The 2009 outburst of magnetar 1E 1547-5408: Persistent radiative and burst properties
aa r X i v : . [ a s t r o - ph . H E ] O c t Accepted June 27, 2011
Preprint typeset using L A TEX style emulateapj v. 11/10/09
THE 2009 OUTBURST OF MAGNETAR 1E 1547 − P. Scholz & V. M. Kaspi Accepted June 27, 2011
ABSTRACTThe magnetar 1E 1547 − Swift
X-ray Telescope(XRT). We find that the 1–10 keV flux increased by a factor of ∼
500 and hardened significantly,peaking ∼ Subject headings: stars: neutron — pulsars: individual (1E 1547 − − − INTRODUCTION
Originally classified as two distinct types of ob-jects, anomalous X-ray pulsars (AXPs) and soft gammarepeaters (SGRs) are now generally accepted to behighly magnetized neutron stars, magnetars, with mag-netic fields
B > − G (for reviews, seeWoods & Thompson 2006; Mereghetti 2008). Magne-tars differ from ‘normal’ rotation-powered pulsars inthat their X-ray luminosities exceed the rate of en-ergy released from their spin-down. They are believedto be powered by the decay of their magnetic fields.Another common property of magnetars is that theyexhibit episodes of violent bursting activity. At onepoint thought to be only observed in SGRs, X-raybursts in AXPs were first detected from 1E 1048.1 − − Einstein satellite (Lamb & Markert 1981).It was identified only recently as a magnetar byGelfand & Gaensler (2007) based on its X-ray spectrumand infrared flux, as well as its possible association withthe supernova remnant G327.24 − ∼ . An XMM-Newton observation in 2007 showed asignificant flux enhancement from what was previouslymeasured (Halpern et al. 2008), thus revealing that anX-ray outburst event had occured between 2006 and Department of Physics, Rutherford Physics Building, McGillUniversity, 3600 University Street, Montreal, Quebec, H3A 2T8,Canada ∼ pulsar/magnetar/main.html Swift (Israel et al. 2010) and
Fermi (Kaneko et al. 2010) satellites. 1E 1547 − Swift , INTEGRAL ,and
Fermi . Tiengo et al. (2010) report the appearanceof dust scattering rings around 1E 1547 − XMM-Newton and
Swift . The energy released by the event causing therings was estimated to be 10 − ergs. From theserings, a distance to the source was determined to be 3.9kpc. Ng et al. (2011) report on Chandra observationsbeginning ∼ Swift .The burst study will focus primarily on bursts from the2009 outburst since the number ( ∼ Swift (Israel et al. 2010). A summary of the observa-tions is presented in §
2. The analysis performed on thedata as well as the results for the persistent emission arereported in § § §
4. Finally,our findings are summarized in § OBSERVATIONS
The data presented in this paper were obtained us-ing the X-Ray Telescope (XRT) (Burrows et al. 2005)on the
Swift satellite. The XRT uses Wolter-I opticsand an
XMM-Newton /EPIC MOS CCD detector to pro-vide rapid imaging and spectra of X-ray transients in the0.5–10 keV energy range. During the 2008 outburst of1E 1547 − Swift
Burst Alert Telescope (BAT)triggered at 09:28:08 UT on 2008 October 3 and
Swift promptly slewed to the source position. The XRT begantaking observations 99 s after the trigger. On 2009 Jan-uary 22, the BAT triggered at 01:32:41 UT and the firstXRT observation began ∼
50 min later. Table 1 shows asummary of the observations following the 2009 outburstevent as well as two observations preceding the outburst.The observations following the 2008 outburst were previ-ously presented by Israel et al. (2010); here we focus onthe 2009 event.Cleaned data products in both windowed-timing (WT)and photon-counting (PC) modes were obtained fromthe HEASARC
Swift
Archive. Data were reduced tothe barycentre using the position of 1E 1547 − h m . s − ◦ ′ . ′′ − ANALYSIS AND RESULTS
Persistent flux evolution
In order to investigate the behavior of the persistentflux of 1E 1547 − xselect . Spectralfitting was performed using the XSPEC package ver-sion 12.6. The spectra were grouped with a minimum of20 counts per energy bin. Ancillary response files werecreated using the FTOOL xrtmkarf and the standardspectral redistribution matrices from the Swift
CALDB http://xspec.gfsc.nasa.gov were used. Bursts were then removed from the observa-tions using the method described in § ∼ Swift
XRT. For these observations, the radiative prop-erties of the source cannot be simply disentangled fromthe delayed emission from the dust scattering rings (A.Tiengo & P. Esposito, private communication). Appro-priate modelling of this is currently under investigation(Tiengo et al., in preparation) but beyond the scope ofour work. Following the first day after the outburst, thedust scattering rings were outside of the extraction re-gion and so those observations are not affected by dustscattering.The spectra were fitted with a photoelectrically ab-sorbed blackbody with an added power-law component.To determine N H , we fit observations having an exposuretime greater than 3 ks jointly with a single N H . The pa-rameters kT , Γ and their normalizations were allowed tovary in these fits. This resulted in N H = 3 . × cm − which is consistent with the value of 3 . +0 . − . × cm − reported in Gelfand & Gaensler (2007), though issomewhat lower than the value of 4 . × cm − measured by Ng et al. (2011). All the Swift observationswere subsequently fit with the column density fixed toour best-fit value. The 2008 data have been previouslypresented by Israel et al. (2010), whose results are gen-erally consistent with those of our analysis for the sametime period, so they will not be presented here. Figure1 shows the results of fitting of the observations follow-ing the 2009 outburst. The split observations, affectedby dust scattering, are the first 10 data points in Fig-ure 1 following the trigger, and so should be regardedwith caution. The two observations preceding the out-burst were fit with only a blackbody and no power-lawcomponent. This is because the power-law index couldnot be constrained due to a paucity of counts. The fitsin general were excellent; the goodness-of-fit statistic χ ν ranged between 0.61 and 1.48 with a mean of 1.06.The peak of the persistent emission occured ∼ − kT . After the peak, the spectral index soft-ened and kT fell as the flux dropped. To characterize theflux decay of the outburst, a power law decay was fit tothe data following the first day of the outburst. The firstday was not included as the source decay is superimposedby delayed emission from the dust scattering rings. Thepower law decay is descibed by F = A ( t − t ) α , where F is the unabsorbed flux, A is the normalization, α is thepower-law index, and t is the time of the BAT trigger.The decay was described by a power-law with an indexof − . ± .
