Abstract

We have investigated the long-term flux variation in Cen X-3 using orbital modulation and pulsed fraction in different flux states using observations made with the All-Sky Monitor and the Proportional Counter Array on board the Rossi X-ray Timing Explorer. In the high state, the eclipse ingress and egress are found to be sharp whereas in the intermediate state the transitions are more gradual. In the low state, instead of eclipse ingress and egress, the light curve shows a smooth flux variation with orbital phase. The orbital modulation of the X-ray light curve in the low state shows that the X-ray emission observed in this state is from an extended object. The flux-dependent orbital modulations indicate that the different flux states of Cen X-3 are primarily due to varying degree of obscuration. Measurement of the pulsed fraction in different flux states is consistent with the X-ray emission of Cen X-3 having one highly varying component with a constant pulsed fraction and an unpulsed component and in the low state, the unpulsed component becomes dominant. The observed X-ray emission in the low state is likely to be due to scattering of X-rays from the stellar wind of the companion star. Though we cannot ascertain the origin and nature of the obscuring material that causes the aperiodic long-term flux variation, we point out that a precessing accretion disc driven by radiative forces is a distinct possibility.

1 INTRODUCTION

Several persistent X-ray binaries show large flux variations in their X-ray light curves on time-scales significantly longer than their orbital periods (Wen et al. 2006). Highly periodic flux variations at time-scales larger than their respective orbital periods are seen in Her X-1 (35 d: Tananbaum et al. 1972; Still & Boyd 2004), LMC X-4 (30.28 d: Paul & Kitamoto 2002), 2S 0114+650 (30.75 d: Farrell, Sood & O'Neill 2006), SS 433 (164 d: Eikenberry et al. 2001; Fabrika 2004), XTE J1716–389 (99.1 d: Cornelisse, Charles & Robertson 2006), 4U 1820–303 (172 d: Smale & Lochner 1992; Zdziarski et al. 2007a) and Cyg X-1 (150 d: Kitamoto et al. 2000; Lachowicz et al. 2006). The first three of these seven sources, known with stable superorbital period, are accretion-powered pulsars. Several other X-ray binaries show quasi-periodic long-term flux variations. SMC X-1 has a long-term period in the range of 50–70 d (Gruber & Rothschild 1984; Clarkson et al. 2003; Wen et al. 2006). GRS 1747-312 shows quasi-periodic flux variations of ∼150 d (in't Zand et al. 2003; Wen et al. 2006). Cyg X-2 shows quasi-periodic long-term variability with a period in the range of 60–90 d (Smale & Lochner 1992; Paul, Kitamoto & Makino 2000), and LMC X-3 also shows long-term flux variations with periodicity in the range of 100–500 d (Wen et al. 2006) that is unstable over a longer period (Paul et al. 2000; Wen et al. 2006). Long-term quasi-periodic variations with a period of about 217 d are also seen from the Rapid Burster (X1730–333) which is in the form of recurrent outburst rather than a gradual change in X-ray flux (Masetti 2002).

Cen X-3 is a high-mass X-ray binary pulsar with a spin period of ∼4.8 s and an orbital period of ∼2.08 d, discovered with UHURU (Giacconi et al. 1971). This X-ray binary is known to have strong long-term flux variations with transient quasi-periodicity (Priedhorsky & Terrell 1983). The long-term light curve of this source obtained with the All-Sky Monitor (ASM) of the Rossi X-ray Timing Explorer (RXTE) showed very strong aperiodic flux variations along with two different accretion modes (Paul, Raichur & Mukherjee 2005). In 12 yr of monitoring data obtained with the RXTE-ASM in three different energy bands, the hardness ratio between the 5–12 and 3–5 keV bands was found to have a larger value during 2000 December to 2004 April, compared to the same outside this period. Long-term observations with Burst and Transient Source Experiment (BATSE) of the Compton Gamma-RayObservatory (CGRO) found that Cen X-3 has an alternate spin-up and spin-down intervals which last from about 10 to 100 d (Finger, Wilson & Fishman 1994). However, GINGA observations revealed that there is no correlation between the observed X-ray flux and pulse period derivative of Cen X-3 (Tsunemi, Kitamoto & Tamura 1996), and it was suggested that the observed X-ray flux of Cen X-3 does not represent its mass accretion rate.

