ebook img

Broadband spectroscopy of the eclipsing high mass X-ray binary 4U 1700-37 with Suzaku PDF

0.94 MB·English
Save to my drive
Quick download
Download
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview Broadband spectroscopy of the eclipsing high mass X-ray binary 4U 1700-37 with Suzaku

Mon.Not.R.Astron.Soc.000,000–000 (0000) Printed7January2015 (MNLATEXstylefilev2.2) Broadband spectroscopy of the eclipsing high mass X-ray binary 4U 1700-37 with Suzaku 5 Gaurava K. Jaisawal1⋆ and Sachindra Naik1† 1 0 1Astronomy and Astrophysics Division, Physical Research Laboratory, Navrangapura, Ahmedabad - 380009, Gujarat, India 2 n a J 6 ABSTRACT ] We presentthe results obtained frombroadbandspectroscopyof the high mass X-ray R binary 4U 1700-37 using data from a Suzaku observation in 2006 September 13-14 S covering 0.29-0.72 orbital phase range. The light curves showed significant and rapid . variation in source flux during entire observation. We did not find any signature of h p pulsations in the light curves. However, a quasi-periodic oscillation at ∼20 mHz was - detected in the power density spectrum of the source. The 1-70 keV spectrum was o fitted with various continuum models. However,we found that the partially absorbed r high energy cutoff power-law and Negative and Positive power-law with Exponential t s cutoff (NPEX) models described the source spectrum well. Iron emission lines at 6.4 a keV and 7.1 keV were detected in the source spectrum. An absorption like feature [ at ∼39 keV was detected in the residuals while fitting the data with NPEX model. 1 Considering the feature as cyclotronabsorptionline, the surface magnetic field of the v neutronstarwasestimated to be ∼3.4×1012 Gauss.To understandthe cause ofrapid 2 variation in the source flux, time-resolved spectroscopy was carried out by dividing 2 the observationinto 20narrowsegments.The results obtainedfromthe time-resolved 1 spectroscopy are interpreted as the accretion of inhomogeneously distributed matter 1 in the stellar wind of the supergiant companion star as the cause of observed flux 0 . variationin 4U1700-37.A sharpincreasein columndensity after ∼0.63orbitalphase 1 indicates the presence of an accretion wake that blocks the continuum and produces 0 the eclipse like low-flux segment. 5 1 Key words: Xrays:stars:binaries:eclipsing–neutron–stars:individual–4U1700- : v 37 – stars: individual – HD 153919 i X r a 1 INTRODUCTION opticalspectroscopicobservations,Clarketal.(2002)evalu- atedphysicalparametersoftheopticalcompanionandused 4U 1700-37 was discovered by Uhuru satellite in Decem- in Monte Carlo simulation to estimate the mass of the X- btieorns19w7i0th(JUonheusruetraelv.e1a9le7d3)t.hEexstyesntseimveafsollaonw-eucpliposbinsegrvbai-- raanydsMouorc∼e5a8n±d1o1pMti⊙ca.lTchomedpiasntaionncetoofbtehMebxin∼a2r.y44sy±s0t.e2m7wMa⊙s nary with an orbital period of 3.412 days. One of the most estimated to be1.9 kpc(Ankayet al. 2001). luminous and hottest optical star among the known high massX-raybinaries,asupergiantstar(HD153919) ofO6.5 Atentativedetectionofpulsationat∼67swasreported Iaf spectral type was identified as the optical companion from Tenma observations of 4U 1700-37 (Murakami et al. (Hutchings et al. 1973). Using BATSE data, the orbital 1984).However,laterobservationsdidnotconfirmthedetec- parameters of the binary system such as inclination i=66◦, tionofspinperiodinthesource.AlthoughdetectionofX-ray eccentricity e < 0.01, 48 < ax sin i <82 lt-sec and semi pulsationsarenotyetconfirmed,thespectrumof4U1700-37 eclipse angle θE=28◦.6 were derived (Rubin et al. 1996). has been well described by the standard continuum models Using Monte Carlo simulation, the mass of compact object of the accretion powered X-ray pulsars. Broad-band X-ray and mass and radius of the optical companion star were spectrumof4U1700-37obtainedfromvariousobservatories constrained at Mx ∼ 2.6 M⊙, Mo ∼ 30 M⊙ and Ro ∼ 18 such as HEAO 1, EXOSAT, Ginga, BeppoSAX had been R⊙, respectively (Rubin et al. 1996). Using ultraviolet and describedwithahighenergycutoffpower-lawmodel(White etal.1983;Haberletal.1989;Haberl&Day1992;Reynolds et al. 1999). A soft excess component was also detected in ⋆ [email protected] the spectrum during the eclipse and eclipse ingress obser- † [email protected] vations of 4U 1700-37 (Haberl et al. 1989; Haberl & Day 2 G. K. Jaisawal & S. Naik Orbital Phase 0.30 0.35 0.40 0.45 0.50 0.55 0.60 0.65 0.70 150 XIS−0 0.5−10 keV 100 e at R nt 50 u o C 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 60 PIN 10−70 keV e Rat 40 nt u o 20 C 0 PIN/XIS−0 1 o ati R s s e 0.5 n d ar H 0 0 2×104 4×104 6×104 8×104 105 1.2×105 Time (s) Figure 1. Light curves (top and middlepanels) andhardness ratio(bottom panel) obtained from the Suzaku observation of the high- massX-raybinary4U1700-37.DatafromXIS-0andHXD/PINdetectorsareplottedhere.Fluxvariabilitybyanorderof∼10-15canbe seenintopandmiddlepanelsofthefigure.Thequotednumbersatthetoppanelshowthedurationofsegmentsusedfortime-resolved spectroscopy. TheorbitalphasescoveredduringtheSuzakuobservationaremarkedatthetopofthefigure. 