(Return to course home page)

1. INTRODUCTION

Classical Novae: We see a bright classical nova about once per decade. A very faint binary star system suddenly brightens to about 10⁵ (100,000) Solar luminosities and then fades after a few months. The most recent classical nova visible to the naked eye (magnitude V = 4.4 at maximum brightness) was Nova Cygni 1992. Here is a recent Hubble Space Telescope image of Nova Cygni 1992, showing an expanding shell ejected by the explosion. (Be sure to read the caption. See also Nebula Nova Cygni turns on). When astronomers observe novae several years after outburst, they always find a close binary system containing a white dwarf. Evidently, nova explosions are not powerful enough to destroy the star systems that give rise to them.

Observing classical novae is a popular occupation of many amateur and professional astronomers. The link Novae and Supernovae illustrates the results of recent observations of novae, including light curves (graphs illustrating the fading of novae with time).

Supernovae: Supernovae are far more powerful events than novae. A star suddenly brightens to roughly 10⁹ Solar luminosities. In the Milky Way galaxy, a supernova occurs about once every 30 years or so (on average), but most are hidden behind the dark dust clouds of the Milky Way and are invisible to the naked eye. However, some (about 10%) have been close enough and bright enough that they were visible to the naked eye. The table below lists several such events that have been recorded during the past 2000 years and provides links to a few of the original historical records. A few of these events were very bright. The supernova of 1006 AD, for example, was about 1/10 as bright as the full moon!

The blast waves from supernova explosions produce powerful sources of radio and X-ray emission -- supernova remnants -- that astronomers can still see many thousands of years after the event. Links to modern images of the remnants of the historical supernovae are provided below. Astronomers have confirmed that the sightings reported in the table below were supernova explosions by: (1) finding such supernova remnants at precisely the locations where supernova events were noted in ancient sky charts; and (2) verifying that the historical date of the explosion is consistent with the date estimated from modern observations of the size and expansion rate of the remnant (the age is roughly the size divided by the expansion velocity).

Observing supernovae is a very active field of astronomy. Supernovae in distant galaxies, bright enough to observe with modern telescopes, appear somewhere in the sky at a rate of about one per second -- if only we knew how to find them! Until recently, astronomers found only a few dozen per year. Today, with sophisticated search techniques, they are finding a few hundred each year. The International Supernova Network is a group of amateur and professional astronomers dedicated to finding new supernovae. You should also check the Recent Supernovae Page of the Harvard-Smithsonian Center for Astrophysics. It has several images of recently discovered supernovae, plus a great graphic showing a real supernova image blinking off (before explosion) and on (after) in a galaxy. You can find the light curves of many recent supernovae in the link Novae and Supernovae.

There are two entirely different kinds of supernovae, as summarized in the table below. But first, you should read this: An introduction to supernovae.

The white dwarf star then accumulates a layer of unburned hydrogen on its surface. When about 10^-8 solar masses of hydrogen accumulates, the temperature and pressure at the base of this layer become great enough so that thermonuclear reactions begin. The hydrogen begins to fuse into helium. This reaction is explosive, and ejects the burning layer. The nova phenomenon is illustrated very nicely here.

Since the donor star has far more than 10^-8 solar masses in its outer envelope, it can replenish the layer of hydrogen on the white dwarf. This may take from decades to thousands of years, depending on how fast the donor is swelling up. At any rate, novae can repeat many times as a result. As an example, check this Hubble Space Telescope observation of debris from the recurrent nova T Pyxidis.

Energy is released when light elements fuse into heavier elements: hydrogen into helium; helium into carbon and oxygen; oxygen into silicon; silicon into iron. But since iron is the most tightly bound atomic nucleus, fusion of iron into heavier elements such as lead, uranium, etc., will not release energy. On the contrary, such fusion will absorb energy. (For the same reason, a uranium nucleus releases energy when it splits apart.) Likewise, it requires energy to tear iron nuclei apart. Iron nuclei have no more energy to release, either by fusion or fission. Be sure to check the curve of binding energy, which illustrates this point very clearly.

Now we come to one of the great questions of astrophysics: what accounts for the relative abundances of the heavy elements (i.e., heavier than carbon and oxygen). On Earth, and everywhere in the cosmos, some heavy elements such as silicon (found in dirt, sand, and rocks) and iron, are relatively common; but other elements, such as gold, silver, and uranium are rare, as illustrated here. Today, we can be certain of the answer: the heavy elements are formed in the explosions of stars.

