Crisis at Eta Carinae?
A scientific detective story is unfolding now in the southern sky. Something unusual is expected to happen to the star Eta Carinae in the next few months, but we don't yet know exactly what or why. If the predicted event does occur, it may give astronomers their best chance ever to settle some of the mysteries that surround this extraordinary object.
No other star has a list of credentials like Eta Carinae, which lies about 7,500 light-years away in the far-southern sky. It is 4 million times brighter than the Sun, making it the most luminous star in our region of the galaxy. With roughly 100 times the Sun's mass, it is among the most massive stars believed to exist. About 150 years ago it underwent a giant, 20-year outburst--for a few months in 1843 it outshone Canopus (magnitude -0.7) even though it's 30 times more distant. It was the biggest explosion that any star is known to have survived. Several solar masses of matter were ejected at almost 1,000 kilometers per second, but the central star emerged apparently unscathed. That eruption produced the bipolar nebula famous from Hubble Space Telescope images (opposite).
Meanwhile Eta Carinae and its nebula have become the brightest infrared object in the sky beyond the solar system. It is also a unique stellar X-ray source, and at radio wavelengths the star produces the brightest known stellar wind.
Eta Carinae also has a unique place in the theory of stellar structure, evolution, and instability. It may be the most extreme example of a luminous blue variable--a critical stage in the evolution of the most massive stars. It is also the site of several classic puzzles in stellar astrophysics, the most obvious one being that we don't know what caused the giant eruption of the 1840s. Some theorists think internal radiation pressure somehow triggered an instability, while others suspect that the system may be a close interacting binary. Recent discoveries have tended to deepen the enigma rather than solve it.
A Spectral Discovery
Several times in the past 50 years astronomers have reported cryptic, temporary changes in Eta Carinae's spectrum. The spectrum is a complex forest of hundreds of emission lines, representing hot hydrogen, singly ionized iron and nickel, and other elements. Close examination shows at least two significantly different classes of spectral lines. Broad emission lines are formed in a wind gushing out from the star at hundreds of km per second. Superimposed on them are narrower lines formed in gas moving less than 50 km per second, oddly slow in a system with so much turmoil. Using HST in recent years, we've found that the narrow lines come mainly from a few gas blobs several hundred astronomical units (roughly 200 billion km) from the star, moving outward more or less in its equatorial plane. But we don't understand why those blobs exist.
During the brief changes in Eta Carinae's spectrum, the narrow emission lines fade and some of them disappear for a few weeks. It looks as though the heating of the gas blobs, which prompts the emission, is somehow turned off for a while. In any case, some astronomers feel that these spectroscopic "events" ought to be clues to the star's instability. But nobody has been able to act on this hunch because each event has happened without warning and was over before it could be adequately observed.
Last year, though, Augusto Damineli (University of Sao Paulo, Brazil) made a surprising discovery: the star's spectroscopic changes may be predictable after all. In fact they seem to recur at intervals of 5 1/2 years (S&T: August 1996, page 13). This wasn't noticed earlier because of gaps in the observational record. Damineli's finding makes the events even more important, for several reasons.
First, the 5 1/2-year period must be a strong hint about what is happening, if only we can interpret it. But from a theoretical point of view it seems a strangely long time. Perhaps Eta Carinae is a binary-star system with a 5 1/2-year period. Astronomers have often speculated that it might be a close interacting binary, but the hypothetical orbits they envisioned have always had periods of less than a year. With a 5 1/2-year orbit the two stars would be too far apart to interact strongly.
Or maybe Eta Carinae is a single star with recurrent upheavals. In that case, 5 1/2 years is far longer than the natural pulsation period a star like this would have (months rather than years). So it must represent some novel type of buildup-and-release cycle of heating, like a geyser.
Damineli's discovery is important for another reason --it allows us to predict the next event. Since the last one was seen in mid-1992, a new spectroscopic event is expected at the end of 1997 or the beginning of 1998. Recent indications look good. Damineli reports that as of last September, a key helium emission line was fading on schedule.
Furthermore, a research group at NASA/Goddard Space Flight Center led by Michael F. Corcoran (Universities Space Research Association/NASA-Goddard Space Flight Center) has been monitoring the X-ray flux from Eta Carinae since last year. Spotty data from previous years led the group to expect a gradual decrease in X-rays as Damineli's predicted event approaches, but instead they've found something more dramatic. The hot X-rays from the star's wind have been acting like the stock market--a general increase with a series of progressively bigger "flare" events. It's hard to resist the feeling that this behavior will culminate toward the end of 1997, maybe in a final X-ray crescendo followed by a steep plunge. Kazunori Ishibashi (University of Minnesota), working with Corcoran's group, has further thickened the plot by discovering an 85-day periodicity in the X-ray flares (S&T: November 1997, page 21). This might be Eta Carinae's rotation period, a close binary orbital period, or a pulsation period. We don't know how it might be related to the 5 1/2-year cycle.
