1.1. High frequency quasi-periodic oscillations in low-mass X-ray binaries
One of the greatest successes of RXTE is the discovery of HFQPOs in the X-ray emission of low-mass X-ray binaries (LMXBs). RXTE made it possible by the combination of the large effective area of the PCA, an excellent time resolution, an extended telemetry bandwidth and optimized observing efficiency. HFQPOs ranging from to Hz have been reported so far from 21 neutron star LMXBs (Van der Klis 1999): the 6 Z sources, 12 known atoll sources, the unknown bursting source in the galactic center region, the recently discovered X-ray bursters XTE J2123-058 and XTE J1723-376 (Marshall & Markwardt 1999). Following their discoveries, it has been realized that they could possibly be used to constrain the fundamental parameters of neutron stars, such as their spin periods, their masses, and their radii. Implications of general relativity in the interpretation of these HFQPOs have also been discussed by several authors like, e.g., Kaaret et al. (1997), Kluzniak (1998), Miller et al. (1998b).
In most systems, twin HFQPOs are observed with a separation in the range 220-360 Hz. Changes in the source luminosity cause the twin peaks to shift simultaneously in a way that their separation remains approximatively constant. For the Z source Sco X-1 and the atoll sources 4U1608-52, 4U1728-34 and 4U1735-44, however, the frequency separation significantly decreases as the HFQPOs frequencies increase (Van der Klis et al. 1997; Méndez et al. 1998c; Méndez & Van der Klis 1999; Ford et al. 1998b; Psaltis et al. 1998). In a few X-ray bursters (e.g. 4U1702-43, KS1731-260), which are atoll sources, coherent pulsations have been seen during type I X-ray bursts at frequencies equal or twice the frequency difference between the twin HFQPOs (Markwardt et al. 1999; Smith et al. 1997). In 4U1636-53 and 4U1728-34, burst oscillations have been detected at a frequency significantly greater than the twin HFQPOs separation (Méndez et al. 1998a; Méndez & Van der Klis 1999).
Specific also to the atoll sources, HFQPOs are detected when the source is in the so-called island state and on the lower branch of the banana state. These two states are easily identified in color-color diagrams and there is conclusive evidence that the accretion rate increases from the island to the banana state (Hasinger & Van der Klis 1989). They are associated with different types of power density spectra (PDS) at low frequencies (in the remainder of this paper we define low frequencies as frequencies below 100 Hz). In the island state, the source displays high frequency noise (HFN also called "flat topped noise") and the PDS can be approximated by a zero-centered Lorentzian (Olive et al. 1998, for 1E1724-3045), or alternatively by broken power laws (Wijnands & Van der Klis 1999a, for example). Superposed to the noise, a QPO-like feature is generally seen around Hz (see Table 1 for a review). In this state, the root mean square (RMS) values of the fluctuations (to the mean flux) reach 50% (Van der Klis 1994). On the other hand, on the banana branch, the shape of the PDS is roughly a power law with an index of -1.0; the so-called very low frequency noise (VLFN) or red noise (Hasinger & Van der Klis 1989). In this state, the RMS is around 10% or less. In the lower part of the banana branch, a broad structure, often refered to as "peaked HFN", or a QPO is seen at frequencies in the range 1-70 Hz (see Table 1). A correlation between the centroid frequency of the bump, and the frequency of the upper kHz QPO has been found in several LMXBs (e.g. KS1731-260), and interpreted in the framework of a Lense-Thirring effect (Stella & Vietri 1998; Stella et al. 1999) or a two-oscillator model (Titarchuk & Osherovich 1999; Osherovich & Titarchuk 1999a).
Table 1. Sources for which a QPO (or a QPO-like feature or peaked HFN) has been reported in addition to either flat topped noise or red noise in their low frequency ( Hz) PDS.
In the context of these results, we have analyzed our RXTE observations of the type I X-ray burster 4U1915-05. This source is known as the first dipping X-ray source discovered, providing the first reliable evidence for the binary nature of such sources (Walter et al. 1982; White & Swank 1982; Swank et al. 1984; Smale et al. 1988).
The period of the dips in X-rays is 50 minutes, the shortest among dipping sources (White & Swank 1982; Walter et al. 1982; Yoshida et al. 1995; Chou et al. 1999). It is also optically identified with a 21st V magnitude blue star (Grindlay et al. 1988). A modulation is seen in the optical but at a period 1% larger than the X-ray dip period (Callanan et al. 1995). The discrepancy between the two periods is confirmed by Chou et al. (1999) who combined historical and recent data (including the RXTE data used in this paper). This discrepancy has led to several interpretations among which the SU UMa 1 model is prefered, although the hierarchical triple system model proposed by Grindlay et al. (1988) is still possible (Chou et al. 1999).
To date, the only timing study of the non-dipping and non-bursting emission of 4U1915-05 was carried out with the Ginga data, which represented the most extensive X-ray observations of the source before the launch of BeppoSAX and RXTE. The 1-37 keV luminosity varied between and ergs s-1, in the range of luminosities previously observed for the source. A V-shape was observed in the hardness (7-18/1-7 keV) intensity (1-18 keV) diagram (Yoshida 1992). The Hz PDS were fitted with a power law of index -0.6 for the left branch of the V data and of index roughly -1.1 for the right branch data. Yoshida (1992) suggested that 4U1915-05 was an atoll source. Yoshida (1992) interpreted the left branch of the hardness intensity diagram as the island state and the right branch as the banana state, although the two branches overlapped in the color-color diagram.
From an X-ray burst showing photospheric radius expansion in the GINGA data, the distance to the source was estimated as 9.3 kpc (Yoshida 1992). 4U1915-05 was observed at energies up to 100 keV by BeppoSAX (Church et al. 1998). This was the first observation of a hard X-ray tail in the spectrum of 4U1915-05. Indeed, previous pointed and monitoring observations at high energies with OSSE and BATSE aboard the Compton Gamma-Ray Observatory had failed to detect the burster (Barret et al. 1996; Bloser et al. 1996). 4U1915-05 is the first dipping source to belong to the growing group of LMXBs detected up to keV (Barret et al. 2000).
In this paper we report on the timing analysis of the X-ray emission which led to the discovery of HFQPOs and LFQPOs in 4U1915-05. In the next section, we describe the observations and the data analysis. Then, we present the results. Finally, we compare 4U1915-05 to similar sources and discuss the results in the framework of currently proposed models of HF and LFQPOs.
© European Southern Observatory (ESO) 2000
Online publication: September 5, 2000