02. Although the χ ν /ν of the fit was 4.1/23,a power-law model fit the data much better than an ex-ponential decay. To determine the blackbody radius ofthe emitting region as shown in Figure 1, a distance of3 . ± .
07 kpc from Tiengo et al. (2010) was assumed.To measure pulsed fractions and fluxes, the burst-removed time series (see § RXTE observations of 1E 1547 − ν = 0 . ν = − . × − s − at refer-ence epoch MJD 54854.0 or 2009 January 23. The pulsedflux was calculated using an RMS method according tothe formula in Dib et al. (2008), with 7 harmonics. Thepulsed fraction was determined by dividing the pulsedflux by the phase-averaged flux. A pulsed flux and frac-tion were measured for each WT observation with anexposure time > ∼
40 days, the pulsed fraction hadnot yet recovered to its pre-burst level of ∼ Swift data confirm the evolution that is observed inthe
RXTE data which are presented briefly in Ng et al.(2011) and in detail in Dib et al. (2011). The pulsed fluxenhancement ∼
11 days following the initial trigger of the2008 event, although not noted by Israel et al. (2010), isclearly present in the
Swift data. Thus, the 2009 eventshowed a much smaller increase in pulsed flux than didthe 2008 event, while the opposite is true of the totalflux.Figure 3 shows an anti-correlation between pulsed frac-tion and unabsorbed flux. It includes observations fromboth the 2008 and 2009 outbursts. Ng et al. (2011)present this trend using
Chandra and
XMM data as wellas the burst-removed
Swift data presented in this paper.
X-Ray Bursts
Bursts were identified in a manner similar to that de-scribed in Gavriil et al. (2004). The event lists werebinned at a 1/16 s time resolution and a mean numberof counts per bin was calculated for each Good TimingInterval (GTI). The number of counts in each bin, n i ,was compared to the GTI mean counts, λ , according tothe probability P i of n i occuring randomly, P i = λ n i e − λ n i ! . (1)Time bins that had P i ≤ . /N , where N is the totalnumber of time bins in the GTI being searched, wereidentified as part of a burst. Since the mean number ofcounts in a GTI can be overestimated due to contam-ination from bursts in other time bins, the procedureabove was repeated iteratively, each time removing thebins that were identified as containing bursts, until nofurther bursts were identified. The above procedure wasthen repeated for 1/32 s and 1/64 s time resolutions toimprove sensitivity to bursts of different durations.Since several bins identified with bursts can be part ofthe same burst, a burst was defined by its peak. The peakof a burst was determined by first finding the minimumtime to accumulate 10 counts, using unbinned event data.The midpoint of the time spanned by these 10 counts wasdefined as the peak. A search for a peak was done within0.5 s on each side of an identified bin. This definition ofburst peak is independent of binning and, for bursts with durations shorter than 1 s, will merge all of the identifiedbins into a single burst. Bursts within 1 s of each otherthat were merged into a single burst were identified whenselecting the background (see below) and their propertieswere measured separately.Once identified, to remove the bursts from the eventlists, the lists were divided into full periods of the pulsar,starting with the first event of the observation as a refer-ence point. The periods that contained bursts were iden-tified and all counts which arrived in that interval wereremoved from the event list. This was done to ensureequal exposure to all pulse phases in the pulse profile. Burst Statistics
For each burst, a fluence, T , rise time ( t r ), and falltime ( t f ) were measured with the same analysis as inGavriil et al. (2004), using the unbinned event data. Thefluence was determined by first measuring a backgroundcount rate in hand-picked regions on either side of theburst. The background region by default was between 1 sand 2 s from the burst peak to either side of the burst, butwas manually adjusted for most bursts in order to avoidcontamination from other nearby bursts. The cumulativebackground-subtracted counts were then fit with a stepfunction, using data point from the hand-picked back-ground region. The height of the step function in countscorresponds to the fluence of the burst. The T dura-tion of the burst is the time between when 5% and 95%of the fluence has been accumulated. The burst rise andfall times were determined using a maximum likelihoodfit to a piecewise function with an exponential rise andan exponential decay. Peak fluxes in counts per secondwere determined by passing a 62.5 ms boxcar integratorthrough a 250 ms interval (4 boxcar widths) centered onthe burst peak. The highest count rate measured by theintegrator was defined to be the peak flux.In total, for the 2009 outburst, 424 bursts were iden-tified in 86 ks of observations from 2009 January 22 to2009 September 30 using 1/16 s time resolution. Thir-teen additional bursts were identfied using the 1/32 sand 1/64 s time resolutions for a total of 437 bursts. Ofthose identified, 34 had too few counts to permit a reli-able measure of fluence, and for 32, the exponential riseand decay were not successfully fit. For 64 of the bursts,properties could not be measured because they were tooclose to another burst to allow a reliable background es-timate between them. Four bursts were too close to theedge of a GTI to calculate