XI. Supernovae and The Universe

1. Probing the Size, Shape, and Fate of the Universe with Supernovae

A. Supernovae as Signposts

Apart from their intrinsic interest as star-destroying explosions, supernovae have other uses simply because they are so bright. Their great luminosity means that they are visible across the Universe. More specifically, supernovae are signposts that determine the distances to their host galaxies. Careful measurements of those distances allow astronomers to map out how fast the Universe is expanding and hence how old it is, the curvature of space, and clues to the fate of the Universe. The use of supernovae in this way has expanded extensively in the last few years and the results have been dramatic. Supernovae have provided clues that the Universe will expand forever and that it is even now in the grip of powerful repulsive forces that accelerate its outward rush.

To use supernovae as a technique to measure distances requires some perspective on what we are trying to accomplish and how we are doing the task. Recall the various two-dimensional analogs we have employed to picture curved space. The two-dimensional space around a gravitating object is funnel-like when viewed from the perspective of three dimensions. The two-dimensional analog of the Universe itself, at one moment of time, can be represented as the surface of a sphere, an infinite flat plane, or a saddle extending "upward" to infinity fore and aft and "downward" to infinity sideways as shown in Figure 11.1. These two-dimensional analogs are formally embedding diagrams. Informally, they help picture curvature in three dimensions. We have stressed that looking down from a higher, three-dimensional, perspective is cheating in a sense, because there is no way we can look "down" on our three-dimensional curved space from an "outside." That "outside" would, by analogy, itself have to be a fourth spatial dimension. If there were an observer in that fourth spatial dimension, that observer could "see" the curvature of our Universe or that around the Earth or around a black hole in much the same way that we can "see" the curvature of the surface of a sphere. On a more direct and personal level, such an observer would also not be limited to viewing our surfaces, our skin and facial features as we do one another. An observer from a hypothetical fourth dimension would also be able simultaneously to "see" our volume, our guts and bones, much as we can see the interior of a circle inscribed on a sheet of paper. This is an amusing perspective, but not one of physics.

Rather, the proper perspective is to recognize that a two-dimensional creature living in any of these curved two-dimensional spaces could determine that the space curves and by how much by doing geometry, by carefully measuring distances and angles. That is now our task! We are three-dimensional supernova observers trapped in our three-dimensional Universe. We must determine the curvature of our three-dimensional space without stepping "outside" of three dimensions, something we simply cannot do. Fortunately, we do not need to step outside. We just have to be careful with our geometry and our astrophysics.

FIGURE 11.1: Einstein’s theory tells us that the Universe must have one of three basic shapes. The two-dimensional analogies (embedding diagrams) for these cases are a spherical surface (a "closed universe"), a flat plane extending to infinity in all directions (a "flat" universe), and a "saddle shape" that also extends to infinity in all directions (an "open" universe). Two-dimensional astronomers in two-dimensional universes cannot stand "outside" their universes to "see" the nature of the curvature the way a three-dimensional "hyperspace" observer can. Rather, they can do geometry in the context of their own space and determine the shape of their universe. Triangles in flat space will have their interior angles sum to 180 degrees, but the answer will be more than 180 degrees in the spherical universe and less than 180 degrees in the saddle-shaped case. As three-dimensional astronomers in our own three-dimensional Universe, we cannot stand outside of it in hyperspace, but we can do geometry to determine the nature of the Universe we occupy.

B. Type Ia Supernovae as "Calibrated Candles" and "Understood Candles"

The use of supernovae to measure distances is based on a simple principle: things further away look dimmer. Turned around, how dim a supernova appears to be is a measure of how far away it is. The basis for this intuitively reasonable notion is that when light spreads out from a central source equally in all directions, the locus of the photons emitted at a given time defines a larger and larger surface. The light falling on a detector of a given area, a human eyeball or a telescope, then captures a smaller and smaller fraction of the total the further away the detector is from the source. The fraction decreases just as the total area into which the radiation floods increases and that goes like the distance squared (the area is 4¹D2, where D is the distance). This means that the apparent brightness of a source of a given total luminosity decreases like the inverse of the square of the distance. In simple terms, the fainter a given kind of object appears, whether it is a porch light, a star, or a supernova, the further away it must be. If you know how bright the object really is, then you can tell from how bright it apparently is how far away it must be. This gives us a powerful tool for measuring distances. The key is to figure out how bright a given object really is.

Recall that Type Ia supernovae are generally the brightest of all the different types. This makes them especially good sign posts for measuring large distances. If we knew exactly how bright they were, the task of measuring distances would be rather easy. We would just see how bright a supernova looked in a given telescope and read off the distance. The immediate problem is to determine the intrinsic brightness of a given supernova.

For a long time there was some reason to believe that Type Ia supernovae were all equally bright. That would have made the task of measuring their distances particularly easy. The jargon for this is that such identical supernovae would represent a standard candle. The idea is that if you have a set of "candles" of identical, known brightness they can serve as a "standard" with which to compare other sources of luminosity and to measure distances. In the last decade we have determined that Type Ia supernovae are not exactly the same, but that the differences are systematic. That allows astronomers to make allowances for the differences between individual Type Ia supernovae.

In particular, astronomers have found that the Type Ia supernovae that are intrinsically brighter decline in brightness more slowly than those that are intrinsically dimmer. We believe that we even have a basic understanding of why this is true. Some variation in the exploding white dwarf causes variation in the amount of radioactive nickel-56 produced in the explosion. The extra energy from radioactive decay not only makes the supernova brighter, it keeps the expanding matter opaque longer. The radiation takes longer to leak out, giving the slower decay. The trend that relates the brightness of the supernova to the rate of decline from peak light gives the means to determine the brightness of the supernova. One just needs to see how fast the supernova declines and that tells you how bright it really is. Comparison with how bright it seems in the telescope then gives the distance.

There are two ways of doing this comparison. One uses only the empirical data from the supernova with no attempt at a theoretical understanding. This method requires some comparison with other astronomical objects for which the distances are established in some other way. This calibration sets the overall scale of just how bright a Type Ia supernova with a given rate of decline really is. This must be done for as many supernovae as possible for which the distance is already known (in practice a dozen or so, with the sample growing as this book is being completed). Then the brightness-decline relationship gives the intrinsic brightness and hence the distance from a measurement of the decline rate alone. This technique uses Type Ia not as "standard candles," but as light sources for which the brightness of each supernova can be calibrated compared with known sources, hence the phrase "calibrated candles."

The other technique to employ Type Ia supernovae to measure distances uses theoretical models of the explosions to determine how bright the supernova must be to produce a given light curve. This technique thus attempts to employ "understanding" rather than "calibration" to provide the necessary information to turn the decline rate into a known intrinsic brightness. This technique thus uses Type Ia supernovae as "understood candles."

The first technique, using the Type Ia supernovae as "calibrated candles" is only as good as the calibration and the implicit assumptions that underlie the empirical relation between peak brightness and the rate of decline. A key assumption is that the brightness-decline relation is unique. Two supernovae with identical decline rates are assumed to have the same intrinsic peak brightness. The second technique, using Type Ia as "understood candles" is only as good as the rather complex underlying theory of the explosion and of the production of luminosity. This method can, in principle, allow for cases where, because of more subtle circumstances, other variables enter and two supernovae with the same decline rate do not have the same peak brightness. The two methods agree rather well. They both give the same age of the Universe.

D. The Age of the Universe

The Universe we see around us began in what we call the Big Bang. There are still mysteries surrounding how the Universe came to be. We will touch on some of them in the last chapter. There is, however, no doubt that the visible Universe arose in a very dense, hot state, and expanded outward. Although the first instants are murky, ordinary particles, protons and electrons formed very quickly and the Universe was pure hydrogen for a while. The light elements --helium, lithium -- formed when this expansion was a few minutes old. When it was a million years old, the matter got sufficiently dilute that the radiation from its heat could stream freely. We see that radiation as the cosmic background radiation that comes at us from all directions. This cosmic radiation is red shifted by the expansion that pulls everything in the Universe away from everything else. We understand this process very well. Further expansion of the Universe brought the agglomeration of matter into galaxies, stars, and planets in ways we are still striving to understand. Continued expansion pulls all the distant galaxies apart. Understanding the expansion of the Universe allows us to measure its age.

