Astronomy Supplement 21.

Galaxies


Latest Modification: October 26, 1998

Table of Contents


Although our Galaxy is home to us, it is an extremely small entity in the vastness of the Universe. It is only one of billions of galaxies, which are the building blocks or basic units of which the Universe is composed. Like stars in our Galaxy that form binary systems and clusters, galaxies also can and do exert sufficient gravitational attraction to be bound to one another permanently. Consequently, galaxies occur in pairs, small groups of a few tens, great clusters of hundreds to thousands, and superclusters, which are composed of a number of clusters, small groups, and individual galaxies. Galaxies constitute the bulk of the matter that is emitting visible light. And through most of this century, the visible galaxies in all their diversity were thought to contain the bulk of the Universe' matter. However, evidence is growing, as we shall see in later chapters, that the visible galaxies may be but a small percentage of all the matter in the Universe. In any regard, galaxies and their organization in the form of clusters and superclusters lead us to a consideration of the largest scales of structure in the Universe.

The place to begin our discussion of galaxies is with their classification which was done in the first half of this century.

21.1. Hubble's Classification of Galaxies

21.1.1. Naming the Galaxies

Astronomers had observed, described, and cataloged small, nebulous patches and bright knots of light even before the advent of photography. The French comet hunter Charles Messier (1730-1817), in order to avoid confusion with the comets for which he was searching, carefully described these objects and assembled them in a catalog (Appendix 4). Completed in 1781, the catalog, known as the Messier Catalogue, listed 103 objects that we know today as galaxies, star clusters, and gaseous nebulae. Between 1888 and 1908, Danish astronomer John Dreyer (1852-1926) compiled the New General Catalogue (NGC), the most comprehensive of the older catalogs still used, and two supplemental Index Catalogues (IC). The catalog numbering system is simple: M13 is Messier's thirteenth catalog entry, for the globular cluster in Hercules, which is NGC 6205, or entry 6205 in Dreyer's New General Catalogue. Besides their catalog designations, a number of galaxies are identified by a proper name, such as the Andromeda galaxy or the Large and Small Magellanic Clouds. Somewhat less confusion exists in identifying galaxies than in identifying stars.

The New General Cataloque and the two Index Cataloques list about 13,000 star clusters, planetary nebulae, diffuse nebulae, and galaxies. A more recent catalog includes about 200,000 extended optical sources, which are galaxies, star clusters, nebulae of all descriptions, or quasistellar objects. Even this number is dwarfed by the total number, running into the hundreds of millions or billions, that can be photographed with modern telescopes.

21.1.2 The Hubble Sequence

From his extensive collection of photographs, Hubble chose about 600 well-defined bright galaxies on which to base a classification scheme for galaxies. He arranged them in an orderly progression, which we now call the Hubble sequence. Shaped like a tuning fork, his sequence ran from essentially spherical configurations through lens-shaped systems to very flat spiral systems to irregular ones, as shown in Figure 21.1.

[Figure 21.1]

Today astronomers, having elaborated on Hubble's original scheme, classify galaxies according to several criteria:

With these characteristics in mind, let us look at the four major classes of galaxies in the Hubble sequence: elliptical (E), lenticular (SO), spiral (S and SB), and irregular (Irr).

[Biography - Edwin Powell Hubble (1889-1953)]

21.2. Types of Galaxies

21.2.1. Elliptical Galaxies

Elliptical galaxies are distinguished from the other three classes in that they are characterized by a very smooth and symmetrical texture (with no evidence for internal structure) and are tightly bound, like NGC 4486 in Figure 21.2. That is, they possess a bright center from which the brightness fades in a reasonably uniform fashion; in short photographic exposures they can appear quite small, whereas in longer exposures they can be enormously larger. They do not show evidence for a disk, or plane, containing spiral structure as in our own Galaxy, which is characteristic of the spiral class. Elliptical galaxies are designated by an E followed by a number that increases from 0 to 7 as the shape appears more elongated, that is, spheroidal to ellipsoidal.

[Figure 21.2]

Ellipticals progress in size from dwarf systems of tens of millions of stars up to giant systems containing trillions of stars covering a diameter that is thousands of times that of the dwarf ellipticals. In elliptical galaxies, the constituent stars are analogous to the spheroidal population stars (old stars) in our Galaxy, and most such galaxies do not have the pronounced dust streak across their middle that is so prevalent in spirals seen edge on (an exception is NGC 5266 in Figure 21.3). Astronomers interpret these two observations as evidence that star formation has virtually ceased in most elliptical galaxies; that is, there are no stars analogous to the disk population stars of our Galaxy.

[Figure 21.3]

The closest ellipticals are reddish in color because their brightest stars are the older spheroidal population stars, such as globular cluster stars. Dwarf ellipticals have a stellar content similar to that of giant ellipticals, but their stars seem to contain less of the heavy elements than do the stars in giant ellipticals. Dwarf ellipticals are also a much less compact grouping of stars compared with their larger counterparts. It is interesting to note that if one considers only the spheroidal stars in our Galaxy, they form a collection of stars that is very much like an elliptical galaxy in the nature of its stars and its size, except that an elliptical galaxy contains many more stars.

Rotation rates have been measured for some elliptical galaxies, and it appears that as a class they are slow rotators compared with spiral galaxies. For some time, most astronomers had thought that the more nearly spherical ellipticals were probably the slowest rotators, whereas the more pancakelike ones rotated faster: It was the rate of rotation that caused flattening in an elliptical galaxy. However, recent data indicate that there is no significant difference between spherical and ellipsoidal galaxies in rate of rotation. Thus their shape may not be the result of rotation, so this still remains a perplexing question.