a background count rate. Insummary, 303 bursts could be fully analysed and onlythese 2009 bursts will be considered henceforth. Exam-ples of different bursts are shown in Figure 4. Only twobursts were found in the XRT observations of the 2008event; their properties were similar to those during the2009 outburst.Figure 5 shows the distributions of the burst proper-ties, grouped in logarithmic bins. The T , rise time, falltime, and the ratio of rise to fall time distributions havebeen fit with log-normal distributions using maximumlikelihood fitting. The T distribution has a mean of305 ms and a range for one standard deviation of 140 -662 ms. For the rise time distribution, we find a mean of39 ms and a range for one standard deviation of 14 - 109ms. For the distribution of fall times, we find a meanof 66 ms and a range of 24 - 182 ms for one standarddeviation. The mean of the t r /t f distribution is 0.59with a range for one standard deviation of 0.21 - 1.66. Asummary of the measured quantities is provided in Ta-ble 2. There we also provide these quantities, when avail-able, for the three other sources for which such statisticalanalyses have been done. We note in particular that theaverage burst duration for 1E 1547 − § − . ± .
1. For the peak flux distribution,the power-law index is − . ± .
12. This index is quiteshallow and, despite the omission of the first two pointsin the fit, may still be affected by the bias in the burstsearch.For other AXPs, there is evidence that bursts tend toarrive on pulse (e.g. Gavriil et al. 2002, 2004). For 1E1547 − − . < t r /t f <
2. Bursts with t f greater than 2 t r were defined as slow-fall bursts. Of the303 well measured bursts, 158 were classified as symmet-ric and 116 as slow falls. The remainder, 29 bursts, werethose with rise times greater than twice their fall times.Curiously, the folded slow-fall burst counts exhibit amuch stronger pulse than do the folded symmetric burstcounts, with a χ ν = 43 for the null hypothesis, com-pared to χ ν = 6 for the folded symmetric bursts. Thissymmetric/slow-fall definition is somewhat arbitrary andthese two classes of bursts are by no means distinct pop-ulations. We use this distinction only to demonstratethat the more symmetric bursts tend to be more pulsedthan those that are less symmetric. Burst Spectroscopy
Spectra within the T interval of each burst peak wereextracted and grouped with 20 counts per bin. Back-ground spectra were taken from 1 min on either side ofthe burst peak extracted from the burst-removed eventdata. The 46 bursts with fluences over 200 background-subtracted counts were fitted with a photoelectrically ab-sorbed power law with N H fixed to 3 . × cm − as measured from the fit to the persistent emission (see § .
17, with a standard deviation of 0.33, where N ( E ) ∝ E − Γ . Fitting with more complicated modelswas attempted, but parameters could not be successfullyconstrained because of the low number of counts. Asshown in Table 2, the average Γ we measure is signifi-cantly harder than that measured in the only other mag-netar outburst for which the value is reported. This isdiscussed further in § − Swift data may be due to the limited spectral rangeof the XRT (0.5–10 keV). However, an anti-correlationbetween power-law index and average flux over T fromthe spectral fits was observed (Fig. 8). This is consistentwith a correlation between the hardness and the magni-tude of a burst.Spectral features in magnetar bursts have beenreported for some sources (Strohmayer & Ibrahim2000; Ibrahim et al. 2002; Gavriil et al. 2002;Ibrahim et al. 2003; Woods et al. 2005; Gavriil et al.2009; Kumar & Safi-Harb 2010), although most ofthe features have been discovered above 10 keV. Theindividual burst spectra for 1E 1547 − addspec . Spectralbins were grouped with a minimum of 20 counts perbin. The average spectrum was fit with a blackbody,a power law, a blackbody with an added power law,and a Comptonized blackbody; all four models werephotoelectrically absorbed. All of the models gaveacceptable fits with χ ν ∼
1; thus, no significant spectralfeatures were observed in any of the residuals.Fluences and peak fluxes in counts and counts per sec-ond, respectively, were converted into cgs units using theresults of the spectral fits. The fluxes from the power-law fits were multiplied by the burst durations and com-pared to the fluences measured in counts. The propor-tionality constant between the two was determined to be2 . × − ergs cm − counts − . In using this singleconversions factor between counts and ergs cm − we areeffectively assuming an average spectral model for thebursts which is not generally true. The conversion fromcounts to cgs units in reality is different from burst toburst depending on their spectra. Values in cgs units,using this single conversion factor, are indicated on thetop axes of the fluence and peak flux distributions inFigure 6. DISCUSSION
We have presented
Swift
XRT observations of the 2008and 2009 outburst events of 1E 1547 − Swift followingthe 2009 event. Noteworthy results we have found in-clude a correlation between burst flux and hardness. Wealso note that burst counts tend to contribute very sig-nificantly to the pulsed fraction, even if the phases ofburst peaks are randomly distributed in pulse phase.Next we discuss our results in the context of the mag-netar model, as well as in comparison with those for othersimilar outbursts.