How can we know this? The proof is a triumph of nuclear physics and astrophysics. Since the 1950s, physicists have been doing experiments to measure the rate of fusion reactions by accelerating atomic nuclei (such as carbon, oxygen, silicon, etc.) to very high speeds in machines such as cyclotrons. The fast nuclei then strike a target made of some other material, and the physicists measure the products of the fusion reactions that occur. After measuring the rates and temperature dependence of hundreds of such fusion reactions, the physicists can incorporate their results into hundreds of equations that characterize how a mix of all sorts of elements would evolve if it was heated to billions of degrees in a supernova explosion. They then solve these equations on a computer. A typical result (taken from W. D. Arnett, Supernovae and Nucleosynthesis) is illustrated below.

Instead of becoming a classical nova, a white dwarf in a binary system can have a far more violent fate as a result of mass transfer. When the layer of unburned fuel on the surface of the white dwarf explodes, the surface explosion may be violent enough to raise the temperature of the carbon/oxygen core of the white dwarf above 7 x 10⁸ K. If so, the carbon and oxygen burn explosively, making mostly iron (actually, the explosion makes radioactive nickel-56 that decays into cobalt-56 and finally into iron after a few years). The energy released by the explosion blows the white dwarf to smithereens, leaving nothing behind. We call this event a Type I supernova. It is an entirely different phenomenon than a Type II supernova (the result of core collapse of a massive star), to be discussed in the next section.

The scenario I have described above is illustrated very well in the following two places: Type II Supernovae, and And then it just blew up. Don't fail to review these sites. They have some diagrams that you will be expected to recognize and an excellent review of much of the material of this lesson.

On the evening of February 23, 1987, a young Canadian astronomer named Ian Shelton walked outside of the dome at the Cerro Tololo InterAmerican Observatory and took a photo of the Large Magellanic Cloud (LMC) with his Nikon Camera. Shelton developed the photo and immediately noticed a bright star where none had been seen before. He told his colleagues, and within hours the word had sped around the world, by phone, e-mail and fax. This was a nearby supernova -- the brightest one to be seen since 1604 AD, when Johannes Kepler observed a supernova in our own galaxy.

Astronomers immediately began to observe the supernova (called SN1987A) with every telescope they could use. Unfortunately, the great telescopes in Hawaii and the Northern hemisphere could not observe the supernova because it was far to the south. But astronomers could observe it almost continually with telescopes in Chile, Australia, and South Africa. In this image of the LMC, you will see a big bright nebula of glowing hydrogen gas in the upper left, called 30 Doradus or the Tarantula Nebula. In this magnified image of the Tarantula Nebula with SN1987A you can see the supernova to the lower right, and in this pair of images of SN1987A before and after the explosion, the arrow on the right panel points to the star that blew up. Before it blew up, it was already a fairly luminous (about 10⁵ times the Sun) blue star; but after the explosion it suddenly became 1000 times brighter (about 10⁸ times the Sun), as seen on the left panel.

Neutrino observations: Actually, there were two "observatories" in the Northern hemisphere that did observe SN1987A. They were huge tanks of water deep underground, all sides of which were covered with instruments designed to detect flashes of light that might occur in the water due to the interactions of fundamental particles. One, called the Kamiokande experiment, was in Japan; and the other, called the IMB experiment, was in Ohio (deep under Lake Erie). These experiments were designed to detect rare proton decay events predicted by theories of elementary particles, and the builders had no idea that they might detect neutrinos from supernova explosions. The neutrinos from SN1987A came right through the Earth and entered these tanks of water from beneath. A very few of them (12 in the Kamiokande detector and 8 in the IMB detector) interacted with atomic nuclei (protons) in the water to make very fast positrons, which in turn made light flashes in the water that the experiments detected. You can see the results here. From these observations, scientists were able to infer three properties of the neutrino burst: (1) the total energy (about 0.1M_Sunc², or 1/10 the mass-energy equivalent of the Sun); (2) the temperature of the neutrino source (40 billion K); and (3) the duration of the neutrino signal (about 10 seconds). All these values were just what scientists expected from theoretical calculations of the collapse of the core of a star to form a neutron star. So these observations of neutrinos from SN1987A provided very strong evidence that the supernova explosion was accompanied by the formation of a neutron star, as illustrated here.

Despite this evidence, and ten years of intensive observations, nobody has found any further evidence of a neutron star at the center of the debris of SN1987A. Why? We don't know. Read more about this puzzle in this article from Discover Magazine: Mystery of the Missing Star.