The periodic long-term variations seen in X-ray binaries are explained by variable obscuration of the central X-ray source by an accretion disc precessing under tidal force. The tidal precession can be in the form of forced precession of accretion discs in the gravitational field of the companion star (Katz 1973) or slaved precession of accretion disc due to the rotation axis of the star being inclined (Roberts 1974). Tilted and twisted discs due to coronal winds (Schandl & Meyer 1994; Schandl 1996) and radiation pressure-induced warped precessing discs (Iping & Petterson 1990; Maloney, Begelman & Pringle 1996; Wijers & Pringle 1999; Ogilvie & Dubus 2001) can have variable or chaotic precession periods. Iping & Petterson (1990) simulated the temporal evolution of radiatively warped accretion disc, and found that it can give rise to aperiodic variability in the X-ray light curves. They suggested that the long-term variability of Cen X-3 could be caused by such a radiatively warped accretion disc. Other more detailed disc warping models have also been developed where warping instabilities are incorporated into the models which give rise to unstable and chaotic disc precession with no stable long-term period (Wijers & Pringle 1999; Ogilvie & Dubus 2001). Accretion torque-induced precession of magnetic axis can also be a possible explanation for the observed flux variations (Truemper et al. 1986) in some X-ray binaries. Third-body mechanism is also known to cause a periodic X-ray flux variation by modulating the mass accretion rate (Zdziarski, Wen & Gierliński 2007b).

In this work, we investigate the orbital modulation of the 1.5–12 keV X-ray flux of Cen X-3 when the source is in high, intermediate and low states to understand if the aperiodic variations are occurring due to a variable mass accretion rate or the variable obscuration of the central X-ray source by an accretion disc. The eclipse structure is useful to know the size of the observed X-ray emission region in different flux states. We have also measured the pulsed and total X-ray emission of the source in its different flux states using many observations by the Proportional Counter Array (PCA) of the RXTE. Evolution of the pulsed fraction with observed X-ray flux is useful to know the relative importance of scattered X-ray emission in different flux states.

2 OBSERVATIONS AND ANALYSIS

The ASM on board RXTE has three detectors which scan the sky in a series of 90 s dwells in three energy bands, namely 1.5–3, 3–5 and 5–12 keV (Levine et al. 1996). We have used the 1.5–12 keV light curves of four sources Cen X-3, Her X-1, SMC X-1 and LMC X-4, covering about 4100 d from 1996 January to study and compare the orbital modulation of these sources in three different flux states. The light curves of the four sources binned with their respective orbital periods after excluding the eclipse data are shown in Fig. 1, for 500 d. The distinctly aperiodic flux variation of Cen X-3 is in sharp contrast with the periodic modulation in Her X-1 (with one main-on and one short-on state), LMC X-4 (with smaller signal-to-noise ratio) and the quasi-periodic modulation in SMC X-1 (with some scatter within each high state).

Figure 1

The ASM light curves of Cen X-3, Her X-1, SMC X-1 and LMC X-4 are shown here for 500-d binned with the orbital period of the respective sources. The Cen X-3 light curve clearly shows aperiodic superorbital variations, Her X-1 and LMC X-4 light curves show periodic flux variations whereas SMC X-1 light curve shows quasi-periodic flux variations. The numbered arrows in the Cen X-3 light curve represent the times during which the respective pulse profiles of Fig. 4 are taken.

Figure 1

The ASM light curves of Cen X-3, Her X-1, SMC X-1 and LMC X-4 are shown here for 500-d binned with the orbital period of the respective sources. The Cen X-3 light curve clearly shows aperiodic superorbital variations, Her X-1 and LMC X-4 light curves show periodic flux variations whereas SMC X-1 light curve shows quasi-periodic flux variations. The numbered arrows in the Cen X-3 light curve represent the times during which the respective pulse profiles of Fig. 4 are taken.

