Probing the stellar wind environment of Vela X-1 with MAXI

0.44 MB·

by  C. Malacaria

Astronomy&Astrophysicsmanuscriptno.velamalacaria_revised cESO2016
(cid:13)
February19,2016
Probing the stellar wind environment of Vela X-1 with MAXI
C.Malacaria1,2,T.Mihara2,A.Santangelo1,K.Makishima2,3,M.Matsuoka2,M.Morii2,andM.Sugizaki2
1 InstitutfürAstronomieundAstrophysik,Sand1,72076Tübingen,Germany
e-mail:[email protected]
2 MAXIteam,RIKEN,2-1Hirosawa,Wako,Saitama351-0198
3 DepartmentofPhysics,TheUniversityofTokyo7-3-1Hongo,Bunkyo-ku,113-0033Tokyo,Japan
February19,2016
6
1
0 ABSTRACT
2
Context.VelaX-1isoneofthebest-studiedandmostluminousaccretingX-raypulsars.Thesupergiantopticalcompanionproduces
b
astrongradiativelydrivenstellarwindthatisaccretedontotheneutronstar,producinghighlyvariableX-rayemission.Acomplex
e
phenomenology that is due to both gravitational and radiative effects needs to be taken into account to reproduce orbital spectral
F
variations.

8 Aims.WehaveinvestigatedthespectralandlightcurvepropertiesoftheX-rayemissionfromVelaX-1alongthebinaryorbit.These
1 studiesallowconstrainingthestellarwindpropertiesanditsperturbationsthatareinducedbythepulsatingneutronstar.
Methods.WetookadvantageoftheAllSkyMonitorMAXI/GSCdatatoanalyzeVelaX-1spectraandlightcurves.Bystudyingthe

] orbitalprofilesinthe4 10and10 20keVenergybands,weextractedasampleoforbitallightcurves( 15%ofthetotal)showing
E adiparoundtheinferio−rconjunction−,thatis,adouble-peakedshape.Weanalyzedorbitalphase-averaged∼andphase-resolvedspectra
H ofboththedouble-peakedandthestandardsample.
Results.Thedipinthedouble-peakedsampleneedsN 2 1024cm 2tobeexplainedbyabsorptionalone,whichisnotobserved
h. inouranalysis.WeshowthatThomsonscatteringfromHan∼exte×ndedand−ionizedaccretionwakecancontributetotheobserveddip.Fit
p byacutoffpower-lawmodel,thetwoanalyzedsamplesshoworbitalmodulationofthephotonindexthathardensby 0.3aroundthe
- inferiorconjunction,comparedtoearlierandlaterphases.Thisindicatesapossibleinadequacyofthismodel.Incontr∼ast,includinga
o
partialcoveringcomponentatcertainorbitalphasebinsallowsaconstantphotonindexalongtheorbitalphases,indicatingahighly
r
t inhomogeneous environmentwhosecolumndensityhasalocalpeakaroundtheinferiorconjunction. Wediscussourresultsinthe
s frameworkofpossiblescenarios.
a
[ Keywords. X-rays:binaries–stars:neutron–stars:winds–accretion–pulsars:individual:VelaX-1

2
v
51. Introduction center,andDthedistanceofthecompactobjectfromthecenter
3 ofthecompanion.
7VelaX-1istheprototypeoftheclassofwind-fedaccretingX-ray
TheobservedaverageX-rayluminosityof 4 1036erg/sis
7binaries.Thesourcehasbeendiscoveredin 1967(Chodiletal. ∼ ×
consistentwithawind-fedaccretingpulsarscenario(seeEq.1).
01967)andislocatedatadistanceof 2kpc(Nagase1989).Itis
∼ However,itistypicalofVelaX-1toshow"off-states"inwhich
.aneclipsingsupergiantHighMassX-rayBinary(HMXB),con-
1 it becomes faint for several 102 103s, and "giant flares",
sisting of a neutron star (NS) with mass M 1.77M (Rawls
0 NS −
et al. 2011) and a B0.5Ib optical companion,∼HD 775⊙81, with when the source flux suddenly increases up to a factor of 20
6
(Kreykenbohmetal.2008;Sidolietal.2015).Densityinhomo-
1mass MB 23M and radius RB 30R (van Kerkwijk et al.
:1995). ∼ ⊙ ∼ ⊙ geneities are observed in the matter during the accretion stage
v (seeMartínez-Núñezetal.2014,andreferencestherein)andare
AsketchofthebinarysystemisshowninFig.1.Thebinary
i considered a possible reason for the observed X-ray flux vari-
Xorbitis almostcircular (eccentricitye 0.089),with a periodof
∼ ability.
8.9d(Kreykenbohmetal.2008).TheNSorbitsataclosedis-
r
a∼tancefromthecompanion’ssurface,0.7R attheapastron,and Orbital phase-resolved studies of VelaX-1, both observa-
B
itiscontinuouslyembeddedinitsstrongstellarwind,character- tional and theoretical (see Blondin et al. 1991, and references
izedby M˙ = 1.5 2 10 6M yr 1 (Watanabeetal.2006) therein),provideevidenceofastrongsystematicmodulationof
wind − −
andterminalvelocity−v ×1100km⊙s 1(Prinjaetal.1990).Pass- theabsorbingcolumndensitywithorbitalphase.Thisisalsoob-
−
ingwithintheNSaccre∞ti∼onradius,r =2GM /v2 ,thiswind servedinotherHMXBs(seetheseminalcaseofCenX–3,Jack-
acc NS rel
is accreted onto the compact object, powering the X-ray emis- son 1975, and more recently, the study of IGR J17252–3616,
sion,withluminosity Manousakis et al. 2012). Such an effect is generally attributed
totheNSinfluenceonthestellarwindbyvariousmechanisms.
These include the formation of a tidal stream trailing the NS
L = (GMNS)3M˙wind , (1) (dueto the almostfilled Roche lobe of the companion),the X-
X R v4 D2 rayphotoionizationofthewind(whichleadstotheformationof
NS rel
aStrömgrensphereandaphotoionizationwake),andanaccre-
whereGisthegravitationalconstant,R 10kmistheNSra- tionwakethatsurroundsthecompactobject(duetothefocusing
NS
∼
dius,v therelativevelocityofthewindseenfromtheaccretion ofthestellarwindmediumbytheNSgravitationalinfluence).A
rel
Articlenumber,page1of10page.10
A&Aproofs:manuscriptno.velamalacaria_revised
Table1.Observationtimeandephemerisusedforlightcurvesfolding.
150
0
0. Parameter Value
9
100 0. ∆T [MJD]a 55133 56674
0. Obs ÷
1
P[d] 8.964357 0.000029
±
T [MJD]b 55134.63704
50 0.8 Mid−Eclipse
nds ] 0 0.2 N(b)oTteask.en(a)asTtihmeephsapsaen-zoerfot.he MAXI observations used in this work;
o
c
-se 0.7 sample of double-peaked orbital profiles in two energy bands,
t
h -50 whosedipisaroundtheinferiorconjunction.Weperformorbital
g
[ li 0.3 phase-averagedand phase-resolvedspectroscopywith different
modelsandanalyzeourresultsusingthemostrecentstellarwind
0.
