2013MNRAS.429.2366S

This paper proposes a simple model for the 19th century eruption of Eta Carinae that consists of two components: (1) a strong wind ({dot}M=0.33   M_☉/yr; v∞ = 200k/s), blowing for 30 yr, followed by (2) a 1050 erg explosion (10 M☉; 750-1000k/s) occurring in 1844. The ensuing collision between the fast ejecta and the dense circumstellar material (CSM) causes an increase in brightness observed at the end of 1844, followed by a sustained high-luminosity phase lasting for 10-15 yr that provides a close match to the observed historical light curve. The emergent luminosity is powered by converting kinetic energy to radiation through CSM interaction, analogous to the process occurring in more luminous Type IIn supernovae, except with ∼ 10 times lower explosion energy and at slower speeds (causing a longer duration and lower emergent luminosity). We demonstrate that such an explosive event not only provides a natural explanation for the light-curve evolution, but also accounts for a number of puzzling attributes of the highly scrutinized Homunculus, including: (1) rough equipartition of total radiated and kinetic energy in the event, (2) the double-shell structure of the Homunculus, with a thin massive outer shell (corresponding to the coasting cold dense shell) and a thicker inner layer (between the cold dense shell and the reverse shock), (3) the apparent single age and Hubble-like flow of the Homunculus resulting from the thin swept-up shell, (4) the complex mottled appearance of the polar lobes in Hubble Space Telescope images, arising naturally from Raleigh-Taylor or Vishniac instabilities at the contact discontinuity of the shock, (5) efficient and rapid dust formation, which has been observed in the post-shock zones of Type IIn supernovae, and (6) the fast (3000-5000k/s) material outside the Homunculus, arising from the acceleration of the forward shock upon exiting the dense CSM. In principle, the bipolar shape could be explained borrowing from earlier studies of interacting winds, except that here the requisite pre-existing `torus' may be provided by periastron collisions occurring around the same time, and the CSM interaction occurs over only 10 yr, producing a thin shell with the resulting structures then frozen-in to a homologously expanding bipolar nebula. This self-consistent picture has a number of implications for other eruptive transients, many of which may also be powered by CSM interaction. A key remaining unknown is the ultimate source of the 1050 erg of energy required in the explosion.