91av

Astronomers who probe the death throes of stars are having a successful run. Most of the catastrophic stellar explosions known as supernovae appear in distant galaxies, so they are difficult to study. But during the past seven years, two supernovae have erupted close to our own Galaxy, the Milky Way. And the most recent of these violent deaths, which astronomers observed just a year ago, has given us an unprecedented insight into the way stars evolve.

In February 1987, a star exploded in the galaxy next door, the Large Magellanic Cloud, which lies just 160 000 light years away. SN 1987A, the first supernova of 1987, was the first to be visible to the naked eye since 1604. Then on 28 March 1993, an amateur astronomer in Spain, Francisco Garcia, spotted a suspicious point of light in the inner spiral arm of M81, one of the brightest galaxies in the sky. Like the Milky Way, M81 is a giant spiral galaxy. Garcia hurriedly checked a picture of the galaxy and found that the point of light was new. Another amateur astronomer checked to see whether the point was an asteroid passing in front of M81. It wasn’t. Garcia rushed word of his supernova find to professional astronomers.

The next night, the point of light was twice as bright. By measuring its spectrum, professional astronomers confirmed that the object was indeed the blast of a dying star. Last year, dozens of supernovae appeared elsewhere, but Garcia’s has generated more excitement than all the rest combined. For of all the supernovae seen this century, SN 1993J – the tenth supernova found that year – was one of the nearest.

M81 lies only 11 million light years away, in a group adjacent to our own Local Group of galaxies. Because the galaxy is close, astronomers know its distance accurately, and for the same reason supernova 1993J was bright enough for astronomers to study the light for clues about the explosion. Moreover, M81 appears in the far northern sky, near the Plough, just 21 degrees south of the Pole Star. So, viewed from most of the northern hemisphere, M81 never sets. Astronomers could, therefore, observe its supernova all night and all year long.

The star that exploded lay 16 000 light years from the galaxy’s centre, or three-fifths the distance we are from the centre of the Milky Way. The fated star seems to have exploded around midnight on the evening of 27 March, less than 24 hours before Garcia discovered it. (Because M81 is 11 million light years away, the star actually exploded 11 million years ago, but the convention among astronomers is to say that an event occurs at the time it can be seen from Earth.) The short time between the explosion and its discovery meant that astronomers could observe the supernova almost from the moment the explosion began. In contrast, many supernovae are not discovered until weeks after the explosion, by which time they have started to fade.

BRIGHTER THAN BRIGHT

But as astronomers around the world turned their telescopes towards SN 1993J, the supernova brightened, reaching its greatest luminosity on 30 March. At that time it was intrinsically a billion times brighter than the Sun. From Earth it appeared as a star of magnitude 10.6, or one-seventieth as bright as the faintest star that can just be seen by the naked eye. It looked brighter than all the other supernovae that had appeared in the northern sky since 1954. SN 1987A, though brighter, had appeared in the southern sky. Many supernovae remain bright for weeks but SN 1993J faded quickly.

Then, to the surprise of observers, it brightened again, reaching a second peak on 18 April that rivalled the first. Since then, the supernova has steadily faded, but its legacy has revealed much about how stars explode. Its progress can be charted by a light curve, a plot of how the supernova’s brightness changes over time. The figure on the opposite page, which was drawn from data produced at Princeton University in New Jersey by Michael Richmond and his colleagues, clearly shows SN 1993J’s two peaks and steady decline. Because M81 is nearby, astronomers can see individual stars in it. By examining pre-explosion photographs of the galaxy, they were able to pinpoint the star that had exp-loded. The only other time this had been possible was when SN 1987A exploded.

Identifying the progenitors of the 1987 and 1993 supernovae confirmed some aspects of the theory of supernova explosions. According to this theory, a typical supernova results from a star that is born with more than eight times the Sun’s mass. The pre-explosion photographs confirmed that the progenitors of the 1987 and 1993 supernovae were indeed massive, having 15 to 20 times the Sun’s mass.

When a star is born, it consists mainly of hydrogen. The life of a high-mass star is fast and furious: it acts as a huge nuclear fusion reactor. When young, the star fuses hydrogen nuclei into helium nuclei at its core, just as the Sun does. Such stars are called main sequence stars and 90 per cent of stars in our Galaxy are of this type. In a massive star the core is more compressed and hotter. So the nuclear reactions proceed faster, generating much more energy, which makes its surface hot. As a star’s colour is directly related to its surface temperature, the star appears blue.

FAST BURNER

The star consumes its fuel so fast that it remains a main sequence star for only about 10 million years, a tiny fraction of the Sun’s 4.6-billion-year age. When the core runs out of hydrogen, the star burns hydrogen in a shell surrounding the core. The star’s outer layers expand and cool, causing it to change from a blue to a red star. Betelgeuse in the constellation Orion is an example of this type of star, which is known as a red supergiant. If the largest known red supergiant were placed at the centre of the Solar System, the star would engulf every planet out to Saturn.

