Bursts and flares from magnetic Pulsars

Transcription

1 Bursts and flares from magnetic Pulsars GianLuca Israel (INAF - RomA ASTRONOMICAL OBSERVATORY)...and many many many others

2 Why Magnetars? Loss of rotational energy is orders of magnitudes too small (1030erg/s) with respect to the observed persistent Lx No accretion from a companion (M>Mjup) No Doppler modulation/shift in the spin pulses In analogy with isolated rotation-powered Definition: NS (Isolated) neutron stars where π R ns PP = 3 3 c I 2 B0 sin 2 α Gauss MAGNEtic stars the main source of energy is the magnetic field [most observed NS have B = G and are powered by accretion, rotational energy, or residual internal heat ]. Several indirect evidences (for high B) collected through years!

3 Two (?) classes of Magnetars (historically) Soft Gamma-ray Repeaters o Discovered in 1979 as sources of hard X-ray bursts and giant flares o 6 confirmed SGRs Pdot ~ P~2-12s Anomalous X-ray Pulsars o Identified in the 90 s as a peculiar class of soft and persistent X-ray pulsar with no signs of binary companions o Spin periods: 2-12 s o 9 confirmed AXPs -10 o o o o (See review: Mereghetti 2008, A&A Rev. 15, 225) Period derivatives: ss-1 Short (~100ms) X/γ bursts / Glitches Rare Giant Flares (>1047 ergs; minutes) Associated with SNRs (2) and massive open clusters (3) with M turn-off of 30-40Msolar and b<0.5 ~ yr

4 BURSts/outbursts

5 SGR : QPOs in the GF tail 27th December 2004 hyperflare Up to ergs released during the first ~ 0.6s (@ a distance of 815kpc), 1 erg /cm2 at Earth!! Similar phenomenology and frequencies in the two sources during GFs 18 30, 61, 92.5,Hz Easily accounted for TOROIDAL GSOs (l=2,4,7) (Israel et al. 2005)

6 Transient QPOs QPOs detection confirmed by RHESSI obs. Additionally transient QPOs at 720, 976. Moreover 625 and 1840 Hz were detected. A re-analysis of the 1998 giant flare of SGR1900 revealed QPOs at the following frequencies: 28.5, 52.5,84 and Hz. SGR SGR (Strohmayer & Watts. 2005; Watts & Strohmayer 2006) The highest freq. signal ever detected in (X-ray) astronomy!!

7 B and the Cavallo-Fabian-Rees limit No direct way to infer B from obs. Several indirect ways. Many of then are model dependent. The Cavallo-Fabian-Rees L limit violation is independent from any assumption or models and allows to infer a B lower limit: B 4x1015G(10km/Rns)(0.1/η)1/2 L/ t = η 4π/3 R3 n mp c2 η=extr. efficiency total energy released within t is related to the total mass within the source dimension R t > R/c(1+τt) τt=σtnr (n baryon density) t must exceed the time over which photons escape from the source t > (3/2π) (σt/mpc4) ( L/η) L/ t=η 2 x 1042 erg/s2 (Cavallo & Rees 1978) For QPOs ( Hz) in SGR1806 the highest L is L/ t=23/2 π L arms νqpo ~ 1044 erg/s2 σ may differ from the Thomson cross section due to B. In the E mode (Meszaros 79) σt is σ reduced by a factor (εbq/(mec2b))2, thus we imposed (εbq/(mec2b))2<η/20 which gives B 3 Bq (0.1/η)1/ G ε~14kev from the BB component with kt~5kev and R~30km as measured in the ringing tail (Vietri et al. 2006)

8 SGR1900: The 2006 burst Forest SGR as observed by Swift on 29th March 2006 More than 100 single bursts were detected in 20min, with 40 in less than 30s. Few of them have intermediate duration (200ms-2s). Total Energy: few x 1042 ergs (one of the more energetic events recorded ever after the 1998 giant flare)

9 the 06 burst forest: spectroscopy Different simple models considered in analogy with previous work done by Feroci et al. (2004), Olive et al. (2004) and Nakagawa et al. (2007). [Bremss not able to fit the low energy part; 2BB better choice] Fits carried out on BAT (on short timescales; 729 spectra; keV) and BAT+ XRT (burst average; 8 spectra; 2-150keV) data. BAT spectra have ~4000 photons each and Dt in the 8ms-400ms range 2BB and CompTT are, by far, the models which give the smallest reduced χ2

10 The KT - R2 plane Sharp edge / saturation present for the brightest part of the bursts The 2BB distributions identify a natural separation surface at 20-25km 20-25km kt-3 kt-4 Olive keV A sort of turn/cut-off is present for the BBh around 10-13keV and 5-15km The locus identified by the relation R2kT3=c can be regarded as constant number of emitted γ per unit time (R2kT4=c identifies a constant L) (Israel et al. 2008)

11 2BB: Lsoft vs Lhard BBs and BBh Luminosities correlate below 3x1041 erg/s (iso-l). Above such value only LBBh evolves (we assumed 10kpc distance). t Max LBBh is ~3x1041 erg/s at 10keV and 15km = magnetic Eddington luminosity (Paczynsky 1992) for B of 8e14 G [similar to that inferred form P and Pdot]. LEdd,B(r) 2 LEdd (B(r)/1012)4/3 Max LBBs ~ 1041 erg/s at 100km hints to the maximum efficiency of the MF to substain the radiation pressure.

