INTEGRAL & Magnetars: a high energy approach to extreme neutron stars

Transcription

1 INTEGRAL & Magnetars: a high energy approach to extreme neutron stars Diego Götz CEA - Saclay - Irfu/Service d Astrophysique N. Rea (UvA), S. Zane (MSSL), R. Turolla (Uni Padova), M. Lyutikov (Purdue Univ.) P. Esposito, S. Mereghetti, A. Tiengo, G.L. Israel (INAF), K. Hurley (UCB), E.V. Gotthelf (Columbia Univ.)

2 Main manifestations of Neutron Stars: (Radio) Pulsars - Rotational energy >1500 pulsars observed in radio (+ several Pulsar Wind Nebulae) the youngest seen also at higher energies mostly isolated typical rotation periods: 1.5 ms 5 s Accreting X-ray binaries - Gravitational energy several hundreds in High Mass and Low Mass X-ray binaries many are transients typical rotation periods s Magnetars do not fit in these two categories! 2

3 AXPs Originally identified as a class based on: Periods in a narrow range and other properties that distinguished them from the classical X-ray pulsars in High Mass X-ray Binaries 3

4 Summary of AXP properties No evidence for companion stars (very faint IR counterparts, no Doppler delays in pulses) Rotational period of a few seconds (2-12 s) Secular spin-down (0.05-4)x10-11 s/s L x erg s -1 >> Rotational Energy Loss Very soft X-ray spectrum below 10 kev (kt~0.5 kev) 3 (or 4?) are in Supernova Remnants 3 (or 4?) are transients 4

5 AXP Census - 10 confirmed 2 candidates P (s) dp/dt (10-11 s/s) 4U E CTB 109 1E E Kes 73 AX J G Tr. RXS CXO J in SMC XTE J Tr./R CXO J in Wes 1 Tr. 1E Tr./R PSR J1846 (Kes 75) PWN 5

6 SGRs: Initially considered a peculiar class of Gamma-Ray Bursts short, soft, repeating, L peak >>> L Eddington Durations 836 GAMMA-RAY BURSTS Spectra GAMMA-RAY BURST NUMBER OF EVENTS SOFT GAMMA REPEATERS NUMBER OF EVENTS FLUX, photons/cm 2 s kev SOFT GAMMA REPEATER kt~30 kev DURATION, SECONDS ENERGY, kev 6

7 SGRs bursting activity is not continuous 7

8 Bursts from SGR observed with INTEGRAL kev kev Götz, et al. (2004), A&A 417, L45 8

9 3 Giant Flares from 3 SGRs 1979 March 5 - SGR August 27 - SGR December 27 SGR

10 5 confirmed SGRs Soft Gamma Repeaters 4 are in the Galactic plane typical distance ~several kpc one is in the N49 supernova remnant in the Large Magellanic Cloud (d=55 kpc) ? 10

11 SGR-like bursts seen from six AXPs 1E E 2259 XTE J1810 4U 0142 CXO J1647 PSR J bursts in 8 yrs >80 bursts in few hours 4 bursts in 3 yrs 5 burst in 8 yrs 1 burst 5 bursts in 10 years Gavriil et al E E Kaspi et al

12 Transient Phenomena in AXPs Gotthelf & Helfand (2007) XTE 1810 Gavriil at al. (2008) PSR J1846 Israel et al. (2007) CXO in Wes 1 Swift/BAT 12

13 SGR cooling Γ ~1.5 The persistent spectrum is much brighter and harder 2 components flux decay Γ ~3.3 Burst active phase Esposito et al. 2008, MNRAS, in press, arxiv:

14 Magnetar model Duncan & Thompson 1992, ApJ 392, L9 Thompson & Duncan 1995, MNRAS 275, 255 Thompson et al. 2000, ApJ 543, 340 Thompson, Lyutikov & Kulkarni 2002, ApJ 574, 332. If the proto-ns is initially spinning at ~few ms an efficient dynamo can produce B~10 15 G Magnetars spin-down quickly to P>10 s in 10 4 /B 2 15 yrs 14

