neutron star science with astrosat/laxpc: periodic x-ray timing and spectroscopy · 2017-02-03 ·...

Neutron Star Science with ASTROSAT/LAXPC:Periodic X-Ray Timing and Spectroscopy

Deepto ChakrabartyMIT and TIFR

Workshop on Data Analysis and LAXPC ScienceTata Institute of Fundamental Research

Mumbai, IndiaJanuary 18, 2017

Neutron Stars and Pulsars

• Neutron stars: compact stars supported against gravitational collapse by neutron degeneracy pressure

• Pulsars: rotating, highly magnetized neutron stars that emit pulsed radiation at their rotation period

• Formed in core-collapse supernova explosions from massive stellar progenitors (8-20 solar masses). Neutron star birth mass ~1.4 solar masses. Possibly also formed in accretion-induced collapse of white dwarf.

• Can be isolated (single) stars or in binaries (or even higher-order multiples)

• Millisecond pulsars: extremely fast rotators, but inferred to be old

• Classes of pulsar (based on power source for radiation):• Rotation-powered pulsars (~2000 known, including ~200 ms pulsars)• Magnetars (~20 known)• Accretion-powered pulsars (~100 known, including 15 ms pulsars)• Nuclear-powered pulsars (17 known, including 16 ms pulsars)

Spin and Magnetic Evolution of Neutron Stars1. Pulsars born with B~1012 G,

P~20 ms. Spin-down due to radiative loss of rotational K.E.

1

2 If in binary, then companion may eventually fill Roche lobe (“X-ray binary”).

1 R☉

10 R☉

100 R☉



1

2 If in binary, then companion may eventually fill Roche lobe (“X-ray binary”).

accretion disk

Accretion-Powered X-Ray Pulsars

spin axis magnetic axisdipole magnetic field

~ rm • Magnetically-channeled flow onto polar caps, hits at ~ 0.1 c. (Requires B > 108 G)

• Gravitational potential energy released as X-rays,

• Misaligned magnetic dipole axis: pulsations at spin period from X-ray hot spots at poles.

• Accretion adds mass and angular momentum (measure torque). �

L = ˙ M GMR

⎛ ⎝ ⎜

⎞ ⎠ ⎟



12

3

4

€

Peq ≈1 s B1012 G#

$ %

&

' ( 6 /7 ˙ M

10−9 MSun /yr

#

$ % %

&

' ( (

−3/7

In regions 2, 3, 4 in above plot, accretion-powered X-ray pulsations expected. Thus, expect to see millisecond X-ray pulsations in LMXBs.

2 If in binary, then companion may eventually fill Roche lobe (“X-ray binary”). Accretion spins up pulsar to equilibrium spin period

3 Sustained accretion (~109 yr) attenuates pulsar magnetic field to B~108 G, leading to equilibrium spin P~few ms

4 At end of accretion phase (companion exhausted or binary disrupted), millisecond radio pulsar remains

High-Mass X-ray Binaries (HMXBs)

(Taken from Wilms & Kreykenbohm 2012)Most neutron stars in HMXBs are X-ray pulsars.

NOTE: HMXBs are young systems (<30 Myr)! (Evolutionary argument)

1 R☉

10 R☉

100 R☉

(Slide from Wilms & Kreykenbohm 2012.Figure from Bildsten et al. 1997, after Corbet 1986)

Subclassification of HMXBs: The Corbet diagram

New class identified by INTEGRAL: “Super-fast X-ray transients”, see reviews by Sidoli. Appear to be wind-fed HMXBs.

