Erratum: ``A Potential Supernova Remnant-X-Ray Binary Association in M31'' (ApJ, 634, 365 [2005])

Abstract

We have discovered a flaw in the light curves we previously generated using the XMM-Newton Science Analysis Software (SAS), which can produce spurious Type A power density spectra (R. Barnard et al. 2007, preprints astro-ph/0610035, astro-ph/0703120). This is the type of variability associated with disk-accreting binary systems (R. Barnard et al. ApJ, 634, 365 [2005]) that was reported to be of high-significance in the supernova remnant (SNR) r3-63 in M31. We interpreted this variability as strong evidence that the SNR contained an associated X-ray binary (XRB).

Using a work-around, we have reanalyzed the XMM-Newton data of r3-63 and find no evidence for variability. The power density spectrum of the revised light curve is shown in our revised Figure 3. It is flat at a value of 2, the value expected for Poisson noise. Therefore, our detection of variability in this source, which led to the suggestion that it may harbor an XRB, was false.

Supporting evidence for a possible XRB association (possible hard spectral component and variability in the Chandra data) are by themselves not significant enough to claim an XRB association. Because the only >3 σ detection of variability was that from the XMM-Newton data and because that detection has been shown to be an artifact of the light-curve extraction, there is no significant evidence that this SNR harbors an XRB. This conclusion is also consistent with the recent findings of S. Trudolyubov et al. (2006, preprint astro-ph/0610809) who, with the addition of more XMM-Newton data, found no evidence for a hard component to the spectrum of r3-63.