If one were to consider the 1000 or so brightest galaxies, one would find that most are spiral galaxies (to be discussed below). Since the dwarf elliptical galaxies are so faint and difficult to see at great distances, they might be expected to be the most numerous. Yet a study of all galaxies within 30 million ly (Mly), a distance 600 times the radius of our Galaxy, found that only about 20 percent are elliptical, including dwarf ellipticals. The percentage of each class of galaxy among all the galaxies in the Universe is an important question yet to be answered.

21.2.2. Lenticular Galaxies

A type of galaxy that seems to represent a transition between elliptical and spiral galaxies is the lenticular galaxy (S0). Lenticular galaxies, such as NGC 1201 and NGC 2859 in Figure 21.4, possess a smooth, abbreviated, disklike extension out from a nuclear bulge that shows no sign of spiral structure. Seen edge on, lenticular galaxies appear reasonably flat and are generally without the dark streak indicating the presence of interstellar dust. In contrast to S0 galaxies, SB0 galaxies, such as NGC 2859, exhibit a stump, or faint suggestion of a disk without spiral arms, in place of a clearly defined bar, and such galaxies are often surrounded by a faint halo or ring structure. Unlike the elliptical galaxies, which as a class are slow rotators, S0 and SB0 galaxies rotate more rapidly, as do the spiral galaxies. S0 galaxies contain a disk population group of stars like those of our Galaxy in addition to a spheroidal population in the nuclear bulge, which is like the elliptical galaxies.

[Figure 21.4]

21.2.3. Spiral Galaxies

Among the conspicuous galaxies in the neighborhood of our Galaxy, most are spirals. Two distinct kinds are recognizable: One is the normal spiral (S) galaxy, as shown in Figure 21.6, and the other, which is somewhat less numerous, is the barred spiral (SB) galaxy, as shown in Figure 21.5. Normal spirals generally have two (sometimes more) arms emerging from opposite sides of a nucleus and winding through a disk in much the same fashion as that observed for our Galaxy. Extending out of opposite sides of the nucleus in barred spirals is a barlike group of stars, from whose ends spiral arms emerge, usually one from each end. Like elliptical galaxies, spirals are brightest at the center, with a gradual fading in brightness radially out through the disk and a very rapid fading perpendicular to the disk.

[Figure 21.5]

Normal and barred spirals vary in size, as do elliptical galaxies, but they do not exhibit the enormous differences in size and mass that are observed for ellipticals. The largest spiral recognized has a visible diameter of almost 500,000 ly, about 5 times that of our Galaxy and the Andromeda galaxy, M31 (Figure 23.1). There are recognizable categories of both giant and dwarf spirals.

Both normal and barred spirals are graded by the lower-case letters a, b, c, and occasionally d for the tightness of the arms (a, tight, to c, open) and the size of the nucleus (a, large, to c, small). That is, the more tightly wound arms generally go with the largest nuclei, and the loosely wound arms go with the smallest nuclei, as depicted in Figure 21.1. Our Galaxy appears to be a normal Sb spiral.

In the nuclei of spiral galaxies, both normal and barred, the brightest stars are normally red giants or supergiants, whereas in the arms the brightest stars are the highly luminous spiral-arm population stars that are blue in color. The brightest objects in the spiral arms are the H II emission nebulae that are interspersed with O associations, and both are strung along the arms like brilliant beads in subclasses b and c. Obscuring dust lanes usually lie on the inner edges of spiral arms, as can be seen in M51 in the frontispiece to this chapter. In spirals seen edge on, such as NGC 4565 in Figure 21.4, this opaque interstellar dust is clearly evident as a long, dark streak threading across a galaxy's middle. Figure 21.6 shows the spiral galaxy NGC 4622 in a face-on attitude so that the long, well-wrapped arms are clearly evident. Subclass a has a somewhat smooth and unbroken texture made of older disk population stars and sparse occurrences of H II emission nebulae or O associations. The halo is more conspicuous in subclass a than it is in c.

[Figure 21.6]

Just as spirals in visible light can be seen to possess varying amounts of interstellar dust, radio observations reveal that they also contain varying amounts of interstellar gas, primarily atomic and molecular hydrogen, and some interstellar molecules. As in our Galaxy, we have been able to detect atomic hydrogen in nearby spirals by its emission of 21-cm radiation and at the appropriate wavelengths emission from such molecules as carbon monoxide and water. Whereas visible light from stars in spiral galaxies is centrally concentrated, the 21-cm radiation from atomic hydrogen (H I) is not centrally concentrated but instead shows a pronounced minimum, or hole, at the galaxy's center (somewhat larger than the size of the nucleus). In our own Galaxy, the location of atomic hydrogen is correlated with the location of spiral arms seen in visible light. This same situation more or less occurs in other spirals. We can summarize the presence of interstellar matter by saying that the amount increases from subclass a to subclass c.

Just as globular clusters have been observed in elliptical galaxies, they have also been discovered in the halos of nearby spiral galaxies. This raises the question of whether or not other spirals besides our Galaxy show evidence for a massive corona. The strongest argument for such coronae is that the rotational velocity measured for several spirals does not decline in the outer reaches of the visible disks as it should if most of the mass lies toward the nucleus of galaxies, but instead is approximately constant out to the edge of the visible disks. This is interpreted as meaning that there are significant amounts of mass at large distances from the nucleus. It is assumed that this mass is very faint matter contained in roughly spherical coronae that engulf the visible galaxies and whose radii are several times that of the visible galaxies.

