Roger D. Blandford

Quasars, which were first discovered in 1963, are active regions in the centers of galaxies that are so luminous that they outshine the surrounding stars. Many astronomers have concluded that the source of their prodigious power is a massive black hole that attracts surrounding gas by its gravitational force and liberates gravitational energy as radiation. In the context of the black hole model for quasars, this inward drift of matter is known as accretion and the rate and the manner of the accretion in individual active galactic nuclei is a major factor in determining the type of object we observe.


There is a natural limit, known as the Eddington limit and named alter the famous astronomer, Sir Arthur Eddington, to the luminosity L that can be radiated by a compact object of mass M. This limit arises because both the attractive gravitational force acting on an electron-ion pair and the repulsive force due to radiation pressure decrease inversely with the square of the distance from the black hole. When the luminosity exceeds the Eddington limit, which is given by

LEdd = 4piGMmpc / sigmaT,

the gas will be blown away by the radiation. (In this equation, G is the gravitational constant, mp is the mass of a proton, c is the speed of light, and sigmaT is the Thompson cross section or the effective area of an electron when it is illuminated by radiation.) Note that the Eddington limit is independent of distance from the compact object. Numerically, if we express the mass in units of mass of the sum Msmsun and the luminosity in units of the luminosity of the sun Lsmsun, then

LEdd = 30,000 (M / Msmsun) Lsmsun.

A bright quasar has a luminosity of about 1013 Lsmsun, and so if it is to to continue to attract gas to power itself, the central mass must exceed about 3 x 108 Msmsun.

Most galaxies display evidence for central activity and so they may also contain black holes although they need not be as massive as the black holes in bright quasars. Indeed, we know that the maximum mass of a hypothetical black hole in the center of our milky way is 3 x 106 Msmsun. Correspondingly, most galaxies that contain massive black holes do not necessarily accrete gaseous fuel at a sufficient rate to maintain their luminosities at the Eddington value.


Stars and gas are observed to rotate about the centers of galaxies. This implies that if and when gas accretes toward a black hole, the centrifugal force acting upon the gas will increase in importance relative to the gravitational force. The gas is then expected to settle into a rotating disk, known as an accretion disk, just like the gas in some mass transfer binary stars. If the gravitational field were spherically symmetrical, then there would be no reason for the orbiting gas to move in any particular plane. However, the stars in the centers of galaxies are not spherically distributed and probably define a preferred plane into which the disk can settle. In addition, the black hole itself would probably be spinning rapidly and general relativistic effects may twist the accretion disk into its equatorial plane. This need not coincide with that defined by the stars. The orbital period of the gas in an accretion disk will change with radius, just like the orbital periods of the planets in the solar system. This implies that adjacent rings of gas will rub against each other and be subject to friction which will allow the gas to move toward the black hole. Several sources of this decelerating frictional force have been suggested. It may be caused by turbulent motions of the gas; alternatively, it has been attributed to magnetic field lines that are stretched between one ring and the next. Other possibilities, which are more likely to operate in the outer parts of accretion disks, include the development of gas clouds, bars, and spiral arms. Many active galactic nuclei produce a pair of jets - two collimated outflows that carry gas away from the nucleus to the outer parts of the galaxy and beyond. It is widely believed that these jets are launched perpendicular to the central accretion disks. Other objects exhibit outflowing winds. It is possible that the creation of either jets or a wind might also produce a reaction force on the gas in the disk allowing the gas to sink inward toward the central black hole.

Whatever its origin, this frictional force is responsible for heating the gas in the disk which can then radiate. The source of the radiant energy is ultimately gravitational and up to about 1020 erg of energy may be released for every gram of gas that is accreted onto a blackhole. (This is several hundred times more efficient than the nuclear processes occurring in stars.) Most of this energy will be released fairly close to the black hole, within a radius of typically 1015 cm for massive black hole in a quasar. In order to fuel a bright quasar, gas must accrete at a rate of up to 10 Msmsun yr-1.