-100 6 accretionscenarios.
4
0. 0
.5
2. Observationsanddata
-150
The all-sky monitor MAXI (Matsuoka et al. 2009) is operating
from the International Space Station (ISS) since August 2009.
To Earth
-200 It consists of two slit cameras working in two complementary
-150 -100 -50 0 50 100 150
energy bands: the Solid-state Slit Camera (SSC, Tomida et al.
[ light-seconds ]
2011)inthe0.7 10keV,andtheGasSlitCamera(GSC,Mihara
−
etal.2011)inthe2 20keVenergybands.Hereweonlyused
Fig. 1. Orbital sketch of the binary system VelaX-1 as observed −
archival GSC data. The GSC module is composed of 12 pro-
from Earth. The blue disk centered at (0,0) is the optical companion
portionalcounters,eachofwhichconsistsofaone-dimensional
HD77581. The black solid line is the NS elliptical orbit. The gray
dashed rays and the respective numbers indicate the orbital phases. slit-slatcollimatorandaproportionalcounterfilledwithXe-gas.
Phasezeroistakenatthemid-eclipsetime.Theredstraightsolidline Assembled,theyallowcoveringafieldofviewof1.5 160deg2,
×
crossing the ellipse represents the major axis, pointing the periastron observing about 75% of the whole sky at each ISS revolution
andtheapastronattheintersectionswiththeNSorbit. ( 92min), and 95% per day. The GSC typically scans a point
∼
sourceontheskyfor40 150sduringeachISSrevolution(the
−
actualtimedependsonthe sourceincidentangleforeachGSC
sketchof thebinarysystem whereallthesestructuresareillus- counter).Thedailyfluxsourcesensitivityat5σc.l.inthefullen-
tratedisshowninFig.2. ergybandisabout20mCrab,andthe1σc.l.energyresolutionis
Many authors have reported on the X-ray light curve and 18%at5.9keV(Sugizakietal.2011).Thesepropertiesallowed
spectral modulation of VelaX-1 around the inferior conjunc- frequently observing VelaX-1 along its orbit and to collecting
tion (e.g., Eadie et al. 1975, Watson & Griffiths 1977, Nagase datafromthissourceformorethanfouryears.
1989). At the same time, the emission from the optical com- WeextractedspectraandlightcurvesofVelaX-1fromGSC
panion also shows modulation with the orbital phase. An ab- data covering almost the whole MAXI operation time, that is,
sorption feature was reported by Smith (2001), who observed from55133to56674MJD(seeTable1).Ourdatawereproperly
a dip in the VelaX-1 UV light curve around the transit phase correctedtoaccountforthetime-dependenteffectiveareaofthe
(0.46 < φ < 0.70 in their work, where φ = 0 is the mid- cameras.Allextractedspectrawererebinnedtocontainatleast
eclipse). Moreover, Goldstein et al. (2004), analyzing spectro- 200 photonsper energybin. Following the MAXI team recom-
scopicChandradatatakenatthreedifferentphases(i.e.,eclipse mendation,we addeda systematic errorata levelof 1% to the
φ = 0,φ = 0.25andφ = 0.5),foundanabsorbed“eclipse-like” finalcountratesoftheextractedspectra.
spectrum at φ = 0.5 in the soft X-ray energy range. Similarly,
Watanabeetal.(2006)foundanextendedstructureofdensema-
terial that partially covers the X-ray pulsar, which is between 3. Orbitalprofilessamplesanalysis
the X-ray emitter and the observer at the inferior conjunction.
The mission-long light curve of VelaX-1 as observed by
Morerecently,Naiketal.(2009)analyzedorbitalphase-resolved
MAXI/GSC is providedby the MAXI team1 at differentenergy
spectraofVelaX-1observedwithRXTE,findinglargevariation
ranges.Itexhibitsarichvarietyoftimingphenomena.However,
of the absorbing column density along the orbital phase (up to
weherefocusontheorbitalpropertiesofthesource.We there-
1024cm 2),buttheirobservationsdonotcoverthewholeorbit,
− foreextractedMAXI/GSClightcurveswithatimeresolutionof
that is, their data do not include observations between the in-
6hintwoenergybands,4 10and10 20keV.Then,wefolded
ferior conjunction and the apastron. Finally, Doroshenko et al. − −
the light curves to produce average orbital profiles in the two
(2013) have shown that the observed increasing absorption at
chosenenergybands,usingtheorbitalephemerisfromKreyken-
lateorbitalphasedoesnotagreewithexpectationsforaspheri-
bohmetal. (2008)thatwe reportin Table1. The phasezerois
callysymmetricsmoothwind.
takenat the mid-eclipsetime (see Fig. 1). The resulting orbital
Inthispaperwe presentthe spectroscopicalandlightcurve
light curves are shown in Fig. 3, where the strongest changes
analysis of VelaX-1 observed with the Monitor of All-sky X-
rayImage(MAXI).Ourdatashowanexcellentcoverageofthe 1 TheentireMAXI/GSCVelaX-1lightcurvecanbeexploredathttp:
entireorbitandextendovermorethanfouryears. We extracta //maxi.riken.jp/top/index.php?cid=1&jname=J0902-405
Articlenumber,page2of10page.10
C.Malacariaetal.:VelaX-1observedwithMAXI
Fig.2. Sketchof thebinary systemVelaX-1whereboth radiativeandgravitational effectsareillustratedat threedifferent orbital phases. The
photoionizationwakeisrepresentedasadisk-likestructuretoshowboththecumulativeeffectalongoneentireorbit(theredring)andthemore
localizedeffect(thefainterredtail)trailingtheNSattheinferiorconjunction.Aclumpystellarwind,notshowninthepicture,mayalsobetaken
intoaccountforamorecompletescenario.
ofdatapointsineachphasebinequalto4(i.e.,1dobservation),
4-10 keV
whichleavesuswithatotalof84statisticallyacceptablecycles.
