.5.Colorsindicatethepartsoftheobservationsusedintheanalysis(seeFig.1),
orb orb orb
whiledatainblackwereexcludedfromtheanalysis.
ThedatawereprocessedwiththestandardChandraInterac- and its modulation with orbital phase. As shown in Fig. 3,
tiveAnalysisofObservations(CIAO)software,version4.2.Fur- dipping is clearly present for most of the orbit, but the dip
ther analysis was done with the Interactive Spectral Interpreta- frequency seems to be phase dependent. Light curves around
tionSystem(ISIS),versions1.6.1and1.6.2(Houck&Denicola φ ∼ 0.0 are modulated by strong and complex absorption
orb
2000). Cross-sections were taken from Verneretal. (1996), dips. Dipping already occurs at φ ∼ 0.7 and has not ceased
orb
abundancesfromWilmsetal.(2000),andatomicdatafromthe at φ ∼ 0.2. The dip events become shorter and shallower as
orb
AtomicDatabase,AtomDBv.2.0.1(Fosteretal.2012).Owing the black hole moves away from superior conjunction which is
totheverylowChandrabackground(comparedtothesource), expected,giventhatthelineofsightthroughthedensestregions
nobackgroundwassubtractedfromthefinalspectra.Datawere ofthe(focused)stellarwindislongestforφ ∼ 0.0.Thedata
orb
grouped to a minimum S/N = 10. Unless noted otherwise, all at φ ∼ 0.5, which probe only the outer regions of the stel-
orb
uncertainties are at the 90% level for one parameter of interest larwind,donotshowanydipping.Thisdistributionofdipping
(∆χ2 =2.71;Lamptonetal.1976). is consistent with theoretical predictions and observations that
For spectra from observations performed in TE mode, the see the high-density focused wind close to the binary axis at
effectofpile-uphastobeconsidered.Inthefirstorderspectra, φ = 0.Aconsequenceisahighprobabilityofseeingdipping
orb
this causes a pure reduction of count rate. It is stronger in the eventsatφ = 0andamuchsmallerprobabilityofdippingat
orb
MEG spectra than in the HEG spectra owing to the lower dis- φ =0.5(Bałucin´ska-Churchetal.2000;Poutanenetal.2008;
orb
persionandhighereffectiveareaoftheMEG.Theapparentflux Boroson&Vrtilek2010).
reductionismostsignificantnear2keV(6–7Å)wherethespec- Themaingoalofthispaperistostudytheeffectsofthestel-
trometerhasthelargestefficiencyandthehighestcountratesare
lar wind on the X-ray spectrum. The time intervals where the
obtained.AsdescribedingreaterdetailinPaperI,pileupinthe
data are distorted by dips need to be removed from the anal-
gratingscanbemodeledinISISusingthenonlinearconvolution
ysis. Properties, dynamics, and origin of the dips will be dis-
model simple_gpile2. This model describes the reduction of
cussedbyHelletal.(2013,andinprep.).Followingthediscus-
thepredictedsourcecountrate,S(λ),bypileupas sion of Hankeetal. (2008), different stages of dipping can be
S(cid:48)(λ)=S(λ)·exp(−β·S (λ)), (1) defined based on different count rate levels in the light curve,
tot
oronitsdependenceinthecolor-colordiagram,oronthesoft-
wherethetotalcountrateS (λ)isestimatedfromtheeffective ness ratio. According to Kitamotoetal. (1984), and confirmed
tot
areaandtheassumed(model)photonflux,andwherethescale by the lightcurves of our observations, dips can last from sev-
parameterβisafitparameter.S (λ)includesthecontributions eral seconds to more than 10minutes, especially around su-
tot
of the 1st, 2nd, and 3rd order spectra at the detector location perior conjunction. We therefore extracted light curves with
correspondingtoλ. a 25.5s resolution, except for ObsID 3814 where 12.25s were
used (PaperI), to be able to also identify short dipping in-
tervals. For ObsIDs 3815, 8525, and 9847 the selection of
3. Continuummodeling non-dip intervals was based on the hardness ratio defined as
(0.5–1.5keV)/(3–10keV). See Table 1 for the exact selection
3.1. Overview
criteria, which vary between observations owing to differences
A detailed look at the light curves and spectra of the five hard inthecontinuumshapeandtheresultingnon-dipexposuretimes.
state observations enables us to probe the structure of the wind As shown in Fig. 3, the selection criteria work well for all
A114,page4of24
I.Miškovicˇováetal.:ChandraspectroscopyofthefocusedwindofCygnusX-1.II.
observations,withonlyasmallcontributionduetoresidualdips Table2.ContinuumparametersofObsID8525,9847,and11044.