1992). 1991 April Ginga observation of the source showed plasma (Boroson et al. 2003). Many recombination lines as a clear difference in the temperature corresponding to the well as fluorescence emission lines were also detected in the softexcesscomponentbeforeandaftertheeclipse(0.47keV spectrum of XMM −Newton observation of 4U 1700-37 and0.74keV).Theriseinsoft excesstemperatureafterthe during eclipse, eclipse egress and low flux segments of the eclipse was explained on the basis of the bow shock forma- binary (van der Meer et al. 2005). Presence of recombina- tion in front of compact object (Haberl et al. 1994). tion lines from H and He atoms in eclipse phase suggested an extended ionization region around the source. As in the The nature of the compact object in 4U 1700-37 bi- caseofaccretion poweredX-raypulsars,severalmHzQPOs nary system is quite unclear . Due to observed hard X- werealsodetected inthepowerdensityspectraof4U1700- ray spectrum and non-detection of pulsation, Brown et al. 37, obtained from the Chadra observation (Boroson et al. (1996) suggested theX-ray source as a low mass black hole 2003). candidate. However, the 2-200 keV BeppoSAX spectrum of 4U 1700-37 was observed by Suzaku in 2006 Septem- 4U 1700-37 was found to be well fitted with a high energy berduringoutofeclipse phaseofthebinary.Wereport the cutoff power-law model representing the classical spectrum time-averagedandtime-resolvedbroadbandspectroscopyof ofanaccretionpoweredX-raypulsars(Reynoldsetal.1999). 4U1700-37tounderstandthenatureofthecontinuumemis- Apart from this, Reynolds et al. (1999) also reported the sion and its orbital dependency, flaring activities, emission possiblepresenceofcyclotronabsorptionfeatureat∼37keV lines and cyclotron features in thespectrum. in BeppoSAX spectrum. These results discarded the possi- bilityofablackholeasacompact object inthe4U1700-37 binary system. High resolution spectra from Chandra and XMM−Newtonobservatoriesweredescribedbyusingtwo 2 OBSERVATION AND ANALYSIS component absorption model as used in 1991 April Ginga observation (Boroson etal. 2003; vanderMeer etal. 2005). Suzaku is the fifth Japanese X-ray satellite which was The recombination lines from H and He like species and launched in 2005 July by Japan Aerospace Exploration fluorescence emission lines from neutral atoms were seen in Agency(Mitsudaetal.2007).Itcoversabroadenergyrange theChandraobservations duringintermittentflarestateof (from 0.2 keV to 600 keV) in theX-ray band with thehelp 4U1700-37.Thestrengthofthelineswerefoundtobevary- oftwosetsofdetectors,X-rayImagingSpectrometers(XIS; ingovertheobservation.Thedetectionoftripletstructurein Koyamaetal.2007)andHardX-rayDetectors(HXD;Taka- SiandMgindicatedthedisequilibriumofthephoto-ionized hashietal.2007).XISsareimagingCCDcamerasthatwork Broadband spectroscopy of 4U 1700-37 with Suzaku 3 in0.2-12keVrange.ThreeCCDcameras(XIS-0,XIS-2and 104 XIS-3) are front-illuminated whereas the other one (XIS-1) is back-illuminated. The effective areas of front-illuminated 1000 and back-illuminated XISs are 340 cm2 and 390 cm2 at 100 1.5keV,respectively.FieldofviewofXISis18′×18′ infull wofintdwoowinmstorudme.eHntXsDsucishaasnsoinli-ciomnagPiInNg ddieotdeecsto(rHcXoDns/iPstIiNng) Hz−1 10 and GSO crystal scintillators (HXD/GSO) working in 10- n) 2 1 a 70 keV and 40-600 keV ranges, respectively. Effective area me of the HXD/PINis 145 cm2 at 15 keVwhereas for GSO,it ms/ 0.1 is315cm2at100keV.FieldofviewofHXD/PINis34′×34′ (r 0.01 and is similar for HXD/GSO up to 100 keV. 4U1700-37wasobservedwithSuzakuin2006Septem- 10−3 ber 13-14. The observation was carried out during out of 10−4 eclipse phase of the binary covering 0.29-0.72 orbital phase range(considering mid-eclipsetimeasphasezero; Rubinet 10−4 10−3 0.01 0.1 1 Frequency (Hz) al.1996).Theobservationwasperformedin“XISnominal” positionwithaneffectiveexposureof∼81.5ksand∼82.1ks Figure2.Powerdensityspectrum(PDS)of4U1700-37obtained forXISandHXD,respectively.XISdetectorswereoperated from HXD/PIN light curve in 10-70 keV energy range. Absence ofpulsationsintherangeof10−5Hzto2Hzrangecanbeseenin in the “burst” clock mode with “1/4 window” option pro- thefigure.AQPOat∼20mHzisdetected inthePDSoftheX- viding 1 s time resolution during the observation. We used raysource.Thesolidlineinthefigurerepresentsthefittedmodel publicly available data (version 2.0.6.13) of the Suzaku ob- comprisingofapower-lawcontinuumandaGaussianfunctionat servation in the present work. HEASoft software package theQPOfrequency. (version 6.12) and calibration database (CALDB) released on 2012 February 10 (for XIS) and 2011 September 13 (for HXD)were used for the data analysis. in the PIN spectra as suggested by the instrument team. Unfiltered event files were processed by using Epoch2responsefile(20080129) forHXD/PINwasusedin ‘aepipeline’packageofFTOOLSalongwithstandardscreen- thespectralanalysis.DatafromallfourXISs(XIS-0,XIS-1, ingcriteriatocreatecleanedXISandPINeventfiles.These XIS-2 and XIS-3) and HXD/PIN were used in the present reprocessed clean event files were used in further analysis. study. “XSELECT”packageofFTOOLSwasusedtoextractlight curvesandspectrafromthereprocessedXISandHXD/PIN eventdata.Barycentriccorrectionwasappliedontherepro- 3 RESULTS cessed clean event files by using the task ‘aebarycen’. Atti- tude correction was applied to the reprocessed XIS event Source and background light curves in soft (XIS - 1 s time data files by using S-lang script aeattcor.sl1. The repro- resolution) and hard X-ray (HXD/PIN - 0.1 s time resolu- cessed XIS event data were checked for the possible pres- tion)energyrangeswereextractedasdescribedabove.Back- ence of photon pile-up. Photon pile-up in XIS event data groundsubtractedlightcurvesin0.5-10 keVand10-70keV was estimated by using S-lang script pile estimate.sl2 and rangesareshownintopandmiddlepanelsofFigure1.From found to be 10%, 11%, 11% and 11% at the center of the the figure, significant and rapid flux variability by a factor image obtained from XIS-0, XIS-1, XIS-2 and XIS-3 event of ∼10-15 can be seen in soft and hard X-ray bands. The data, respectively. An annulus region with inner and outer presence of flaring episodes along with stable low flux seg- radii of 30′′ and 180′′ from the source position was selected mentscan also beclearly seen in XISandPIN light curves. to reduce the photon pile-up to 64%. Source light curves Hardnessratio(ratiobetweenthelightcurvesobtainedfrom and spectra were extracted from the reprocessed XIS event HXD/PINandXIS-0eventdata)plot(bottompanelofFig- data by selecting above annulus region around the central ure1)wasgeneratedtocheckthespectralstateofthesource source. Background light curves and spectra were accumu- duringtheflaringepisodesaswellaslowfluxsegmentsinthe latedbyselectingcircularregionsawayfromthesource.The lightcurve.However,apartfrommarginalhardeningduring response and effective area files for all the XIS detectors theextendedlowfluxsegmenttowardstheendoftheobser- were generated byusingthetask ‘xisrmfgen’ and‘xissimar- vation,anysignificant changein thevalueofhardnessratio fgen’ of FTOOLS. Source light curve and spectra were cre- (spectral state) was absent. ated from reprocessed HXD/PIN event file by using “XS- To investigate the presence of any periodicity (pulsa- ELECT”. However, the HXD/PIN background light curves tion) in the X-ray source, power density spectrum (PDS) and spectrum were accumulated in a similar manner from wasgeneratedbyusingHXD/PINlightcurvewith0.1stime ‘tuned’ non X-ray background (NXB3) event file. A correc- resolution and shown in Figure 2.Absenceof any clear and tionforcosmicX-raybackground(CXB4)wasincorporated sharppeaksinthePDSin0.5sto105 srangesuggestedthe non-detection of pulsation in above time range. To confirm the non-detection of pulsation, we generated pulse profiles 1 http://space.mit.edu/ASC/software/suzaku/aeattcor.sl by assuming 50 s (corresponding to ∼20 mHz peak in the 2 http://space.mit.edu/ASC/software/suzaku/pile estimate.sl PDS) and earlier reported 67 s (from Tenma observation) 3 http://heasarc.nasa.gov/docs/suzaku/analysis/pinbgd.html asthespinperiodofthesource.Wedefinedpulsefractionas 4 http://heasarc.nasa.gov/docs/suzaku/analysis/pin cxb.html theratio between thedifference in the maximum and mini- 4 G. K. Jaisawal & S. Naik 10 For spectral analysis, the source and background spec- V−1 tra, response matrices and effective area files for all instru- nts s ke−1 0.11 minepnrtesvwioeuresgseecnteiroant.edAfbtyerfoalplopwroinpgritahteepbraocckegdruoruensddseuscbrtirbaecd- ou tion, spectra from all the detectors were fitted simultane- C d ously in 0.8-70 keVenergy range using XSPEC v12.7 pack- malize0.01 age. Due to the presence of known Si and Au edge feature or 10−3 in XIS spectra, data in 1.7-1.9 keV and 2.2-2.4 keV energy N 4 2 rangeswereignoredin thespectralfitting.TheXISspectra χ 0 werebinnedbyafactorof6from0.8to10keVwhereasthe −2 −4 PIN spectrum was binned by a factor of 2 up to 25 keV, a 4 2 factor of 4 from 25 keV to 50 keV and a factor of 6 from χ −20 50to70keV.Allthespectralparametersweretiedtogether −4 during the fitting except the relative normalization of de- 1 2 5 10 20 50 tectors which were kept free. Standard continuum models Energy (keV) for X-ray pulsars like high energy cutoff power-law (White, Figure 3. Energy spectrum of 4U 1700-37 obtained with the Swank & Holt 1983), Fermi Dirac cutoff power-law (FD- XISs and PIN detectors of the Suzaku observation, along with CUT; Tanaka 1986), NewHcut (a third order polynomial the best-fit model comprising a partial covering NPEX contin- uum model, three Gaussian functions for emission lines and a function with continuous derivatives; Burderi et al. 2000), cyclotron absorption component. The middle and bottom pan- cutoff power-law, NPEX (Makishima