It is important to realize that the Big Bang did not occur as an explosion in a pre-existing space, like a bomb in outer space. Rather space itself expanded, carrying the matter with it. One popular analogy is the behavior of spots on the surface of an expanding balloon. The spots do not move with respect to the rubber surface as the balloon expands, but they become ever further apart, as shown in Figure 11.2. A three-dimensional analogy is raisins in a rising loaf of bread that never drift in the dough, but again move ever further apart until the loaf stops rising. The second analogy is limited and a little deceptive because the loaf of bread is finite. The three-dimensional loaf of bread is surrounded by ordinary three-dimensional space into which it expands, whereas the space of the Universe is all-encompassing. The first analogy is limited because it is restricted to two dimensions, but it is more revealing in a way. One can see that the two-dimensional surface of the balloon has no two-dimensional "outside," neither the outside as we understand it from our three-dimensional perspective nor what we regard as inside the balloon, which still requires going off into a third-dimensional "hyperspace" from the perspective of a two-dimensional creature inhabiting the two-dimensional surface. Likewise, the loaf of bread is perceived to have a center, whereas (ignoring the opening though which one blows) there is no two-dimensional center to the two-dimensional surface of a perfect sphere to which the balloon is an approximation. Unlike the loaf of bread, the balloon shows that if attention is restricted to the confines of the dimensions of the space, two for the surface of the balloon, three for our Universe as we perceive it, there is no "center" and there is no "outside." These are tricky and fascinating issues and we will return to them in the last chapter.

FIGURE 11.2: A small piece of any two-dimensional universe will appear flat. As the universe expands after its Big Bang, this piece of the universe will expand, drawing all the galaxies in it further apart with time. A three-dimensional hyperspace observer could "see" this expansion, but two-dimensional astronomers resident in the two-dimensional universe could determine the expansion by registering the Doppler red shift as all distant galaxies move apart from all others. As three-dimensional astronomers in our own three-dimensional Universe we cannot stand outside, but we can measure Doppler shifts of distant galaxies and determine how fast the Universe is expanding.

For our current purposes, it is sufficient to picture the expansion of the balloon and its dots or the bread and its raisins to understand how to measure the age of the Universe. The effect of the expansion of the Universe is still much the same as an explosion in pre-existing space even if the concepts are radically different. If you can measure how far away something is from you, say a distant supernova, and determine how fast it is traveling away from you, by measuring its Doppler shift to the red, then you can tell how long it has been traveling to get as far as it has. You get the same answer for every supernova and every galaxy. The faster they move away from us, the more distant they are, but they took the same time to get there, drawn by the expansion of the underlying space.

The parameter that is measured in this way is called the Hubble Constant, after Edwin Hubble who pioneered this sort of measurement of distances and determined the nature of the Universal expansion. The Hubble Constant tells you how fast something will be moving away from you at a given distance. Both techniques for measuring the distances to Type Ia supernovae outlined in Section 2, and other techniques as well, say that velocity will be about 65 kilometers per second for every million parsecs in distance. The age is related to the inverse of the Hubble Constant. Obtaining the age of the Universe from the Hubble Constant involves another subtlety since it depends on the curvature of space and the acceleration of the Universe. Neglecting that subtlety for the moment, the corresponding age of the expanding Universe is roughly just the inverse of the Hubble Constant. If a supernova moving at 65 kilometers per second is a million parsecs away, it must have been moving away from us for about 10 to 15 billion years. If another supernova is moving away from us at 650 kilometers per second and is at 10 million parsecs, then the time for it is get there is just the same, 10 to 15 billion years. We get the same answer for every supernova, as we must since we are measuring the same age in every case, the age of the Universe.

E. The Fate of the Universe

The game is not over with the measurement of the Hubble Constant. It is not enough to measure how old the Universe is. We want to know what will happen to it in the future. Since the days of Hubble, astronomers, cosmologists, have been engaged in a grand quest to determine the "fundamental parameters of the Universe." This quest was shaped by Einstein’s theory of gravity. The first attempts to apply Einstein’s theory to the whole Universe showed that there were three parameters that would describe the whole shebang: The Hubble Constant, the overall curvature of the Universe, and the rate at which the Universe is decelerating its expansion due to the gravitational pull of the matter and energy within it. The issue of curvature is whether the Universe is the three-dimensional analog of the surface of a sphere, a flat plane, or a saddle, as shown in Figures 11.1 and 11.2. Einstein’s theory showed that it had to be one of the three and that if it were "sphere-like" it would have a finite life and recontract to a singularity, if it were "flat" it would expand forever, just reaching zero expansion rate at the end of time, and if it were "saddle-like" it would expand forever at finite velocity. The models of the Universe that defined these fundamental, and measurable, parameters, made some simplifying assumptions. We will see in the last chapter that these parameters may not tell the whole story, but they make up a critical part of it. Determining these parameters has occupied cosmology for most of the Twentieth Century.

There are various ways of going about measuring the other two parameters in addition to the Hubble Constant. Using supernovae, the underlying theory requires the constraint of two specific quantities. One is the mass density of the Universe at the current epoch. In its simplest guise, this means determining the total mass of all kinds of stuff that has a finite mass and does not move at the speed of light. This means stars, planets, and dust, but it also means any component of the mysterious dark matter that consists of particles, no matter how exotic. The other quantity to be constrained (and ultimately measured) is the value of what is called the vacuum density. Recall that even a vacuum has an energy associated with it. This energy underlies the emission of Hawking radiation from black holes. The vacuum may have even more subtle properties that would only be manifested when its effects are determined on the scale of the whole Universe.

There is a story behind this vacuum energy. The vacuum energy is, in principle, related to the quantum properties of the vacuum, but it arises in Einstein’s theory of gravity where it is called the Cosmological Constant. Astronomers who write the history of this subject tend to quote Einstein himself in this regard with great glee. Einstein called the Cosmological Constant "the greatest blunder of my life." The historian’s glee and Einstein’s self-criticism are probably unfair. The cosmological constant emerges from the mathematics of Einstein in a perfectly natural way (it appears as a constant of integration, for those who know calculus). It is not a question of whether it exists in this mathematical sense, but whether it is zero or not and whatever its value, including zero, what is the physics that determines that value?

The reason Einstein regarded his treatment of the Cosmological Constant a "blunder" is that the first mathematical models for the Universe showed that the Universe would expand. Einstein’s intuition told him that the Universe could not possibly do such a radical thing. To render the solution static, Einstein went back to the equations and realized that he had implicitly set the value of the Cosmological Constant to zero. If he assigned it just the right non-zero value, then the Cosmological Constant could serve as an extra source of gravity and balance the tendency of the Universe to expand. Shortly afterward, Hubble proved that the Universe is expanding. It appeared to Einstein that the Cosmological Constant was unnecessary, a blunder.

Einstein may have blundered in guessing the Universe was static, and hence in the value to which he set the Cosmological Constant, but he did not blunder in the introduction of the idea. In the long run, it is the latter that is more important, and another tribute to the power of Einstein’s theory. The blunder was much less than it is often made out to be. We now see that even the issue of whether the Cosmological Constant might be exactly zero is not a trivial one, but one that involves some of the deepest thinking about the Universe. More than that, there are hints that the Cosmological Constant is not zero, and that definitely raises profound issues of physics and cosmology.

Using supernovae to determine the other fundamental parameters of the Universe has been a dream for decades. Many people have worked for a long time to bring it to pass. One of the pioneers, Stirling Colgate of the Los Alamos National Laboratory, estimated that to get the job done when he started working on an automated supernova search telescope in the early 1970’s he would have had to invent seven or eight brand new technologies. These included digital control of the telescope and its instrumentation, electronic detectors to replace photographic plates (Colgate called all this "dig-as" for digital astronomy; the tide of the digital revolution has fully enveloped astronomy by now, but the term never caught on), thin light-weight mirrors, time-sharing computers necessary for many people to work cooperatively on the complex computer code required to control the telescope and scan images, and cheap microwave links to allow remote control of the telescope from a distant site. The telephone company wanted $3,000,000 for a microwave link from his telescope to the headquarters in Socorro, New Mexico. Colgate had only $3000 for the job. He invented a simple method of error checking and installed the link with the funds and equipment he had.

