Astronomy&Astrophysicsmanuscriptno.IGRJ17544˙pulsations (cid:13)c ESO2012 January12,2012 X-ray pulsations from the region of the Supergiant Fast X-ray Transient IGR J17544−2619 S. P. Drave∗1,A. J. Bird1,L. J. Townsend1,A. B. Hill1,V. A. McBride1,2,3,V. Sguera4,5,A. Bazzano5,andD. J. Clark6,7 1 SchoolofPhysicsandAstronomy,UniversityofSouthampton,UniversityRoad,Southampton,SO171BJ,UK 2 Astronomy,GravityandCosmologyCentre,DepartmentofAstronomy,UniversityofCapeTown,Rondebosch,7701,SouthAfrica 3 SouthAfricanAstronomicalObservatory,POBox9,Observatory,7935,SouthAfrica 4 INAF-IASF,IstitutodiAstrofisicaSpazialeeFisicaCosmica,ViaGobetti101,Bologna,Italy 5 INAF-IASF,IstitutodiAstrofisicaSpazialeeFisicaCosmica,ViadelFossodelCavaliere100,00133Roma,Italy 6 Centred’EtudeSpatialedesRayonnements,CNRS/UPS,BP4346,31028Toulouse,France 2 7 CREATEC,Unit8,DerwentMillCommercialPark,Cockermouth,Cumbria,CA130HT,UK 1 0 Recieved−25/08/2011/Accepted−06/01/2012 2 ABSTRACT n a J Phase-targetedRXTEobservationshaveallowedustodetectatransient71.49±0.02ssignalthatismostlikelytobeoriginatingfrom thesupergiantfastX-raytransientIGRJ17544−2619.Thephase-foldedlightcurveshowsapossibledouble-peakedstructurewitha 1 pulsedfluxof∼4.8×10−12ergcm−2s−1(3−10keV).Assumingthesignaltoindicatethespinperiodoftheneutronstarinthesystem, 1 theprovisionallocationofIGRJ17544−2619ontheCorbetdiagramplacesthesystemwithintheclassicalwind-fedsupergiantXRB region.SucharesultillustratesthegrowingtrendofsupergiantfastX-raytransientstospanacrossbothoftheoriginalclassesof ] E HMXBinPorb−Pspinspace. H Keywords.X-rays:binaries-X-rays:individual:IGRJ17544−2619-stars:winds,outflows-stars:pulsars . h p 1. Introduction long respectively, were observed on the same day, indicating - o fastandrecurrenttransientbehaviour.Asubsequentdetectionon r Supergiant Fast X-ray Transients (SFXTs) are a new class of 2004 March 8 (Grebenev et al. 2004) further illustrated the re- st high mass X-ray binary (HMXB) system that have been un- currentnatureoftheX-rayoutburstsinIGRJ17544−2619.After a veiled over the lifetime of the INTEGRAL mission (Winkler the INTEGRAL detection, IGR J17544−2619 was associated [ etal.2003).ThesesourcesarecharacterisedbyrapidX-rayout- withthesoftX-raysource1RXSJ175428.3-262035(Wijnands bursts, with durations of the order of tens of minutes to tens of 1 2003, Voges et al. 2000) and discovered in archival Beppo- v hours (Sguera et al. 2005), and an association with OB super- SAX (in’t Zand et al. 2004) and XMM-Newton data (Gonza´lez- 4 giant companion stars (Negueruela et al. 2006b). To date there Riestra et al. 2004). A Chandra observation (in’t Zand 2005) 8 are10confirmedSFXTsclusteredalongtheGalacticplane(see precisely located the source with a positional accuracy of 0.6” 2 individual papers for details: Pellizza et al. 2006, Negueruela (RA=17:54:25.284,DEC=-26:19:52.62,J2000.0),confirming 2 et al. 2006a, Romano et al. 2007, Masetti et al. 2008, Nespoli theassociationofIGRJ17544−2619with2MASSJ17542527- 1. etal.2008,Rahoui&Chaty2008,ZuritaHeras&Walter2009, 2619526. Pellizza et al. (2006) classified the companion as an 0 Romano et al. 2009). There are also several candidate SFXTs O9Ib star with a mass of 25−28M at a distance of 2−4kpc. (cid:12) 2 which exhibit the same X-ray flaring behaviour but for which SubsequentlyRahouietal.(2008)performedSEDfittingtothe 1 an optical/IR counterpart is yet to be determined (e.g. Sguera mid-IRspectrumandrefinedthedistanceestimatetothesystem v: et al. 2006). With the determination of source distances from as∼3.6kpc. optical/IR spectroscopy (Rahoui et al. 2008), peak outburst lu- Xi minositiesof1036 −1037ergs−1 havebeendeduced(Grebenev Using long baseline IBIS/ISGRI light curves, Clark et al. (2009) identified the orbital period of IGR J17544−2619 as r etal.2004).Combinedwiththeobservationofquiescencestates a at 1032erg s−1 (Bozzo et al. 2008, in’t Zand 2005) this illus- 4.926±0.001d,oneoftheshortestorbitalperiodsobservedin tratesaveryhighX-raydynamicrangeof104−105 withinthese anSFXT. systems.X-raypulsationshavebeendetectedinfourconfirmed ThespectralpropertiesofIGRJ17544−2619havebeenstud- SFXTs implying there are accreting neutron stars in these sys- iedatalllevelsofemission.Theoutburstspectraareoftenwell tems (for individual sources see: Lutovinov et al. 2005, Sguera fit with powerlaw models that show variations in column den- etal.2007,Sidolietal.2007a,Sidolietal.2008). sity, 1.1−3.3×1022cm−2, and photon index, 0.75−1.3, between IGRJ17544−2619wasfirstdiscoveredasahardX-raytran- outbursts (Rampy et al. 2009, Sidoli et al. 2009, Romano et al. sient source on 2003 September 17 (Sunyaev et al. 2003) with 2008). The quiescence spectra are far softer than those of out- theIBIS/ISGRI(Ubertinietal.2003/Lebrunetal.2003)instru- bursts with photon indices between 2.1 and 5.9 (Sidoli et al. ment aboard INTEGRAL. Two short outbursts, 2 and 8 hours 2008, in’t Zand 2005). The softness of these spectra has led to the conclusion that the compact object in IGR J17544−2619 is ∗ [email protected] mostlikelyaneutronstar(in’tZand2005,Pellizzaetal.2006). 