HD 5980 AB(C?)

Jon Morse, Kris Davidson,
STScI, NASA


Larger infrared image.



Luminous Blue Variables
like Eta Carinae, at left,
appear to eject huge
amounts of their outer
gas envelopes in quick
but violent outbursts.

Once a massive star has fused most of its core hydrogen to helium ash, it becomes very unstable, and they eventually may blow up as spectacular Type-II supernovae. Some of the most massive ones may first pass through an LBV stage when they appear to eject huge amounts of mass from their outer layers of gas in a very short time (even more mass in a shorter period than Wolf-Rayet stars). The ejected gas (mostly hydrogen) moves outwards from the star at speeds of 50 to 500 kilometers per second and may be observed as nebulae that are bright in radio wavelengths. (Radio images of some LBVs and more discussion are available.)

Yves Grosdidier, Anthony Moffat,
Gilles Joncas, Agnes Acker,
STScI, NASA



Larger false-color image.




At least one of the stars of the
HD 5980 system is a Wolf-Rayet
star, like WR124 at left, which
is vigorously ejecting its outer
envelope of gas and dust (more
from STScI and APOD).

The chemical peculiarities of W-R stars are attributed to the presence of nuclear-processed material at the stellar surface. According to the standard evolutionary scenarios (A. Maeder, 1983; and Maeder and Meynet, 1994, and 1987), a massive O star (of 50-120 Solar-masses) is expected to rapidly swell when evolving toward lower temperatures after core hydrogen exhaustion. In this hypothesized evolution, such stars are believed to reach an instability limit when they enter a phase with even more rapid and massive loss of mass, that has been identified with the stars known as luminous blue variables or LBVs (Nota and Lamers, 1997; and Davidson, Moffat, and Lamers, 1989). LBVs are very luminous and unstable, hot hyper- or supergiants that undergo irregular eruptive events that lead to the formation of a very extended photosphere at maximum visual brightness. Generally at that stage, their winds are expanding slowly (100-200 km per second) and relatively low-effective surface temperatures are observed (7000-9000 Kelvins). Subsequent evolution with continued mass loss (of its outer hydrogen layer) leads to the creation of a WN star, followed in succession by a WC and a WO just prior to the supernova explosion that ends the life of the star. Hence, the expected evolutionary sequence is:

O --> Of --> LBV --> WNL --> WNE --> WC --> WO --> Supernova

John Bally, Dave Devine, and
Ralph Sutherland, STScI, NASA



Larger false-color image.



Astronomers believe that Wolf-Rayet stars begin
life as main-sequence O-type dwarf stars, like
Theta1 Orionis C, the brightest of the four
central stars of the Trapezium Cluster in Orion
(at left), located around 800 light-years away.

According to Koenigsberger et al (1998), the WNE stage is believed to be reached only after the helium core of the star is fully exposed, given the low abundance of hydrogen in WNE stars relative to WNLs. More recently, Langer et al (1994) proposed a modified scenario for the evolution of massive stars that incorporates theoretical results for violent pulsational instabilities (Kiriakidis et al, 1993; and Glatzel et al, 1993). Presumed to be driven by the "iron opacity bump," these instabilities could be producing pulsationally driven mass loss near the end of the core hydrogen-burning phase of massive stars. Langer et al proposed that at this stage, prior to the LBV phase, a massive star could have a WN-type spectrum, and so their proposed evolutionary sequence is:

O --> H-rich WN --> LBV --> H-poor WN --> H-free WN --> WC --> Supernova

Andrea Moneti, MSX Spirit-III



Larger true-color composite image.




O-type stars evolve very quickly, and
the most massive ones may soon become
eruptive, luminous blue variables, like
the Pistol Star, at left, which may have
already ejected half of its initial 200
Solar-masses since its birth as much
as three million years ago.

Koenigsberger et al (1998) argue that a close relationship between some of the WNLs (mainly WN9-11) and LBVs is now firmly established (Pasquali et al, 1997; (Crowther et al, 1995; and Stahl et al, 1983), supporting the connection between hydrogen-rich WNs and LBVs. In contrast to these theoretical models, the stars in the R136a region of the Large Magellanic Cloud's that display WNE-type spectra have recently been found to be extremely massive main-sequence objects that mimic the more hydrogen-depleted W-R stars (de Koter et al. 1997). Thus, WNE characteristics are not exclusively associated with postmain-sequence stars.

