XIV Advanced School on Astrophysics

Topic III: Observations of the Accretion Disks of Black Holes and Neutron Stars

III.3: Accretion Disks of Non-Magnetic Neutron Stars

Ron RemillardKavli Institute for Astrophysics and Space ResearchMassachusetts Institute of Technology

http://xte.mit.edu/~rr/XIVschool_III.3.ppt

IV.3 X-ray States of Accreting NSs

Classifying Atolls, Z-sources, and X-ray Pulsars Subclass Inventory and Spectral Shapes Color-Color and Hardness-Intensity Diagrams X-ray Spectra and Power-Density Spectra

Soft and Hard States of Atoll Sources X-ray Spectra and the Model Ambiguity Problem The L vs. T4 question for Neutron Stars Interpreting the Boundary Layer and the Hard State

Z Sources Z Source Properties and the Two Subroups XTE J1701-462: the first Z-type transient Phenomenological and Spectral Results for XTEJ 1701 Physical Models for Z-Branches and Vertices

Inventory of Neutron-Star X-ray Sources

Subtype Typical Characteristics Number Transients

Accretors:

Atoll Sources LMXBs; X-ray bursters ~100 ~60

Msec X-ray Pulsars (182-599 Hz) ; atoll-like X-spectra 8 8

Z-sources high- Lx LMXBs; unique spectra/timing 9 1

HMXB or Pulsars hard spectrum; P > 3 d.; many X-pulsars ~90 ~50---------------

Non-accreting:

Magnetars Soft Gama Repeaters (4 + 1 cand.) 14 7

Anomalous X-ray Pulsars (8 + 1 cand.)

Other Isolated Pulsars young SNRs; X-detect radio pulsars 70? 0?

---------- ---------

Totals 291 126

X-ray Transients in the Milky Way

RXTE ASM:

47 Persistent Sources > 20 mCrab (1.5 ASM c/s)

83 Galactic Transients(1996-2008; some recurrent)

Transients: timeline of science opportunities.

Accreting NS Subclasses

HMXB/pulsar (o)

Hard spectra: e.g., power-law photon index < 1.0 at 1-20 keV;

easiest distinguished via gross spectral shape

weakly magnetized, accreting NS ()

BH Binaries andcandidates (squares)

filled symbol: persistentopen symbol: transient

Cackett et al. (2006)

Accreting NS Subclasses

Atolls and Z-sources: X-ray spectra are soft when source is bright ; types distinguished with color-color and hardness-intensity diagrams.

choose 4 energy bands {A, B, C, D} in order of increasing energy soft color = B/A hard color = D/C

atoll transient bright atoll source Z source

extreme island, island, banana branch horizontal, normal, andand banana branches (upper and lower) flaring (here dipping) braches

Top to bottom:

Accreting NS Subclasses

Atolls and Z-sources: LMXBs with binary periods < 2 d. diverse and complex phenomenology (van der Klis 2006; Strohmayer & Bildsten 2006)

Spectra in different states/branches disk & boundary layer

Power rms/shape in each state/branch disk & boundary layer

Type I X-ray bursts NS & thermonuclear burning

Burst Oscillations (show NS spin) NS & thermonuclear burning

Superbursts NS & thermonuclear burning

Low-frequency QPOs (0.1 – 50 Hz) disk?

kHz QPOs (200-1300 Hz) disk?

Accreting NS Subclasses

Atolls and Z-sources: LMXBs with binary periods < 2 d. diverse and complex phenomenology (van der Klis 2006; Strohmayer & Bildsten 2006)

Spectra in different states/branches disk & boundary layer

Power rms/shape in each state/branch disk & boundary layer

Type I X-ray bursts NS & thermonuclear burning

Burst Oscillations (show NS spin) NS & thermonuclear burning

Superbursts NS & thermonuclear burning

Low-frequency QPOs (0.1 – 50 Hz) disk?

kHz QPOs (200-1300 Hz) disk?

