A model of an expanding giant that swallowed planets for the eruption of V838 Monocerotis

ABSTRACT In early 2002 V838 Monocerotis had an extraordinary outburst whose nature isstill unclear. The optical light curve showed at least three peaks and imaging revealeda light echo around the object – evidence for a dust shell which was emitted severalthousand years ago and now reflecting light from the eruption. Spectral analysis sug-gests that the object was relatively cold throughout the event, which was characterizedby an expansion to extremely large radii. We show that the three peaks in the lightcurve have a similar shape and thus it seems likely that a certain phenomenon wasthree times repeated. Our suggestion that the outburst was caused by the expansionof a red giant, followed by the successive swallowing of three relatively massive plan-ets in close orbits, supplies a simple explanation to all observed peculiarities of thisintriguing object.Key words:stars: evolution – stars: individuals: V838 Mon – planetary systems 1 INTRODUCTIONThe outbust of V838 Mon was discovered in 2002 January(Brown 2002). The object was about 6 mag brighter than itspresumable quiescent brightness level (Munari et el. 2002a).After a short phase of slow decline it had a second episodeof fast brightening by nearly the same factor. V838 Mongradually faded again in February,but in March a third peakwas observed. Figs. 1 & 2 display the visual light curve of theoutburst compiled from various sources. To our knowledgesuch a light curve has never been observed before.Spectra of V838 Mon taken by several groups (Munariet el. 2002a; Goranskii et el. 2002; Banerjee & Ashok 2002;Wisniewski et al. 2003) suggested a very cool and extendedphotosphere throughout the eruption. V838 Mon becameredder during the outburst and the following decline. Thespectrum near the peak of the outbust was fitted with aK5 giant while several months later an M10 giant can beconcluded (Munari et el. 2002b; Munari & Desidera 2002).Later infrared observations in 2002 October led to the sug-gestion that this star is the first L-supergiant (Evans et al.2003). This evolution in time corresponds to a gradual de-crease in the effective temperature from ∼6000 K to less