0.15 10-20 keV
Thisallowedustoperformasystematicanalysisofthosemem-
bers that differ from the template in each of the three defined
phasebins.Forthesystematiccomparisonofthemembers’pro-
fileagainstthetemplate,weconsideredonlytherelativeshape,
0.10 avoiding any bias from flux variation. For this reason, we nor-
e
at malized each member with a normalization factor equal to the
unt r ratiooftheaveragecountratebetween0.10 0.85andthetem-
o −
C plateaveragecountrateinthesamephaseinterval.Successively,
0.05 wesubtractedthecountratesofthetemplatefromthecountrates
of each memberin the X, Y, and Z phase bins. To evaluatethe
templateand the membercountrate in each phase bin,we cal-
X Y Z
culated the median of the data points. The obtained count rate
differenceisreferredtoastheresidualswithrespecttothetem-
0.00
plate.TheresidualsofthemembersfromthetemplatefortheY
0.0 0.2 0.4 0.6 0.8 1.0 phase-bininthe4 10keVenergybandareshowninFig.4.This
Orbital Phase −
plotshowsa scattered distributionof the residuals,as expected
Fig.3. VelaX-1folded orbital light curves in4 10 (bluedots), and becauseof thehighvariabilityofthe source.However,theplot
10 20keV (red triangles) energy ranges. The−horizontal bars indi- alsoclearlyshowsthatresidualstendtoclustertowardnegative
cate−the phase-bin. The mid-eclipse time is taken as the phase zero. values.Therefore,weperformedanewanalysisandstackedonly
Thethreeorbitalphasebins(X,Y,andZ)usedforthephase-resolved those members that had high (> 30%) negative residuals (i.e.,
spectroscopy ofthedouble-peaked sampleareindicated(seetext).(A < 0.0345cnt/s)in the Y phase bins andat the same time low
colorversionofthisfigureisavailableintheonlinejournal). (<−30%) residuals in the X and Z phase bins. This allowed us
toobtainanewaverageorbitalprofileinthe4 10keVenergy
−
band,madeof12outof84totalorbitalcycles(i.e.,about15%
areobservedinthesofterbandlightcurve.Thisisexplainedas oftheanalyzedcycles).ThisisshowninFig.5.Usingadifferent
strongabsorptioncolumnmodulationalongthebinaryorbit.
definition of the residuals, where the ratio between the median
Successively,weconsideredthedifferencebetweenindivid- valuesistaken(insteadoftheirdifference),showstheverysame
ualorbitallightcurves(definedas the membersof the sample) trendasthatinFig.4,andleadstothesameselectedprofiles.
andtheaverageorbitalprofile(definedasthetemplate).Toob-
tain a smooth template of the two folded light curves, we in-
terpolatedwith a cubic spline functionbetween all the pairs of Since our normalization procedure could remove flux-
consecutivedatapoints(theknots)ofthelightcurves. dependentfeatures,wecarefullyverifiedthattheseselectedpro-
From our line of sight, the accretion wake effects are more fileswerenotassociatedwithaspecificluminositylevelbefore
prominent for about 0.2 in orbital phase (i.e., about 1.5days) normalization.In addition, for all of these 12 selected profiles,
around the inferior conjunction (see Fig. 1). To properly sam- thenullhypothesiswastestedagainstthetemplatewithnormal-
ple these effects, we defined three orbital phase bins that we ity tests (χ2, Kolmogorov-Smirnov,Shapiro-Wilk) and was re-
call X, Y, and Z, correspondingto phase intervals 0.15 0.35, jectedata significancelevelhigherthan3σ. Furthermore,they
−
0.35 0.55,and0.55 0.75,respectively(seeFig.3).Foranalo- all show a median value in the Y phase-bin that deviates at
− −
gousreasonsweimposedforeachmemberaminimumnumber least3σfromthetemplate.Theseprofilesarecharacterizedbya
Articlenumber,page3of10page.10
A&Aproofs:manuscriptno.velamalacaria_revised
0.06 0.20
4-10 keV
0.04 0.15
0.02 0.10
]
/s] 0.00 2m
esiduals[cnt −−00..0042 ate [ ph / s / c 000...100405 10-20 keV
R R 0.12
Y nt
−0.06 ou 0.10
C
0.08
−0.08 0.06
0.04
−0.10 0.02
55000 55200 55400 55600 55800 56000 56200 56400 56600 56800
0.00
Time[MJD] 0.0 0.2 0.4 0.6 0.8 1.0
Orbital Phase
Fig. 4. Residuals of single orbital profiles from the template in the
Fig.5.Averageorbitalprofilesofthedouble-peakedandofthestandard
4 10keV energy band for the Y phase-bin (see text). The residuals
− samplesin4 10keV(top)and10 20keV(bottom).Thebluedotsare
arecomparedtotheobservationtimeonthex-axis.Thethinblackline − −
thedatapointsobtainedbystackingthememeberswithlargeresiduals
isthezero-value,whilethereddashedlineisthe30%limitbelowwhich
intheYphasebin(seetext).Theverticalerrorbarscorrespondtothe
theselectioncriterionintheYphase-binisfullfilled.Errorbarscorre-
standarddeviationoftheaverageofthecountratesineachphasebin,
spondtothesamplestandarddeviation(σ/√N,withNnumberofdata
whilethehorizontalbarsindicatethephase-binwidth.Thereddashed
points). Not all the profiles below the red dashed line are selected as
linesrepresentsthefunctioninterpolatedbetweenthedatapointsofthe
double-peakedprofilesbecauseoftheconditionsimposedontheXand
standardsamplesineachenergyband.
Zphase-bins.
sity(N )alongtheentirelineofsight.Weassumedsolarabun-
double-peakshapeandafluxdroparoundtheinferiorconjunc- H
dancesby Anders & Grevesse (1989)for this. Using the metal
tion.Werefertothemasthedouble-peakedsample.
abundancesofKaperetal.(1993)forHD77581(i.e.,sub-solar
Similar results are obtained if the same procedure is per-
C, andsuper-solarN, Al) doesnotaffectourresults, whichare
formedinthe10 20keVenergyband.Inthiscase,theselected
− wellconsistentwithin1σ.Therefore,weusedsolarabundances
profilesare11outof84.Theoverallprofileshapeisverysimi-
throughouttherestoftheanalysis.WealsoaddedtwoironGaus-
lartotheprofileobtainedinthe4 10keVenergyband,witha
− sianlinecomponents(Odakaetal.2013),KαandKβat6.4and
double-peakmorphologyandadiparoundtheinferiorconjunc-
7.08keV.. The same components were included in the models
tion. The extracted sample in the 10 20keV energy band is
− usedforthephase-resolvedspectroscopy.Theresultsofthefits
showninFig.5.
are givenin Table 2. Both modelsreturnacceptablefits for the
Finally,weextractedanotherindependentsampleinthetwo
standardand thedouble-peakedsample.However,forthe stan-
energy bands by stacking all the statistically acceptable orbital
dardsample,thepartialcoveringmodeldescribesthedatanotas
profiles exceptfor the double-peakedmembersin each respec-
wellasthecutoffpower-lawmodel.
tiveenergyband.Werefertothissampleasthestandardsample.
