Most stars lose mass at a very high rate as they reach the end of their active life, namely, as their nuclear reaction is about to cease. By that time the core of the star is compact and very hot, such that nuclear reactions occur at a very high rate. Consequently, the star is very luminous, thousands to million times the present luminosity of the sun, with more massive stars being more luminous. The outer stellar region, the envelope, on the other hand, is very large, tens to hundreds times the present size of the sun.  The star is a giant, with a surface temperature about half that of the present sun, and has a red color. The star is a red giant.

The large radius of red giants implies a relatively weak gravity on their surface. The huge luminosity then, can push material from the very outer layer of the star, and accelerate it to a velocity high enough for the gas to escape the star. This outflow of gas is called a stellar wind. The present sun also has a wind. But in about seven billion years from now, when the sun will by a fully developed red giant star, it will lose about hundred-million times more mass each second.   

Stars, beginning their life with a mass between one and eight times the solar mass live and die in the following way. They convert hydrogen to helium via a nuclear process, termed fusion, in their hot (about ten million degrees Kelvin) core for most of their lives. The sun has been in this stage for 4.6 billion years, and has six billion years more to stay in this stage. After the hydrogen in the core is exhausted, the core contracts and heats up, until it is hot enough, about hundred million degrees, and the star starts to fuse helium into carbon and oxygen.

At the same time hydrogen supplied by the envelope continue to be converted to helium on the outskirts of the core. The star is a red giant. This stage lasts for a length of about ten percent of the length of the previous stage. At the end of this stage the star is a fully developed red giant with a huge mass loss rate.  The massive wind removes the entire envelope, leaving the bare core. Without fresh hydrogen the nuclear reaction turns off. The remnant is a dense core with a mass a little less than the mass of the sun, and having a size of only one percent of the present solar size, namely, about the size of Earth. After nuclear reactions cease, the remnant is very hot and luminous, but then cools and fade for ever, as a white dwarf. This will be the fate of the Sun, starting about seven billion years from now.

The newly born white dwarf is surrounded by the dense wind that once was the stellar envelope of the red giant progenitor of the white dwarf. After about hundred thousands years, which is a brief astronomical period, this matter is mixed with the gas residing between stars in the galaxy.  During these hundred thousand years the white dwarf heats the expanding gas to about ten thousand degrees. The warm expanding gas is glowing in the visible, and is termed a planetary nebula.  The name was coined in the 19^th century, because the  telescopes of that time could not resolve well the nebulae, and showed a small spot of light that reminded the astronomers a planet.  Figure 1 presents several such images taken with the Hubble space telescope. Note that the colors in the images are not real, but rather represent small variations in the physical conditions within the nebulae.  As can be seen, these nebulae are not circular, as their progenitor star was, but rather posses beautiful structures of many different kinds. 

Figure 1: Images of some planetary nebulae observed by the Hubble Space Telescope. From left to right: First row: IC 418; The Eskimo nebula (NGC 2392); NGC 7009; and a round planetary nebulae IC 3568. Second row: M 2-9; Hubble 5; and He 2-47. Note that the colors in the images are not real, but rather represent small variations in the physical conditions within the nebula.

Note:  The images of the planetary nebulae shown in figure 1 were taken from the following sources:
http://hubblesite.org/gallery/album/nebula_collection/pr2000028a/
The Eskimo nebula:
http://hubblesite.org/gallery/album/nebula_collection/pr2000007a/
NGC 7009
http://hubblesite.org/gallery/album/nebula_collection/pr1997038g/
IC 3568
http://hubblesite.org/gallery/album/nebula_collection/pr1997038c/
M 2-9:
http://hubblesite.org/gallery/album/nebula_collection/pr1997038a/
Hubble 5:
http://hubblesite.org/gallery/album/nebula_collection/pr1997038f/
He 2-47
http://hubblesite.org/gallery/album/nebula_collection/pr2007033b/

Figure 2: Left: An expanded view of the planetary nebulae NGC 6543: The Cat’s Eye nebula. Right: A detailed view of the inner region. Note the spherical halo (in green) compared with the non-spherical inner region. The filaments and blobs in the halo results from the stochastic nature of the wind that formed the nebula.

Note:  figure 2 is taken from: google images


The main unsolved question is how planetary nebulae acquire their non-spherical structure? There are several models proposed by different researchers, but there is no acceptable theory yet for the formation of the impressive nebular structures observed in planetary nebulae.  These types of structures have axi-symmetry, and hint to the role of rotation. As I have been arguing for 15 years now, a single star cannot sustain enough angular momentum to have the required degree of rotation. For that a stellar, or even a massive planet, companion is required. Indeed, in recent years, a consensus started to emerge, and most researchers in the field agree now that a binary companion must be involved in forming these shapes.