12 THE proposed/qualitative scenario A possible interpretation: different way with which Eand O-mode polarised photons propagate into the O-mode γ photosphere magnetosphere; scattering cross section of E-mode may be reduced by B and the scattering photosphere is R~100km smaller (~R NS). E-O-mode γ splitting photosphere R~25-30km E-mode γ photosphere R~15km The R~25km corresponding to the max radius of the BBh and the B=Bcrit O-e--e+ BBh E min of the BBs identify a critical surface at which (QED). BBs O O B<BQED =4.4x1013 G

13 1E outbursts 2 outbursts in less than 4 months (Oct 08 / Jan 09)

14 1E BAT spectra (kt vs R2) ~1800 spectra fitted in a way similar to that of SGR spectra selected after fitting (with smaller uncertainties) Very similar results and behavior to those Of SGR1900! Data are normalized to know distance (and it accounts for the shift in R2)

15 1E BAT spectra (Lsoft vs Lhard) similar but with different saturation values for L soft Can we learn something? B r 40 LEdd,B (r) 2 x 10 B R QED NS 2/3 erg s-1 Deviding Ledd,B measured for the two sources BSGR1900 B1E1547 r 7 1E1547 rsgr R~45km R~15km Ledd,B~ erg/s Ledd,B~ 1041 erg/s 2/3 erg s-1 [ not that dissimilar to what expected from the timing measurements: B~ (PPdot)1/2 BSGR1900 B1E ]

16 SGR1900 BAT spectra: Implications If correct (function of B) LEdd,B should be different for each burst (occurring at different NS surface regions) Lmax~ erg/s Lmax~ erg/s Lmax~ erg/s A ratio of 2 in Lmax corresponds to a factor of 3.2 in B

17 Outburst monitoring 8 outbursts in ~4 years Swift XMM

18 Putting together the timing and spectral domain XTEJ DDT/GO obs ( Fx~50 ) Mar06 H H S S sep06 Hard BB disappeared between MAR and SEP 06 -> PF flattens (Israel et al. 2008; Bernardini et al. 2009) 2BB -> 3BB Ftest gives p~10-12 Perna & Gotthelf (2008) obtained similar results by using 2BBs and fitting light curves and spectra together

19 Putting together The timing and spectral domain New models are used: The NTZ consider the radiative transfer of thermal photons emitted on the NS surface and propagating in a twisted magnetosphere by means of a 3D Montecarlo code. Simultaneous spectral and timing fits (Nobili e t al. 2006; Albano et al. 2010) Not a fit!! only Norm is free to vary. Timing parameters superimposed!!

20 Putting together The timing and spectral domain Physics described by three parameters: φ twist angle β (average) charge velocity kt thermal seed photons +2 angles ( ξ spin-mag. pole axis angle and χ spin line of sight) (Albano et al. 2010) - XTEJ1810 is an aligned rotator - φ decays after the outburst onset to a non 0 value - 3 thermal components needed. The hottest ones decrease with time

21 XTeJ : a transient radio pulsar The brightest radio (!!) pulsar in the sky for several months! Transient emission A new observational windows opened

22 constraints from XTEJ mw obs. No significant X-ray variability during the radio transitions. Flux~const however Fr~0.01Fx Bright Fading Dim Mar 07 Fading Sep 06 X-ray peak phase : \ Radio peak phase : \ Simultaneous detection of radio and Xray pulsations is one of the strongest evidence against accretion being at work! X-ray/radio alignment X-rays and radio are coming from the same portion of the NS. A larger X-ray duty cycle indicates a larger emitting area for X-rays. Radio variability not correlated to X-rays - no significant flux variations In agreement with the idea that most of the X-rays originated deep in the crust after thermalization.

23 1E : the 09 radio Switch Not detected on 22nd (Camilo et al. 09) and 23rd (Burgay et al. 09) Jan 09 weak radio pulsations detected on 25th from Parkes in simultaneity with Chandra X-ray Radio (Burgay et al. 2009) 2007 observations

24 1E bursts and radio X-ray burst coincides with radio pulsations and X-ray peak!! Only simultaneous X-ray/radio observations allow to obtain these results.

25 ... 1E P P P solution P= (7) s,. P= 2.9(2) s s 1.. P= (2) s, P= 2.4(4) s s 2 inc. spin-down P consistent with 0 Oct Jan P= 2.4(1) s s 1.. Implications (preliminary): standard scenario (radio pulsarlike).p > or <.0 depending on whether Ppost / Psec < or > 1 magnetar/twist scenario in disagreement unless the twist is implanted locally and it have not yet reached its maximum value (after that it decreases) to be confirmed in the future For a review on glitches and Pdot see Dib et al Sant Cugat, 14th Apr 2010

26 related objects? magnetar-like X-ray bursts from the young pulsar PSR J , at the center of the supernova remnant Kes 75 (surface dipolar magnetic field of G) (Gavriil et al. 2008) Pdot ~ P~2-12s

27 Conclusions. Not that conclusive Almost all AXPs/SGRs are Transient and emit burst and flares!! At all wavelengths!! Peculiar events (flares and outbursts) seems to be a very powerful tool in the study the emission mechanism(s) from AXPs/SGRs. The greatest part of these results have been obtained thanks to the complementary capabilities of small and large X-ray missions (Swift XMM Chandra). Fermi mission is also detecting new Magnetars (SGR )! Main issues which have been addressed: - Model independent measure of B (giant flares) - The effect of B (trapped fireball/confinement) during intense burst (IFs) - The effect of B (twisted magnetosphere) ToO and Multi-Wavelenght (simultaneous) obs has revealed to be a fundamental strategy

28