Be/X-ray Transients (HMXBs)

(Taken from Wilms & Kreykenbohm 2012)

• Accretion torques and disk-magnetosphere interaction (see, e.g., Bildstenet al. 1997)

• Cyclotron line spectroscopy: measure magnetic field strength• Orbital decay (rapid HMXB evolution; see papers by Levine, Rappaport,

et al.)• Eclipsing systems (dynamical mass measurements, supergiant atmosphere

structure)• Physics of accretion column in strong magnetic field

Applications of Pulsars in HMXBs

accretion disk

Accretion Torques on X-Ray Pulsarsspin axis magnetic axis

dipole magnetic field

~ rm

Important length scales:

rm = magnetospheric radius, where

r

vφ

rco

Keplerian

corotating

�

B2(rm )8π

~ ρ v2(rm )

rco = corotation radius, where

�

Ωkep(rco) = Ωrot

Characteristic torque:

�

N0 = ˙ M GMrco

�

Peq ≈1 s B1012 G⎛ ⎝ ⎜

⎞ ⎠ ⎟

6 / 7 ˙ M 10−9 Msun /yr⎛

⎝ ⎜

⎞

⎠ ⎟ −3 / 7

Equilibrium spin period (rm ≈ rco):

Spin Evolution of an HMXB Pulsar Due to Accretion Torques

Chakrabarty et al. 1993

ASTROSAT simulation of cyclotron lines from 4U 0115+63

Ecycn = 11.6

✓nB

1012 G

◆(1 + z)�1 keV

Harmonic structure. Pulse phase spectroscopy. Probe magnetosphere.

Orbit of eclipsing HMXB pulsar Orbital decay of an HMXB pulsar

Chakrabarty et al. 1993 Levine et al. 1993

Low-Mass X-ray Binaries (LMXBs)

(Taken from Wilms & Kreykenbohm 2012)

(in most cases; but sometimes there are millisecond pulsations)

• Low-mass X-ray binaries with low accretion rates are subject to an ionization instability in their accretion disk. This leads to episodic accretion: X-ray transients. Duty cycle is low: X-ray transients lie dormant for months or years, then become active for a few days or weeks when disk instability is triggered.

• Active systems sometimes transition between distinct spectral states with very different emission characteristics and behavior. Certain phenomena only present during particular states.

• Need sky monitoring capability to detect transient outbursts and identify spectral state

• Need flexible scheduling of main timing instrument to catch source in correct active state of interest.

X-ray Binaries: Episodic Temporal/Spectral Variability

Millisecond Variability in NS/LMXBs

1. “Bona fide” accretion-powered millisecond X-ray pulsars (low-luminosity; measure orbit)

2. X-ray burst oscillations (low/medium luminosity; orbit inaccessible)3. Kilohertz quasi-periodic oscillations (kHz QPOs) (all luminosities; spin uncertain, orbit inaccessible)

Three distinct types of rapid X-ray variability, first identified by the Rossi X-Ray Timing Explorer (RXTE), occurring at different ranges of X-ray source luminosity:

1 2

3

Kilohertz quasi-periodic oscillations (kHz QPOs)• QPO pairs with roughly constant frequency separation (~300 Hz). QPO frequencies drift by

hundreds of Hz as X-ray flux changes (200-1200 Hz). Originally thought that separation frequency (∆ν) is a characteristic of a given source, no longer clear.

• Separation frequency ≈ νspin or ≈ (νspin/2) for cases where spin known (Fast vs Slow). Not usable as a precise spin tracer, but may give clue to spin range. However, connection to spin has been challenged as a statistical artifact (Mendez & Belloni 2007).

• Seen in over 20 bright LMXBs. Believed to originate in accretion disk. Mechanism?

4U 1608-52 (Mendez et al. 1998)

Sco X-1 (van der Klis et al. 1997)

300 1000Frequency (Hz)

200 1000 2000Frequency (Hz)

• Thermonuclear X-ray bursts due to unstable nuclear burning on NS surface, lasting tens of seconds, recurring every few hours to days.• Millisecond oscillations discovered during some X-ray bursts by RXTE (Strohmayer et al. 1996). Spreading hot spot on rotating NS surface yields “nuclear-powered pulsations”.

