(Background, above) This artist's conception depicts one possible scenario for the "central engine" of an active galaxy. Two jets of matter are shown moving outward perpendicular to a flattened accretion disk surrounding a supermassive black hole. Such jets are seen at many places in the universe.

(Inset A) Numerical simulations--i.e., done on computers--can help us visualize key events in the universe. Here a uniform distribution of some thousand stars is spread over a dimension of several hundred parsecs.

(Inset B) As the simulation proceeds, the hypothetical stars congregate preferentially toward the core of the star cluster, forming a "luminosity cusp" of light.

(Inset C) If we could see inside the cusp, as shown in this piece of art, an accretion disk would probably be evident, swirling around a giant black hole.

(Inset D) As the accretion disk rotates, tilts, and acquires more matter--probably as whole stars fall in--jets of matter emerge to cool the environment surrounding the hole.


Studying this chapter will enable you to:

Specify the basic differences between active and normal galaxies.

Describe the important features of Seyfert and radio galaxies.

Explain what drives the central engine thought to power all active galaxies.

Describe the observed properties of quasars, and discuss the special properties of the radiation they emit.

Discuss the place of active galaxies in current theories of galactic evolution.

Our journey from the Milky Way to the Great Wall in the past two chapters has widened our cosmic field of view by a factor of 10,000, yet the galaxies that make up the structures we see show remarkable consistency in their properties. The overwhelming majority of galaxies fit neatly into the Hubble Classification Scheme, showing few, if any, "unusual" characteristics. However, sprinkled through the mix of normal galaxies, even relatively close to the Milky Way, are galaxies that are decidedly abnormal in their properties. Although their optical appearances are often quite ordinary, these abnormal galaxies emit huge amounts of energy--far more than a normal galaxy--mostly in the invisible part of the electromagnetic spectrum. Observing such objects at great distances, we may be seeing some of the formative stages of our own galactic home.

Back Astronomers have recognized and cataloged spiral, elliptical, and irregular galaxies as far away as several hundred megaparsecs from Earth. Beyond this distance, galaxies appear so faint that it is difficult to discern their characteristic shapes, so their types are largely unknown. Nevertheless, according to their observed redshifts (and Hubble's law), we know that many galaxies lie well beyond this distance, in the farthest reaches of the observable universe. But what kinds of objects are they? Are they normal galaxies--close relatives of the galaxies that populate our local neighborhood--or are they somehow different? The answer is that they often seem to be different from the galaxies found in our cosmic backyard. By and large, very distant objects tend to be more active--more luminous--than those found closer to home. The most energetic objects, which can emit hundreds or thousands of times more energy per second than our entire Galaxy, are known collectively as active galaxies.

Not all active galaxies are distant, and only a small fraction of all distant galaxies are active. Some active galaxies are found locally, scattered among the normal galaxies that make up most of our cosmic neighborhood, while many faraway "normal luminosity" galaxies are known. As a general rule, though, active galaxies are more common at greater distances, and the most active objects lie farthest from the Earth. We might wonder whether this predominance of energetic objects with large redshifts is just an observational effect, resulting from our inability to detect relatively faint normal galaxies at great distances. After all, the apparent brightness of any astronomical object decreases as the square of its distance from us, so even with the very best telescopes, we would expect to observe only the more energetic and powerful galaxies in remote regions of space. Although this observational bias does play a role, it turns out that it can only partly explain the apparent predominance of energetic objects among those having large redshifts. We are led to conclude that bright active galaxies really are more common at great distances.

In addition to their great brightness, there is something else that is basically different about active galaxies. It seems that their radiative character differs fundamentally from that of normal galaxies. Most of a normal galaxy's radiated energy is emitted in or near the visible portion of the electromagnetic spectrum, much like the radiation from ordinary stars. Indeed, to a large extent, the light we see from a normal galaxy is just the accumulated light of its many component stars. For example, our entire Milky Way has a luminosity of about 1037 W at optical frequencies--20 billion Suns' worth of radiation--but only 1031 W at radio frequencies--a million times less. By contrast, as illustrated in Figure 25.1, the radiation observed emitted by active galaxies does not peak at optical frequencies--far more energy is emitted at longer wavelengths than in the visible range. The radiation from active galaxies is inconsistent with what we would expect if it were the combined radiation of myriad stars. The radiation is thus said to be nonstellar in nature.

Figure 25.1 The nature of the energy emitted from a normal galaxy differs from that of an active galaxy. (This plot illustrates the general run of intensity for all galaxies of a particular type and does not represent any one individual galaxy.)

The two most important categories of active galaxies are Seyfert galaxies and radio galaxies, although other classes exist (see, for example, Interlude 25-1 below). Even more extreme in their properties are the quasars. Astronomers conventionally distinguish between active galaxies and quasars based on their spectra, appearance, and distance. Active galaxies typically look like galaxies, whereas quasars are often so distant that little structure can be discerned. In addition, quasars are generally even brighter than active galaxies (which are themselves much brighter than normal galaxies like the Milky Way). The distinction is largely historical, however--it dates back to the days when the connection between quasars and active galaxies was not understood. As we will see later, most astronomers now believe that quasars are simply an early stage of galaxy formation, and that there is really no sharp dividing line between quasars and active galaxies. Many researchers now go so far as to include quasars in the "active galaxy" category.

Physical conditions were undoubtedly different at earlier times than they are now. Perhaps, then, we should not be surprised that remote astronomical objects, which emitted long ago the radiation we observe today, differ from nearby objects, which emitted their radiation much more recently. What is surprising--in fact, astounding--is the amount of energy radiated by some of the most luminous objects. Their tremendous power, nonstellar radiation, and abundance at great distances suggest to many astronomers that the universe was once a much more violent place than it is today.

Back In 1943 Carl Seyfert, an American optical astronomer studying spiral galaxies from Mount Wilson Observatory, discovered the type of active galaxy that now bears his name. Seyfert galaxies are a class of astronomical objects whose properties lie between those of normal galaxies like the Milky Way and those of the most violent active galaxies known. This fact suggests to some astronomers that Seyferts represent an evolutionary link between these two extremes. The spectral lines of Seyfert galaxies are usually substantially redshifted--most Seyferts seem to reside at large distances (hundreds of megaparsecs) from us, although a few are as close as 20 or 30 Mpc.

