The column density towards LMC X-1

We measure the neutral absorption towards the black hole X-ray binary system LMC X-1 from six archival soft X-ray spectra obtained with the gratings and/or CCD detectors on Chandra, XMM-Newton, and Swift. Four spectral models for the soft continuum have been investigated. While the powerlaw model may overestimate NH considerably, the others give consistent results. Taking the lower metalicity of the Large Magellanic Cloud into account, we find equivalent hydrogen column densities of N_H = (1.0-1.3)*10^22 cm^-2, with a systematic dependence on the orbital phase. This variation of the neutral absorption can nearly explain the orbital modulation of the soft X-ray flux recently detected with the All Sky Monitor (ASM) on the Rossi X-ray Timing Explorer (RXTE).

💡 Research Summary

The authors present a detailed study of the neutral hydrogen column density (N_H) toward the black‑hole X‑ray binary LMC X‑1, using six archival soft‑X‑ray observations obtained with the gratings and CCD detectors on Chandra, XMM‑Newton, and Swift. The data set includes a Chandra HETGS observation (C1), two XMM‑Newton pointings (X1 with EPIC‑pn and RGS, X2 with RGS only), and three Swift XRT observations (S1 in PC mode, S2 and S3 in WT mode). Standard reduction pipelines (CIAO/TGCat for Chandra, SAS v7.1 for XMM‑Newton, and HEASOFT for Swift) were applied, and problematic energy ranges (e.g., 1.5–2 keV in Swift WT) were excluded to minimise calibration uncertainties.

A central methodological point is the adoption of LMC‑specific elemental abundances in the absorption model. Table 2 lists the logarithmic abundances (ε) for He, C, N, O, Ne, Mg, Si, S, Ar, and Fe, showing that the LMC metallicity is roughly 40–90 % of the Galactic values. The authors therefore used the tbvarabs model with these abundances rather than the default Galactic composition, because soft‑X‑ray absorption is dominated by metals and using Galactic abundances would underestimate the true N_H.

Four continuum models were tested: (1) a multicolour disk blackbody (diskbb) plus a simple power‑law, (2) the empirical convolution model simpl applied to diskbb, (3) simpl applied to the relativistic disk model kerrbb, and (4) the physical Comptonisation model eqpair. The simple power‑law model, with a very steep photon index (Γ≈3.7 for the best‑covered X1 spectrum), forces an artificially high low‑energy absorption to compensate for the excess low‑energy flux, leading to N_H values as high as ~1.8×10^22 cm⁻². In contrast, the simpl‑based models introduce an intrinsic low‑energy cutoff, breaking the spurious N_H–Γ correlation (Fig. 2) and yielding consistent column densities in the range (1.0–1.3)×10^22 cm⁻². The eqpair fits give comparable results, confirming that the choice of a physically realistic continuum does not bias the absorption measurement.

The authors then examined the dependence of N_H on orbital phase (φ_orb). Using the ephemeris from Orosz et al. (2009), they assigned phases to each observation. The column density shows a clear modulation: observations taken near φ≈0 (when the black hole is behind the O‑type donor) have higher N_H (≈1.25×10^22 cm⁻²), whereas those near φ≈0.5 (black hole in front) have lower values (≈1.05×10^22 cm⁻²). A sinusoidal fit to the six measurements yields a mean N_H of 1.15×10^22 cm⁻² and a semi‑amplitude of ≈0.15×10^22 cm⁻², corresponding to a full amplitude of ~3×10^21 cm⁻². The systematic spread among the different continuum models is <8×10^20 cm⁻², indicating that the modulation is robust against model choice.

High‑resolution RGS spectra reveal absorption lines of Ne IX (13.45 Å) and possibly Ne II (13.62 Å), suggesting a modest contribution from ionised material, although the signal‑to‑noise is insufficient for a detailed ionisation analysis.

Finally, the authors compare the inferred N_H modulation with the orbital modulation of the soft X‑ray flux reported by the RXTE‑ASM (Levine & Corbet 2006). The ASM shows a ~7 % modulation in the 1.5–5 keV band and a smaller ~4 % modulation at higher energies. The authors calculate that a column density variation of 3×10^21 cm⁻² would produce flux changes of 7.7 % (1.5–3 keV), 1.6 % (3–5 keV), and 0.4 % (5–12 keV), which matches the observed amplitudes within uncertainties. This suggests that neutral absorption, rather than pure Thomson scattering in the stellar wind, can largely account for the observed X‑ray variability. However, the phase of the sinusoidal N_H fit does not perfectly align with the ASM light curve, implying that the wind structure may be more complex than a simple sinusoid.

In summary, the paper demonstrates that (i) incorporating LMC‑specific metallicities is essential for accurate N_H determination, (ii) continuum model choice can significantly bias N_H if a simple power‑law is used, (iii) LMC X‑1 exhibits a genuine orbital‑phase‑dependent neutral absorption likely arising from the donor’s stellar wind, and (iv) this absorption can explain most of the soft X‑ray flux modulation seen by RXTE‑ASM. The authors recommend further phase‑resolved soft X‑ray observations (e.g., additional Chandra pointings) to map the wind geometry more precisely and to test whether ionised absorption contributes significantly.