• Oscillations in burst tail not yet understood. Along with frequency drift, may be due to surface modes on NS. (Heyl; Piro & Bildsten; Cooper & Narayan)

• Burst oscillations suspected to be tracingspin, but coherence not really testable owingto short duration of oscillation.

Nuclear-Powered Millisecond X-Ray Pulsars (X-Ray Burst Oscillations)

thermonuclearburst

quiescent emission due to accretion

contours of oscillation power asfunction of time and frequency

X-ray burstcount rate

SAX J1808.4-3658 (Chakrabarty et al. 2003)

4U 1702-43 (Strohmayer & Markwardt 1999)

Burst oscillations in accretion-powered millisecond pulsarsSAX J1808.4-3658 (Chakrabarty et al. 2003) XTE J1814-334 (Strohmayer et al. 2003)

HETE J1900.1-2455 (Watts et al. 2009)

• Verifies that burst oscillations basically trace the spin frequency (at least within a few Hz)

• Nuclear burning ignition at a point on stellar surface should lead to pulsations in early phase of burst. This is consistent with measurements of burst oscillation during burst rise(Strohmayer et al. 1996, 1997).

• Oscillations during burst tail occur after burning has spread over entire surface, and so must have a different explanation. May be due to surface modes on NS. (Heyl; Piro & Bildsten; Cooper & Narayan). Could modes also explain early-phase oscillations?

• What causes frequency drift?

What causes X-ray burst oscillations, and how are they coupled to spin?

Accreting Millisecond X-ray Pulsar (AMXPs)

RXTE Power spectrum of SAX J1808.4-3658 (April 1998)

Wijnands & van der Klis 1998

Chakrabarty & Morgan 1998

• First example discovered in April 1998.

• Persistent X-ray pulsations at 401 Hz spin rate of neutron star.

• Circular binary orbit with 2 hr period

• Magnetic field strength estimate ~108 G.

• X-ray transient (turns out to be typical). This one now seen in 6 outbursts with RXTE.

• Other aspects: secular spin-down, orbital evolution. Optical counterpart (brown dwarf donor). Energy-dependent pulse phase lags.

• ~15 AMXPs known. Why aren’t other NS/LMXBs pulsed?

Two-hour orbit of SAX J1808.4-3658

401 Hz

Old problem: Why is it difficult to find accretion-powered millisecond pulsars?Updated: Why do we only find them in low-accretion-rate LMXB transients?

• Orbital Doppler smearing?• Many searches using acceleration techniques

• Detected systems only mildly affected; for same pulsed fraction, should not be serious

• Scattering/obscuration? (Brainerd & Lamb 1987; Kylafis & Phinney 1989)

• Suggestion that detected pulsars have low optical depths (Titarchuk et al. 2002)

• More detailed analysis finds no correlation (Gogus et al. 2007). Scattering not the answer.

• Non-magnetic accretion flow?• Intrinsically weak magnetic fields? How to reconcile with millisecond radio pulsar population?

• Screening of magnetic field by accretion? (Cumming, Zwiebel, & Bildsten 2001) Fields only penetrate for very low accretion rates, intrinsic field reemerges when accretion ends.

• Gravitational self-lensing? (Wood, Ftaclas, & Kearney 1988; Meszaros, Riffert, & Berthiaume 1988)

• Suggestion that detected pulsars, being in transients, are systematically less massive. Higher accreted mass leads to increased self-lensing, suppressing pulsation. (Ozel 2009).

• New issue: Intermittency of pulsations.

Intermittent pulsar HETE J1900.1-2455 (Galloway et al. 2007): Active outburst for >1 yr, but pulsations only during first few months. Enhanced amplitude shortly after X-ray bursts during that time.

“Intermittent” accretion-powered millisecond pulsars

Intermittent pulsar SAX J1748.9-2021 (Altamirano et al. 2007; Gavriil et al. 2007; Patruno et al. 2009). Pulsations detected intermittently during 2 of 3 outbursts. Some correlation with presence of X-ray bursts.

