X-Ray Timing of the Young Pulsar in 3C 58

0.1 MB·

by  Scott Ransom

X-Ray Timing of the Young Pulsar in 3C 58
Scott Ransom∗†, Fernando Camilo∗∗, Victoria Kaspi∗†, Patrick Slane‡, Bryan
Gaensler‡, Eric Gotthelf∗∗ and Stephen Murray‡
∗DepartmentofPhysics,McGillUniversity,Montreal,QCH3A2T8,Canada
4 †CenterforSpaceResearch,MassachusettsInstituteofTechnology,Cambridge,MA02139
0 ∗∗ColumbiaAstrophysicsLaboratory,ColumbiaUniversity,550West120thStreet,NewYork,NY10027
0 ‡Harvard-SmithsonianCenterforAstrophysics,60GardenSt,CambridgeMA,02138
2

n
Abstract. PSRJ0205+6449isayoungpulsarintheCrab-likepulsarwindnebula3C58whichisthoughttobearesultof
a
thehistoricalsupernovaSN1181CE.The65.7-mspulsaristhesecondmostenergeticoftheknownGalacticpulsarsandhas
J
beenshowntoberemarkablycoolforitsage,implyingnon-standardcoolingprocessesintheneutronstarcore.Wereport
8
onRXTEtimingobservations takenduringAO7andsupplemented bymonthlyradioobservations ofthepulsarmadewith