15 15

16 MAGNETIC ENERGY ROTATIONAL ENERGY E B ~ (1/12) B 2 R 3 E R = ½ I Ω 2. B = 3.2 x (PP) 1/2. E B ~ (P/5 s) P -11 E R = (P/5 s) -2 Magnetic energy dominates over rotational energy after the NS has slowed down to periods of a few seconds 16

17 Known manifestations of Neutron Stars: (Radio) Pulsars - Rotational energy >1500 pulsars observed in radio (+ several Pulsar Wind Nebulae) the youngest seen also at higher energies mostly isolated - typical periods s Accreting X-ray binaries - Gravitational energy several hundreds in High Mass and Low Mass X-ray binaries many are transients - typical periods s Magnetars Magnetic energy - B~10 15 Gauss 4 Soft Gamma-ray Repeaters + ~10 Anomalous X-ray Pulsars middle aged isolated NS - Thermal energy a few nearby NS T~ K Type I X-ray bursts - Nuclear energy a subclass of Low Mass X-ray binaries 17

18 Magnetars hard tails discovered by INTEGRAL Clear evidence for non-thermal persistent emission. Energetically important contribution: L(>10 kev) ~10 36 erg/s Spectrum above 10 kev hardens for AXPs, while for SGRs it softens SGRs No clear physical model has yet been developed for the broad-band spectra of Magnetars. Persistent hard X-ray emission can be due to: Bremsstrahlung photons produced in a thin layer close to the neutron star (Thompson & Belobodorov 2005). Cutoff at ~100 kev. at 100 km altitude in the magnetosphere through multiple resonant cyclotron scattering (Thompson et al. 2002). Cutoff at ~1 MeV A third scenario involving resonant magnetic Compton up-scattering of soft X- ray photons by a non-thermal population of highly relativistic electrons has been proposed by Baring et al. (2007) Götz et al. (2006) AXPs 18

19 Modelling Magnetars High Energy Emission Our immediate goal: kev emission well represented by a blackbody plus a power law: WHY?? Correlation in spectral hardening, luminosity, spin down rate - as in SGR 1806, during the pre (and post)-giant flare (24 Dec 2005) evolution Evolution of transient AXPs Model the hard tails 19

20 Twisted Magnetospheres Thompson, Lyutikov and Kulkarni (2002): Magnetars (AXPs and SGRs) differ from radiopulsars since their internal magnetic field is twisted up to 10 times the external dipole. At intervals, it can twist up the external field A key feature of twisted MSs is that they support current flows (in excess of the Goldreich-Julian current). Thermal seed photons (i.e. emitted from the star surface) travelling through the magnetosphere experience efficient resonant cyclotron scattering onto charged magnetospheric particles (e - and ions) the thermal surface spectrum get distorted! 20

21 Twisted Magnetospheres While the twist grows, charged particles (e - and ions) produces both : an extra heating of the star surface (by returning currents) -> X-ray luminosity increases and a large resonant cyclotron scattering depth -> spectral hardening increases The B-field flares out slightly -> open field flux > then in a dipole -> spin down torque increases a) Crustal cracks occur when the crust cannot bear the stress anymore or b) a global rearrangement of the field lines. -> a forced opening of the field outwards -> launch of an hot fireball And everything is reversed during the aftermath (simplification of the external B-field and by a partial magnetospheric untwisting -> rapid drop in the flux, spectral softening, period derivative decrease, etc..) Qualitatively ok, and quantitatively? 21

22 SGR PULSE PERIOD POWER LAW INDEX (2-10 kev) 2-10 kev FLUX XMM kev FLUX INTEGRAL IPN BURST RATE 22

23 Resonant Cyclotron Scattering in Magnetars Main point: closed field lines in NS magnetospheres are not dead. Populated by hot, highly over-dense plasma, n >> ngj (e.g. Thompson, Lyutikov & Kulkarni 2002; Liutykov & Gavriil 2006) Twisted magnetospheres are filled with plasma which may modify radiation properties. # " RCS ~ R & L % (" T ~ 10 5 " T $ ' at 1 kev with " B R L ~ 8R NS NS $ # r e B crit 1/ 3 % " 1keV ' $ & # h( B % ' & 1/ 3 23