Intermittent pulsar Aql X-1 (Casella et al. 2007): Detected for 150 s out of ~1.5 Ms!!! Are these accretion-powered pulsations or something else?

How to explain intermittency in accretion-powered pulsations?

• Must be related to something that changes on short time scales (not orbital smearing or self-lensing)

• Magnetic screening by matter (controlled by short-term accretion rate)? Perhaps in HETE J1900.1-2455, but probably not in SAX J1748.9-2021 given location of pulsations in outburst history.

• Changes in size, shape, and/or location of “hot spot” due to accretion flow instabilities?

• At small magnetic inclination angles, 3-D MHD simulations suggest that there is a regime of unstable accretion flow such that the X-ray “hot spot” moves around erratically near the magnetic pole (Romanova et al. 2003, 2004, 2008).

• Also at these small magnetic inclinations, small changes in the location of the hot spot can lead to suppression of observable X-ray pulsations (Lamb et al. 2009).

• Such behavior might explain the abrupt jumps in pulse phase and changes in pulse shape observed in the short-term X-ray timing of several accretion-powered millisecond pulsars, particularly SAX J1808.4-3658 (Burderi et al. 2006; Hartman et al. 2008, 2009)

• What role does thermonuclear X-ray burst activity play?

• Is the extremely rare Aql X-1 pulsation the same phenomenon, or something different?

• Is the mechanism responsible for intermittency the same reason that all the known accretion-powered millisecond pulsars are in low-accretion-rate transients?

• Discovery of an eclipsing accreting millisecond X-ray pulsar• Swift J1749.4-2807, Porb=8.8hr (Markwardt & Strohmayer 2010; Altamirano et al. 2011)• Allows determination of binary inclination angle• Measurement of optical/IR radial velocity curve of companion would allow

dynamical measurement of neutron star mass

• Discovery of an 11-Hz pulsar in Terzan 5 • Terzan 5 X-2 = IGR J17480-2446• 91 ms pulsar, 21.3 hr binary (Strohmayer & Marwardt 2010; Altamirano et al. 2010;

Cavecchi et al. 2011; Papitto et al. 2011...)• Missing link in magnetic field evolution of LMXBs• Interesting burst behavior, shows importance of rotation and magnetic field

in burst properties (Linares et al. 2011, 2012; Chakraborty & Bhattacharyya 2011)

• Possible detection of non-radial mode in an AMXP• Possible detection of non-radial r-mode or g-mode in the accretion-powered

millisecond pulsar XTE J1751-305 using archival RXTE data. Also, upper limits on other systems. At edge of detection sensitivity for late-era RXTE. (Strohmayer & Mahmoodifar 2013).

Other Interesting Results on Accreting Millisecond X-ray Pulsars

• 27 systems, including both burst oscillation sources and accretion-powered pulsars

• Most objects clustered in the 100-700 Hz range. Cluster is consistent with a uniform distribution.

• One object with slow spin (11 Hz; Terzan 5 X-2 = IGR 17480-2446; Strohmayer & Markwardt 2010, Altamirano et al. 2010). Inconsistent with uniform distribution at 2-sigma level, but arguably drawn from a different population.

• No objects with rapid spin >700 Hz. Compare with expected break-up limit of ~2000 Hz. Assuming emission properties are not spin-dependent, there is no significant selection effect here.

The RXTE-measured NS/LMXB spin distribution as of 2012

Why does the spin distribution not extend well above 700 Hz?