Figure 25.2 shows two optical images of a typical Seyfert galaxy. A casual glance at a long-exposure photograph of a Seyfert (such as Figure 25.2a) reveals nothing strange. Superficially, Seyferts resemble normal spiral galaxies. However, closer study of Seyferts reveals some peculiarities not found in normal spirals.

Figure 25.2 Photographs of a Seyfert galaxy (NGC 5728) (a) from the ground and (b) from Earth orbit. The enlarged view in (b) of two cone-shaped beams of light shows a group of glowing blobs near the galaxy's core, perhaps illuminated by radiation arising from the accretion disk of a black hole. This object is about 40 Mpc away--one of the closest active galaxies known.

First, maps of Seyfert energy emission show that nearly all of the radiation stems from a small central region known as the galactic nucleus. This region can be seen at the center of the overexposed core of Figure 25.2(a), and is shown in more detail in Figure 25.2(b). Astronomers suspect that a Seyfert nucleus may be quite similar to the center of a normal galaxy, such as the Milky Way or the Andromeda galaxy, but with one very important difference. The nucleus of a Seyfert is 10,000 times brighter than the center of our Galaxy. Indeed, the brightest Seyfert nuclei are 10 times more energetic than the entire Milky Way.

Second, Seyfert galaxies emit their radiation in two broad frequency ranges. The stars in the Seyfert's galactic disk and spiral arms produce about the same amount of visible radiation as those of a normal spiral galaxy. However, most of the energy from the Seyfert's nucleus is emitted in the form of invisible radio and infrared radiation, which cannot be explained as coming from stars--it must be nonstellar in origin.

Third, Seyfert spectral lines bear little or no resemblance to those produced by ordinary stars, although they do have many similarities to the spectral lines observed toward the center of our own Galaxy. Seyfert spectra contain strong emission lines of highly ionized heavy elements, especially iron. The lines are very broad, indicating either that the galaxy's gases are tremendously hot (more than 108 K) or that they are rotating very rapidly (at about 1000 km/s) around some central object. The first possibility can be ruled out, since such a high temperature would cause all the gas to be ionized, so that no spectral lines would be produced. Thus, the broadening indicates rapid internal motion in the nucleus.

Finally, extensive monitoring of Seyfert radiation over long periods of time has shown that the energy emission often varies over time. Figure 25.3 shows an example of luminosity variations for a typical Seyfert. These radiative changes are unlike anything found in the Milky Way or any other normal galaxy. A Seyfert's luminosity can double or halve within a fraction of a year.

Figure 25.3 The irregular variations of a particular Seyfert galaxy's luminosity over a period of two decades. Because this Seyfert, called 3C 84, emits most strongly in the radio part of the electromagnetic spectrum, these observations were made with large radio telescopes. The optical and X-ray luminosities vary as well.

These rather rapid fluctuations lead us to conclude that the source of energy emissions must be quite compact. As mentioned in Chapter 22 (in the context of luminosity variations in neutron stars and black holes), for astronomers to be able to detect a coherent variation in brightness within a certain time interval, the source of radiation must be smaller in size than the distance traveled by light in that interval. Otherwise, the intensity variations would be blurred, not sharp, as observed. Simply put, an object cannot "flicker" in less time than radiation takes to cross it. Because the rise and fall of a Seyfert's radiation usually occurs within one year, we can confidently conclude that the emitting region must be less than one light year across--an extraordinarily small region, considering the amount of energy emanating from it. High-resolution interferometric radio maps of Seyfert cores generally confirm this reasoning.

Seyfert galaxies are apparently experiencing huge explosions within their cores. Their time variability and their large radio and infrared luminosities together strongly imply violent nonstellar activity. The nature of this activity may well resemble processes occurring in the center of our own Galaxy, but its magnitude is thousands of times greater than the comparatively mild events within our own Galaxy's center.

Seyfert galaxies are not the only kind of active galaxy known. Radio observations have uncovered another class of extremely energetic sources that also seem to be galactic in nature. Radio galaxies differ from Seyferts both in the wavelength at which their energy is emitted and in the appearance and extent of the emitting region.


One common type of radio galaxy is often called a core-halo radio galaxy. As illustrated in Figure 25.4, the energy from such an object comes mostly from an extremely small central nucleus, or core, less than 1 pc across, with weaker emission coming from an extended halo surrounding it. The halo typically measures about 50 kpc across, similar in size to the surrounding visible galaxy, which is usually elliptical and often quite faint. The radio luminosity from the core can be as great as 1037 W--about the same as the emission from a Seyfert nucleus and comparable to the output from our entire Milky Way Galaxy at all wavelengths.

Figure 25.4 Radio contour map of a typical core-halo radio galaxy. The radio emission from such a galaxy comes from a bright central nucleus, or core, surrounded by an extended, less intense halo. The radio map is superimposed on an optical image of the galaxy and some of its neighbors, shown previously in Figure 24.15.

Figure 25.5 shows two optical photographs of a core-halo radio galaxy, along with false-color images of its radio and infrared emissions. This object is a giant elliptical galaxy known as M87--the eighty-seventh object in Messier's catalog. (We can be sure that this eighteenth-century Frenchman had no idea what he was really looking at. Nor perhaps do we!) M87 is roughly 20 Mpc distant, making it a prominent member of the Virgo Cluster, and one of the closest active galaxies. Its nearness and interesting activity have made it one of the most intensely studied of all astronomical objects.

Figure 25.5 The giant elliptical galaxy M87 (also called Virgo A) is displayed here at several different wavelengths. (a) A long optical exposure of its halo and embedded central region. (b) A short optical exposure of its core and an intriguing jet of matter, on the same scale as (a). (c) A radio image of its jet, on a somewhat expanded scale compared with (b). The red dot at left marks the bright nucleus of the galaxy; the red and yellow blob near the center of the image corresponds to the bright "knot" visible in the jet in (b). (d) A recent image of the jet taken in the near-infrared, at roughly the same scale as (c).