withlowN remaininginthelightcurves.Choosingslightlydif-
H
ferentselectioncriteriashowsthattheseresidualdipsdonotaf- Parameter ObsID8525 ObsID9847 ObsID11044
fectouranalysis.ForObsID3814,thecount-ratebasedselection φ ∼0.05 φ ∼0.2 φ ∼0.5
orb orb orb
ofPaperIisused.SincenodipsarepresentinObsID11044,the Power-law
full∼30ksofexposurecanbeusedintheanalysis. A [s−1cm−2keV−1] 1.23+0.03 1.45±0.02 1.38±0.01
PL −0.02
Residualdippingpresentintheso-callednon-dipspectracan Γ 1.43+0.02 1.45±0.01 1.59±0.01
−0.01
influence our fitting results. To gauge the influence of dips on TBnew
these spectra, we relaxed the hardness criterion and also ex- N [1022cm−2] 0.55+0.03 0.47±0.02 0.56±0.01
tractedaspectrumofObsID8525,whichincludesmoderatedip- NH [1018cm−2] 0.77−±0.002.18 1.13+0.14 0.58±0.05
pingandrefittedthecontinum.Thecontaminatedspectrumwas NNe[1017cm−2] – – −0.15 1.5±0.2
chosenveryconservatively,itcorrespondstoahardness≥0.449 Fe
Simple_gpile2
in the lower panel of Fig. 3. In the combined spectrum, NH β 4.40+0.26 3.8+0.1 3.8±0.1
changedby15%from5.5×1021cm−2to6.3×1021cm−2.Because HEG−1 −0.41 −0.2
ofourmuchmoreconservativeselectioncriteria,thesystematic βHEG+1 4.74+−00..2460 4.01±0.16 4.08±0.08
β 6.01+0.15 5.69±0.09 5.46±0.04
errorinournon-dipspectraissignificantlysmallerthanthat.We MEG−1 −0.20
estimateittobeofthesameorderofmagnitudeasthestatistical βMEG+1 6.41+−00..1137 6.41±0.07 6.15±0.03
errorsofthefits. χ2 3174.49 3661.78 12142.47
d.o.f. 2985 3382 11040
χ2 1.06 1.08 1.10
red
3.2. Continuummodel
Notes. A –fluxdensityofthepowerlawat1keV;Γ–photonindex
Theextractednon-dipspectraarecharacterizedbyanabsorbed PL
ofthepowerlaw;N –hydrogencolumndensity,N andN –neu-
continuum onto which a large number of absorption lines are H Ne Fe
tralcolumndensitiesofNeandFe;simple_gpile2β–pile-upscale
superimposed. As we are not focusing on a physical interpre-
parametersinunitsof10−2sÅ.
tation of the continuum, we describe it with a simple empir-
ical model that is flexible enough to give an accurate repre-
sentation of the proper continuum spectrum, i.e., an absorbed
power law. Cross-checks with simultaneous broadband spectra
parameter, gives A = 1.61 ± 0.01 and the χ2 = 2.34
from RXTE performed during our Chandra observations show PL red
thatourcontinuumparametersareinreasonableagreementbe- (χ2/d.o.f.=14.07/6).Theratiobetweenthedataandthemodel
tween the two satellites. We look at two extreme observations, again shows deviations of ≤1%, however, the residuals suggest
ObsID3815and11044.InObsID11044theChandraspectrum that adjustment of the slope of power law would improve the
is neither distorted by calibration features, nor influenced by fit. A fit with Γ left free, APL=1.72 ± 0.06, Γ = 1.45 ± 0.02,
strong absorption lines. Fitting the Chandra continuum model is consistent with the original fit to within the error bars. The
given in Table 2 to the simultaneous 3–6keV PCA spectrum, data-to-model ratio deviations lie below 0.5%. This best-fit has
leaving the normalization value free to vary to take into ac- an unphysically good χ2 = 0.21 (χ2/d.o.f. = 1.04/5), indi-
red
count the well-known flux-cross calibration issues between the cating that the systematic error in the PCA has been overesti-
PCA and other satellites (Nowaketal. 2011), gives a reason- mated.FittingthePCAdatawithoutapplyingasystematicerror
able χ2 = 1.65 (χ2/d.o.f. = 9.95/6) with consistent photon givesχ2 =0.87(χ2/d.o.f.=4.34/5).SeeNowaketal.(2011),
red red
indicesforthetwoinstruments.Whileformallynotaverygood Wilmsetal. (2006), and Gierlin´skietal. (1999) for physical
fit, the ratio between the data and the model in the PCA shows continuumdescriptions.
deviations of <1% and is therefore consistent with the calibra- Neutral absorption is modeled with the TBnew model
tion uncertainty of the PCA in this energy band (Jahodaetal. (Juettetal. 2006, Paper I, and http://pulsar.sternwarte.
2006;Shaposhnikovetal.2012).Thelargeχ2isthereforedueto uni-erlangen.de/~wilms/research/tbnew). Compared to
PCAcalibrationsystematics.Wenotethattheenergybandcho- the absorption model of Wilmsetal. (2000), TBnew contains a
sen,3–6keV,representsthemaximumoverlapbetweenthePCA better description of the absorption edges and allows a simple
and the HETGS, we deliberately do not extend the PCA data fittingofcolumnsofindividualelements(see,e.g.,Hankeetal.
to higher energies because we are only interested in determin- 2010).Ingratingsdatathistypeofapproachispossiblewhena
ing how well the continuum is described by the model in the strongabsorptionedgeispresentinthespectrum.Forthewave-
HETGSband. length range studied here, the most important edge is the Ne
As a second example, we consider the continuum of Chan- K-edge. Where indicated below, we therefore fitted the column
draObsID3815,whichhadtobemodeledusingacomplicated ofneutralNeindependentlyof N .AlthoughK-edgesofS,Si,
H
andnonphysicalcontinuum.Adirectcomparisonofthismodel Mg,andNaarealsopresent,theyarenotasclearlyvisibleinthe
with the contemporaneous RXTE-PCA data is complicated by spectraandtheabundancesoftheseelementswerefixedtotheir
the fact that these data were unfortunately taken during one of interstellar values (Wilmsetal. 2000). Taking into account pile
the deep dips in the lightcurve. We therefore extracted a Chan- up,theadoptedcontinuumshapewas
dra spectrum from the aforementioned dip and made a com-
parison between only strictly simultaneous data. We modeled
N (E)=simple_gpile2⊗(TBnew×powerlaw). (2)
it with the same continuum model that was also used for the ph
non-dip continuum, giving A =1.38±0.04, Γ = 1.40±0.03,
PL
N = 0.33±0.01×1022cm−2 and a χ2 = 1.07 (χ2/d.o.f. = Inthefollowing,webrieflydiscussthecontinuumpropertiesfor
H red
10550/9855). Applying the same fit to the PCA, leaving N the fourobservations modeled here. Werefer to PaperI for the
H
fixed at the Chandra value, leaving only the normalization a continuumdescriptionofthenon-dipspectrumofObsID3814.
