Table Of Content

Proceedingsofthe363.WE-HeraeusSeminaron:“NeutronStars and Pulsars”(Postersand contributedtalks) Physikzentrum BadHonnef, Germany,May. 14−19,2006, eds.W.Becker,H.H.Huang, MPEReport291,pp.201-204 Long term spectral variability in the soft gamma-ray repeater SGR 1900+14 Paolo Esposito1,2, Sandro Mereghetti2, Andrea Tiengo2, Diego Götz3, Lara Sidoli2, and Marco Feroci4 1 Università diPavia, Dipartimento diFisica Nucleare e Teorica and INFN-Pavia,via Bassi 6, I-27100 Pavia, Italy 2 INAF-IASFMilano, via Bassini 15, I-20133 Milan, Italy 7 3 CEA Saclay, DSM/DAPNIA/Serviced’Astrophysique,F-91191, Gif-sur-Yvette,France 0 4 INAF-IASFRoma, via Fosso del Cavaliere 100, I-00133 Roma, Italy 0 2 n a Abstract. We present a systematic analysis of all the Table 1. Log of the BeppoSAX observations of J BeppoSAX data ofSGR1900+14.The observationsspan- SGR 1900+14 9 ning five years show that the source was brighter than usual on two occasions: ∼20 days after the August 1998 2 v giant flare and during the 105s long X–ray afterglow fol- Obs Date Instrument/exposure 7 lowingtheApril2001intermediateflare.Inthelattercase, 8 weexplorethepossibilityofdescribingthe observedshort A 1997May12 LECS/20ks MECS/46ks PDS/20ks 7 term softening only with a change of the temperature of B 1998Sep15 LECS/14ks MECS/33ks PDS/16ks 2 a blackbody-like component. In the only BeppoSAX ob- C 2000Mar30 LECS/14ks MECS/40ks PDS/18ks 1 D 2000Apr25 LECS/17ks MECS/40ks PDS/19ks servation performed before the giant flare, the spectrum 6 E 2001Apr18 LECS/20ks MECS/46ks PDS/17ks 0 of the SGR1900+14 persistent emission was significantly F 2001Apr29 LECS/26ks MECS/58ks PDS/26ks / harder and possibly detected also above 10 keV with the h G 2002Mar09 - - PDS/48ks p PDSinstrument.InthelastBeppoSAXobservation(April H 2002Apr27 - MECS/83ks - - 2002) the flux was ∼25% lower than the historical level, o suggestingthatthesourcewasenteringaquiescentperiod. r t s field (B∼1014–1015 G) rather than by rotation. a : Herewepresenttheanalysisofthepersistentemission v 1. Introduction of SGR1900+14 by means of the BeppoSAX satellite, i X both in the soft and hard X–ray range, and its evolution The soft gamma-ray repeater (SGR) SGR1900+14 was r across the giant flare and in relation to its bursting a discovered in 1979 through series of short and soft activity. gamma-raybursts(Mazets et al.1979).Manyyearslater, its persistent pulsating X–ray counterpartwas discovered in the 2–10 keV energy band (Hurley et al. 1999b). More 2. Soft X–ray emission recently, it was also detected in the hard X–ray range 2.1. Observations and data analysis (20–100 keV) with the INTEGRAL satellite, becoming the second SGR, after SGR1806−20, established as a We have analyzed all the X–ray observations of persistent hard X-ray source (Götz et al. 2006). SGR1900+14performedwiththeBeppoSAXsatellite(see The rather discontinuous bursting activity of Table 1). The spectra were extracted from the MECS SGR1900+14 (see Fig. 1, bottom panel) raised to a (Boella et al. 1997) and LECS (Parmar et al. 1997) in- summit on 1998 August 27 with the emission of a giant struments using circles with radii 4′ and 8′, respectively. flare,when morethan1044 ergsofγ–rayswere emitted in The background spectra were extracted in all cases from less than one second (Hurley et al. 1999a). This was one nearby regions and time filters were applied to both the of the three giant flares detected up to now from three sourceandbackgroundspectratoexcludetheSGRbursts different SGRs. The extreme properties of these events detected during observation B, E and F. are the main motivation for the magnetar interpretation. In this model (Thompson & Duncan 1995, 1996), the 2.2. Spectral results SGRs and the Anomalous X–ray Pulsars (AXPs, another class of X–ray sources with similar properties, see e.g. We have first tried to fit the spectra with an absorbed Mereghetti et al. 2002) are believed to be neutron stars power-law model, but three observations give unaccept- poweredby the decayoftheir extremely intense magnetic able values of the