1 Draveetal.:X-raypulsationsfromtheregionoftheSFXTIGRJ17544−2619 In this paper we outline a new study performed on the IGR 10 J17544−2619 system. Our data set is described in Sect. 2, fol- lowed by temporal and spectral results in Sect. 3. A discussion of these results is given in Sect. 4 followed by conclusions in 8 Sect.5. !1CU 6 P Obs 1 Obs 2 Obs 3 2U.siDngattaheSeortbaitnaldeAphneamlyesriissof Clark et al. (2009) three ob- !1counts sec 4 Phase 0.412 Phase 0.720 Phase 0.811 servations of the region around IGR J17544−2619 were per- formed through one half of the compact object orbit using the 2 Proportional Counter Array (PCA) instrument aboard RXTE (Jahodaetal.2006,Swank1994respectively).Theobservations 0 wereperformedon2010May15,16and17at04:47,16:54and 55331.00 55331.50 55332.00 55332.50 55333.00 55333.50 03:53UTCwitheachobservationhavinganexposureof∼10ks. MJD In all subsequent analysis we now use an updated orbital pe- Fig.1. Background subtracted, binned standard mode one light rioddeterminationandperiastronephemerisof4.9278±0.0002d curvesoftheRXTE/PCAobservations(bintime:500s).Theor- andMJD53732.632respectively,placingperiastronatanorbital bital phase of each observation is shown using the zero phase phase of 0.0. These improved values were determined through ephemeris MJD53732.632 and orbital period 4.9278d where utilisingtheanalysismethodsoutlinedinClarketal.(2009)with periastronoccursatorbitalphase0.0. anINTEGRAL/IBISdatasetof13.3Ms,representingan∼66% increase in exposure compared with the original study. The or- bital phase of the three observations are then 0.412, 0.720 and Table 1. Statistical properties of the 0.125s resolution 2 0.811respectively. −120keVstandardmodeonelightcurves. The data were analysed using the standard tools within Obs AverageFlux χˆ2 Exposure HEASOFT v6.9. The analysis was performed on both the sci- countss−1PCU−1 s encearray(standardmodeone)andscienceevent(goodxenon) 1 7.02±0.06 12.1 9102 data. PCU 2 was the only detector active throughout all three 2 7.19±0.06 10.9 9722 observations,withPCU4activeforasmallpercentageofeach, 3 6.87±0.06 12.2 8872 henceextractionwasperformedonPCU2dataonly.TheGTIs oftheobservationsweregeneratedwiththeftoolMAKETIME using the observation filter files and the standard event selec- tioncriteriadescribedintheRXTE dataanalysismanual.Light anditisseenthatthefirstobservation(Obs1)occursjustbefore curvesofthestandardmodeonedata,whichcoversthefullen- apastronwhilstthesecondandthirdobservations(Obs2and3) ergy response of the PCA at 0.125s time resolution, were ex- wereperformedasthecompactobjectisapproachingperiastron. tractedusingtheftoolSAEXTRCT.Thelightcurveswereback- Theshapeandintensityoftheemissioninthethreeobserva- groundsubtractedusingthemostrecentPCAfaintbackground tionswasfirstcharacterisedbyassessingthestatisticalproperties modelandbarycenteredusingtheFXBARYftool.Similarlythe of the finely binned (0.125s) standard mode one light curves. eventmodedataweretreatedinthesamemanner,wherebythe The mean fluxes are seen to vary by a maximum of ∼5% be- light curves were extracted using SEEXRCT at 0.125s resolu- tween the observations. The level of variation within each ob- tion.ThreelightcurveswereextractedcoveringthefullPCAen- servationwasalsoinvestigatedbymeansofthegoodnessoffit ergyresponse,3−10keVand10−120keVwithalllayersofPCU to a constant flux at the average count rate. All three observa- 2 used in each case. Again the light curves were barycentered tions showed a poor fit, indicating variations in excess of sta- usingFXBARY.Thescienceeventdatawasalsousedtocreate tistical noise within the light curves. These statistics are out- energyspectraoftheobservations.Thespectra,backgroundand linedinTable1.Thepropertiesoftheobservedemissionsuggest responsefilesweregeneratedforeachscienceeventfileandfor thattherearenosourcesundergoingbright,fastoutburstswithin