et al. 1999), Thermal els show the contributions ofthe residualsto χ2 foreach energy Comptonizationmodel(CompTT;Titarchuk1994)wereap- binforthepartialcoveringNPEXcontinuummodelwithoutand pliedinthespectralfitting.However,thehighenergycutoff withcyclotroncomponent inthemodel,respectively. power-law,NewHcutandNPEXmodelwithpartialcovering component described thesource source spectrum well. Additionofpartialcoveringcomponenttoallthreecon- 10 tinuum models improved the spectral fitting yielding the keV−1 1 reduced χ2 values from >4 to <2. In this model compo- nts s −1 0.1 ncoemntp,otnheenrte(aerqeuitvwaolendtiffheyrednrtogaebnsocroplutimonn cdoemnspitoyneanlotsn.gOthnee u Co source direction) absorbs the entire spectrum, whereas the d malize0.01 otothtehrecXom-rpaoynseonutr(cien)haobmsoogrbensetohueslsyoudricsterisbpuetcetdrummatptaerrtcialollsye. Nor 10−43 This model has been used to describe spectra of other 2 HMXBs (Jaisawal et al. 2013; Pradhan et al. 2014) and χ 0 −2 Be/X-ray binary pulsars which show the presence of sev- −4 4 eral absorption dips at various pulse phases during Type I 2 χ 0 outbursts(Naiketal.2011;Paul&Naik2011;Maitraetal. −2 2012; Naik et al. 2013 and references therein). Apart from −4 1 2 5 10 20 50 thesespectralcomponentsinthecontinuummodels,theiron Energy (keV) fluorescencelines at 6.42 keV(FeKα) and 7.1 keV(FeKβ) Figure 4. Energy spectrum of 4U 1700-37 obtained with the were detected in the spectrum of 4U 1700-37. An emission XISs and PIN detectors of the Suzaku observation, along with linelikefeatureat3.19keVwasseeninthespectralresidue thebest-fitmodelcomprisingapartialcoveringhighenergycutoff ofallthreecontinuummodels.AdditionofaGaussiancom- power-law continuum model, three Gaussian functions for emis- ponent at ∼3.19 keV to above three continuum models im- sionlinesandacyclotronabsorptioncomponent.Themiddleand proved the spectral fitting further. The line at ∼3.19 keV bottom panels show the contributions of the residuals to χ2 for was identified as the fluorescence emission from S XV as each energy bin for the partial covering power law continuum seen in EXO 2030+375 (Naik et al. 2013) or ArK . model without and with cyclotron component in the model, re- β spectively. In contrast to earlier findings, the thermal component (softX-rayexcess)wasnotseeninthespectrumduringthe Suzaku observation of 4U 1700-37. XMM-Newton observa- mum intensities to the sum of themaximum and minimum tionsofthesource, however,showedthepresenceof soft X- intensities in the pulse profile. We estimated pulse-fraction rayexcessoverthecontinuummodelin 0.22-0.31, 0.72-0.79 fromeachofthepulseprofilesobtainedbyassuming50sand and 0.07-0.17 orbital phase ranges. Apart from the eclipse 67sasspinperiodofthesourceandfoundtobe∼1%.The phasewhentwosoftX-rayexcesscomponentsweredetected, negligiblevaluesofpulsefractionindicatethenon-detection thesourceshowedasingleandweaksoftX-rayexcessduring ofX-raypulsationinthesource.Ontheotherhand,theob- otherorbitalphases(vanderMeeretal.2005). Thoughthe served weak and broad feature at ∼20 mHz confirmed the sourcewasobservedwithXMM-Newtonatfourepochs,only detectionofaquasi-periodicoscillation (QPO)intheX-ray one observation (in 0.48-0.59 phase range) overlaps partly source. The significance of QPO feature was determined by with the binary orbital phase covered during Suzaku obser- fitting the PDS with a power-law continuum along with a vation.Thenon-detectionofsoftX-rayexcessduringSuzaku Gaussian function at QPO frequency and found that the observationofthesource,therefore,canbeexplainedasdue detection was more than 3σ level. tothe relatively low sensitivity of Suzaku instrumentscom- Broadband spectroscopy of 4U 1700-37 with Suzaku 5 Table1.Best-fitparametersobtainedfromthespectralfittingofSuzakuobservationof4U1700-37with90%errors.Model-1:Partial covering NPEX model with Gaussian components, Model-2 : Partial covering NPEX model with Gaussian components and cyclotron absorption line, Model-3 : Partial covering high energy cutoff power-law model with Gaussian components, Model-4 : Partial covering highenergycutoffpower-lawmodelwithGaussiancomponentsandcyclotronabsorptionline,Model-5:PartialcoveringNewHcutmodel withGaussiancomponents andModel-6:PartialcoveringNewHcutmodelwithGaussiancomponents andcyclotronline. Parameter Value Model-1 Model-2 Model-3 Model-4 Model-5 Model-6 NH1a 1.9±0.1 2.0±0.1 2.2±0.3 2.2±0.1 2.2±0.1 2.2±0.1 NH2b 4.1±0.1 4.5±0.2 4.9±0.1 4.9±0.1 4.5±0.1 4.8±0.2 Cov.Fraction 0.6±0.1 0.6±0.1 0.7±0.1 0.7±0.1 0.7±0.1 0.7±0.1 Photonindex 0.2±0.1 0.3±0.1 1.0±0.1 1.0±0.1 0.9±0.1 0.9±0.1 Ecut (keV) 7.5±0.1 8.8±1.5 7.1±0.1 7.1±0.1 7.0±0.2 7.0±0.2 Efold (keV) – – 19.1±0.1 19.7±0.4 18.6±0.4 19.0±0.5 FeKα line Lineenergy(keV) 6.46±0.01 6.46±0.01 6.46±0.01 6.46±0.01 6.46±0.01 6.46±0.01 Eq.width(eV) 81±2 82±3 75±2 75±2 77±2 77±2 FeKβ line Lineenergy(keV) 7.13±0.01 7.13±0.01 7.15±0.01 7.15±0.01 7.14±0.01 7.14±0.01 Eq.width(eV) 21±2 22±2 14±1 14±1 20±2 20±2 Cyclotronline