DRAFTVERSIONFEBRUARY2,2008 PreprinttypesetusingLATEXstyleemulateapj XMM-NEWTON SPECTROSCOPYOFTHEACCRETION-DRIVENMILLISECONDX-RAYPULSAR XTEJ1751-305INOUTBURST J.M.MILLER1,2,3,R.WIJNANDS1,4,5,M.MÉNDEZ6,E.KENDZIORRA7A.TIENGO8,M.VANDERKLIS8,D.CHAKRABARTY1, B.M.GAENSLER2,ANDW.H.G.LEWIN1 Subject headings: binaries:close – pulsars: general – pulsars: individual: XTE J1751- 305 – stars: neutron – x-rays:binaries–whitedwarfs DraftversionFebruary2,2008 3 ABSTRACT 0 We presentan analysisof the first high-resolutionspectra measuredfroman accretion-drivenmillisecond X- 0 ray pulsar in outburst. We observed XTE J1751–305 with XMM-Newton on 2002 April 7 for approximately 2 35 ks. Using a simple absorbed blackbody plus power-law model, we measure an unabsorbed flux of 6.6± n 0.1×10- 10 erg cm- 2 s- 1 (0.5–10.0 keV). A hard power-law component (Γ=1.44±0.01) contributes 83% of a the unabsorbedfluxinthe0.5–10.0keV band,butablackbodycomponent(kT =1.05±0.01 keV)isrequired. J Wefindnoclearevidencefornarroworbroademissionorabsorptionlinesinthetime-averagedspectra,andthe 7 sensitivityofthisobservationhasallowedustosetconstrainingupper-limitsonthestrengthofimportantfeatures. 1 The lack of line features is at odds with spectra measured from some other X-ray binaries which share some 2 similaritieswithXTEJ1751–305. We discusstheimplicationsofthesefindingsontheaccretionflowgeometry v inXTEJ1751–305. 6 6 1. INTRODUCTION verylow(∼0.01M⊙). 1 After the discovery of XTE J1751–305, we submitted a Millisecond radio pulsars are thought to be created in neu- 8 target-of-opportunityrequestto XMM-Newton for the purpose 0 tronstarlow-massX-raybinaries(LMXBs). InthoseLMXBs, ofstudyingthissourcewithhigh-resolutionspectroscopy. The 2 accretingmattermayspin-upthe neutronstar (see, e.g.,Bhat- X-ray spectrum of the first system — SAX J1808.4–3658— 0 tacharya & van den Heuvel 1991). Therefore, it is expected couldonlybestudiedduringoutburstusingRXTEandwiththe / that millisecond pulsars should also be found in LMXBs dur- h WideFieldCamerasaboardBeppoSAX,whichhaveonlymod- ingtheaccretionphaseofthebinaryasX-raypulsars.Although p erate spectral resolution. In this Letter, we present an analy- evidenceforrapidlyspinningneutronstarsinLMXBswasin- - sisofthetime-averagedEPIC-pnandReflectionGratingSpec- o ferredfromtheburstoscillationswhichwereseenduringtype- trometer (RGS) data. The spectra obtained represent the first r I X-ray bursts in several systems (see Strohmayer 2001 for a t CCD- and gratings-resolution measurements from a millisec- s review), the detection of millisecond pulsations in persistent a ondX-raypulsarinoutburst. emission remained elusive for many years. In 1998, the first : v suchsystemwasdiscovered(SAXJ1808.4–3658,whichhasa Xi spin frequency of 401 Hz; Wijnands & van der Klis 1998a). 