every combination of active detectors using the standard meth- theseobservations. ods. As the majority of the exposure in each observation was achievedwhilstPCU2wastheonlyactivedetector,thesespec- tra, and backgrounds, were summed using SUMPHA and the 3.1. PeriodicityAnalysis responsescombinedusingADDRMF. Tosearchforpulsationsinthedata,thebackgroundsubtracted, 0.125s resolution standard data mode one light curves of each observation were subjected to a Lomb-Scargle analysis (Lomb 3. Results 1976, Scargle 1982). The analysis of Obs 2 and 3 showed no Figure1showsthePCAstandardmodeonelightcurvesofthe significantsignalswithintheirperiodograms.However,asignif- observationsat500sbinning.InthismodePCAdatahasnoen- icantpeakat71.49swasobservedwithapowerof22.039inthe ergyresolutionandhencetheserepresenttheemissiondetected periodogram of Obs 1, as shown in Fig. 2. This peak does not acrossthefull2−120keVPCAenergyrange.Weseethatsteady, correspondtoabeatfrequencybetweenanycharacteristictime lowlevelemissionatanintensityof∼7countss−1PCU−1isbe- scales within the data and is present when the analysis is per- ing detected across all three observations. This corresponds to formedonlightcurveswithawiderangeofdifferenttimebin- afluxof∼5.0−5.5×10−11ergcm−2 s−1 inthe3−10keVenergy nings. The 99.99% and 99.999% confidence levels were calcu- rangeusingthespectralmodelsoutlinedinTable2.Theorbital latedatLomb-Scarglepowersof20.133and22.227respectively phase of the centre point of each observation is also indicated usingarandomisationtestasoutlinedinHilletal.2005whilst 2 Draveetal.:X-raypulsationsfromtheregionoftheSFXTIGRJ17544−2619 25 8.5 99.999% Confidence Level 99.99% Confidence Level 20 8.0 15 !1U 7.5 C Power 10 !1Counts s P 7.0 5 6.5 0 6.0 0 100 200 300 400 500 0.0 0.5 1.0 1.5 2.0 Period (Seconds) Phase Fig.2.Lomb-Scargleperiodogramofthebackgroundsubtracted 0.30 Obs10.125sresolutionstandardmodeonelightcurveshowing apeakataperiodof71.49s.Thisperiodiscalculatedtohavea 0.25 V sayilsgsionsiotfiafckatihnnecgeliinogtfhot4ac.3cu7crσvoeu.sntfrtohmeetxhterathtrreiaelsobresseurvltaintigonfsro.mBythlienaenarally- Ratio 10 ! 120 / 3 ! 10 ke 00..1250 sisnaigtteinoripnfiocblaaantsicenedgotbfeesttthwewe7ae1sn.4uth9seesdsepteocroieonsdfitidamesan4tce.e3t7lheσev.eeAlrsroswrimeoinclaatrlhcriuaslnapdteeormitohide- Hardness 00..0150 as ±0.02s (see Drave et al. 2010 for details). Figure 3 shows 0.00 the phase-folded light curve of Obs 1 using the 71.49s period. 0.0 0.5 1.0 1.5 2.0 Theshapeisdominatedbyalargepeakontopofanunderlying Phase flux at ∼6.4counts s−1 PCU−1, representing a pulse fraction of ∼13%,wherethepulsefractionisdefinedas(C −C )/(C Fig.3.Top:Obs1lightcurvephase-foldedonthe71.50speriod. +C ).Apossiblesecondpeakisalsoobservemdaaxtlowmeinrsignmiafx- The profile shows a pulse fraction of ∼13%. The dashed line min icanceandisoffsetfromthelargerpeakbyaphaseof∼0.3−0.4. shows the relative exposure of each phase bin (a relative expo- sureof1equatestoacountrateof8.3and0to0withinthisscal- Adiscussionofthisshapeandthenon-detectionsinObs2and ing).Bottom:Hardnessratiobetweenthe3−10and10−120keV 3isgiveninSect.4 energybandsshowingahardeningoftheobservedemissiondur- For consistency the event mode data were also investigated ingthepeaksintheabovephasefoldedlightcurve.Bothcurves for periodicities using the extracted 0.125s resolution back- posses the same phase binning and arbitrary zero pulse phase ground subtracted light curves. Using the full energy response ephemeris. lightcurve(2−120keV)theonlysignificantfeatureintheperi- odogramisdetectedat71.52swithapowerof20.37,inagree- mentwiththeresultfromthestandardmodeonedata.Theevent 3.2. SpectralAnalysis mode data light curves in the 3−10 and 10−120keV energy bands were also searched for pulsations (Note: When using re- Theenergyspectraderivedfromthetotaldataineachobserva- stricted energy ranges 3keV is used as the low energy cut-off tion were investigated to characterise the spectral shape of the pointduetothedegradingofthePCAenergycalibrationbelow emission and aid in the identification of its source. The spectra thisvalue).Inbothcasesapeakat∼71.5sisobservedinthepe- werefitwithmodelsinthe3−20keVenergyrangeusingXSPEC riodogram but it is not at a significant level. The sum of these v12.4.Allreportederrorsfromthespectralfitsarequotedatthe periodogramsdoeshowever,produceasinglesignificantpeakat 90% confidence level. The spectra were initially fit with sim- the correct periodicity. We take this to illustrate that due