Lineenergy(keV) – 38.9±3.2 – 38.9∗ – 38.9∗ Width(keV) – 19.3+6.1 – 9.8+8.1 – 7.3+8.3 −4.3 −5.1 −5.5 Depth – 0.4±0.1 – 0.1±0.1 – 0.1±0.1 Fluxc (1-10keV) 2.1±0.1 2.1±0.1 2.1±0.1 2.1±0.1 2.1±0.1 2.1±0.1 Fluxc (10-70keV) 5.6±0.3 5.7±0.7 5.6±0.2 5.6±0.1 5.6±0.2 5.6±0.2 Norm.Const.d 1/1/1.04/0.98/1.02 −−−− 1/1/1.04/0.98/1.02 −−−− 1/1/1.04/0.98/1.01 −−−− χ2 (dofs) 1551(888) 1363(885) 1389(888) 1375(886) 1349(887) 1343(885) a :Equivalenthydrogencolumndensityinthesourcedirection(in1022 atoms cm−2 units), b :Additionalhydrogencolumndensity(1022 atomscm−2 units), c :in10−9 ergscm−2 s−1 unit. d:Quotedrelativenormalizationconstants areforXIS-0,XIS-1,XIS-2,XIS-3andPIN,respectively.Thevaluesremainsamewhile addingCRSFwiththerespectivecontinuum models(Model-1,Model-3andModel-5). ∗ :ThevalueswerefixedatthevalueobtainedfromspectralfittingwithModel-2. paredtothatofXMM-Newtontodetecttheweaksoftcom- outandwiththeadditionofcyclotronabsorptionlineinthe ponent in thespectrum. continuummodel, respectively. To check the statistical significance of the absorption feature, F-test routine of IDL, mpftest5, was applied on Apart from the soft X-ray excess, earlier reported ab- the χ2. As in case of 4U 1909+07 (Jaisawal et al. 2013), sorption like feature at ∼37 keV in the source spectrum probability of chance improvement (PCI) was evaluated by (Reynoldsetal.1999) wasalsomarginally seeninthespec- considering the χ2 without and with the cyclotron absorp- tral residuals of all three continuum models in the present tion component in continuum models (Press et al. 2007). work. However, the absorption feature was clearly detected The estimated PCI was found to be 3%, 46% and 49% af- when the source spectrum was fitted with partial covering ter adding the cyclotron component in the partial covering NPEX continuum model. Addition of cyclotron absorption NPEXmodel,thepartialcoveringhighenergycutoffpower- component to thepartial coveringNPEX continuum model lawmodelandthepartialcoveringNewHcutmodel,respec- improvedthespectralfittingyieldingbettervalueofreduced tively.AtsuchhighPCIvalues(46%and49%),thedetection χ2whichwasdecreasedfrom1.75to1.54.Energyandwidth of cyclotron absorption feature is statistically insignificant ofthecyclotronabsorptionfeaturewerefoundtobe∼39keV in spectral fitting with partial covering high energy cutoff and ∼19 keV, respectively. The cyclotron absorption com- power-law and partial covering NewHcut models. Though ponentwasaddedtothepartialcoveringhighenergycutoff thePCIvalueforpartialcoveringNPEXmodel(3%)suggest power-lawandNewHcutmodelsyieldingalmostidenticalre- the detection of cyclotron absorption feature in the source, sults. The parameters obtained from the spectral fitting of the broad width of the feature (∼19 keV) makes the de- Suzaku observation of4U 1700-37 are givenin Table1.The tection tentative. Considering the different values of PCI valuesofrelativeinstrumentnormalizationsoffourXISsand obtained for different models, the use of F-test in checking HXD/PINare also given in thetable and found tobecom- the statistical significance of the presence of cyclotron ab- parable to that obtained during the detector calibrations. sorptioncomponentinthespectrumof4U1700-37 (present The energy spectra of the source are shown in Figures 3 case) may not be reliable enough. Observations with high and 4 along with the best-fit models of partial covering NPEXmodelandpartialcoveringhighenergycutoffpower- law model, respectively. The middle and bottom panels in each figure show the residuals to the best-fit models with- 5 http://www.physics.wisc.edu/∼craigm/idl/down/mpftest.pro 6 G. K. Jaisawal & S. Naik sensitive hard X-ray detectors for long exposures can con- with continuumfluxwas interpretedas duetothepresence firmthepresence/absence ofcyclotron resonancescattering of precessing tilted accretion disk causing modification in feature in 4U 1700-37. thegeometryandvisibilityofironlineemittingregioninthe binary systems. To investigate the geometry in 4U 1700-37 binary system, we plotted the equivalent widths of 6.4 keV 3.1 Time-resolved spectroscopy and 7.1 keV iron lines with the observed column density in During the Suzaku observation of 4U 1700-37, observed Figure 7. The equivalent widths of both thelines are found sourcefluxwasfoundtobehighlyvariableatdifferenttime to be marginally variable with the column density. Similar scales. Several flare like episodes lasting for ∼10 ks and kindofvariationofequivalentwidth(belowof200eV)with low flux segments were seen in soft and hard X-ray light absorptioncolumndensity(orderof1022 cm−2)wasseenin curves (Figure 1). To probe the changes in spectral param- HerX-1andVelaX-1(Figure8ofMakishimaetal.1986).In eters during these flaring episodes and low flux segments such configuration, the X-ray source is expected to be sur- at such short intervals, we divided the entire observation round by inhomogeneously distributed absorbing material into20segmentsasmarkedinthetoppanelofFigure1.As