A long time exposure (Figure 25.5a) shows a large fuzzy ball of light--a fairly normal-looking type E1 elliptical galaxy, about 100 kpc across. A shorter exposure of M87 (Figure 25.5b) captures only the brightest regions of the galaxy and reveals a compact central region only a few hundred parsecs in diameter. Beyond the core, a long thin jet of matter ejected from M87's center appears. The jet is about 2 kpc long and is traveling outward at a very high velocity, possible as great as half the speed of light. Computer enhancement shows that the jet is made up of a series of distinct "blobs," more or less evenly spaced along its length. This high-speed jet, which emits energy at the rate of almost 1035 W, has been imaged in the radio and infrared as well as in the optical regions of the spectrum (Figures 25.5c and d). Jets such as this are vital to our understanding of these energetic radio sources, as we will see in a moment.


Many radio galaxies are not of the core-halo type. While they do emit most of their radiation in the long-wavelength part of the spectrum, like Seyferts and core-halo galaxies, very little of this emission arises from a compact central nucleus. Instead, most of the energy comes from giant extended regions called radio lobes--roundish clouds of gas up to a megaparsec across, lying well beyond the center of the galaxy itself. For this reason, these objects are known as lobe radio galaxies.

The radio lobes of all lobe radio galaxies are truly enormous. From end to end, an entire lobe radio galaxy typically is more than ten times the size of the Milky Way, comparable in size to the entire Local Group. The lobes emit no visible light, but their radio luminosity can range from 1036 to 1038 W--between one-tenth and 10 times the total energy emitted by our Galaxy. Several of these strange objects are located relatively nearby, so we can study them at close range. One such system, known as Centaurus A, is shown in Figure 25.6. It lies only 4 Mpc from the Earth.

Figure 25.6 Lobe radio galaxies, such as Centaurus A shown here, have giant radio-emitting regions extending a million parsecs or more beyond the central galaxy. The lobes cannot be imaged in visible light and are observable only with radio telescopes. (The lobes are filled with false color to indicate decreasing intensity from red to yellow to green to blue.)

Figure 25.7 is an optical image of Centaurus A, with another representation of Figure 25.6 superimposed to show the relation between the optical and radio emission. In visible light, Centaurus A is a rather peculiar-looking object, apparently an E2 galaxy bisected by an irregular band of dust. Numerical simulations suggest that this system is probably the result of a merger between an elliptical galaxy and a smaller spiral galaxy about 500 million years ago. The radio lobes are roughly symmetrically placed, jutting out from the center of the visible galaxy. Note, too, that the jets are roughly perpendicular to the dust lane. The elliptical galaxy itself is very large--some 500 kpc in diameter. Relatively little radio emission is observed from the location of the optical image, however. Most of the radio radiation arises from the giant lobes well beyond it.

Figure 25.7 An optical photograph of Centaurus A, one of the most massive and peculiar galaxies known and believed to be the result of a galaxy collision some 500 million years ago. The pastel false colors mark the radio emission shown in Figure 25.6, in this case more recently acquired and with higher resolution. Note that the radio lobes emit no visible light.

The lobes of radio galaxies vary in size and shape from source to source, but they maintain their alignment with the center of the optical galaxy in nearly all cases. This fact suggests that the lobes are actually "blobs" of material that were somehow ejected in opposite directions by violent events in the galactic nucleus. In the case of Centaurus A, this argument is strengthened by the presence of an additional pair of secondary lobes, smaller than the main lobes (about 50 kpc in length, marked in Figure 25.7) and closer to the visible galaxy. Both pairs of lobes share the same high degree of linear alignment. Astronomers believe that the inner lobes were expelled from the nucleus by the same basic process as the outer ones, but more recently, so they have not had time to travel as far. Still higher-resolution studies reveal the presence of a roughly 1-kpc-long jet in the center of Centaurus A, again aligned with the larger lobes.

If material is ejected from the nucleus at close to the speed of light (which seems likely), it follows that Centaurus A's outer lobes were created a few hundred million years ago, possibly around the time of the collision thought to be responsible for the galaxy's peculiar optical appearance. Apparently some violent process at the center of Centaurus A started up around that time and has been intermittently firing jets of matter out into intergalactic space ever since.

Further evidence in favor of the interpretation of radio lobes as material ejected from the center of a galaxy is provided by another object, Cygnus A, shown as an optical image in Figure 25.8(a) and as high-resolution radio map in Figure 25.8(b). The filamentary structure evident in the radio lobes and the thin, radio-emitting line joining the right lobe to the center of the visible galaxy (the dot at the center of the radio image) strongly suggest that we are seeing two oppositely directed jets of material running into the intergalactic medium.

Figure 25.8 (a) Cygnus A also appears to be two galaxies in collision, although it is not completely clear that that is what is really happening. (b) On a much larger scale, this cosmic object displays radio-emitting lobes on each side of the optical image. To put these images into proper perspective, the optical galaxy in (a) is about the size of the small dot at the center of the "radiograph" in (b). Notice the thin line of radio-emitting material joining the right lobe to the central galaxy. Frame (c) shows even greater detail in one of the lobes.

In some systems, known as head-tail radio galaxies, the lobes seem to form a "tail" behind the main galaxy. For example, the lobes of radio galaxy NGC 1265, shown in Figure 25.9, appear to be "swept back" by some onrushing wind. In fact, this is the most likely explanation for the galaxy's appearance. If NGC 1265 were at rest, it would be just another double-lobe source, perhaps looking quite similar to Centaurus A. However, the galaxy is traveling through the intergalactic medium of its parent galaxy cluster (known as the Perseus Cluster), and the outflowing matter forming the lobes tends to be left behind as the galaxy moves.

Figure 25.9 (a) Radiograph, in false color, of the active "head-tail" galaxy NGC 1265. (b) The same radio data, in contour form, superposed on the optical image of the galaxy. Astronomers reason that this object is moving rapidly through space, trailing a "tail" behind as it goes--a little like a comet, but on a vastly larger scale.

Radio galaxies share many characteristics with Seyfert galaxies. They emit comparably large amounts of energy, and there is good evidence that the energy source is a compact region at the center of an otherwise relatively normal-looking galaxy. In the lobe radio galaxies, that energy is fired out from the nucleus in the form of jets of matter and is ultimately emitted (in the form of radiation) from far beyond the galaxy itself. But the central compact nucleus is still thought to be the place where the energy is actually produced. Before going on to probe even more distant, and more violent, objects, let us now consider the current view of the "engine" that powers all this activity.

