Confirmation of GRS 1747-312 as the active transient in Terzan 6

ATel #4915; A. Bahramian, C. O. Heinke, D. Altamirano, G. R. Sivakoff, C. Markwardt, J. Homan, D. Pooley
on 25 Mar 2013; 19:46 UT
Credential Certification: Craig Heinke (cheinke@virginia.edu)

Tweet

The globular cluster Terzan 6 hosts the known recurrent (~ every 6 months) neutron star transient GRS 1747-312, which has shown regular eclipses that identify its 12.36 hour orbital period (e.g. in't Zand et al. 2003, A&A, 406, 233). Recently transient LMXBs in the globular clusters NGC 6440 and Terzan 5 have been shown to be different objects than the previously known transients in these clusters (ATel #2139, #2974, #4302). Since Terzan 6 has a comparable stellar encounter rate to those clusters (Bahramian et al. 2013, ApJ, 766, 136), it is not certain that new outbursts from Terzan 6 are necessarily from GRS 1747-312.

INTEGRAL/JEM-X detected emission from Terzan 6 on 2013 March 11 (ATel #4883), as also reported by a MAXI transient alert on 2013 March 11 12:45:25 UT. We obtained two pointed Swift/XRT observations, from 2013-03-18 20:59:58 to 21:16:34 and 2013-03-24 02:12:03 to 02:28:39. We timed the second observation to coincide with a predicted eclipse ingress, based on the orbital period ephemeris of in't Zand et al. (2003). We clearly see the predicted eclipse ingress, with the source count-rate falling by a factor of ~60 over ~25 seconds, with a mid-ingress time of 02:20:23 (UTC) compared to the predicted 02:21:25 (UTC). Therefore, we confirm this outburst is due to GRS 1747-312.

We extract spectra from the first Swift/XRT observation of GRS 1747-312, and from the pre-eclipse and eclipse portions of the second observation, using standard procedures, including excising the PSF core due to pileup (except during the eclipse). We find that an absorbed power-law PHABS*PEGPWRLW in XSPEC describes all three datasets reasonably. Since we measure N_H=(1.6+/-0.2)*10^{22} cm^{-2} for the first observation and N_H=(1.8+/-0.3)*10^{22} cm^{-2} for the pre-eclipse portion of the second observation, we fix N_H to the mean value of 1.66*10^{22} cm^{-2} for the eclipse interval, which only has 46 counts. We measure power-law photon indices of 1.5+/-0.2,1.8+/-0.3 & 1.5+/-0.8 , and (unabsorbed) 0.5-10 keV fluxes of 6.0*10^{-10}, 4.9*10^{-10}, and 1.0*10^{-11} erg cm^{-2} s^{-1} for the first, second pre-eclipse, and second eclipse portions, respectively. The non-eclipse fluxes are consistent with the typical post-peak fluxes of GRS 1747-312 during outburst (in't Zand et al. 2003), and translate (for a 6.8 kpc distance, Harris, W.E. 1996, AJ, 112, 1487) to luminosities of 3.3*10^{36}, 2.7*10^{36}, 5*10^{34} erg s^{-1}.

Although the Swift PSF encompasses the cluster core, the eclipse flux and spectrum are probably too high to be due to emission from a menagerie of quiescent LMXBs and cataclysmic variables in this globular cluster. This indicates that the observed photons in eclipse are probably scattered X-rays from an accretion disk corona. The position of the emission during eclipse as derived with xrtcentroid, J2000=17:50:46.37, -31:16:29.6 (4.7"), is consistent with the position of GRS 1747-312 when bright during the same observation, J2000=17:50:46.25, -31:16:29.4 (3.6"), and also consistent with the cluster center from in't Zand et al. (2003), so does not distinguish between the possibilities.