A114,page5of24
A&A590,A114(2016)
3.3. ThecontinuumofObsIDs8525and9847(φ ∼0.05 Table3.ContinuumparametersofObsID3815.
orb
andφ ∼0.2)
orb
Parameter HEG−1 HEG+1 MEG−1 MEG+1
Afterfilteringfordips,only∼4.4ksofnon-dipdataremainfor
power-law
each of the two observations. To stay above a S/N of 10, the
A 2.70±0.01 2.68±0.01 2.24±0.01 2.53±0.01
continua of these two observations were modeled in the wave- PL
Γ 1.81±0.01 1.73±0.01 1.64±0.01 1.69±0.01
lengthrange2Å–15ÅforMEGand2Å–12ÅforHEG.Best-fit
tbnew
parametersarelistedinTable2. (cid:16) (cid:17)
Both observations show many short, but strong, dips in the NH[cm−2] (cid:16)0.352+−00..00(cid:17)0053 ×1022
light curve. It is probable that even after the exclusion of dips, N [cm−2] 0.76+0.01 ×1018
Ne −0.04
thenon-dipspectrumiscontaminatedbyfaintdips,whichmay GaussE ∼1.12keV(λ∼11Å)
haveaninfluenceonparametersobtainedintheanalysis.
A 0.23 0.17 0.15 0.04
σ 0.178 0.163 0.159 0.099
3.4. ThecontinuumofObsID11044(φ ∼0.5) GaussE ∼2.25keV(λ∼5.5Å)
orb
A 0.03 0.01 0.07 0.04
ThehighS/Nofthis∼30ksobservationallowsustomodelthe σ 0.232 0.117 0.283 0.196
continuum in the range of 1.7Å–20Å. The continuum is well
χ2/d.o.f. 29805/14406
described by Eq. (2). The Ne- and also Fe-column densities
χ2 2.07
were allowed to vary and are mainly constrained by the Ne K red
andFeL /L edgesat14.3Åand17.2Å/17.5Å,respectively,
2 3 Notes.A –fluxdensityofthepowerlawat1keVincm−2s−1keV−1;
which are very prominent in the spectrum. Best-fit parameters PL
Γ–photonindexofthepowerlaw;A–fluxofthebroadlinefeaturein
areagainshowninTable2.Asalsoindicatedbysimilarbehav-
phcm−2s−1;σ–linewidthinkeV.
ior in RXTE-ASM (Fig. 2), ObsID 11044 was performed in a
similarstateasObsID8525(φ ∼0.05)and9847(φ ∼0.2),
orb orb
andthereforeitisnotsurprisingthatthespectralparametersare
quantity, the Ne column density could be measured indepen-
verysimilar.
dentlyfromthetotalcolumn.
3.5. ThecontinuumofObsID3815(φ ∼0.75)
orb
The analysis of the ∼45ks non-dip spectrum of this CC-mode 4. Modelingoflinefeatures
observation is complicated by the lack of imaging informa-
tion and by calibration issues1. Below 2Å, both spectra show 4.1. Introduction
an excess of up to 50phcm−2s−1Å−1 for HEG and up to
Visual inspection of the non-dip spectra reveals that all spectra
150phcm−2s−1Å−1intheMEG.Thisexcessisprobablycaused showdiscreteline absorptionowingtohighlyionized material,
bycontaminationofthespectrabythedustscatteringhalothat mostly from H-like and He-like ions of S, Si, Al, Mg, Na, and
surroundsthesource,whichisclearlyvisibleinthedetectorim- Ne.Wealsoobserveintercombination(i)andforbidden(f)emis-
ages of other observations (see also Xiangetal. 2011), or con- sion lines of He-like ions. Various L-shell transitions of Fe are
tamination from the wings of 0th order image. As there is no alsopresent.Table4ofPaperIgivesacompletelistoftransitions
imaginginformationavailableinCC-mode,itisnotpossibleto fromH-andHe-likeionspresentinthespectrumofObsID3814
correct for this contamination. Since the low S/N of the data (φ ∼ 0.95).Mostoftheseabsorptionlinesarealsopresentin
orb
above 15Å does not allow for detailed spectral modeling, only thespectraofObsID8525,9847,and3815(φ ∼ 0.0–0.2and
orb
the2–15ÅHEGand2.5–15ÅMEGdataaretakenintoaccount φ ∼ 0.75). Since the wavelength range used here is smaller
inthefurtheranalysis. thoarbninPaperI,linesfromOviii,Ovii,Fexxvi,andNixxviii
UnlikefortheTE-modedata,thecontinuumherecannotbe areoutsideoftheinvestigatedspectralrange.
describedbythesimplepowerlawofEq.(2),sincenon-physical
TodescribethelinesfromtheH-andHe-likeions,wefitall
curvaturecausedbycalibrationissuesispresentinthespectrum.