In just the last few years, the technical capability, the development of critical techniques, and willingness to devote a great deal of hard work have come together to bring this dream to fruition if not in quite the fully automated way Stirling Colgate envisioned. A key development has been the construction of large new telescopes and the special electronic detectors to record faint images over relatively large patches of the sky. Another was the launch, repair, and updating of the Hubble Space Telescope.

A team of astronomers at the Lawrence Berkeley Lab of the University of California, now headed by Saul Perlmutter, pioneered the breakthrough in technique. One of the inhibitions of research on supernovae is that their eruption is always a surprise. This means astronomers have to scramble to get data when an explosion occurs. Telescopes are often in the wrong configuration with the wrong instrumentation, the Moon is too bright to see the faint supernova light, or the weather is poor. The result is that we still do not get adequate information on most supernovae.

The Berkeley team realized that in certain circumstances they could discover supernovae "on cue." They could then schedule procedures in advance to follow them up. These techniques work in precisely the context where one can use the resulting discoveries to do cosmology with supernovae. The trick is that if one looks out to very large distances, a given image obtained with a telescope spans a huge volume containing a huge number of galaxies. It is impossible to predict which of the many galaxies will produce a supernova, but if enough galaxies are in the image, one can be confident that some supernovae will erupt. It turns out that one does not even have to know which specific galaxies are there in advance. If one looks distant enough, there will always be plenty of galaxies and plenty of supernovae. The distances involved, billions of light years, are also just the distances astronomers needed to probe to learn about cosmology.

More particularly, the technique developed by the Berkeley team is to schedule time on a large telescope when the Moon is not up and the sky is dark. They obtain a first image of the sky. They then return and take another image of the same patch of sky two or three weeks later after the Moon has passed through its bright phase and is no longer a problem. They compare the second image to the first and look for any new lights in the faint images. This is not trivial since both the galaxies and the supernovae are very faint. Many person-decades have been invested in the computer codes that can automate this process and detect and eliminate flashes of man-made light, cosmic rays that strike the detector, asteroids, and other things that are just a nuisance for this project.

Nothing can be done about bum weather, but these procedures have brought the other factors under control. In addition to the Berkeley group, another group has sprung up in competition led by Brian Schmidt of Mt. Stromlo Observatory near Canberra and comprising astronomers in Chile, Harvard and elsewhere. The results have been striking. The two groups of astronomers can guarantee the discovery of roughly a dozen very distant supernovae each time they return to take the second image. Since they know far in advance when they will take the second image, they can coordinate the prior scheduling of other telescopes. In this way, they are prepared to get critical spectral and photometric information as soon as they determine the precise location of the new discoveries. Rapid global communication, including the Internet, also plays a key role here. Both teams have also made use of the Hubble Space Telescope to closely examine the host galaxies after the supernovae have faded. This is a critical step, because one must subtract off the light of the host galaxy to get a pure signal from the supernova. Determining the light of the galaxy alone can be done efficiently after the supernova has faded, but not when the supernova first goes off and the light is a complex admixture of supernova and galaxy emission. This technique requires patience. Several months have to pass before the supernova has faded sufficiently, and many more months are required for careful calibration and analysis. Using these techniques, the number of supernovae discovered per year has shot up to around 100, most of them at distances that span a good fraction of the observable Universe.

The results of these efforts are just now becoming known. Taken at face value, they are very surprising. Recall from Section 2 that for a given intrinsic luminosity, the apparent brightness of a supernova declines as the inverse of the distance squared. This result, like the ratio of the circumference to the radius of a circle and the sum of the interior angles of a triangle, depends on the curvature of the underlying space. The power of the method of using supernovae is that they can, in principle, give such precise measurements of the distance at such great distances that the effects of the curvature of the space can be gleaned.

As mentioned above, the amount of mass of all kinds in the Universe affects the curvature of the Universe and tends to slow down the expansion because of the mutual self-gravity of all the mass-energy. If the Universe is slowing down, then it was expanding more rapidly in the past. This means that when we look at supernovae long, long ago and far, far away, with a given Doppler red shift, they will be a little closer and a little brighter than if the Universe had just been coasting at a constant speed, as shown in Figure 11.3. The Universe will also be younger than one would estimate from a given value of the Hubble Constant and the assumption that the Universe had always expanded at the current rate.

FIGURE 11.3: The size of the Universe as measured by the distance and Doppler shift of distant supernovae as a function of the age of the Universe. The three lines represent, schematically, the behavior of a closed Universe that is destined to recollapse, a flat Universe that will slowly coast to a halt in infinite time, and an accelerating Universe. The lines all have the same slope at the epoch marked "now." The slope of the lines at that point gives the Hubble Constant. The beginning of the lines represent the origin of the Big Bang for each case. For a given slope of the lines now, the closed Universe gives the shortest time since the Big Bang, and the accelerating Universe gives the longest.

That is all there is to it if the value of the Cosmological Constant is zero. If the Cosmological Constant is not zero, then the effect depends on whether the Cosmological Constant is positive or negative. If it were negative, the energy of the vacuum would add to the gravity of the matter and slow the expansion even more. If the Cosmological Constant were positive, the vacuum energy has the effect of a repulsive force causing the Universe to fly apart ever faster as it ages. That sounds like a strange effect, but it is possible within the framework of Einstein’s theory and another measure of why the introduction of the Cosmological Constant was not a blunder, but a very fascinating step. The mass density in the Universe must be positive, but the value of the Cosmological Constant could be positive or negative or zero and must be determined by observation or theory. If the Cosmological Constant were positive it would act in the opposite way to the mass density. A positive Cosmological Constant would tend to make the Universe accelerate rather than decelerate, as shown in Figure 11.3. This means that a supernova at a given red shift will be a little further away and a little dimmer than if the Universe had expanded at a constant rate. Likewise, the Universe would be a little older than one would estimate for a given Hubble Constant and the assumption of a constant rate of expansion.

Because the effect of the positive mass density and of a positive Cosmological Constant work in opposite directions to determine the dynamics of the Universe, the measurement of distances to supernovae tends to constrain the difference between the two effects. Using supernovae alone, the effects can not be easily separated. Careful measurement of the apparent brightness and red shift of Type Ia supernovae of a given rate of decline and hence intrinsic brightness can, however, constrain the values of the mass density and the Cosmological Constant. From those constraints and a knowledge of the Hubble Constant, the curvature of space and the rate of change of the speed of expansion of the Universe can also be estimated.

The current work on distant supernovae has given two surprises. One is that there does not seem to be enough matter to close the Universe and give it a finite lifetime. This suggests the Universe will expand forever. There are other astronomical techniques that are giving the same result. They all need to be further refined and considered, but astronomers are taking the possibility seriously. There is a more surprising result from the supernova work. Compared to the local sample of supernovae on which the calibration is done and compared to a Universe for which the Cosmological Constant is zero, the distant supernovae seem to be a bit too dim. If this is caused purely by cosmological effects, then the implication is that the supernovae are a bit further away for a given red shift. This effect, in turn, can only be explained by a finite and positive Cosmological Constant. The implication is that not only is the Universe not closed and finite, but it is accelerating, not decelerating! This is a striking and unexpected result.

A positive Cosmological Constant raises profound questions about what the nature of the vacuum must be that it contains a quantum property that acts as a repulsive force. The most popular models of the Big Bang now are so-called "inflationary" models. In this picture, when the Universe was first born it had a vacuum energy that did act as a repulsive force, an anti-gravity, that caused a piece of the Universe to rapidly expand to form the Universe we see today. According to the theory, this energy of the vacuum should have decayed away to zero by now. If the vacuum still has some of this repulsive energy, new theories of the vacuum will have to be developed.

