URGENT: THE IMPENDING EVENT IN ETA CARINAE

K. Davidson$^1$, M. Corcoran$^2$, and A. Damineli$^{3,4}$


$^1$ Astronomy Dept., University of Minnesota
$^2$ Laboratory of High Energy Astrophysics, Goddard Space Flight Center
$^3$ Inst.\ Astr.\ e Geof., Universidade de S\~{a}o Paulo
$^4$ JILA, University of Colorado

Eta Carinae is expected to undergo an unusual spectroscopic event at the end of this year and the beginning of 1998. Although the precise nature of the event is not yet understood, it is likely to be an unprecedented opportunity to settle certain long-standing major problems regarding this enigmatic object. Unfortunately, there is serious cause for concern that observational coverage of this event, ground-based but especially with HST, will not be adequate. {\it The situation is so extraordinary that additional observations must be encouraged even though time is now growing short.\/} In the following notes we provide some background information and solicit help. (A review containing more general references about ${\eta}$ Car will appear soon: \ Davidson \& Humphreys 1997, {\it Ann.Revs.Astr.Ap.,\/} in press).

The historical development leading up to the predicted event has been as follows. On several occasions in recent decades a particular type of spectroscopic event in ${\eta}$ Car has been observed: see, e.g., Zanella et al.\ 1984, {\it A{\&}A\/} 137, 79; Ruiz et al.\ 1984, {\it ApJ\/} 285, L19; Bidelman et al.\ 1993, {\it PASP\/} 105, 785; Damineli et al.\ 1993, {\it Space Sci.\ Revs.\/} 66, 211. During each event some bright narrow emission lines temporarily change dramatically, especially those of higher excitation. [Ne~III], He~I ${\lambda}10830$, and the narrow component of H$\alpha$, for example, can disappear entirely for several weeks. Most of the narrow-line emission is formed in strange slow-moving blobs of ejecta several hundred a.u.\ from the star, probably near the equatorial plane (Davidson et al.\ 1995 and 1997, {\it AJ\/} 109, 1784, and 113, 335). Zanella et al.\ conjectured that the observed type of event signals a temporary decrease in the supply of hard UV photons. Whether this is due to a shell ejection, or geometrical phenomena in a binary system, or other causes (see below), it seems a very promising clue to the nature of Eta and in particular to the basic instability. But the observations were neither intensive nor extensive and the reports tended to be anecdotal, because the events had occurred unexpectedly and were brief in duration.

Last year Damineli (1996, {\it ApJ\/} 460, L49) found good evidence that the spectroscopic events recur with a 5.5-year period. All the known examples were consistent with this periodicity, as were some prominent near-infrared photometric fluctuations noted by Whitelock et al.\ (1994, {\it MNRAS\/} 270, 364). If this periodicity is confirmed (and we now expect that it will be), then {\it it greatly enhances the importance of the spectroscopic events.\/} They had already seemed useful as noted above; but in addition the probable recurrence period is likely to be of great theoretical significance, while it also enables us to prepare for suitable observations of the next event! {\it This is predicted to occur near 1998 Jan 1,\/} definitely within a few months of that date and probably within a few weeks of it. So far, according to Damineli's continuing observations, the He~I $\lambda$10830 emission line has been fading on schedule during 1997, behaving as it did before the 1992 event. The X-ray behavior has also been peculiar lately, as noted below.

A period of 5.5 years in $\eta$ Car is a quantitative surprise. If this represents an orbital period in a binary, then the separation between components is far too large for significant interactions except perhaps very near periastron in a high-eccentricity model. In a single-star scenario, 5.5 years is far longer than any obvious dynamical timescale and probably must represent a thermal timescale for some particular fraction of the star. (The mere fact of a well-defined period does not rule out a single-star model; Eta has often been likened to a geyser, a simile consistent with the idea of a thermal timescale, and of course some geysers are fairly periodic.)

Damineli, Conti, and Lopes (1997, {\it New Astr.\/} 2, 107) have proposed a specific 5.5-yr binary model based on some apparent changes in radial velocity. Davidson (1997, {\it New Astr.\/} to appear soon) suggests a more eccentric orbit but also notes that the binary models have some awkward features, while a single-star shell ejection event may produce a similar ``velocity curve''. The important points are that the binary--vs.--single-star question is still quite unresolved, and that the basic parameters of either type of model are also unsettled. There is good reason to hope that spectroscopy during and after the predicted event can answer the first question and help a great deal with the second.