transitionsoftheseriesfromagivenionsimultaneouslyusinga
We model this curvature by adding two non-physical Gaussian
curveofgrowthapproach(seePaperIforadetaileddescription
components, centered at ∼1.12keV and ∼2.25keV. Calibration
ofthismodel).Thisapproachallowsustodescribeweakerand
issues causing slope differences between both instruments also
oftenblendedlines,whiletheirdescriptionwithseparateGaus-
necessitatedseparatemodelingofthecontinuaoftheHEGand
sianprofileswouldoftenbeimpossible.Foreachlineseries,the
theMEG.Thefinalparametersofthecontinuumfitaresumma-
fitparametersarethecolumndensityoftheionresponsiblefor
rizedinTable3.
the line, N, the Doppler shift, v, and the thermal broadening
i i
We note that despite the fact that the broadband CC mode
parameter,ξ.AlllineshapesaremodeledusingVoigtprofiles.
i
calibration is suboptimal, its relative calibration over small
Inadditiontothelinesfromlineseries,otherabsorptionand
wavelength intervals is still good. This means that parameters
emissionlinesarealsopresentinthespectra.Wherethesetypes
ofabsorptionlinescanneverthelessbemeasured.Forexample,
oflineswereidentified,theywereaddedbyhand.Unlessnoted
equivalent widths of narrow absorption lines do not depend on
otherwise,lineshapesweredescribedusingVoigtprofiles,with
the overall shape of the continuum and are therefore not af-
fit parameters being the thermal broadening parameter, ξ, the
fected by local fitting or the shape of the non-physical contin-
naturallinewidth,Γ,i.e.,thefullwidthathalfmaximumofthe
uum model. Since the Ne column is, as in the other observa-
LorentziancomponentoftheVoigtprofile,andthelineflux, A.
tions,obtainedmainlyfrommodelingtheNeK-edge,i.e.,alocal
Negativefluxesdenoteabsorptionlines.
1 http://cxc.harvard.edu/cal/Acis/Cal_prods/ccmode/ The following sections discuss the details and peculiarities
ccmode_final_doc02.pdf ofeachobservation.
A114,page6of24
I.Miškovicˇováetal.:ChandraspectroscopyofthefocusedwindofCygnusX-1.II.
Table4.LowionizationSilinesinObsIDs8525(φ ∼0.05)and9847(φ ∼0.2).
orb orb
Line λ ObsID λ A ∆v ξ
lab obs
[Å] [Å] [10−4phs−1cm−2] [kms−1] [kms−1]
Sixii(Li) 6.7200±0.0003 8525 6.716±0.004 −3.4+1.0 −162±162 130+320
−1.2 −100
9847 6.715±0.003 −5.9±1.3 −244±114 380+200
xi −220
Si (Be) 6.7848±0.0003 8525 linenotpresent
9847 6.7750±0.0001 −3.2+0.8 −434±7 15+334
Six(B) 6.8565±0.0002 8525 6.858±0.008 −2.7−+01..92 43±332 600−+4600
−1.3 −500
9847 6.859±0.001 −1.2+1.3 126±33 ≤2595
Siix(C) 6.9285±0.0003 8525 6.9228±0.0001 −11+−111.4 248±2 ≤749
−15
9847 linenotpresent
Notes.λ :laboratorywavelength(Helletal.2013),λ :observedwavelength,A:lineflux(negative:absorption),∆v:velocityshift,ξ:thermal
lab obs
broadeningparameter.
x x
4.2. LinespectroscopyofObsIDs8525and9847 modeled properly, then Ne Lyβ (10.24Å), Ne Lyγ (9.71Å),
(φorb ∼0.05,φorb ∼0.2) andNexLyδ(9.48Å)arepredictedtobeweakerthanobserved.
Tofindthereasonforthisdiscrepancy,weexcludedaverynar-
The line identifications for these observations are shown x
row region of the spectrum where the Ne Lyα is located and
in Figs. A.1 and A.2. While we are able to fit all line series in x
xvii xix fittedthespectrumwithoutthisline.AllotherlinesoftheNe
ObsID9847,Ar andCa couldnotbeconstrainedinOb-
seriesweredescribedverywell.Thereis,however,noreasonto
sID8525.InadditiontotheabsorptionlinesfromH-andHe-like x
assumethatthedataintheregionof Ne Lyαareoflowqual-
ionsandFe,thespectrashowevidenceforlowerionizationab-
xii xi x ity,especiallynotfornon-dipspectrawiththeirgoodstatistics,
sorption lines of Si (Li-like), Si (Be-like), Si (B-like),
ix andthisdiscrepancyisprobablyduetocontaminationbynearby
andSi (C-like)inthe6.6Å–7.0Åband.Aswewilldiscussin x
Fe lines. Ne Lyα also appears to be asymmetric, possibly be-
greater detail in our analysis of the dip spectrum (Hirsch et al.,
inganindicationofPCygniprofiles(seeSect.4.4).Milleretal.
in prep.), these features become very strong during the deep-
(2005)3 reportinconsistenciesinequivalentwidthsbetweenin-
est phases of absorption dips. Their appearance in the non-dip x
dividuallinesintheNe series.Apartfrom(partial)saturation
spectrum therefore reveals the presence of cooler dense mate-
oftheresonancelineorblendsoflines,theauthorsarguethatan-
rial along the line of sight. This is not surprising given that
otherpossibleexplanationcouldbethatlinescomefromanin-
ObsIDs 8525 and 9847 represent the densest part of the wind, homogeneous region, or from different regions simultaneously.