These results are still new as this book is being written and need to be carefully considered. One of the critical issues is whether it is not the properties of the Universe that were different in the past, but the properties of the supernovae themselves. We know that not all Type Ia supernovae are alike and that their properties tend to correlate with the age of the stellar systems and host galaxies from which they emerge. The teams analyzing the supernovae have made some allowances for this. The question is whether the allowances are adequate. The current techniques to determine the brightness of the distant supernovae use the "calibrated candle" methods. An important task will be to check this result with the "understood candle" method. In addition, both of these methods currently use only rather gross properties of the supernovae, how much light is emitted at a given time, to determine the properties of the event. Much more information about subtle differences is contained in the details of the evolution of the spectrum of each event. This information has not yet been effectively tapped, but will be in the next few years. These early results of the distant supernova searchers are very stimulating. The implied results of low mass density and finite, positive Cosmological Constant may even be correct, but it will probably take several years of hard work by many people before we know that with confidence.

Other techniques will also come into play in the future to determine the state of the Universe. The Cosmic Background Explorer (COBE) satellite launched in 1989 revealed that the background radiation is of an exceedingly well-defined temperature, as expected. COBE also revealed faint irregularities in the intensity of the radiation from different parts of the sky. These regularities were also expected and even inevitable, given our understanding of the Big Bang. The Big Bang grew out of a singularity. That singularity must have been subject to quantum fluctuations in its properties that would be imposed on the expansion of the Universe and hence on the density and temperature of the matter in the Universe. Detection of these irregularities was another major vindication of the Big Bang picture.

New satellites to measure the properties of the cosmic background radiation will be launched early in the next century. Careful measurement of the fluctuations in the background radiation has great promise to further constrain the matter density and the Cosmological Constant. These techniques will be a critical complement to the research based on supernovae. For this effect, the mass density and the Cosmological Constant tend to work in concert. The larger the mass density, the stronger the gravity and the faster the fluctuations will tend to grow. On the other hand, if there is a finite and positive Cosmological Constant, so the Universe tends to accelerate, then the Universe will be a little older than it otherwise would be, other things being the same, and this gives the fluctuations more time to grow, again making them larger. The result is that the measurement of the fluctuations in the cosmic background will tend to measure the sum of the mass density and the effect of the Cosmological Constant, whereas the supernova technique measures the difference between these quantities. Neither technique by itself is apt to give the full picture. If, however, we have independent measures of both the sum and the difference of the mass density and the Cosmological Constant, then, in an algebraic sense, we can solve for both unknowns. At this writing, the exercise of joining the data from the supernovae with the data from the background radiation is consistent with the Universe being exactly flat, as the inflation theories demand, with about one-third of the total density being matter (mostly "dark" matter), and two-thirds being due to the vacuum energy associated with the Cosmological Constant.

F. The Past in our Future: The Dark Ages

Looking to the future brings yet another exciting possibility. After the epoch when the Universe was a million years old, the cosmic radiation streamed freely. The matter cooled and became dark. During the subsequent eons of expansion, the matter agglomerated into lumps that became galaxies. At some point, the gas in the lumps condensed and heated and started the first production of stars. The long interval between the release of the cosmic background radiation and the lighting up of the first stars has come to be called the "Dark Ages." After a long period with no light, stars winked on and the Universe started to take the form we recognize around us now. The processes involved in forming the first stars and galaxies, the emergence from the Dark Ages, is one of the frontiers of modern astronomy. It can be probed by the new generation of telescopes in the 8 to 10 meter class. The end of the Dark Ages will be the prime target of the New Generation Space Telescope currently under design by NASA.

Some of those first stars to form will be massive. They will evolve, collapse, and explode in just the way described in Chapter 6. When they do, their host galaxies will still be embryonic, small and dim. There is a chance that when astronomers peer back to the beginning of the end of the Dark Ages what they see will be supernovae, the brightest beacons in the young Universe.

The first supernovae to arise should be from massive, short-lived stars. They should be predominantly Type II supernovae, although there could also be an admixture of Type Ib and Type Ic supernovae. The Type II supernovae might all resemble SN 1987A by exploding as blue supergiants. As explained in Chapter 7, we do not fully understand why SN 1987A was a blue rather than a red supergiant when it exploded. Theoretical studies have shown, however, that when the amount of heavy elements in the atmosphere of an evolving massive star is low, the hydrogen envelope is likely to remain relatively compact so the star will look hot and blue rather than expanding so that the star will look cool and red. In the very young Universe at the end of the Dark Ages, there will not have been much time to make heavy elements. Whatever caused SN 1987A to be a blue supergiant, the paucity of heavy elements in the young Universe may cause all the exploding stars to be blue supergiants, even if they retain their hydrogen envelopes against the ravages of winds and binary companions.

If the first supernovae at the end of the dark ages explode in blue supergiants, the resulting explosions, like SN 1987A, may be relatively dim and somewhat harder to see. As the Universe ages and more heavy elements collect in the interstellar gas from which new stars are born, then at some point, massive stars may begin to evolve into fully formed red supergiants before they die. They will then explode as what we consider to be "normal" Type II supernovae. With the full power of new telescopes to scan from the present epoch back to the end of the Dark Ages, we should be able to see that epoch when the normal Type II supernovae turn on.

This discussion has omitted Type Ia supernovae. That is because we think they have a "fuse" that must burn before they explode. As discussed in Chapter 6, we do not understand the binary evolution that leads to the explosion of a white dwarf as a Type Ia supernova. All the indications are, however, that considerable time must pass before these binary processes, perhaps the evolution of the smaller mass companion, perhaps the decay of orbits through emission of gravitational radiation, lead to the explosion. That Type Ia supernovae have a long fuse compared to Type II means that when supernovae begin to explode at the end of the Dark Ages, they should all be due to the collapse of the cores of massive stars. There should be no thermonuclear explosions of white dwarfs and hence no Type Ia.

As the Universe ages and the binary evolution fuse burns, there will eventually be an epoch when the Type Ia supernovae begin to explode. Using the big new telescopes on the Earth and in space as time machines to probe these distant times, we should also be able to see this onset of Type Ia events. This would be a very exciting result because the time of the onset will give us critical new information on just what type of binary evolution constitutes the fuse. This, in turn, may finally teach us what binary evolution leads to Type Ia.

2. Gamma Ray Bursts

A. Yet Another Cosmic Mystery

There was a revolution in astronomy in the first few months of 1997. A major breakthrough occurred in one of the outstanding mysteries of modern astrophysics, the cosmic gamma-ray bursts. This story began in the 1960’s. The U.S. launched a series of satellites that orbited the Earth at great distance, halfway to the Moon. They were called the Vela series and they were designed to detect gamma rays. If it strikes you that you have never heard of these satellites and that there must be something special about them to be so far from Earth, you are on the right track. They were not designed for astronomy, but to detect treaty-violating terrestrial nuclear bomb tests. None of us non-classified folks know whether they ever served that purpose, but they did detect a series of outbursts of an extraterrestrial nature. The scientists at Los Alamos required some time before they were able to convince themselves that the gamma-ray signals were from outer space. The discovery was finally announced by Klebesadel, Strong, and Olson in a paper in the Astrophysical Journal in 1973. This paper created a new scientific industry.

The bursts of gamma rays from beyond the Earth were seen at irregular intervals. These bursts lasted for 10 to 30 seconds and showed variations on times as short as a thousandth of a second. Subsequent investigations showed that the gamma-ray bursts were primarily a gamma-ray phenomenon, with relatively little energy in the X-ray band, unlike other sources of gamma rays that emit abundantly at lower energies as well. That the dominant emission mode is gamma-rays means that a high energy is involved. Gamma-ray bursts probably require high gravity and motion at nearly the speed of light.

The quest for an explanation of gamma-ray bursts was long handicapped by a lack of direct

knowledge of the distance to the bursts. A debate raged as to whether they are in the Galaxy or at the furthest reaches of the Universe. This debate was brought into sharp focus by the immensely successful Burst And Transient Source Experiment (BATSE) on the Compton Gamma

Ray Observatory. The Compton Gamma Ray Observatory was launched in 1990 as one of the series of Great Observatories planned by NASA. The Hubble Observatory was the first. Two others, the Advanced X-ray Astronomy Facility and the Space Infrared Telescope Facility have been downsized, de-scoped and delayed for over a decade. Both AXAF and SIRTF should fly before the end of the century. BATSE has had many roles, but a principle component of its design was to search for gamma-ray bursts. The BATSE experiment has discovered over 1000 new gamma-ray bursts.

