Quasi-periodic flares in EXO 2030+375 observed with INTEGRAL

Context: Episodic flaring activity is a common feature of X-ray pulsars in HMXBs. In some Be/X-ray binaries flares were observed in quiescence or prior to outbursts. EXO 2030+375 is a Be/X-ray binary showing “normal” outbursts almost every ~46 days, near periastron passage of the orbital revolution. Some of these outbursts were occasionally monitored with the INTEGRAL observatory. Aims: The INTEGRAL data revealed strong quasi-periodic flaring activity during the rising part of one of the system’s outburst. Such activity has previously been observed in EXO 2030+375 only once, in 1985 with EXOSAT. (Some indications of single flares have also been observed with other satellites.) Methods: We present the analysis of the flaring behavior of the source based on INTEGRAL data and compare it with the flares observed in EXO 2030+375 in 1985. Results: Based on the observational properties of the flares, we argue that the instability at the inner edge of the accretion disk is the most probable cause of the flaring activity.

💡 Research Summary

EXO 2030+375 is a well‑studied Be/X‑ray binary that exhibits regular “normal” outbursts roughly every 46 days, coinciding with periastron passage. While occasional single flares have been reported in the past, a clear quasi‑periodic flaring episode had only been seen once before, with EXOSAT in 1985. In this paper the authors present INTEGRAL observations of the source during the rise of a normal outburst in November–December 2010. Using IBIS/ISGRI (20–300 keV) and JEM‑X (3.5–35 keV) data, they constructed a high‑time‑resolution light curve that revealed at least five distinct flares. The flares recur with a mean period of ≈0.3 days (≈7 h), a value significantly longer than the ≈4 h spacing reported for the 1985 event.

To quantify the variability, the authors calculated the normalized excess variance (NXS) in successive time intervals. NXS is high during the rising part of the outburst, where the flares occur, and drops sharply as the source reaches its peak flux, remaining low throughout the decay. A Lomb‑Scargle periodogram applied to the rising‑phase data shows a clear peak at 0.293 days, confirming the quasi‑periodic nature. Folding the light curve on this period yields an asymmetric flare profile with a steep rise and a slower decay, reminiscent of a “draining reservoir” model.

Spectrally, the combined JEM‑X and ISGRI spectrum of the whole outburst is well described by a cutoff power‑law (photon index Γ≈1.6, folding energy ≈30 keV) with a high absorption column (n_H≈1.1×10^23 cm⁻²). The limited low‑energy coverage prevents a firm conclusion about the unusually large n_H. Because the data quality does not allow spectral fitting of individual flares, the authors investigated how the photon index varies with flux. During the flaring (rising) phase the spectrum hardens noticeably as the flux increases (Γ decreases), whereas during the decay phase Γ remains roughly constant. Linear fits give a slope of (−9.0 ± 3.7)×10⁻³ (cts s⁻¹)⁻¹ for the rise and (2.8 ± 4.2)×10⁻³ (cts s⁻¹)⁻¹ for the decay, indicating a clear difference in spectral behaviour.

The discussion evaluates three possible mechanisms for the flares: (1) magnetospheric instabilities at the inner disk boundary, (2) inhomogeneities in the donor’s stellar wind, and (3) thermonuclear burning on the neutron‑star surface. The authors argue that wind clumping cannot produce the observed quasi‑periodicity because the viscous time of the accretion disk (several days) would smooth out any short‑term mass‑transfer variations. Thermonuclear bursts are unlikely given the high accretion rate surrounding the flares, which would suppress the necessary temperature gradients. The most plausible explanation is an instability at the inner edge of the accretion disk, where the magnetospheric radius r_m lies close to the corotation radius r_c. In this regime, matter can pile up at r_m, increasing the local pressure and moving r_m inward until it crosses r_c, at which point the stored material is rapidly accreted, producing a flare. The cycle then repeats.

Using the standard α‑disk prescription (α≈0.1, H/R≈0.05) and the known spin period (P≈40 s), the viscous time at r_c is τ_c≈1/