to the ple absorbed models (e.g. phabs(bremss)), however the large faintnessoftheemissionasignificantdetectioncanonlybeob- reduced Chi-Squared values (6.8−12.8 for 37 degrees of free- tainedwhenthefullenergyrangeisincludedinthedatasetand dom) showed a need for an additional Gaussian component in- hence the periodicity is present (to some extent) over a broad terpreted as an iron emission line. It was found that absorbed rangeofenergies. power-law models again with this additional Gaussian compo- Variationsinthehardnessratioasafunctionofpulsephase nent,gavethebestfitsacrossallthreeobservations.Power-laws have been investigated. Figure 3 shows the 2−120keV phase- with high energy cut-offs and thermal bremsstrahlung models folded light curve (top) above the phase-folded 10−120 to werealsofoundtoproducestatisticallysimilarfitsforsomein- 3−10keVhardnessratio(bottom).Itcanbeseenthattheemis- dividualobservations,butneithercouldproducegoodfitsacross sion hardens during the two pulse phase regions that are coin- allthreeobservations. cident with increased flux in the phase-folded light curve. This Theresultsofthepower-lawfitsareoutlinedinTable2.Itis suggests a physical origin for both the high and lower signifi- seenthatthespectraarequitehighlyabsorbedanddonotshow cancepeaksseeninthephase-foldedlightcurveshowninFig.3 asignificantvariationincolumndensityacrosstheobservations. (top). Theypossessphotonindicesthatvaryfrom2.37+0.09to2.69+0.13 −0.09 −0.13 3 Draveetal.:X-raypulsationsfromtheregionoftheSFXTIGRJ17544−2619 between Obs 1 and 3. The inferred 3−10keV fluxes also show 4. Discussion some variation across the three observations, declining signif- icantly from 5.43+0.08×10−11 in Obs 1 to 5.13+0.10×10−11 erg Following the spectral analysis presented in Sect. 3.2 we have cm−2 s−1 in Obs 3−.0.1T0he flux in Obs 2 lies betw−0e.1e2n these two concluded that the emission observed in Obs 2 and 3 is most likely originating from the diffuse Galactic Ridge emission. values, suggesting a decline throughout the observations. The HoweverinObs1thereappearstobeevidenceforanadditional spectraoftheindividualobservationswerealsosummedandthe flux component that is generating a periodic signal. This emis- totalspectrumfitted.ThebestfitparametersareshownTable2 sionisattributedtoafurtheractiveX-raysourcewithintheFOV andasimilarshapeisagainobserved. duringthisobservation.Asnosignificantperiodicsignalswere The large equivalent width of the iron line component in observed during Obs 2 and 3, and no structure was seen when each spectrum, varying from 0.78 to 0.88keV, is of interest. theselightcurveswerefoldedontheknown71.49speriod,we Thesevaluesareconsistentwiththosequotedasresultingfrom concludethattheadditionalX-raysourcewasnotactiveduring Galactic Ridge emission (Koyama et al. 1986), suggesting that theseobservations.Thisinterpretationissupportedbythesignif- GalacticRidgeemissioncouldaccountforpartoralloftheemis- icant decrease in 3 − 10keV flux observed between Obs 1 and sion detected in these observations. To investigate this we fit 3,seeSect.3.2.Wenotehoweverthatthisinterpretationrepre- thespectrawiththemodelofthecentralridgeregionpresented sentsthesimplestsituationwherebyweonlyusetwofluxcom- in Table 3 of Valinia & Marshall (1998). The Raymond-Smith ponentsthatwecandefinitivelyidentify,namelythepulsedflux plasma temperature and power-law photon index were fixed to in Obs 1 and a contribution from the Galactic Ridge emission kT=2.9keV and Γ =1.8 respectively, whilst the absorption and in all three observations. In fact it may be that there are addi- normalisations were left as free parameters. As Table 3 shows, tional faint, non-pulsating sources within the PCA FOV during Obs 2 and 3 are well fit by this model. However Obs 1 is not allthreeobservationsthatcontributetowardsthedetectedfluxin well fit by this model, indicating that Galactic Ridge emission each.Changesinthefluxemittedbyorthenumberofanysuch doesnotfullydescribethespectralshapeofthisobservation.A sources could cause the variations in the average 2 − 120keV decreasingunabsorbedfluxtrendisalsoobservedbetweenObs1 flux detected in each observations as outlined in Table 1. Due and3.CombiningthepoorfitgivenbytheGalacticRidgemodel thenon-imagingnatureofthePCAhoweveritisnotpossibleto inObs1andthedecreasingfluxtrendseenacrossthethreeob- performidentificationofanynon-pulsatingsourceotherthanthe servationswiththefactthatapulsationisonlyseenduringObs GalacticRidgeemission,identifiedbythelargeequivalentwidth 