The surprising result was that the sources are, to great accuracy, distributed uniformly on the sky. There is no statistical evidence for any tendency to lie toward the plane of the disk of our Galaxy or toward the Galactic center. This contradicts any model in which the sources are distributed throughout the Galaxy and viewed from the off set position of the Earth, 25,000 light years from the Galactic center. This result fueled increasing conviction that the sources of the gamma-ray bursts were in galaxies at cosmological distances since the distant galaxies are naturally distributed uniformly on the sky, on average. In addition, fainter sources are more abundant. The precise number of faint sources shows a pattern that is close to what one would expect if the bursts constituted a gamma-ray "standard candle" viewed in ever larger volumes of space in an expanding Universe. There might, however, be other explanations for this pattern. There is no particular reason to think that gamma-ray bursts are a standard gamma-ray "candle."

The problem is that if the gamma-ray bursts are at cosmological distances, the intrinsic source of energy must be huge, comparable to or exceeding that of a supernova, but radiated essentially entirely in gamma-rays. The energy requirements are more moderate if the energy is directed in a beam, but then the number of sources must be correspondingly higher. Although some models of colliding neutron stars or neutron stars colliding with black holes, or black holes consuming magnetized accretion disks produce such energies, such models tend to strain credibility. Everything about the cosmic gamma-ray bursts strains credibility, yet there they are.

One of the clearly defined problems in the study of gamma-ray bursts was the complete lack of counterpart events at other wavelengths, especially optical wavelengths. Without optical counterparts, the full weight of astronomical lore, much of it derived from optical astronomy, could not be brought to bear on the issue. The problem was that the gamma-ray detectors could not provide sufficiently good locations. It is a difficult technical feat to bring gamma rays to focus. The gamma-ray sky has typically been "fuzzy," a situation somewhat analogous to a nearsighted person looking around with their glasses off. A given gamma-ray burst could be said to be "over there," but "there" could not be precisely defined. The uncertainties in position are typically several to 10’s of degrees in radius (the full Moon subtends about one-half degree in angular diameter). In an area of the sky of that size there can be thousands of stars. Finding the point of light that corresponds to a given 10-second long gamma-ray burst has been like seeking the proverbial needle in a haystack, a needle that was likely to vanish if you did not find it in less than a minute.

The nature of these events has puzzled astrophysicists for nearly 30 years. Without the fetters of any relation to classical astronomy, theorists have had a field day trying to account for the observations. The requirements for a theory in these circumstances are that it account for the observations and be self-consistent. Plausibility is not necessarily a constraint, since gamma-ray bursts represent a new and unprecedented phenomenon. At a meeting shortly after their discovery, the theorist who was giving the review talk on gamma-ray bursts announced that it was easier to give a list of the people who had not presented a theory of gamma-ray bursts than it was to give a list of those who had. He showed a slide consisting of one name, another prominent theorist who, for whatever reason, had not jumped on the gamma-ray burst bandwagon.

Theories have ranged from black hole collapse to "relativistic bb's." The latter were supposed to be little grains of dust accelerated to near the speed of light and then arriving at the solar system to crash energetically into the solar wind. Remember all the billion pulsars that have died in the Galaxy? One of the first theories, and one that generated more than a few chuckles, postulated that gamma-ray bursts were generated by comets falling onto those neutron stars. One of the little-known but supportive ideas of this hypothesis is that clouds of comets may very well spread nearly from one star to another. Space may be filled with comets, and the chance that one of them would occasionally fall onto one of those billions of neutron stars is not so low.

The argument that swayed some people into taking this idea more seriously is the problem of generating gamma rays at all with a neutron star. The problem is related to the Eddington limit. If energy is released on the surface of a neutron star, the material expands and cools in response to the radiation pressure. Under normal circumstances such matter can get hot enough to emit X-rays, as we have seen in Chapter 8, but not hot enough to emit the more energetic gamma rays. The importance of the impact picture is that the material arrives in a lump and is compressed much more than would be either a dribble of gas or material just sitting on the surface. The effect might be enhanced if the infalling matter were a rock, so asteroids have been considered as well as comets. After a hiatus of a number of years, a similar idea is still around, as will be outlined below. This shows the benefit of allowing the imagination of the theorists to run the bounds of the known data.

What was really needed were more data so that theory and observation could march hand in hand in some fruitful direction.

B. The Revolution

All this changed with the launch of a Dutch-Italian X-ray satellite, BeppoSAX on April 30, 1996. This wonderful name derives from the nickname of a pioneering Italian physicist and X-ray astronomer, Giuseppe Occhialini, known as Beppo to friends and colleagues with the appendage for X-ray satellite in Italian, Satellite per Astronomia a Raggi X. BeppoSAX was designed to look everywhere on the sky for the weaker X-ray signal that characterizes gamma-ray bursts and to give a first coarse location, more accurate than BATSE. The key innovation for BeppoSAX was a second instrument that can be brought to focus by quickly slewing the satellite in an attempt to rapidly find the X-ray flare from the gamma-ray burst and to provide a much more accurate location, with an uncertainty of a few minutes of arc. At that point, ground-based optical telescopes can be brought to bear to search the much smaller location to see if there is any optical component. All this was a bit of a gamble. If the whole gamma-ray burst phenomenon in lower energy X-rays and in the optical faded in the 10’s of seconds that characterized the gamma-ray bursts themselves, then there would be no time to slew the satellite, a process that would take at least hours, never mind time to obtain optical images, a process that might take a day (or night) even in the best of circumstances.

Another chapter of this story is worth telling if only to recognize the great effort and ingenuity that goes into the scientific enterprise that sometimes fails to pay off. At a meeting on gamma-ray bursts in Santa Cruz in the 1970’s the attendees recognized that studies of gamma-ray bursts were stymied by the lack of observations at other wavelengths. A project was born to design a satellite that would contain a gamma-ray detector, but also ultraviolet and optical detectors to look in the same direction and hence to get simultaneous information on the burst at other wavelengths. The project was named HETE for High Energy Transient Explorer and the arduous process began of design, winning NASA competitions to build and launch, and suffering the inevitable delays. HETE was finally scheduled to launch on November 4, 1996, a date that would have put it in competition with BeppoSAX. The Pegasus rocket that carried HETE failed and the satellite was destroyed. That opened the way for BeppoSAX. To their credit, the HETE team regrouped, took the plans and spare parts and built a new satellite. It should launch in 1999 and will be a valuable tool for the study of gamma-ray bursts.

BeppoSAX scored its coup on February 28, 1997, when it localized a burst sufficiently well that an optical follow-up was feasible. The result was the discovery of the first optical counterpart by a team led by Dutch astronomer Jan Van Paradijs. The fashion has been to label gamma-ray bursts by the year and day that they were discovered. So far, no two have been discovered on the same day to mess up this scheme. The breakthrough gamma-ray burst was thus named GRB 970228. Two months later, in early May, BeppoSAX found another event, GRB 970508, enabling another optical identification. In this case, absorption lines of matter in front of this source prove that the source is at a cosmological distance, of order a billion light years or greater. In December of 1997, yet another optical counterpart was discovered associated with GRB 971214. After the gamma-ray burst faded, a faint galaxy was revealed. The red shift of this galaxy was immense, with the wavelength of the detected light shifted by more than a factor of 3 from its natural wavelength. This galaxy is estimated to be 12 billion light years away. If GRB 971214 is radiating equally into all directions and hence following the basic inverse-square law for apparent brightness, then estimating the distance from the red shift (and adopting specific values of the cosmological parameters) implies that the energy of this source is fantastically large. More energy is required than the entire collapse and neutrino energy of a supernova and more then even the most exotic theories of colliding neutron stars and black holes can support.

This story is still rapidly unfolding, and even GRB 971214 is not the record. That belongs to the first burst localized by BeppoSAX in 1999, GRB 990123. This burst brings in yet another interesting chapter in the saga. Many people realized that if an optical counterpart were ever to be seen, then an especially rapid response was needed. A special Email notice system was set up run by Scott Barthelmy and his colleagues at the NASA Goddard Space Flight Center in Maryland. Even more extreme, some people began to wear beepers that were triggered electronically by a signal from a satellite, BATSE or BeppoSAX, so that they were get buzzed the instant (allowing for the finite travel time of light and relay switches) a gamma-ray burst was detected. One of the things that this rapid response allowed was communication with automatically-controlled robotic telescopes that would very quickly swivel to look for an optical counterpart, perhaps in the time frame of the original gamma-ray burst. This was the mission of ROTSE, the Robotic Optical Transient Search Experiment.