2. OBSERVATIONANDDATAREDUCTION This source was extensivelystudied due to its obviousimpor- r XMM-NewtonobservedXTEJ1751- 305beginningon2002 tanceforbinaryevolutionscenariosandaccretionflowgeome- a April 07.5 (UT). The EPIC-pn camera was operated in “tim- tries and dynamics(see, e.g., Wijnands & van der Klis 1998b ing”modetopreventphotonpile-upeffectsandtoallowpulse- andreferencestherein). phase-resolved spectroscopy . The RGS was operated in the Inthespringof2002,Markwardt&Swank(2002a)reported standard“spectroscopy”mode. The“thin”opticalblockingfil- thediscoveryofthesecondaccretion-drivenmillisecondpulsar ter was selected for the EPIC cameras. The EPIC-MOS cam- XTE J1751–305(435Hz) in the Rossi X-rayTiming Explorer eraswerenotoperatedinmodesoptimizedforasourceofthis (RXTE)bulgescanobservationprogram. Aboutonemonthaf- intensity,andwedonotconsidertheMOSspectrahere. ter the discoveryof this system, XTE J0929–314was discov- TheEPIC-pnandRGSdatawerereducedusingSASversion ered (185 Hz; Remillard, Swank, & Strohmayer 2002.). The 5.3. We appliedthe standardreductionprocedure“epproc”to neutron stars in both systems are in orbit around a compan- ion star, with anorbitalperiodof∼42minutes(Markwardt& produceacalibratedpneventlist. In“timing”mode,thespatial information is compressed into one dimension. We extracted Swank2002b;Gallowayetal. 2002). Thesesystemsarevery sourcedatainarectangle(withawidthof37”)alongthelength tightbinariesandtheinferredmassoftheircompanionstarsis oftheCCD, andbackgrounddata fromadjacentregions; only 1CenterforSpaceResearchandDepartmentofPhysics,MassachusettsInstituteofTechnology,Cambridge,MA02139–4307 2Harvard-SmithsonianCenterforAstrophysics,60GardenStreet,Cambridge,MA02138,[email protected] 3NationalScienceFoundationAstronomyandAstrophysicsFellow 4SchoolofPhysicsandAstronomy,UniversityofSt.Andrews,StAndrewsKY169SS,UK 5ChandraFellowwhileatMIT 6SRON,NationalInstituteforSpaceResearch,Sorbonnelaan2,NL-3584CA,Utrecht,NL 7InstitutfürAstronomieundAstrophysik,Abt.Astronomie,UniversitätTübingen,Sand1,D-72076,Tübingen,DE 8AstronomicalInstitute“AntonPannekoek”,UniversityofAmsterdam,Kruislaan403,NL-1098SJAmsterdam,NL 1 2 XTEJ1751- 305 “single” and “double” events were selected. The net pn ex- posure was 33.7 ks. Using LHEASOFT version 5.1, the tool “grppha”wasusedtorebinthespectrumtorequireaminimum of20countsperbin. Weusedthetrial“timing”moderedistri- butionandancillaryresponsefunctionsdevelopedbytheEPIC- pncalibrationteam. We appliedthestandardreductionproce- dure“rgsproc”toproducecalibratedRGSeventlists,first-order spectra, and response functions. Periods of high instrumen- tal backgroundwere excluded, giving a net RGS exposure of 34.2ks. The RGS spectra were groupedto requireatleast 20 countsperbin. 3. ANALYSISANDRESULTS Preliminary spectroscopic results from this XMM-Newton observationwerereportedinMilleretal.2002aandMilleretal. 2002b. TheaccuratepositionofXTEJ1751–305wasreported FIG. 1.— The EPIC-pn spectrum of XTE J1751- 305, fit with a byEhleetal. (2002),alsobasedonanalysisofthisobservation model consisting of blackbody and power-law components modified (R.A.=17h51m13.5s,Decl.