For each of the three sources Cen X-3, Her X-1 and SMC X-1, we have made three separate light curves, one each for the high-, intermediate- and low-flux states. To do this, we first calculated the average count rate per orbit after excluding the eclipse data (see Paul et al. 2005 for more details). Depending on the average count rate in a binary orbit, the data points available during that orbit were collected in one of the three high-, intermediate- or low-state light curves (see Table 1 for the interval of count rates defining different flux states for the three sources). These three light curves of each source were then folded with the respective orbital period of the source to get the orbital modulation light curves.

Table 1

The source parameters.

Source Orbital period1 (d) Eclipse duration (d) Orbit-averaged count rate for: 
High Intermediate low 
Cen X-3 2.087 06 ± 0.000 09 0.52 ≥18.0 18.0 − 2.0 ≤2.0 
Her X-1 1.700 15 ± 0.000 09 0.22 ≥2.5 2.5 − 1.0 ≤1.0 
SMC X-1 3.8921 ± 0.0004 0.62 ≥1.3 1.3 − 0.7 ≤0.7 
LMC X-42 1.408 40 ± 0.000 06 0.23 – – – 
Source Orbital period1 (d) Eclipse duration (d) Orbit-averaged count rate for: 
High Intermediate low 
Cen X-3 2.087 06 ± 0.000 09 0.52 ≥18.0 18.0 − 2.0 ≤2.0 
Her X-1 1.700 15 ± 0.000 09 0.22 ≥2.5 2.5 − 1.0 ≤1.0 
SMC X-1 3.8921 ± 0.0004 0.62 ≥1.3 1.3 − 0.7 ≤0.7 
LMC X-42 1.408 40 ± 0.000 06 0.23 – – – 

1Orbital periods are measured from RXTE-ASM light curves.

2The high, intermediate and low states of LMC X-4 were determined using the phases of its stable superorbital period. See Section 2 for details.

We have adopted a slightly different analysis for LMC X-4 since its signal-to-noise ratio is small. We first folded the ASM light curve with the superorbital period and determined the ephemeris for superorbital modulation as follows  

1
formula
where N is an integer. The high, intermediate and low states were determined based on the superorbital phase being 0.35 to 0.55 (0.20 to 0.35, 0.55 to 0.80) and −0.20 to 0.20, respectively. Data points from the full ASM light curve belonging to the three different states were then used to make three different light curves and folded with the orbital period of LMC X-4 to get the orbital modulation light curves.

In Fig. 2, we have shown the orbital modulation of the four sources in different flux states. For each source, the orbital modulation light curves of the high- and intermediate-flux states are shown in the top panel, high-state points are indicated by circles. The low-state modulations are shown in the bottom panels. All the orbital modulation curves are normalized by dividing the original light curves by the average count rate calculated over an orbital phase of 0.2 near the peak flux of the respective orbital modulation curve. The low-state ASM light curve for LMC X-4 does not show any orbital modulation. In Table 1, we have also given the orbital periods and eclipse durations of these four sources determined from the RXTE-ASM light curves. In Fig. 3, we have shown the flux-dependent orbital modulation light curve of Cen X-3 separately for the hard spectral state that started in 2000 December and ended in 2004 April (Paul et al. 2005).

Figure 2

The orbital modulation in high, intermediate and low states is shown here for four X-ray sources Cen X-3, Her X-1, SMC X-1 and LMC X-4. The high- and intermediate-state plots are shown in the upper panel, with the circles denoting the high-state points. The plot in the lower panel shows the low-state orbital modulation. All the plots are normalized by dividing the original curve with the respective maximum count rate of the curve.

Figure 2

The orbital modulation in high, intermediate and low states is shown here for four X-ray sources Cen X-3, Her X-1, SMC X-1 and LMC X-4. The high- and intermediate-state plots are shown in the upper panel, with the circles denoting the high-state points. The plot in the lower panel shows the low-state orbital modulation. All the plots are normalized by dividing the original curve with the respective maximum count rate of the curve.