ROTSE is a small telescope situated at the Los Alamos National Laboratory. It was designed and operated by astronomers at the University of Michigan, Los Alamos, and the Lawrence Livermore National Laboratory. ROTSE was constructed to receive signals directly from the satellites that detect gamma-ray bursts to rapidly swivel and look at the location of a gamma-ray burst. ROTSE is not very sensitive as telescopes go since it has only four wide-angle camera lenses, but it can see a fairly large portion of the sky at one time to look for variable sources. The advantage is that it is quick! Quickness does not count if the weather does not cooperate or if the discovered gamma-ray burst is only visible from the Southern Hemisphere or if it is "up" in the North during daylight hours. This was the tale for the first number of BeppoSAX bursts.

Finally, on January 23rd of 1999, everything came together and ROTSE had its first clean shot at a gamma-ray burst. On its first clean opportunity, ROTSE detected the immediate optical counterpart of GRB 990123. The results were dramatic. ROTSE saw a burst rise in about 10 seconds to 9th magnitude and then fade over the next minute or so. This peak apparent brightness was only about a factor of ten dimmer than can be seen with the naked eye! Associated work on this gamma-ray bursts revealed it to be at yet another immense distance. This makes GRB 990123 the intrinsically brightest optical event ever recorded in scientific history. Ho hum, another record for gamma-ray bursts. Actually, there is nothing to be blasé about here. If radiated uniformly in all directions, the implied peak optical luminosity of GRB 990123 was equivalent to 10 million supernovae or ten thousand very bright quasars. This optical burst did not last long, but its intensity was very impressive.

Most of the energy emitted by GRB 990123 was in the gamma-ray range. Here again, GRB 990123 set a record. The detected gamma-ray intensity was among the strongest ever seen at the Earth. At the distance observed, the total energy in gamma-rays was 10 times higher than the previous record-setters like GRB 971214. If this gamma-ray energy poured out equally in all directions, the energy involved was equivalent to the complete annihilation of two solar masses of matter! One runs out of exclamation points.

These optical counterparts of the cosmic gamma-ray bursts have thus revolutionized the field and proven the power of focusing optical astronomy on this decades-old problem. They have opened a new era in the study of gamma-ray bursts that promises not only rapid progress in understanding the bursts themselves, but also their use to explore the nature of the Universe at great distances.

The emission witnessed in the X-rays by BeppoSAX, in the optical by ground-based telescopes, and in the radio by radio telescopes was discovered to last much longer than the original gamma-ray burst. Rather than 10’s of seconds, the X-rays last for days and the optical and radio can stay above limits of detectability for weeks or months. This delayed emission of energy has been termed the afterglow of the gamma-ray burst. The general interpretation is that the process that energizes the event, whatever that process is, sends a powerful explosion out into the interstellar gas surrounding the event. The explosion generates a strong shock wave that moves at very nearly the speed of light. The interaction of this shock wave with the interstellar gas can produce gamma rays, X-rays, optical emission, and radio emission in appropriate circumstances. The general process leading to this afterglow is called a relativistic blast wave. Models based on this process have been successful in accounting for many of the observations of the afterglow, including the spectrum of the radiation and the rate of decay that tends to drop off as one over the time since the original gamma-ray burst. If you wait twice as long, the glow is half as bright.

Many new pieces to the puzzle of gamma-ray bursts have been put on the table since the first BeppoSAX discovery. Many fundamental mysteries still remain. Despite the new information and the growing understanding of the afterglow phase, we still do not know the basic mechanism that releases the energy and converts so much of the energy into gamma-rays. Some of the BeppoSAX events show optical afterglows, but others apparently do not. Most of the afterglows decay so that the power fades inversely with time, but some decay more rapidly. In a real sense, the field is just beginning and is likely to explode with activity.

C. The Shape of Things

One of the issues to be confronted in the study of gamma-ray bursts is the manner in which the energy is released into the surroundings. There are a number of tightly intertwined issues here. Theoretical models of relativistic blast waves and the afterglow demand that a shock wave moves out from the source at speeds very close to the speed of light. To do this, the flow of energy must carry along with it very few ordinary particles, protons, or more generally baryons. Too many of these particles of ordinary matter would slow the shock wave down so that it could not propagate with the deduced speeds. That is one thing that must distinguish an ordinary supernova and a gamma-ray burst. Both events have roughly the same amount of energy, but supernovae put their energy into moving a lot of ordinary matter at high, but not relativistic speeds. Gamma-ray bursts must put as much or more energy into a very small amount of mass.

Given the expansion at nearly the speed of light, a number of issues arise that come from Einstein’s Special Theory of Relativity. When motion with respect to an observer is high, lengths are foreshortened and times are constricted. An event that takes a month to play out in the rest frame of the gamma-ray burst may take only hours or days as observed at Earth. In particular, it may take a month for the relativistic shock wave to expand out from the source of energy, pile up mass in the interstellar medium, and slow to ordinary speeds. An observer on Earth would see all this playing out in a day or so. Turned around, when we see a gamma-ray burst afterglow fading over a few days, it might have taken a month in a far galaxy.

Another interesting effect is that if a source of radiation moves toward an observer at a high speed, the radiation is thrown in the direction of the observer. This "beaming" can make the radiation seem brighter than it would otherwise be. In addition, if the source of energy is moving toward the observer there is a very large blue shift, a "boost" of the energy of each photon that is detected. This can again make the source look brighter.

Such issues arise in trying to determine how bright a given gamma-ray burst really is and how much energy it emits. Even if the energy from a gamma-ray burst is emitted equally in all directions, it will be beamed and boosted and look brighter for a shorter time to an observer standing still on the Earth compared to an observer at the same distance who moved with the velocity of the shock. Trying to figure out how bright a given gamma-ray burst "really" is in its own rest frame is a rather tricky business that requires an understanding of just how the boosting and beaming is working.

One can get a measure of the total energy emitted in the radiation independent of the beaming and boosting if the energy is emitted equally in all directions. The procedure is to add up all the energy received at Earth over the course of the burst event. That energy might have been emitted over a different time span in the frame of the explosion, but all the energy is all the energy, and it must all go somewhere eventually. If one assumes it goes off equally in all directions and corrects for the fact that things look dimmer by the inverse square of the distance (plus perhaps some corrections for cosmological warping), then the total energy in radiation of the explosion can be determined. For the BeppoSAX events for which there is a measure of the red shift and hence the distance, the results are imposing, as mentioned earlier. For the event at the largest distance, 12 billion light years, the energy is more than the entire flow of neutrinos from a supernova, a huge amount of energy and for GRB 990123 it is 10 times the neutrino energy of a supernova.

There is an important caveat to this method of measuring energy. If the flow of energy does not come out equally in all directions, if it is collimated in some way, if it flows out in a jet, then less total energy is required for a given observed burst, just in proportion to the amount of collimation, as shown in Figure 11.4. If the energy flows only into 10 percent of all available directions, then a given energy received on Earth requires only 10 percent as much total energy at the source. If the energy flows in a jet filling only 1 percent of the area into which it expands, then the energy at the source is only 1 percent of that deduced from the assumption that equal energy goes in all directions.

FIGURE 11.4: (top) If the energy in a gamma-ray burst flows out equally in all directions, then it does not make any difference where the observer is. All observers at the same distance will see the same brightness and deduce the same energy. (bottom) If the energy is collimated into a jet, however, the observer (#1) who looks down the jet will see a much higher luminosity than the observer (#2) looking from the side. If observer #1 assumes that the energy is emitted equally in all directions he will deduce too large a total energy for the event.

This collimation effect is not a fantasy. It is almost the rule rather than the exception. We see collimated flows from the Sun, protostars, planetary nebulae, binary black holes, and quasars. If the energy of a gamma-ray burst comes out in a collimated relativistic blast wave in only certain directions, then one must be careful in making estimates of luminosities and energies.

