Apparent Glitch in SGR 0418+5729

ATel #2152; Peter M. Woods (Dynetics), Ersin Gogus (Sabanci Univ.) & Chryssa Kouveliotou (NASA/MSFC) report on behalf of a larger collaboration:
on 6 Aug 2009; 21:31 UT
Credential Certification: Peter Woods (Peter.Woods@dynetics.com)

Tweet

Since its discovery on June 5, 2009 (GCN #9499, ATEL #2077), SGR 0418+5729 has been diligently observed with the RXTE PCA for the last 2 months. As noted by Kuiper & Hermsen (ATEL #2151), the pulse timing of this object is unusual for soft gamma repeaters. Specifically, we too find no evidence for measurable spin-down across this ~54 day time span. However, we do find evidence for the effects of a glitch near the onset of burst activity in early June.

Using standard phase-coherent timing techniques, we measured the phase offsets for 30 data segments within this time span by folding the data on a constant frequency. The phase residuals are consistent with a rapid exponential decay versus time - similar to what was observed from 1E 2259+586 directly following the 2002 outburst and glitch (Woods et al. 2004).

We fit the phase residuals to the glitch model defined in equation 5 of Lyne, Pritchard & Smith (1993). Since only the exponential frequency rise was clearly visible in the data and no pre-glitch data exists, we froze all parameters in the glitch model with the exception of the exponential frequency rise and the nominal phase and frequency. The epoch of the putative glitch was frozen at the onset of detected burst activity. Using this model, we obtain a very good fit to the data with a chisquare of 28.2 for 26 degrees of freedom. Within the time span of the RXTE data collection, we detect an exponential frequency rise of 2.5 microsec on a time scale of 2.2(3) days. This corresponds to a fractional frequency change of dNu/Nu = 3.2E-6. Assuming that the resolved glitch component in SGR 0418+5726 constituted a similar fraction of the instantaneous frequency change detected in the 2002 glitch from 1E 2259+586 (~20%), the total glitch amplitude for SGR 0418+5726 is inferred to be dNu/Nu ~ 2E-5. This would be consistent with the largest glitches ever detected from magnetars (Dib, Kaspi & Gavriil 2008) or any pulsar for that matter.

Formally, we find a 3 sigma lower limit on the frequency derivative of -3.8E-15 Hz/s. This would correspond to an upper limit on the dipole magnetic field of 6E+14 G. As pointed out by Kuiper & Hermsen, this is extremely small for a magnetar - particularly a soft gamma repeater. Since glitch recovery is typically followed by an increase in the instantaneous spin-down rate, the magnetic field should only be smaller than this limit. However, we note that glitch recovery (particularly in magnetars) varies from source to source and even within individual sources (Dib et al. 2008). Given the limited number of recorded magnetar glitches and our limited understanding of this recovery in general, we caution that our frequency derivative measurement (limit) and corresponding magnetic field limit may be significantly biased.

We thank Jean Swank, Divya Pereira and the RXTE team for carrying out these ToO observations.