Uncertain Status of the HD 5980 System

In the case of HD 5980, Koenigsberger et al (1998) described the system as a very massive and luminous binary within which a major LBV-type eruptive event was observed (Barbá and Niemela, 1994; Barbá et al. 1995; Bateson & Jones 1994; Koenigsberger et al, 1994 and 1995). The magnitude of the eruption (with a change of magnitude greater than three) identified it as a major eruption, similar to those that occurred in Eta Carinae and P Cygni (where reviews of previous investigations can be found in Barbá et al, 1996, and 1997; Breysacher, 1997; Koenigsberger et al, 1996, and 1998; Moffat, 1997, and Moffat et al, 1998, and Niemela et al, 1997). Moreover, the binary system's orbit is eccentric, with a period of 19.3 days and an inclination of greater than 86° (Breysacher 1997; Moffat et al, 1997), where eclipses occur at orbital phases 0.0 and 0.36 (the star "in front" at orbital phase 0.00 (O7 I) is star A, while the star "in front" at phase 0.36 (WN4) is star B. According to Koenigsberger et al's "crude estimate" of the stellar masses and luminosities of the two stars within the binary system, both stars are very massive and extremely luminous, and assuming evolution with constant luminosity, the erupting star lies near the theoretical evolutionary track of a star with an initial mass of 120 Solar-masses.

Since the first observations in the early 1970s, the observed spectral type of the system has evolved in the following manner (Koenigsberger et al. 1994, 1996):

WNE --> WNL --> LBV-type eruption --> B1.5Ia --> WNL

Changes in the wind velocity leading up to the eruptive event (Koenigsberger et al, 1998) show a systematic decrease prior to the eruption and an increase in velocity after the eruption, where both were correlated with the spectral variations. Although the timescale over which the transition has been observed is very short (18 years) in terms of evolutionary timescales, it is possible that stellar evolution into the instability strip at its earliest stages is being observed.

However, there are several complications which conflict with the hypothesis that the system's eruptive behavior is actually a direct consequence of stellar evolutionary processes. First, the spectral type of HD 5980's primary has evolved from WNE (believed to be hydrogen-poor W-Rs) to WNL (which are usually hydrogen-rich), although the evolutionary sequence is in the opposite direction. Second, the pulsational instabilities referred to above are driven by the iron opacity peak. Given the significantly lower iron abundance in HD 5980 with respect to its Galactic counterparts (Koenigsberger, Moffat, & Auer, 1987), it is not clear whether these instabilities can be very effective in a system such as HD 5980. Finally, HD 5980's eccentric orbit has led some to propose that the eruption could have been induced by the oscillating gravitational forces within the system (Koenigsberger and Moreno, 1997; and Moreno, Georgiev, and Koenigsberger, 1997).

The fact that HD 5980 is a binary system has complicated analysis of the evolutionary state of its component stars. Foremost among these is the uncertainty as to which of the two stars in the system is the source of the outburst (Barbá et al, 1996; Koenigsberger et al, 1995; and Moffat et al, 1998). As mentioned above, in the late 1970s the system was composed of WN4 + O7 I stars (Breysacher et al, 1982). According to this classification, the first question would be avoided if the eruption were to have occurred in the O7 supergiant. However, a subsequent classification of the stars in the binary was given by Niemela (1988) and Niemela et al (1997) to be WN4.5 + WN3 (but see Moffat et al, 1998), so that regardless of which of the two stars underwent the eruption, the precursor would have had WNE characteristics. While the presence of a third stellar object within the system is now suspected (Koenigsberger et al, 2002; and Heydari-Malayeri et al, 1997), there is no question that the erupting star forms part of the 19.3 day eclipsing-orbit binary (Koenigsberger et al, 1995). The situation is confused further by the fact that two such massive and close binaries have very strong interaction effects, including wind-wind collisions (cf. Moffat et al, 1996; and 1998), producing phase-dependent variations in the spectrum that do not have a clear interpretation in terms of orbital motion. (For an extended discussion from which the previous exposition was taken, see: Koenigsberger et al, 1998).