In the supergiant, the helium formed as a result of the hydrogen burning simply sits in the star’s core, because helium burns at a higher temperature than hydrogen does. But as the pressure and temperature at the core build up, it becomes hot enough to fuse the helium nuclei, to create carbon and oxygen. Meanwhile, hydrogen continues to fuse into helium in a layer surrounding the core. During this stage, the star may contract and heat up, becoming a blue supergiant star like Rigel in Orion. Though dozens of times the size of the Sun, which has a diameter of 1.4 million kilometres, a blue supergiant is about a tenth the size of a typical red supergiant.

The star then burns other elements and again expands into a red supergiant. It fuses carbon into neon and magnesium, neon into oxygen and magnesium, oxygen into silicon and sulphur, and silicon and sulphur into iron. The red supergiant resembles an onion in which different layers are burning different elements: hydrogen burns in the outer layer, helium in a lower layer, and so on down to the core, where silicon and sulphur burn to form iron (see Figure). But iron means death.

At this point, the red supergiant faces a colossal energy crisis. It needs the outward pressure of the energy generated by the burning elements to balance the enormous inward pressure of the star’s mass. When iron nuclei fuse, the reaction does not produce energy – it takes in energy. So when the star’s core turns into iron, it can no longer generate the energy needed to support the outer layers of the star. Suddenly the star collapses inwards and explodes, producing a supernova. Most of the heavy elements of which we and our planet are made, such as the oxygen we breathe and the iron in our blood, were cast into the Galaxy by supernovae that exploded billions of years ago, before the Sun and Earth were born.

This catastrophic explosion was what happened in the Large Magellanic Cloud in 1987 and in M81 in 1993. After SN 1987A, however, the supernova theory needed fine-tuning. By examining pre-explosion photographs of the Large Magellanic Cloud, astronomers discovered that the star that exploded was not a red supergiant but a blue supergiant. The discrepancy arose, astronomers now believe, because the Large Magellanic Cloud has a lower abundance of heavy elements – such as carbon, nitrogen, oxygen and iron – than the Milky Way. Although heavy elements constitute a small fraction of a star’s mass, they affect the star’s structure such that a high-mass star with few heavy elements can explode while it is a blue rather than a red supergiant.

Like the Milky Way, and unlike the Large Magellanic Cloud, M81 has a high abundance of heavy elements, so according to theory the star that exploded in 1993 should have been a red supergiant. But to the astronomers’ surprise, the star turned out to be an orange supergiant, which is smaller and hotter than a red supergiant. This was closer to the mark than the blue supergiant, but still a puzzle.

PECULIAR SPECTRUM

Meanwhile, astronomers looking at the spectrum of the radiation emitted by the 1993 supernova found it had some peculiar features. Supernovae are classified into two types, type I and type II, depending on whether their spectra show the presence of hydrogen, the most common element in the Universe. Type I supernovae lack hydrogen; type II have large quantities of it.

Last April, astronomers obtained spectra of the M81 supernova. These revealed that hydrogen was present. So, like 1987A, the M81 supernova was of type II. All type II supernovae arise from massive stars. When a star that starts life 15 to 20 times more massive than the Sun evolves into a red supergiant, its outer atmosphere consists of 10 to 15 solar masses of hydrogen. This hydrogen shows up in the supernova’s spectrum when the star explodes.

Last May, however, Alex Filippenko and Thomas Matheson of the University of California at Berkeley discovered that the spectrum of SN 1993J was changing. The hydrogen lines began to disappear and the spectrum began to look more like that of a type I supernova.

In the 1980s, astronomers recognised that type I supernovae split into three classes: type Ia, Ib and Ic. Unlike all other supernovae, type Ia supernovae do not arise from massive stars but from dense Earth-sized stars called white dwarfs. A white dwarf explodes if it receives too much material from a companion star. But because a white dwarf is a burnt-out star that has little hydrogen, there will be no hydrogen lines in its spectrum when the star explodes.

Type Ib and Ic supernovae lack hydrogen too. But before the 1993 supernova, most astronomers believed that type Ib and Ic supernovae are more closely related to type II supernovae than they are to type Ia objects. The 1993 supernova confirmed this belief. Like type II supernovae, type Ib and Ic supernovae originate from high-mass stars. But they lack hydrogen because the exploding star has lost its outer atmosphere of hydrogen, either because the star has thrown out the hydrogen or because a companion star has taken it. Type Ib supernovae have helium, whereas type Ic supernovae do not. This suggests that type Ic progenitors have lost their hydrogen and their helium.