=- 30◦37′22′′,equinox2000.0). byphotoelectricabsorption(seeSection3fordetails). Thespectrum TheinternalcalibrationaccuracyoftheEPICcamerasisbet- anddata/modelratioshownabovearerebinnedforvisualclarity. terthan5%foron-axissources(Kirsch2002). Wefinddevia- in the 0.5–10.0 keV band (an extrapolation from the 0.6– tionsupto0.6keVwhicharenotremovedbyspectralmodel- 10.0keVbandasthe0.5–10.0keVrangeismoreconventional). ing,andthereforerestrictouranalysisofthepnspectrumtothe 0.6–10.0keVband.Theaveragecountrate(79.36countss- 1)is Motivated by evidence for a Comptonizing volume in SAXJ1808.4–3658(Gierlinski,Done,&Barret2002),wealso wellbelowthatatwhichphotonpile-upoccursintimingmode (1500countss- 1). Testsoftheeffectiveareacalibrationofthe fit the EPIC-pn spectrumwith the “compTT”Comptonization model (Titarchuk 1994), which attempts to account for the RGS reveal accuracies of 5% in the 7–36Å (0.35–1.77 keV) Compton-upscattering of cool photons in a hot corona self- rangeand10%atwavelengthsshorterthan7Å (energiesabove consistently. Thismodeldoesnotprovideanacceptablefitby 1.77 keV; den Herder 2002). Although our best-fit model for itself(χ2/ν=2- 4,ν=1884).Wefindadegeneracybetweena the EPIC-pn spectrum (see below) providesan adequatefit to the RGS spectrum(χ2/ν =1.314,ν=1463), broad deviations coolcorona(kT ∼3keV)withhighopticaldepth(τ ∼8),and a hotcorona(kT ∼40 keV)with loweropticaldepth(τ ∼1), are apparent at energies above 1.8 keV. We therefore rely on eachup-scatteringaseedphotondistributionpeakingbetween theEPIC-pnspectrumtocharacterizethecontinuumemission, kT =0.5–0.6keV.Finally,wetriedfittingablackbodycompo- and search the RGS bandin contiguous3Å slices for narrow nentinadditiontocompTT.Thismodelmightcorrespondtoa features. scenario in which the blackbody emission region is only par- TheEPIC-pnandRGSspectrawereanalyzedusingXSPEC tiallycoveredbythecorona. Thismodelalsoyieldedapoorfit version11.1.Allerrorsquotedinthispaperare90%confidence (χ2/ν∼2). errors. Notethatsystematicerrorsarenotaddedtothespectra The strength, ionization state, and profile of Fe Kα emis- toaccountforfluxcalibrationuncertainties. sion lines — if produced through irradiation of the accretion disk—canrevealagreatdealabouttheaccretiongeometryof 3.1. TheEPIC-pnSpectrum low-massX-raybinaries. Asaietal. (2000)reportthepossible detectionofanFeKαlinein4U1820–30(P =11.4minutes; We first considered a model consisting of blackbody and orb Stella, White, &Priedhorsky1987). Thelinewasfoundtobe power-law components, modified by neutral absorption in the ionized(E=6.6±0.1keV),broad(FWHM=0.7+0.2keV),and ISM (using the “phabs” model in XSPEC). We measure a - 0.5 best-fit equivalent hydrogen column density of N = (9.8± weak(WKα=33+- 1121eV).AssumingalineofequivalentFWHM 0.1)×1021 atoms cm- 2. The blackbody tempeHrature ob- inthe 6.40–6.97keVrange(FeI —Fe XXVI),we measurea tained is moderate: kT =1.05±0.01keV. For spherical sym- 95%confidenceupper-limitofjust6 eVin theEPIC-pnspec- metry, this translates into a blackbody radius of R = f2× trum of XTE J1751–305. Using a Gaussian with zero width 3.01+0.09 (d/10 kpc) km, where the spectral hardening fac- tomodelalinenarrowerthantheEPIC-pnresolution,the95% - 0.06 confidenceupperlimitontheequivalentwidthofanysuchnar- tor f is the ratio of the color temperature and the effective rowFeKαlinesislessthan4eV(6.40–6.97keV). temperature, and d is the distance to the source. The mea- sured power-law index is Γ= 1.44±0.01; the normalization is (6.4±0.1)×10- 2 (photonskeV- 1 cm- 2 s- 1 at 1 keV). This 3.2. Pulse-Phase-ResolvedSpectroscopy:FirstResults model produces an adequate fit to the data: χ2/ν = 1.107 We applied the SAS task “barycen”, and using the binary (whereνisthenumberofdegreesoffreedom;ν=1884forthis systemparametersreportedbyMarkwardtetal.