The same question applies to nebulae around other types of evolved stars, including stars as massive as eight times as the sun or more, that end their lives in a supernova explosion that forms a neutron star or a black hole.  The nebulae around these stars are formed before the explosion. In Figure 3 two nebulae around massive stars are presented, together with two planetary nebulae. The triple ring system of SN 1987A, a supernova that exploded 21 years ago, was formed thousands of years before the explosion. The exploding star is the very small dot in the center. The progenitor of SN 1987A was a star about twenty times as massive as the sun.
      
The nebula of the massive binary system Eta Carinae was formed in an eruption during the years 1837-1856. The primary star is about 150 times more massive than the sun, and its companion is about 40 times as massive as the sun. They will explode as supernovae within few millions years; first the more massive star, and later the less massive companion. The orbital period is five and a half years, from which we conclude that the two stars are close enough to interact with each other. Namely, during the 19^th eruption the more massive star transferred mass to the less massive companion. The companion then blew two jets that shaped the nebula (see below). The Red Rectangle nebula (lower left in fig. 3) might evolve to become a planetary nebula. The central star is known to have a binary companion with an orbital period of a bout a year. The binary systems at the center of the Red Rectangle and Eta Carinae strengthen the binary model for the shaping of nebulae around evolved stars.

FIGUR 3: Four images showing non-spherical nebulae around evolved stars (Colors are not real). Upper left: Supernova 1987A with its triple ring system. This star exploded 20 years ago. The rings were formed thousands of years before the explosion. Note that one large ring is in front of the star and the other large ring is behind the star. The small ring is located around the star. All rings are circular, but seen elliptical because the symmetry axis is tilted from us.  Upper right: The massive binary star Eta Carinae and its bipolar nebulae called Homunculus.  The massive star is 150 times as massive as the sun, and the orbital period is 2024 days. The star is expected to become a supernova in the relatively near future. (Credit: J. Hester/Arizona state University) Lower left: The Red Rectangle. A star is evolving to become a planetary nebula. It has a stellar companion with an orbital period of 322 days. (Credit: NASA; ESA; Hans Van Winckel (Catholic University of Leuven, Belgium); and Martin Cohen (University of California, Berkeley) Lower right: The Hourglass planetary nebula (MyCn 18). This nebula was blown by the central star, now a dying sun-like star seen as a white dot in the center. The central star heats the nebula to ten thousand degrees. (Credits: Raghvendra Sahai and John Trauger (JPL), the WFPC2 science team, and NASA).

The formation mechanism of these types of nebulae is connected to a wider question, the most puzzling one regarding stellar evolution: What is the role of rotation in stellar evolution, and how does stellar rotation evolve from the birth to the death of stars. 

For about a hundred years, from the middle of the 19^th century to the middle of the 20^th century, the major question in the field of stellar structure and evolution was the energy source of stars. For half a century it was thought that the energy source of stars is gravitational energy. This view was supported, among others, by William Thomson (Lord Kelvin), who was involved at the end of the 19^th century in fierce arguments with other scientists on this fundamental question: Is it gravitational energy or another type of energy that power the stars.  At the beginning of the 20^th century it became clear that nuclear processes are the source of energy in the sun and other stars. It took another 50 years and the efforts of great scientists, such as Arthur S. Eddington and Hans A. Bethe, to derive the major nuclear reactions in stars.

During last forty years the major open questions in the field of stellar evolution are related to the role and evolution of rotation (or angular momentum) of stars and gaseous disks around stars. Are magnetic fields required to transport angular momentum in gaseous disks around stars?  How two oppositely well collimated jets are launched by such disks? How are jets that are responsible to the gamma ray burst phenomena are formed from collapsing of rotating massive stars? What is the shaping mechanism of gaseous nebulae (as presented in this article) around evolved stars?

For about forty years there is a debate on whether single stars can form such non-spherical structures (as seen in the figures) or whether a binary companion is required for this non-spherical mass loss (“wind”) geometry from evolved stars.  
With the high detailed images brought by the Hubble Space Telescope and several other large telescopes, and the triple ring system observed around SN 1987A, this debate intensified about 15 years ago (see Balick & Frank 2002). It is now becoming clear that a singly evolved star cannot rotate fast enough to form non-spherical nebulae, and a binary companion is required. In some cases even a massive planet is sufficient to lead to a non-spherical wind. However, in the solar system Jupiter is to light and too far from the sun to influence its evolution. The sun will form a faint circular nebula, a `boring’ nebula.