Figure 3

Flux-dependent orbital modulation of Cen X-3 in hard spectral state during 2000 December to 2004 April. Upper panel plots show high- and intermediate-state orbital modulation, with high-state points being marked with circles. Lower panel plot shows low-state orbital modulation.

Figure 3

Flux-dependent orbital modulation of Cen X-3 in hard spectral state during 2000 December to 2004 April. Upper panel plots show high- and intermediate-state orbital modulation, with high-state points being marked with circles. Lower panel plot shows low-state orbital modulation.

We have also investigated the pulsation characteristics of Cen X-3 at different flux levels in the 2–60 keV band. Cen X-3 was observed by RXTE-PCA many times during the years 1996, 1997 and 1998. All these observations have been obtained when Cen X-3 was in the soft spectral state. We have chosen 18 PCA observations depending on the orbit-averaged ASM count rates at that time such that a wide range of X-ray flux is covered. The 2–60 keV band light curves were obtained from the Standard-1 mode data of the PCA. Background light curves were generated using the background models provided for RXTE-PCA by the High Energy Astrophysics Science Archive Research Centre (HEASARC) and were subtracted from Standard-1 light curves to get the source light curves. The source light curves were first barycentre corrected and then searched for the spin period of the neutron star. We did not detect any pulsations when the orbit-averaged ASM count rate of Cen X-3 was less than 0.8 count s−1 (equivalent to about 50 count s−1 per proportional counter unit) with a 90 per cent upper limit of 0.8 per cent on the pulsed fraction. The pulse profiles were then generated by folding the barycentre-corrected light curves by the respective spin period found in that light curve. To avoid smearing of the pulse profiles due to the orbital motion of the pulsar, the pulse profiles were generated from short data segments of duration of a hundred pulses. In Fig. 4, we have shown six pulse profiles obtained at different source flux levels including a light curve folded in the low-flux state with a period of 4.81 s when no pulses were detected. Fig. 5 is a plot of the maximum–minimum count rate per Proportional Counter Unit (PCU) against the maximum count rate per PCU for the 18 pulse profiles. The points marked with the circles are for the pulse profiles shown in Fig. 4. Epochs of the six PCA observations, pulse profiles from which are shown in Fig. 4, are marked in the top panel of Fig. 1. A two-component function was fitted to the points in Fig. 5:  

2
formula

Figure 4

Pulse profile of Cen X-3 is shown here in different flux states of the source. The count rate is per PCU. Hundred consecutive pulses are folded with arbitrary pulse phase and the local spin period to get each of the above pulse profiles.

Figure 4

Pulse profile of Cen X-3 is shown here in different flux states of the source. The count rate is per PCU. Hundred consecutive pulses are folded with arbitrary pulse phase and the local spin period to get each of the above pulse profiles.

Figure 5

Pulsed emission of Cen X-3 is plotted here against the maximum count rate per detector over a range of X-ray flux of the source. The points marked with circles correspond to the pulse profiles shown in Fig. 4. The solid line in the figure is a fit to the function given in the text.

Figure 5

Pulsed emission of Cen X-3 is plotted here against the maximum count rate per detector over a range of X-ray flux of the source. The points marked with circles correspond to the pulse profiles shown in Fig. 4. The solid line in the figure is a fit to the function given in the text.

The pulse fraction f of the pulsating component was determined to be 90 per cent, while the unpulsed component grows up to a count rate of F0= 175.5 per PCU.

As seen in Fig. 4, apart from the changing ratio of the unpulsed component of the flux to the total flux, the pulse shape is also varying from one observation to another. To see whether the pulse shape changes are related to the flux, we selected three different observations with very similar eclipse-subtracted orbit average ASM count rates of 17.12 ± 1.35, 17.52 ± 1.21 and 17.73 ± 1.47. Within these three observations, the pulse shape changed from a double peak profile to a broad single peak profile but the pulsed fraction remained the same. We therefore conclude that while the pulsed flux of Cen X-3 is related to the total flux, the pulse shape is independent of the X-ray flux.

