08 May 7 SAIt08

Fenomeni Transienti nei Magnetars (i piu' potenti magneti dell'universo)

GianLuca Israel (osservatorio astronomico di Roma/monteporzio)

INAF (Roma/Milan/Palermo): P. Romano, V. Mangano, F. Bernardini, M. Burgay, A. Possenti, P. Esposito, D. Gotz, S. Mereghetti, G. Chincarini, L. Stella, S.

Campana, T. Belloni, G. Tagliaferri, A. Moretti, C. Guidorzi, A.Albano (also Univ. Padova) Univ. Pisa: S. Dall’ Osso (also INAF Roma) MSSL (London): S. Zane, J. Blustin, R.

MignaniNASA/GSFC: T. Sakamoto, N. Gehrels, H. Krimm PSU: D. Burrows LANL: D. Palmer Berkeley: K. Hurley JILA: R. Perna Caltech: M. Muno GSFC: J. Cummings

08 May 7 SAIt08

Studying Magnetars

Identifying their nature.Observational properties did not fit into any of the known class of astrophysical objects.

Understanding their emission mechanism(s). Theoretical models involve peculiar objects. If correct, AXPs/SGRs would be the first class of astrophysical objects the emission of which is mainly driven by decay of high magnetic fields (>10^15Gauss !!).

AXPs/SGRs are rare objects (<15) which might allow us to test the physics under extreme conditions (grav. field, B, density/pressure, etc.)

08 May 7 SAIt08

OUTLINETransient phenomena in AXPs/SGRs

QPOs in GFTimescales of Frac./few msSGR1806-20 discovery of QPOs (L. Stella’ s talk)

Bursts: Type A and BTimescales of 10-100ms4U0142+614, 1E1048.1-5937, 2E2259+586: Persistent AXPsCXOUJ164710.2-455216, XTEJ1810-197: Transient APXs

Intermediate flares: Timescales of 1-40s

SGR1900+14 SGR1900+14 sampling of the trapped sampling of the trapped fireballfireball

OutburstsTimescales of months/years

XTE J1810-197XTE J1810-197 the outburst fading [1-4yr]the outburst fading [1-4yr]CXOU J164710.2-455216CXOU J164710.2-455216 the outburst onset [0- the outburst onset [0-1.5yr]1.5yr]

08 May 7 SAIt08

Why transient and outbursts ?

In analogy with studies done in the past on transient X-ray pulsarsin binary systems the (identification and) study of transient AXPs might give insights on :

• the emission mechanisms (mainly unknown) as a function of flux (continuously changing main parameters) for a constant distance, geometry and magnetic field B. Outburst timescales (years) are such that monitoring observations are easy to set.

08 May 7 SAIt08

From Quiescence to outburst

On 2006 September 21st BAT recorded a 20ms long and intense (for an AXP… Lx~1039 erg/s) burst from CXOU J164710-455216 (after 2E2259+58, 1E1048-59 and 4U0142+61)(Note that in 3 years and ~300 events no GRB – or false event - was ever detected within 3’ from the position of any AXPs/SGRs).

Pre or post outburst ??

No XRT pointing due to the presence of a bright X-ray sourceGX 340-2 at 20’ . Moreover the burst was tagged as not real……

First pointing was finally carried out after 13hours !

kT~10keVkT~10keVoror=1.8=1.8

Swift BATSwift BAT(15-150keV)(15-150keV)

(Israel et al. 2007)

08 May 7 SAIt08

The Awakening of the AXP in Westerlund 1

The source was detected at a Lx level of 1036 ergs/s ~300-400 times larger than the week before !! The largest flux enhancement ever detected for an AXP.The spectrum was different form the quiescent level one

We started an X-ray (Swift/XMM) and Radio (Parkes) ToO monitor campaign

PL~2.2PL~2.2(60%of total)(60%of total)

BB~0.65keVBB~0.65keV R ~ 2kmR ~ 2km

PL~3.5 PL~3.5 (60%of total)(60%of total)