Thestandardsampleorbitalprofilesinthetwoenergybandsare
shownwithdashedlinesinFig.5. 5. Phase-resolvedspectralanalysis
5.1.Orbitalphase-resolvedspectroscopyofthe
4. Phase-averagedspectralanalysis double-peakedsample
The orbital phase-averaged spectral analysis of VelaX-1 was Toinvestigatethenatureofthedipinthedouble-peakedsample,
performedseparatelyforthetwosamplesdefinedinSect.3.Be- weextractedX-rayspectraofVelaX-1intheX,YandZorbital
cause the physical problem is complex, a fully physical model phasebins.Forthis,wefilteredMAXI/GSCdatawithGTIscor-
of the X-ray production mechanism in accreting X-ray pulsars respondingtothethreephasebinsinthedouble-peakedprofiles
hasnotbeenimplementedsofar.Therefore,anempiricalmodel (seeSect.3).BasedontheresultsfromSect.4,weusedthesame
hastobeused.Followingthecommonempiricalmodelusedin modelsaswereappliedinthephase-averagedcasetofittheor-
theliterature(see,e.g.,Schanneetal.2007),we adoptedacut- bitalphase-resolvedspectra:(a) a cutoffpower-lawmodelplus
offpower-lawmodel(cutoffplinXSPEC)towhichablackbody a blackbody component, and (b) a partial covering model. To
component was added to account for soft excess (see Hickox avoidunphysicalbest-fitsolutionswhenmodel(a)wasused,we
et al. 2004 and references therein). On the other hand, to ac- fixedthecutoffenergyandtheblackbodytemperaturetotheirre-
countfortheexpectedinhomogeneousstructureoftheambient spectivephase-averagedvalues.Formodel(a)wealsocompared
windinVelaX-1,wealsotestedthepartialcoveringmodel(pc- theresultsforafixed(tothephase-averagedvalue,Γ=0.27,see
fabs*powerlaw in XSPEC) characterized by a local absorbing Table 2) or free photon index. We found that the model with
component (Npc) that affects only a fraction f of the original freephotonindexisstatisticallypreferable(∆χ2 >20).Further-
H
spectrum.ThismodelhasalsobeensuccessfullyusedtofitNuS- more, fitting the Y phase bin spectrum, the modelwith a fixed
tardataofVelaX-1(Fürstetal.2014).Tobothmodels,photo- photonindexleadstounphysicallyhighvaluesoftheblackbody
electric absorption by neutral matter (Morrison& McCammon normalization(tocompensateforthesoftexcess).Thus,wedid
1983) was added, to account for the absorption column den- notconsiderthecutoffpower-lawmodelwithafixedphotonin-
Articlenumber,page4of10page.10
C.Malacariaetal.:VelaX-1observedwithMAXI
0.8
0.6
φ2
0.5 Xbin 0.7 φ7
0.4
0.6 φ3 φ1
0.3
Γ 0.2 Zbin Γ 0.5 φ6
φ4
0.1
0.4
0.0
−0.1 Ybin 0.3 φ5
−0.2
2 4 6 8 10 12 14 0.2
2 4 6 8 10 12 14
NH [1022 cm−2] NH [1022 cm−2]
Fig. 6. χ2 contour plots for the VelaX-1 double-peaked sample fitted
withanabsorbedcutoffpower-lawmodelincludingablackbodycom- Fig. 7. χ2 contour plots for the VelaX-1 standard sample fitted with
ponent(modela,seeTable3).Theellipsesareχ2-contoursfortwopa- anabsorbedcutoffpower-lawmodelincludingablackbodycomponent
rameters(N andΓ).Thecontours correspond toχ2 +1.0 (thepro- (seeTable4).SymbolsarethesameasinFig.6.Therespectiveorbital
H min
jections of this contour to the parameter axes correspond to the 68% phasebinsareindicated.
uncertaintyforoneparameterofinterest),χ2 +2.3(68%uncertainty
min
fortwoparametersofinterest),andχ2 +4.61(90%uncertaintyfortwo
min
parametersofinterest).Therespectiveorbitalphasebinsareindicated. a blackbody component, and (b) a partial covering model. To
avoid unphysical best-fit solutions, we fixed the cutoff energy
and the blackbodytemperatureof model(a) to their respective
dexa viable modelin our analysis(similar considerationshold phase-averagedvalue.AsalreadymentionedinSect.5.1,model
forthephase-resolvedspectroscopyofthestandardsample,see (b)returnstwo statistically equivalentsolutions:one where the
Sect.5.2).Itisoftenpossibletofindastatisticallyequivalentso- partial covering component is statistically significant, and the
lutionformodel(b)wherethepartialcoveringcomponentisnot otherwherethesamecomponentisnotnecessary.Insuchcases,
necessary.Wetestedthesignificanceofthepartialcoveringcom- we adopted the solution in which the partial covering compo-
ponentwith the F-test, includingsuch a componentonly when nentwasnotrequired,unlessitwassignificantata99%c.l.The
it was significantat a 99%c.l. Otherwise, we preferredthe so- results are given in Table 4. All phase-resolved spectra return
lutionswherethepartialcoveringcomponentwasnotincluded. goodfitsforeachofthetwotestedmodels.However,thecutoff
Thesamecriterionwasadoptedforthestandardsamplecase(see power-lawmodelwithblackbodyshowsamarkedmodulationof
Sect.5.2) thespectralphotonindexΓalongtheorbitalphase.Thetrendof
Models(a)and(b)bothreturnagoodfit,andtheresultsare thismodulationissimilartothatobservedforthedouble-peaked
shown in Table 3. Model (a) shows orbital modulation of the sample(seeSect.5.1).Foradeeperinvestigationoftheseresults,
spectralphotonindexΓ.Tobetterexplorethespectralvariation, χ2-contourplotsofΓandNH wereproduced(seeFig.7).
we produced χ2-contour plots of Γ and N in the three phase Thepartialcoveringmodelshowsonlymarginalvariationof
H
bins(seeFig.6). the photon index Γ, while the column density clearly changes
Conversely, model (b) does not show photon index modu- along the orbital phase, and the covering fraction is consistent
lation along the orbital phase, although it shows variation of with a constant value. A plot of the standard sample phase-
thecolumndensity.Aplotofthedouble-peakedphase-resolved resolvedspectrafitwith thepartialcoveringmodelisshownin
spectrafitwiththepartialcoveringmodelisshowninFig.8. Fig.9.
5.2.Orbitalphase-resolvedspectroscopyofthestandard 6. Discussion
sample
Themainobservationalresultsofourworkcanbesummarized
Similarly to the double-peaked sample, we performed orbital asfollows:
phase-resolved spectroscopy of the standard sample. However,
forthissample,thestatisticsismuchhigherthanintheprevious – About 15% of the orbital light curves in VelaX-1 shows a
case, so that a finer binning is possible. We divided the stan- double-peakedorbitalprofile,withadiparoundtheinferior
dard sample profile into seven phase bins, each one wide 0.1 conjunction,bothinthe4 10keVand10 20keVenergy
− −
in phase, from 0.1 to 0.8. Phase bins around the eclipse time bands;
are notsuitable for our purpousesand suffer from lower statis- – orbitalphase-resolvedspectraofboththedouble-peakedand
tics,thereforetheywerenotanalyzedinourwork.Theanalyzed thestandardsamplearewellfitbyacutoffpower-lawmodel
phasebinsarelabeledφ,withi = 1,...,7.Thenwefilteredour andbyapartialcoveringmodel;
i
data with GTIs corresponding to the seven orbital phase bins – the cutoff power-law model leads to orbital modulation of
andextractedorbitalphase-resolvedspectraofthestandardsam- boththespectralphotonindexandthecolumndensity;
ple.BasedontheresultsfromSect.4,wefitthephase-resolved – apartialcoveringmodelfitsthedataequallywell.Noorbital
spectra with the same two models: (a) a cutoff power-lawwith modulationof the spectral photonindex is observedin this
Articlenumber,page5of10page.10
A&Aproofs:manuscriptno.velamalacaria_revised
Table2.Best-fitspectralparametersoftheorbitalphase-averagedspectrafromthedouble-peakedandstandardsamples.Theerrorscorrespondto
the1σ-uncertainties.
