A catalog of quasars near and far.
MORE THAN three decades have passed since Allan Sandage announced in 1960 that he and Thomas Matthews had identified a strange source of radio emission that, in visible light, looked like a faint star. Soon other such objects were found; they were eventually dubbed quasi-stellar objects (QSOs), or quasars.
In 1963 Maarten Schmidt identified the mysterious emission lines that had been found in one of them, 3C273, which at magnitude 13 is still the brightest known quasar. He showed that the lines are from ordinary hydrogen but redshifted by an extraordinary amount, 15.8 percent. This value implied that the object was receding from us at 44,000 kilometers per second. Then Jesse Greenstein and Tom Matthews showed that the spectral lines of 3C48 are redshifted by 36.7 percent, or 91,000 km per second. (Such a redshift is written as z = 0.367.) By 1967 astronomers had discovered and measured emission-line redshifts for about 150 quasars. Thousands more have been found since.
THE 1993 QSO CATALOG
Last year we published a catalog of all QSOs with measured redshifts -- 7,315 of them -- complete to April 1993. The redshifts range from z = 0.1 (our catalog's low-end cutoff) to nearly 4.9, while the apparent brightnesses range from magnitude 13 to about 22.5.
Before about 1980 most quasars that were identified were found by their radio emissions. Because radio surveys have been made over the entire sky, the radio QSOs are found everywhere too. With the advent of optical methods of identification, however, the situation changed. Optical surveys for candidate QSOs produce far more finds per square degree, and it soon became apparent that for every radio-emitting QSO there are about 100 radio-quiet ones. The optical surveys, however, have been made in very limited areas of the sky. Overall we believe that our catalog contains close to 20 percent of all QSOs brighter than magnitude 17.5, but only 0.7 percent of the total to magnitude 20. It is estimated that down to that limit, there are a million QSOs distributed fairly evenly across the sky.
One of the most remarkable features of QSOs, something we have known since the early years but which shows up with impressive clarity from our catalog, is their distribution in the "Hubble diagram" (page 34), a plot of apparent brightness versus redshift. It is very different from the Hubble diagram for galaxies originally produced by Edwin Hubble and others, which showed that normal galaxies become steadily fainter at increasing redshifts. This was an early sign that redshifts reliably indicate the distances of galaxies. For QSOs, however, the diagram shows a wide scatter in apparent brightness at every redshift. In fact there is little correlation of brightness to redshift at all! Either QSOs come in an extremely wide range of intrinsic luminosities, as most people believe, or their redshifts do not indicate distance.
In choosing between these interpretations, we have a personal dilemma. A scenario has been put together over the last 20 years in which QSOs are explained from birth to death, at least on a superficial level, as extreme cases of active galactic nuclei (S&T: August 1992, page 138). While there are many gaps in this model, it is accepted by all but a few. We are among the few.
Why do we feel this way? We shall briefly explain and then turn to the majority view.
Our interpretation is based on evidence that has circulated since the late 1960s. Quite a number of bright QSOs lie close to relatively bright, nearby galaxies (nearer than several hundred million light-years) that have much lower redshifts. This statistical evidence, and signs of optical or radio connections between galaxy and QSO, lead us to conclude that they are physically associated. Many spectacular cases of these apparent associations were discovered by Halton C. Arp and are discussed in his book Quasars, Redshifts and Controversies (S&T: January 1988, page 38). Contrary to what you are often told, the statistical evidence for association is well documented and has held up since the first proper analysis of QSOs in the 3C catalog was made in 1971.
An ingenious way of explaining away this phenomenon is to argue that the QSOs only appear to favor the halos of nearby galaxies because their images are amplified by gravitational microlensing due to dark stars in the halos. This was first proposed by Claude Canizares in 1980, but it fails quantitatively, essentially because there aren't enough faint QSOs for their images to be amplified.
Thus for us the only conclusion that can be drawn is that at least some QSOs are relatively nearby, and that a large fraction of their redshift is due to something other than the expansion of the universe. This applies even to some of the QSOs that show absorption lines in their spectra in addition to the emission lines. In these cases the QSOs must lie in and not far behind the halos of the galaxies.
Therefore we do not believe that QSOs can be used generally to investigate the distant universe. Perhaps there are two classes of QSOs. For one the cosmological redshift may dominate, but unless and until this is established we consider that the conventional picture is doubtful in the extreme. But here it is.
THE CONVENTIONAL VIEW
If the redshifts are due to the expansion of the universe, the QSO luminosities must be very large compared with those of galaxies. Because their emission-line spectra and overall energy distributions are rather similar to those of the nuclei of Seyfert and other active galaxies, the argument for a continuum of these objects with quasars at one end is very attractive.
