No Arabic abstract
X-ray reverberation lags have recently been discovered in both active galactic nuclei (AGN) and black hole X-ray binaries. A recent study of the neutron star low-mass X-ray binary 4U 1608-52 has also shown significant lags, whose properties hint at a reverberation origin. Here, we adapt general relativistic ray tracing impulse response functions used to model X-ray reverberation in AGN for neutron star low-mass X-ray binaries. Assuming relativistic reflection forms the broad iron line and associated reflection continuum, we use reflection fits to the energy spectrum along with the impulse response functions to calculate the expected lags as a function of energy over the range of observed kHz QPO frequencies in 4U 1608-52. We find that the lag energy spectrum is expected to increase with increasing energy above 8 keV, while the observed lags in 4U 1608-52 show the opposite behavior. This demonstrates that the lags in the lower kHz QPO of 4U 1608-52 are not solely due to reverberation. We do note, however, that the models appear to be more consistent with the much flatter lag energy spectrum observed in the upper kHz QPO of several neutron star low-mass X-ray binaries, suggesting that lower and upper kHz QPOs may have different origins.
Studying the reflection of X-rays off the inner edge of the accretion disk in a neutron star low-mass X-ray binary, allows us to investigate the accretion geometry and to constrain the radius of the neutron star. We report on a NuSTAR observation of 4U 1608-52 obtained during a faint outburst in 2014 when the neutron star, which has a known spin frequency of 620 Hz, was accreting at ~1-2% of the Eddington limit. The 3-79 keV continuum emission was dominated by a Gamma~2 power law, with a ~1-2% contribution from a kTbb~0.3-0.6 keV black body component. The high-quality NuSTAR spectrum reveals the hallmarks of disk reflection; a broad iron line peaking near 7~keV and a Compton back-scattering hump around ~20-30 keV. Modeling the disk reflection spectrum points to a binary inclination of i~30-40 degrees and a small `coronal height of h<8.5 GM/c2. Furthermore, our spectral analysis suggests that the inner disk radius extended to Rin~7-10 GM/c2, close to the innermost stable circular obit. This constrains the neutron star radius to R<21 km and the redshift from the stellar surface to z>0.12, for a mass of M=1.5 Msun and a spin parameter of a=0.29.
We test the proposed 3-component spectral model for neutron star low mass X-ray binaries using broad-band X-ray data. We have analysed 4 X-ray spectra (0.8-30 keV) obtained with Suzaku during the 2010 outburst of 4U 1608-52, which have allowed us to perform a comprehensive spectral study covering all the classical spectral states. We use a thermally Comptonized continuum component to account for the hard emission, as well as two thermal components to constrain the accretion disc and neutron star surface contributions. We find that the proposed combination of multicolor disc, single-temperature black body and Comptonization components successfully reproduces the data from soft to hard states. In the soft state, our study supports the neutron star surface (or boundary layer) as the dominant source for the Comptonization seed photons yielding the observed weak hard emission, while in the hard state both solutions, either the disc or the neutron star surface, are equally favoured. The obtained spectral parameters as well as the spectral/timing correlations are comparable to those observed in accreting black holes, which support the idea that black hole and neutron star low mass X-ray binaries undergo a similar state evolution during their accretion episodes.
(abridged) We studied the energy and frequency dependence of the Fourier time lags and intrinsic coherence of the kHz QPOs in the NS LMXBs 4U 1608-52 and 4U 1636-53 using RXTE data. In both sources we confirmed energy-dependent soft lags of 10-100 mu s for the lower kHz QPO. We also found that the time lags of the upper kHz QPO are independent of energy and inconsistent with the soft lags of the lower kHz QPO. The intrinsic coherence of the lower kHz QPO remains constant at 0.6 from 5 to 12 keV, and then drops to zero, while for the upper kHz QPO the intrinsic coherence is consistent with zero across the full energy range. The intrinsic coherence of the upper kHz QPO is consistent with zero over the full frequency range of the QPO, except in 4U 1636-53 at ~780 Hz where it increases to 0.13. In 4U 1636-53, for the lower kHz QPO the 4-12 keV photons lag the 12-20 keV ones by 25 mu s in the QPO frequency range 500-850 Hz, with the lags decreasing to 15 mu s at higher frequencies. In 4U 1608-52 the soft lags of the lower kHz QPO remain constant at 40 mu s. In 4U 1636-53, for the upper kHz QPO the 12-20 keV photons lag the 4-12 keV ones by 11 +/- 3 mu s, independent of QPO frequency; we found consistent results for the time lags of the upper kHz QPO in 4U 1608-52. The intrinsic coherence of the lower kHz QPO increases from ~0-0.4 at 600 Hz to 1 and 0.6 at 800 Hz in 4U 1636-53 and 4U 1608-52, respectively. In 4U 1636-53 it decreases to 0.5 at 920 Hz, while in 4U 1608-52 we do not have data above 800 Hz. We discuss our results in the context of scenarios in which the soft lags are either due to reflection off the accretion disc or up-/down-scattering in a hot medium close to the neutron star. We finally explore the connection between, on one hand the time lags and the intrinsic coherence of the kHz QPOs, and on the other the QPOs amplitude and quality factor in these two sources.
