Program: GS-2005A-C-9

During 2003, $\eta$ Carinae experienced a dramatic ``spectroscopic event'', when high-excitation lines in its UV, optical, and IR spectrum disappeared, and its hard X-ray and radio continuum flux crashed. This behavior is part of a 5.5 year cycle, the periodicity has been attributed to a binary system, but the changes also resemble a shell ejection. UV spectra obtained during the event with HST suggest that {\it a shell ejection did indeed occur}, so we expect dust to form, and mid-IR images and spectra with T-ReCS are needed to measure the changing mass of dust and the current bolometric luminosity. Now is a critical time to obtain unique IR measurements of the ejecta as $\eta$ Carinae recovers from this event and the shell expands, cools, and forms dust grains. This will provide a direct estimate of the mass ejected. Near-IR emission lines trace related changes in the post-event wind, and can test specific models for the cause of $\eta$ Car's variability as it recovers from its recent ``event''. High resolution near-IR spectra with GNIRS will continue the important work of HST/STIS, investigating changes in the direct and reflected spectrum of the stellar wind, and ionization and excitation changes in the nebula. The complex kinematic structure of $\eta$ Car's ejecta also holds important clues to its mass ejection history, and is essential for interpreting other data. Phoenix can provide a unique kinematic map of the complex density and ionization structure of $\eta$ Car's nebula, which is our best example of the pre-explosion environment of very massive stars.