Recurrent Nova U Sco Shows Deep Optical Eclipses During Plateau Phase
ATel #2452; Bradley E. Schaefer, Ashley Pagnotta (Louisiana State University), Bill Allen, Tut Campbell , Tom Krajci, Thomas Richards, George Roberts, William Stein, Chris Stockdale (Center for Backyard Astrophysics), Shawn Dvorak, Tomas Gomez, Barbara G. Harris, George Sjoberg, Thiam Guan Tan (AAVSO), Arto Oksanen (Caisey Harlingten Observatory), and Gerald Handler (University of Vienna)
on 20 Feb 2010; 08:44 UT
Credential Certification: Bradley E. Schaefer (schaefer@lsu.edu)
U Sco is a recurrent nova whose long-predicted eruption started on 2010 Jan 28.4385 UT (IAUC # 9111 ), peaked brighter than V=7.8, faded with the fastest rate of all known novae until Feb 9 whereupon the light curve leveled off in a plateau at V=14.0. The optical light curve in 2010 is closely following those of all nine previous eruptions (Schaefer 2010, ApJSupp, in press, arXiv:0912.4426). Flickering in the light curve was first reported on Feb 5.174 (IAUC # 9114 ). U Sco started as a weak relatively hard X-ray source (ATEL #2419), then became a supersoft X-ray source (ATEL #2430) at the same time the plateau started. U Sco produces deep eclipses in the UV (with depth 0.5-0.7 mag and FWHM duration around 5.3 hours) and probably shows broad X-ray eclipses with ~40% amplitude on Feb 12.46, 13.69, and 14.92 (ATEL #2442).
We have measured 5109 magnitudes, from the discovery images up to Feb 18.30. Observations were made with many telescopes in the United States (Florida, Arkansas, and New Mexico), New Zealand, Australia, South Africa, Spain, and Chile. The magnitudes are all from CCD images, either through a V-band filter or without a filter (resulting in effective wavelengths of near 6100A). Each unfiltered data set was normalized to the Johnson V-band by means of simultaneous magnitudes taken with classical V-band photometry. We constructed a smooth light curve template and subtracted this from all magnitudes.
Before Feb 5.174, our folded light curve (see the first figure in the link below) has a RMS scatter of 0.10, which is consistent with the uncertainties in the detrending of the light curve. On Feb 3.86, we see a flicker with duration of ~1 hour and amplitude of at least 0.15 mag. No eclipses are evident.
From Feb 5.174 to Feb 11.0, U Sco had frequent large amplitude flickering (second figure). No flickers were seen after Feb 9.8. The flickers have a typical time scale of 30 minutes and amplitude of 0.4 mag. These flickers have the same amplitude and time scale as in quiescence, but with U Sco being ~60 times brighter than in quiescence, so the energy in the flickers is greatly larger than what is available in the normal disk. The RMS scatter during this time interval is 0.28 mag. No eclipses are visible in the folded light curve.
Between Feb 8.8 and Feb 11.3, U Sco goes from no eclipse to deep eclipse. The folded light curve for Feb 11.0-18.3 (third figure) shows a highly significant eclipse roughly 0.6 mag deep. The light curve shape appears to be constant over this entire interval. The FWHM of the eclipse is 0.18+-0.02 in phase (5.3+-0.6 hours), while the total duration of the eclipse is 0.29+-0.06 in phase (8.6+-1.8 hours). This can be compared to the eclipses in quiescence (see Figures 45-47 in Schaefer 2010), where the V-band depth is 1.3 mag, the FWHM is 0.10 in phase (3.0 hours), and the total duration is 0.18 in phase (5.3 hours). We attribute the sudden onset of eclipses to the recession of the photosphere to a radius smaller than the binary separation (i.e., ~6.4 R_sun), whereas the relative small depth and long FWHM of the eclipse show that the photosphere is somewhat larger than the size of the companion star (~2.3 R_sun). Another implication is that the early flickering arises from the shell, not the accretion disk.
The folded light curve also appears to show a shallow secondary 'eclipse'. While the companion star (a G5 IV star) is normally at V=18.9 alone, perhaps the apparent secondary minimum is caused by the photosphere around the white dwarf covering the brilliantly irradiated companion star. Another intriguing feature is the asymmetry where U Sco is brighter at phase 0.25 by 0.11+-0.04 mag than at phase 0.75. This poorly-understood situation is similar to that of the recurrent novae CI Aql and V394 CrA in quiescence (Figs 44 and 48, Schaefer 2010).
U Sco detrended light curve folded on the orbital period for three time intervals