BB~0.5keVBB~0.5keV R~ 0.4kmR~ 0.4km

08 May 7 SAIt08

PHASE-coherent timing: PULSE shapesSw

ift

XMM-Newton

XMM-Newton

chandra

Peak identificationPeak identification based on the assumption of maximizing the constant parameters: 2 peaks almost constant ( © )1 peak highly variable ( ® )

After 6 months the pulse shapeis still very different from thatin quiescence

Pulse variability testifies to changes in the emission pattern,From rather simple (sinusoid) to complex and energy dependent

® ©©

Tim

e

S

H

08 May 7 SAIt08

PHASE-coherent timing: a glitching AXP

A glitch ( ~ 6.5x10-5) occurred within 0.8 days from the burst epoch (the largest ever detected in a NS). We also revealed the presence of the secular spin-down trend.

Our Pdot of 9.2(4)x10-13 s/s implies a dipole field strength of ~ 1014 G.

(Israel e

t al. 2

006,2

00

7)

(Woods et al. 2003)

~ 4x10-6

2E2259+58Energy: the glitch par. imply an energy release of 5x1041 ergs (three months) 0.004% emitted in the burst6% stored in the NS and released in 2days90% went to increase the persistent flux

08 May 7 SAIt08

From outburst to quiescence: XTeJ1810-197

Israel et al. 2006

kT = 0.67 +/- 0.01 keV

= 3.7 +/- 0.2 (kT=0.3keV)

NH = 1.05(5) x 1022 cm-2 P=5.5sPdot=2 10-11 s/sB = 2.4x1014 GPF=46(3)%

A 5.5s transient X-raypulsar was dicovered in 2003

Timing and spectral properties consistent with those of AXPs

08 May 7 SAIt08

XMM (2003)

+ XMM (2004)

+ ROSATFx ~ 5 x 10-13 erg s-1 cm-2

BB with kT=0.18(2) keV and R~10kmNo pulsations with PF>20-25%

Lx~1033erg/s @ 4kpc not distinguishable from hundreds of unidentified ROSAT sources

The spectral evolution

Open Issues: • Does the XMM Soft BB evolve into the ROSAT BB component ? • Is the quiescent ROSAT BB component always there (and constant) ?• Which is (will be) the XMM Hard BB component evolution ? • Is there any other component present/appearing/disappearing through the outburst ? • Which is the PF at quiescence ? Is the modified BB really needed ?

2BBs used to account for the X-ray emission from the NS Surface (Gotthelf & Halpern 2006). kT1= 0.25(1) keV kT2= 0.67(1) keV

Federico Bernardini thesis

08 May 7 SAIt08

the SPECTRAL evolution – 3BBs Mar06

sep06

Tail component present between 03-04 possiblyrelated to a HE PL observed in other AXPs (3 c.l.)

Hard BB disappearedbetween MAR and SEP 06 -> PF flattens(Bernardini et al. 2008)

H

H

S

S

2BB -> 3BB Ftest gives p~10-12

9 DDT/GO obs (Fx~50 )

08 May 7 SAIt08

XMM/Parkes: NO correlated variability !

X-ray peak phase : 0.151+\-0.005 Radio peak phase : 0.151+\-0.019

No significant X-ray variability during the radio transitions. Flux~const however Fr~0.01Fx

Bright FadingDim

Fading

Sep 06Mar 07

No significant variations in the X/radio shift between Sep 06 and Mar 07.

X-ray/radio alignment X-ray/radio alignment X-rays and radio are coming from the same X-rays and radio are coming from the same portion of the NS. portion of the NS. A larger X-ray duty cycle indicates a larger emitting area A larger X-ray duty cycle indicates a larger emitting area for X-raysfor X-rays. . PPS analysis confirms that 2BB emitting regions are in phase PPS analysis confirms that 2BB emitting regions are in phase

Radio variability not easily 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.

08 May 7 SAIt08

Conclusions - I

TAXP studies seems to be a very powerful tool to study the emission mechanism(s) from AXPs and their evolution. The exceptional case of XTEJ1810-197 and its radio emission is even more important allowing to identify a common emission region for X-rays and radio

Open questions to be addressed in the near future:

- Are the outbursts permanently modifying the timing/spectral parameters of the source ?

