[ Previous | Next | ADS ]

Large amplitude variability and detection of a 35 sec modulation of the soft X-ray flux of nova KT Eri by Swift

ATel #2423; A. P. Beardmore (U. Leicester), S. Balman (METU), J. P. Osborne, K. L. Page (U. Leicester), M. F. Bode (Liverpool JMU), J. J. Drake (CfA), J.-U. Ness (ESA/ESAC), M. Orio (INAF-Padova and U. Wisconsin) and S. Starrfield (ASU)
on 5 Feb 2010; 16:43 UT
Credential Certification: Andy Beardmore (apb@star.le.ac.uk)

Subjects: Binary, Cataclysmic Variable, Nova

Referred to by ATel #: 2434, 2558, 5573, 5897

Swift X-ray Telescope (XRT) observations of Nova Eri 2009 (KT Eri) show that the supersoft source (SSS) reported by Bode et al (ATel #2392) first increased, and then decreased in variability. The 0.3-1 keV count rate showed >500x variations on a timescale of ~1.5 days, from <0.1 c/s to ~70 c/s between days 66 to 68.5 from the time of peak optical brightness (2009 Nov 14.632). After day 69.4 the variation became less dramatic, with the count rate varying between about 5 and 150 c/s, until after day 77.2 where the source has remained almost consistently above ~100 c/s. Similar high X-ray variability phases were seen in the novae RS Oph, V458 Vul and LMC 2009 (e.g. Osborne et al., ATel #770; Drake et al., ATel #1721; Bode et al. ATel #2025).

We have searched the XRT WT mode light curves for variability on shorter timescales. A Fourier analysis of the data from day 66.60 to 79.25, whereby power spectra of 102 intervals of 1024 s segments of data at 1 s binning were averaged, has revealed the presence of a quasi-periodic oscillation (QPO) and low frequency noise (LFN). Fitting the averaged power spectrum with a power-law for the LFN, a Lorentzian for the QPO and a constant for the statistical noise, gives a best fit QPO period of 35.09 +/- 0.25 s, a QPO FWHM of (1.7 +/- 0.5)e-3 Hz and a LFN power-law index of -2.03 +/- 0.06. The integrated power in the Lorentzian corresponds to a fractional r.m.s. of 1.7%. The coherence of the QPO as measured from the averaged power-spectrum is ~16 cycles. Continuous 512 s sections of data, in which the periodicity is occasionally detected strongly, show fractional amplitudes up to 7%. A similar analysis of the equivalently bright AGN Mkn 421 (also XRT WT mode) did not reveal any such period to very low modulation limits, which gives us confidence in this result.

Chandra observed KT Eri with the LETG (ATel #2418). A time series analysis of this dataset, spanning day 70.283 to 70.457, revealed a marginal (<90% confidence) detection of the 35s periodicity, with a 90% fractional amplitude upper limit of 0.9%. The XRT did not detect the periodicity significantly in its nearest snapshots at day 70.228 and 70.564, with fractional amplitude limits of 2.3% and 4.5%, respectively.

Most surprisingly, KT Eri is the second nova, after the recurrent nova RS Oph (Osborne et al Atel #770), to exhibit a ~35 s oscillation in its SSS X-ray emission. In RS Oph the modulation was occasionally seen until the SSS turned off (Beardmore et al., 2008 ASPC 401 296). The 35 s oscillation in RS Oph was confirmed independently with XMM (Ness et al., 2007 ApJ 665 1334; Nelson et al. 2008 ApJ 673 1067).

The presence of apparently the same QPO in KT Eri as in RS Oph would appear to argue against a rotation-based origin for this modulation, while a possible residual nuclear-burning white dwarf pulsation might be more constrained in frequency.

We are grateful to the Swift PI, Neil Gehrels and the Swift Mission Operations team for their support of these observations.