No less a significant question for spiral galaxies is why the spiral arms, which appear to trail as the galaxy rotates, do not wind up (this question was discussed in relation to our Galaxy in Section 20.4). Of the two mechanisms proposed for maintaining spiral structure--the spiral-density-wave theory and the succession of supernova outbursts--most of the evidence supports the spiral-density-wave theory. It is more likely to produce symmetrical two-armed spirals such as M51 (chapter opener) or NGC 4622 (Figure 21.6), whereas the supernova-outburst theory should explain a more chaotic or ragged arm structure, such as that for M33 in Figure 28.18.

As a means of summarizing the structure of spirals, Figure 21.7 shows schematically how a normal spiral galaxy, such as our own Galaxy, is composed of different subunits. Although the figure is somewhat idealized, it is a reasonable representation of the composition of spiral galaxies.

[Figure 21.7]

21.2.4. Irregular Galaxies

About 3 percent of the brightest galaxies in the sky are classified as irregular galaxies. They exhibit little symmetry, as can be seen in Figure 21.8, and can be divided into two kinds: The group Irr I contains highly luminous blue stars, star clusters, and some interstellar gas with very little dust (a prime example is the nearby Magellanic Clouds, described in Section 21.5); the second group, Irr II, is marked by more deformity in its structure, fairly conspicuous dust lanes, and a composite spectrum of unresolved stars.

[Figure 21.8]

21.3. Properties of Normal Galaxies

21.3.1. What Does The Classification of Galaxies Mean?

Hubble's classification arranged galaxies in a progressive morphological sequence, but galaxies do not necessarily evolve from one form into another. We think it likely that different forms in the sequence reflect differences in their initial situations at the time of formation and how galaxies have evolved under a variety of conditions and environments--not different stages in evolution.

What then is the Hubble sequence? It is, first, a dynamic arrangement, ordering galaxies according to the degree of rotation and structure, and it is, second, a population sequence, characterized by the progress of stellar evolution in each galaxy.

21.3.2. Distance Indicators, Inverse Square Law of Light

It is hardly surprising, in view of the enormous expanses involved, that distances to remote galaxies are at best crude estimates. Distances to neighboring galaxies, however, are fairly well established. Accuracy, of course, diminishes as the distance grows. Just as precise determination of the astronomical unit sets the scale for our Solar System, so we must have an accurate yardstick for measuring distances of galaxies. Essential cosmic data depend on how reliable this yardstick is. The linear dimensions, spatial distribution, intrinsic luminosities, and masses of the galaxies; the physical and evolutionary differences among them; the average density of matter in the Universe; the rate of expansion of the Universe; and the type of cosmological world model all depend on the correct scale of distance.

To find the distances of galaxies, we assume that similar objects in our Galaxy and in other galaxies have the same physical characteristics (an assumption whose accuracy can be checked in a variety of ways). From investigations in our own Galaxy and nearby galaxies we know the absolute luminosity of these objects (or distance indicators, as they are called). To determine the distance to a galaxy, we take the absolute luminosity of one of the distance indicators within the galaxy and its observed apparent brightness. Then, using the inverse-square law of brightness, we find the distance by way of the distance modulus (m-M), as described in Section 13.2.

Table 21.1 lists some distance indicators and the maximum distances to which they theoretically may be applied in estimating distances to galaxies with the 5.1-m Hale reflector. Except perhaps for supernovae, in practice astronomers cannot measure distances beyond roughly 50 Mly by using individual objects in a galaxy. This is so because the objects are so hard to resolve in more remote galaxies. If they cannot be resolved in a galaxy, then they cannot be used. There are a variety of things that affect the resolution, such as the orientation of a galaxy. Cepheid variables are still our most dependable criteria in finding distances. We can use them today for only the closer galaxies (less than about 15 Mly), where they can be identified readily.

[Table 21.1]

21.3.3. Other Distance Methods

Even when we cannot distinguish isolated objects in a galaxy, astronomers can often estimate the distance from the galaxy's total luminosity, which they obtain from its surface brightness, apparent size, or mass-luminosity ratio. They must be careful to recognize exactly the class of galaxy involved because galaxies differ greatly in brightness and size. Lumping all galaxies into one luminosity class would introduce large errors in the estimated distance. We can distinguish between the image of a galaxy at the threshold of visibility and a star's pointlike image by the galaxy's fuzziness or nonspherical shape. However, this still may not allow it to be classified, as the images in Figures clearly show.

Another method for determining distance is to take advantage of a correlation between the luminosity of a galaxy and the wavelength width of the 21-cm emission line produced by atomic hydrogen in the interstellar medium of the galaxy. The correlation was established for nearby galaxies of known distance and can now be used as a means of determining distances of more remote galaxies.

At large distances, the distance to a large cluster of galaxies can be determined much more accurately than that to its individual galaxies. We can select, say, the observed average apparent magnitude of the 10 brightest galaxies as a criterion for luminosity instead of depending on one galaxy. If we assume that their mean absolute magnitude is a particular value, such as M = -21, based on knowledge gained from similarly constructed nearer clusters, the distance of the cluster can be found from the inverse-square law.

21.3.4. Cosmic Distance Scale

Astronomers have set up a cosmic distance scale by building a chain of overlapping distances, proceeding from the nearest objects to the farthest. Figure 21.9 depicts this overlapping system of distance indicators and shows practical distances that are less than the theoretical limits in Table 21.1. The cosmic distance scale begins with the distances of relatively close stars set by trigonometric parallax and by the distance of the nearby Hyades open-star cluster. These distances serve as a basis for the next step, which comes from variable-star data, chiefly from the Cepheids, and from the spectroscopic and intrinsic brightness of stars in our Galaxy. This sequence in turn serves as a basis for distances of neighboring galaxies of the Local Group, which are determined from characteristics of their brightest stars, Cepheid variables, and other stellar data.