An example of this phenomenon is the "blazars." Blazars are a certain sub-class of quasars that are especially bright and highly variable. The common interpretation is that in these objects we happen to be looking right down the nozzle of a jet of matter ejected at nearly the speed of light. By the accident of the Earth’s position in the beam, we see an especially bright source of radiation because of the beaming and boosting associated with the rapid motion toward us. We also see especially rapid time variability of the radiation that is thought to be associated with the shrinkage of time due to the relativistic motion. No one suggests that this energy is flowing out equally in all directions, thus requiring unprecedented amounts of energy, even for quasars. Rather it is assumed that if we happened to observe the same object from the side it would resemble an "ordinary" quasar. Understanding whether, how, and how much gamma-ray bursts are collimated is one of the key tasks facing the field.

D. Supernovae and Gamma-Ray Bursts

Despite the lack of understanding of the source of energy of gamma-ray bursts, they are widely thought to be at least vaguely related to supernovae. Many models require objects that are produced in gravitational collapse -- neutron stars or black holes -- to collide or merge or spin or swallow matter in some dramatic way. In the onrush of events that followed from the BeppoSAX discoveries another surprise made the relation of gamma-ray bursts and supernovae explicit.

On April 25, 1998, BeppoSAX discovered a gamma-ray burst, GRB 980425, of otherwise ordinary properties in terms of its apparent brightness, energy, and time scale. BeppoSAX then swung to bring its fine position sensor X-ray detector into position and detected a couple of X-ray sources, one of which diminished in time. A day later, optical astronomers caught up and found a strongly variable object. This object was not, however, the afterglow that one had quickly learned to expect. It was, rather, a supernova, one of rather strange properties. The supernova, SN 1998bw, was not exactly at the position of either of the two X-ray sources reported by BeppoSAX. This raised some question about the association of SN 1998bw with GRB 9980425. In the next few months, the BeppoSAX team recalibrated the positions of the X-ray sources they detected. The source that was at first observed to vary was determined to be much to far from SN 1998bw to be associated. The other source, at first thought to be constant, was shifted so that an association with SN 1998bw could not be ruled out. Then this source was discovered to be variable, if only slightly. This has left the issue of the association of SN 1998bw with the BeppoSAX X-ray sources somewhat befuddled. One must be wary of other sources of variable X-ray emission, such as active galactic nuclei that could accidentally fall near the supernova, but an association of one of the X-ray sources with SN 1998bw cannot be ruled out.

A few days after the detection of the SN 1998bw, radio astronomers found a very bright radio source. This radio source is precisely at the position of SN 1998bw, so there is no question of their association. Analysis of the radio data showed that the radio source is brighter than can be easily explained without expansion of a shock wave at nearly the speed of light. Independent of the gamma-ray burst, SN 1998bw clearly produced a relativistic blast wave. All this evidence taken together suggests that the SN 1998bw and the gamma-ray burst GRB 980425 are one and the same thing. The likelihood of finding both GRB 980425 and SN 1998bw in the same part of the sky in the brief interval of time when they erupted is very low, so many astronomers think the connection must be real. In particular, while gamma-ray astronomers tend to be leery of the association, supernova mavens have embraced it with full passion.

Observations of SN 1998bw and its host galaxy showed that it was at a distance of about 40 million parsecs or about 120 million light years. That is a great distance, but far less than the record setting 12 billion light years of GRB 971214. At 40 million parsecs, the total energy in the gamma-ray burst is deduced to be much less than that of the most distant gamma-ray bursts, by a factor of about 1 million. On the other hand, at the same distance SN 1998bw is exceptionally bright for a supernova. Both of these results are puzzles that must be assimilated in the ongoing attempt to understand gamma-ray bursts.

Although it is a step along an esthetically ugly path, one idea that emerges from this new event is that there are at least two kinds of gamma-ray bursts, one of very high energy seen at cosmological distances and one of lower energy seen relatively nearby. This is an uncomfortable hypothesis given that the gamma-ray properties of GRB 980425 were seemingly unexceptional. The similar nature of far away energetic and nearby lower energy gamma-ray bursts may arise because any physical events that can emit gamma-rays will have certain properties in common whether the total energy involved is high or low, but this remains to be shown.

SN 1998bw brings its own set of questions. The early spectra were unlike any other supernovae we have discussed, Type Ia, Ib, Ic or II. With hindsight, there were a few other supernovae, SN 1997ef is a conspicuous example, that did bear some resemblance to SN 1998bw, so there may be some precedent. As it evolved, SN 1998bw looked more and more like a Type Ic with no evidence for hydrogen or helium. It certainly did not look like either a Type II or a Type Ia.

The first models of the light curve and spectra have assumed that SN 1998bw resulted from core collapse and that enough radioactive nickel was produced to power the peak of the light curve. Since SN 1998bw was about as bright as a Type Ia (even though the spectrum is completely different), a comparable amount of nickel is required, about 0.7 solar masses. Basic spherically symmetric models can produce this amount of nickel in a core collapse explosion by shocking silicon layers, but they are extreme. Models that make this much nickel and that produce the observed light curve and spectra at some level of agreement (not perfect in the first models) require an exploding carbon/oxygen core of about 10 solar masses and an energy of expansion of the matter of more than 10 times that normally associated with supernovae. These models suggest that SN 1998bw was a "super" Type Ic, and the term hypernova has been adopted in some circles.

If this interpretation of the observations is right, then SN 1998bw is an exceptional event. It must involve processes unlike all the other supernovae discussed in this book. Speculations run toward the creation of a black hole in the explosion and subsequent accretion of matter to fuel the large energy release. Other possibilities are that the associated gamma-ray burst afterglow contributed some of the optical light, so less nickel is required or that the explosion is asymmetric so that the explosion looks brighter at some viewing angles than others. All these possibilities need to be explored.

E. The Possibilities

This two years of living dangerously after the first BeppoSAX discovery has left a large range of possibilities for the origin of gamma-ray bursts. They cannot all be true, but it will take a lot of work and ingenuity to show which ideas will survive.

The idea that gamma-ray bursts are not at cosmological distances, but relatively nearby still has supporters. People of this bent argue that the apparent association of gamma-ray bursts such as GRB 971214 with a very distant galaxy is an accidental alignment, the eternal bugaboo of astronomers who peer out into the huge volume of space. If you look far enough, so this argument goes, you are bound to see a galaxy, so what? Estimates of the probability are about one in a thousand, small, but not impossible.

An elaborate and inventive argument for nearby gamma-ray bursts comes from Stirling Colgate. Colgate envisions that some neutron stars are born in binary systems. The burst of neutrinos when the neutron star forms gives a jet out one side that selectively spins the neutron star counter to its original orbit and sends it rocketing out past its companion. The neutron star grabs a little matter from the star on the way. This matter, about 1 percent of the mass of the Sun, settles into a small disk around the neutron star. The neutron star then sails out of the galaxy, given sufficient time. The matter collected by the neutron star spreads out, some of it slipping inward, some moving outward. The inwardly moving matter falls on the neutron star rather quickly. Colgate argues this is the source of the soft gamma-ray repeaters discussed in Chapter 9, objects that give bursts of low energy gamma rays (or high energy X-rays, depending on your point of view).

A million years passes as the neutron star moves at high speed out into the far halo of the Galaxy. At this distance, Colgate argues, the apparent homogeneity of the gamma-ray bursts on the sky can be satisfied without requiring cosmological distances. The gamma-ray bursts arise when the disk matter that moved outward cools, loses its hydrogen and helium and collects in "rubble piles" in just the way planets form around stars. These rubble piles, about a million of them, scatter off one another and some, half, are scattered into orbits that take them near the neutron star, like the collision of comets with the Sun or comet Shoemaker-Levy with Jupiter (remember the original idea of comets falling on neutron stars?). Near the neutron star, the rubble pile is trapped, heated and compressed. It forms a hot, quantum-pressure supported thin disk. This disk is an excellent electrical conductor spinning in the magnetic field of the neutron star. Huge electrical fields build up, create gigantic electron-positron pair sparks and a burst of gamma rays. After about a million rubble piles have created a million gamma ray bursts, the rubble piles that can be gobbled up have been, and the process fades out. Colgate contends that this mimics the fading of the number of gamma-ray bursts with apparent brightness that has otherwise been argued as strong support for cosmological models. People who espouse such "local" galactic halo models for gamma-ray bursts have not yet assimilated the implications of GRB 980425 and SN 1998bw as of this writing.