(2002)wepro- model;seeFigure1). We notethatasimplepower-lawmodel ducedspectrafromthe“high”and“low”partsofthepulse. We returnsaverypoorfit(χ2/ν>4,ν=1886);thesoftblackbody findthatthemodelappliedtothetime-averagedspectrumisan componentisstronglyrequiredbythedata. acceptable descriptionof these spectra (no convincingnarrow With this simple but standard model, we obtain an unab- or broad emission or absorptionfeatures are found). Interest- sorbed 0.5–10.0 keV flux of (6.6±0.1)×10- 10 erg cm2 s- 1, ingly,theslopeofthepower-lawcomponentisconsistentinthe or(0.170±0.004)photonscm- 2s- 1. Thehardpower-lawcom- two spectra butthe normalizationchanges, and the blackbody ponentcontributes83%oftheunabsorbedflux componentchangesintemperaturebutnotinnormalization. J.M.Milleretal. 3 FIG. 2.— TheRGS spectra of XTE J1751- 305 (RGS-1inblack, thFeING.e3X.—Ly-TαhelinReGaSt-122s.1p3ecÅtr,ufimtwoifthXaTEsimJ1p7l5e1p-o3w0e5r-ilnawtheplruesgiGoanuosf- RGS-2 in red; both are rebinned for clarity and narrow features are sianmodel.Ourbestfitmodeldoesnotrequireanemissionlinewithin instrumental)fitlocallyintheregionoftheNeKedge(14.25Å)with 12.13±0.1Å,assumingaFWHMequivalent totheprominent NeX asimplepower-lawplusedgemodel. Theedgedepthwasfixedatthe Ly-α emission line seen in the 42-minute X-ray pulsar 4U 1626–67 solarvaluebutthepower-lawindexandnormalizationwereallowedto (Schulzetal.2001).Narrowfeaturesareinstrumental. vary;atrendinthedata/modelratioisclearlyvisible. Fitsmadewith avariableedgedepthsuggestNeislikelyunder-abundant(seetext). the 95%confidenceupper-limitisthattheabundanceofNe isonly77%ofthesolarvalue. AsolarNeabundancedoesnot Similar behavior was discovered in pulse-phase-resolved spectra of SAX J1808.4- 3658 (Gierlinski, Done, & Barret allowforanoptimalcharacterizationofthelocalspectrum(see Figure2). 2002). ThebinaryX-raypulsar4U1626–67alsohasa42-minuteor- Wenotethatapossiblefeatureatapproximately2.9keVap- bitalperiod.Moreover,thedistancetothissystemmaybeclose pears to vary in width, intensity, and centroid energy in the to 8 kpc (Chakrabarty 1998); we anticipate a Galactic center “high”and“low”spectra;however,theerrorlimitsontheline location for XTE J1751- 305and therefore a similar distance. parameters overlap. This feature is also visible in the time- A Chandra/HETGS spectrum of 4U 1626- 67 reveals broad- averaged spectrum (see Figure 1). As there are no astrophys- enedemissionlinesandDoppler-shiftedlinepairs(Schulzetal. icallyabundantelementswithtransitionsnearthisenergy,itis temptingtoassociatethisfeaturewithared-shiftedabsorption 2001).TheNeLy-αline(12.13Å)isparticularlystrong(aflux linefromtheneutronstarsurface(giventhetemperatureofthe of 8.15±0.93×10- 5 photons cm- 2 s- 1 is measured; FWHM blackbody,perhapsFeXXV orFeXXVI).AlthoughanF-test =2860±330 km s- 1). We report the 95% confidence upper- finds that the addition of a Gaussian to model this feature is limits on emission lines from helium-like and hydrogenicNe, significant at the 7σ level of confidence in the “low” part of Mg,andSiinTable1(upperlimitsonOemissionlinesarenot the pulse (3σ in the “high” part), a similar featureis apparent constrainingduetopoorstatisticsatlowenergies). when the EPIC-pn spectrum of the Galactic black hole candi- dateXTEJ1650- 500(Milleretal. 