[Figure 21.9]

The next sequence uses the distances of more remote galaxies (in what astronomers refer to as "Nearby Groups"), taking as criteria their brightest stars, surface brightness of the galaxy, and the apparent size of their bright gaseous nebulae. The next link in the chain is a cluster of galaxies such as the Virgo cluster, using the cluster's brightest galaxy or its luminosity type as a standard of comparison. Finally, we connect this distance scale to distances of more remote clusters of galaxies by means of the Hubble constant derived from expansion of the Universe, which will be discussed in Section 21.4.

21.3.5. Galactic Luminosities and Masses

Galaxies have a very large range in size, brightness, and mass, as shown in Table 21.2. As noted earlier, the largest and smallest galaxies are elliptical. The true dimensions of galaxies are derived from their apparent diameters and known distances. By measuring galactic apparent magnitude and combining that with known distances, we obtain luminosity.

[Table 21.2]

The simplest way for astronomers to estimate the masses of galaxies, particularly for spirals that are seen more or less edge on, is to find the Doppler shifts of their spectral lines. From this the difference in radial velocity at opposite ends of the disk (the end approaching and the one receding along our line of sight) can be measure. With the velocity difference and the known size of a galaxy we can calculate the mass from Newton's modified form of Kepler's third law. Of course, we find only the mass within the region observed, since the mass of any matter outside this region has no effect gravitationally.

Binary galaxies, in which two galaxies orbit each other, provide another means of measuring masses of galaxies. The difference in radial velocity for several pairs, combined with their known separation, leads to a statistical estimate of their average masses through the use of Kepler's third law. This method is similar to the one we use in finding the masses of binary stars (Section 14.2).

A third method makes use of the correlation between mass and luminosity known as the mass-luminosity ratio (representative values are given in Table 21.2): The mass of a galaxy can be estimated from its observed luminosity. The mass-luminosity ratio is an important indicator of the stellar populations in galaxies. The variations in the mass-luminosity ratio reflect differences in the spectral types of stars that make up various classes of galaxies. Letting the Sun's mass divided by its luminosity be a unit of measure equal to unity, a typical mass-luminosity ratio for giant ellipticals, lenticulars, and spirals is 6, whereas for irregular galaxies it is closer to 3. A higher value means that a larger proportion of stars are faint dwarfs whose mass contribution to the galaxy is much greater than their luminosity contribution. Remember from the discussion of stars in our Galaxy that the luminosity of a galaxy is provided by the intrinsically bright stars. Thus in pictures of galaxies the distribution of light does not necessarily, although it may, reflect the distribution of mass in those galaxies.

We cannot close this section without again mentioning the growing evidence for huge amounts of dark matter in galaxies. As we mentioned when discussing our Galaxy (Section 20.5), several lines of evidence from nearby galaxies suggest that a visible galaxy may be only a small percent of the total mass of the matter composing the galaxy. The rest of the mass consists of dark or nonluminous matter in an extended halo surrounding the visible galaxy, analogous to the halo's dark component in our own system. These findings raise the question as to what should be included when one talks about the mass of a galaxy. Should one include only the matter that can be detected because it emits radiation over some region of the electromagnetic spectrum, or should one include all the matter that dynamical arguments indicate is present even though some of it apparently emits little, if any, radiation? Although no definitive answers presently exist to this and many other questions about dark matter in galaxies, we have tried to cover the range of possibilities in the numbers in Table 21.2. It is not unlikely, however, that many of the numbers, such as the mass-luminosity ratio, in this table will be revised one way or another in coming years.

21.4. Expansion of the Universe

21.4.1. Receding Galaxies, the Cosmological Redshift

In 1912, it was found that absorption lines in the composite spectrum of all the stars in the Andromeda galaxy were Doppler-shifted toward the blue, indicating a velocity of approach between that galaxy and ours. But by 1928 large redshifts in the absorption lines of all but 5 of 41 nearby galaxies had been found (the 5 having blueshifted spectra). Even larger redshifts have since been found for fainter galaxies, as one can see in Figure 21.10.

[Figure 21.10]

After Hubble succeeded in estimating distances for a number of galaxies whose redshifts had been measured, he found that a straight-line relationship existed between their redshifts interrupted as recessional velocities and their distances, in the sense that the farther away a galaxy is, the faster it is moving away from us (Figure 21.11). The only exceptions were several nearby galaxies that exhibited velocities of approach. Hubble was able to show that for the intrinsically more luminous galaxies, the recessional velocity is also correlated with its apparent brightness. This is in the general sense that the greater the recessional velocity, the fainter a galaxy appears and the more distant it is. Hence, unlike stars in our Galaxy, where those which appear brighter are not necessarily closer to us but are intrinsically bright, faint galaxies in general are truly distant ones.

[Figure 21.11]

21.4.2. Hubble's Velocity-Distance Law

If we interpret it literally, the relationship in which velocity is proportional to distance, or Hubble's velocity-distance law, indicates that the Universe is expanding (we shall discuss cosmological implications in Chapter 24). Hubble's original results have been extended by other investigators and now include thousands of more distant galaxies and several hundred distant clusters of galaxies.

Hubble's Velocity-Distance Law: The farther away a galaxy is from our Galaxy, the faster that galaxy is receding from us; that is, recessional velocity equals a constant times distance.