Optical light and gamma rays: The graph below shows a record of the optical light curve of SN1987A for the first 1400 days after the explosion was first observed. The crosses are the actual data. The graph shows that the supernova

The light from SN1987A faded at almost exactly the same rate observed in laboratories for the decay of the radioactive nucleus Cobalt-56 into the stable nucleus Iron-56 (the half-life of Cobalt-56 is 77 days). This was not a great surprise, because astronomers had long suspected that supernova explosions were responsible for the formation of the heavy elements in the universe. Theoretical calculations of the formation of heavy nuclei at billions of degrees (the temperature expected during the explosion) indicated that about this much Cobalt-56 would be formed. But the fact that the supernova light decayed just as expected was the strongest confirmation to date of the idea that supernova explosions really did make the heavy elements -- and, for the first time, we could measure exactly how much Cobalt-56 was made (0.07 Solar masses).

Spectrum evolution: The graph below shows the optical spectrum of SN1987A at three dates after outburst.

The top panel shows the spectrum on February 25, 1987, two days after outburst. You can see that the spectrum has a strong continuum that is rising at short wavelengths, punctuated by very broad emission and absorption lines of hydrogen and helium. This spectrum is very similar to that of a hot blue star (spectral type O). The broad absorption lines are caused by rapidly expanding gas above the photosphere; they are blue-shifted because we see the near side of the supernova, which is expanding toward us. As you will see in homework 4, you can measure the expansion velocity of the supernova debris from the blue shifts of these absorption lines.

Six months after the explosion (bottom panel), the continuum was almost completely gone from the spectrum, which was now dominated by emission lines. This kind of spectrum is called a nebular spectrum, because it resembles that of a gaseous nebula, such as the Orion Nebula. As the name suggests, the expanding debris of the supernova has now become so thin that one can see right through it.

The mystery of the circumstellar rings. A few months after SN1987A went off, a small telescope on the International Ultraviolet Explorer (IUE) satellite observed that narrow emission lines began to appear in its ultraviolet spectrum. This was a surprise: the supernova debris was expanding with velocities of thousands of kilometers per second; yet, the ultraviolet emission lines were narrow, indicating expansion velocities of less than 10 km/s. The ultraviolet emission lines continued to brighten, reaching maximum brightness 400 days after outburst, and then began to fade.

Several years later, the Hubble Space Telescope obtained this image of SN1987A (be sure to read the caption). The supernova was surrounded by a remarkable system of three rings of glowing gas! Where did these rings come from? They must be related to the supernova because the supernova is at their center. But they are expanding so slowly that they could not have been ejected by the explosion. In fact, if one divides the radius of the inner ring (6 x 10¹¹ km, or 0.6 light-years) by the expansion speed (10 km/s, or 1/30,000 times the speed of light), one finds that the inner ring must have been expanding for 20,000 years to grow to this radius. Thus, the inner ring must have been ejected 20,000 years before the supernova explosion.

If the gas was ejected by the star, why does it have the triple-ring shape seen by the Hubble Space Telescope? Astronomers really don't know -- this triple ring system is one of the outstanding mysteries of SN1987A. Some physical effect must determine the polar axis of the rings. We suspect rotation. But rotation of what? Many astronomers now believe that the parent star of SN1987A was actually a close binary system. Perhaps the rings of gas were ejected while the merger took place, 20,000 years before the explosion, as illustrated here.

In homework 4, you will see how the time variation of the ultraviolet emission lines seen by the IUE satellite is the result of illumination of the inner ring by the initial X-ray flash of the supernova. You will also see how to analyze this time variation to infer the physical diameter of the ring. By comparing the physical diameter to the angular diameter of the ring seen in the Hubble Space Telescope Image, you can infer the distance to the supernova. Finally, by dividing the radius of the ring by the expansion speed of the supernova debris, you can estimate when the debris will hit the inner ring.

The answer is: today! This is exciting, because we have never before seen a crash like this. With my students and colleagues, I have been working for a few years to predict what we will see when the crash occurs and what we will learn from such observations. We predict that the ring will become hundreds of times brighter than it is today at optical wavelengths. It will also become very bright in the radio, infrared, ultraviolet, and X-ray bands of the electromagnetic spectrum. That's great, because astronomers are now building telescopes that can observe SN1987A in these bands with unprecedented sensitivity, angular resolution, and spectral resolution. The Hubble Space Telescope will be the telescope of choice at optical and ultraviolet wavelengths. Next year, NASA will launch the AXAF X-ray telescope, and SN1987A will be one of the first targets it will observe. A few years later, we'll be able to observe SN1987A at infrared wavelengths with NASA's Stratospheric Observatory for Infrared Astronomy and Space Infrared Telescope Facility, and at radio wavelengths with the Millimeter Array Telescope.