Back The behavior of active galaxies is contrary to that expected from vast collections of stars. The lobe radio galaxies in particular, with their huge energy emission from far beyond the optical galaxy, are among the most powerful objects in the universe. Can we explain this enormous nonstellar energy output in terms of known physics? Remarkably, the answer is "Yes." The present consensus among astronomers is that, despite the great differences in appearance, these objects--Seyferts and radio galaxies--may share a common energy-generation mechanism. The energy can be reprocessed into many different forms before it is finally emitted into intergalactic space, but the engine is probably the same in either case.

As a class, active galaxies (and quasars too, as we will see) show some or all of the following properties.

  1. They have high luminosities, generally greater than the 1037 W characteristic of a fairly bright normal galaxy like the Milky Way.
  2. Their energy emission is nonstellar--it cannot be explained as the combined radiation of even trillions of stars.
  3. Their energy output can be highly variable, implying that it is emitted from a small central nucleus much less than a parsec across.
  4. They often exhibit jets and other signs of explosive activity.
  5. Their optical spectra may show broad emission lines, indicative of rapid internal motion within the energy-producing region.

The principal questions then are: How can such vast quantities of energy arise from these relatively small regions of space? Why is so much of the energy radiated at low frequencies, especially in the radio and infrared? And what is the origin of the extended radio-emitting lobes and jets? Let us first consider how the energy is produced.


To develop a feeling for the enormous emissions of active galaxies, consider for a moment an object with a luminosity of 1038 W. In and of itself, this energy output is not inconceivably large. The brightest giant ellipticals are comparably powerful. Thus, some 1012 stars--a few normal galaxies' worth of material--could equivalently power a typical active galaxy. The difficulty arises when we consider that in an active galaxy this energy production is packed into an object much less than a parsec in diameter!

It is difficult to imagine how several Milky Way Galaxies could be compressed into a space no larger than a parsec. Even if we could somehow squeeze that much mass into such a volume, it would immediately collapse to form a huge black hole, and none of the light it produced could escape to the outside! Thus, even neglecting its nonstellar spectrum, the total energy output of an active galaxy simply cannot be explained as the combined energy of many stars. We must think of something else.

The twin requirements of large energy generation and small physical size bring to mind our discussion of X-ray sources in Chapter 22. The presence of the jet in M87 and the ejection of matter to form radio lobes in Centaurus A and Cygnus A strengthen the connection. Recall that the best current explanation for those "small-scale" phenomena involves the accretion of material onto a compact object--a neutron star or a black hole. Large amounts of energy are produced as matter spirals down onto the central object, and high-speed jets may well be a common by-product of the process. In Chapter 24, we suggested that a similar mechanism, involving a supermassive black hole--one with a mass of around a million suns--may also be responsible for the energetic radio and infrared emission observed at the center of our own Galaxy.

As illustrated in Figure 25.10, the leading model for the central engine of active galaxies is essentially a scaled-up version of the same accretion process, now involving black holes with masses between a few million and a billion times the mass of the Sun. As with its smaller-scale counterparts, infalling gas forms an accretion disk and spirals down toward the hole. It is heated to high temperatures by friction within the disk and emits large amounts of radiation as a result. In this case, however, the origin of the accreted gas is not a binary companion, as in stellar X-ray sources, but entire stars and clouds of interstellar gas that come too close to the hole and are torn apart by its strong gravity.

Active Galaxy

Figure 25.10 The leading theory for the energy source in active galactic nuclei (and quasars) holds that these objects are powered by material accreting onto a supermassive black hole. As matter spirals toward the hole, it heats up, producing large amounts of energy. At the same time, high-speed beams of gas may be ejected perpendicular to the accretion disk, forming the jets and lobes seen in many active objects. Magnetic fields generated in the disk are carried by the jets out to the radio lobes, where they play a crucial role in producing the observed radiation.

The accretion process is extremely efficient at converting infalling mass (in the form of gas) into energy (in the form of electromagnetic radiation). Detailed calculations indicate that as much as 10 or 20 percent of the total mass energy of the infalling matter can be radiated away before it crosses the hole's event horizon and is lost forever. Since the total mass energy of a star like the Sun--the mass times the speed of light squared--is about 2 × 1047 joule, if follows that the 1038 W luminosity of a bright active galaxy can be accounted for by the consumption of only 1 solar mass of gas per decade by a billion-solar-mass black hole. Less luminous active galaxies would require correspondingly less fuel--for example, a 1036 W Seyfert galaxy would devour only one star's worth of material every thousand years.

In this picture, the small size of the emitting region is a direct consequence of the compact nature of the central black hole. Even a billion-solar-mass hole has a radius of only 3 × 109 km, or 10-4 pc--about 20 A.U.--and theory suggests that the part of the accretion disk responsible for most of the emission would be much less than a parsec across. Instabilities in the accretion disk can cause fluctuations in the energy released, leading to the variability observed in many objects. The broadening of the spectral lines observed in the nuclei of Seyferts (and in quasars) results from the rapid orbital motion of the gas in the hole's intense gravity.

Recent observations of galaxies in the Virgo Cluster by the Hubble Space Telescope lend strong support to this general picture. Figure 25.11 shows an image of a disk of gas and dust apparently feeding a possible black hole at the core of a giant elliptical galaxy. As expected from the theory just described, the disk is perpendicular to the huge jets emanating from the center of the active galaxy.

Figure 25.11 (a) A combined optical/radio image of the giant elliptical galaxy NGC 4261, in the Virgo Cluster, shows a white visible galaxy at center, from which red-orange (false color) radio lobes extend for about 60 kpc. (b) A close-up photograph of the galaxy's core reveals a 100-pc-diameter disk surrounding a bright hub that presumably harbors a black hole.

Hubble has also allowed astronomers to probe the fine details of the Virgo Cluster's most prominent object--the huge M87 galaxy (Figure 25.5)--and what they have found is in excellent agreement with the idea that the energy is produced by accretion onto a large black hole. At M87's distance, Hubble's resolution of 0.05 arc second corresponds to a distance of about 5 pc, so we are still far from seeing the (solar-system-sized) central black hole itself, but the improved "circumstantial" evidence has convinced many doubters of the correctness of the theory. Figure 25.12 shows imaging and spectroscopic data that suggest the existence of a rapidly rotating disk of matter orbiting the galaxy's center, perpendicular to the jet. Measurements of the gas velocity on opposite sides of the disk indicate that the mass within a few parsecs of the center is approximately 3 × 109 solar masses--we assume that this is the mass of the central black hole.