The redshifts of distant galaxies represents the amount the Universe has expanded since the time the galaxy's light was emitted. Think of a wave moving across the surface of a pond. If we could cause the surface of the pond to expand laterally as the wave moves across the surface, then the wavelength of the wave would grow larger as the surface area of the pond increased. Thus the galactic redshifts are the result of the expansion of the Universe: an expansion of space leading us to call them cosmological redshifts to distinguish them from Doppler or gravitational redshifts.

In addition to a recessional velocity, galaxies also exhibit a small peculiar velocity superimposed on the expansion velocity. For nearby galaxies, the peculiar velocity is larger than the recessional velocity, and thus we observe some blueshifted galaxies. However, for very distant galaxies, the recessional velocity is much larger than the peculiar velocity, and the peculiar velocity may be neglected in comparison. Hubble's law deals with the recessional velocity caused by the Universe's expansion and not a local Doppler effect producing the peculiar velocity.

In the mathematical expression of Hubble's law (v = Hr), in which the recessional velocity v is proportional to the distance r, the constant of proportionality H is called Hubble's constant. Its value is the slope of the line in Figure 21.11. In a straight-line relationship between v and r, we see that a galaxy 2 billion ly (2000 Mly) away is receding twice as fast as one that is 1 billion ly away. As we shall discuss in Chapter 23, the numerical value of Hubble's constant and any departure from a straight-line relationship at great distances are important aspects of the Universe.

Recent estimates of Hubble's constant place its value between 15 and 30 km/s/Mly (50 to 100 km/s/Mpc). When the distance indicators of Table 21.1 are unresolvable, astronomers can use Hubble's velocity-distance law for a specific value of H to estimate distances of remote galaxies from their redshifted spectral lines. For example, the largest redshifts z (= / ) yet found for supposedly normal galaxies have values equal to about 1.2, which corresponds to a recessional velocity of about 200,000 km/s. Assuming a Hubble constant of 17 km/s/Mly, these galaxies are about 11 billion ly away and we are observing them as they were 11 billion years ago. That is, over half the estimated age of the Universe has passed since light that we observe today left those galaxies. Some of the quasars to be discussed in Section 22.4 have redshifts over 3 and a few over 4, corresponding to distances of almost 15 billion ly.

[Box - Relativistic Doppler Redshifts and Cosmological Redshifts]

21.5. Peculiar and Active Galaxies

Over the last 30 or so years with the ability to carry out astronomical investigations using electromagnetic radiation in other wavelength regions besides the visible, it has become evident that some tremendously energetic events are occurring in the Universe. Most of these events occur in and around galaxies, so that we recognize that galaxies can be extremely violent objects. The existence of these active galaxies bears heavily on the question of what is a galaxy and what role it plays in the large-scale structure of the Universe.

21.5.1. Galaxies That Do Not Fit the Hubble Scheme

Not all galaxies fit into Hubble's classification scheme. In recent years, astronomers have discovered over 10,000 peculiar galaxies, which do not fit the Hubble classification scheme because of either their optical appearance or their excessive radio emission as compared with a Hubble-type galaxy. It is extremely difficult at this juncture to estimate what percentages of distant galaxies (and thus of all galaxies) are peculiar galaxies. But we can say at this point that the percentage is not insignificant.

Figures 21.12 and 21.13 show some typical examples of peculiar galaxies. Other peculiar galaxies have tidally distorted forms, many of them connected to each other by luminous bridges of stars and interstellar matter. From computer-generated models for gravitational interactions between grazing galaxies, such encounters apparently can produce streams of stars and interstellar matter curving in opposite directions from the galaxies. Many of these simulated appearances closely resemble actual forms, as shown in Figure 21.14.

[Figure 21.12]

[Figure 21.13]

[Figure 21.14]

A subgroup of the peculiar galaxies are the active galaxies, with their compact appearance and bright nuclei. Many active galaxies are strong radio emitters; some are strong X-ray and infrared emitters of the type associated with violent or energetic events. Let us look at some of the abnormalities associated with active galaxies.

21.5.2. Violent Events in the Cosmos

When radio astronomers started probing the heavens in the late 1940s, our perception of the Universe began to change abruptly from a quiet, orderly cosmos to one punctuated by extraordinarily violent events. There are galaxies with extremely active nuclei, exploding galaxies, galaxies with jets, and quasars. Because the radiation from these active galaxies may be very intense throughout, or in particular portions of, the electromagnetic spectrum, they stand out from normal galaxies. One of their most prevalent features is a continuous spectrum produced by nonthermal synchrotron radiation.

Astronomers distinguish two kinds of radiation source: thermal and nonthermal. The intensity of thermal radiation grows weaker with increasing wavelength in accordance with the blackbody distribution of energy. Most thermal radiation from a galaxy comes from its stars and is situated in the visible and ultraviolet portions of the spectrum.

Radiant energy from isolated Galactic sources, such as the Crab Nebula and other supernova remnants, or from the Galactic nucleus is primarily nonthermal. Unlike the intensity of thermal radiation, the intensity of nonthermal radiation grows stronger--or at least does not drop so rapidly--with increasing wavelength. Most often nonthermal radiation is what is referred to as synchrotron radiation, which is radiation emitted by free electrons spiraling around magnetic lines of force (Figure 21.15).

