Properties of the 24 Day Modulation in GX 13+1 from Near-Infrared and X-ray Observations

A 24 day period for the low-mass X-ray binary GX 13+1 was previously proposed on the basis of 7 years of RXTE ASM observations (Corbet 2003) and it was suggested that this was the orbital period of the system. This would make it the one of the longest known orbital periods for a Galactic low-mass X-ray binary powered by Roche lobe overflow. We present here the results of: (i) K-band photometry obtained with the SMARTS Consortium CTIO 1.3 telescope on 68 nights over a 10 month interval; (ii) Continued monitoring with the RXTE ASM, analyzed using a semi-weighted power spectrum instead of the data filtering technique previously used; and (iii) Swift BAT hard X-ray observations. Modulation near 24 days is seen in both the K-band and additional statistically independent ASM X-ray observations. However, the modulation in the ASM is not strictly periodic. The periodicity is also not detected in the Swift BAT observations, but modulation at the same relative level as seen with the ASM cannot be ruled out. If the 24 day period is the orbital period of system, this implies that the X-ray modulation is caused by structure that is not fixed in location. A possible mechanism for the X-ray modulation is the dipping behavior recently reported from XMM-Newton observations.

💡 Research Summary

The paper revisits the claim that the low‑mass X‑ray binary GX 13+1 exhibits a ∼24‑day periodicity, originally identified by Corbet (2003) from seven years of RXTE All‑Sky Monitor (ASM) data and interpreted as the orbital period. Such a long orbital period would place GX 13+1 among the longest‑period Roche‑lobe‑overflow systems in the Galaxy. To test the robustness of this claim, the authors combine three independent data sets and employ a more rigorous timing analysis. First, they obtained near‑infrared K‑band photometry with the SMARTS 1.3‑m telescope at CTIO on 68 nights spread over a ten‑month interval. The infrared light curve clearly shows a sinusoidal modulation with a period close to 24 days and an amplitude of roughly 0.1 mag, indicating that the system’s optical/infrared emission is also varying on this timescale. Second, they re‑analysed the RXTE ASM light curve using a semi‑weighted power spectrum rather than the data‑filtering technique used in the original study. This method explicitly accounts for the uneven sampling and the individual measurement uncertainties, thereby providing a statistically sound detection of any periodic signal. The semi‑weighted spectrum reproduces a peak near 24 days, confirming the presence of a modulation in the X‑ray band. However, the peak is broader and its phase drifts over the 15‑year ASM baseline, suggesting that the signal is not strictly coherent. Third, they examined hard X‑ray data from the Swift Burst Alert Telescope (BAT) covering 15–50 keV. No statistically significant 24‑day peak is found in the BAT power spectrum, although the sensitivity limits are such that a modulation of the same fractional amplitude as seen in the ASM cannot be ruled out. The authors therefore conclude that the 24‑day signal is real but quasi‑periodic. If the 24‑day period indeed corresponds to the binary orbit, the X‑ray modulation must arise from a structure that is not fixed in the binary frame. They propose that the recently reported dipping behaviour seen in XMM‑Newton observations—temporary obscuration of the X‑ray source by irregular, possibly warped, material in the outer accretion disc—provides a natural explanation. Dipping can produce a modulation that repeats roughly each orbit but with variable depth and timing, matching the observed ASM behaviour. The long orbital period also implies a relatively large donor star, likely a late‑type giant, consistent with the system’s high infrared luminosity and the presence of strong absorption features in its spectrum. In summary, the paper presents (i) independent confirmation of a ∼24‑day modulation in both infrared and soft X‑ray bands, (ii) evidence that the X‑ray modulation is not strictly periodic, (iii) a non‑detection in hard X‑rays that nevertheless does not exclude a comparable signal, and (iv) a plausible physical scenario—variable disc‑related dipping—rather than a static orbital eclipse. The work highlights the importance of multi‑wavelength monitoring and advanced timing techniques for disentangling orbital signatures from transient accretion phenomena in long‑period X‑ray binaries.