3 INTERPRETATION AND DISCUSSION

We have found that in the four sources Cen X-3, Her X-1, SMC X-1 and LMC X-4, the binary orbital modulation of the X-ray flux shows remarkable dependence on the X-ray flux state (Fig. 2). X-ray eclipses are found to be sharp in the high-flux state which becomes more gradual in the intermediate state. In the low state, instead of sharp eclipse ingress and egress, there is a smooth flux variation with orbital phase. Though the orbital modulation of LMC X-4 is not detectable in the low state of LMC X-4 with ASM data, we note that a weak and smooth orbital modulation in the low state of LMC X-4 was clearly detected earlier with BeppoSAX observations (Naik & Paul 2003). The Her X-1 orbital light curves show some pre-eclipse dips (Fig. 2). In the intermediate state of Cen X-3, the eclipse egress starts earlier in phase by 0.03 compared to the high state as is shown with a vertical dashed line.

Using a subset of the RXTE-ASM light curve for Her X-1, Scott & Leahy (1999) also found that the orbital modulation is shallower in the short-on state compared to the same in the main-on state. In the BeppoSAX observations of LMC X-4, the absence of clear eclipse transitions in low state was interpreted as due to obscuration of the central X-ray source by the precessing accretion disc. The weak orbital modulation of the light curve in the low state is due to an extended X-ray scattering region, emission from which dominates the detectable X-rays in the low state.

The flux-dependent orbital modulations of these four sources indicate that at lower flux, an increasing fraction of the observed X-rays are from a larger region, comparable to the size of the companion star. The larger emission region may have different visibility at different orbital phases, leading to the smooth orbital modulation in the low state. Reprocessing of X-rays emitted from the compact star by scattering from stellar wind of the companion star is one likely scenario. However, for Her X-1, which has a low-mass companion star, the scattering medium is more likely to be a part of the accretion disc and disc outflows than the stellar wind. Zdziarski et al. (2007a) have performed a similar analysis of RXTE-PCA light curve of the X-ray binary 4U 1820–303, and found a significant dependence of the profile of the orbital modulation on the average count rate. However, in 4U 1820–303, the superorbital modulation is associated with an accretion rate modulation, probably due to third-body interaction. Poutanen, Zdziarski & Ibragimov (2008) discovered a superorbital phase dependence of the soft X-ray orbital modulation in Cyg X-1 in its hard spectral state, that is related to the size of a bulge in the outer accretion disc.

The ratio of the X-ray flux when the source is in eclipse and when it is out-of-eclipse is a measure of the relative scattering efficiency. Only for Cen X-3 there is good enough statistics to compare the ratios. We find that the out-of-eclipse count rates differ by a large factor (22.02 ± 0.07, 7.56 ± 0.02 and 1.23 ± 0.02 for high, intermediate and low states, respectively) while the in-eclipse count rates of the three states are comparable (0.69 ± 0.07 for high, 0.48 ± 0.03 for intermediate and 0.27 ± 0.03 for low state). The eclipse count rate varies by a factor of only about ∼2.5 while the out of eclipse count rate varies by a factor of ∼18. The ratio of X-ray flux of Cen X-3 during eclipse and out-of-eclipse is larger in the low state by a factor of 7.0 ± 1.3 compared to the same in the high state. This behaviour is similar to SMC X-1, in which the eclipse count rate was found to be comparable in high and low states whereas the out-of-eclipse count rate varied by more than a factor of 20 (Wojdowski et al. 1998).