Model CutoffPL+BB Part.Cov.
Sample Double-peaked Standard Double-peaked Standard
N /1022cm 2 3.91+0.43 4.29+0.31 2.55+2.18 2.02+0.68
H − 0.36 0.30 2.55 0.79
N pc/1022cm 2 –− –− 14.9−+16.0 12.7−+2.5
H − 5.4 1.7
Cov.Fract.pc – – 0.63−+0.21 0.64+−0.07
0.18 0.06
Γ 0.27+0.04 0.48+0.06 1.09+−0.10 1.13+−0.03
0.03 0.06 0.07 0.03
E [keV] 15.3−+0.4 20.1−+2.5 –− –−
cut 3.4 2.0
kT [keV] 0.16+−0.04 0.14+−0.06 – –
BB 0.04 0.08
χ2/D.o.F.=χ2 285/311−=0.92 393/351−=1.12 318/311=1.02 480/352=1.36
red
Notes.(pc)PartialCoveringModel.
Table3.Best-fitspectralparametersofthephase-resolvedspectrafromthedouble-peakedsample.Theerrorscorrespondtothe1σ-uncertainties.
Phase-bins X Y Z
Model CutoffPL+BB Part.Cov. CutoffPL+BB Part.Cov. CutoffPL+BB Part.Cov.
N /1022cm 2 2.02+0.41 3.68+0.29 3.89+0.93 2.49+2.07 11.9+1.1 16+1
H − 0.31 2.16 0.87 2.49 1.0 3
N pc/1022cm 2 –− –− –− 16.1+−11.3 –− –−
H − 5.3
Cov.Fract.pc – – – 0.64+−0.17 – –
0.09
−
E [keV] 15.3(fixed) – 15.3(fixed) – 15.3(fixed) –
cut
Γ 0.44+0.04 1.12+0.02 0.02+0.06 0.91+0.29 0.29+0.06 1.10+0.05
0.03 0.04 − 0.06 0.11 0.06 0.05
Norm 0.110−+0.009 0.235−+0.014 0.003+−0.005 0.129−+0.151 0.089−+0.015 0.256−+0.042
PL 0.008 0.017 0.005 0.038 0.01 0.031
kT [keV]b 0.16(fi−xed) –− 0.16(fi−xed) –− 0.16(fi−xed) –−
BB
χ2/D.o.F.=χ2 113/121=0.93 121/120=1.00 113/108=1.05 117/107=1.09 101/102=0.99 113/101=1.11
red
Notes.(pc)PartialCoveringModel.(b)Blackbodycomponent.
case.Incontrast,orbitalmodulationofthecolumndensityis Thesampleofdouble-peakedorbitalprofiles,obtainedfrom
observed. the very long MAXI/GSC monitoring of Vela X-1, shows a re-
markabledip at phase Y, also in the 10 20keVband. This is
In the following we discuss our results within the context −
challengingtoexplainsolelybyinvokingabsorptionfromneu-
ofthemostrecentwindaccretionscenariosthathavebeenpro-
tralmatter.TheX-rayfluxobservedatthedipphaseisabouthalf
posedforVelaX-1.
ofthefluxinthestandardsampleatthesamephase(seeFig.3).
Tohalvealuminosityof 4 1036erg/sinthe10 20keVband,
6.1.Double-peakedorbitallightcurve a column density value ∼of N×H 2 1024cm−2 i−s needed, 10
∼ × ∼
timeshigherthanwhatisobservedinourstudy.
Indications of a double-peaked orbital light curve in Vela X-1
In addition,hydrodynamicalsimulationsof stellar windsin
havealreadybeenreportedbyearliestobservationsofthesource
HMXBs(Blondinetal.1991,Manousakisetal.2012)predicta
withArielV(Eadieetal.1975;Watson&Griffiths1977).Based
localpeakofthecolumndensityN aroundtheinferiorconjunc-
onTenmadata,Nagase(1989)foundanincreaseofthecolumn H
tionontheorderof1023cm 2.Thisagreeswithourobservations.
density along the NS orbit, in particularafter the transit phase. −
We also mention that the INTEGRAL analysis of Fürst et al.
Blondinetal.(1991)tookbothgravitationalandradiativeeffects
(2010)highlightsapossibledip-likefeaturearoundthetransitin
into accountand reproducedthe observedorbitalphase depen-
the phase-resolvedflux histogramsof VelaX-1 at 20 60keV,
denceofthecolumndensity.Theyfoundthatanaccretionwake
−
that is, at an energy band that is almost unaffected by absorp-
that extends toward the observer’s line of sight at the inferior
tion. Our conclusion is therefore that some additional mecha-
conjunctionproducesalocalpeakinthecolumndensityabsorp-
nism must be at work that contributes to the flux reduction at
tionalongthebinaryorbit.
the inferior conjunction,and that is triggered only in a limited
Intriguing evidence on the accretion wake in VelaX-1 was
sampleoforbitalcycles.
obtainedbySmith(2001),whoinferredthatahotsourcearound
thetransitphaseoftheNSispresentbyanalyzingHeλ1640ab- Acontributiontotheobservedreducedfluxcouldcomefrom
sorptionlinestrengths.BasedonChandraX-raydata,Watanabe theComptonscatteringatthelow-energylimit,wherethecross
etal.(2006)alsohaveinferredthepresenceofadensecloudob- sectionischaracterizedbyonlyaweakdependenceonthepho-
scuringthecompactsourceattheinferiorconjunction.Allthese tonenergyandcanbeapproximatedbytheThomsonscattering
results are tightly related to the observationalfeaturesreported crosssection(σ = 6.65 10 25cm2).ToefficientlyscatterX-
T −
×
inthiswork. raysfromtheaccretingpulsar,thesurroundingsofthecompact
Articlenumber,page6of10page.10
Table4.Best-fitspectralparametersoftheorbitalphase-resolvedspectrafromthestandardsample.Theerrorscorrespondtothe1σ-uncertainties.
Sample CutoffPL+BB(E =20.1keVfixed,kT b =0.14keVfixed)
cut BB
Phase-bin φ φ φ φ φ φ φ C
1 2 3 4 5 6 7 .