φ ∼ 0.0–0.2, where most dipping is observed. The spectra
orb Marshalletal. (2001)4 describe an opposite problem whereby,
takenfartherawayfromthesuperiorconjunction,whicharefar
while the Lyα lines were strong, no Lyβ lines were detected,
lessdominatedbydips,arevirtuallyfreeoflowerionizationSi
xii concludingthatthelineswerenotsaturatedandthattheabsorb-
features,withtheexceptionofpossibledetectionsofaSi line
xi inggascoveredmorethan∼50%ofthesource.
in ObsID 11044 (φ ∼ 0.5) and a Si line in ObsID 3815
orb Emission lines are present in the spectrum, but are gener-
(φ ∼ 0.75). We note, however, that in all cases the optical
orb ally much weaker than the absorption lines. Since full width at
depthofthelinesismuchsmallerthanthatseenduringdipsand halfmaximum,Γ,andthethermalbroadening,ξ,ofthelineare
considerourresultstoberepresentativeofthenon-dipspectrum. correlatedforVoigtprofiles,itwasdifficulttoconstraintheirpa-
Best-fit line parameters of the Si features are listed in Table 4. rameters (Table 5). We therefore set Γ to a fixed value that is
xii xii
WealsonotethatSi λ6.72ÅblendswithMg Lyγandthe
locatedattheintersectionoftheconfidencecontoursforallin-
xiii xi
Si f emission line, while Si λ6.785Å blends with a line dividualemissionlinesfromthefreefits,whichappearedtobe
fromFexxiv.Theparametersoftheselinesarethereforediffi- split into two groups: for stronger lines, Γ was fixed to 15 eV
culttoconstrain2. andthethermalbroadeningreachesvaluesaround∼500kms−1.
xiii
Si forbiddenlineat6.74Åisasingleemissionfeaturein Thefullwidthathalfmaximumoftheweakerlineswasfixedto
thespectrumanditsmodelingisambiguousowingtothepossi- 1.5eVandξismostlyconsistentwithzero.
bleblendsdescribedabove.FixingΓandfittingthefeaturegives
line fluxes of 0.29+0.15 × 10−3phs−1cm−2 in ObsID 8525 and
0.56±0.16×10−3p−h0.1s3−1cm−2inObsID9847,whilethethermal 4.4. LinespectroscopyofObsID11044(φorb ∼0.5)
broadeningparameter,ξ,isalmostunconstrained. Whereas resonance transitions were always detected in absorp-
tion at early orbital phases, the spectrum at φ ∼ 0.5, taken
orb
duringtheinferiorconjunctionoftheblackhole,showsPCygni
4.3. LinespectroscopyofObsID3815(φ ∼0.75)
orb profileswithanemissioncomponentattherestwavelengthand
The full range spectrum is given in Fig. A.3. In this observa- a weak blue-shifted absorption component. Owing to the com-
tion,thelineseriesofArxvii,Caxix,Caxx,andFexviicould plex profiles, line series fitting is not feasible in this observa-
not be fitted. The approach of line series fitting does not al- tion. Describing the P Cygni profiles as the sum of a positive
lowustodescribeeverysinglelineperfectly,butthisismostly andanegativeVoigtprofileallowsustomodeltheshapesofthe
xiv xii x x
only the case for weak emission lines or Fe blends. Notable is Lyαtransitionsof Si ,Mg ,andNe ,Ne β,andδquite
x x
the discrepancy in the case of Ne . If Ne Lyα (12.13Å) is
3 ObsID2415:∼32ksatφ ∼0.76,CCmode,∼twiceashighfluxas
orb
2 Becauseofblends,theSiabsorptionlinesvisibleinthenon-dipspec- common.
trumofObsID3814werenotidentifiedassuchinFig.10ofPaperI. 4 ObsID1511:12.7ksatφ ∼0.84,CCmode.
orb
A114,page7of24
A&A590,A114(2016)
Table5.Intercombination(i)andforbidden(f)emissionlinesfromHe- 4.5. Linespectroscopy:Best-fitresults
likeionsatφ ∼0.75(ObsID3815).
orb
Figure4showstheparametersofindividuallineseriessumma-
rized in Table A.1: Doppler velocities v, column densities N
λ A ξ i i
and thermal broadening ξ for ObsIDs 8525, 9847, and 3815
[Å] [10−3phs−1cm−2] [kms−1] i
(φ ∼ 0.0–0.2and 0.75).ParametersobtainedinPaperIfrom
ix orb
Ne f 13.726 0.12±0.09 ≤200 ObsID 3814, φ ∼ 0.95, are shown for comparison. Because
Neixi 13.561 0.59+0.09 140+70 ofthedifferentofirbttingapproach,theparametersofObsID11044
Naxf 11.199 0.61+−00..1180 590−+130700 (φ ∼0.5)areexcludedfromthiscomparison.
Naxi 11.089 0.84−+00..3612 1200−+920600 orbAll series in ObsID 3815 (φ ∼ 0.75) are consistently
Mgxif 9.316 0.17−+00..1097 ≤2−05000 redshifted, with values falling in othrbe interval between 100 and
Mgxii 9.235 0.47±−00.0.067 ≤180 400kms−1. The exception is the Nax line, which is Doppler
Alxiif – shifted by ∼870kms−1. The column densities of all line series
Alxiii 7.81 0.4±0.08 550+200 (Fig.4)areintheinterval(0–5)×1016cm−2withathermalbroad-
Sixiiif 6.742 0.79+0.05 <1−41950 eningofξ (cid:46) 500kms−1,suggestingthelinesoriginatefromthe
Sixiiii 6.685 0.27−+00..0078 360+140 sameregion.
xv −0.07 −310 Because of the short exposure of ObsIDs 8525 and 9847,
S f –
Sxvi 5.054 0.05+0.09 ≤200 many fit parameters show large uncertainties or cannot be con-
−0.05 strained. The velocities in ObsIDs 8525 and 3814 (φ ∼ 0.0)
orb
Notes.Fullwidthathalfmaximum,Γ,wasfixedto15eVforthelines aremuchlowercomparedtoObsID3815,infact,theyarecon-
indicatedinbold.Foralltheotherlines,itwasfixedto1.5eV.Emission sistentwithzero,whiletheonesfromObsID9847(φorb ∼ 0.2)
linesofArxviiandCaxixwerenotpresentinthespectrum. showmostlyblueshiftedvalues.Overall,thevelocitiesspanthe
range±500kms−1.