There are several processes by which a companion can lead to the different structures observed. They can be classified into two main groups. In one group of processes the companion is swallowed by the primary red giant star. The companion can be a massive planet, an unevolved star similar to the sun, or a white dwarf that is itself a remnant of a sun-like star, and itself formed a planetary nebula in the past.  The companion then spirals inside the red giant’s envelope, toward its core. This process is termed “common envelope” formation.  In some cases the companion survives, in other it is destroyed in the envelope. An example is the progenitor of SN 1987A, which is believed to have swallowed a companion about 20 thousands years before its explosion.  In the common envelope process the companion expels mass in the equatorial plane as it enters the red giant’s envelope. This might be the explanation for the small inner ring in the triple ring system of SN 1987A (Figure 3). Another example is the planetary nebulae NGC 2346 (see Balick’s site); its central white dwarf has a companion very close to it, with an orbital period of only 16 days. This system went through a common envelope evolution when the progenitor was a red giant. 

In the second type of binary interaction the companion stays outside the envelope and influences the primary stellar wind from outside. This is the case in the massive binary system Eta Carinae, and in the Red Rectangle (both nebulae are presented in Figure 3).  In the main shaping process the companion accretes mass from the giant star via its gravitational attraction. The mass is not accreted directly onto the companion, but rather it spirals in an accretion disk.  This disk blows two jets, one into each of its two sides. The physical process that forms jets is unknown.

Each jet inflates a lobe, or a bubble, as it expands into the wind of the giant star. Namely, we have two kinds of wind here: a slow wind that is blown by the giant star, and a collimated fast flow. Two opposite jets, blown by the companion, expand into the slow wind. This kind of interaction explains the two lobes observed in many systems, such as Eta Carinae (figure 3), and the planetary nebulae M 2-9, and Hubble 5 (Figure 1). Sometimes the system has several episodes of jets formation and in different directions. In such case several pairs of lobes are formed, such as in He 2-47 (figure 1).

Interestingly, jets are observed in completely different astrophysical objects, where in some cases they form structures similar to those in planetary nebulae. Most intriguing is the similarity to clusters of galaxies. These systems are about million times larger than a typical planetary nebula. The medium between the galaxies is filled with tenuous hot gas, tens of million of degrees Kelvin. Opposite jets blown by a massive black hole inflate a pair of bubbles. The black hole is about a billion times as massive as the sun, and it resides at the center of a large galaxy at the center of the cluster. Because of its very high temperature the gas in clusters of galaxies glows in the X-ray band rather than in the visible. For that, an X-ray observatory from space must be used. An example is given in Figure 4. An X-ray image of the hot gas in the Perseus cluster of galaxies as taken by the Chandra X-ray space telescope is presented. It is compared to the visible image of the Owl planetary nebula. We therefore have at hand a unified mechanism for the formation of pairs of bubbles over many orders of magnitude in system size. A unified mechanism is always considered an achievement in science.

Figure 4: Blowing bubbles by jets. At the centre of both the Owl planetary nebula (left) and the Perseus cluster of galaxies (right) is a pair of bubbles. These bubbles are filled with low density gas, and therefore look fainter than their surroundings. The bubbles in the galaxy cluster are about 100,000 times larger than those in the planetary nebula. In clusters of galaxies, such bubble pairs are known to be formed by two jets. The Owl nebula is shown in visible light (From M. Guerrero and collaborators); the Perseus image was taken by the Chandra X-ray Observatory.

Astrophysicists must continue their quest for the different shaping mechanisms of nebulae formed by evolved stars. Projecting from the hundred years that were required to understand the production of energy in stars, we may expect that this research will take tens of years more before we fully understand the shaping mechanism of nebulae around dying stars.

To read more:

1. For more popular information, with many other images of planetary nebulae and links, See the WEB page of Bruce Balick:http://www.astro.washington.edu/balick/WFPC2/
2. A scientific paper reviewing the subject of asymmetrical planetary nebulae is Balick, B., & Frank, A. 2002, Annual Review of Astronomy and Astrophysics, Vol. 40, p. 439-486
3. For a paper comparing images of planetary nebulae with that of clusters of galaxies, see the Appendix of Soker, N. & Bisker, G http://arxiv.org/PS_cache/astro-ph/pdf/0601/0601032v1.pdf