Swift observations of two supergiant fast X-ray transient prototypes in outburst

We report on the results from observations of the most recent outbursts of XTE J1739-302 and IGR J17544-2619, which are considered to be the prototypes of the supergiant fast X-ray transient (SFXT) class. They triggered the Swift/BAT on 2011 February 22 and March 24, respectively, and each time a prompt Swift slew allowed us to obtain the rich broad-band data we present. The XRT light curves show the descending portion of very bright flares that reached luminosities of ~2x10^{36} and ~5x10^{36} erg/s, respectively. The broad-band spectra, when fit with the usual phenomenological models adopted for accreting neutron stars, yield values of both high energy cut-off and e-folding energy consistent with those obtained from previously reported outbursts from these sources. In the context of more physical models, the spectra of both sources can be well fitted either with a two-blackbody model, or with a single unsaturated Comptonization model. In the latter case, the model can be either a classical static Comptonization model, such as COMPTT, or the recently developed COMPMAG model, which includes thermal and bulk Comptonization for cylindrical accretion onto a magnetized neutron star. We discuss the possible accretion scenarios derived by the different models, and we also emphasize the fact that the electron density derived from the Comptonization models, in the regions where the X-ray spectrum presumably forms, is lower than that estimated using the continuity equation at the magnetospheric radius and the source X-ray luminosity, and we give some possible explanations.

💡 Research Summary

This paper presents a detailed analysis of the most recent outbursts from the two prototypical supergiant fast X‑ray transients (SFXTs), XTE J1739‑302 and IGR J17544‑2619, as observed by the Swift satellite. Both sources triggered the Burst Alert Telescope (BAT) on 22 February 2011 (XTE J1739‑302) and 24 March 2011 (IGR J17544‑2619). The automatic slew of Swift allowed the X‑Ray Telescope (XRT) to start observing within minutes, providing continuous coverage of the decaying phase of each flare over a broad energy range (0.3–150 keV).

The XRT light curves reveal that the flares reached peak luminosities of roughly 2 × 10³⁶ erg s⁻¹ for XTE J1739‑302 and 5 × 10³⁶ erg s⁻¹ for IGR J17544‑2619, consistent with the extreme variability that defines the SFXT class. The authors first fitted the broadband spectra with the phenomenological models traditionally used for accreting neutron stars (power‑law with high‑energy cut‑off, cutoff power‑law). The derived cut‑off energies (E_cut ≈ 12–18 keV) and e‑folding energies (E_fold ≈ 15–25 keV) match those obtained in previous outbursts, indicating that the basic spectral shape is stable across events.

To explore the underlying physics, the authors then applied more physically motivated models. A two‑blackbody description reproduces the data with a cool component (kT₁ ≈ 0.6–0.8 keV) and a hotter component (kT₂ ≈ 1.8–2.2 keV), suggesting emission from a small hot spot on the neutron‑star surface together with a surrounding, slightly cooler region. The spectra are also well described by unsaturated Comptonization models. Using the static COMPTT model yields electron temperatures of kT_e ≈ 5–7 keV and optical depths τ ≈ 8–12. The more recent COMPMAG model, which incorporates both thermal and bulk Comptonization in a cylindrical accretion column onto a magnetised neutron star, provides an equally good fit. In this framework, bulk motion of the infalling plasma (β ≈ 0.2–0.3) contributes significantly to the high‑energy tail, naturally accounting for the hard X‑ray emission observed early in the flare.

A striking result emerges from both Comptonization fits: the inferred electron density in the region where the X‑ray spectrum forms is n_e ≈ 10¹⁸ cm⁻³, an order of magnitude lower than the density expected from the continuity equation at the magnetospheric radius (R_m ≈ 10⁹ cm) given the measured X‑ray luminosity (L_X ≈ 10³⁶ erg s⁻¹). The authors discuss several possible explanations. One possibility is that the radiative region lies well inside the magnetospheric boundary, in the free‑fall zone where the density is reduced. Another is that anisotropic radiation patterns and a reduced effective optical depth lead to an apparent low density. A third scenario involves a clumpy stellar wind combined with a more ordered stream, producing a mixed flow where the average density is lower than the simple spherical‑infall estimate.

In summary, the Swift observations provide high‑quality, time‑resolved broadband spectra of two SFXT prototypes during bright outbursts. The data confirm that phenomenological cut‑off models remain valid, while physically motivated Comptonization models offer deeper insight into the accretion geometry and plasma conditions. The discrepancy between the electron density derived from spectral modeling and that predicted by simple continuity arguments highlights a key unresolved issue in our understanding of SFXT accretion physics. The authors conclude that future observations with more sensitive hard X‑ray instruments (e.g., NuSTAR, HXMT) together with three‑dimensional magnetohydrodynamic simulations will be essential to resolve this tension and to clarify the mechanisms that drive the rapid, extreme variability of SFXTs.