that covers a fraction of radiation along theline of sight. mentionedearlier,sourcespectraforthese20segmentswere extracted from all four XISs and PIN detectors. For time- resolvedspectroscopy,weusedsamebackgroundspectraand 4 DISCUSSION responsematricesforcorrespondingdetectorsthatwereused for time-averaged spectroscopy. As all threemodels used in The continuum X-ray emission in the neutron star X-ray time-averagedspectroscopywellfittedthesourcespectrum, binaries is understood to be emitted through the inverse wechooseoneofthemodele.g.thehighenergycutoffpower- Comptonization of the photons originated from the mag- law model tofit thetime-resolved spectra. Iron Kα and Kβ netic poles as well as within the accretion column of the lines were detected in each of the 20 time-resolved spectra. neutron star (Becker & Wolff 2007). The emitted photons Thebest-fitspectralparameters(with90%errors)obtained interact with the materials present in the surroundings of from the simultaneous spectral fitting of each of the seg- the neutron star. The effect of interactions can be seen in ments are plotted in Figure 5 along with XIS-0 and PIN theobservedspectrumintheform ofphoto-electricabsorp- light curvesin left and right top panels, respectively. tion, soft excess emission, line emissions and cyclotron ab- TheequivalenthydrogencolumndensityNH wasfound sorptionfeatures.Detectionofsoftexcessandlineemissions to be high (7×1022 cm−2) in the beginning of observation. canreflectthereprocessingofhardX-rayphotonswithmat- However, the values of NH gradually decreased to a low ter (neutral or ionized plasma) in accretion disk, accretion value beyond which again showed gradual increase during column etc. The effect of the neutron star magnetic field theobservation.ThesystematicandsmoothvariationofNH can be seen in the X-ray spectrum through the interaction irrespective of source intensity during the Suzaku observa- of quantized electrons with the source photons. Depending tion suggested the orbital dependence of the matter distri- on thestrength of theneutronstar magneticfield,theelec- bution in 4U 1700-37. This can be confirmed with further trons are quantized in to Landau levels through the rela- longobservationsofthesourcewithupcomingobservatories tion Ea = 11.6B12(1+zg)−1 (keV), where zg is gravita- such as ASTROSAT. A sharp increase in the value of NH tional red-shift and B12 is magnetic field strength in units (13×1022 cm−2) wasobserved duringtheextendedlow flux of 1012 Gauss. The absorption like feature, cyclotron res- segmenttowardstheendoftheobservation(after∼0.63or- onance scattering feature (CRSF), is detected generally in bital phase). The power-law photon index was found to be 10-100keVspectrum.InthepresentcaseofhighmassX-ray variable e.g. higher during the low flux segments compared binary 4U 1700-37, we detected an absorption like feature to theflaring episodes. at∼39keV,aswasearlierreportedinBeppoSAXspectrum Flux of Iron Kα and Kβ emission lines were found to ofthesource(Reynoldsetal.1999).DuringtheBeppoSAX vary with the source flux whereas the corresponding equiv- observation, the cyclotron absorption feature was detected alent widths showed the opposite trend.Source flux as well whenthesourcespectrumwasfittedwithahighenergycut- asthefluxofironemissionlineswerefoundtobelowduring off power-law model.However,in thepresent study,thecy- theextendedlowfluxsegmentattheendoftheobservation. clotron absorption feature was detected when thespectrum The variation in the iron line parameters (flux and equiva- was fitted with a NPEX model. Though the detections of lent width) with the absorbed source flux in 8-70 keV are CRSF during BeppoSAX observation as well as Suzaku ob- plotted in Figure 6. Though the flux of both the emission servation did not provide any concluding results, the pres- lines increased along with the source flux, flux of Kα line enceofweakabsorption-likefeatureat∼39keVinboththe was found to increase faster compared to that of the K cases can not be entirely ruled out. Assuming the detected β line.However,theequivalentwidthofboththelinesshowed feature as CRSF in 4U 1700-37, the magnetic field of the nosystematicvariationwiththesourcefluxthoughtheval- X-ray source (neutron star) in the binary system was esti- ues were higher at low flux level. Dependence of emission matedtobe∼3.4×1012Gauss.Thepresenceofthecyclotron line flux and equivalent width with the source flux indi- featurecanbeconfirmedbyusingdatafromfutureobserva- cates the fluorescence origin of the lines from the matter tionswithlongexposureandgoodhardX-rayspectroscopic nearbytheneutronstar.Variationofironemissionlineflux instrumentssuch as NuSTAR, ASTROSAT & Astro−H andequivalentwidth withhardX-raycontinuumflux(8-70 missions. keV) in 4U 1700-37 are found to be similar to that found 4U 1700-37 was found to be significantly variable dur- in LMX X-4 and Her X-1 (Naik & Paul 2003). In case of ingtheSuzakuobservation.Suchkindoffluxvariabilityare