theGreenBankTelescope(GBT).Thetotaldurationcoveredwiththetimingsolutionsis450days.Wemeasureveryhigh
1 levelsoftimingnoisefromthesourceandfindevidencefora“giant”glitchofmagnitudeD n /n ∼1×10−6thatoccurredin
v
2002October.Wehavealsomeasuredthephase-resolvedspectraofthepulsationsandfindthemtobesurprisinglyhard,with
27 photonindicesG =0.84−+00..1056forthemainpulseandG =1.0−+00..34fortheinterpulseassuminganabsorbedpower-lawmodel.
1
1
0 1. INTRODUCTION andhasbeenspinningdownwiththecanonicalbraking
4
index (due to magnetic dipole braking) of n= 3, then
0
PSR J0205+6449was discoveredin the heartof 3C 58 it must have been born spinning relatively slowly with
/
h in Chandra HRC data and immediately confirmed in P ∼60ms. Interestingly, the 65-ms PSR J1811−1925
0
p archival RXTE data by Murray et al. [1]. Camilo et al. in G11.2−0.5 seems to be an identical twin in many
-
o [2]soonthereafterfoundextremelyfaintradiopulsations respects(seeRobertsetal.,theseproceedings).
r from the pulsar using the new Green Bank Telescope
t
s (GBT). Chandra ACIS-S observations taken by Slane,
a Helfand, & Murray [3] showed that the temperature of 2. OBSERVATIONSAND DATA
:
v theneutronstar(NS)issignificantlybelowthatpredicted PREPARATION
i bystandardNScoolingmodelsgiventheveryyoungage
X
ofthepulsar(∼820yrs).Measurementofthespinperiod
We analyzed all available RXTE PCA data on 3C 58,
ar (65.686ms) and spin-down rate (P˙ =1.94×10−13s/s) which totaled more than ∼410ks of on-source time.
ofthepulsarhaveallowedtheestimationofseveralkey ThedatacamefromOBSIDsP20259(∼20ks),P60130
parametersusing the canonicalexpressionsfrom pulsar (∼100ks), and P70089 (∼300ks), and were taken in
theory:
GoodXenonmodewhichprovidedeventswithtimeres-
• Spin-downenergy olutionsof ∼122m sand256energychannelsintheen-
ergy range ∼2−60keV. Most observations were taken
E˙ =4p 2IP˙/P3=2.7×1037erg/s (1) using3PCUs.
WeprocessedtheRXTEdatausingstandardFTOOLs1
• Surfacemagneticfieldstrength
inordertofilteroutnon-Level1eventsorthoserecorded
outside of Good Time Intervals and to converttheir ar-
3c3IPP˙
B= =3.6×1012Gauss (2) rival times to the Solar System barycenter (using the
r 8p 2R6
DE200 planetary ephemeris [4]). We determined the
• Characteristicage X-ray pulse times-of-arrival (TOAs) with a maximum-
t =P/(2P˙)=5400yr (3)
Each of these parameters is typical for young pulsars. 1 http://heasarc.gsfc.nasa.gov/docs/software/
ftools/ftools_menu.html
Oneunusualpointisthatifthepulsarreallyis822yrsold
likelihoodtechniquethatwedevelopedwhichassumeda
two-GaussianmodeloftheX-raypulseprofile.Asignal-
to-noise ratio optimization procedure allowed us to de-
termine that the best energy range to use for the tim-
ing analysis is ∼2−16keV (energy channels 4−39 in-
clusive),andsowefilteredouteventsoutsideofthisen-
ergyrange(seetheprofileinFigure1).Asingletime-of-
arrival(TOA)typicallycomprisedoneortwo RXTEor-
bitsdependingonthebackgroundlevelsandthenumber
ofPCUson.Sincetheobservationsconsistedofbetween
4−6 orbits, we measured 3−4 TOAs each month with
precisionsof100−200m seach.
The GBT data were taken with the Berkeley-Caltech
Pulsar Machine (BCPM)2 backend at either 820MHz
or1400MHz.TheBCPMisananalog/digitalfilterbank FIGURE 2. A “Giant” Glitch. The thick lines correspond
which samples each of 2×96 channels using 4-bits at tothephase-coherenttimingmodelsdescribedin§3.Thethin
flexiblesamplingratesandchannelbandwidths[5].The grey lines correspond to the average spin-down as measured
data were recorded using 134MHz of bandwidth and duringthe∼210daysbeforeandafterMJD52555.Thefrac-
o5f0bmasnsdawmidptlhesafnodrt7h2em1s40sa0mMpHlezsofobrsethrvea8ti2o0nsMoHr5z0oMbsHerz- tliuotnioanlsdiisffDerne/nnce∼in1×sp1in0−f6re,qwuhenicchyibsectowmeepnartahbeletwinomtimaginnigtusdoe-
totheglitchesfromtheolderandalso“cool”Velapulsar.
vations.Typicalintegrationslastedbetween4−8hrs.We
de-dispersed and folded all of the data using the pulsar
analysispackagePRESTO[6].Eachobservationyielded 3. TIMING
2−3TOAswhichwemeasuredbycorrelating(inthefre-
quencydomain)thefoldedpulseprofilewithaGaussian
The timing behavior and therefore the spin-down rate
of fractionalwidth 0.04 in phase. The typical precision
of PSR J0205+6449is extremely noisy. Determining a
oftheradioTOAswas200−400m s.
phase-connected timing solution would have been im-
possibleusingtheRXTEorGBTdataalone.Thankfully,
thecadenceofthemonthlyobservationsintheradioand
X-ray bandswas sufficiently differentsuch that the ob-
servationscomplementedeachotherandusuallyallowed
unambiguouspulsenumberingbetweenobservations.
WeappliedtimecorrectionstotransformalltheTOAs
toUTCattheSolarSystemBarycenterusingfaxbary
and TEMPO3 for the RXTE and GBT data respectively.
ThisabsolutetimingshowsthattheX-raymainpulseis
delayedrelativetotheradiopulse(Camiloetal.,inprep).
WeperformedalltimingmodelfitswithTEMPO.
Ourdatashowthata“giant”glitch(D n /n ∼1×10−6)
