Astron. Astrophys. 319, 515-524 (1997)

1. Introduction

The X ray source OAO 1657-415 was first detected by the Copernicus satellite (Polidan et al. 1978) in the 4-9 keV range. Initially association of this source with a massive star V861 Scorpii was considered. Subsequent observations by HEAO 1 satellite (Bryne et al. 1979; Armstrong et al. 1980) and the Einstein Observatory (Parmar et al. 1980) did not confirm the association with V861 Scorpii. The HEAO 1 observations also showed 38.22 sec pulsations in the 1-40 keV and 40-80 keV bands (White & Pravdo 1979; Byrne et al. 1981). Observations with Ginga and GRANAT (Kamata et al. 1990; Gilfanov et al. 1991; Mereghetti et al. 1991; Sunyaev et al. 1991) have shown episodic changes in the pulse period. Timing observations of this source with the Burst and Transient Source Experiment (BATSE) on the Compton Gamma Ray Observatory (CGRO) have shown that OAO 1657-415 is in an 11 days binary orbit with an X-ray eclipse by the stellar companion (Chakrabarty et al. 1993). The observed orbital parameters imply that the companion is a supergiant of spectral class B0-B6.

Massive X-ray binaries fall into three separate groups when the pulse periods are compared with orbital periods (Corbet 1986). The systems with Be companions show correlations between the orbital and spin periods (Corbet 1986; Waters & van Kerkwijk 1989), while systems with OB supergiant companions fall into two separate regions. X-ray pulsars with Be companions show transient behavior with episodic spin-ups (i.e EXO 2030+375, Parmar et al., 1989; A0 535+26, Finger et al., 1996) which suggest that the low velocity equatorial stellar wind forms an accretion disk during the periastron passages. Systems with OB giants with pulse periods sec and orbital periods days show optical photometric evidence for accretion disks (LMC X-1, Cen X-3, and SMC X-1) and nearly steady spin-up on longer time scales days (Finger et al., 1993). BATSE observations of Cen X-3 have shown a lot of spin-up/down episodes on time scales days, this behaviour in LMC X-1 and SMC X-1 is not known because of the long gaps between adjacent observations. Systems with longer spin periods sec) showed short term spin-up/down episodes at days (i.e Vela X-1, see Deeter et al., 1989) which can be explained in terms of flip flop instabilities of wind accretion (or very short time scale accretion disk formation; see Anzer et al., 1987; Matsuda et al., 1987; Blondin et al., 1990). OAO 1657-415 falls between these groups in terms of orbital and spin period. In order to improve our understanding of the spin history and the accretion process of the source, in this work we study the statistical properties of the angular velocity changes and investigate the possible correlations of angular acceleration with X-ray flux.

The angular velocity fluctuations in accretion powered neutron stars are produced by torques originating outside and inside the object. The external torque is carried by the accretion flow; the internal torque depends on the coupling between the core superfluid and the solid outer crust. External fluctuations of the torque are filtered by the coupling between the crust and superfluid interior to produce an output represented by the observed changes in the angular velocity. The theoretical description of torque variations in terms of noise power spectral analysis was first studied by Lamb et al., (1978a, b), who analyzed the response of a two component star to external fluctuations. Their model makes it possible to diagnose theoretically the properties of accretion flows and the internal structure of neutron stars. Techniques for estimating the noise power spectra in the case of nonuniformly sampled pulsar timing data were developed by Deeter & Boynton (1982) and Deeter (1984). They were applied to Vela X-1 and Her X-1 using the data obtained by HEAO-1 and UHURU (Boynton 1981; Deeter 1981; Boynton et al., 1984; Deeter et al., 1989). The results showed that the angular velocity time series of Vela X-1 and Her X-1 can be modelled as a random walk (or white noise in the angular accelerations with a power law index ). Recent results of BATSE observations are indicating that 4U 1626-67 has an angular velocity time series which is also consistent with the random walk model (Chakrabarty et al., 1995). Random walk in angular velocity can be characterized by a rate of torque events R of steps . The rate and the step size of the events depend on the type of accretion (Prince et al., 1994). For example in the wind accretor Vela X-1, the spin-up/down time scales are around day (Deeter et al., 1989), therefore the rate of the torque events may be days^-1 while in Her X-1 the rate is lower due to the disk accretion (Wilson et al., 1994). Time domain techniques such as autoregressive time series (Scargle 1981) were applied by Baykal & Ögelman (1993) and de Kool & Anzer (1993) in order to model the angular velocity time series of accretion powered X-ray binaries. They found that some of these systems have angular velocity time series which are consistent with random walk. Recent BATSE results for Cen X-3 and GX 1+4 have shown that the noise power spectrum is redder then the random walk model in angular velocity time series (or flicker noise in the angular accelerations with power law index ; see Finger et al., 1994; Chakrabarty et al., 1995). The reason for flicker noise in the angular accelerations may be the smooth transitions from spin-up to spin-down (or vice versa) (see Finger et al., 1994) which makes the noise power spectra steeper (or redder).

The sign and magnitude of the torque physically depends on the magnetosphere of the neutron star and the type of accretion flow exterior to magnetosphere (Ghosh & Lamb 1979 a, b; Blondin et al., 1990). One way to study the accretion dynamics of neutron stars is to observe the nature of the correlation between changes in rotation rate () and mass accretion rate () (or the related quantity X ray flux). Such a study may provide more detailed information on the physics of accretion and helps to discriminate between disk and wind-type accretion (Ghosh & Lamb 1979 a, b; Blondin et al., 1990).

In this paper, we apply various techniques in order to study the torque fluctuations of OAO 1657-415. In Sect. 2, the data base used in this analysis is described. In Sect. 3, we construct the low resolution power density spectrum using the mean-squared residuals technique developed by (Cordes 1980) and Deeter (1984). In Sect. 4, to see the sharp changes of rotation rate, a structure function is calculated Cordes (1980); the two component neutron star model and the response of the neutron star to external white torque noise is compared with the observed data. In Sect. 5, the correlations between the angular acceleration (), X-ray flux, and specific angular momentum are studied. These correlations are compared with the disk accretion hypothesis (Ghosh & Lamb 1979 a, b) and the wind accretion hypothesis (Blondin et al., 1990).

© European Southern Observatory (ESO) 1997

Online publication: July 3, 1998
helpdesk@link.springer.de