1,istakentoshowthatinthisobservationthereisanadditional oftheironlinecomponentinthespectralfits,andhenceweuse sourceofX-rayemissionthatisgeneratingthepulsedsignalthat themostsimplisticinterpretationfortheremainderofthispaper. isobservedinadditiontotheGalacticRidgeemission. Asecondconsequenceofthenon-imagingnatureofthePCAis that we are also required to give further consideration as to the To further investigate the possible nature of the additional sourceoftheexcess,pulsedemissionseeninObs1. X-raysourceweusedpulsephaseresolvedspectroscopytoex- Figure 4 shows the fourth INTEGRAL/IBIS survey signifi- tractspectraduringthepulseon,phase0.4−0.7inthetoppanel cance map of the region in the 18−60keV energy band (Bird of Fig. 3, and pulse off, 0.7−1.0, phase regions. The spectrum et al. 2010). Overlaid are the PCA half and zero collimator re- collectedfromthepulseoffregionwasthensubtractedfromthe sponse contours, at 0.5o and 1o respectively, for the pointing pulse on region, this calculation was performed in ‘count rate’ used, along with the sources in the INTEGRAL general ref- space to compensate for the different exposure times accumu- erence catalog (squares) (v.31, Ebisawa et al. 2003) and fur- lated for each phase region. However, whilst there were resid- therX-raydetections(circle).SimilarlytheROSAT allskysur- ual counts in the subtracted spectrum the signal-to-noise was vey photon map of the region in the 0.1−2.4keV energy range not sufficient to allow the fitting of spectral models to charac- is shown in Fig. 5 (Voges et al. 1999). It can be seen that terise the pulsed emission and provide a direct estimate of the the only two sources significantly detected within the zero re- residual flux. Figure 3 (bottom) does show a hardening of the sponse contour by INTEGRAL are IGR J17544−2619 and IGR emissionduringthepulseonphaseregionhowever,suggesting J17507−2647. The latter of these sources, which is at the edge thepresenceofanadditionalactive,pulsingX-raysourcewithin of the FOV, was characterised using Chandra observations by theFOVastheGalacticRidgeemissiondoesnotvaryonthese Tomsick et al. (2009) as a weak, persistent source with a flux timescales. of 4.5×10−12erg cm−2 s−1 (0.2−10keV) that is most likely We can make a refined estimate of the pulsed flux in Obs a distant HMXB at ∼8.5kpc. The Chandra spectrum of the 1 by using the phase-folded light curve shown in the top panel source reported showed a high level of absorption with nH = ofFig.3.Undertheassumptionthattheminimumcountratein 1.34×1023cm−2 and the source was not detected in the ROSAT the phase-folded light curve corresponds to zero pulsed emis- map. As IGR J17507−2647 is located at the edge of the PCA sion, which is supported by the consistency of the count rate FOV it is in a region of low collimator response (< 5%) and betweenphases0.70and1.0inFig.3,wecalculatethepercent- would therefore require a flux of at least a factor of 20 greater age excess above this minimum count rate in each phase bin. than that reported by Tomsick et al. 2009 to generate the ob- Taking the average of the excesses observed in each phase bin servedpulsedflux.Asthisisapersistentsourcewithnoreported then produces an estimate of the pulsed flux fraction as 8.9%. outbursts we conclude that the emission observed by RXTE is Using the Obs 1 flux value obtained from the spectral fits out- unlikelytobecontaminatedbyIGRJ17507−2647. linedinTable2thiscorrespondstoafluxof4.8×10−12ergcm−2 ThesoftX-raysource1RXSJ175454.2-264941(Vogesetal. s−1 (3−10keV) originating from the pulsed signal with the re- 1999)isreportedintheROSAT brightsourcecatalogandislo- mainder resulting from the constant Galactic Ridge emission. catednearthehalfresponsecontour;itistheonlybrightsoftX- AstheGalacticRidgeemissioncannotgenerateapulsedsignal raysourcedetectedwithinthePCAFOVbyROSAT.Thehard- orspectralvariationsonatimescaleoftensofsecondswethere- nessratioreportedsuggeststhesourceismoderatelyhard(0.82 foreattributethispulsedfluxcomponenttoanotheractiveX-ray forthe0.5−2/0.1−0.4keVenergybands)andcouldbedetected sourcewithintheFOVduringObs1. by the PCA. There are also four sources from the ROSAT faint 4 Draveetal.:X-raypulsationsfromtheregionoftheSFXTIGRJ17544−2619 Table2.Spectralfitstothetotalandthreeindividualobservationsusingthemodel:Phabs(Powerlaw+Gaussian).Errorarequoted atthe90%confidencelevel Obs χˆ2/d.o.f. Γ n LineEnergy LineSigma LineEquivalent Flux(3−10keV) H 1022cm−2 keV keV WidthkeV