[Figure 21.15]

Abundant amounts of nonthermal radiation comes from many galaxies whose optical appearances and spectra are unusual, the active galaxies. Even though active galaxies possess a variety of different structural forms, in general they exhibit large redshifts in their spectra, some being among the largest found for any type of galaxy. Because active galaxies tend to be very far from our Galaxy, they must be emitting immense amounts of energy to have their observed brightnesses. Astronomers feel reasonably confident that the same energy source, whatever it is, is present in varying degrees in all of them. The most luminous active galaxies are quasars, which emit up to 10,000 times more light than does a normal galaxy such as the Milky Way. Indeed, as pointed out in Chapter 20, the same energy source as in quasars may operate at the center of our Galaxy but on a greatly reduced scale. In the following sections we shall discuss the various types of active galaxies, whose general properties are summarized in Table 21.3. The Seyfert galaxies are our first example.

[Table 21.3]

21.5.3. Seyfert Galaxies

More than 120 Seyfert galaxies have been identified since Carl Seyfert (1911-1960) first called attention in the 1940s to this subclass of spiral galaxies. These galaxies have unusually small and bright nuclei, yet with a reasonably normal looking disk (NGC 4151 is shown in Figure 21.16). Their redshifts range up to a few tenths, the largest corresponding to a fifth the velocity of light and a distance of about 1 billion ly.

[Figure 21.16]

Visible spectra of Seyfert galaxies have emission lines of such elements as hydrogen and highly ionized elements like iron with Doppler shifts that indicate that very hot matter is either swirling around or being expelled at high velocities (several thousand kilometers per second) from their nuclei. Supporting this is the fact that Seyfert nuclei are sources of intense and variable X-ray, infrared, and radio radiation.

The luminosity of Seyfert galaxies is compared with that of other active galaxies in Table 21.4. Although in visible light their absolute brightness is comparable with that for normal spirals, Seyfert galaxies are a hundred or so times brighter than normal spirals when compared in infrared and radio radiation. Seyfert galaxies are also sources of X rays, with about 25 reliable identifications among them. Figure 21.17 shows the X-ray sky (the same as Figure 19.14) with several of the extragalactic sources identified, including several Seyfert galaxies. Because the X-ray emissions vary in periods of hours or days, the X-ray emitting region must be very small. For example, the emitting region in NGC 4151 is estimated to be about twice the Earth-Sun distance. Since this Seyfert galaxy emits thousands of times more energy in the form of X-rays than does our Galaxy, the nature of the emitting region is extremely mysterious. Adding to the mystery is the fact that NGC 4151 is also a source of gamma rays.

[Table 21.4]

[Figure 21.17]

There is some thought that Seyfert galaxies may be in an early or recurrent stage through which all spirals, including our own, must pass. Thus, according to this theory rather than a few percent of all spirals' being Seyfert galaxies, all spirals spend a few percent of their lives being Seyfert galaxies.

21.5.4. N Galaxies

Originally recognized in 1958 as the optical counterparts of some powerful radio sources, N galaxies have some properties in common with Seyfert galaxies and quasars, although they are somewhat less luminous. They have a bright, sharply defined nucleus surrounded by a small nebulous envelope. Of the couple of dozen or so that have been identified, all are at great cosmological distances according to their large redshifts, which indicate distances up to about 4 billion ly. Some astronomers have pointed out that a very distant Seyfert galaxy might resemble an N galaxy. N galaxies fluctuate rapidly in brightness and color, at times within a few days, suggesting that the variable source must be less than a light week (the distance light travels in a week) in diameter. Other similarities to quasars and Seyfert galaxies are their large output of energy in the X-ray and radio regions of the spectrum.

21.5.5. BL Lacertae Objects

A small number of extremely compact galaxies that closely resemble Seyfert and N galaxies constitute a group of active known as BL Lacertae objects. They are named after their prototype, BL Lacertae, which in 1929 was incorrectly identified as a variable star. Its true nature was not revealed until 1969, when radio astronomers identified it as a very active radio source. About 40 such objects are presently known. They are characterized by a sharply defined and brilliant nucleus that emits nonthermal radiation and whose continuous visible spectrum has no emission or absorption lines. Surrounding the bright nucleus is a faint halo whose spectrum resembles that of a typical elliptical galaxy--at least this is true in the case of BL Lacertae. The spectrum of this fuzzy halo has absorption lines from which redshifts can be measured. Such measurements place some BL Lacertae objects with halos at great distances, comparable to those of the nearer quasars.

Although emissions from BL Lacertae objects are strong in all wavelengths, they radiate most of their energy in the optical and infrared wavelengths. They undergo rapid changes in brightness in visible light, the infrared, and X-rays. At peak brightness, their luminosity rivals that of the brightest quasars. Radio measurements by very-long-baseline interferometry point to a central source that is at most a few light years in diameter. This is also inferred from the rapid fluctuations in brightness. For example, the X-ray brightness varies in periods of several hours for a particular BL Lacertae object seen in the southern skies, suggesting that the emitting region is less than 6 light hours in extent. For comparison, the Pluto-Sun distance is about 5.5 light hours.