Figure 25.12 Recent imaging and spectroscopic observations of M87 support the idea of a rapidly whirling accretion disk at its core. (a) An image of the central region of M87, similar to that shown in Figure 25.5(d), shows its bright core and jet. The scale is comparable to the scale of Figure 25.5(c). (b) A magnified view of the core suggests a spiral swarm of stars, gas and dust. (c) Spectral-line features observed on opposite sides of the core show opposite Doppler shifts, implying that the material there is coming toward us on one side and moving away from us on the other. The strong implication is that an accretion disk spins perpendicular to the jet, and that at its center is a black hole having some 3 billion times the mass of the Sun.

Even more compelling evidence for a supermassive black hole has recently come from studies with radio telescopes. Using the Very Long Baseline Array--a continent-wide network of ten radio telescopes--a U.S.-Japanese team was able to achieve spectacular angular resolution, in fact hundreds of times better than with the Hubble Space Telescope. Observations of NGC 4258, a spiral galaxy about 6 Mpc away, have uncovered a group of molecular clouds that are swirling in an organized fashion about the galaxy's core. Armed with an understanding of the Doppler effect, the astronomers have been able to detect the red and blue shifts of water-vapor spectral lines in those faraway clouds. As depicted in Figure 25.13, the pattern revealed by the clouds is that of a slightly warped and spinning disk centered precisely on the galaxy's heart. The rotation velocities imply the presence of more than 40 million solar masses all packed within a region less than 0.2 pc across--a mass density more than ten times that of any previously observed black-hole candidate.

Figure 25.13 A network of radio telescopes has probed the core of the spiral galaxy NGC 4258, shown here in the light of mostly hydrogen emission. Within the innermost 0.2 pc (inset), observations of Doppler-shifted molecular clouds (designated by red, green, and blue dots) show that they obey Kepler's Third Law perfectly, and have revealed a slightly warped disk of rotating gas (shown here in artist's conception). At the center of the disk presumably lurks a huge black hole.


The model just described is now widely accepted as the correct picture of how active galaxies and quasars generate their enormous power. It provides a natural explanation of the observed facts, and has the added advantage of being essentially similar to the processes thought to power smaller-scale phenomena, such as stellar X- and gamma-ray sources and normal galactic nuclei. Having thus accounted for the source of the energy, let us now turn to the way in which it is eventually emitted into intergalactic space.

In order to account for the details of the observed radiation spectra of some Seyfert galaxies, it is necessary to assume that the energy emitted from the accretion disk is "reprocessed"--that is, absorbed and reemitted--by gas and dust surrounding the nucleus. The jets and lobes seen in many systems consist of material (mainly protons and electrons) blasted out into space--and out of the galaxy entirely--from the inner regions of the disk. The details of how these jets form remain uncertain, but, as we have seen, there is a growing consensus among theorists that jets are a very common feature of accretion flows, large and small. The jets also contain strong magnetic fields (shown in Figure 25.10), possibly generated by the gas motion within the disk itself, which accompany the gas as it leaves the galaxy.

As sketched in Figure 25.14(a), whenever a charged particle (here an electron) encounters a magnetic field, it tends to spiral around the field lines. We have encountered this idea several times previously, in a variety of different contexts (see, for example, the discussion of planetary magnetospheres in Chapters 7 and 11, or solar activity in Chapter 16. As the particles whirl around, they emit electromagnetic radiation, as discussed in Chapter 3. The faster the particles move, or the stronger the magnetic field, the greater the amount of energy radiated. In most cases, the fastest-moving particles are the low-mass electrons, so they are responsible for essentially all of the radiation we observe.

Figure 25.14 (a) Charged particles, especially fast electrons (red), emit synchrotron radiation (blue) while spiraling in a magnetic field (black lines). This process is not confined to active galaxies. It occurs, on smaller scales, when charged particles interact with magnetism in the Earth's Van Allen belts (see Chapter 7), when charged matter arches above sunspots on the Sun (see Chapter 16), in the vicinity of neutron stars (see Chapter 22), and at the center of our own Galaxy (see Chapter 23). (b) Variation in the intensity of thermal and synchrotron (nonthermal) radiation with frequency. Thermal radiation, described by a black-body curve, peaks at some frequency that depends on the temperature of the source. Nonthermal synchrotron radiation, by contrast, is most intense at low frequencies. It is independent of the temperature of the emitting object. Compare this figure with Figure 25.1.

The radiation produced in this way--called synchrotron radiation--is nonthermal in nature. There is no link between the emission and the temperature of the radiating object, so the radiation is not described by a black-body curve. Instead, its intensity increases with decreasing frequency, as shown in Figure 25.14 (b). This is just what is needed to explain the overall spectrum of radiation recorded from active galaxies (compare Figure 25.14b with Figure 25.1). Observations of the radiation received from the jets and radio lobes in active galaxies are completely consistent with this process.

Eventually, the jet is slowed and stopped by the intergalactic medium, the flow becomes turbulent, and the magnetic field grows tangled. The result is a giant radio lobe, like those pictured in Figures 25.6-25.8, emitting virtually all of its energy in the form of synchrotron radiation. Even though the radio emission comes from an enormously extended volume of space that dwarfs the visible galaxy, the source of the energy is still the (relatively) tiny accretion disk--a billion billion times smaller in volume than the radio lobe--lying at the galactic center. The jets serve merely as a conduit to transport energy from the nucleus, where it is generated, into the lobes, where it is finally radiated into space.

The existence of the inner lobes of Centaurus A and the blobs in M87's jet imply that jet formation may be an intermittent process. There is also evidence to suggest that much, if not all, of the activity observed in nearby active galaxies could have been sparked by recent interaction with a neighbor. Many nearby active galaxies (Centaurus A, for example) appear to have been "caught in the act" of interacting with another galaxy, suggesting that the fuel supply can sometimes be turned on by a companion. Just as tidal forces can trigger star formation in starburst galaxies (mentioned in the previous chapter), they may also divert gas and stars into the galactic nucleus, triggering an outburst that may last for millions or even billions of years.