ergcm−2s−1 1 0.95/34 2.37+0.09 4.4+1.3 6.54+0.05 0.03+0.19 0.82 5.43+0.08×10−11 −0.09 −1.2 −0.05 −0.03 −0.10 2 0.39/34 2.40+0.13 5.1+1.7 6.56+0.07 0.15+0.18 0.78 5.28+0.10×10−11 −0.13 −1.7 −0.07 −0.15 −0.11 3 0.78/34 2.69+0.13 5.9+1.6 6.67+0.07 0.14+0.17 0.88 5.13+0.10×10−11 −0.13 −1.6 −0.06 −0.14 −0.12 Total 1.01/34 2.48+0.06 5.2+0.8 6.58+0.03 0.11+0.11 0.80 5.32+0.06×10−11 −0.06 −0.8 −0.03 −0.10 −0.06 Table 3. Spectral fits to the three individual observations using themodelPhabs(Raymond+Powerlaw)ofValinia&Marshall (1998) Obs χˆ2 Flux(3−10keV) ergcm−2s−1 1 1.735 5.41+0.09×10−11 −0.08 2 0.700 5.26+0.10×10−11 −0.08 3 0.869 5.13+0.07×10−11 −0.09 sourcecatalog(Vogesetal.2000)withintheFOV,howeverwe would not expect a detection of these sources using the PCA. FinallytheASCAsourcesshowninFig.4arereportedasX-ray pointsourcesbySugizakietal.(2001).AXJ1753.5−2538was detected at 3σ in the 0.7−2keV band but was not found in the 2−10keVbandindicatingthatitisasoftsource,hencewewould not expect a detection with PCA. However the non-detection in Fig. 5 does suggest a transient nature for this source. AX J1754.0−2553 is not detected in the 0.7−2keV band but does Fig.4.TheIBISsurvey18−60keVsignificancemapoftheIGR have a 3.8σ detection in the 2−10keV energy range. This sug- J17544−2619region,exposure∼8Ms,withthePCAFOVhalf geststhesourceisharderand,asfor1RXSJ175454.2−264941, and zero response contours overlaid (Bird et al. 2010). The couldbedetectedinthePCAdata. sourcesinthisregionthatarecontainedintheINTEGRALgen- Ofallthesourcesdiscussedabove,IGRJ17544−2619isthe eralreferencecatalogareshownassquarepoints,whilstfurther mostactiveshowingalargenumberofoutburststhathavebeen X-raysourcesareshownascircles. detectedbyalargevarietyofmissions(Clarketal.2009).Taking into account the nature of the known sources within the PCA FOVmakesIGRJ17544−2619themostlikelyknownsourceof theemissiondetectedbythePCAinstrument.Fortheremainder alsobeenseentovaryasafunctionofpulsephaseintheSFXT of this paper we assume this to be the case, although we can- IGRJ11215−5952(Sidolietal.2007b),showingahardeningof notruleoutthepossibilitythattheemissioniscomingfromone emission during the pulse-on phase region as is also observed of theother knownsources, 1RXSJ175454.2−264941 andAX forIGRJ17544−2619inthelowerpanelofFig.3. J1754.0−2553inparticular,oranewunknownsourcewithinthe The detection reported here, combined with past observa- FOV. tions by other observatories, indicates that the pulsation signal Ifweconsiderthe71.50ssignalasapulsationfromtheIGR produced by IGR J17544−2619 is not always observable. IGR J17544−2619system,thenthisconfirmsthatthecompactobject J17544−2619 has had snapshot observations taken by XMM- inthesystemisaneutronstar,ashasbeensuggestedfromqui- Newton (three ∼10ks exposures, Gonza´lez-Riestra et al. 2004) escencespectra(in’tZand2005,Pellizzaetal.2006).Assuming andChandra(19.6ks,in’tZand2005)alongwithamonitoring a source distance of ∼3.6kpc this makes the estimated source campaignbySwift/XRT(Sidolietal.2008),noneofwhichshow fluxequivalenttoanunabsorbedluminosityof∼1×1034ergs−1 detectionsofthe71.49ssignal.Howevertheobservationalprop- (3−10keV),indicatingthatIGRJ17544−2619wasobservedin erties of SFXTs make the detection of pulse periods difficult. alowX-raystateasopposedtoduringoneofitslargeoutbursts ThebiggestobstacleistheX-rayflaringtimescaleobservedat (suchasthe1036ergs−1eventreportedinGrebenevetal.2004). softX-rayenergies,thatissimilartolikelypulsationperiods(i.e. X-ray pulsations have also been detected in other SFXT sys- tenstohundredsof seconds).Flaresdominatemanyofthe soft temsobservedduringsimilarlowluminositystates,∼1×1034erg X-raydatasetsandmaskpulsationsignalsduetothelargerflux s−1 (0.5−10keV) in IGR J18483−0311 (Giunta et al. 2009) variationstheyinduce.Manyoftheremainingdatasetsthatare and 2.3×1034erg s−1 (2−10keV) in AX J1841.0-0536 (Bamba notdominatedbyflaresareshortSwift/XRTexposuresthathave et al. 2001) for example. Additionally the hardness ratio has too short baselines or insufficient statistics for accurate timing 5 Draveetal.:X-raypulsationsfromtheregionoftheSFXTIGRJ17544−2619 10 4 IGR J16418!4532 IGR J11215!5952 10 3 IGR J16465!4507 d (s) 10 2 IGR J17544!2619 o eri 10 1 IGR J18483!0311 P AX J1841.0!0535 e uls P 10 0 SFXTs 1100!!21 IGR J16479!4514 SAX J1818.6!1703 XTE J1739!302 BWReLiXn FdRi!lBlfiesndg SSggXXRRBBss 10 0 10 1 10 2 10 3 Orbital Period (d) Fig.6.TheCorbetDiagramshowingthelocationsoftheSFXT systems with at least one