Persistent Flux
Figure 1 shows that as the X-ray flux increased, theblackbody temperature increased and the power-law in-dex decreased during the 2009 outburst. Such a hard-ness/flux correlation was seen also in the 2008 outburst(Israel et al. 2010) and ubiquitously in other magne-tar outbursts (e.g. Kaspi et al. 2003; Woods et al. 2004;Israel et al. 2007). In the case of 1E 1547 − kT and Γ as shown in Fig-ure 1 on the first day after the outburst with caution, asthey are very likely biased by dust scattering. The hard-ness/flux correlation for this outburst on the day of thetrigger is likely to be robust, since dust scattering softensphotons. However, note that if the dust scattering ringsare caused by a burst that is much harder than, and sep-arate from, the persistent emission, the dust scatteringcould effectively cause the overall emission to harden inrelation to the true persistent emission.The thermal emission of magnetars, in the twistedmagnetosphere model (Thompson et al. 2002), has asits origin heating from within the star, due to the de-cay of the strong internal magnetic field. The resultingthermal surface photons are thought to be scattered bycurrents in the atmosphere, resulting in a Comptonized blackbody-like spectrum which is often modelled with ablackbody plus a power law (Lyutikov & Gavriil 2006).The magnetospheric currents are present due to ‘twists’in the field structure, either global (Thompson et al.2002), or, more likely, in localized regions (Beloborodov2009). In either case, the current strength, hence de-gree of scattering, increases with increasing twist magni-tude. Moreover, return currents provide an additional,external source of surface heating in addition to the in-ternal source. The increase in flux during a magnetaroutburst is thus theorized to be caused by an internalheat-releasing event that may significantly increase thesurface temperature (see, for example, ¨Ozel & Guver2007), further twist the magnetospheric field, and in-crease the external return-current heating. Thus, a corre-lation between hardness and flux is generically expectedin the twisted-magnetosphere model (Lyutikov & Gavriil2006), in agreement with observations of 1E 1547 − − flux command and the hardness was calculated fromtheir ratio. For references that did not report a 2–10keV unabsorbed flux, one was determined using XSPECusing the reported model paramenters. As is clear fromthe Figure, an overall trend is apparent, although thereis significant variation from source to source. For ex-ample, for similar flux enhancements over the quies-cent level, 1E 2259+586 became much harder than did1E 1048.1 − − − Chandra and
XMM-Newton data in Ng et al. (2011), who suggest it could be due toan increase in the emitting area of a thermal hot spot,which would cause the emission to be observable during agreater portion of the phase of the pulsar. We note thatboth correlations (Gotthelf et al. 2004; Zhu et al. 2008)and anti-correlations (Israel et al. 2007; Tam et al. 2008)between pulsed fraction and flux have been observed inmagnetar outbursts. Different behaviors can result de-pending on the location of the emission region and theviewing geometry. Although detailed modeling of spec-tral, pulse morphology and pulsed flux changes duringmagnetar outbursts is beyond the scope of this paper,recent attempts at unified modelling of magnetar sur-face and magnetosphere emission geometries and emis-sion mechanisms (Albano et al. 2010) could in principleuse data like ours to constrain the hot spot location andsize. This seems to us to be a good future avenue forinvestigation.Interestingly, Kaneko et al. (2010) report a ∼ ∼
100 keV as observed by
Fermi
GBM ∼ ∼
100 keV for an SGR .However, AXPs have previously been shown to exhibitpulsations at ∼
100 keV with pulsed fractions as highas 100% (e.g. Kuiper et al. 2006). Thus it is possiblethat the measured 55% actually decreased from a higherpulsed fraction leading up to the outburst event, i.e. the ∼
100 keV pulsed fraction may have behaved similarly tothat in the 1–10 keV band. Future hard X-ray telescopeswith focusing optics, such as
NuSTAR , will allow mucheasier measurements of the pulsed fraction of magnetarsat high X-ray energies.