We simulated the impact of the supernova debris with the inner ring by solving the equations of gas dynamics on a supercomputer. We assumed that the ring was a torus (i.e., shaped like a donut). Click here, to see the results of the simulation, and be sure to watch the MPEG movie. (It will take a few minutes to download this MPEG over a modem, but then you can run it again quickly to see the action.) In this simulation the yellow circle represents a cross-section cut through part of the ring and the supernova is off to the left. The blast wave from the supernova enters the ring 12.7 years after the explosion, or AD2000. It compresses the gas to high density and high temperature (shown as red). It continues to crush and shred the ring, until about 30 years after the explosion, when the ring is totally destroyed. Once we know the temperature and density of the shocked gas, we can calculate the intensity and spectrum of the radiation it emits in the optical, infrared, ultraviolet, and X-ray bands. With such calculations, we predicted that the ring will brighten by a factor of several hundreds after the impact occurs.

Today, all these predictions are coming true. With the Hubble Space Telescope (HST) we are seeing ultraviolet and optical emission from the fast-moving debris of the supernova. We also see a rapidly brightening "hot spot" on the ring, evidently the spot where the supernova blast is making first contact with the ring. Check here for the details of a press conference that some friends and I gave at NASA headquarters last month. But that's already old news. Just two days ago (Tuesday March 3), a friend of mine from the University of Texas sent me new optical spectra of SN1987A that he and his colleagues observed last week at the European Southern Observatory in Chile, showing that the optical spectrum of the ring is changing fast. We haven't even identified the spectral lines yet. Moreover, more observations with the HST are scheduled for Saturday, March 7. Keep watching this page. This story will change fast!

All this action is summarized in this movie that the Space Telescope Science Institute made. The first part shows the flash of the supernova, and then the ring lighting up due to illumination by X-rays from the supernova flash. We see the near side of the ring first because it is closer (see homework 4). About 400 days after the initial flash, the whole ring was bright -- hundreds of times brighter than it is today. This would have been a great show; but nobody could see it. The HST hadn't been launched yet and no ground-based telescope could see the ring at that time. The next segment of the movie shows the ring fading over 10 years, while the invisible debris of the explosion races out toward the ring. Then the first bright spot appears on the ring, as we see today. Soon, several other spots will appear, growing and merging. Finally, 10 years from now, the entire ring will become hundreds of times brighter than it is today.

In fact, we can still see remnants of supernovae that occurred in prehistoric times. One famous example is the Cygnus Loop (also called the Veil Nebula), the remnant of a supernova that occurred approximately 15,000 years ago at a distance of about 2,500 light years. With a diameter of 3^o (six times that of the Moon), it is one of the most spectacular objects in the sky. Here are some images of the entire Cygnus Loop as seen in optical, X-ray, and infrared bands. Here's an optical close-up of part of the Cygnus Loop seen from the Hubble Space Telescope. Another famous example is the Vela supernova remnant (optical, X-ray) the remnant of a star that exploded 11,300 years ago at a distance of about 1,500 light years. (Watch out! Several sites on the Web say 6,000 light years. That's wrong.) We can date the Vela SNR more accurately than the Cygnus Loop because it contains a pulsar, a spinning neutron star that is the remnant of the supernova explosion (we will discuss pulsars in the next lesson).

As you can see, supernova remnants look different depending on the wavelength band in which they are observed. That is true because the radiation comes from different constituents of the compressed gas. For example, the optical radiation comes from gas heated to temperatures of order 10⁴ - 10⁵ K, while the X-rays come from still hotter (10⁶ - 10⁷ K) gas. You can see that the X-ray emission comes from inside the shell of optical and radio radiation, indicating that the hottest gas is inside the blast wave (for example, check the superposed radio and X-ray image R,X of IC433, the remnant of SN837). The infrared radiation from a SNR comes from grains that are heated by the gas. The radio emission is most interesting; it comes from cosmic ray electrons that are accelerated up to velocities nearly equal to the velocity of light. They gyrate in turbulent magnetic fields compressed by the blast, emitting synchrotron radiation, which we will discuss in the next lesson.

After 20,000 years or so, the blast slows down enough that the remnant becomes invisible. But the cavity of low density hot gas created by the explosion will persist for a long time -- perhaps millions of years -- before the interstellar gas flows back in to fill it. Since massive stars are born in clusters, there's a good chance that another star in the same cluster will blow up before the cavity from the previous supernova is filled in ... and another, and another. The repeated action of many supernovae may cause the cavity to continue to grow, to diameters of hundreds of light years. Astronomers have observed many such supershells in the Milky Way and other galaxies. The compressed interstellar gas on the periphery of the supershell may begin to fragment and collapse under its own gravity, giving rise to propagating star formation, as illustrated in your text, Fig. 19.19.

(Return to course home page)

Last modified May 1, 1998
Copyright by Richard McCray