LMC X-4 and Her X-1, the change in iron line parameters seen in other wind-accreting high mass X-ray binaries such Broadband spectroscopy of 4U 1700-37 with Suzaku 7 150 0.5−10 keV 100 10−70 keV CountRate10500 CountRate 50 0 0 NH10 units)22 11055 Eq. Width(eV)11055000 6.4 keV (1.2 80 1 h 60 7.1 keV PhotonIndex000...468 Eq. Widt(eV) 2400 8 8 Ecut(keV) 46 Line Flux10 units)−11 246 6.4 keV ( Efold(keV) 112050 Line Flux10 units)−11 01..5512 7.1 keV ( Flux0 units)−9 246 1−10 keV Flux0 units)−9 1120505 10−70 keV 1 1 ( ( 0 5×104 105 0 5×104 105 Time (s) Time (s) Figure 5.Spectral parameters obtainedfromthetime-resolvedspectroscopy forSuzakuobservation of4U1700-37. Thetoppanelsin boththesidesshowlightcurvesof4U1700-37in0.5-10keV(XIS-0)and10-70keV(HXD/PIN)energyranges.ThevaluesofNH,power- lawphoton index,cutoff (Ecut)andfoldingenergy(Efold)areshowninsecond, third,fourthandfifthpanels inleftside,respectively. The iron emission line parameters such as the equivalent widths and flux in for 6.4 keV and 7.1 keV iron emission lines are shown in second, third, fourth and fifth panels in right side, respectively. The source flux in 1-10 keV (left side) and 10-70 keV (right side) are showninbottom panels.Theerrorsshowninthefigureareestimatedfor90%confidence level. 6.4 keV 6.4 keV 7.1 keV 7.1 keV 6 150 m)−2 g s c−1 h (eV) 10 er−11 4 nt Widt 100 ux ( vale ne Fl 2 Equi 50 Li e F 0 0 5 10 15 20 0 5 10 15 20 Flux (10−9 erg s−1 cm−2) Flux (10−9 erg s−1 cm−2) Figure6.Changeinthe6.4keVand7.1keVironemissionlineflux(leftpanel)andequivalentwidths(rightpanel)withrespecttothe estimatedfluxin8-70keVenergyrange. 8 G. K. Jaisawal & S. Naik withthereversalofstorageaccretionflowthatincreasesthe 6.4 keV mass accretion rate. The flow reversal occurs in the range 7.1 keV of few hours that generates the flares of 15 m to 1 hr time 150 scales as seen in several segments (such as 3, 4, 7, 10 etc.) of present Suzakuobservation. The“flip-flop instability” in V) theaccretiondiskcanpossiblyalsoexplainshorttimeflaring e h ( activities as observed in 4U 1700-37 (Matsuda et al. 1991). Widt 100 However, in an alternate scenario, the hydrodynamics sim- nt ulationresultsforwind-fedsourcesshowedtheformationof e al non-steadyaccretion wakeconsisting thedensefilamentsof v ui compressed gas where thedensity reaches∼100 times more q E 50 comparedtoundisturbedstellarwind(Blondinetal.1990). Accretion of these filaments with fluctuating density may generatetheabruptvariationintheX-rayluminosityasob- served in thepresent case. 0 5 10 Suzaku observation of 4U 1700-37 was taken during N (1022 cm−2) out of eclipse of binary. However, an eclipse-like low flux H segment was observed towards the end of the observation Figure7.Changeinthe6.4keVand7.1keVironlineequivalent in0.63-0.73orbitalphaserange.Duringthissegment,anin- widths with respect to the column density during the Suzaku creaseincolumndensitywasfound.Thesourcefluxandline observation. flux of both the iron emission lines were decreased to min- imum values compared to the rest of the observation. The as Vela X-1 (Kreykenbohm et al. 2008), Cen X-3 (Naik et presence of dense matter in this orbital phase range can be al. 2011), 4U 1907+09 (River et al. 2010) etc. The source thepossiblereasonfortheeclipselikesegments.Suchtypeof luminosity in these wind-fed systems depends on the den- eclipse-like segments (quiescence period) was also observed sity and velocity of stellar wind as Lx ∝ ρv−3 (Bondi & during Chandra observation of 4U 1700-37 around ∼0.68 Hoyle 1944). Any fluctuations either in density or velocity orbitalphase(Borosonetal.2003).Asignificantincreasein can produce the variation in luminosity. Flux variability in thecolumndensityafterphase0.5wasalso reportedearlier time scales of kilo-seconds as seen in 4U 1700-37, were seen duringtheCopernicusobservationof4U1700-37(Masonet in Vela X-1and was explained on thebasis of clumpywind al. 1976). During EXOSAT observation, an increasing col- with the fluctuating density causing variation in the accre- umn density was also noticed after 0.6 orbital phase of the tion rate (Kreykenbohm et al. 2008; Odaka et al. 2013). binary (Haberl et al. 1989). The sharp increase in the col- Presence of low and high flux levels were also seen in the umndensityataboveorbitalphaserangecanbeinterpreted pulsar4U1907+09duringSuzakuobservationsin2006and as the formation of accretion wake as observed in Vela X- 2007 (Rivers et al. 2010). Though the low flux levels were 1 (Blondin et al. 1990). Haberl et al. (1994) also reported presentinthelightcurvesofboththeobservations,theone the presence of accretion wake based on the temperature presentin2006observationwasconsistentwithearlierobser- difference observed in the soft excess component. However, vations of 4U 1907+09 and interpreted as due to change in the spectroscopic evidences confirmed the formation of ac- whole continuum rather than obscuration/absorption of X- cretionwakeatlateorbitalphasesofbinarythatblocksthe