Bursts
Previous studies of SGR bursts have shown that theirenergies, and thus fluences, follow a power-law distribu-tion dN/dE ∝ E − α with α equal to ∼ − − − . ± . dN/dE ∝ E − . . Gavriil et al. (2004) find an α of 1 . ± . INTEGRAL observations whichcover an energy range of >
80 keV, show different burstproperties from those determined using the
Swift
XRTobservations. Savchenko et al. (2010) report a 68-msmean duration derived from a log-normal distributionwith a scatter of 30 - 155 ms. This is much shorter thanthe 305-ms duration determined in this work (see Ta-ble 2). This discrepency may be due to the differencein energy coverage; perhaps bursts have different mor-phologies at different energies. However, the definitionof duration used by Savchenko et al. (2010) differs fromthe T definition used here. Savchenko et al. (2010) de-fine the burst duration as the time between the moment Note that Kaneko et al. (2010) use the designation SGR1550 − − when the count rate rises above 5 σ to when the countrate drops below 3 σ . When applying their definition ofduration to the bursts identified in our study, we find,for a log-normal distribution, a mean duration of 101 msand a range for one standard deviation of 59 - 173 ms,closer to but still somewhat longer than their measure-ment, suggesting a possible energy dependence of burstduration. We do not, however, detect any significantdifference in the durations measured using 0.5–10 keVcounts with those measured using 2–10 keV counts.The properties of bursts from the 2009 outburst eventof 1E 1547 − Swift was notobserving 1E 1547 − − − RXTE . The larger collecting area of
RXTE allows it todetect bursts that are fainter than those
Swift can de-tect. If faint, short bursts are missed by the XRT, themean burst duration (as well as t r and t f ) may be over-estimated. Also, the energy range of RXTE (2–60 keV)probes higher energies and so differences may be due tothe energy dependance of burst properties. A detailedstatistical study of 1E 1547 − RXTE isneeded to clarify this point.Although
Swift
XRT is not ideal for probing the spec-tra of magnetar bursts, which have significant flux above10 keV, we were still able to draw the following con-clusions from our analysis. While we do not observea hardness-fluence correlation for 1E 1547 − INTEGRAL data of the2009 outburst, find a correlation between burst hardnessand count rate. Their hardness ratio is defined as theratio between the Anti-Coincidence Shield (ACS) flux,which is sensitive to photons above 80 keV, and a 20-60keV flux from the ISGRI instrument. This is also con-sistent with a correlation between hardness and burstmagnitude for 1E 1547 − − Swift are much harderthan those from 1E 2259+586. In light of the observedhardness-fluence correlation in AXP bursts, this is nota suprising result. The 28 most fluent bursts from1E 2259+586 for which spectral indices were measured,have fluences of ∼ − − − erg cm − (Gavriil et al.2004). This is to be compared with the 46 most fluentbursts for which Γ was measured here for 1E 1547 − ∼ − − − erg cm − .This may account for the harder average spectral indexfor bursts from 1E 1547 − − Swift
XRT could be classified as Type B. However, the twobursts with long pulsating tails found in
INTEGRAL data by Savchenko et al. (2010) and Mereghetti et al.(2009) have typical Type B morphology. For the burstsin this work, over half of the bursts were classified assymmetric, which nominally corresponds to Type A (seetop-right panel of Figure 4 for example). About 40% ofthe bursts were classified as slow-fall bursts, which areactually closer to Type A in morphology than Type B,although they are not very symmetric. Thus, the burstsfrom 1E 1547 − Swift -detected burstswas ∼ × erg in the 1-10 keV range. This ismuch lower than the 1-10 keV energy released from thepersistent emission between 2009 January 22 and 2009September 30 of ∼ × erg. For reference, the en-ergy released from the spin-down in that same period is ∼ × erg. Woods et al. (2004) find that for SGRs,the energy released in the bursts is higher than that re-leased in the persistent emission, but for 1E 2259+586,the opposite is true. In this regard, the 2009 outburst of1E 1547 − CONCLUSIONS
We have presented an analysis of the persistentradiative evolution of the 2009 January outburst of1E 1547 − Swift