More on Gravitational Lenses


Back In the early days of radio astronomy, many radio sources were detected for which no corresponding visible object was known. By 1960, several hundred such sources were listed in the Third Cambridge Catalog, and astronomers were scanning the skies in search of optical counterparts. Their job was made difficult both by the low resolution of the radio observations (which meant that the observers did not know exactly where to look) and by the faintness of these objects at visible wavelengths.

In 1960, astronomers detected what appeared to be a faint blue star at the location of the radio source 3C 48 (the 48th object on the Cambridge list) and obtained its spectrum. Containing many unknown broad emission lines, the unusual spectrum defied interpretation. 3C 48 remained a unique curiosity until 1962, when another similar-looking, and similarly mysterious, faint blue object with "odd" spectral lines was discovered and identified with the radio source 3C 273. Several of these peculiar objects are shown in Figure 25.15.

Figure 25.15 (a) Optical photograph of 3C 275, one of the first quasars discovered. Its starlike appearance shows no obvious structure and gives little outward indication of this object's enormous luminosity. However, 3C 275 has a much larger redshift than any of the other stars or galaxies in this image; it is about 2 billion parsecs away. (b) Optical image of a field of quasars (marked), including QSO 1229+204, one of the most powerful quasars yet discovered, shown enlarged in (c). Its starlike appearance shows only a hint of structure, and gives little outward indication of the object's enormous luminosity. Like 3C 275, its distance from the Earth is about 2000 Mpc.

The following year saw a breakthrough, when astronomers realized that the strongest unknown lines in 3C 273's spectrum were simply familiar spectral lines of hydrogen redshifted by a very unfamiliar amount--about 16 percent! This large redshift indicated a recessional velocity of about 48,000 km/s. Figure 25.16 shows the spectrum of 3C 273. Some prominent emission lines, and the extent of their redshift, are marked on the diagram. Once the nature of the strange spectral lines was known, astronomers quickly found a similar explanation for the spectrum of 3C 48. Its 37 percent redshift implied that it is receding from the Earth at almost one-third the speed of light.

Figure 25.16 Optical spectrum of the distant quasar 3C 273. Notice both the redward shifts and the widths of the three hydrogen spectral lines marked as H, H, and H. The redshift indicates the quasar's enormous distance. The width of the lines implies rapid internal motion within the quasar itself. (Note that, in this figure, red is to the right and blue is to the left.)

These huge speeds mean that neither of the two objects can possibly be members of our Galaxy. Applying the Hubble law (with our adopted value of the Hubble constant H0 = 75 km/s/Mpc), we obtain distances of 640 Mpc for 3C 273 and 1300 Mpc for 3C 48. Clearly not stars (with such enormous redshifts), these objects became known as quasi-stellar radio sources ("quasi-stellar" means "starlike"). The term has been shortened to quasars. We now know that not all such highly redshifted, starlike objects are strong radio sources, so the term quasi-stellar object (or QSO) is more common today. However, the name quasar persists, and we will continue to use it here.


The most striking characteristic of the several hundred quasars now known is that their spectra all show large redshifts, ranging from 0.06 up to the current maximum of about 4.9. (See the More Precisely feature below for an explanation of the meaning of redshifts greater than 1.) Thus, all quasars lie at large distances from us--the closest is 240 Mpc away, the farthest nearly 4700 Mpc (according to Table 25-1). The majority of quasars lie more than 1000 Mpc from the Earth. We see most quasars as they existed long ago--they represent the universe as it was in the distant past.

Thus, despite their unimpressive optical appearance--see for example, Figure 25.17, which compares a quasar and a spiral galaxy that happens to lie close to it on the sky--the large distances implied by quasar redshifts mean that these faint "stars" are in fact the brightest known objects in the universe! 3C 273, for example, has a luminosity of about 1040 W. More generally, quasars range in luminosity from around 1038 W--about the same as the brightest radio galaxies--up to nearly 1042 W. A value of 1040 W, comparable to 20 trillion Suns or 1000 Milky Way Galaxies, is fairly typical. Thus quasars outshine the brightest normal and active galaxies by about a factor of 1000.

Figure 25.17 Although quasars are the most luminous objects in the universe, they are often rather unimpressive in appearance. In this optical image, a distant quasar (marked by an arrow) is seen close (on the sky) to a nearby spiral galaxy. The quasar's much greater distance makes it appear much fainter than the galaxy.

Quasars display many of the same general properties as active galaxies. Their radiation is nonthermal, and some show evidence of jets and extended emission features (although few quasars are more than a ball of luminous fuzz in visible-light images). Figure 25.18 is an optical photograph of 3C 273. Notice the jet of luminous matter, reminiscent of the jet in M87, extending nearly 3 kpc from the quasar itself. Often, as shown in Figure 25.19, quasar radio radiation arises from regions lying beyond the bright central core, much like the core-halo and lobe radio galaxies studied earlier. In other cases, the radio emission is confined to the central optical image. Quasars have been observed in the radio, infrared, optical, ultraviolet, and X-ray parts of the electromagnetic spectrum, and some have even been found to emit gamma rays. However, most quasars emit most of their energy in the infrared.

Figure 25.18 (a) The bright quasar 3C 273 displays a luminous jet of matter, but the main body of the quasar is starlike in appearance. (b) The jet extends for about 30 kpc, and can be seen better in this high-resolution image.

Figure 25.19 Radio image of the quasar 2300-189 showing radio jets feeding faint radio lobes. The bright (red) central object is the quasar, some 400 Mpc away.

Interestingly, in addition to their own strongly redshifted spectra, many quasars also show additional absorption features that are redshifted by substantially less than the lines from the quasar itself. For example, the quasar known as PHL 938 has an emission-line redshift of 1.955, placing it at a distance of some 3400 Mpc, but it also shows three sets of absorption lines, with redshifts of 1.949, 1.945, and 0.613, respectively. The first two sets may well come from high-speed gas within the quasar itself (the velocity differences are only a few hundred kilometers per second), but the third is interpreted as arising from intervening gas that is much closer to us (only about 1700 Mpc away), which explains why it has a smaller redshift than the quasar itself. The most likely possibility is that this gas is part of an otherwise invisible galaxy lying along the line of sight. Quasar spectra, then, afford astronomers a means of probing previously undetected parts of the universe.