known period (Orbital or Pulse) (Corbet 1986). IGR J17544−2619 lies in the region of param- eterspacepopulatedbytheclassicalSgXRBs. Fig.5. The ROSAT all sky survey photon map of the IGR supergiant companion, for example IGR J11215−5952 (Sidoli J17544−2619 region in the 0.1−2.4keV energy range (Voges et al. 2007b) and XTE J1739−302 (Drave et al. 2010). Such etal.1999).TheannotationsarethesameasthoseinFig.4. characteristics are more akin to the BeXRB class of HMXBs. Currently the SFXT class is split into classic and intermediate SFXTs via the X-ray luminosity dynamic range observed from analysis to be performed (e.g. Sidoli et al 2008). Given the na- thesystem:>100foran‘intermediate’and>1000fora‘clas- ture of the previous soft X-ray observations of this source it is sic’ SFXT. It is now becoming apparent that a distinction can plausible that the pulsation of IGR J17544−2619 has not been also be drawn from the level of similarity of individual SFXTs detected prior to these observations, which show a steady flux toeachoftheclassicHMXBfamilymembersandsuchsimilar- at the 1034 erg s−1 level. These fluxes are consistent with those itiesmaybevisualisedbytheCorbetdiagram.Thisdistinction, duringwhichpreviousSFXTpulsationshavebeendetected(e.g. and the level to which some systems may again be ‘intermedi- Giuntaetal.2009). ate’ between the two classes, should become more apparent as thenumberofSFXTsthatcanbeplacedontheCorbetdiagram Furthermore, the observation was performed as the neutron increases. Such a distinction could be an indication of a vari- star was approaching apastron in the system which could also ety of stellar wind geometries present in SFXT systems and/or helpexplainwhythepulsationsweredetectedhere.Clarketal. ofvaryingevolutionarypathsfollowedintheircreation(seeLiu (2009) showed that at this orbital phase there is still a non- etal.2011forfurtherdetails). zero probability of stellar wind clump interaction in this sys- tem(Clarketal.2009,Fig.7).Additionallyunderthemodelof Duccietal.(2009)thestellarwindclumpsexpandastheymove 5. Conclusions out from the companion supergiant star, likely becoming more homogeneous in density as they travel. As a result we may ex- Using observations from RXTE we have detected a transient pectthattheX-rayemissiongeneratedduringtheinteractionof 71.49±0.02s signal that is most likely from the spinning neu- theneutronstarandtheexpandedstellarwindclumpwouldalso tron star in the SFXT IGR J17544−2619. The phase-folded besmootherandlesspronetoundergothefastflaresthatdom- light curve shows a double peaked structure where the emis- inateothersoftX-rayobservationsofthissource.Untilalarger sion is observed to harden during the peaks in the profile. numberofdetectionshavebeenachievedhowever,formalcon- The source was observed in a steady state of emission with a clusionsonanylinkbetweenthedetectionofthepulsationsand luminosity of ∼1×1034erg s−1 (3−10keV). This pulse period theorbitalphaseoftheobservationscannotbedrawn. places IGR J17544−2619 in the same region as the classical Combining this detected pulsation with the 4.926d orbital wind-fedSgXRBsontheCorbetdiagram.ThecandidateSFXT period of Clark et al. (2009) allows the placement of IGR IGRJ16418−4532alsooccupiesthisregion,howeverothersys- J17544−2619ontheCorbetdiagram(Corbet1986),Fig.6.We tems appear to be in intermediate (e.g. IGR J18483−0311) and seethatIGRJ17544−2619islocatedclosetotheclassical,wind- BeXRB like positions (e.g. IGR J11215−5952) suggesting that fed SgXRBs. Under the ‘clumpy wind’ model of SFXTs (in’t SFXTseitherspanthegapbetweentheclassicaltypesofHMXB Zand2005)thedifferenceinbehaviourseenwhencomparedto or represent extreme examples of the known classes. A greater the classical systems is explained by an enhanced eccentricity population of SFXTs is required on the Corbet diagram to in- which results in the compact object spending only a fraction vestigate this further. We encourage further observations of the of its orbit within a dense stellar wind environment (Walter & IGR J17544−2619 system using focusing X-ray telescopes to ZuritaHeras2007).InthisrespectSFXTscanbeconsideredas removethecurrentuncertaintyintheoriginofthedetectedpul- anextensionoftheclassicalSgXRBsthatresultfromvaryingor- sations and definitively prove that this is the pulsation of IGR bitalparameters.However,someSFXTshavelongerorbitsand J17544−2619.Phase-targeted,high-cadenceobservationsofthe show orbital emission profiles that could be explained by the source will also allow for a better understanding of the physi- presence of a disk-like structure within the stellar wind of the cal processes that could be causing the transient nature of the 6 Draveetal.:X-raypulsationsfromtheregionoftheSFXTIGRJ17544−2619 pulsationdetection.FurtherobservationsofallSFXTswithone Sugizaki,M.,Mitsuda,K.,Kaneda,H.,&etal.,.2001,ApJS,134,77 known periodicity (pulsation or orbital) are also vital to allow Sunyaev, R. A., Grebenev, S. A., Lutovinov, A. A., & et al.,. 