J Donald Fernie. American Scientist. Volume 84, Issue 5. Sep/Oct 1996.
Readers may recall that a year ago I wrote a column celebrating the 75th anniversary of what has become famous in 20th-century astronomy as the Great Debate. The debaters were Harlow Shapley, at the time a staff member of the Mount Wilson Observatory of the Carnegie Institution of Washington, and Heber Curtis of the Lick Observatory, University of California. Their debate took place on April 20,1920 and concerned the question of whether spiral nebulae are associated with the Milky Way or are independent galaxies: “coequal empires or … colonies,” as Arthur Eddington put it in typical 1920s British style. The Great Debate, not surprisingly, came to no firm conclusion on the question, which was settled mainly by Edwin Hubble some five years later. Nevertheless, for reasons perhaps more human than scientific, the debate caught the popular imagination and has been a staple of elementary astronomy texts ever since.
It was the idea of Robert Nemiroff of NASA and George Mason University that the diamond jubilee of the Great Debate should be celebrated in 1995 by having another such debate, this time over a puzzle that, like its predecessor, has exercised the minds and tongues of astronomers for decades without a firm solution having been found. It is the puzzle of what and where are the sources of brief but violently energetic bursts of gamma rays that come at us every day from somewhere beyond the earth.
The contenders in this new debate (held on April 22, 1995 in the same hall as the first debate) were Donald Lamb of the University of Chicago and Bohdan Paczynski of Princeton University, with Sir Martin Rees of Cambridge University as moderator. Basically, it was Lamb’s position that the gamma-ray bursters (as the sources have come to be called) are neutron stars in a giant corona surrounding our galaxy, whereas Paczynski argued that the bursters, of unknown identity, must lie at cosmological distances some 10,000 times greater than that.
Like so many discoveries in astronomy, the discovery of the gamma-ray bursters was serendipitous. The story begins in the 1960s during the Cold War, when the two sides in the war anxiously monitored each other’s activities, particularly with regard to nuclear bomb tests. Fearing the Soviet Union might go so far as to carry out nuclear weapon tests on the far side of the moon, where the explosion would be invisible from earth, the United States put into orbit pairs of satellites that circled the earth opposite one another at a separation of about 250,000 kilometers. Between them, they could watch almost the whole sky and surface of the earth continuously. An actual nuclear explosion on the far side of the moon might still not be seen directly, but because the subsequent blast cloud of dust and debris would be radioactive and emit x rays and gamma rays well above the surface of the moon, the satellites were equipped with specially designed detectors of such wavelengths.
Bursters in Our Neighborhood
The years went by, and, of course, no nuclear explosions on the moon were ever detected. But the satellites did report occasional bursts of gamma rays coming from random directions at a rate of 10 or 20 a year. Timing the moment of arrival of a given burst at various satellites allowed scientists to detect not only the direction of the source but also whether it was nearby. It turned out that the sources must lie beyond the inner solar system; the major planets and sun were therefore ruled out.
The gamma-ray reports were not classified by the military and eventually came to the attention of astronomers, the first announcement of them being a paper in the Astrophysical Journal in 1973. Naturally, although the direction determinations were relatively poor, everyone wanted to know whether the bursts could be identified with one or another of the many strange objects that populate the astronomical zoo. The answer was no; as far as one could tell they come from no clearly identifiable type of object. An even greater surprise was that reports of gamma-ray bursts were not accompanied by reports of bursts at any other wavelength. So far as we know even today, the bursters, whatever they may be, emit bursts only at these most energetic frequencies and at no other, not even x rays.
The situation was a paradise for theorists, untrammeled by facts. Only two years after that first Astrophysical Journal paper, an exasperated reviewer remarked that there were already more theories than the total number of bursts reported to that time! And by now, a quarter-century later, more than 2,000 papers have been written about gamma-ray bursters. Yet, as Gerald Fishman, a veteran of the field, puts it, “They remain perhaps the least understood of all objects that have been observed in the Universe.”
The observational situation, however, has improved over the years and brought new constraints to the theorists. As more sensitive detectors have been placed on satellites, the rate of detection has gone up from a dozen or two a year to more like 300 a year, or nearly one a day. For years some of the best data came from an experiment running on the Pioneer Venus Orbiter (PVO), but in April of 1991 the Burst and Transient Source Experiment (BATSE) was put into orbit on the Compton Gamma-Ray Observatory, and this has almost revolutionized the subject.
With the advent of the PVO data it seemed clear that the bursters are randomly distributed on the sky, and that their observed intensities followed a so-called three-halves law. That is to say, the cumulative number N having intensities above some level P depended on P to the power of -1.5. This implied that the sources are homogeneously distributed in three-dimensional Euclidean space around us, suggesting that we are detecting objects that are relatively local, say within a few kiloparsecs. (A parsec is a unit of distance equal to 3.26 light-years, or about 3 x 10^sup 16^ meters.)
Theories based on these results included the possibility that the sources lie in the Oort Cloud of comets on the distant outskirts of the solar system. This spherical shell of perhaps 10^sup 11^ cometary nuclei starts at about 50,000 times the earth’s distance from the sun. How to produce the gammaray bursts there is a bit tricky, but comet-comet and comet-primordial-blackhole collisions have been posited. In any case, this theory seems to have fallen from favor. The cloud cannot be exactly spherical because of tidal effects of our galaxy, and the number density of comets almost certainly must vary with distance.
The most favored theory at the height of the PVO era was that the bursts somehow involve neutron stars on the order of a kiloparsec away in the thick disk of our galaxy A neutron star is the remnant core of a massive star that has exploded as a supernova. It is typically 10 kilometers in diameter, but with a mass several times that of the sun, so that the star is at nuclear densities, hence its name. Its magnetic field may range to 1012 gauss or more, and all in all its exotic properties make a good starting point for theoretical models of gamma-ray bursters.
