Fourier resolved spectroscopy of 4U 1728-34: New Insights into Spectral and Temporal Properties of Low-Mass X-ray Binaries

TL;DR

This study uses Fourier-resolved spectroscopy of 4U 1728-34 to analyze its spectral and timing properties across different states, revealing that the thermal component dominates variability and that the iron line shows little short-term variability.

Contribution

It provides new insights into the spectral components responsible for variability in low-mass X-ray binaries, especially the role of the boundary layer and the behavior of the iron line.

Findings

The blackbody component explains source variability in banana states.

The iron line at 6.6 keV shows little variability on 10^{-2} to 10^2 second timescales.

Spectral models vary between states, with powerlaw components needed in the island state.

Abstract

Using archival RXTE data we derive the 2-16 keV Fourier-resolved spectra of the Atoll source 4U 1728-34 in a sequence of its timing states as its low QPO frequency spans the range between 6 and 94 Hz. The increase in the QPO frequency accompanies a spectral transition of the source from its island to its banana states. The banana-states' Fourier-resolved spectra are well fitted by a single blackbody component with $kT \sim 2-3$ keV depending on the source position in the color -- color diagram and the Fourier frequency, thus indicating that this spectral component is responsible for the source variability on these timescales. This result is in approximate agreement with similar behavior exhibited by the Z sources, suggesting that, as in that case, the boundary layer -- the likely source of the thermal component -- is supported by radiation pressure. Furthermore, it is found that the iron line at $\sim$6.6 keV, clearly present in the averaged spectra, not apparent within the limitations of our measurements in the frequency-resolved spectra irrespective of the frequency range. This would indicate that this spectral component exhibits little variability on time scales comprising the interval $10^{-2}-10^2$ seconds. In the island state the single blackbody model proved inadequate, particularly notable in our lowest frequency band ($0.008-0.8$ Hz). An absorbed powerlaw or an additive blackbody plus hard powerlaw model was required to obtain a satisfactory fit. Statistics do not allow unambiguous discrimination between these possible scenarios.