The measurement of pulsed X-ray flux as a function of the peak X-ray flux of Cen X-3 as presented in the Figs 4 and 5, is also consistent with a scenario in which the measured X-ray flux has two components. One component is highly variable with a pulsed fraction of about 90 per cent and a second component that is unpulsed. Similar result was obtained for SMC X-1 over a wide range of its measured X-ray flux (Kaur et al. 2007). In the low state, the unpulsed component becomes dominant leading to non-detection of pulses. We also note that in all the four sources mentioned here, the X-ray pulsations (pulse period of ∼4.8 s in Cen X-3, 1.24 s in Her X-1, 0.7 s in SMC X-1 and 13.5 s in LMC X-4) have never been detected during the low state of the superorbital period, indicating that most of the radiation observed in low state is probably reprocessed emission from a large region (Wojdowski et al. 1998; Naik & Paul 2003). The non-detection of pulsations is not due to faintness of the sources. Even in the low state, Cen X-3, Her X-1 and SMC X-1 are bright enough for detection of a few per cent pulse modulation with the RXTE-PCA.

We point out the possibility that the intrinsic X-ray luminosity of the central X-ray source may remain unchanged for a long time. It can be seen in Fig. 1 that except during the anomalous low-state events of Her X-1, the peak luminosity of superorbital modulation of Her X-1, SMC X-1 and LMC X-4 does not change very much and the peak of the 5–12 keV X-ray luminosity of Cen X-3 has also a ceiling (see fig. 1 of Paul et al. 2005). Though the results presented here indicate that the different flux states of Cen X-3 are largely due to varying degree of obscuration, as is the case with Her X-1, SMC X-1 and LMC X-4, we cannot completely rule out some contribution to the variability from a varying mass accretion rate, especially the variations associated with spectral change. But as shown in Fig. 3, even during the hard spectral state of Cen X-3 from 2000 December to 2004 April, the orbital modulation light curve indicates an extended X-ray source at low flux level. Detailed observations over a wide-spectral range with future missions like ASTROSAT can throw more light on these aperiodic variations and spectral mode changes seen in Cen X-3 light curves.

4 CONCLUSIONS

  • The binary orbital modulation of X-ray from Cen X-3 is similar to that seen in the other three accreting X-ray pulsars. From sharp eclipse in high state, it turns to a gradual modulation in the low state. Cen X-3 eclipse egress starts earlier in the intermediate state compared to the high state. These observations indicate a larger emission region in the low state of Cen X-3. The ratio of X-ray flux of Cen X-3 during eclipse and out-of-eclipse is larger in the low state by a factor of 7.0 ± 1.3 compared to the same in the high state.

  • A measurement of the pulsed X-ray flux in different flux states of Cen X-3 is consistent with the X-ray flux having two components, one with a large pulsed fraction and a second unpulsed component that dominates in the low state.

  • We propose that the long-term intensity variations in Cen X-3 are mostly due to aperiodic obscuration of the compact source by the accretion disc. The unpulsed X-ray emission from an extended region appears to be due to scattering of the X-rays from the central source by the stellar wind.

We thank the referee A. Zdziarski for very useful suggestions that helped us very much to improve this paper. This research has made use of data obtained from the High Energy Astrophysics Science Archive Research Centre (HEASARC), provided by NASA's Goddard Space Flight Centre.

