Fading and probable quasi-periodic oscillation in the Swift X-ray observations of symbiotic recurrent nova V745 Sco

ATel #5897; A. P. Beardmore, J. P. Osborne, K. L. Page (University of Leicester)
on 18 Feb 2014; 10:51 UT
Credential Certification: Julian P Osborne (julo@star.le.ac.uk)

Tweet

The soft (0.3-2 keV) Swift XRT count rate of nova V745 Sco peaked at day 5.5 after discovery (CBET #3803), and has been declining continuously since then. At day 7 it was 40 c/s, at day 10 it was 6 c/s. The hard X-ray count rate (2-10 keV) has been declining since day 4, at day 10 it was 0.4 c/s.

While the nova was a bright super-soft X-ray source (ATEL #5877) it was feasible to search for short period X-ray count rate oscillations, such as those seen at 35 sec in the related object RS Oph (Osborne et al. 2011, ApJ 727 124), at 35 sec in nova KT Eri (ATEL #2423) and at 54 sec in Nova Del 2013 (ATEL #5573).

We report the probable discovery of a transient quasi-periodic oscillation with a period of ~163 sec in the 0.3-1.5 keV count rate. In the 5 WT mode snapshots taken between day 5.2 and 5.6 (when the count rate was above 50 c/s) Fourier analysis of continuous time series of 750-1000 sec duration with 0.018 sec binning reveals a broad peak which has a significance of 5 sigma above a power law red noise model. Fitted with a Lorentzian, the peak has a central frequency of 0.00613 Hz (P = 163 sec) and a FWHM of 0.0036 Hz. We consider this to be evidence of a low-coherence QPO. Some time series show distinct pulses throughout, and a higher coherence power spectrum (e.g. a modulation at 178 +/-4 sec with an amplitude of 5.4+/-0.7% in data taken at a mid-time of 2014-02-12T00:13 = day 5.4). Modulation is not present after day 5.6 even though the source remains at roughly the same brightness. There is no modulation with an amplitude above 2% in the period range 1 to 50 sec in the day 5.2-5.6 interval. The observed period length is such that only 5-6 cycles are present in any time sequence, and we note that confidence in the reality of this QPO depends upon the applicability of the red noise model.

Swift UVOT event mode data taken in the uvm2 filter (central wavelength = 2246 A) at the time of the 178 sec X-ray detection does not show a modulation at this period, with a 90% upper amplitude limit of 4.0%. No similar period is detected in any other UVOT data when it is seen in the XRT data.

The period seen in V745 Sco is significantly longer than the QPO seen in the novae listed above, and is only seen in a small portion of the super-soft phase. Still longer X-ray periods have been reported using other facilities, e.g. 22 min in V4743 Sgr (Ness et al. 2003, ApJ 594, L127), 42 min amongst others in V1494 Aql (Drake et al. 2003, ApJ 584, 448); the low observing duty cycle of Swift (Porb = 96 min) makes efficient detection of such longer periods difficult. The origin of all these modulations is unclear. If the very fast super-soft source turn-on time and 2-magnitude decline time indicate a higher mass white dwarf in V745 Sco, then the period reported here, longer than that of RS Oph etc, may argue against an atmospheric oscillation as its origin, as the characteristic timescales of higher mass and density white dwarfs are shorter. The low coherence of the modulation is clearly inconsistent with material tightly bound to a rapidly spinning white dwarf.