M
N /1022cm 2 4.86+0.55 2.57+0.31 1.85+0.35 3.24+0.48 6.25+0.68 11.5+0.9 11.9+1.0 a
H − 0.42 0.29 0.33 0.46 0.64 0.8 0.9 la
Γ 0.58−+0.03 0.72−+0.02 0.58−+0.03 0.49−+0.03 0.33−+0.04 0.55+−0.04 0.63+−0.05 ca
NormPL 0.111−+000...00030098 0.194−+000...00021009 0.126−+000...00030077 0.095−+000...00030076 0.066−+000...00030055 0.11−+000...000411 0.12−+000...000511 riaet
χ2/D.o.F.=χ2red 173/212−=0.82 233/247−=0.94 244/237−=1.03 216/228−=0.95 246/228−=1.08 216/205−=1.05 204/199−=1.03 al.:
V
Model PartialCovering ela
X
Phase-bin φ1 φ2 φ3 φ4 φ5 φ6 φ7 -1
o
NH/1022cm−2 6.3+10..30 1.7+11..27 0.+1.4 2.2+01..953 5.6+00..96 13.3+01..77 13.7+01..77 bse
N pc/1022cm 2 –− 7.1−+7.9 6.4+4.4 13−+8 32+−10 –− –− rv
H − 2.0 1.0 4 8 ed
Cov.Fract.pc – 0.49−+0.29 0.68−+0.02 0.51+−0.19 0.61−+0.05 – – w
Γ 1.09+0.03 1.26−+00..2045 1.13−+00..2055 1.12−+00..1008 1.23−+00..0138 1.08+0.04 1.15+0.04 ith
0.02 0.03 0.04 0.06 0.19 0.04 0.04 M
Nχ2o/rDm.oP.FL.=χ2 1920/2.2111−−+00=..00110.91 2500/.24407−+−00=..00641.01 2480/2.2366−+−00=..00321.05 2120/.22267−+−00=..00740.94 2510/2.3297−+−00=..21771.11 2370/2.2024−+−00=..00221.16 2190/1.2938−+−00=..00221.11 AXI
red
Notes.(b)Blackbodycomponent.(pc)PartialCoveringModel.
A
r
tic
le
n
u
m
b
e
r,
p
a
g
e
7
o
f
1
0
p
a
g
e
.1
0
A&Aproofs:manuscriptno.velamalacaria_revised
x1 0.01 x1
0.01
normalized counts s keV−1−1 10−3 x0.2 normalized counts s keV−1−1 1100−−43 xxxx0x0000....0.000125325
x0.01
x0.1 10−5
2 2
χ 0 χ 0
−2 −2
5 Energy (keV) 10 5 10
Energy (keV)
Fig.8.Orbitalphase-resolvedspectraofthedouble-peakedsamplefit-
Fig.9.Orbitalphase-resolvedspectraofthestandardsamplefitwitha
tedwithapartialcoveringmodel,whensignificant(seetext).Thetop
partialcoveringmodel,whensignificant(seetext).Thetoppanelshows
panelshowsfromtoptobottomdataandfoldedmodelofthethreespec-
fromtoptobottomdataandfoldedmodelofthespectrainphasebins
trainphasebinsX,Y,andZ.Thespectrahavebeenmultipliedbythe
from φ to φ . The bottom panel shows the residuals for the best-fit
factorshownaboveeachcurveontheleftsideforbettervisualization. 1 7
model.SymbolsarethesameasinFig.8.SeeTable4forallspectral
Thebottompanelshowstheresidualsforthebest-fitmodel.SeeTable3
parameters.
forallspectralparameters.
6.2.Spectralresults
6.2.1. Orbitalphase-averagedspectroscopy
objectneedtobeionizedanddenseenough.Hydrodynamical2D
and 3D simulations of the stellar wind in VelaX-1 by Mauche Orbital phase-averagedspectra in the 2 20keV energy range
−
et al. (2008) showed that the interaction of the wind with the of both the double-peakedand the standard sample were mod-
NSleadstostrongvariationsofthewindparameters.Theseau- eledwithacutoffpower-lawmodel,addingablackbodycompo-
thors found that the accretion wake reaches high temperatures nenttoaccountforthesoftexcess,andwithpartialcovering(Ta-
( 108K) and a high ionization parameter (ξ = L /(nR2 ) ble2).Forthedouble-peakedsample,bothmodelsreturnagood
1∼04ergcms 1, where n is the number density at axgiveniponoin∼t, fit. For the standard sample the partial covering model is only
−
while R is the distance from this point to the X-ray source). marginallyacceptable,whilethecutoffpower-lawdescribesthe
ion
The accretion wake extends to scales on the order of the op- datawell.
tical companion size (R 1012cm) with a number density of Asaresultoftheinhomogeneousenvironmentinwhichthe
109 12cm 3,which,taki∼nganaveragevalueofn 1010cm 3, NS travels along the binary orbit, our phase-averaged results
− − −
∼leads to a column density of 1023cm 2. This i∼s similar to are not straightforward in terms of physical interpretation. In-
−
the values obtained from our s∼pectral analysis of the double- stead, these results were used to avoid unphysicalbest-fitsolu-
peakedsampleintheYphasebin(seeSect.6.2.2andTable3). tions of the orbital phase-resolved spectra, fixing some of the
Moreover,thebremsstrahlungcoolingtimeisabout1011√T/ns modelparameterstotheirphase-averagedvalue(seeSect.6.2.2
(Rybicki & Lightman 1979) that is, 105s, or about one day, andSect.6.2.3).
∼
whichagreeswiththeobservedtimedurationofthedip.Inother
words, the gas stays ionized for a time long enough to allow 6.2.2. Double-peakedsamplephase-resolvedspectroscopy
Thomsonscattering. The ionizedgas is nottraced by the spec-
tralabsorptionin the softX-rayband.Nonetheless,its electron Fitting the phase-resolved spectra of the double-peaked sam-
density n is high enough to result in an optical depth value of plewith a cutoffpower-lawresultsin orbitalmodulationofthe
τ=nlσ =0.5 0.7(wherel=2 1012cmisthelengthofthe photon index (see Table 3), which has never been observed in
T
accretionwake).−Thus,whentheio×nizedstreamoftheaccretion VelaX-1 to the best of our knowledge.The spectrum becomes
wakepassesthroughtheobserver’sdirection,thatis,duringthe harder at the inferior conjunction (Y phase bin) than at earlier
transit,asubstantialpartoftheX-rayradiationcanbeefficiently and later phases (X and Z phase bins, respectively). Contour
scattered out of our line of sight. This scenario is further sup- plotsforNH andΓ(showninFig.6)appearclearlydistinctfrom
portedbythenormalizationofthepower-lawcomponent,which one phase bin to another, indicating that the modulation is not
appearstobemodulatedalongtheorbit(seeTable3),indicating duetotheintrinsiccouplingbetweenthetwoparameters.
anintrinsicallydimmerfluxattheinferiorconjunction(although Prat et al. (2008) reported photon index modulation along
largeparameteruncertaintiesneedtobeconsidered). thebinaryorbitofHMXBIGRJ19140+0951.Theseauthorsas-
cribedtheobservedorbitalmodulationtoavariablevisibilityof
Stellar-windvariabilitymaythenberesponsibleforthemod- theregionthatproducesthesoftexcess.Sincewedonotobserve
ifications of the accretion wake, which in turn modify the or- anymodulationoftheblackbodycomponent,ourstudydoesnot
bitallightcurveshape,alternatingbetweenadouble-peakedand supportthisinterpretation.Tothebestofourknowledge,wedid
a standardmorphology.We thereforeconcludethatthe ionized notfindanyreliablephysicalinterpretationforthisphenomenon.
gas in the hot accretion wake, combined with its high column Modeling with partial covering does not show an orbital
density, is a possible cause for the dip that is observed at the modulation of the photon index. This model results in absorp-
inferiorconjunctioninthedouble-peakedsamplelightcurves. tionincreasealongtheorbitalphase(seeSect.6.2.3foramore
Articlenumber,page8of10page.10
C.Malacariaetal.:VelaX-1observedwithMAXI
comprehensivediscussion of this result) and in a partialcover- High-resolution spectroscopic observations carried out at
ingcomponentwithhighcoveringfractionneededattheYphase critical orbital phases are necessary in future to distinguish
bin,whichisnotrequiredatearlierorlaterphases,however(see amongdifferentmodelsandconstrainstellarwindscenariosand
Table3).Apossibleinterpretationoftheseresultsisgivenbythe accretionmodalitiesinVelaX-1.
oscillatingbehavioroftheaccretionwake(Blondinetal.1990).
