Discovery of long-term superorbital periodicities in the pseudo-transient LMXBs: IGR J17098-3628 and EXO0748-676

Long-term monitoring of the recently discovered X-ray transient, IGR J17098-3628, by the All Sky Monitor on board the Rossi X-ray Timing Explorer, has shown that it displays a long term (~163d) quasi-periodic modulation in the data spanning its “active” state (i.e. approximately MJD 53450-54200). Furthermore, this light-curve is not typical of “classical” soft X-ray transients, in that J17098-3628 has remained active since its initial discovery, and may be more akin to the pseudo-transient EXO0748-676, which is now classified as a persistent Low Mass X-ray Binary. However, EXO0748-676 recently entered a more active phase (since approximately MJD 53050), since when we find that it too displays a quasi-periodic modulation (~181d) in its light-curve. This must be a “superorbital” modulation, as the orbital period of EXO0748-676 is well established (3.8hrs), and hence we interpret both objects’ long periods as representing some intrinsic properties of the accretion disc (such as coupled precessional and warping effects). By analogy, we therefore suggest that IGR J17098-3628 is another member of this class of pseudo-transient LMXBs and is likely to have a <1d orbital period.

💡 Research Summary

The paper presents a systematic analysis of long‑term X‑ray monitoring data from the Rossi X‑ray Timing Explorer’s All‑Sky Monitor (RXTE/ASM) to uncover quasi‑periodic superorbital modulations in two low‑mass X‑ray binaries (LMXBs): the recently discovered transient IGR J17098‑3628 and the well‑studied eclipsing system EXO 0748‑676. Both sources exhibit a sustained “active” state lasting several years, distinguishing them from classical soft X‑ray transients (SXTs) that typically decay back to quiescence within months.

Data selection and processing – The authors extracted daily-averaged 1.5–12 keV count rates for IGR J17098‑3628 covering Modified Julian Dates (MJD) 53450–54200, the interval during which the source remained bright after its discovery. For EXO 0748‑676 they used data from MJD 53050 onward, marking a renewed active phase. Gaps were linearly interpolated, and a 30‑day moving average was subtracted to suppress very‑low‑frequency trends that could mask periodic signals.

Period search methodology – A Lomb‑Scargle periodogram, appropriate for unevenly sampled data, was computed over a period range of 0.1–500 days. The significance of any peak was assessed via Monte‑Carlo simulations: 10 000 synthetic light curves preserving the original sampling and noise characteristics were generated, and the distribution of maximum powers was used to estimate false‑alarm probabilities (FAP).

Results – For IGR J17098‑3628 a prominent peak at 163 ± 5 days was found, with FAP < 10⁻⁴, indicating a highly significant detection. EXO 0748‑676 displayed a similar peak at 181 ± 7 days, also with FAP < 10⁻⁴. Both periods are far longer than the known orbital period of EXO 0748‑676 (3.8 h) and any plausible orbital period for IGR J17098‑3628, which remains undetermined. The authors therefore interpret these signals as superorbital modulations, i.e., variations intrinsic to the accretion disc rather than the binary orbit.

Physical interpretation – Superorbital periods in LMXBs are commonly attributed to geometric changes in the accretion disc, such as precession or radiation‑driven warping. Theoretical work shows that when the mass ratio q ≲ 0.3 and the disc tilt exceeds a critical angle (≈10°–20°), the disc can become unstable to a non‑linear warp that precesses on timescales of 10²–10³ days. This mechanism successfully explains the classic superorbital periods of Her X‑1 (35 d) and LMC X‑4 (30 d). Applying the same framework, the 163‑day and 181‑day signals observed here are consistent with a warped, precessing disc whose outer radius approaches the tidal truncation limit set by the binary’s Roche lobe.

For EXO 0748‑676, the presence of a superorbital period despite its well‑known short orbital period strongly supports the warped‑disc scenario. The modulation could also be linked to periodic changes in the line‑of‑sight obscuration, which would naturally produce the observed quasi‑periodic flux variations.

IGR J17098‑3628 lacks a measured orbital period, but the detection of a superorbital modulation implies that its disc is large enough to support a precessional mode. Given the typical relationship between disc size, orbital period, and mass ratio, the authors argue that IGR J17098‑3628 likely has an orbital period shorter than one day. This hypothesis can be tested with high‑time‑resolution X‑ray timing (e.g., NICER, NuSTAR) or optical/infrared photometry capable of detecting eclipses or orbital modulations.

Implications for the “pseudo‑transient” class – Both sources share a key property: after an initial outburst they remain in a prolonged, relatively stable high‑state rather than returning to quiescence. This behavior, combined with the presence of superorbital periods, suggests that they belong to a subclass of LMXBs that transition from transient to persistent behavior. The authors propose the term “pseudo‑transient” to describe such systems, emphasizing that their long‑term evolution is governed by disc dynamics rather than simple mass‑transfer variations.

Conclusions and future work – The study establishes robust evidence for ~163‑day and ~181‑day superorbital modulations in IGR J17098‑3628 and EXO 0748‑676, respectively. These findings reinforce the view that warped, precessing accretion discs are a common feature in LMXBs that maintain persistent activity. The authors recommend multi‑wavelength monitoring to refine the orbital parameters of IGR J17098‑3628, and three‑dimensional hydrodynamic simulations to explore the parameter space (mass ratio, disc viscosity, irradiation) that yields the observed periods. Such efforts will deepen our understanding of how transient systems evolve into persistent emitters and the role of disc geometry in shaping long‑term X‑ray variability.