We collect the data of twin kilohertz quasi-periodic oscillations (kHz QPOs) published before 2012 from 26 neutron star (NS) low-mass X-ray binary (LMXB) sources, then we analyze the centroid frequency ( u) distribution of twin kHz QPOs (lower frequency u_1 and upper frequency u_2) both for Atoll and Z sources. For the data without shift-and-add, we find that Atoll and Z sources show the different distributions of u_1, u_2 and u_2/ u_1, but the same distribution of Delta u (difference of twin kHz QPOs), which indicates that twin kHz QPOs may share the common properties of LXMBs and have the same physical origins. The distribution of Delta u is quite different from constant value, so is u_2/ u_1 from constant ratio. The weighted mean values and maxima of u_1 and u_2 in Atoll sources are slightly higher than those in Z sources. We also find that shift-and-add technique can reconstruct the distribution of u_1 and Delta u. The K-S test results of u_1 and Delta u between Atoll and Z sources from data with shift-and-add are quite different from those without it, and we think that this may be caused by the selection biases of the sample. We also study the properties of the quality factor (Q) and the root-mean-squared (rms) amplitude of 4U 0614+09 with the data from the two observational methods, but the errors are too big to make a robust conclusion. The NS spin frequency ( u_s) distribution of 28 NS-LMXBs show a bigger mean value (about 408Hz) than that (about 281 Hz) of the radio binary millisecond pulsars (MSPs), which may be due to the lack of the spin detections from Z sources (systematically lower than 281 Hz). Furthermore, on the relations between the kHz QPOs and NS spin frequency u_s, we find the approximate correlations of the mean values of Delta u with NS spin and its half, respectively.
It is commonly assumed that the properties and geometry of the accretion flow in transient low-mass X-ray binaries (LMXBs) significantly change when the X-ray luminosity decays below $sim 10^{-2}$ of the Eddington limit ($L_{rm Edd}$). However, there are few observational cases where the evolution of the accretion flow is tracked in a single X-ray binary over a wide dynamic range. In this work, we use NuSTAR and NICER observations obtained during the 2018 accretion outburst of the neutron star LMXB 4U 1608-52, to study changes in the reflection spectrum. We find that the broad Fe-K$alpha$ line and Compton hump, clearly seen during the peak of the outburst when the X-ray luminosity is $sim 10^{37}$ erg/s ($sim 0.05$ $L_{rm Edd}$), disappear during the decay of the outburst when the source luminosity drops to $sim 4.5 times 10^{35}$ erg/s ($sim 0.002$ $L_{rm Edd}$). We show that this non-detection of the reflection features cannot be explained by the lower signal-to-noise at lower flux, but is instead caused by physical changes in the accretion flow. Simulating synthetic NuSTAR observations on a grid of inner disk radius, disk ionisation, and reflection fraction, we find that the disappearance of the reflection features can be explained by either increased disk ionisation ($log xi geq 4.1$) or a much decreased reflection fraction. A changing disk truncation alone, however, cannot account for the lack of reprocessed Fe-K$alpha$ emission. The required increase in ionisation parameter could occur if the inner accretion flow evaporates from a thin disk into a geometrically thicker flow, such as the commonly assumed formation of an radiatively inefficient accretion flow at lower mass accretion rates.