It is argued that at very small redshifts, and when the nuclear luminosities are comparatively low, we can see the outer parts of the galaxies. In such cases we classify the objects as Seyferts. As the luminosities or distances increase it becomes harder to see the outer parts of the galaxies, so only the starlike nuclei will be visible. These are the QSOs, and the assumption is that each is embedded in a host galaxy too faint to be seen with current instruments. Echoing this view, some researchers use the term "active galactic nucleus" in place of quasar or QSO.
Why are QSOs so bright, and where does their energy come from? Even before the discovery of QSOs, it had become clear that the energy released in known active galaxies cannot come from the normal slow, evolutionary process of hydrogen burning in stars. Instead it appeared that the energy most likely has a nonthermal origin and is in fact incoherent synchrotron radiation, produced by electrons moving near the speed of light in magnetic fields. The high-energy end of this continuum radiation is adequate to explain, through photoionization, the emission-line spectra.
Where do the magnetic field and the electrons' energy come from? Many QSOs vary in brightness over time scales of years, months, or even days, so the size of the energy-generating volume must be very small, possibly no larger than the solar system. In the 1960s astronomers concluded that the energy must be gravitational in origin -- due to the falling together of matter -- or else it must be created somehow in situ. Because the latter suggestion requires a modification of the general theory of relativity, almost everyone argued that the gravitational model should be explored in every detail before any consideration is given to what is called "new physics."
Thus the gravitational-energy model has been developed into what is known as the black-hole accretion disk paradigm. It is believed that in the center of a galaxy a supermassive black hole is formed by evolutionary processes. The black hole is surrounded by an accretion disk containing gas, dust, and stars, and gravitational energy is released as matter spirals through the inner part of the disk down to the black hole. This energy is transformed into particle and magnetic energy and nonthermal radiation. Phrases like "feeding the monster" have been used by those who believe in these scenarios. There is no proof that the model is correct.
Recently there has been a revival of the old idea that violent stellar activity, such as cascades of supernovae, can explain at least the optical spectra. This is possibly correct, but the radio properties cannot be derived this way. To explain the radio phenomena it is argued that twin jets of relativistic particles are somehow ejected from the accretion disk. Also, some blobs must be ejected at bulk velocities greater than 99 percent of the speed of light in some objects -- those exhibiting so-called apparent superluminal motions.
What about the life history of QSOs? Some are found at very high redshifts -- greater than z = 4, meaning they were radiating within the first billion years of the universe. This requires that some galaxies formed and evolved to give rise to central black holes and accretion disks by that time. No one really knows how galaxies formed. The conventional wisdom (or lack of it) is that density variations were already present in the very early universe, and that they were able to condense into protogalaxies in time. Stars formed, the chemical elements were synthesized and ejected, and the galaxies evolved. Sometimes new physics in the form of cosmic strings or other topological defects in space-time is invoked to get galaxy formation started early enough.
The QSOs and the powerful radio galaxies are the only direct probes we have of the universe as it was many billions of years ago, since few normal galaxies have had redshifts measured out beyond about z = 1. This is why QSOs and radio galaxies are so important in the eyes of most cosmologists.
We can also use the absorption lines in the spectra of QSOs to investigate the universe. When the first absorption was discovered, its redshift turned out to be nearly the same as that of the emission lines. This means that the absorbing gas is very close to the QSO, if not part of it. But as absorption lines in more QSOs were detected, it was found that many sets of these lines have various lower redshifts. One possible explanation is that gas is being ejected from the QSO. However, the popular interpretation is that we see the QSO through many intervening gas clouds -- perhaps galactic halos -- at a variety of lesser distances.
Many of these systems of narrow absorption lines include heavy elements, and there is evidence that some of the lines do arise in the halos of intervening galaxies. There is a second system of narrow absorptions, which is due simply to the Lyman series of hydrogen lines; it was named the "Lyman-alpha forest" by Roger Lynds. Very large numbers of such lines appear in high-redshift QSOs, extending from immediately to the short-wavelength side of the observed Lyman-alpha emission line all the way to the Lyman cutoff corresponding to zero redshift. On the assumption that these too are due to intervening gas, large numbers of small hydrogen clouds must populate the intergalactic medium.
Are QSOs really telling us about the universe as it was in the distant past, or is there a deeper mystery? After 30 years some of us, at least, still don't know.
Geoffrey Burbidge is former director of Kitt Peak National Observatory and professor of physics at the University of California at San Diego. He spends most of his time nowadays "attempting to distinguish between fact and fiction in extragalactic astronomy." Adelaide Hewett has worked on quasars with Margaret and Geoffrey Burbidge since 1963.