Bursters and Neutron Stars
BATSE changed all that. Some 30 times more sensitive than the PVO detectors, those on BATSE were expected to finally reveal the anisotropy resulting from there being an increasing density of neutron stars towards the plane of the Milky Way. If BATSE could “see” the faint sources outside the disk it should find fewer there, and the sky maps would begin to show a higher density of points in the galactic plane. In fact, since we are located at a considerable distance from the galactic center, it was expected that the direction to the center would show an even heavier concentration of bursters on the sky maps.
To the astonishment of almost all burst watchers, BATSE did nothing of the kind. Certainly it detected many more weaker sources, but the isotropy was only enhanced, and (horror of horrors) the three-halves law broke down: For really weak sources N-P^sup -0.8^. This meant that either BATSE had penetrated to the edge of the burster distribution, or it was dealing with a distribution on a scale that required non-Euclidean spacethat is, a cosmological scale of perhaps a gigaparsec. Where were the theorists to turn?
Here we come back to the debate between Lamb and Paczynski. Lamb favored a theory that still invoked neutron stars as burster sources, but now located them in an immense halo or corona with an approximately 100-kiloparsec radius around our galaxy, whereas Paczynski much preferred the gigaparsec cosmological scale, even though there is currently no known satisfactory model for the sources themselves at such distances.
Neutron stars in general have very rapid rates of rotation, some measured in milliseconds per rotation. The intense magnetic field of such a star beams the non-thermal electromagnetic radiation it emits (particularly at radio wavelengths), and the rotation then sweeps this beam around as a lighthouse sweeps its beam of light around. If we happen to lie near the plane of the beam, radio telescopes can detect the train of pulsed radiation as it repeatedly sweeps over us. Such neutron stars are called pulsars. This, in fact, is almost the only way of detecting neutron stars.
Lamb based his preference mainly on two facts. First, it has recently been discovered that perhaps as many as half of all pulsars have velocities over 800 kilometers per second, and in some cases at least twice this value. At such speeds they would move out into the postulated corona in relatively short times of less than 10^sup 8^ years and would appear isotropically distributed. Second, there are three known cases where a gamma-ray burst has been repeated by the same source, and all three are associated with young supernova remnants (the dispersed material of the exploded star), which, of course, are the birthplaces of neutron stars. In particular, one of these three is located in the Large Magellanic Cloud at a known distance of 50 kiloparsecs, comparable to the postulated corona at 100 kiloparsecs. Among its 17 recorded bursts is one whose characteristics overlap many of those of the nonrepeating bursters.
Lamb cited other evidence linking bursters to neutron stars, in particular early evidence that some bursts have shown spectral lines, so-called cyclotron lines. These require a strong magnetic field to explain them, and again, it is neutron stars that have such fields.
Paczynski, however, discounted much of this evidence. He argued against there being a corona of sources, noting that the Andromeda galaxy, much like our own, but 700 kiloparsecs away, presumably also would have such a corona, yet does not show up at all on sky maps of the bursts. A counterargument would be that BATSE lacks the sensitivity to detect neutron star bursters at that distance, but Paczynski responds by saying that at their postulated speeds some of the Andromeda bursters would have reached our galaxy by now, yet we see no increase in the direction of Andromeda, so either there are none or they switch themselves off after a certain time. More particularly, however, Paczynski disagreed that sources in such a corona, if they had originated in the galactic disk, would appear isotropic. His adamant claim was that the signature of their birthplace would always be revealed in their present distribution.
As to the repeating bursters, Paczynski noted that while there may be some overlap in characteristics, those of the repeaters are sufficiently different from the non-repeaters to suggest they come from a different kind of source. He went on to note that the claims for cyclotron lines were uncertain, and BATSE, the most reliable and sensitive of available instruments, has so far not found any evidence for such lines.
Paczynski’s favoring of the cosmological scenario seems mainly a matter of default; his dissatisfaction with the galactic corona theory leaves only the cosmological one as an alternative. The strong point of this theory is that it easily explains the isotropy of sources as well as the breakdown of the three-halves law, which would result from the expansion of the universe (a redshift effect.) There is more positive evidence in that weaker bursts seem to last longer, which could be a relativistic time-dilation effect for more distant sources, but as yet this is not really certain. A significant weakness of the theory is that it offers no clue as to what actually causes bursts so powerful that we see the signal from across the universe.
Of course much evidence and discussion in the debate cannot be presented here, but I think it fair to say that, as in the first Great Debate, there was no clear victory for either side, and the moderator, wisely, did not call for a vote.
What of the future? A critical test will be to check whether or not the Andromeda galaxy has a corona of bursters. If it does, then the corona theory wins; if it doesn’t, the cosmological theory wins. It will take detectors one to two orders of magnitude more sensitive than BATSE to check this, but presumably that will come. Meanwhile, new, suitably located satellites will improve direction finding of bursts and testing for counterpart bursts at other wavelengths. Already Nature (4 April 1996, p. 377) reports a rapid-response system that flashes a burst detection from BATSE to the Goddard Space Flight Center in Maryland. There the burst coordinates on the sky are instantly calculated and sent across the Internet to a network of some 30 fast-slewing optical and radio telescopes around the world, so that within seconds of a gamma-ray burst the same area of the sky is being searched at longer wavelengths. So far, though, no luck!
We smile now when looking back at the original Great Debate, knowing as we do what the answer really was, and I’m confident that when the time comes to celebrate the 100th anniversary of that debate there will be smiles again in looking back to the 1995 debate.