χ2 and structured residuals. For 202 Paolo Esposito et al.: Long term spectral variability in thesoft gamma-ray repeater SGR1900+14 Fig.2. Background subtracted MECS 2–10 keV light curve andblackbody temperature observedon2001April 18about7.5hoursafter the flare.Thelatter valuescorre- spond to an additionalblackbody component that can be interpretedasduetoaportionoftheneutronstarsurface heated during the flare (see Section 2.2). ∼20times lowerthanthatofthe 2001April18giantflare (Guidorzi et al. 2004). Theafterglowfollowingthisbrightburstisclearlyvis- ible during the BeppoSAX observation as a decrease in the X–ray flux (see Fig. 2, top panel), accompanied by a Fig.1. Long term evolution of the 2–10 keV unabsorbed significant softening of the spectrum (Feroci et al. 2003). flux, the spectral parameters (for an absorbed power-law plus blackbody model,assumingn =2.6×1022cm−2) and In addition to the afterglow analysis already reported by H Feroci et al. (2003), we have performed a time resolved the burstactivity(asobservedby the InterplanetaryNet- spectroscopy of the afterglow by dividing observation E work)ofSGR1900+14.Theverticaldashedlines indicate into five time intervals. Under the assumption that the the giant and intermediate flares (1998 August 27 and variable “afterglow” emission is present on top of a “qui- 2001 April 18, respectively). escent” emission that shows only moderate variations on long time-scales, we fitted them with a model consisting of a power-law plus blackbody with fixed parameters (as these observations, good fits are obtained with the ad- representative parameters of the fixed quiescent emission dition of a blackbody component. Since such a two- we used values consistent with those seen in the last ob- components model is typical of the magnetar candidates servationsbeforetheflare:CandD),plusathirdvariable (Woods & Thompson 2004), we have used this model to componenttomodeltheafterglowemission.Althoughthe fitallthe availablespectra,obtainingthe resultsreported “afterglowemission” in the five spectra canbe fitted by a in Fig. 1. The blackbody parameters are compatible in variety of models, the spectral evolution of the afterglow all the available observations, except for that taken dur- iswellrepresentedbyanadditionalblackbodycomponent ing the afterglow of the intermediate flare (observation with fixed emitting area (radius of ∼1.5 km, for a source E). This indicates that a constant blackbody component distance of 15 kpc) and progressively decreasing temper- with k T∼0.4 keV and emitting area with R∼6–7 km, ature (k T from ∼1.3 to ∼0.9 keV, see Fig. 2, bottom B B might be a permanent feature of the X-ray spectrum of panel), that can be interpreted as due to a portion of the SGR1900+14. neutron star surface heated during the flare. As can be seen in the upper panel of Fig. 1, the Excluding the two observations taken after the excep- flux varies by a factor >5, with the highest values ob- tionalexplosiveevents(BandE),thefluxofSGR1900+14 served during observations B and E. These two observa- hada rather constantvalue of ∼10−11 ergcm−2 s−1 from tions were taken shortly after extreme bursting events. 1997to 2001.Onthe other handa significantlylowerflux The former was performed 20 days after the giant flare, level was seen in the following observations. The flux de- that was followed by a ∼2 months period of enhanced crease actually started when the source was still moder- X–ray flux (Woods et al. 1999). The latter started only ately active (the flux in observation H is at least ∼20% 7.5 hours after the intermediate flare which had a fluence lowerthanin allthe previousquiescentobservations)and Paolo Esposito et al.: Long term spectral variability in the soft gamma-ray repeater SGR1900+14 203 has been interrupted by a slight rise in coincidence with Table 2. Status of the three transient sources within the March 2006 burst reactivation, as shown by recent the PDS field ofview during the BeppoSAX observations. XMM-Newton observations (Mereghetti et al. 2006). The presence of the two X–ray pulsars (4U 1907+97 and Although the flux of the only pre-giant flare observa- XTEJ1906+09)isconfirmedbythepresenceoftheirpul- tioniscompatiblewiththatofthequiescentpost-flareob- sations in the PDS data, while the black hole candidate servations taken before 2002, its spectrum is significantly (XTEJ1908+94)byitsdetectionintheMECSandLECS harder,asshowninthesecondpanelofFig. 