XRT observations. We foundthat ∼ ∼ × − ergs cm − s − , an increase of more than 500times the quiescent flux leading up to the outburst. Thisflux evolution is not due solely to the source; in the firstday, there is also emission from dust scattering rings thatis delayed emission from an energtic event near the onsetof the outburst. There was significant spectral harden-ing at the outburst as seen in other magnetar outbursts.Note that the absence of spectral variation reported byNg et al. (2011) is consistent with our results, as theymissed the bulk of the spectal changes which occured inthe first day of the outburst and their Chandra observa-tions did not begin until the next day. The pulsed frac-tion showed an anti-correlation with the phase-averagedflux for both the previous 2008 and 2009 outbursts, withboth sets of data following the same trend. We have com-piled data from this and five other magnetar outburstsfor four different sources and find a generic X-ray hard-ness/flux correlation overall, but with no clear universalquantitative relationship between the two.We have also presented a detailed statistical analy-sis of the several hundred bursts detected during the2009 event. The bursts do not easily fall into theType A/Type B classification put forth by Woods et al.(2005). We found that the peaks of the bursts were ran-domly distributed in pulse phase, but that when the in-dividual photon counts were folded at the pulsar emphe-meris, a clear pulse was present. This phase dependenceis stronger for those bursts with longer decays than risetimes than for those bursts that are symmetric. We alsoreport a correlation between burst hardness and burstflux.In many ways, these observations yield more questionsthan answers. The range of observed phenomenologyin magnetar outbursts seems to increase with eachevent, with few overall trends to assist in constrainingmodels emerging. Nevertheless given the paucity ofevents studied in detail, we remain hopeful that throughperserverence, eventually physical insight will emerge.The fast response of telescopes like
Swift is crucial tothis endeavor. For 1E 1547 − − Swift
XRT data. Apreliminary comparison of their results with ours showsgeneral agreement. In particular their reported overallsource behavior, namely spectral hardening correlatedwith flux, as well as pulsed fraction anti-correlated withflux and blackbody radius, are in broad agreement withour results. We further note that they omitted reportingon data taken with
Swift
XRT in the 24 hrs followingthe 2009 trigger, due to the difficulty in handling thedust-scattered emission (P. Esposito, A. Tiengo, privatecommunication).We are grateful to P. Esposito and A. Tiengo for shar-ing with us their preliminary work on modelling of thedust-scattered emission on the first day following the on-set of the 2009 outburst. We thank A. Archibald, P.Lazarus, C.-Y. Ng, and S. Olausen for useful discus-sions. We thank R. Dib for providing an ephemerisand pulsed count rate measurements from
RXTE . Wethank P. Woods for helpful comments. V.M.K. holdsthe Lorne Trottier Chair in Astrophysics and Cosmol-ogy and a Canadian Research Chair in ObservationalAstrophysics. This work is supported by NSERC via aDiscovery Grant, by FQRNT, by CIFAR, and a KillamResearch Fellowship.
REFERENCESAlbano, A., Turolla, R., Israel, G. L., Zane, S., Nobili, L., &Stella, L. 2010, ApJ, 722, 788Beloborodov, A. M. 2009, ApJ, 703, 1044Bernardini, F., Israel, G. L., Stella, L., Turolla, R., Esposito, P.,Rea, N., Zane, S., Tiengo, A., Campana, S., Gotz, D.,Mereghetti, S., & Romano, P. 2011, A&A, in press;arXiv:1102.5419Burrows, D. N., Hill, J. E., Nousek, J. A., Kennea, J. A., Wells,A., Osborne, J. P., Abbey, A. F., Beardmore, A., Mukerjee, K.,Short, A. D. T., Chincarini, G., Campana, S., Citterio, O.,Moretti, A., Pagani, C., Tagliaferri, G., Giommi, P., Capalbi,M., Tamburelli, F., Angelini, L., Cusumano, G., Br¨auninger,H. W., Burkert, W., & Hartner, G. D. 2005, Space Sci. Rev.,120, 165Camilo, F., Ransom, S. M., Halpern, J. P., & Reynolds, J. 2007,ApJ, 666, L93Cheng, B., Epstein, R. I., Guyer, R. A., & Young, A. C. 1996,Nature, 382, 513Crosby, N. B., Aschwanden, M. J., & Dennis, B. R. 1993, SolarPhysics, 143, 275Dib, R., Kaspi, V. M., & Gavriil, F. P. 2008, ApJ, 673, 1044—. 2011, ApJ, submittedGavriil, F. P., Dib, R., & Kaspi, V. M. 2009, ApJ, submitted;arXiv:0905.1256Gavriil, F. P., Kaspi, V. M., & Woods, P. M. 2002, Nature, 419,142—. 