Support ATel At Patreon

[ Previous | Next | ADS ]

Chandra Localization and Detection of a Burst and Pulsations from GRO J1744-28

ATel #5895; Deepto Chakrabarty (MIT), Peter G. Jonker (SRON), Craig B. Markwardt (NASA/GSFC)
on 17 Feb 2014; 23:56 UT
Credential Certification: Deepto Chakrabarty (deepto@space.mit.edu)

Subjects: X-ray, Binary, Neutron Star, Transient, Pulsar

Referred to by ATel #: 5896, 5901, 5904, 5963, 10079

As part of an ongoing Chandra program for precise localization of X-ray transients in low-mass X-ray binaries, we obtained a short Chandra/HRC-I observation of the bursting accretion-powered pulsar GRO J1744-28 during its recent outburst (ATel #5790, #5810, #5845, #5858, #5883). Our observation was made on 2014 February 16 06:55 TT with an exposure time of 1139 s.

A single source was detected in the field of view, with a total of 9581(98) counts measured. The best-fit X-ray source position is:

RA(J2000) = 17h 44m 33.06s
Dec(J2000) = -28d 44' 27.1"

with an uncertainty radius of 0.6 arcsec (90% confidence). This is consistent with the 2.2 arcsec (90% confidence) Swift/XRT error circle measured on February 2 (ATel #5845), lying 0.8 arcsec from its center. Our new position is also consistent with the quiescent X-ray counterpart detected in 2001 by both Chandra (Wijnands and Wang 2002, ApJ, 568, L93) and XMM-Newton (Daigne et al. 2002, A&A, 386, 531), lying 0.3 arcsec from the Chandra position (0.7 arcsec uncertainty) and 2.7 arcsec from the XMM-Newton position (4 arcsec uncertainty).

The mean persistent count rate from the source was 8.4(1) count/s. A single burst was detected about 830 s into our observation, with a peak intensity of 178(13) count/s and a duration of 17 s. A total of 1383(37) counts were detected during the burst. The burst had a double-peaked time profile, with the secondary peak reaching an intensity of 124(11) count/s. The quoted count rates were computed after applying standard HRC dead-time corrections to the data.

Our HRC-I observation does not provide useful spectral information. However, assuming the absorbed power-law X-ray spectrum measured by Swift/XRT on February 5 (ATel #5858; Gamma=1.0, N_H=1.05e23 cm^(-2)), our measured persistent count rate corresponds to a persistent absorbed flux of 2.7e-9 erg/cm^2/s (0.2-10 keV). For our observed burst, we assume the blackbody spectrum of the burst measured by Swift/BAT on February 11 (ATel #5883; kT=5.4 keV) and again adopt an absorption column N_H=1.05e23 cm^(-2). Our observation then implies an absorbed burst fluence of 6.4e-7 erg/cm^2 (0.2-10 keV) and an absorbed mean burst flux of 3.8e-8 erg/cm^2/s (0.2-10 keV).

Our HRC-I timing data are subject to a known problem wherein the time of an event is recorded with the following event, which may or not be telemetered (see Section 7.11 of the Chandra Proposer's Observatory Guide, v16.0, 2013). For the present count rate, this results in the timing accuracy being degraded from 16 microseconds to about 120 milliseconds, a significant fraction of the pulsar spin period. Despite this limitation, we detected coherent pulsations in the 830 s of pre-burst persistent emission with a barycentric pulse frequency of 2.1410(1) Hz (corresponding to a period of 467.07(3) ms), uncorrected for the binary orbit. This is consistent with the pulse frequencies measured by Fermi/GBM on January 19-23 (ATel #5810). Accounting for the number of frequency bins searched, the false alarm probability for our detection was 9e-5, equivalent to 3.9 sigma Gaussian significance. We made no attempt to correct for the effects of the HRC wiring error. We did not detect pulsations during the 17 s burst or during the 290 s post-burst interval, but these non-detections are not very constraining.