The column densities of ObsIDs 8525, 9847, and 3814 are
xi
well (Table A.2). In addition, the data show the Mg triplet, much higher than those of ObsID 3815, supporting the idea of
fourpureemissionlines,andonlytwopureabsorptionlines.See the denser focused wind around φ ∼ 0.0. We do not see any
orb
Fig.A.4forthefullspectrum. particular trend in the thermal broadening measured in these
Although the interpretation of P Cygni profiles is challeng- observations, although with (cid:46)300kms−1 it seems to be gener-
ing,weattempttodeterminecolumndensitiesfortheabsorption ally lower than in ObsID 3815 (even though the broadening
partoftheprofilesusing(Mihalas1978,seealsoPaperI) shows a rather large scatter). For a plasma with a temperature
of T ∼ 106K, thermal velocities are on order of ∼ 10kms−1.
mc2W 1.13×1017cm−2 (cid:32)λ (cid:33)−2 (cid:32)W (cid:33) The higher observed values therefore suggest that most of the
N = λ = · 0 · λ , (3) broadeningisduetomicroturbulence.
i πe2· fijλ20 fij Å mÅ Closer investigation shows that for lines with lower S/N a
higher column appears to imply smaller thermal broadening.
assumingthatthelinesarenotsaturated(τ(ν)(cid:28)1).Here,W is This kind of behavior could explain the outlier lines. For ex-
λ xiv
theequivalentwidthoftheline,and f istheoscillatorstrength ample,inthecaseof Si inObsID8525,assumingatypical
ij
of the transition, in this case a sum of all transitions contribut- value of ξ ∼ 300kms−1 would reduce the column density by a
ing to the absorption line. The values of N are summarized in factorof∼10.Forotherlinesthiseffectwillbesmaller.
x i
TableA.2.ColumndensitiesforNe βandδareoutliers.Values
of∼75and∼ 85×1016cm−2 arefartoohigh,giventhecolumn
x 5. TheX-rayorbitalvariabilityofCygX-1
density of Ne α. This deviation is most likely due to a partial
fillingoftheabsorptioncomponentbythestrongemissionwing, Having described the properties of the individual observations,
whichcannotbedeterminedreliablyowingtothelowS/Nofthe wenowturntolookattheminthecontextofthedifferentlines
line(Fig.A.4).Columnswerealsodeterminedfortheabsorption of sight onto the black hole to better understand how the prop-
xii
lineatλ6.705Å,assumingitisfromLi-likeSi ,usingtheos- erties of the wind depend on the orbital phase. We first study
cillatorstrengthofBehar&Netzer(2002). thevariationofthecolumndensityobtainedfromcontinuumor
The fact that the columns of the absorption lines are much edgemodelingwithorbitalphaseandtheninvestigatethevaria-
smaller in this observation compared to earlier orbital phases tionoftheionizingabsorberbylookingatcolumndensitiesand
demonstratesthatthecolumndensityoftheionizedabsorberis velocityshiftsofindividuallineseries.
stronglymodulatedwithorbitalphase,althoughtheblendingof
the absorption and the emission component strongly influences
5.1. Columndensityvariationoftheneutralabsorber
the interpretation of the columns found for this observation. A
more detailed study would require fits of P Cygni profile mod- TheequivalenthydrogencolumndensityN foundfromcontin-
H
els using line profile shapes in which the exciting source is not uumfittingis,inprinciple,atracerforthematerialintheinter-
situated at the center of the complex, asymmetric flow. To our stellar medium along the line of sight to the X-ray source and
knowledge,suchcalculationsarenotavailableasyet. for the moderately ionized material in its vicinity. The reason
Doppler shifts, v, of the absorption and the emission com- isthat0.5–10keVX-raysaremainlyabsorbedbyK-shellelec-
i
ponents of the P Cygni profiles were determined separately trons. As the cross-section of the K-shell is only mildly depen-
(Table A.2). As the emission components are located rather dent on ionization stage, absorption in moderately ionized ma-
x
close to the rest wavelength and, except for lines from Ne , terialcanalsousuallybedescribedreasonablywellwithcross-
areredshifted,andtheabsorptiontailsare,incontrast,strongly sections for neutral atoms, at least at the level that is usually
blueshifted,theyprovideonlyupper/lowerlimitsofthevelocity used when modeling broadband continua. N is thus a mea-
H
values. sure for both the ISM foreground absorption and the column
A114,page8of24
I.Miškovicˇováetal.:ChandraspectroscopyofthefocusedwindofCygnusX-1.II.