Meanwhile, a group led by Corcoran has found remarkable results while monitoring the hard X-rays of $\eta$ Car. The origin of these X-rays is unknown, especially since they indicate high temperatures requiring shock speeds of 2000 km/s or more, much faster than Eta's wind speed observed at visual wavelengths (see Corcoran et al.\ 1997, submitted to {\it ApJ\/}). Colliding winds in a binary system and fast localized wind streams from a single star are among the possible explanations. Since mid-1996 the hard X-ray flux has been increasing, but the most interesting development is a series of progressively higher peaks (``flares'') that recur with a period of 85 days, reported in IAU Circulars 6668 and 6701. Such a period is speculatively fascinating against the background of the longer 5.5-year period, but the main point here is that the rising, pulsating X-ray flux curve informally gives the impression of an impending crisis.

The expected event may therefore give us the best observational opportunity we've ever had to attack several major problems alluded to above: the binary--vs.--single-star question for $\eta$ Car, the mystery of the star's basic instability, and the origin of two different observed periodicities. But those are not all; other unusual problems may also be included in the list, especially concerning peculiar excitation mechanisms in the narrow-line-gas near Eta. This gas is the most intense known site for certain fluorescent processes reviewed by Johansson \& Hamann (1993, {\it Physica Scripta\/} T47, 157). Especially noteworthy are the amazingly intense Fe~II features near 2507 {\AA}, which may even represent the only known natural UV laser (Johansson et al.\ 1996, in {\it Science with the HST, STScI/ST-ECF Workshop,\/} p. 361). These are expected to be very sensitive to the UV illumination and observations of their behavior in a spectroscopic event are extremely desirable.

At this time only a limited set of observations are scheduled, not nearly as extensive as the event deserves. A group including Damineli will monitor $\eta$ Car with ground-based spectroscopy at some wavelengths, and Corcoran's group plan to continue their intensive X-ray monitoring; contact us for details. Below we list a few types of observations that may prove valuable or even essential; interested observers may devise alternative approaches. Needed are: (1) Coverage at other wavelength ranges, especially radio and mm-wave, during the event and in the following months. (2) Additional ground-based spectroscopy. There are two obvious aspects to watch for, the intensity behavior of the narrow emission lines with various excitations and also the velocity behavior of the broad lines that come from the star itself. (3) Precisely repetitive large-telescope long-slit spectroscopy of the homunculus nebula during the weeks and months following the event. Since the homunculus is essentially a scattering nebula about half a lightyear across, reflected spectrum changes may be detectable after suitable delay times. If detectable, these would give information about the appearance of the event in various directions. We must acknowledge that such observations are expected to be difficult. (4) Repeated large-telescope imaging for photometric purposes. Photometric changes are not expected to be dramatic, but it would be prudent to have some careful data ``just in case''. Emission-line photometry of the hydrogen lines may prove useful. (5) HST/STIS observations with very high spatial resolution. These are needed for at least two reasons. First, the spectrum of the star itself must be spatially resolved in order to reliably test the predicted orbital velocity changes for binary models. Second, in some models the changes in the narrow emission lines should move across a region about 0.3 arcsec across, for example due to shadowing in a binary scenario. (This region is described in Davidson et al.\ 1997, {\it AJ\/} 113, 334, and refs.\ therein.) Such observations would of course represent a {\it tour de force\/} of HST spectroscopy, but they seem attainable. (6) HST/STIS observations in the UV, again with at least two different specific motivations. First, the spectral changes in the star itself are expected to be far more noticeable in the UV; in a binary model that is the only likely wavelength region for detecting the partial eclipse of the companion star as it moves behind the primary's dense wind. Second, it would be extremely disappointing to allow the bizarre Fe~II $\lambda$2507 emission (mentioned above) to go unobserved during such an event.

It is frustrating to report that {\it no\/} HST observations of the impending event have been approved at this time. More than once we have requested emergency target-of-opportunity STIS observations of the event, emphasizing the extraordinariness of the situation and suggesting that such data should become public as soon as they are obtained. Surprisingly, STScI has denied all these requests; but we hope they may yet be persuaded to reconsider. In our opinion, considering the unique nature of the event and the unique nature of $\eta$ Car, a failure to obtain UV spectroscopy and high-spatial-resolution data would constitute a significant and unnecessary Missed Opportunity for stellar and nebular astrophysics.

In summary, we welcome and solicit community interest and participation in the expected event, and we hope to organize expressions of support for adequate observations. Time is short; please mention the situation to your colleagues, and for further information contact us.

kd@ea.spa.umn.edu,
corcoran@barnegat.gsfc.nasa.gov,
damineli@casa.colorado.edu

August 1997