2002c)isfitusingthesame 4. DISCUSSION responsematrix. Itislikelythatthisfeatureis duetoa defect We have analyzed the first CCD- and grating-resolutionX- inthetiming-moderesponsematrix. rayspectraof a millisecondX-raypulsarin outburst. We find noconvincingevidenceforbroadornarrowemissionorabsorp- 3.3. TheRGSSpectra tionfeaturesinthetheEPIC-pnandRGSspectra(seeFigure1 andTable1).Withoursimpleblackbodypluspower-lawmodel We analyzed the RGS spectra in contiguous 3Å slices for fortheEPIC-pnspectrum,we findthatthepower-lawcompo- narrowemissionorabsorptionlines.Ineach3Å slice,asimple nentcomprises83%oftheunabsorbedfluxinthe0.5–10.0keV power-lawplusISMedge(s)modelwasfittothedata. Wefind band. The power-law component is hard: Γ = 1.44±0.01, no convincingevidence for narrow absorption features in this butwithinarangetypicalforX-raypulsars(e.g.,Γ=1.0–1.5; band (0.3–2.5 keV, or 5–40Å). At energies below ∼0.6 keV White,Swank,&Holt1983). TheEPIC-pnspectrumrequires (wavelengths longer than ∼21Å) the sensitivity is low due to ablackbodywithkT =1.05±0.01keVandanimpliedemitting therelativelyhighcolumndensityalongthelineofsighttoXTE radiusofR= f2×3.01+0.09(d/10kpc)km. Itisfairtosaythat - 0.06 J1751–305. ourknowledgeofthespectralhardeningfactor f isratherpoor Recently, evidence for enhanced Ne abundances has been (foradiscussion,seeLewin,vanParadijs,&Taam1993). If f foundincompactbinaries(Schulzetal. 2001;Juett,Psaltis,& isless than1.3, itwouldindicatethatonlypartofthe neutron Chakrabarty2001).WefittheRGSspectraaroundtheposition starsurface(a“hotspot”)isinvolved. of the Ne K edge(14.25Å) with a power-lawmodelbetween Markwardt et al. (2002) fit a simple absorbed power-law 13.25Å and 15.25Å(seeFigure 2). The power-lawindex and model to the RXTE spectrum of XTE J1751–305, and mea- normalizationwere bothallowed to varyin this fit. Using the sure a softer power-law index (1.7 < Γ < 1.9). The effec- cross-sectionsofVerneretal. (1993)andsolarabundancesrel- tivelowersensitivityboundoftheRXTEProportionalCounter ativetoHasperMorrison&McCammon(1983),wefindthat Array (PCA) is 3 keV; fitting the 3–10 keV EPIC-pn spec- Nemaybeunder-abundantinXTEJ1751–305: trumwithonlyapower-law,wefindΓ=1.72±0.01forN = H 4 XTEJ1751- 305 (9.8±0.1)×1021atomscm- 2andΓ=1.90±0.02allowingN spectrum of XTE J1751–305 is also consistent with a highly H tofloat(inthiscase,N =[2.5±0.1]×1022atomscm- 2).Thus, ionized disk in this source. Secondarily, the absence of such H thediscrepantindicesareeasilyexplainedintermsofdifferent lines may be due to differences in the accretion flow geome- instrumentalranges. try due to the neutron star magnetic field. Whereas the mag- InmodelingsimultaneousChandra/LETGSandRXTEspec- netic field in XTE J1751- 305is likely to be similar to that in traofXTEJ0929–314,Juett,Galloway,&Chakrabarty(2002) SAX J1808.6- 3658(B=(2- 6)×108 G; Wijnands& vander measuresimilarlydiscrepantpower-lawindices. Theysuggest Klis 1998a),the