2003, The thelargestpossiblesampletobeplacedontheCorbetdiagram, Astronomer’sTelegram,190,1 Swank,J.H.1994,inBulletinoftheAmericanAstronomicalSociety,Vol.26, furtheringourknowledgeofthenatureoftheclassaswhole. BulletinoftheAmericanAstronomicalSociety,1420–+ Tomsick,J.A.,Chaty,S.,Rodriguez,J.,&etal.,.2009,ApJ,701,811 Ubertini,P.,Lebrun,F.,DiCocco,G.,&etal.,.2003,A&A,411,L131 Acknowledgements Valinia,A.&Marshall,F.E.1998,ApJ,505,134 Voges,W.,Aschenbach,B.,Boller,T.,&etal.,.1999,A&A,349,389 Theauthorswishtothanktheanonymousrefereefortheirhelp- Voges,W.,Aschenbach,B.,Boller,T.,&etal.,.2000,IAUCirc.,7432,3 fulcommentsandsuggestions.TheauthorswishtothankR.H. Walter,R.&ZuritaHeras,J.2007,A&A,476,335 D.CorbetandthePCAinstrumentteamfortheirdiscussionson Wijnands,R.2003,TheAstronomer’sTelegram,191,1 thetimescalesofsystematiceffectswithinthePCAbackground Winkler,C.,Courvoisier,T.,DiCocco,G.,&etal.,.2003,A&A,411,L1 ZuritaHeras,J.A.&Walter,R.2009,A&A,494,1013 model.TheauthorsalsowishtothankJ.J.M.in’tZandforhis helpfuldiscussionontheGalacticRidgeemission. S. P. Drave acknowledge support from the Science and Technology Facilities Council, STFC. L. J. Townsend is sup- ported by a Mayflower scholarship from the University of Southampton. A. Bazzano and V. Sguera acknowledge sup- port from ASI/INAF contract n.I/009/10/0. This research has made use of the SIMBAD database, operated at CDS, Strasbourg, France. This research has made use of the IGR Sources page maintained by J. Rodriguez & A. Bodaghee (http://irfu.cea.fr/Sap/IGR-Sources/). References Bamba,A.,Yokogawa,J.,Ueno,M.,&etal.,.2001,PASJ,53,1179 Bird,A.J.,Bazzano,A.,Bassani,L.,&etal.,.2010,ApJS,186,1 Bozzo,E.,Campana,S.,Stella,L.,&etal.,.2008,TheAstronomer’sTelegram, 1493,1 Clark,D.J.,Hill,A.B.,Bird,A.J.,&etal.,.2009,MNRAS,399,L113 Corbet,R.H.D.1986,MNRAS,220,1047 Drave,S.P.,Clark,D.J.,Bird,A.J.,&etal.,.2010,MNRAS,409,1220 Ducci,L.,Sidoli,L.,Mereghetti,S.,&etal.,.2009,MNRAS,398,2152 Ebisawa,K.,Bourban,G.,Bodaghee,A.,&etal.,.2003,A&A,411,L59 Giunta,A.,Bozzo,E.,Bernardini,F.,&etal.,.2009,MNRAS,399,744 Gonza´lez-Riestra,R.,Oosterbroek,T.,Kuulkers,E.,&etal.,.2004,A&A,420, 589 Grebenev, S. A., Rodriguez, J., Westergaard, N. J., & et al.,. 2004, The Astronomer’sTelegram,252,1 Hill,A.B.,Walter,R.,Knigge,C.,&etal.,.2005,A&A,439,255 in’tZand,J.,Heise,J.,Ubertini,P.,&etal.,.2004,inESASpecialPublication, Vol. 552, 5th INTEGRAL Workshop on the INTEGRAL Universe, ed. V.Schoenfelder,G.Lichti,&C.Winkler,427–+ in’tZand,J.J.M.2005,A&A,441,L1 Jahoda,K.,Markwardt,C.B.,Radeva,Y.,&etal.,.2006,ApJS,163,401 Koyama,K.,Makishima,K.,Tanaka,Y.,&etal.,.1986,PASJ,38,121 Lebrun,F.,Leray,J.P.,Lavocat,P.,&etal.,.2003,A&A,411,L141 Liu,Q.Z.,Chaty,S.,&Yan,J.Z.2011,MNRAS,415,3349 Lomb,N.R.1976,Ap&SS,39,447 Lutovinov,A.,Revnivtsev,M.,Gilfanov,M.,&etal.,.2005,A&A,444,821 Masetti,N.,Mason,E.,Morelli,L.,&etal.,.2008,A&A,482,113 Negueruela,I.,Smith,D.M.,Harrison,T.E.,&etal.,.2006a,ApJ,638,982 Negueruela, I., Smith, D. M., Reig, P., & et al.,. 2006b, in ESA Special Publication,Vol.604,TheX-rayUniverse2005,ed.A.Wilson,165–+ Nespoli,E.,Fabregat,J.,&Mennickent,R.E.2008,A&A,486,911 Pellizza,L.J.,Chaty,S.,&Negueruela,I.2006,A&A,455,653 Rahoui,F.&Chaty,S.2008,A&A,492,163 Rahoui,F.,Chaty,S.,Lagage,P.,&etal.,.2008,A&A,484,801 Rampy,R.A.,Smith,D.M.,&Negueruela,I.2009,ApJ,707,243 Romano, P., Cusumano, G., Sidoli, L., & et al.,. 2008, The Astronomer’s Telegram,1697,1 Romano,P.,Sidoli,L.,Cusumano,G.,&etal.,.2009,MNRAS,392,45 Romano,P.,Sidoli,L.,Mangano,V.,&etal.,.2007,A&A,469,L5 Scargle,J.D.1982,ApJ,263,835 Sguera,V.,Barlow,E.J.,Bird,A.J.,&etal.,.2005,A&A,444,221 Sguera,V.,Bazzano,A.,Bird,A.J.,&etal.,.2006,ApJ,646,452 Sguera,V.,Hill,A.B.,Bird,A.J.,&etal.,.2007,A&A,467,249 Sidoli, L., Mereghetti, S., Vercellone, S., & et al.,. 2007a, The Astronomer’s Telegram,997,1 Sidoli,L.,Romano,P.,Mangano,V.,&etal.,.2008,ApJ,687,1230 Sidoli,L.,Romano,P.,Mangano,V.,&etal.,.2009,ApJ,690,120 Sidoli,L.,Romano,P.,Mereghetti,S.,&etal.,.2007b,A&A,476,1307 7