*3.3. Eta Carinae*

The classic example of a supernova impostor, of course, is *η* Carinae. It merits a separate subsection here, because it has been observed in far more detail than any other relevant object. Following the

tradition of classic examples in astronomy, it has abnormal properties. Its event seen in 1830–1860 persisted much longer than other known giant eruptions, and it has a companion star that approaches rather closely at periastron. Many authors reviewed *η* Car in ref. [57], and later developments have not altered the main facts.

The star's luminosity is roughly 106.6 *L*. Before 1830 its mass was probably in the range 140–200 *M*, with an apparent temperature close to 20,000–25,000 K [58]. Then its 30-year Great Eruption ejected 10-40 *M* at speeds averaging 500 km s<sup>−</sup>1. It converted more than 10<sup>50</sup> ergs of energy to roughly equal portions of radiation, kinetic energy of ejecta, and potential energy of escape. The resulting "Homunculus" ejecta-nebula [59] is famously bipolar, indicating a complex role for angular momentum [58]. The ejected material is clearly CNO-processed, with helium mass fraction *Y* ∼ 0.4–0.6 [60,61]. The star's subsequent recovery has been unsteady, including a smaller eruption around 1890 and two later disturbances [58,62,63]. Until recently its wind was opaque in the continuum with *M* ˙ ∼ 10−<sup>3</sup> *M* y<sup>−</sup>1, and most likely above 10−<sup>2</sup> *M* y<sup>−</sup><sup>1</sup> a century ago [62]. That rate was too large for a line-driven stellar wind.

The companion object is most likely an O4-type star with a highly eccentric 5.54-year orbit, see many articles in [57]. The primary star is indeed the eruption survivor, since it has an extremely abnormal wind that has been diminishing since the event. But the hot secondary star alters the situation in several ways.


For historical reasons [5], *η* Car is often called an LBV despite its location in the HR Diagram (Figure 1). The high-luminosity end of the LBV strip in Figure 1 may be misleading, and in principle every star with *L* > 106.3 *L* and *T*eff < 25,000 K might be an LBV; there are not enough examples to know. But that is only a possibility, and we have no definite reason to classify *η* Car as an LBV. An LBV-like eruption would be complicated for this object, because the radius of a 10,000 K photosphere would exceed the companion star's periastron distance.

### **4. The Spectrum of an Opaque Outflow**

This section has four main points: (1) Giant stellar eruptions usually have similar colors and spectra even if they're caused by different processes. (2) Certain emission line profiles indicate an opaque outflow. (3) Stellar spectral types are not reliable indicators for outflow temperatures. (4) LBV outflows are not homologous with giant eruptions.