Many quasars have been observed to vary irregularly in brightness over periods of months, weeks, days, or (in some cases) even hours, in many parts of the electromagnetic spectrum. The same reasoning as we used earlier for active galaxies leads to the conclusion that the region generating the energy must be very small--not too much larger than our solar system in some cases. The More Precisely feature below discusses another curious aspect of quasar variability that may have much to tell us about the details of the energy source.

Figure 25.20 illustrates how the light history for one quasar reveals evidence for variations in its optical radiation. In part (a), a 1937 photograph clearly shows a starlike object (marked by arrows). It has since been identified as a quasar, labeled 3C 279. In part (b), taken in 1976, the quasar has nearly disappeared. Part (c) shows the light history of 3C 279, based on photographs taken since 1930. 3C 279's great distance and measured apparent brightness imply that, in 1937, this faint speck of light was intrinsically one of the most luminous objects ever observed in the universe. At that time, its luminosity exceeded 1041 W. Two years later, its brightness had dropped by almost a factor of 250, making it "merely" 10 times brighter than the brightest radio galaxies.

Figure 25.20 (a) Quasar 3C 279 (at the intersection of the arrows), whose luminosity in 1937 made it the most intrinsically brilliant object known in the universe. (b) The same quasar in 1976, when its luminosity was much diminished. Part (c) shows this quasar's optical variations since 1930. Its 1937 outburst later gained 3C 279 a (temporary) place in the Guinness Book of Records for the greatest absolute brightness of any known cosmic object.


Back Quasars exhibit all of the properties described earlier for active galaxies--large luminosities, nonthermal emission, jets, lobes, and rapid variability (implying small size). In many respects, a quasar looks like a "souped-up" active (Seyfert or radio) galaxy, so it should come as no surprise that the best current explanation of the quasar engine is basically a scaled-up version of the mechanism powering lower-luminosity active galaxies--accretion onto a supermassive black hole residing at the galactic core.

A 108- or 109-solar-mass black hole can emit enough energy to power even the brightest (1038 W) radio galaxy by swallowing stars and gas at the relatively modest rate of 1 star every 10 years. To power a 1040 W quasar, which is 100 times brighter, the hole simply consumes 100 times more fuel--10 stars per year. The "reprocessing" mechanisms that convert the quasar's power into the radiation we actually detect--namely, the ejection of matter in jets and lobes and the reemission of radiation by surrounding gas and dust--probably operate in much the same manner as the mechanisms we described earlier for Seyferts and radio galaxies. The most likely explanation for the large luminosities of quasars is simply that there was more fuel available at very early times, perhaps left over from the formation of the galaxies in which the quasars reside. At the distances of most quasars, the galaxies themselves cannot be seen. Only their intensely bright nuclei are visible from the Earth.

In this picture, the brightest known quasars devour about 1000 solar masses of material every year. A simple calculation indicates that if they kept up this rate of energy production for the roughly 10-billion-year age of the universe, a total of 1013 stars would have to be destroyed. Unless the galaxies housing quasars are much larger than any other galaxy we know of, most of the quasar's parent galaxy would be completely consumed, and the universe should contain many 1013-solar-mass black holes--"burned-out" quasars. We have no evidence for the existence of any such objects. One way around this problem is to suppose that a quasar spends only a fairly short period of time in this highly luminous phase--perhaps a few tens of millions of years. There is theoretical evidence to suggest that black holes tend to eat out "cavities" at the centers of their host galaxies, effectively cutting off their fuel supply through their own greed. Alternatively, as with nearby active galaxies, the high luminosities may be the result of interactions between galaxies in the early universe. The fact that quasars have been observed in some distant galaxy clusters argues in favor of this latter view. For a radically different view of quasars, however, see Interlude 25-2 below.


In 1979, astronomers were surprised to discover what appeared to be a binary quasar--two quasars with exactly the same redshift and very similar spectra, separated by only a few arc seconds on the sky. Remarkable as the discovery of such a binary would have been, the truth about this pair of quasars turned out to be even more amazing. Closer study of the quasars' radio emission revealed that they were not in fact two distinct objects. Instead, they were two separate images of the same quasar! Optical views of such a so-called twin quasar are shown in Figure 25.21.

Figure 25.21 (a) This "twin" quasar (designated AC114 and located about 2 billion parsecs away) is not two separate objects at all. Instead, the two large blobs (at upper left and lower right) are images of the same object, created by a gravitational lens.These two "L" shaped blobs of light have striking symmetry. The lensing galaxy itself is probably not visible in this image--the two objects near the center of the frame are thought to be unrelated galaxies in a foreground cluster. (b) A larger perspective of AC114 in the nighttime sky.

What could produce such a "doubling" of a quasar image? The answer is gravitational lensing--the deflection and focusing of light from a background object by the gravity of some foreground body. In Chapter 23 we saw how lensing by brown dwarfs or other compact objects in the halo of the Milky Way may temporarily cause the light from a distant star to be amplified, allowing astronomers to detect otherwise invisible stellar dark matter in our Galaxy. In the case of quasars, the idea is the same, except that the foreground lensing object is an entire galaxy or galaxy cluster, and the deflection of the light is so great (an arc second or so) that several separate images of the quasar may be formed, as illustrated in Figure 25.22. About a dozen likely gravitational lenses are known. Figure 25.23 shows an image of a lensed system in which four images of the same quasar can be seen, neatly arranged around the central image of the lensing galaxy.

Figure 25.22 When light from a distant object passes close to a galaxy or cluster of galaxies along the line of sight, the image of the background object (here, the quasar) can sometimes be split into two or more separate images (here, A and B). The foreground object is a gravitational lens.

Figure 25.23 (a) The "Einstein Cross," a multiple imaged quasar. In this Hubble view, spanning only a couple of arc seconds, four separate images of the same quasar have been produced by the galaxy at the center. (b) An artist's conception of what might be occurring here, with the Earth at right and the distant quasar at left.