Becauseoftheorbit-to-orbitvariationsoftheambientwind,the
6.2.3. Standardsamplephase-resolvedspectroscopy
tail of the accretion wake is expected to oscillate with respect
toourlineofsight.Averagingthisbehavioroverseveralorbital Modeling phase-resolved spectra of the standard sample with
periodswouldresultinalargecoveringfraction,wherethewob- a cutoff power-law leads to a hardening of the spectra around
blingaccretionwakesimulatesaninhomogeneousabsorbingen- theinferiorconjunctionofabout∆Γ 0.35.Aswepointedoutin
vironment. Sect.6.2.2,aphysicalmechanismfor∼suchanorbitaldependence
A different interpretation might be that an intrinsically in- ofthephotonindexisnotknownandisdifficulttoexplain.
homogeneous structure is present around the inferior conjunc- On the other hand, fitting the spectra with a partial cover-
tionandpartiallycoverstheX-raysource.Inhomogeneousstel- ingmodelshowsthattheneutralmatter columndensity N in-
H
larwindsfromsupergiantstars(so-calledclumps)areexpected creasestowardlaterphases,varyingfrom 1022 aftertheegress
(Runacres&Owocki2005)andobserved(Goldsteinetal.2004, (φ ),to 1023cm 2beforetheingress(φ ).∼Thisagreeswithear-
1 − 7
and references therein), and have been suggested as the cause lier stud∼ies of VelaX-1 (see, e.g., Nagase 1989), and a similar
for VelaX-1 flaring activity (see Kreykenbohmet al. 2008and phenomenonhasbeenobservedinotherHMXBsaswell,forin-
Martínez-Núñezetal.2014).Thetimescaleonwhichclumpshit stance,in4U1700-37(Haberletal.1989).Basedontheanalysis
theX-raypulsarisontheorderofthewind’sdynamictimescale ofbothoptical(Carlberg1978;Kaperetal.1994)andX-raydata
(Feldmeieretal.1996),ithasbeenshownthattheaccretionwake
0.5R 1011cm
B 1000s (2) is notadequateto explainthe later phase (φ > 0.5)absorption,
v ∼ 1100km/s ∼ andthatamoreextendedphotoionizationwakeisnecessary.
∞
where 0.5R is the average separation between clumps Whenthismodelisapplied,thespectralphotonindexmod-
B
(Sundqvist et al. 2010). Thus, if we consider the X-ray source ulation along the orbital phase is only marginal. Based on the
on longer timescales, the resulting effect is a partially covered best-fitresults,wecanadopthereasimilarinterpretationasfor
spectrum. However, if the partial covering is due to the inho- the double-peaked sample results, that is, an inhomogeneous
mogeneousstructureoftheambientwind,nomodulationofthis ambient medium that produces strong absorption of the X-ray
componentwith orbital phase would occur. In contrast, our re- emissionandwhoseabsorptionefficiencydependsontheorbital
sults underline the necessity of partial coveringonly for the Y phase(see Sect.6.2.2).Thecolumndensityshowsa localpeak
phase bin spectra, where the obscuration effect of the accre- of NHpc∼3 × 1023cm−2 around the inferior conjunction, similar
tionwakeismaximized.Hydrodynamicalsimulations(Blondin towhatisobservedforthedouble-peakedsample,strengthening
etal.1990,1991)revealedthattheaccretionwakeiscomposed the connectionwith the accretion wake. Moreover,this sample
ofcompressedfilamentsofgasandrarefiedbubbles.Moreover, needsthepartialcoveringcomponentalsotofitthespectrumof
recent models also showed that density perturbations in stellar thephasebinφ2,whichislocatedaroundtheNSgreatesteastern
windsofOBstars onlydevelopatdistanceslargerthanthe NS elongation(seeFig.1).Regardlessofwhetherthepartialcover-
orbitalradiusinVelaX-1(r 1.1R ,seeFig.2inSundqvist& ing is dueto a wobblingor an inhomogeneousaccretionwake,
⋆
Owocki 2013). Therefore,in∼trinsic stellar wind clumps around these results indicate at a wake that affects the observations at
the X-ray pulsar could be not structured enough to trigger the earlierphasesthantheinferiorconjunction.
observedhighX-rayfluxvariability.Instead,theinhomogeneous We therefore conclude that the absorbing (and likely inho-
structurecharacteristicoftheaccretionwaketrailingtheNSin- mogeneous)materialthatpartiallycoverstheX-raypulsarmight
dicatesthatthecompactobjectcouldbeabletoperturbtheambi- beduetoadifferentinfluenceoftheaccretionwakeatdifferent
entwindsuchthatclumpsareformedatitspassage.Thesphere orbitalphases.
of influence of the X-ray pulsar is dictated by its accretion ra-
dius,r 1011cm(Feldmeieretal.1996)andbytheStrömgren
sphereadcce∼finedbythephotoionizationparameterξ = L /(nR2 ), 7. Summary
x ion
whichleads(seeSect.6.1forparameterdefinitionsandvalues) We have presented the results of a very long MAXI observa-
to a photoionizationradiusRion 0.2RB 6 1011cm racc. tionofVelaX-1,whichallowedustoperformdetailedspectral
∼ ∼ × ∼
SincetheorbitalvelocityoftheNSisvorbit 280km/s,theper- (phase-averaged and phase-resolved, in 2 20keV) and light
turbationtimescaleisabout ≈ curve(in4 20keV)analysis.Ourresultsca−nbesummarizedas
−
R follows:
t = ion 2 104s 5.5h (3)
pert
v ≈ × ∼
orbit – About 15% of the orbital light curves in VelaX-1 shows a
which can be considered as an upper limit for the X-ray vari- double-peakedorbitalprofile,witha 1.5daydiparoundthe
∼
ability induced by accretion of locally formed clumps. This inferiorconjunctioninthe4 10keVand10 20keVenergy
− −
is consistent with observations which showed that flares in bands.