occurred near MJD 52555 (see Figure 2). The timing
solution prior to the glitch consists of a 6 parameter fit
(spinfrequencyn plus5frequencyderivatives)resulting
in RMS timing residuals of ∼0.6ms. The post-glitch
timing solution consists of the spin frequency plus 2
FIGURE1. RXTEPulseProfile.The2−16keVpulsepro-
frequencyderivativesresultinginRMStimingresiduals
fileforPSRJ0205+6449forRXTEOBSIDsP20259,P60130,
and P70089. The total integration time included over 400ks of∼1.2ms.Additionalfrequencyderivativesinthepost-
with3PCUsandover100kswitheither4or5PCUs.Thesolid glitchsolutionimprovetheresidualsbutarenotrequired
greybandsshowtheportionsofthepulseusedforthephase- forphaseconnection4.TheresidualsfromtheGBT and
resolvedspectroscopydescribedin§4withthepulseontheleft RXTEforbothtimingsolutionsareshowninFigure3.
andtheinterpulseontheright.Thelightlyhashedregionswere
usedtocomputethespectralbackground.
3 http://pulsar.princeton.edu/tempo
4 While we believe that the post-glitch data is phase-connected, the
2 http://www.gb.nrao.edu/~dbacker sparsesamplingandhighlevelsoftimingnoisemakeitpossiblethat
wehavemiscountedpulsesbetweenoneormoreoftheobservations.
FIGURE 3. Post-fit Timing Residuals. Timing residuals for pre-glitch (left) and post-glitch (right) timing solutions. TOA
precision isabout 3× better withRXTE (the “X”s) than with the GBT at 820MHz (the circles) at 1400MHz (the squares) per
unitoftelescopetime.Theobvioussystematictrendsaretheresultofverylargelevelsoftimingnoisefromthepulsar.Welost
phaseconnectionnearMJD52555duetheglitchshowninFigure2.Inboththepre-andpost-glitchsolutions,themeasuredbraking
indexismuchgreaterthanthecanonicalvalueof3,whichistypicaloftimingnoiseorglitchrecovery.
4. PHASE-RESOLVED SPECTROSCOPY 5. FUTUREPROSPECTS
Figure4showstheresultsofabsorbedpower-lawfitsto With its young age, low temperature, very noisy spin-
thephase-resolvedspectraofthepulsar’smainpulseand down,highE˙,hardpulsedspectrum,andlargeglitches,
interpulse as determined using the FTOOL fasebin. the pulsar at the center of 3C 58 is certainly an object
We analyzed events from each PCU independently for worthwatchinginthefuture.Istheextremelyhighlevel
each OBSID, combined these for each AO, and then of timing noise from the pulsar somehow related to its
combined the three AOs together. We fit the spectra “cool”temperature?Arebothofthesestrangeproperties
withxspecaftersubtractingtheaverageoff-pulseback- relatedto a morefundamentalcharacteristicoftheneu-
ground as shown in the hatched region in Figure 1 and tronstarsuchasitsmass?Weareoptimisticthatcontin-
freezing the value of N =3.62×1021cm−2 based on uedmonitoringofthisunusualsourcewilltellussome-
H
the recent Chandra ACIS observation of 3C 58 (Slane thing about the interior structure and internal dynamics
et al., in prep).Fits madeto the summedPCU data and of neutron stars. In addition, we note that the narrow
simultaneously to the 5 PCUs independentlywere con- pulse profile and hard spectrum of the pulsed emission
sistentwithinthestatisticalerrorstooneanother. fromJ0205+6449makeitanidealandpossiblyunique
The measured photon indices G = 0.84+0.06 for the target for INTEGRAL and other upcoming hard X-ray
−0.15
mainpulseandG =1.0+0.4fortheinterpulsearesurpris- missions.
−0.3
inglyhard.Thesevaluesaredifficulttoreconcilewiththe
G =1.73±0.07 photon index measured with Chandra
forthecentralpointsource[3],althoughsignificantneb- REFERENCES
ular emission within the selection region could explain
some of the spectral softening. Recent results from the 1. Murray,S.S.,Slane,P.O.,Seward,F.D.,Ransom,S.M.,
350ks Chandra observationof 3C 58 place an approxi- andGaensler,B.M.,ApJ,568,226–231(2002).
mateupperlimitof1.5×1014ergcm−2s−1 foranyhard 2. Camilo,F.,Stairs,I.H.,Lorimer,D.R.,Backer,D.C.,
Ransom,S.M.,Klein,B.,Wielebinski,R.,Kramer,M.,
componentin the range 0.5−10keV. But our fits to the
McLaughlin,M.A.,Arzoumanian,Z.,andMüller,P.,ApJ,
mainpulsealonepredictfluxesinthesameenergyband 571,L41–L44(2002).
atleastonehundredtimeslarger.Wenotethatourmea- 3. Slane,P.O.,Helfand,D.J.,andMurray,S.S.,ApJ,571,
surementsareconsistentwithintheerrorswiththephase- L45–L49(2002).
resolved spectral fits of Gotthelf [7] (G = 1.11±0.34 4. Standish,E.M.,A&A,114,297–302(1982).
5. Backer,D.C.,Dexter,M.R.,Zepka,A.,Ng,D.,Werthimer,
for the pulsed emission) based on the RXTE data from
D.J.,Ray,P.S.,andFoster,R.S.,PASP,109,61–68(1997).
P60130.
6. Ransom,S.M.,NewSearchTechniquesforBinaryPulsars,
Ph.D.thesis,HarvardUniversity(2001).
7. Gotthelf,E.V.,ApJ,591,361–365(2003).
FIGURE 4. Phase-Resolved Spectroscopy. The spectra of the main X-ray pulse (top) and interpulse (bottom) as fit from
3−16keVwithanabsorbedpower-lawmodel.TheNH wasfixedat3.62×1021cm−2 basedontherecent350ksChandraACIS
observationof3C58(Slaneetal.,inprep).Thebestfitphotonindicesaresurprisinglyhard:G =0.84−+00..1056forthemainpulseand
G =1.0+−00..43fortheinterpulse.