REFERENCES

Clarkson
W. I.
Chales
P. A.
Coe
M. J.
Laylock
S.
Tout
M. D.
Wilson
C. A.
,
2003
,
MNRAS
 ,
339
,
447
Cornelisse
R.
Charles
P. A.
Robertson
C.
,
2006
,
MNRAS
 ,
366
,
918
Eikenberry
S. S.
Cameron
P. B.
Fierce
B. W.
Kull
D. M.
Dror
D. H.
Houck
J. R.
Margon
B.
,
2001
,
ApJ
 ,
561
,
1027
Fabrika
S.
,
2004
,
Astrophys. Space Phys. Rev.
 ,
12
,
1
Farrell
S. A.
Sood
R. K.
O'Neill
P. M.
,
2006
,
MNRAS
 ,
367
,
1457
Finger
M. H.
Wilson
R. B.
Fishman
G. J.
,
1994
, in
Friedlander
M. W.
Gehrels
N.
, eds, AIP Conf. Proc. Vol. 280,
Compton Gamma-Ray Observatory
 .
Am. Inst. Phys.
,
New York
, p.
304
Giacconi
R.
Gursky
H.
Kellogg
E.
Schreier
E.
Tananbaum
H.
,
1971
,
ApJ
 ,
167
,
67
Gruber
D. E.
Rothschild
R. E.
,
1984
,
ApJ
 ,
283
,
546
In't Zand
J. J. M.
et al
,
2003
,
A&A
 ,
406
,
233
Iping
R. C.
Petterson
J. A.
,
1990
,
A&A
 ,
239
,
221
Katz
J. I.
,
1973
,
Nature Phys. Sci.
 ,
246
,
87
Kaur
R.
Paul
B.
Raichur
H.
Sagar
R.
,
2007
,
ApJ
 ,
660
,
1409
Kitamoto
S.
Egoshi
W.
Miyamoto
S.
Tsunemi
H.
Ling
J. C.
Wheaton
W. A.
Paul
B.
,
2000
,
ApJ
 ,
531
,
546
Lachowicz
P.
Zdziarski
A. A.
Schwarzenberg-Czerny
A.
Pooley
G. G.
Kitamoto
S.
,
2006
,
MNRAS
 ,
368
,
1025
Levine
A. M.
Bradt
H.
Cui
W.
Jernigan
J. G.
Morgan
E. H.
Remillard
R.
Shirey
R. E.
Smith
D. A.
,
1996
,
ApJ
 ,
469
,
33
Maloney
P. R.
Begelman
M. C.
Pringle
J. E.
,
1996
,
ApJ
 ,
472
,
582
Masetti
N.
,
2002
,
A&A
 ,
381
,
L45
Naik
S.
Paul
B.
,
2003
,
A&A
 ,
401
,
265
Ogilvie
G. I.
Dubus
G.
,
2001
,
MNRAS
 ,
320
,
485
Paul
B.
Kitamoto
S.
,
2002
,
JA&A
 ,
23
,
33
Paul
B.
Kitamoto
S.
Makino
F.
,
2000
,
ApJ
 ,
528
,
410
Paul
B.
Raichur
H.
Mukherjee
U.
,
2005
,
A&A
 ,
442
,
L15
Poutanen
J.
Zdziarski
A. A.
Ibragimov
A.
,
2008
,
MNRAS
 , submitted (arXiv:0802.1391)
Priedhorsky
W. C.
Terrell
J.
,
1983
,
ApJ
 ,
273
,
709
Roberts
W. J.
,
1974
,
ApJ
 ,
187
,
575
Schandl
S.
,
1996
,
A&A
 ,
307
,
95
Schandl
S.
Meyer
F.
,
1994
,
A&A
 ,
289
,
149
Scott
D. M.
Leahy
D. A.
,
1999
,
ApJ
 ,
510
,
974
Smale
A. P.
Lochner
J. C.
,
1992
,
ApJ
 ,
395
,
582
Still
M.
Boyd
P.
,
2004
,
ApJ
 ,
606
,
L135
Tananbaum
H.
Gursky
H.
Kellogg
E. M.
Levinson
R.
Schreier
E.
Giacconi
R.
,
1972
,
ApJ
 ,
174
,
L143
Truemper
J.
Kahabka
P.
Oegelman
H.
Pietsch
W.
Voges
W.
,
1986
,
ApJ
 ,
300
,
63
Tsunemi
H.
Kitamoto
S.
Tamura
K.
,
1996
,
ApJ
 ,
456
,
316
Wen
L.
Levine
A. M.
Corbet
R. H. D.
Bradt
H. V.
,
2006
,
ApJS
 ,
163
,
372
Wijers
R. A. M. J.
Pringle
J. E.
,
1999
,
MNRAS
 ,
308
,
207
Wojdowski
P.
Clark
G. W.
Levine
A. M.
Woo
J. W.
Zhang
S. N.
,
1998
,
ApJ
 ,
502
,
253
Zdziarski
A. A.
Gierliński
M.
Wen
L.
Kostrzewa
Z.
,
2007a
,
MNRAS
 ,
377
,
1017
Zdziarski
A. A.
Wen
L.
Gierliński
M.
,
2007b
,
MNRAS
 ,
377
,
1006