[ Previous | Next | ADS ]

XMM-Newton X-ray spectra of V407 Lup (Nova Lup 2016)

ATel #10756; Jan-Uwe Ness (ESA/ESAC), Sumner Starrfield (Arizona State U.), Chick E. Woodward (U. of Minn.), Paul Kuin (UCL/MSSL), Kim Page, Andy Beardmore, and Julian Osborne (U. Leicester), Gloria Sala (U. Barcelona), Margarita Hernanz (CSIC-IEEC), Marina Orio (U. of Wisconsin and Padova Observatory), Bob Williams (U. Washington)
on 18 Sep 2017; 21:52 UT
Distributed as an Instant Email Notice Novae
Credential Certification: Jan-Uwe Ness (juness@sciops.esa.int)

Subjects: Ultra-Violet, X-ray, Nova

Nova Lup 2016 (V407 Lup) was observed by XMM-Newton from 11 March 2017, 11:45 to 17:08 UT, 168 days after outburst (ATel #9538) with an exposure duration of 23,000 s. The EPIC pn was operated in Timing Mode with Medium filter. The absorbed flux at Earth, as measured with the RGS over the 14-38A range, was (1.2+/-0.5)E-9 erg/cm^2/s. This is a factor of 44 brighter than 172 days later (day 340 after outburst) as reported in ATel #10722.

The RGS spectrum is dominated by a bright Super Soft Source (SSS) continuum with relatively weak absorption lines covering the wavelength range 14-38A. The most obvious features are absorption edges from NVI (18.6A) and OI (22.8A) and absorption lines from NVII 1s-{2p,3p,4p,5p} (lam_0=24.74A, 20.9A, 19.83A, and 19.36A), OVII 1s-2p (lam_0=21.6A), and NVI 1s-{2p,3p} (lam_0=28.78A and 24.9A). These lines are shifted by at most -400 km/s. In addition, the OVII and NVII 1s-2p lines contain a fast component of -3200km/s. A line at 28.5A may either be CVI, shifted by -400km/s, or NVI, shifted by -3200km/s. Further we find interstellar absorption lines of OI 1s-2p (23.5A) and NI 1s-2p (31.3A).
The continuum can be parameterized surprisingly well by a blackbody fit plus an absorption edge at 19A to reproduce the NVI edge (tau=0.94, column density 1.5E17 cm^-2) yielding Teff=6.1E+5K (kT=53eV), NH=3.12E+21cm^-2 with interstellar O and N reduced to 0.6 solar. The normalization corresponds to a radius of 5.2E+4km (assuming spherical symmetry) at 10kpc assumed a distance which is currently unknown. While a bloated white dwarf is a possible interpretation, overestimates of radius is quite common when using blackbody fits.
We also attempted finding physical NLTE atmosphere models but found no good agreement yet. Comparison to wind-type atmosphere models by van Rossum (2012, ApJ 756,43) suggests Teff between 5.5E+5K and 6.0E+5K (kT=47-52eV). The closest approximation with a plane-parallel, static, non-local thermal equilibrium TMAP model yields Teff=9.8E+5K (kT=85eV), NH=1.4E+21cm^-2, and an absorption line velocity of 290 km/s. We notice that this TMAP model also contains few, weak, absorption lines.

Since the blackbody fit currently reproduces the spectrum better than any of the atmosphere models, we also fit a blackbody to the Chandra spectrum taken on day 340 (ATel #10722) and determined Teff=4.9E+5K (kT=43eV), NH=1.6E+21cm^-2, and R=1.1E+4km. The reduction in flux by a factor 44 can thus not be explained by the lower NH and the reduction in temperature alone but requires also a reduction of the surface area. If assumed spherical, the radius would have to be a factor 5 smaller.

The Optical Monitor (OM) took five exposures, one with the visible grism (3000-6000A) and four in Fast Mode. The grism spectrum contains about a dozen emission features above a weak continuum that is about a factor two above the expected Rayleigh-Jeans tail of the SSS spectrum, reaching out to about 4400A. In the wavelength range 3000A-4500A, the spectrum is contaminated by the first order spectrum of another star. While some of the emission features may be attributed to contamination from zero orders of other stars, we identify emission lines from He II (3203A), [Ne III] (3343A,), [Ne III] (3869A and 3968A), H delta (4101A), [O III] (4363A + some H gamma), He II (4542A + some contamination), He II (4686A or [Ne IV] 4721A since the auroral lines are strong), [O III] (5007A+4959A), and [N II] (5755A). We do not see [Ne V] (3426A) and [O I] (5577A).
The OM fast mode light curve shows a slow sinusoidal variation that is consistent with the 3.6 hour modulation in the UVOT light curve (ATel #10632) and time between the dips in the RGS light curve (ATel #10749). However, the 9.4-minute period reported in ATel #10749 is not present in the OM light curve.