21.5.6. Classification of Radio Galaxies

Catalogs now list more than 10,000 discrete radio sources that are emitting significant amounts of energy by nonthermal radiation processes. Those which are extragalactic objects are called radio galaxies, and they emit from 10 to 100 million times more radio radiation than any normal galaxy. Radio galaxies can be divided into two groups according to their physical appearance on radio maps: compact sources and extended sources. Compact sources are very small, and most coincide with galactic nuclei or quasars (Section 21.4); most extended sources consist of two (sometimes more) immense radio-emitting regions located on opposite sides of a normal-looking elliptical galaxy. In this section we shall concentrate on the extended sources. At least four kinds of extended radio galaxies are recognized (Figure 21.18):

[Figure 21.18]

21.5.7. Radio Lobes

Positions for radio-emitting regions, referred to as radio lobes, vary with the different radio galaxies. In most radio galaxies, the radio lobes appear to be completely separate from the optical galaxy, but for several radio galaxies, optical emissions have been detected within the radio lobes or bright, threadlike features connect the optical galaxy to the radio lobes. Presumably high-energy subatomic particles are ejected along these visible features by the central source. For those radio galaxies in which one can see a sequence of jets or blobs along the lobes, astronomers infer that a number of explosions have occurred at different times. The axis along which the matter flows out of the galaxy following an explosion is probably the visible galaxy's axis of rotation, although there is no certainty that it is the rotation axis. As matter is ejected, it sweeps up magnetic fields and carries them along. These fields in turn help to confine particles and to produce synchrotron radiation, which is the characteristic signature of radio galaxies.

21.6. Quasars, A Cosmic Enigma

21.6.1. Discovery of Quasars

By 1960, positions of extragalactic radio sources could be determined with reasonable accuracy through radio interferometry, so that some compact radio sources could be identified with visible starlike objects, which astronomers named quasistellar radio sources, popularly called quasars. Quasars, which resemble ordinary stars in photographs, had been photographed many times before attention was drawn to them, but no one had suspected how peculiar they were. The first recognition of their unusual nature came in 1960, with an object labeled 3C 48 (the forty-eighth entry in the Third Cambridge Catalog; Figure 21.19). To Allan Sandage of the Mount Wilson and Palomar Observatories, it looked like an ordinary sixteenth-magnitude star, but its spectrum, undecipherable at that time, consisted of emission lines superimposed on a continuous spectrum.

[Figure 21.19]

In 1962, Maarten Schmidt (b. 1929) made a spectrogram of a thirteenth-magnitude quasar 3C 273 (it is also shown in Figure 21.19) from positions accurately pinpointed with the 64-m radio telescope at Parkes, Australia. Within weeks he had untangled the puzzle of the optical spectrum of 3C 273, when he recognized in three emission lines the characteristic spacing of the second, third, and fourth lines of the Balmer series of hydrogen (Figure 21.20). To his amazement, they were displaced toward the red end of the spectrum by a large redshift, corresponding to a velocity of 16 percent of the velocity of light. With this key to the optical spectrum in hand, spectrograms of other quasars were soon decoded, revealing even larger redshifts.

[Figure 21.20]

21.6.2. Appearance and Distribution

More than 1500 quasars have now been identified, and the number continues to grow. Although our knowledge about quasars has increased greatly, they are still very enigmatic with their starlike optical appearance and large redshifts. Work over the last 25 years has shown that fewer than 10 percent of the quasistellar radio sources exhibit strong radio emissions. The majority are not intense radio sources, but they are nevertheless still referred to as quasars. The so-called radio-quiet quasars can be identified by their unusually bright ultraviolet emissions and large redshifts. Many quasars exhibit emissions that vary in periods of days or weeks in one or several regions of the spectrum.

The redshifted spectral lines of quasars indicate that they are very distant cosmologically and thus must emit extraordinarily large amounts of radiant energy to be as bright as they appear in the sky. In the radio region, quasars emit as much energy as the brightest radio galaxy, whereas in visible light, they are more luminous than the brightest giant elliptical galaxies. Quasars outshine all other types of galaxies in gamma rays, X-rays, and infrared radiation. As shown in Table 21.4, luminosities of quasars can be the equivalent of 100,000 billion Suns, or a thousand times that of normal galaxies. However, their brightness variations make them the striking cosmic objects that they are, because such variations indicate that these immense radiation sources are located in volumes as small as a light day across.

Since quasars are in general quite distant objects (assuming their redshifts are cosmological), there is speculation that, like Seyfert galaxies, which possess smaller redshifts, quasars may be the brilliant nuclei of faint galaxies and that it is because of their great distances that we are unable to resolve the surrounding galaxy. In a few of the nearer quasars, some nebulous extensions have been photographed whose faint starlike spectra have the same redshifts as the quasars themselves. This provides some evidence for the idea that visible quasars are overly luminous centers of galaxies. Figure 21.21 shows in negative photographic reproduction the quasar 3C 273, whose redshift of 0.16 corresponds to about 3 billion ly. The quasar, marked Q, seems to have not only a faint, nebulous halo but also a jet to the lower right. It is hoped that Space Telescope, which will be able to resolve objects 50 times smaller in angular size than the best ground-based telescope can, will permit astronomers to decide whether or not most quasars are the nuclei of very distant galaxies.

[Figure 21.21]

Despite their starlike appearance in visible light, radio images of quasars are frequently large, structured, and noncircular. Quite a few quasars have a radio structure composed of minute discrete components, whereas others have extended radio lobes on either side of the optical image, somewhat like radio galaxies. Measurements by very-long-baseline radio interferometry disclose that in some quasars there is apparently a rapid separation of close, minute sources that--if interpreted literally--is occurring at velocities greater than the velocity of light. Figure 21.22 shows a radio image of 3C 273 in which, over a 3-year period, first one and then a second knot of radio emissions can be seen separating from the main lobe of radio emissions; the separation velocity appears to be about eight times the velocity of light. However, as discussed in Chapter 16, relativity is predicted on the assumption that neither mass nor energy can be transported at speeds in excess of the velocity of light, and the number of observational tests of relativity is sufficient that its foundations are secure. Therefore, the problem lies not with relativity, but with the interpretations of quasar observations. A number of alternative interpretations have been advanced, but so far none is without some difficulty in explaining these bizarre observations.