Energy Spectra & Power Spectra of Accreting NS

Atoll-type Transients: Aql X-1, 4U1608-52

RXTE ASM:10 outbursts per source

Atoll-type Transients: combine all outburstshard color: 8.6-18 / 5.0-8.6 keV ; soft color 3.6-5.0 / 2.0-3.6 keV

soft (banana), transitional (island), hard (extreme island) states

Atoll Spectra: Model Ambiguity (25 year debate)

Eastern Model: A multi-color disk (MCD) + Comptonized blackbody (BB)

Western model: BB + Comptonized MCD

For each, Comptonization can be a simple slab model (Tseed, Tcorona), or an uncoupled, broken power law (BPL).

All fits are good!

Hard state: hot corona; moderate opt. depth; cool BB or MCD; Compton dominates Lx

Soft state: 3 keV corona; high opt. depth; thermal and Compton share Lx

Performance Test: L (MCD. BB?) vs. T

Eastern Model: MCD behavior unacceptable in soft state Western model: BB Lx is not T4, in soft state, but physics of boundary layer evolution is a complex topic. Never see disk!!

hard state: Lx growth is closer to T4 line (i.e., constant, radius).

LMCD

(1038 erg/sat 10 kpc)

--------------

LBB

(1038 erg/sat 10 kpc)

Solution to problem with atoll soft state?

Lin, Remillard, & Homan 2007

soft state: BB+MCD+weak BPL (constrained < 2.5 ; Ebreak = 20)like double-thermal model of Mitsuda et al. 1984

hard state: Western (BB+BPL)….like BH hard state + boundary layer!

LMCD

and

LBB

(1038 erg/sat 10 kpc)

top line:R = Rburst

lower line:R = 0.25 Rburst

Rns< RISCO?TMCD and TBB TMCD and TBB

Power rms vs. Comptonization fraction

Double-themal model: atolls and BH very similarIn rms power vs. Comptonization fraction

rms power in power density spectrumvs.

fraction of energy (2-20 keV) for Comptonization

Black Holes:

2 Atoll transients

Double-thermal Model: States vs. LBB

If dm/dt (disk) = dm/dt (BL), then hard state has higher rad. efficiency than thermal state.

Alternatively, along L(BPL+MCD), the hard state shows 6X less dm/dt reaching the NS surface, compared to the soft state.

Neither conclusion may hold if there are important geometry issues, e.g. distributing some mass outside the visible boundary layer area during the hard state.

Does LBB track

M-dot at the

NS surface ?

ASM Light Curves of bright Z Sources

GX5-1

GX340+0

Cyg X-2

Sco X-1

GX349+2

GX17+2

Two groups of Z sources (Kuulkers et al. 1994)

RXTE Obs. (several ks) 1996-2005; This group mainly occupies Normal Branch (NB) and Flaring Branches (FB)

GX349+2 GX17+2

Z Sources: Sco X-1 group

FB

NB

HB

RXTE Observations 1996-2005 (each several ks)

GX340+0 GX5-1

Z Sources: Cyg X-2 group

HB

NB

FB

Cyg X-2

RXTE observations

“Z” moves around

more than other sources

Z Source: Cyg X-2

Properties of Z-branches in GX 5-1

Flaring Branch (FB)

Normal Branch (NB)

Horizontal Branch (HB)

Spectral Fits for Z SourcesBeppoSAX Obs. of GX17+2 (Di Salvo et al. 2000)

Horizontal Branch: 8% power law (1-200 keV). ; Normal branch: no hard tail

upper HB lower NB

Spectral Fits for Z SourcesBeppoSAX Obs. of GX349+2 (Di Salvo et al. 2001; see also D’Amico et al. 2001)

Normal Branch vertex has hard tail ; Flaring branch is usually very soft

2006-2007

First and only

Z-type transient

RXTE: 866 obs.

3 Ms archive

Transient Z-Source, XTE J1701-462

RXTE: 866 obs.

3 Ms archive

Horizontal (HB)

Normal (NB)

Flaring (FB)

NB-FB Vertex

Transient Z-Source, XTE J1701-462

Cyg-like

………….... Sco-like Z source…..…….

atoll

6 samples of the

evolving Z pattern

over the outburst

Homan et al. 2007

Lin, Remillard & Homan 2008

XTE J1701-462 Samples of Z’s

Light curve color-color HID-steady HID-variable

double-thermal

model

(disk+BB+CBPL)

XTE J1701-462 Spectral FitsColor-color spectral fit: Lx vs. T

Cyg-Like Z

Sco-like Z

Atoll Stage Reference lines:Radius from burstsFit to constant RBB

XTE J1701-462 Spectral Fits

FB: disk shrinks at constant dM/dt

TR (M dM/dt R-3)1/4

L R2 TR4

L (M dM/dt)2/3 T4/3

not much change in disk NB: BB increases Rat constant T

HB: Cyg-likeAnd Sco-like ZsAppear different?