If there is enough gas around the black hole, then the escaping photons will be absorbed and reemitted several times before they escape. The characteristic frequencies of the radiation can be calculated from Stefan's law just as is done for a stellar atmosphere. For an active galactic nucleus, this turns out to be in the ultraviolet part of the spectrum, which is where most observed objects appear to be most luminous. However, not all this ultraviolet radiation need escape. Some of it will be intercepted by dense clouds and converted into the emission lines by which active galaxies are frequently recognized. More of it may be intercepted by dust grains in the outer parts of the disk and transformed into infrared radiation.

Accretion disks are probably endowed with a magnetic field and their orbital velocity is necessarily supersonic. It is therefore expected that they are embedded in very hot, though transparent, coronae, analogous to the solar corona. The magnetic field lines will be twisted and torn by the motion of the disk and this may lead to the acceleration of relativistic electrons, which some astronomers believe emit the x-rays and gamma rays that are seen coming from some active galaxies.


The large fueling rates required by the most energetic quasars demand a more copious source than normal stars, evolving in the body surrounding galaxy. Some astronomers believe that direct collisions between individual stars in the nucleus of the galaxy is responsible for releasing the gas. Another possibility, particularly relevant to lower power objects, is that stars that pass too close to the black hole itself may be torn apart by the tidal gravitational force exerted by the black hole.

However, recent observations suggest that quasar activity is actually triggered by interactions between galaxies. In some instances, a small galaxy makes a direct hit on a larger galaxy and is ingested and falls to the center of the larger galaxy. In other collisions the incident galaxy may only strike a glancing blow, and ther will only be a small transfer of mass. In fact, no mass transfer is necessary, and in most cases, just the gravitational perturbation due to the incident galaxy can be sufficient to trigger the formation of spiral arms and bars in the galaxy surrounding the active nucleus which may, in turn, drive the gas inward.


The previous description of the workings of a quasar is still largely conjectural. This is mainly because it is not possible to resolve the smallest regions where most of the energy is released. Indeed, it is proving to be very hard to produce clear and unambiguous proof that massive black holes are present. In addition, it is still quite uncertain what is the source of the accreting gas and by what mechanism does it settle toward the black hole.

Fortunately, observations scheduled over the next five years may test the accreting black hole model and provide answers to these difficult questions. Very long baseline interferometry, performed with VLBA and from space, should reveal finer detail in radio maps of nearby and distant active galactic nuclei and may even be able to trace the outer parts of accretion disks. The Hubble space telescope, with its unprecedented resolution at optical wavelengths, should be able to trace the central velocity dispersions of stars in nearby galaxies and thereby measure the central mass which ought to be a fuel-starved black hole in most instances. However, the greatest progress in our understanding may be less direct and come from observing gas in the outer parts of galaxies either accreting onto or flowing away from the nuclei. Only when we understand the accretion process in physical terms will we be able to account for the evolution of quasars, Seyfert galaxies, and radio galaxies.

Additional Reading
  1. Balick, B. and Heckmann, T. M. (1982). Extranuclear clues to the origin and evolution of activity in galaxies. Ann. Rev. Astron. Ap. 20 431.
  2. Meyer, F., Duschl, W. J., Frank, J., and Meyer-Hofmeister, E., eds. (1989). Theory of Accretion Disks. Kluwer Academic Publishers, Dordrecht.
  3. Frank, J. H., King, A. R., and Raine, D. I. (1986). Accretion Power in Astrophysics. Cambridge University press, Cambridge.
  4. Shapiro, S. L., and Teukolsky, S. A. (1983). Black Holes, White Dwarfs and Neutron Stars: The Physics of Compact Objects. Wiley, New York.
  5. See also Accretion; Active Galaxies and Quasistellar Objects, Jets, Black Holes, Theory, Galaxies, Nuclei.