The existence of these multiple images provides astronomers with several useful observational tools. First, the lensing tends to amplify the light of the quasar, making it easier to observe. Second, because the light rays forming the images usually follow paths of different lengths, there is often a time delay, ranging from several days to several years, between them. This delay provides advance notice of explosive events, such as sudden flare-ups in the quasar's brightness--if one image flares up, astronomers know that in time the others will too, so they have a second chance to study the event. The time delay also permits astronomers to determine the distance to the lensing galaxy by carefully timing the measurements. If enough lenses can be found, this method may provide a reliable alternative means of measuring the Hubble constant that is independent of any of the techniques discussed in Chapter 24.

Third, so-called microlensing--lensing by individual stars in the foreground galaxy--can cause large fluctuations in a quasar's brightness. Microlensing allows astronomers to study the stellar content of the lensing galaxy.

Finally, by studying the lensing of background quasars and galaxies by foreground galaxy clusters, astronomers can obtain a better understanding of the distribution of dark matter in those clusters, an issue that has great bearing on the large-scale structure of the cosmos, as we will see in the next chapter. Figure 25.24 shows the images of some faint, blue background galaxies (see Chapter 24) bent into arcs by the gravity of a nearby galaxy cluster. The degree of bending allows the total mass of the cluster (including the dark matter) to be measured.

Figure 25.24 (a) The blue arcs (marked above and to the right of the central yellowish region) are images of young, blue background galaxies, warped by the gravitational field of a foreground galaxy cluster, which deflects their light and distorts their appearance. By measuring the extent of this distortion, astronomers can estimate the mass of the intervening cluster. (b) This spectacular example of gravitational lensing shows more than a hundred faint arcs from very distant galaxies. The wispy pattern spread across the intervening galaxy cluster (A2218, several billion parsecs distant) resembles a spider's web, but it is really an illusion caused by the gravitational field of the lensing cluster.

Dressing the Naked Quasars

In Chapters 23 and 24 we addressed the issue of evolutionary change among normal galaxies. Let us now briefly consider the possibility of evolutionary links among active galaxies and between normal and active galaxies. We emphasize that this section is really mostly speculation. Although the consensus is that galaxies began to form about 8 billion years ago and that quasars were an early stage of galaxy evolution, the details of the connections among different types of active and normal galaxies are still very uncertain.

Most quasars are very distant, indicating that they were more common in the past than they are today. At the same time, "normal looking" galaxies seem to be less common in the distant past. These two pieces of evidence suggest to many astronomers that, when galaxies first formed, they probably were quasars. This opinion is strengthened by the fact that the same black-hole energy-generation mechanism can account for the luminosity of quasars, active galaxies, and the central regions of normal galaxies like our own. Large black holes do not simply vanish, at least in the 10-20 billion years that the universe has existed (see the More Precisely feature on p. 477). Thus, the presence of supermassive black holes in the centers of many, if not all, normal galaxies is consistent with the idea that they started off as quasars, then "wound down" to become the relatively quiescent objects we see today.

In this picture, the gradual reduction in violence from a quasar to a Seyfert galaxy to a normal spiral, for example, occurs primarily because the fuel supply is reduced as the galaxy evolves. A similar sequence might connect quasars to BL Lac objects (Interlude 25-1 above) to radio galaxies to normal ellipticals. These possible evolutionary connections among the active and normal galaxies are illustrated in Figure 25.25. Adjacent objects along this sequence are nearly indistinguishable from one another. For example, weak quasars share some characteristics with some very active galaxies, and the feeblest active galaxies often resemble the most explosive normal galaxies.

Figure 25.25 A possible evolutionary sequence for galaxies, beginning with the highly luminous quasars, decreasing in violence through the radio and Seyfert galaxies, and ending with normal spirals and ellipticals. The central black holes that powered the early activity are still there at later times; they simply run out of fuel as time goes on.

If we accept this appealing (but still unproven) view, we can construct the following possible scenario for the evolution of galaxies in the universe: Galaxies formed about 8 billion years ago. The early round of massive star formation that may have expelled galactic gas and helped determine a galaxy's Hubble type--spiral or elliptical--could also have given rise to many large, stellar-mass black holes, which sank to the center of the still-forming galaxy and merged into a supermassive black hole there. Alternatively, the supermassive hole may have formed directly by gravitational collapse of the dense central regions of the protogalaxy. Whatever the cause, large black holes appeared at the centers of many galaxies at a time when there was still plenty of fuel available to power them, resulting in many highly luminous quasars. The brightest quasars--the ones we see from Earth--were those with the greatest fuel supply. The young galaxies themselves were so faint compared with their bright quasar cores that we simply cannot see them. (Unfortunately, this last point is still conjecture, and recent Hubble observations have failed to clarify the issue. One group of astronomers working with the telescope failed to detect the expected galactic "fuzz" around several quasars. However, another group, working with a different on-board camera, has apparently seen clear evidence for host galaxies.)

As the galaxy developed and the black hole used up its fuel, the luminosity of the central nucleus diminished. While still active, it no longer completely overwhelmed the emission from the surrounding stars. The result was an active galaxy--a radio galaxy or a Seyfert--still emitting a lot of energy, but now with a definite "stellar" component in its spectrum.

The central activity continued to decline. Eventually, only the surrounding galaxy remained visible--a normal galaxy, like the majority of those we now see around us. Today, the black holes that generated so much youthful energy lie dormant in galactic cores, producing only a relative trickle of radiation. Occasionally, two nearby normal galaxies may interact with one another, causing a flood of new fuel to be directed toward the central black hole of one or both. The engine starts up for a while, giving rise to the nearby active galaxies we observe.

Should this picture be correct, then many normal galaxies, including perhaps our own Milky Way Galaxy, were once brilliant quasars. Perhaps some alien astronomer, thousands of megaparsecs away, is at this very moment observing our Galaxy--seeing it as it was billions of years ago--and is commenting on its enormous luminosity, nonstellar spectrum and high-speed jets, and wondering what exotic physical process could possibly account for its violent activity!

When they were first discovered, active galaxies and quasars seemed to present astronomers with insurmountable problems. For a time, their dual properties of enormous energy output and small size appeared incompatible with the known laws of physics and threatened to overturn our modern view of the universe. Yet the problems were eventually solved, and the laws of physics remain intact. Far from jeopardizing our knowledge of the cosmos, these violent phenomena have become part of the thread of understanding that binds our own Galaxy to the earliest epochs of the universe we live in.