magneticfield in 4U 1626- 67is likely much anastrophysicaloriginforthediscrepancyinXTEJ0929–314 higher (B=3×1012 G, Orlandini et al. 1998). The stronger and XTE J1751–305in the form of a power-law with a break magneticfieldin4U1626–67maydisrupttheinnerdisk,allow- or a roll-over in the 1.4–4.4 keV range. A model consist- ingcoolmaterialintheouterdisktobeirradiatedbyaweaker ing of blackbodyand brokenpower-law components(E = corona.Emissionlineshavealsobeenseeninneutronstarsys- break 3.7±0.1 keVx, Γ = 1.34±0.03, Γ = 1.60±0.04) temswhichareviewededge-on(so-calledaccretiondiskcoro- E<3.7 E>3.7 provides an improved fit to the spectrum of XTE J1751–305 nae or “ADC” sources; see, e.g., Kallman et al. 2002); the (χ2/ν=1.090,ν=1882).Althoughtheseresultssuggestthatan absenceofemissionlinesinXTEJ1751–305mayalsobepar- astrophysicaloriginfor the apparentspectralevolutionis pos- tially due to a low system inclination (Markwardtet al. 2002 sible, anumberofconcernsremain. Giventhat: (1)the effec- reportnoevidenceofdipsoreclipses). tivearea(flux)calibrationoftheEPIC-pnhas5%uncertainties Studies of some ultra-compact systems suggest a Ne-rich (Kirsch 2002), (2) the LETGS (plus ACIS-S array) flux cali- companion(forobservationalresults, see: Schulzetal. 2001; brationhas10%uncertaintiesabove1keV (H.Marshall, priv. Juett, Psaltis, & Chakrabarty2001; for theoreticaldiscussions comm.), (3) that RXTE PCA and High EnergyX-Ray Timing see Yungelson, Nelemans, & van den Heuvel 2002; Bildsten Experiment(HEXTE)observationsofthe“Crab”nebula(taken 2002). FitstotheNephotoelectricabsorptionedgeintheRGS tobea purepower-lawforcalibration)differinpower-lawin- spectra of XTE J1751–305 constrain the abundance of Ne to dex by δ(Γ)≃0.1 (see, e.g., Wilms et al. 1999), and (4) that beatmost77%ofthesolarvaluealongthislineofsight(95% Markwardtetal. (2002)didnotfita blackbodycomponentto confidenceupper-limit;seeSection3.3).Thismaysuggestthat the RXTE spectra of XTE J1751–305, we do not consider an aNe-richcompanionisunlikelyinthecaseofXTEJ1751–305. astrophysicaloriginforthediscrepantindicestoberequiredby thepresentdata. 5. ACKNOWLEDGMENTS The upperlimits on the strength of narrow (consistentwith instrumentresolution;FWHM∼0.1keV)andbroad(FHWM= We are grateful to the XMM-Newton Project Scientist Fred 0.7keV)FeKαemissionlinesare4eVand6eV,respectively Jansen for granting time to observe this source. We thank (95% confidence). This stands in contrast the detection of a Michael Nowak for useful discussions. J. M. M. is grateful weakFeKαlineinoutburstspectraofthe401Hzmillisecond for support from the NSF. R. W. was supported by NASA X-raypulsarSAX J1808.4–3658obtainedwithRXTE(Heindl through Chandra fellowship grant PF9-10010, which is oper- &Smith1998;Gierlinski,Done,&Barret2002). Itispossible atedbytheSmithsonianAstrophysicalObservatoryforNASA thatastrong,hotcorona(orthemagnetosphere,ifthecoranae undercontractNAS8-39073. W. H. G. L. gratefullyacknowl- in these LMXBs is small or disrupted) which producesto the edgessupportfromNASA.Thisworkwassupportedinpartby power-lawcomponentmayionizethediskinXTEJ1751- 305 the Netherlands Organization for Scientific Research (NWO). toadegreewhichpreventsFeKαemission. Thisresearchhasmadeuseofthedataandresourcesobtained TheabsenceofH-likeandHe-likeresonanceemissionlines through the HEASARC on-line service, provided by NASA- like thoseobservedin 4U1626–67(Schulzetal. 