XMM-Newton multi-year campaign on NGC 55 ULX-1: Resolving the wind and its variability with RGS

Abstract

Winds are an important ingredient in the evolution of X-ray binary (XRB) systems, particularly those at high accretion rates such as ultra-luminous X-ray sources (ULXs), because they may regulate the accretion of matter onto the compact object. We aim at understanding the properties of ULX winds and their link with the source spectral and temporal behavior. We performed high-resolution X-ray spectroscopy of the variable source NGC 55 ULX-1 to resolve emission and absorption lines as observed with XMM-Newton at different epochs. Optically-thin plasma models are used to characterise the wind. We confirmed and thoroughly strengthened previous evidence of outflows in NGC 55 ULX-1. The presence of radiative recombination signatures and the ratios between the fluxes of the emission lines favours photoionisation balance and low-to-moderate densities, which confirm that the lines originate from classical XRB disc winds. An in-depth parameter space exploration shows line emission from a slowly moving, cool, and variable plasma perhaps associated with a thermal wind. Mildly-relativistic Doppler shifts (about -0.15c) associated with the absorption lines confirm, at higher confidence, the presence of powerful, radiatively-driven, winds. The comparison between results obtained at different epochs revealed that the wind responds to the variability of the underlying continuum and these variations may be used to understand the actual accretion regime and the nature of the source.

Figures (18)

• Figure 1: Swift/XRT long-term light curve (top panel) and hardness-ratio curve of NGC 55 ULX-1 (bottom panel). The hardness ratio is computed as the ratio between the counts in 2--10 keV / 0.3--2 keV energy bands. The values for the 4 on-axis deep XMM-Newton observations are shown with horizontal lines and the times of the events with "X" marks followed by the corresponding obsid.
• Figure 2: XMM-Newton combined spectra of ULX-1 obtained by stacking those from the four deep, on-axis, observations: RGS $1^{\rm st}$ (light-blue) and $2^{\rm nd}$-order (black), EPIC-pn (grey), MOS 1 (red) and MOS 2 (green).
• Figure 3: XMM-Newton time-averaged (on-axis) spectra and continuum model. Labels are same as in Fig. \ref{['fig:spec_tavg']}. The EPIC data were ignored between 0.5-1.8 keV in order to fully employ the high spectral resolution of RGS and decrease degeneracy between models of line emission / absorption.
• Figure 4: Gaussian line scan of the XMM-Newton time-averaged spectrum (top panel) and spectra of the individual observations. Labelled are the energy centroids of some among the most common X-ray spectral lines with Fe K referring to the Fe I fluorescence and the Fe xxv resonant lines. The grey vertical line indicates the separation between RGS and EPIC. The grey-shaded areas indicate the $3\,\sigma$ single-trial significance.
• Figure 5: XMM-Newton RGS time-averaged spectrum (order 1 in blue, order 2 in black) and gaussian emission lines: zooming on the Mg-Ne K, Fe L and O K edges. Notice the weakness of Fe lines, the stronger O vii forbidden line and the tentative RRC profiles (hints of photoionisation).