- Are different outbursts behaving similarly ? More than one type of outbursts ?

Application of more physical spectral models

The Swift detection of a burst from an AXPs increased the chances to identify other outburst form known AXPs and to select new AXPs.

08 May 7 SAIt08

The 2006 burst Forest

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)

SGR1900+14 as observed by Swift on 29SGR1900+14 as observed by Swift on 29thth March 2006 March 2006

Time resolved spectroscopy Resulted in 729 spectra with an average # of photons of 4000 and t in the 8-400m range

08 May 7 SAIt08

2BB: time-resolved evolution

<kTs>=4.8+\-0.3 keV<Rs>=30+\-2 kmPossibly related to the Magnetosphere

<kTh>=9+\-0.3 keV<Rh>=5.7+\-0.5 kmPossibly related to the NSsurface

kTs/kTh ~ 0.5Rs/Rh ~ 6

Given the statistics and the time resolution it rather interesting to look at the T and R distributions (Olive et al. 2004 on a 4s-long IF)

08 May 7 SAIt08

◊ Olive 2004

The KT - R2 plane Sharp edge / saturation present for the brightest part of the bursts

20-25km

10-13keV

The 2BB distributions identify a natural sepa- ration surface at 20-25km

A sort of turn/cut-offis present for the BBharound 10-13keV and 5-15km

The locus identifiedby the relation R2kT3=c can be regarded as constant number of emitted per unit time (R2kT4=c identifies a constant L)

kT-3

kT-4

08 May 7 SAIt08

2BB: Lsoft vs Lhard

BBs and BBh Luminosities correlatebelow 3x1041 erg/s. Above such value only BBh increases. BBs saturation effect

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.

08 May 7 SAIt08

THE proposed/qualitative scenario

O-mode photosphere

E-O-mode splitting photosphere

E-mode photosphere

OE O

O-e--e+

B<BQED =4.4x1013 G

A possible interpretation: different way with which E- and O-mode polarised photons propagate into the magnetosphere (TD95); scattering cross section of E-mode is reduced by B and the scattering photosphere is smaller (~R NS).

The R=30km corresponding to the max radius of the BBh and the

min of the BBs identify a critical

surfaceat which B=Bcrit (QED).

08 May 7 SAIt08

Conclusions - I

TAXP studies seems to be a very powerful tool to study the emission mechanism(s) from AXPs and their evolution. The exceptional case of XTEJ1810-197 and its radio emission is even more important allowing to identify a common emission region for X-rays and radio

Open questions to be addressed in the near future:

- Are the outbursts permanently modifying the timing/spectral parameters of the source ?

- Are different outbursts behaving similarly ? More than one type of outbursts ?

Application of more physical spectral models

The Swift detection of a burst from an AXPs increased the chances to identify other outburst form known AXPs and to select new AXPs.

08 May 7 SAIt08

Conclusions - IIWe carried out a detailed spectroscopic study of the bursting component of SGR 1900+14 thanks to a “dream dataset“ collected by Swift on March 2006(Israel et al. 2008). We tested the magnetar model scenario [detail level not sufficient!].

• a 2BB spectral distribution is confirmed to be the best multi-comp. model

• Discovery of a correlation between T and BB surface (kT3=const) which holds for the brightest (or initial) part of the bursts. It gives info on the different emission mechanisms involved (Compton versus photon splitting)

• a break of the Ls/Lh=const relation, above 1041 erg/s saturation effect in the soft component. Interpreted in the framework of the magnetar model as the existence of E- and O-mode populations

• The maximum L of the hard BB (@ 15km and 10keV) is similar to the inferred LEdd,B at the same radius for B~8x1014G (inferred form P and Pdot). Moreover, the super- shown by the soft BB at large distance (100km) testifies of the presence of a dynamically active magnetic confinement

• short burst and IF do not behave differently each other. They form a continuous in terms od duration and fluence