raysduetothepresenceofadditionalmatterasinlatercase continuumandproducestheeclipselikesegments(Kaperet (Rivers et al. 2010). The argument of presence of clump of al. 1994) which is seen in thepresent case of 4U 1700-37. mattercausinglowfluxlevelsinthelightcurvesduring2007 In summary, the observed flux variability at ks time observationwassupportedbytheenhancementinthevalue scale during Suzaku observation of 4U 1700-37 can be ex- of absorption column density along the line of sight. How- plained on clumpiness in the stellar wind which may cause ever,inthepresentstudy,thevariationofNH wasmarginal fluctuation in mass accretion rate. However, the instability duringflareandnon-flaredurations.Significantlyhighvalue in temporary disk or flip flop instability in accretion disk of NH after orbital phase∼0.6 compared torest of theseg- canalsoproducetheflaresonshorttimeintervalsasseenin mentsoftheSuzakuobservation confirmedthefindingfrom light curves between 0.29-0.63 orbital phase. The extended EXOSATobservationofthesource.ThishighNH valueseg- lowfluxeclipse-likesegmentobservedtowardstheendofthe ment(beyondorbitalphaseof∼0.6) wasinterpretedasdue observationisinterpretedasduetothepresenceofaccretion to the passage of the accretion wake between the neutron wake. star and the observer. As the neutron star moves away, the accretion wake which trails behind the neutron star during the whole orbit, crosses the line of sight of the observer at ACKNOWLEDGMENTS ∼0.6 orbital phase yielding significantly high value of ab- sorption column density. Theauthorswouldliketothanktherefereeforhis/hercon- Taam&Fryxell(1989)performedthesimulationtoun- structivecommentsandsuggestionsthatimprovedthecon- derstandtheinteractionbetweenasymmetricaccretionflow tents of the paper. The research work at Physical Research from OB stars onto the neutron stars. The results showed Laboratory is funded by the Department of Space, Gov- thatatemporarydiskcan beformed duringtheinteraction ernment of India. The authors would like to thank all the of accretion flow with the shock in accretion wake region. members of the Suzaku for their contributions in the in- The destruction of temporary accretion disk is associated strument preparation, spacecraft operation, software devel- Broadband spectroscopy of 4U 1700-37 with Suzaku 9 opment,andin-orbitinstrumentalcalibration.Thisresearch TitarchukL.,1994,ApJ,434,313 has made use of data obtained through HEASARC Online vanderMeerA.,KaperL.,diSalvoT.,M`endezM.,vanderKlis Service,providedbytheNASA/GSFC,insupportofNASA M.,etal.2005,A&A,432,999 High Energy Astrophysics Programs. WhiteN.E.,SwankJ.H.&HoltS.S.,1983,ApJ,270,711 REFERENCES Ankay A., Kaper L., deBruijneJ. H. J., et al.2001, A&A,370, 170 BeckerP.A.&WolffM.T.,2007,ApJ,654,435 BlondinJ.M.,KallmanT.R.,FryxellB.A.&TaamR.E.1990, ApJ,356,591 BondiH.&HoyleF.,1944,MNRAS,104,273 BorosonB.,VrtilekS.D.,KallmanT.&CorcoranM.2003,ApJ, 592,516 BrownG.E.,WeingartnerJ.C.&WijersR.A.M.J.,1996,ApJ, 463,297 BurderiL.,DiSalvoT.,RobbaN.R.,LaBarberaA.,Guainazzi M.,2000, ApJ,530,429 Clark J. S., Goodwin S. P., Crowther P. A., et al. 2002, A&A, 392,909 HaberlF.,WhiteN.E.&KallmanT.R.,1989,ApJ,343,409 HaberlF.&DayC.S.R.,1992,A&A,263,241 HaberlF.,AokiT.&MavromatakisF.,1994,A&A,288,796 Hutchings J. B., Thackeray A. D., Webster B. L. & Andrews P. J.1973, MNRAS,163,13P JaisawalG.K.,NaikS.,PaulB.2013,ApJ,779,54 JonesC.,FormanW.,TananbaumH.,etal.1973,ApJ,181,L43 Kaper L., Hammerschlag-Hensberge G., Zuiderwijk E. J., 1994, A&A,289,846 KoyamaK.etal.,2007,PASJ,59,23 Kreykenbohm I., Wilms J., Kretschmar P., Torrej´on J. M., PottschmidtK.,etal.2008, A&A,492,511 MakishimaK.1986,ThePhysicsofAccretionontoCompactOb- jects,ed.K.P.Mason,M.G.Watson&N.E.White(Berlin: Springer),266,249 MakishimaK.,MiharaT.,NagaseF.,TanakaY.,1999,ApJ,525, 978 MaitraC.,PaulB.,NaikS.,2012,MNRAS,420,2307 MasonK.O.,BranduardiG.,SanfordP.,1976, ApJ,203,29 MatsudaT.,SekinoN.,SawadaK.,etal.1991,A&A,248,301 MitsudaK.etal.,2007,PASJ,59,1 MurakamiT.,KawaiN.,MakishimaK.,MitaniK.,HayakawaS., etal.1984, PASJ,36,691 NaikS.,PaulB.,2003,A&A,401,265 NaikS.,PaulB.,AliZ.2011,ApJ,737,79 NaikS.,MaitraC.,JaisawalG.K.,PaulB.,2013,ApJ,764,158 OdakaH.,KhangulyanD.,TanakaY.T.,WatanabeS.,Takahashi T.,MakishimaK.,2013, ApJ,767,70 PaulB.,NaikS.,2011,BASI,39,429 PradhanP.,MaitraC.,PaulB.,IslamN.,PaulB.C.,2014,MN- RAS,442,2691 Press W. H., Teukolsky S. A., Vetterling W. T., & Flannery B. P.,2007,NumericalRecipes:TheArtofScientificComputing (Cambridge:CambridgeUniv.Press),730 ReynoldsA.P.,OwensA.,KaperL.,ParmarA.N.&SegretoA., 1999,A&A,349,873 Rivers,E.,Markowitz,A.,Pottschmidt,K.,etal.2010,ApJ,709, 179 RubinB.C.,FingerM.H.,HarmonB.A.,etal.1996,ApJ,459, 259 Takahashi T.etal.,2007,PASJ,59,35 TaamR.E.&FryxellB.A.,1989, ApJ,339,297 TanakaY.1986,inProc.IAUColloq.89,Vol.255,RadiationHy- drodynamics in Stars and Compact Objects, ed. D. Mihalas &K.-H.A.Winkler

See more

The list of books you might like

Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.