24 Preliminary investigations (1D) Lyutikov & Gravriil, 2006: A simplified, 1D semi-analytical treatment of resonant cyclotron up-scattering of soft thermal photons Resonant Thomson scattering occurs in a thin, plane parallel slab. Photons can only propagate along the slab normal, i.e. either towards or away from the star. Static, non-relativistic, warm medium; n e constant. No electron recoil (hν << m e c 2 ) The NS surface emits seed photons (blackbody spectrum) Magnetospheric charges have a top-hat velocity distribution centered at zero and extending up to ±β T -> mimics a thermal, 1D, motion (β T» mean e- energy» temperature of the 1D electron plasma). No bulk motion. The e - velocity distribution averages to zero: -> a photon has the same probability to undergo up or down scattering -> no frequency shift due to the thermal motion of e- Photon boosting by particle thermal motion in Thomson limit occurs only due to the spatial variation of the magnetic field. For a photon propagating from high to low magnetic fields, multiple resonant cyclotron scattering will, on average, up-scatter the transmitted radiation -> hard tail. 24

25 Resonant Cyclotron Scattering in Magnetars Same number of free parameters, same as for the empirical as the blackbody+power law model; same statistical significance Optical depth " RCS = $ # RCS n dz e (1-10) Electron thermal velocity " T ( ) Surface Temperature (kev) kt ( ) Distorsion of a seed blackbody spectrum through resonant cyclotron scattering onto magnetospheric electrons, for two values of the blackbody temperature, 0.2 kev and 0.8 kev. Black lines: the RCS model for β T = 0.2 and τ res = 2, 4, 8 (from bottom to top). Grey lines: β T = 0.4 and τ res = 2, 4, 8 (from bottom to top). The normalizations of the various curves are arbitrary. From Rea et al Flux 1 - Energy (kev) - 10 Rea, Zane, Turolla, Lyutikov & Götz (2008) 25

26 RCS: AXPs with hard X-ray emission 4U U Flux BB+PL+PL 1 - Energy (kev) RCS+PL RXS J RXS J BB+PL+PL RCS+PL 1E E BB+PL RCS+PL KT consistent for all sources (0.33 kev) while β (1-2) and τ ( ) 25/11/2008 CEA DSM Irfu Diego Götz - Pulsars Workshop - IAP Paris 26

27 RCS: Transient AXPs Much softer spectra; β increases as flux decreases 1E CXO Flux BB+PL RCS BB+PL RCS 1 - Energy (kev) E XTE BB+PL BB+PL RCS RCS 27

28 RCS: SGRs with hard X-ray emission Harder spectra below 10 kev. Additional PL needed at low energies, which extends up to 200 kev Flux SGR SGR BB+PL BB+PL 1 - Energy (kev) SGR BB+PL SGR RCS+PL Flux Rea, Zane, Turolla, Lyutikov & Götz (2008)) RCS+PL 1 - Energy (kev) /11/2008 CEA DSM Irfu Diego Götz - Pulsars Workshop - IAP Paris 28

29 τ res L 1-10keV (10 34 erg/s) Magnetospheric e- density of n~1.5x10 13 cm -3 = 10 3 n GJ 29

30 Conclusions & Future Developments Twisted magnetosphere model, within magnetar scenario, in general agreement with observations below 10 kev Resonant scattering of thermal, surface photons produces spectra with right properties ESA press release XMM-Newton and INTEGRAL clues on magnetic powerhouses 14/10/2008 More accurate treatment of cross section including QED effects and electron recoil (Nobili, Turolla & SZ MNRAS in press) Many issues need to be investigated further Use the model archive to fit model spectra to observations, investigate what causes the long term variability in AXPS and TAXPS Phase resolved spectroscopy kev tails: up-scattering by (ultra)relativistic (e±) particles? Necessity of more sensitive broad band observations (Simbol-X, NuStar) Detailed models for magnetospheric currents 30