VelaX-1 last from 102s up to 104s (see Ducci et al. 2009; – Explaining the dip around the inferior conjunction by ab-
Martínez-Núñezetal.2014,andreferencestherein). sorption alone requirescolumn density values of about2
Even though the inhomogeneousstructure characteristic of 1024cm 2,whicharenotobservedinouranalysis. ×
−
the accretion wake might be a possible cause for the observed – Thedipinthedouble-peakedsamplemightbeproducedby
partialcoveringaroundtheinferiorconjunction,itcannotbedi- consideringcontributionfromThomsonscatteringbyanion-
rectly probed within the context of our MAXI analysis, which ized accretion wake that is elongated toward the observer
onlyallowsinferringthelong-termpropertiesofthesourceav- during the transit. Intrinsic variability of the stellar wind
eragedoverseveralorbitalphasebins. leads to strong variations of the accretion wake, including
Articlenumber,page9of10page.10
A&Aproofs:manuscriptno.velamalacaria_revised
its ionization degree,thus to alternatingdouble-peakedand Chodil,G.,Mark,H.,Rodrigues,R.,Seward,F.D.,&Swift,C.D.1967,ApJ,
standardprofiles. 150,57
– Whenfitwithacutoffpower-lawmodel,thedouble-peaked Doroshenko,V.,Santangelo,A.,Nakahira,S.,etal.2013,A&A,554,A37
Ducci,L.,Sidoli,L.,Mereghetti,S.,Paizis,A.,&Romano,P.2009,MNRAS,
and the standard samples both show spectral hardening
398,2152
aroundtheinferiorconjunction(by 0.3intermsofthepho-
Eadie,G.,Peacock,A.,Pounds,K.A.,etal.1975,MNRAS,172,35P
∼
ton index). We did not find any reliable physical interpre- Feldmeier,A.,Anzer,U.,Boerner,G.,&Nagase,F.1996,A&A,311,793
tation for this phenomenon, which likely indicates that the Fürst,F.,Kreykenbohm,I.,Pottschmidt,K.,etal.2010,A&A,519,A37
modelisinadequateforVelaX-1. Fürst,F.,Pottschmidt,K.,Wilms,J.,etal.2014,ApJ,780,133
Goldstein,G.,Huenemoerder,D.P.,&Blank,D.2004,AJ,127,2310
– Orbitalphotonindexmodulationisavoidedifapartialcover-
Haberl,F.,White,N.E.,&Kallman,T.R.1989,ApJ,343,409
ingcomponentisusedaroundtheinferiorconjunction.This Hickox,R.C.,Narayan,R.,&Kallman,T.R.2004,ApJ,614,881
component indicates a highly inhomogeneousenvironment Jackson,J.C.1975,MNRAS,172,483
and local column density peak values of 3 1023cm 2 Kaper,L.,Hammerschlag-Hensberge, G.,&vanLoon,J.T.1993,A&A,279,
−
∼ × 485
around the inferior conjunction (orbital phases around the
Kaper,L.,Hammerschlag-Hensberge,G.,&Zuiderwijk,E.J.1994,A&A,289,
eclipsewerenotanalyzed).
846
– Ourresultssuggesteitherawobblingoraninhomogeneous Kreykenbohm,I.,Wilms,J.,Kretschmar,P.,etal.2008,A&A,492,511
accretionwake as the source of the partialcoveringaround Manousakis,A.,Walter,R.,&Blondin,J.M.2012,A&A,547,A20
the inferior conjunction. Gravitational and radiative effects Martínez-Núñez,S.,Torrejón,J.M.,Kühnel,M.,etal.2014,A&A,563,A70
Matsuoka,M.,Kawasaki,K.,Ueno,S.,etal.2009,PASJ,61,999
from the X-ray pulsar may be responsible of density inho-
Mauche,C.W.,Liedahl,D.A.,Akiyama,S.,&Plewa,T.2008,inAmericanIn-
mogeneitiesin the ambient wind that successively feed the
stituteofPhysicsConferenceSeries,Vol.1054,AmericanInstituteofPhysics
observed high X-ray variability. Moreover, the absorption ConferenceSeries,ed.M.Axelsson,3–11
properties of the accretion wake seem to take place at ear- Mihara,T.,Nakajima,M.,Sugizaki,M.,etal.2011,PASJ,63,623
lierphasesthantheinferiorconjunction. Morrison,R.&McCammon,D.1983,ApJ,270,119
Nagase,F.1989,PASJ,41,1
– Someoftheanalyzedorbitalphase-binspectradonotneeda
Naik,S.,Mukherjee,U.,Paul,B.,&Choi,C.S.2009,AdvancesinSpaceRe-
partialcoveringcomponent.Therefore,anyspectralanalysis search,43,900
ofVelaX-1needstoconsidertheorbitalphaseatwhichthe Odaka,H.,Khangulyan,D.,Tanaka,Y.T.,etal.2013,ApJ,767,70
observationhasbeencarriedout. Prat,L.,Rodriguez,J.,Hannikainen,D.C.,&Shaw,S.E.2008,MNRAS,389,
301
Prinja,R.K.,Barlow,M.J.,&Howarth,I.D.1990,ApJ,361,607
Acknowledgements. Wegratefullyacknowledgetheanonymousrefereefornu- Rawls,M.L.,Orosz,J.A.,McClintock,J.E.,etal.2011,ApJ,730,25
merouscommentsthatgreatlyimprovedthemanuscript.Thisworkissupported Runacres,M.C.&Owocki,S.P.2005,A&A,429,323
bytheInternationalProgramAssociate(IPA)ofRIKEN,Japan,andbytheBun- Rybicki,G.B.&Lightman,A.P.1979,Radiativeprocessesinastrophysics(New
desministeriumfürWirtschaftundTechnologiethroughtheDeutschesZentrum York,Wiley-Interscience,1979.)
fürLuft-undRaumfahrte.V.(DLR)underthegrantnumberFKZ50OR1204. Schanne,S.,Götz,D.,Gérard,L.,etal.2007,inESASpecialPublication,Vol.
C.M.gratefullythankstheentireMAXIteamforthecollaborationandhospital- 622,ESASpecialPublication,479
ityinRIKEN.C.M.alsothanksV.DoroshenkoandL.Ducci(IAAT,Tübingen) Sidoli,L.,Paizis,A.,Fürst,F.,etal.2015,MNRAS,447,1299
forusefuldiscussions. Smith,M.A.2001,ApJ,562,998
Sugizaki,M.,Mihara,T.,Serino,M.,etal.2011,PASJ,63,635
Sundqvist,J.O.&Owocki,S.P.2013,MNRAS,428,1837
Sundqvist,J.O.,Puls,J.,&Feldmeier,A.2010,A&A,510,A11
References Tomida,H.,Tsunemi,H.,Kimura,M.,etal.2011,PASJ,63,397
vanKerkwijk,M.H.,vanParadijs,J.,Zuiderwijk,E.J.,etal.1995,A&A,303,
Anders,E.&Grevesse,N.1989,Geochim.Cosmochim.Acta,53,197 483
Blondin,J.M.,Kallman,T.R.,Fryxell,B.A.,&Taam,R.E.1990,ApJ,356, Watanabe,S.,Sako,M.,Ishida,M.,etal.2006,ApJ,651,421
591 Watson,M.G.&Griffiths,R.E.1977,MNRAS,178,513
Blondin,J.M.,Stevens,I.R.,&Kallman,T.R.1991,ApJ,371,684
Carlberg,R.G.1978,PhDThesis(TheUniversityofBritishColumbia(Canada))
Articlenumber,page10of10page.10