Ne IXNe NXa X N a MXIg XIM g AXlIIXIIAl XISIiIXIII Si XIV S X V S X VI Ar X VII Ar X VIII C a XIX C a X X
1000
a)
500
]
1
−
s
m 0
k
[
vi
-500
-1000
b)
100
]
2
−
m
c
6 10
1
0
1
[
Ni 1
0.1
c)
1000
]
1− 100
s
m
k 10 ObsID 3815
[
ξi ObsID 3814
ObsID 8525
1
ObsID 9847
0.1
1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0
Ionization potential [keV]
Fig.4.FitparametersfromTableA.1forlineseriesofH-likeandHe-liketransitionsinObsID3815(squares),ObsID8525(circles),ObsID9847
(stars),andalsoObsID3814(diamonds,PaperI):top:velocityshifts,middle:columndensities,bottom:thermalbroadening.Colorscorrespond
toObsIDsasindicated.Forcompleteness,valuesfaraboveorbelowthedisplayedrangearemarkedbytrianglesclosetoupper/lowerx-axes.
of mildly ionized material in the stellar wind. Earlier monitor- constrain the continuum with Chandra in these absorption line
ing, e.g., with RXTE, showed a modulation of N with orbital dominateddata.Ourexperienceshowsthatfitstosimultaneous
H
phase (Grinbergetal. 2015; Hanke 2011; Wilmsetal. 2006; ChandraandRXTE-PCAdatadohavesignificantlyhigher N .
H
Kitamotoetal. 2000; Holtetal. 1976), although we note that Forexample,inPaperI,wefoundforobservation3814thatN
H
someofthemodulationseenthereisalsocausedbydips,which increased from 3.52±0.041021cm−2 in the Chandra-only fits
werenotremovedintheseanalyses. to N = 5.4±0.4 ×1021cm−2 when including the PCA data,
H
whichwasconsistentwithpreviousmeasurements(Milleretal.
ThevaluesobtainedfromtheChandracontinuumfittingfor 2002; Schulzetal. 2002) as well as with the N obtained from
H
the N (Fig.5,graydatapoints)showasystematicoffsetcom- continuum fitting to observation 8525, which has a similar or-
H
paredtootherdeterminationsofthecolumn,withlowestvalues bitalphase.WethereforeconcludethatN valuesobtainedfrom
of (cid:16)3.52+0.05(cid:17) × 1021cm−2 at φ ∼ 0.75 and (3.52 ± 0.04) × continuummodelingtheChandradataalHonesufferalargesys-
−0.03 orb
tematic uncertainty in these line dominated data, with a clear
1021cm−2 (Paper I) for φ ∼ 0.95. These values are smaller
orb systematicbiastowardssmallerN .
than the total column density of the interstellar medium along H
the line of sight to CygX-1, which dust scattering measure- Anindependent,andlikelybetter,measurefortheabsorbing
mentsdeterminetoN ∼4.6 × 1021cm−2(Fig.5,dashedline; columnispossiblebydirectmeasurementoftheopticaldepthat
Htot
Xiangetal.2011).Thiscolumnsetsthelowerlimittothetotal absorptionedges.Thisapproachislessdependentonsystematics
N toCygX-1,whichisthesumoftheintrinsicabsorptioninthe ofthecontinuummodelingthanfittingN ,especiallyconsider-
H H
systemandtheabsorptionintheinterstellarmedium.LowerN ing the systematics induced by the brightness of the source. In
H
values,asthoseseenhere,thusindicateasystematicerror.This our data, Neon is the only element for which we can measure
is not unlikely, considering that the lowest N values originate the column density directly from the neutral edge in all of our
H
fromanobservationthatwasperformedincontinuousclocking observations (Fig. 5, colored data points). The observed varia-
mode, where the continuum has to be modeled with additional tionsuggeststhatpartoftheneutralN isintrinsictothesource
Ne
broadGaussians(Table3).Afurthersystematicistheinabilityto and varies with orbital phase, with a minimum at φ ∼ 0.5,
orb
A114,page9of24
A&A590,A114(2016)
H-like Ar). Figure 4b shows that these N values are, to within
1.3 i
their uncertainties, consistent with the more tightly constrained
1.2 14 values. Nex and Nax of ObsID 9847 (φ ∼ 0.2) appear as
orb
1.1 (d outliers,althoughcomparedtoothercolumnsdeterminedinthis
12er observationtheyarerelativelywellconstrained.InObsID11044
82cm]−10..90 10ived)N (tNhφieosrbahbo∼swonr0p.i5tni)o,Fnwigph.ae6rrteswotehfreeNsdeterxtoeαnrg,mMliinngeedxsiaiss,hMdoewgscxPrii,bCaeyndgdinnSiSipxerocivfit..l4es.4, tfhoer
100.8 H
1 There are four ions that are strong in all five observations
[e0.7 8 [10 andsocanbetrackedoverthewholeorbit:Nex,Mgxii,Mgxi,
NN0.6 c21 and Sixiv (Fig. 6, right). On closer investigation the lines ap-
m
pear asymmetric, even in cases where they do not show clear
6 −
0.5 2 P Cygni profiles. As asymmetry can affect the column density
]
0.4 measurements,weconsideredonlythoselinesthatdonotshow
4 any asymmetry according to Sect. 6.1 (filled circles in Fig. 6).
0.3
As with the cases discussed above, N is expected to reach its
0.0 0.5 1.0 1.5 2.0 i
maximum at φ ∼ 0.0 and its minimum at φ ∼ 0.5. The
Orbital Phase orb orb
modulationisclearlyvisibleinFig.6.