The locations of the supernovae provided evidence for the link between type II and type Ib and Ic supernovae. All three types appear only in galaxies that have high-mass stars. Type Ia supernovae, on the other hand, can explode in galaxies that have no massive stars. These are galaxies which consist entirely of old stars, since massive stars are short-lived.

The M81 supernova strengthened the link between type II and Ib supernovae. Filippenko and Matheson found that the M81 supernova, once type II, had transformed into a type Ib. This confirmed the belief that type Ib supernovae emerge from high-mass stars, just as type II supernovae do. In 1987, Filippenko had observed a similar transition in a supernova that had occurred in the galaxy NGC 4651. Unfortunately, because that galaxy was several times more distant than M81, and because the Sun moved in front of NGC 4651 shortly after the supernova was discovered, the observations were difficult and uncertain. Some astronomers dismissed it as a fluke, and a few even hinted that Filippenko had observed the wrong object. But the M81 explosion proved that a supernova could transform itself from type II to Ib.

So by the middle of last year, astronomers had found three peculiarities in the M81 supernova. First, the star that had exploded was an orange supergiant, not red. Secondly, the spectrum had switched from type II to Ib. And thirdly, the brightness of the supernova had peaked twice rather than once.

SEEING DOUBLE

Although these features challenged supernova theory, they also sharpened and refined it. Most astronomers now believe that the M81 supernova was unusual because the star that exploded had a companion star that circled it. Many stars are double, but in this case the companion star, which has not yet been seen, altered the main star’s evolution (see Figure on previous page). When the main star expanded into a red supergiant, its hydrogen atmosphere was pulled away by the companion’s gravity. This did two things. First, it made the supergiant smaller, and this made it hotter too, changing its colour from red to orange.

Secondly, because the orange supergiant had lost most of its hydrogen to the companion star, it had little left itself. When the star exploded, some hydrogen appeared in the spectrum, causing astronomers to classify it as type II. But as the material expanded and thinned, the hydrogen lines disappeared to leave a spectrum that looked like a type Ib. If the star had lost all its hydrogen, the supernova would have started as a type Ib and stayed that way. In 1987, Stan Woosley of the University of California at Santa Cruz and his colleagues had proposed that supernovae could transform themselves in just this way.

The small amount of hydrogen also explained the double peak in the supernova’s brightness. Before the explosion, hydrogen in the star’s outer atmosphere was electrically neutral: each atom consisted of one electron and one proton. But the shock of the explosion stripped away the electrons, leaving naked protons in the hydrogen layer. Then as the electrons rejoined the protons, they emitted the light that powered the first peak in the supernova’s light curve. Because there was little hydrogen, however, this peak was short-lived.

But then a second source of energy kicked in. The explosion had made radioactive nickel, which decays over weeks to cobalt, then iron. It was the radiation from this decay that powered the second peak in the supernova’s light. In fact, from the double-peaked light curve, teams led by Ken’ichi Nomoto of the University of Tokyo and Philipp Podsiadlowski of the University of Cambridge had predicted that the supernova would change from type II to Ib, the very metamorphosis that was later observed.

But in a normal type II explosion, the hydrogen layer is so massive and the first peak so prolonged that it blends in with the second, so astronomers see only one peak. And in a type Ib or Ic explosion, the first peak never occurs because there is no hydrogen in the outer layer. So again, only one peak appears.

If this scenario for the M81 supernova is correct, astronomers should soon confirm it by detecting the companion star, which should have survived the explosion. There is no unambiguous image of the companion in pre-explosion photographs because the light of the orange supergiant overpowered the companion’s fainter light. And astronomers cannot see it now because of the supernova’s glare.

But by 1995 or 1996, the supernova will have faded enough for astronomers to find the companion. Moreover, even if it began as a small star, the companion must have gained a large amount of material when it pulled off the main star’s hydrogen atmosphere. So the companion star should itself be bright and massive enough to see.

If astronomers succeed in locating the companion, they can study the exploded star further. A star that starts off with 15 to 20 times the mass of the Sun collapses into a neutron star when it dies, and this should circle the living star and tug on it. The luminous star will therefore move back and forth, and this should be detectable in a periodic shifting of the characteristic frequencies of its spectral lines, due to the Doppler effect. The magnitude of the luminous star’s Doppler shift may help astronomers to calculate the mass of the neutron star. As the supernova fades, astronomers still have things to look forward to.

Even though recent years have spawned two of the closest supernovae of the century, astronomers hope for the ultimate supernova – one in our own Milky Way. On average, several occur every century, but their light usually gets blocked by the dust and gas in our Galaxy. So only those that explode within some 10 000 light years of the Earth can easily be seen. On average, such a nearby supernova appears only once every few hundred years. The last was in 1604, so the next is due soon. It could appear in the sky tonight.

Ken Croswell is an astronomer in Berkeley, California. Further reading: The Supernova Story by Laurence A. Marschall (Princeton University Press, 1994).