1.Theoverall images and in the RossiXTE ASM lightcurve. hardness of the pre-flare observation is confirmed by the factthatthespectraC,D,F,andHcanbesimultaneously fitwiththesameparameters(introducinganormalization Obs 4U 1907+97 XTEJ1906+09 XTEJ1908+94 factor to account for the flux change), while the addition of spectrum A gives an unacceptable fit with structured A OFF OFF OFF residuals. B ON ON OFF C OFF ON OFF D ON ON OFF 3. Hard X–ray emission E OFF ON OFF 3.1. Detection with the PDS instrument F ON ON OFF G OFF ON ON To study the high energy emission from SGR1900+14 we used the BeppoSAX PDS instrument (Frontera et al. 1997), whichoperatedinthe 15–300keVrange.The PDS instrument was more sensitive than INTEGRAL in this energyband,butithadnoimagingcapabilitiesandthere- Aupto∼150keVismostlikelyduetoSGR1900+14and forethe possiblecontaminationfromnearbysourcesmust therefore hereafter we assume that the spectral proper- betakenintoaccount.ThefieldofviewofthePDSinstru- ties discussed in the text are associated to SGR1900+14 ment was 1.3◦ (FWHM) and the background subtraction alone. was performed through a rocking system that pointed to ◦ two 3.5 offset positions every 96 s. 3.2. Spectral analysis In the case of SGR1900+14, the background point- ings were free of contaminating sources, as confirmed by The background subtracted PDS spectrum of observa- the identical count rates observed in the two offset posi- tion A can be well fit by a power-law with photon in- tions during each observation. The field of SGR1900+14 dexΓ=1.6±0.3,significantlyflatterthanthatmeasuredby is instead rather crowded, with three transient sources: INTEGRAL (Γ=3.1±0.5, see Fig. 3) in 2003/2004. The the X–ray pulsars 4U 1907+97 (Liu et al. 2000) and corresponding 20–100 keV flux is 6×10−11 erg cm−2 s−1, XTE J1906+09(Liu et al. 2000), and the black hole can- a factor ∼4 higher than during the INTEGRAL observa- didate XTE J1908+94 (in’t Zand et al. 2002), located at tions, which confirms that before the giant flare the hard ′ ′ ′ angular distances of 47, 33 and 24 from the SGR, re- X–ray tail of SGR1900+14was brighter. spectively. The pulsations of the two pulsars are clearly The INTEGRAL spectrum was collected during visible in the PDS data below 50 keV when they are ac- ∼2.5 Ms of different observations performed between tive,whileXTEJ1908+94,ifinoutburst,isclearlyvisible March 2003 and June 2004, and thus it represents the in the simultaneous MECS and LECS images and, being hard X–ray emission of SGR1900+14 averaged over that verybright,alsointhelightcurvecollectedbytheAllSky long time period. Therefore, its relation to the soft X– Monitor(ASM)onboardtheRossiXTEsatellite.Wehave ray spectrum can be studied only comparing the spec- foundthatatleastoneofthesecontaminatingsourceswas tra taken by other instruments in a similar time pe- on in all the BeppoSAX observations except for the first riod, as shown for example in Fig. 3. The PDS instru- one(seeTable2).Thus,onlythe1997observation(obser- ment, instead, being a high sensitivity hard X–ray detec- vation A), during which a significant signal was detected tor coupled to the MECS and LECS soft X–ray cameras, in the background subtracted PDS data, can be used to gives us the chance to study the broad band spectrum study SGR1900+14 without the problem of contaminat- of SGR1900+14 during a single observation. Fitting the ing sources. We searched for the SGR pulsation period 1–150 keV BeppoSAX spectrum of observation A, we ob- (5.15719s, as measured in the simultaneous MECS data) tain a good result (χ2=1.17 for 136 degrees of freedom) inthe PDSdata,butthe resultwasnotconclusive,giving simply extrapolatingto higher energiesthe