Atoll stage: both disk &BB/boundary layer

exhibit L T4 (constant R)

double-thermal

model

(disk + BB + CBPL)

Lin, Remillard, & Homan 2008

XTE J1701-462 Spectral FitsSpectral Fit Results

Lx vs. T R vs. count rate

Upper and lower vertices form single lines on the HID.

Lower vertex is a key to understanding global evolution and

the physical processes for adjoining branches, i.e. the FB and NB.

XTE J1701-462: Total Hardness-Intensity Diagram

NB:FB Vertex: local Eddington limit in the accretion disk?

Lower Z-vertex (NB:FB)

FB: disk tries to shrink toward ISCO from a point on this curve

NB:FB Flaring Branch NB:FB Flaring Branch

Vertex Vertex

Evolution Speed along the FB

NB:FB vertex appearsmore stable than the FB

HB:NB Vertex: expansion of both disk and boundary layer with Lx

what causes this turning point?

Upper Z-vertex (HB:NB)

Comparing Comptonizarion

(fraction of flux in CBPL)

with rms power fraction

from PDS

Increased continuum power

in Cyg-like HB (only)

tied to boundary layer,

not power-law spectrum

(confirming conclusion

of Gilfanov et al. 2003)

Comptonization & rms in power continuum

Does Compton energy

along the HB

come from the disk?

Top panels: L(disk)

Bottom: L(disk + CBPL)

Comptonization in the HBSamples Ia and IIIa All HB and upper vertex

Hasinger & van der Klis 1990: Increasing dM/dt along

HB NB FB

Sco-like Z sources and dM/dt

HB

NB

FB

Lin, Remilard, & Homan 2008:In a local Z, dM/dt is almost constantwith possible slight increase along NB

Secular increases in dM/dt drive up the Z in the HID, while shifting the emphasis from the FB and lower vertex toward the upper vertex and the HB.

Local Eddington limit is first seen in disk, and the NB:FB vertex maps the disk response of RMCD to Lx (i.e., dM/dt), while RBB ~ constant.

Sco-like Z source phase:

At any point in the RMCD vs. Lx curve, the disk may try to shrink back towards the ISCO, which appears as movement along the FB

Along the NB, the boundary layer brightens independently from the disk, perhaps in the onset of a radial accretion flow (small fraction of total)

the HB shows the onset of Comptonization; the HB:NB vertex appears

to be more stable than the NB, but its nature is somewhat mysterious.

XTE J1701-462: summary

Cyg-like Z source phase (higher dM/dt):

the FB is the dipping type, and the spectral model does not fit the data well, thus preventing our interpretation

along the NB, the boundary layer brightens, similar to the Sco-like phase but there are also changes in the disk, complicating interpretations

HB-upturn shows increased Comptonization, resembling the Sco-like HB

The non-upturn HB shows a large jump in rms without increased CBPL flux. The disk loses energy, while the boundary layer shows a slight gain and appears to be responsible for the rms power.

Next investigations:

Use this spectral model to study kHz QPOs in all Z and atoll sources.

XTE J1701-462: summary

Reviews:Strohmayer & Bildsten 2006 (see reference list in Lecture 1)

Van der Klis 2006 (see reference list in Lecture 1)

Additional References:D’Amico et al. 2001, ApJ, 547, L147

DiSalvo et al. 2000, ApJ, 544, L119

DiSalvo et al. 2001, ApJ, 554, 49

Gilfanov et al. 2003, A&A, 410, 217

Hasinger et al. 1990, A&A, 235, 131

Homan et al. 2007, ApJ, 656, 420

Kuulkers et al. 1994, A&A, 289, 795

Lin, Remillard, & Homan 2007, ApJ, 667, 1073

Lin, Remillard, & Homan 2008, to be submitted Aug. 2008

Mitsuda et al. 1984, PASJ, 36, 741

References