Fig.5.Orbitalphasevariationoffittedcolumns,repeatedtwiceforclar-
ity.Graydatapointsshowtheequivalenthydrogencolumndensity,N ,
H
fromcontinuumfittingtheChandraspectra(righthandy-axis).Asdis- 5.3. Thelocationoftheabsorbers
cussedinthetext,thedatapointsareheavilyinfluencedbysystemat-
ics(theerrorbarsshownarestatisticalonly).Thecoloreddatapoints The previous sections showed that both the moderately ionized
showthevariationoftheNeneutralcolumnfromNe-edgefitting(left material tracked by Ne-edge fitting and the highly-ionized ma-
y-axis)andcorresponding N (righty-axis)derivedassumingaNe:H terial responsible for the H- and He-like species shows orbital
H
abundanceof1:11481asperWilmsetal.(2000).Thephasesofindivid- modulation,i.e.,thematerialislocaltotheX-raybinary.Inthis
ualChandraobservationsaredenotedwithcoloredregions.Thedashed section, we consider the relationship between both absorbers.
lines represent the total equivalent column density of the interstellar First,asalsopointedoutbyMarshalletal.(2001),wenotethat
mediumandofNe(Xiangetal.2011).
inmostobservationstheHe-likelineseriesshowlowercolumns
than their H-like peers, N ≤ N (Fig. 4), which indicates
i,He i,H
that most of the material traced by the H- and He-like lines is
consistent with earlier studies of the N variation in the sys-
H fully ionized. The best example in this respect is ObsID 3815
tem(e.g.,Grinbergetal.2015;Wenetal.1999;Kitamotoetal.
where,forallobservedlineseries,thecolumnoftheH-likeions
2000).Wenote,however,thatapotentialsystematicuncertainty
is larger than the column from the He-like ions. In this ObsID,
isthatonlyObsID11044(φ ∼0.5)doesnotshowanysignif-
orb φ ∼ 0.75,i.e.,thelineofsightisalmostperpendiculartothe
icantabsorptionlinesintherangeoftheNeedge,whileforthe orb
binary axis. For the other observations probing the denser part
otherObsIDsthisregionispopulatedbymanystronglinesfrom
ofthewind,therelationbetweenN andN isnotasclearas
ionizedFe.Itthereforecannotbeexcludedthatsomeofthevari- i,He i,H
forObsID3815,butN ≤ N isfulfilledformostions.
ationisduetoacontaminationoftheNeedgebytheselines.A i,He i,H
To compare the measured variation of the columns with
cross-checkinwhichthecontinuumaroundtheNeedgeismod-
theoretical expectations we use the focused wind model of
eledlocallyrevealscolumnsthataresystematicallyhigherbya
Gies&Bolton (1986) as a toy model. This model consists of
factor of ∼2, such that at least part of the modulation could be
a CAK-model with a longitudinal variation of the wind param-
duetotheFelines.Localandcontinuummodelinggivesimilar eters in a cone ±20◦ degrees from the line between the donor
results only for ObsID 3815 (φ ∼ 0.75). In addition to Ne,
orb andtheblackhole.Outsideofthatregion,weusethewindpa-
thegoodqualityofObsID11044(φorb ∼0.5)allowsustoleave rametersatθ = 20◦,whereθ istheangleintheorbitalplaneof
theFecolumnafreeparameterinthecontinuumfitting,giving
the binary, measured from the line between the donor and the
N =(0.15±0.02)×1018cm−2forthisobservation.Thisvalue
Fe black hole. Figure 7 shows the density structure of the focused
isconsistentwithN determinedfromthelocalmodelingofthe
Fe wind, as well as the line of sight of our observations projected
edgeandwiththeexpectedISMFecontributionusingtheISM
onto the orbital plane and in a side view. The model has been
abundancesofWilmsetal.(2000).
showntobeagoodoverallrepresentationofthestellarwindin
theHDE226868/CygX-1system,eventhoughthedetailedpa-
rameters of the wind are still debated (Giesetal. 2003, 2008;
5.2. Orbitalvariationofthecolumnoftheionizedabsorber
Vrtileketal. 2008, and references therein). We emphasize that
Wenowturntothevariationoftheionizedabsorberwithphase. much of this discussion relates to the wind properties in the
Figure 6 illustrates the variation of the column densities N so-called shadow wind, i.e., the region of the star opposite to
i
of highly ionized elements with orbital phase. The parameter the black hole where X-rays are not presumed to influence the
sample was reduced compared to the complete set of species wind properties (Caballero-Nievesetal. 2009). This region is
present in the spectra by excluding all columns with uncertain- notprobedbythelineofsightsstudiedhere.
ties exceeding 75% of their values. For ObsIDs 3814 and 3815 Figure 8 (solid line) shows that, in the model of
(φ ∼ 0.95and0.76)onlyone N valueperobservation,from Gies&Bolton(1986),thetotal(neutralandionized)columnis
SxorvbiandArxvii,respectively,waislost.ForObsIDs8525and around N ∼ 4 × 1022cm−2, i.e., it is more than a factor of
H
9847 (φ ∼ 0.05 and 0.2, ∼4.4ks exposure) we discarded 10higherthantheN measuredfromfittingthecontinuumorthe
orb H
roughlyhalfofthe N values(ObsID8525,φ ∼ 0.05:H-like N valuesinferredfromfittingtheNeedge.Thelattervarybe-
i orb H
seriesofNe,Al,Si,S,ArandCa,andHe-likeNa;ObsID9847, tween6.8×1021cm−2and1.3×1022cm−2(Fig.5).Owingtothe
φ ∼ 0.2: both H- and He-like series of Al, Si, S, Ca, and expectedhighdegreeofionizationcausedbytheblackholeand
orb
A114,page10of24