best-fit model only a 3σ upper limit of 50% to the pulsed fraction of a found in the soft X–ray range. In fact, a fit with an ab- sinusoidalperiodicity,tobecomparedtothe∼20%pulsed sorbed power-law plus blackbody model gives the follow- fraction observed below 10 keV. ing parameters: photon index Γ=1.04±0.08, blackbody Although we cannot rule out contamination from un- temperature k T=0.50±0.06, radius R =5±2 km, and B bb knowntransientsources,thefluxmeasuredinobservation absorption n =(1.8±0.5)×1022 cm−2. H 204 Paolo Esposito et al.: Long term spectral variability in thesoft gamma-ray repeater SGR1900+14 activity is a key point to try to understand the SGR emission processes. Acknowledgements. Wegratefullyacknowledgethesupportby theWE-Heraeus foundation. References Boella, G., Chiappetti, L., Conti, G., et al. 1997, A&AS, 122, 327 Feroci, M., Mereghetti, S., Woods, P., et al. 2003, ApJ, 596, 470 Frontera,F.,Costa,E.,dalFiume,D.,etal.1997,A&AS, 122, 357 Götz,D.,Mereghetti,S.,Tiengo,A.,&Esposito,P.2006, A&A, 449, L31 Fig.3. Blue points: broad band spectrum of Guidorzi, C., Frontera, F., Montanari, E., et al. 2004, A&A, 416, 297 SGR1900+14 taken on 1997 May 12 (observation A) with BeppoSAX (both MECS and PDS data). Green Hurley, K., Cline, T., Mazets, E., et al. 1999a, Nature, points: INTEGRAL data from March 2003 to June 2004. 397, 41 Hurley,K.,Li,P.,Kouveliotou,C.,etal.1999b,ApJ,510, L111 4. Conclusions in’t Zand,J.J.M.,Miller,J.M.,Oosterbroek,T.,&Par- mar, A. N. 2002, A&A, 394, 553 We have studied the variability of SGR1900+14, both in Liu, Q. Z., van Paradijs, J., & van den Heuvel, E. P. J. thehardandinthesoftX–rayrange,findingthefollowing 2000, A&AS, 147, 25 results: Mazets, E. P., Golenetskii, S. V., & Guryan, Y. A. 1979, Soviet Astronomy Letters, 5, 343 – Except for the observations immediately following ex- Mereghetti, S., Chiarlone, L., Israel, G. L., & Stella, L. ceptionalflares,thefluxlevelinsoftX–rayswasstable 2002, in Neutron Stars, Pulsars, and Supernova Rem- while the source was moderately active and progres- nants, ed. W. Becker, H. Lesch & J. Tru¨mper, [ArXiv: sively decreased when it entered a 3 years long quies- astro-ph/0205122] cent period. Mereghetti, S., Esposito,P., Tiengo,A., et al. 2006,ApJ, – The intermediate flare of 2001 April 18 was followed in press (astro-ph/0608588) by an X–ray afterglow that can be interpreted as due Mereghetti, S., Tiengo,A., Esposito,P., et al. 2005,ApJ, to the heating of a significant fraction of the neutron 628, 938 star surface, that then cools down in ∼1 day. This is Parmar,A. N., Martin, D. D. E., Bavdaz,M., etal. 1997, consistent with the interpretation of similar events in A&AS, 122, 309 other magnetar candidates (Woods et al. 2004). Thompson, C. & Duncan, R. C. 1995,MNRAS, 275, 255 – The soft X–ray spectrum during the only available Thompson, C. & Duncan, R. C. 1996,ApJ, 473, 322 pre giant flare observation was harder than in the fol- Tiengo,A.,Esposito,P.,Mereghetti,S.,etal.2005,A&A, lowing quiescent observation. This is similar to what 440, L63 observed in SGR 1806–20, the only other SGR that Woods, P. M., Kaspi, V. M., Thompson, C., et al. 2004, could be monitored before and after a giant flare ApJ, 605, 378 (Mereghetti et al. 2005; Tiengo et al. 2005). Woods, P. M., Kouveliotou, C., van Paradijs, J., Finger, – Comparing the hard X-ray spectrum of SGR1900+14 M. H., & Thompson, C. 1999, ApJ, 518, L103 recently observed with INTEGRAL to that observed Woods,P.M.&Thompson,C.2004,in“CompactStellar withthe PDSinstrumentin1997,wefindevidencefor X-raySources”,ed.W.H.G.LewinandM.vanderKlis, variations in flux and spectral slope. [ArXiv: astro-ph/0406133] – ThereductionoftheX–raytailincoincidencewiththe giant flare is supported by the count rates detected in the PDS instrument above 50 keV during the different BeppoSAX observations, that indicate how they significantly decreased already in the first post-flare observation.SincethehardX–raytailinthespectrum of SGR1900+14 might contain most of its total emitted energy, its variability in relation to the bursting