2004, ApJ, 607, 959Gelfand, J. D. & Gaensler, B. M. 2007, ApJ, 667, 1111Gotthelf, E. V. & Halpern, J. P. 2007, ApJS, 308, 79 Gotthelf, E. V., Halpern, J. P., Buxton, M., & Bailyn, C. 2004,ApJ, 605, 368G¨o˘g¨u¸s, E., Kouveliotou, C., Woods, P. M., Thompson, C.,Duncan, R. C., & Briggs, M. S. 2001, ApJ, 558, 228G¨o˘g¨u¸s, E., Woods, P. M., Kouveliotou, C., van Paradijs, J.,Briggs, M. S., Duncan, R. C., & Thompson, C. 1999, ApJ, 526,L93G¨o˘g¨u¸s, E., Woods, P. M., Kouveliotou, C., van Paradijs, J.,Briggs, M. S., Duncan, R. C., & Thompson, C. 2000, ApJ, 532,L121Halpern, J. P., Gotthelf, E. V., Reynolds, J., Ransom, S. M., &Camilo, F. 2008, ApJ, 676, 1178Ibrahim, A. I., Safi-Harb, S., Swank, J. H., Parke, W., Zane, S., &Turolla, R. 2002, ApJ, 574, L51Ibrahim, A. I., Swank, J. H., & Parke, W. 2003, ApJ, 584, L17Israel, G. L., Campana, S., Dall’Osso, S., Muno, M. P.,Cummings, J., Perna, R., & Stella, L. 2007, ApJ, 664, 448Israel, G. L., Esposito, P., Rea, N., Dall’Osso, S., Senziani, F.,Romano, P., Mangano, V., G¨otz, D., Zane, S., Tiengo, A.,Palmer, D. M., Krimm, H., Gehrels, N., Mereghetti, S., Stella,L., Turolla, R., Campana, S., Perna, R., Angelini, L., & deLuca, A. 2010, MNRAS, 408, 1387Kaneko, Y., G¨o˘g¨u¸s, E., Kouveliotou, C., Granot, J.,Ramirez-Ruiz, E., van der Horst, A. J., Watts, A. L., Finger,M. H., Gehrels, N., Pe’er, A., van der Klis, M., von Kienlin, A.,Wachter, S., Wilson-Hodge, C. A., & Woods, P. M. 2010, ApJ,710, 1335Kaspi, V. M., Gavriil, F. P., Woods, P. M., Jensen, J. B.,Roberts, M. S. E., & Chakrabarty, D. 2003, ApJ, 588, L93Krimm, H., Barthelmy, S., Campana, S., Cummings, J., Israel,G., Palmer, D., & Parsons, A. 2006, The Astronomer’sTelegram, 894Kuiper, L., Hermsen, W., den Hartog, P. R., & Collmar, W. 2006,ApJ, 645, 556Kumar, H. S. & Safi-Harb, S. 2010, ApJ, 725, L191Lamb, R. C. & Markert, T. H. 1981, ApJ, 244, 94Lu, E. T., Hamilton, R. J., McTiernan, J. M., & Bromund, K. R.1993, ApJ, 412, 841Lyutikov, M. 2003, MNRAS, 339, 623Lyutikov, M. & Gavriil, F. P. 2006, MNRAS, 368, 690Mereghetti, S. 2008, Astron. Astropys. Rev., 15, 225Mereghetti, S., G¨otz, D., Weidenspointner, G., von Kienlin, A.,Esposito, P., Tiengo, A., Vianello, G., Israel, G. L., Stella, L.,Turolla, R., Rea, N., & Zane, S. 2009, ApJ, 696, L74Ng, C., Kaspi, V. M., Dib, R., Olausen, S. A., Scholz, P., G¨uver,T., ¨Ozel, F., Gavriil, F. P., & Woods, P. M. 2011, ApJ, 729, 131¨Ozel, F. & Guver, T. 2007, ApJ, 659, L141Palmer, D. M. 1999, ApJ, 512, L113—. 2002, Memorie della Societa Astronomica Italiana, 73, 578Perna, R. & Pons, J. A. 2011, ApJ, 727, L51+Rea, N. & Esposito, P. 2011, in High-Energy Emission fromPulsars and their Systems, ed. D. F. Torres & N. Rea (SpringerASSP), arXiv:1101.4472v1Romano, P., Campana, S., Chincarini, G., Cummings, J.,Cusumano, G., Holland, S. T., Mangano, V., Mineo, T., Page,K. L., Pal’Shin, V., Rol, E., Sakamoto, T., Zhang, B., Aptekar,R., Barbier, S., Barthelmy, S., Beardmore, A. P., Boyd, P.,Burrows, D. N., Capalbi, M., Fenimore, E. E., Frederiks, D.,Gehrels, N., Giommi, P., Goad, M. R., Godet, O., Golenetskii,S., Guetta, D., Kennea, J. A., La Parola, V., Malesani, D.,Marshall, F., Moretti, A., Nousek, J. A., O’Brien, P. T.,Osborne, J. P., Perri, M., & Tagliaferri, G. 2006, A&A, 456, 917Savchenko, V., Neronov, A., Beckmann, V., Produit, N., &Walter, R. 2010, A&A, 510, A77+Strohmayer, T. E. & Ibrahim, A. I. 2000, ApJ, 537, L111Tam, C. R., Gavriil, F. P., Dib, R., Kaspi, V. M., Woods, P. M.,& Bassa, C. 2008, ApJ, 677, 503Thompson, C. & Duncan, R. C. 1995, MNRAS, 275, 255Thompson, C., Lyutikov, M., & Kulkarni, S. R. 2002, ApJ, 574,332Tiengo, A., Vianello, G., Esposito, P., Mereghetti, S., Giuliani,A., Costantini, E., Israel, G. L., Stella, L., Turolla, R., Zane, S.,Rea, N., G¨otz, D., Bernardini, F., Moretti, A., Romano, P.,Ehle, M., & Gehrels, N. 2010, ApJ, 710, 227
Table 1
Summary of
Swift
XRT observations of the 2009 outburst of 1E 1547 − − . − . Table 2
Burst statistics of magnetarsMagnetar T t r t f Γ B a P a Reference(ms) (ms) (ms) (10 G) (s)SGR 1806 −
20 161.8 - - - 21 7.6 G¨o˘g¨u¸s et al. (2001)SGR 1900+14 93.4 - - - 7.3 8.0 G¨o˘g¨u¸s et al. (2001)1E 2259+586 99.31 2.43 13.21 1.35 0.59 7.0 Gavriil et al. (2004)1E 1547 − a ∼ pulsar/magnetar/main.htmlWoods, P. M., Kaspi, V. M., Thompson, C., Gavriil, F. P.,Marshall, H. L., Chakrabarty, D., Flanagan, K., Heyl, J., &Hernquist, L. 2004, ApJ, 605, 378Woods, P. M., Kouveliotou, C., Gavriil, F. P., Kaspi, M., V.,Roberts, M. S. E., Ibrahim, A., Markwardt, C. B., Swank,J. H., & Finger, M. H. 2005, ApJ, 629, 985Woods, P. M. & Thompson, C. 2006, in Compact Stellar X-raySources, ed. W. H. G. Lewin & M. van der Klis (UK:Cambridge University Press) Zhu, W., Kaspi, V. M., Dib, R., Woods, P. M., Gavriil, F. P., &Archibald, A. M. 2008, ApJ, 686, 520 -10 -9 (cid:0) (cid:1) (cid:2) -1 Days from first BAT Trigger U n a b s F l u x ( e r g s s (cid:3) c m (cid:4) ) k T ( k e V ) (cid:5) BB R a d i u s ( k m ) P u l s e d F r a c t i o n Figure 1.
