2007ApJ...660..301M

We call attention to the fact that one of the brightest red supergiants in the SMC has recently changed its spectral type from K0-1 I (2004 December) to M4 I (2005 December) and back to K0-1 I (2006 September). An archival spectrum from the Very Large Telescope reveals that the star was even cooler (M4.5-M5 I) in 2001 December. By contrast, the star was observed to be an M0 I in both 1978 October and 1979 October. The M4-5 I spectral types is by far the latest type seen for an SMC supergiant, and its temperature in that state places it well beyond the Hayashi limit into a region of the H-R diagram where the star should not be in hydrostatic equilibrium. The star is variable by nearly 2 mag in V, but essentially constant in K. Our modeling of its spectral energy distribution shows that the visual extinction has varied during this time, but that the star has remained essentially constant in bolometric luminosity. We suggest that the star is currently undergoing a period of intense instability, with its effective temperature changing from 4300 to 3300 K on the timescale of months. It has one of the highest 12 µm fluxes of any RSG in the SMC, and we suggest that the variability at V is due primarily to changes in effective temperature, and secondarily to changes in the local extinction due to creation and dissipation of circumstellar dust. We speculate that the star may be nearing the end of its life.