Two leading explanations:

1. Magnetic spin equilibrium? (e.g. Lamb & Boutloukos 2007, Patruno et al. 2011)• Equilibrium spin is set by accretion rate and magnetic field strength in magnetic

accretion• Observed spin distribution reflects range of magnetic field strengths and accretion

history

2. Accretion torque balanced by gravitational radiation? (Wagoner 1984; Bildsten1998)

• Gravitational wave torque scales as ν5, from any of several proposed mechanisms• For non-zero quadrupole, GW emission will dominate above a critical spin rate

but will be negligible at slower spins: sudden “wall”. • Strain of h ~ 10-26 for best case (Bildsten 1998), if accretion torque balanced by

GWs. Direct detection of this signal by LIGO/VIRGO very difficult due to need for multiple-template search over spin/orbital parameters (Watts et al. 2008)

• No evidence for non-magnetic torques in AMXP spin-down (but not excluded)

See Bhattacharyya & Chakrabarty (2017) for a recent update.

What else can we learn from the fast end of the spin distribution?

• What is the distribution of magnetic field strengths in NS/LMXBs? The detailed shape of the spin distribution might answer this question, if one can untangle the accretion rate contribution. (This is independent of what sets the upper rate limit.)

• Is there a minimum magnetic field after sustained accretion?

• How far do the fields decay during the recycling process?

• A minimum field could explain the high-spin cutoff.

• Can a minimum field be reconciled with the presence/absence of persistent pulsations? What about magnetic screening by accretion (e.g. Cumming & Bildsten 2000). Recent paper by Patruno(2012) suggests that screening observed in intermittent pulsar HETE J19001-2455.

• Are there any submillisecond pulsars?• Discovery could significantly constrain NS equation of state

• Are these object rare/non-existent, or are they somehow hidden?

Increasing the sample of accreting millisecond pulsars will help explore these questions.

What about the slow end of the spin distribution?

• The discovery of the 11 Hz pulsar in Terzan 5 opens up an unexplored regime: the transition region between young slow pulsars and recycled millisecond pulsars.

• How does pulsar recycling and accretion-induced magnetic field decay actually work?

Linares et al. (2012)

What can we learn from the slow end of the spin distribution?

• How does accretion-induced field decay work?

• What is the decay time scale?

• How does field strength vary with accreted mass?

• How much time is spent in the recycling transition region?

• Are intermediate-field pulsars rare because of short time scale or because they are somehow hidden? Persistent versus transient systems (super-Eddington vs sub-Eddington accretion?)

• See recent paper by Patruno et al. (2012) for discussion of evolutionary history of Terzan 5 X-2

• What magnetic field strength is required to suppress type-I bursts?• Magnetic channeling expected to increase local accretion rate, stabilize burning (Joss & Li 1980)

• Slowly rotating bursters are ideal targets for detection of photospheric absorption lines (alternate approach to neutron star equation of state measurement)

LMXB/Radio Pulsar Transition Systems(Transitional Millisecond Pulsars or t-MSPs)

Several quiescent LMXBs have been observed to make transitions to a radio pulsar state when they reach very low accretion rates (Lx ~ 1032 erg/s):

• PSR J1023+0038 (Archibald et al. 2009, 2010; Bogdanov et al. 2011; Stappers et al. 2013; Patruno et al. 2014)

• M28I = PSR J1824-2452I (Papitto et al. 2013, Linares et al. 2014)

• XSS J12270-4859 (De Martino et al. 2010, 2013; Bassa et al. 2014; Bogdanov et al. 2014; Roy et al. 2014)

The X-ray spectrum during these radio states is consistent with synchrotron emission.

The sources are millisecond X-ray pulsars (AMXPs) during their LMXB state.

These systems (in radio state) would all be classified as “redbacks”.

• Accretion-powered pulsars: definitely

• Accretion-powered millisecond pulsars (AMXPs): definitely

• Nuclear-powered millisecond pulsars (burst oscillations): definitely

• Transitional millisecond pulsars: possibly – very active area of study right now.

• Luminosities are low, may be near edge of LAXPC sensitivity

Good X-ray Binary Pulsar Targets for ASTROSAT?

neutron star science with astrosat/laxpc: periodic x-ray timing and spectroscopy · 2017-02-03 ·...

Documents