Properties of the persistent emission of 1E 1547 − R X T E P u l s e d c o un t r a t e ( s (cid:6) P C U (cid:7) ) Figure 2. − Swift
XRT and
RXTE . The black crosses show the
RXTE pulsed count rates and the red points are the
Swift pulsed count rates, arbitrarily scaled to the
RXTE values. The dotted verticallines mark the onsets of the 2008 and 2009 outbursts. -10 Unabs Flux (ergs s (cid:8) cm (cid:9) )10 -1 R M S P u l s e d F r a c t i o n Figure 3. C o un t s C o un t s (cid:10) (cid:11) (cid:12) C o un t s (cid:13) (cid:14) (cid:15) C o un t s Figure 4.
Examples of bursts from 1E 1547 − (cid:16) (cid:17) T (s)0510152025303540 N u m b e r o f e v e n t s (cid:18) (cid:19) (cid:20) t r (s) 051015202530354045 N u m b e r o f e v e n t s (cid:21) (cid:22) (cid:23) t f (s)01020304050 N u m b e r o f e v e n t s (cid:24) (cid:25) t r /t f N u m b e r o f e v e n t s Figure 5.
Top left:
Distribution of T duration of bursts. Top right:
Distribution of burst rise times.
Bottom left:
Distribution of burstfall times.
Bottom right:
Distribution of t r /t f . In all panels, the solid line is the best-fit log-normal function. Fluence (counts)10100 N u m b e r o f e v e n t s (cid:26) (cid:27) (cid:28) Fluence (ergs cm (cid:29) ) 10 Peak flux (counts s (cid:30) ) 10100 N u m b e r o f e v e n t s (cid:31) Peak Flux (ergs cm ! s " ) Figure 6.
Left:
Distribution of burst fluences.
Right:
Distribution of burst peak fluxes. The solid line is a linear fit to the filled circles.These distributions are based on burst counts from a 1–10 keV energy band. The open circles are not included in the fits because of reducedsensitivity in detecting such bursts. The top axes show the fluence and peak flux in cgs units which are derived from a single conversionfactor between counts and ergs cm − . In using such a factor, the bursts are assumed to have the same spectrum which is an approximationas each burst has a slightly different spectrum. P e r s i s t e n t c o un t s (e) N u m b e r o f b u r s t p e a k s (a) B u r s t c o un t s (b) S y mm e t r i c b u r s t c o un t s (c) T a il e d b u r s t c o un t s (d) Figure 7. (a) Folded profile of the times of the burst peaks. (b) Folded profile of photon counts from cycles of the pulsar that containbursts. (c) Folded profile of photon counts from cycles containing symmetric bursts. (d) Folded profile of photon counts from cyclescontaining slow-fall bursts. (e) Folded profile of photon counts from cycles of the pulsar that do not contain bursts but not including thefirst two observations following the BAT trigger, illustrating the quiescent pulse profile. Profiles are all in the 0.5–10 keV energy range. -7 Flux (erg cm s $ ) % & Figure 8.
Power-law index as a function of average absorbed flux over T from the spectral fits of the 46 most fluent bursts. The blackpoints are the individual bursts and the blue open squares represent the weighted averages of the power-law indices for bursts in logarithmicfluence bins. F/F q H / H q
1E 1048.1-5937 20071E 1048.1-5937 20021E 2259+586 2002XTE J1810-197 20031E 1547-5408 20091E 1547-5408 2008
Figure 9. H , as a function of 2–10 keV flux, F for magnetar outbursts. Both are normalised to theirquiescent values, H q , F q . For 1E 1547 − −
197 the spectralparameters were taken from Gotthelf & Halpern (2007). For 1E 1048.1 − −−