2001)in the GSFC. REFERENCES Arnaud, K.A., 1996, Astronomical DataAnalysis Software andSystems V, Markwardt,C.B.,&Swank,J.H.,2002a,IAUCirc.7867 eds.G.JacobyandJ.Barnes,p17,ASPConf.Seriesvol.101 Markwardt,C.B.,&Swank,J.H.,2002a,IAUCirc.7870 Asai,K.,Dotani,T.,Nagase,F.,&Mitusda,K.,2000,ApJS,131,571 Markwardt,C.B.,Swank,J.H.,Strohmayer,T.E.,in’tZand,J.J.M,&Mar- Bhattacharya,D.,&vanenHeuvel,E.P.J.,1991,PhysRep.,203,1 shall,F.E.,2002,ApJ,575,L21 Bildsten,L.,2002,ApJ,inpress,astro-ph/0208164 Miller,J.M.,etal.,2002a,ATEL90 Bildsten,L.,&Chakrabarty,D.,2002,ApJ,557,292 Miller,J.M.,etal.,2002b,ATEL91 Campana,S.,etal.,2002,ApJ,inpress,astro-ph/0206376 Miller,J.M.,etal.,2002c,ApJ,570,L69 Chakrabarty,D.,1998,ApJ,492,342 Morrison,R.,&McCammon,D.,1983,ApJ,270,119 Chakrabarty,D.,&Morgan,E.H.,1998,Nature,394,346 Orlandini,M.,etal.,1998,ApJ,500,163L Ehle,M.,etal.,2002,IAUCirc.7872 Remillard,R.A.,Swank,J.,&Strohmayer,T.,2002,IAUCirc.7893 Galloway,D.K.,Chakrabarty,D.,Morgan,E.,&Remillard,R.,2002,ApJ,in Schulz,N.S.,Chakrabarty,D.,Marshall,H.L.,Canizares,C.R.,Lee,J.C.,& press,astro-ph/0206493 Houck,J.,2001,ApJ,563,941 Gierlinski,M.,Done,C.,&Barret,D.,2002,MNRAS,331,141 Stella,L,White,N.E.,&Priedhorsky,W.,1987,ApJ,312,L17 Heindl,W.A.,&Smith,D.A.,1998,ApJ,560,35L Strohmayer,T.E.,2001,AdSpR,28,511 den Herder, J. W., 2002, RGS-SRON-RP-CAL-01/006, available at Titarchuk,L.,1994,ApJ,434,313 http://xmm.vilspa.esa.es/ccf/documents/#general Verner,D.A.,Yakovlev,D.G.,Band,I.M.,&Trzhaskovskaya,M.B.,1993, Juett,A.M.,Psaltis,D.,&Chakrabarty,D.,2001,ApJ,560,L59 ADNDT,55,233 Juett,A.M.,Galloway,D.K.,&Chakrabarty,D.,2002,ApJ,submitted,astro- White,N.E.,Swank,J.H.,&Holt,S.S.,1983,ApJ,270,711 ph/0208543 Wilms,J.,Nowak,M.A.,Dove,J.B.,Fender,R.P.,&DiMatteo,T.,1999, Kallman, T.R., Angelini, L.,Boroson, B., &Cottam, J., 2002, ApJ,subm., ApJ,522,460 astro-ph/0209010 Wijnands,R.,&vanderKlis,M.,1998a,Nature,394,344 Kirsch, M., 2002, XMM-SOC-CAL-TN-0018, available at Wijnands,R.,&vanderKlis,M.,1998b,ApJ,507,63L http://xmm.vilspa.esa.es/ccf/documents/#general Yungelson,L.R.,Nelemans,G.,&vandenHeuvel,E.P.J.,2002,A&A,388, Lewin,W.H.G.,vanParadijs,J.,&TaamR.E.,1993,SSRv,62,223 546 J.M.Milleretal. 5 TABLE 1 EmissionLineUpper-Limits Helium-like Hydrogenic Broada Narrowb Broada Narrowb (10- 5phcm- 2s- 1) (10- 5phcm- 2s- 1) (10- 5phcm- 2s- 1) (10- 5phcm- 2s- 1) Ne 1.6 1.4 0.9 0.5 Mg 1.4 1.3 11 8.8 Si 29 26 3.7 3.5 NOTE.—The95%confidenceupperlimitsonhighlyionizedresonanceemissionlineswithin0.1Å oftheirexpectedrest-framewavelengthsintheRGSspectra ofXTEJ1751–305. aAssumingFWHM=3000km/s,similartotheDoppler-shiftedpairsin4U1626–67(Schulzetal. 2001). bTheupperlimitswerecalculated usingGaussiansofzerowidthtomodelfeaturesnarrowerthantheRGSresolution.LimitsonthestrengthofOlinesarenotconstrainingduetopoorstatisticsatlow energies. Theseupperlimitsaresimilartoorbelowthefluxesmeasuredfromthelinesdetectedin4U1626–67,indicatingaconstrainingsensitivitywasachievedin thisXMM-Newtonobservation.