The author and his colleague, Lifan Wang, have proposed an intermediate picture based on the nature of GRB 980425 and SN 1998bw. The idea is to see how far one could go with using only relatively ordinary supernovae to produce all the gamma-ray bursts. The argument is that all gravitational collapse events produce strong magnetic jets that punch out though the axes of the surrounding carbon/oxygen core. In ordinary Type II supernovae, the outer hydrogen layers would stop these jets. In Type Ic or Type Ib, the jet could escape into interstellar space making the gamma-ray burst. This is shown schematically in Figure 11.5.

FIGURE 11.5: A possible schematic model for a gamma-ray burst in a Type Ic supernovae. The evolution of a massive star that has lost its envelope proceeds to the formation of an iron core that in turn collapses to form a neutron star. The formation of the rotating, magnetized neutron star, perhaps a highly magnetized magnetar (see Chapter 8) sends out jets of material that punch holes in the core as the outer layers of matter hover, waiting to collapse or explode. The newly formed pulsar then fills the inner cavity where the iron core had been with intense electromagnetic radiation. This radiation floods out the wounds in the core punched by the earlier jets and forms the gamma-ray burst. Such a highly collimated gamma-ray burst could only be seen at special angles looking right down the beam. This configuration could, in principle, yield a gamma-ray burst that could be seen across the Universe, but requires only a portion of the supernova energy. A small fraction of the gamma-rays could be radiated sideways, making the weaker gamma-ray burst associated with SN 1998bw. The rotational energy of the pulsar might enhance the explosion of the supernova itself.

In this picture, there are two components to the gamma-ray emission, one that emerges more or less equally in all directions with the energy seen in GRB 980425/SN 1998bw, about 1 to 10 thousand times less than a standard supernova energy, and one component that is highly collimated in a relativistic jet containing perhaps 10 percent of the supernova energy. The lower energy component could be seen if the explosion occurred relatively nearby, at 100 million light years, but would not be detectable with current instruments if the same event were at truly cosmological distances. The other gamma-ray component emerges in the jet so that all the gamma-ray energy contained in it is collimated to flow in a narrow angle. In this way, only some fraction of the supernova energy is required to be channeled into gamma rays. By this argument, the huge energies deduced for the very distant gamma-ray bursts is an artifact of assuming that equal energy is emitted in all directions, rather than being confined to the direction of the jet, as in the "blazar" picture described above. To reduce the required energy from the amount deduced in an "all directions" picture to some fraction of a supernova energy, the jet must be tightly collimated. The area of its cross section must be only one part in a thousand of the area surrounding the burst source. This is about the amount of collimation seen in typical jets from active galaxies and deduced to exist in blazars, so it is not beyond the bounds of credibility. Whether it is produced in a real supernova is another story that will require intensive investigation.

If the jet moves at nearly the speed of light, the gamma rays will be blue shifted and beamed strongly in one direction. This component could, in principle, be seen at cosmological distances if the jet happens to be pointed right at the Earth. Most of the jets will not be pointed at the Earth, so this picture requires many more gamma-ray burst events that are not pointed at the Earth to account for the few that are. If the collimation is to one part in a thousand, then there must be one thousand jets not pointed at the Earth for every one that is. The required rate is roughly that for normal supernovae, approximately one per few hundred years per bright galaxy, giving a crude concordance to the argument.

There is some weak statistical evidence that all Type Ic supernovae and related events like SN 1993J are associated with gamma-ray bursts, but there are many loose ends to this picture. How would the jet form and would it emit gamma rays and the other products of the afterglow in the right way? What is the mechanism of the lower energy gamma-ray component? Why should the weaker gamma-ray components that emerge in all directions have the same observed average properties as the emission from a beamed relativistic jet? The lower energy component should play out in ordinary time, but the jet component should suffer a strong Einsteinian time compression. Why do they both last about a minute? Should every Type Ic show evidence for the jet that could long outlive the supernova? Could SN 1998bw possibly be related to other more normal Type Ic despite the evidence for very high ejecta masses and high explosion energies? There is evidence that "ordinary" Type Ic are not spherically symmetric, but produce explosions that are distorted in a significant way. This already suggests that the core collapse that produces Type Ic, and presumably all other collapse-driven supernovae, is strongly asymmetric, perhaps involving jets shooting out the rotation or magnetic axis. Could such events be brighter in one direction than another, accounting for the apparent excess luminosity of SN 1998bw? Is the evidence for non-spherical ejection in Type Ic related to the jets necessary to make gamma-ray bursts that can be seen across the Universe? If soft gamma-ray repeaters require magnetars (Chapter 8), neutron stars with superstong magnetic fields, what happens at their birth? Are magnetars born in binary star systems that spin up the star? Do they generate a jet when they form? Do they resemble Type Ic supernovae or SN 1998bw? These are all questions that need to be addressed.

Like Type Ic, SN 1998bw showed signs of asymmetry, evidence that the flow of ejected matter departs rather strongly from spherical symmetry. This evidence has been ignored by the first spherically symmetric "hypernova" models that require unprecedented amounts of energy to provide the supernova luminosity. Peter Höflich and the author have considered models that are distorted by a sufficient amount to account for the asymmetries in Type Ic supernovae and in SN 1998bw itself. Preliminary models show that if the ejecta are in the shape of a fat pancake they will be considerably brighter if viewed from the top of the pancake compared to the edge, by about a factor of two. These models have the potential, at least, of accounting for the observed optical properties of SN 1998bw with "normal" amounts of energy and ejected nickel mass. Whether such models, or the hypernova models for that matter, can account for the gamma-ray properties remains to be seen.

Then there is still a plethora of models that address the energy issue head on. Some of these involve colliding neutron stars. That process has plenty of energy, but maybe not enough for the most extreme events. Another principal issue is turning the energy into gamma rays and a relativistic blast wave that is not so overloaded with protons that it can not move rapidly enough to make the afterglow. One possibility is that the neutron stars do not collide directly, but interact through their strong magnetic fields. That way one can think about turning the pure magnetic energy into pure gamma-ray energy without getting the stuff of the neutron stars, those troublesome, slowing baryons, directly involved. Other models envision a ring of matter carrying a large magnetic field that accretes into a newly born, rapidly rotating black hole. If the accretion is fast enough and the field strong enough, large energies can be created. The energy can be scaled up with the mass of the black hole, but these models might also collimate the energy in a jet of some kind, so the total energy requirements would be muted. The open issues are whether the right kinds of black holes form with the right accretion and the origin of the needed magnetic fields. Yet another version of this picture invokes the collapse of supermassive stars of perhaps more than 10 thousand times the mass of the Sun. This process is speculated to occur in any case to form the giant black holes we believe reside in most large galaxies, if not all galaxies, and that give rise to quasars. Because the mass involved in the collapse can be very large, the energy released can also be very large, so that only a small fraction of the total needs to be converted to gamma rays.

All of these pictures have a certain basic plausibility about them, given that we think our Universe is full of magnetic neutron stars and black holes of a range in masses from those of stars to those of galaxies. The devil is in the details. Having accounted for the energy, the first major requirement, can any of these models really account for gamma-ray bursts with the observed properties? All these models designed to give very high energy gamma-ray bursts at cosmological distances must also now confront GRB 980425 and SN 1998bw. How is it that a newly formed accreting black hole in the young Universe produces a gamma-ray burst with the same average observed properties as a relatively nearby, much less energetic, odd supernova?

These are the conundra that make astrophysics so exciting. Gamma-ray bursts will continue to provide all the stimulation an astrophysicist could want for some time to come. As better understanding of the gamma-rays bursts comes, so will a better understanding of the Universe on both stellar and cosmological scales. The gamma-ray bursts give us yet another means look throughout the space and time of our visible Universe.

BATSE detects about one gamma-ray burst per day. If every one were like GRB 990123, there should be a bright optical flash for a minute or so once a day somewhere on the sky that would be easily visible with a decent pair of binoculars. A pair of binoculars allows you to see about one part in a thousand of the total sky. If you looked every night for three years running, you just might get lucky. Backyard cosmology.