Binary or Single Star?
Damineli, Peter Conti (University of Colorado), and Dalton Lopes (Brazilian National Observatory) have proposed a detailed binary-star model for Eta Carinae. Writing in the electronic journal New Astronomy for May 22, 1997, they report wavelength shifts in some spectral lines and interpret these as Doppler shifts due to orbital motion. In their scenario two stars, each weighing about 70 times as much as the Sun, go around each other at an average separation of roughly 15 a.u. (somewhat less than the distance from the Sun to the planet Uranus). Only one of the stars is seen; maybe it is overluminous because it began life with a much larger mass. The astronomers assume that the hot X-rays are formed in shock fronts where the winds from the two stars collide. The proposed orbit is highly eccentric, and a spectroscopic event such as the one expected in January may occur when the two stars pass closest to each other. In a binary model like this, one star can temporarily block the flow of ultraviolet radiation to the gas blobs mentioned earlier (an intrasystem eclipse). This might explain the observed temporary fading of the narrow emission lines during an event.
As reported in New Astronomy for September 14, 1997, I have reexamined the Doppler-shift velocities and find an even more eccentric orbit for the proposed binary. In this version the closest-approach separation between the two stars is reduced to about three a.u. But this still seems too far for them to affect each other much, leaving the connection between the orbit and Eta Carinae's basic instability unclear.
The binary theory is awkward in some respects. The combination of two very massive stars seems a little extreme, and the velocity data require the long axis of the highly eccentric orbit to be pointed suspiciously close to our line of sight. There are other problems as well.
Thus a single-star theory has some appeal. In this scenario Eta Carinae is just one massive star that ejects an unusually large amount of gas every 5 1/2 years. Its ultraviolet light is obstructed at such a time, allowing the narrow-line gas blobs to cool off temporarily. The 5 1/2-year period may represent the time required for a certain portion of the star to accumulate enough heat to trigger an instability. The geyser analogy mentioned earlier may be very apt. The binary- versus single-star question is currently the most pressing problem for Eta Carinae, because the nature of the instability depends on it.
The impending "event" (assuming it does occur) may be of great interest to observers in the Southern Hemisphere. Unfortunately the star's general appearance isn't expected to change much. Its brightness may fluctuate a bit, but in small telescopes this change will be swamped by the nearly constant brightness of the surrounding nebula. We're pinning our hopes on spectroscopy with large instruments. Long-standing questions (and new ones, too) may be answered by specific observations.
First, Doppler-shift velocity data are crucial for determining whether the star is binary. Observations weren't detailed enough near the time of the 1992 spectroscopic event. According to the binary models, wavelengths in the stellar spectrum should change in a particular way during the few months prior to the event. Ground-based telescopes are valuable for testing the predictions, but they don't have enough spatial resolution to separate the star's own spectrum from the narrow-line gas blobs, and they can't see the critical ultraviolet spectrum. Spot checks need to be made with HST, which has adequate spatial resolution and ultraviolet capability. (Even HST cannot resolve the two stars predicted by a binary model; we have to rely on Doppler-shift spectroscopy to detect and measure possible orbital motion.)
Next, the X-ray spectrum can reveal the amount of gas near the star that the X-rays have passed through. How this spectrum changes can also distinguish between a single and a binary star.
Some narrow ultraviolet emission lines represent extremely unusual fluorescent processes near Eta Carinae --possibly the only known natural ultraviolet laser. The lines are especially sensitive to the changing ultraviolet light during an event. We need to find out whether this emission disappears, and if so, when and for how long. This, too, will require HST observations.
Some effects may tell us how the event appears from different viewing directions. In a binary model with internal eclipses, HST should be able to see a shadow region moving across the slow gas blobs only a fraction of an arcsecond from the star. Farther out in the nebula, a few arcseconds from the star and potentially observable in large ground-based telescopes, the reflected spectrum may show delayed changes a few weeks or months after we see the event happen to the star. These delays represent the light's extra travel time to us along indirect, reflected paths.
It is too early to predict exactly what we'll learn from the impending spectroscopic changes in Eta Carinae. The binary theory may be confirmed or disproved, light may be shed on the star's fundamental structure and instability, and other things may happen that we haven't anticipated. Unfortunately there is considerable doubt whether any of the crucial HST observations will be made. If HST doesn't observe the event, it may be a classic lost opportunity.
However, we should still learn much from the ground-based spectroscopy. If Eta Carinae's recent history is any guide, some problems will be solved but unpredictable new questions will appear concerning this remarkable star.
An astronomer at the University of Minnesota, KRIS DAVIDSON studies the most massive stars, particularly luminous blue variables such as Eta Carinae.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||includes related article on possible observations associated with the supergiant star|
|Publication:||Sky & Telescope|
|Article Type:||Cover Story|
|Date:||Jan 1, 1998|
|Previous Article:||Mars Global Surveyor: You ain't seen nothin' yet.|
|Next Article:||The lives of binary stars.|