[Figure 21.22]

The only way to estimate the distances of remote quasars is to use Hubble's redshift-distance law. Sample counts of quasars show that their numbers increase much faster with distance than would be the case if their distribution in space were uniform (Figure 22.11). A plausible interpretation is that quasars were formed in large numbers within a couple of billion years after the Universe began to expand. The Universe may then have had a thousand times more quasars than it does now. Of those quasars within the visible Universe, most would by now have evolved into normal galaxies. Therefore, only very remote quasars, which for us represent an earlier epoch in the Universe, would still be observable.

21.6.3. Mystery of Quasar Size and Distance

Quasars are striking in that an incredible flood of energy gushes out of sources no bigger than a fraction of a light year in diameter. Spectra of the visible portion of their radiation possess broad emission lines of familiar ionized elements (such as carbon, magnesium, oxygen, neon, silicon, and helium) and of hydrogen--all overlying a continuous spectrum. The emission spectrum apparently originates in hot gases surrounding the source of energy, and as noted, the lines are highly redshifted. We can identify the lines from their relative spacings, which are the same as those in the spectra of planetary nebulae, novae, stellar coronae, and other very hot sources of radiation. Measured redshifts for quasars range from a few percent up to 94 percent (z = 4.43) of the velocity of light. For the quasar with a redshift of 4.43, its distance from Hubble's law is 6 billion ly, if Hubble's constant is 30 km/s/Mly, and 12 billion ly, if H is 15 km/s/Mly. Since 1986, eight quasars have been found with redshifts greater than 4, and efforts are being made to find even more distant quasars, if they exist.

Most quasistellar objects also have narrow absorption lines in their spectra, usually with redshifts less than or equal to those of the emission lines. An obvious interpretation is that the absorbing is done by cool shells or clouds moving outward from the quasar at different speeds up to appreciable fractions of the velocity of light. Since the shells are coming toward us, the absorption redshifts are the result of subtracting a Doppler shift (a blueshift) from the true cosmological redshift indicated by the redshifts of the emission lines. However, if the absorption is not physically related to quasars, it presumably must be the result of intergalactic absorption in extended halos of galaxies between us and the quasar. Since the absorption lines are not close to the wavelength of the emission lines in most cases but not all, the second of these hypotheses, intergalactic absorption, seems more acceptable for those cases. But for those few quasars where the emission and absorption wavelengths are close, the absorption lines are apparently produced in part of the quasar itself.

21.6.4. Are the Quasar Redshifts Cosmological?

Most astronomers believe and most of the evidence supports the interpretation that redshifts in the spectra of quasistellar objects correspond to true velocities resulting from the expansion of the Universe, a view known as the cosmological interpretation. However, a few astronomers think that these objects may not be so distant after all. If they were closer to us, their energy output could be smaller, bringing them more in line with other cosmic objects. One noncosmological interpretation, the local Doppler hypothesis, proposes that these bodies were violently expelled at high velocities from nearby normal galaxies or from the center of our Galaxy. Thus quasar redshifts are Doppler redshifts and not cosmological. If quasars were ejected from nearby galaxies, though, some should have velocities of approach. Yet no blueshifted quasars have been found to support the local Doppler hypothesis.

If we consider what appear to be strings of galaxies stretching across part of the sky that include somewhat normal galaxies, peculiar galaxies, and quasars, we have an alternative proposition that quasars have been ejected from the nuclei of abnormal galaxies. In these apparent strings very different redshifts have been measured for the quasar and allied galaxy or galaxies; redshifts may vary by as much as many thousands of kilometers per second (Figure 23.7). In each of these cases, the quasar is always redshifted more than is the allied galaxy or galaxies. (However, several quasars that appear to belong to a cluster of galaxies are known to exhibit the same redshift as the apparently normal galaxies in the cluster.) If enough discrepancies in redshift between physically related galaxies and quasars could be substantiated (for reasons other than chance alignments of bodies at different distances), the cosmological interpretation that redshifts are caused entirely by an expansion of the Universe would come into question. However, the evidence at present for noncosmological redshifts is not much more than suggestive, and the cosmological interpretation seems to be far and away the best interpretation at present for quasar redshifts.

21.6.5. Energy Source for Quasars and Other Active Galaxies

If the cosmological explanation for quasar redshifts is correct, how can we account for the vast amounts of energy pouring out of such distant but incredibly small regions? The energy is immense and is being produced more or less continuously over many millions of years. Something like a trillion solar masses of hydrogen, equivalent to 10 average-sized galaxies, would have to be converted into energy by thermonuclear fusion to equal just the radio energy being emitted by some quasars. Current thinking about the source of energy in quasars and other active galaxies centers, separately or in combination, on such physical processes as intense gravitational fields, powerful magnetic fields, immense explosions, or rapid rotation.

One of the most fascinating speculations on quasar energy sources is one proposing that the source is a rotating black hole with a mass of up to a billion solar masses. Matter, such as gas or even whole stars from crowded surroundings, falls inward toward the black hole's event horizon. In the process, rotational energy from the black hole is extracted, causing it to slow its rotation.

Many astronomers believe, however, that one of several less exotic explanations is able to account for the immense quantity of energy radiated by quasars and other active galaxies. These energy-source mechanisms include:

Each of these energy mechanisms has certain merits in explaining some observed characteristics of quasars and other active galaxies. This is in the sense that they could produce a lot of energy from a very small volume, could produce nonthermal radiation, could be variable in time, and would not last forever. In their present form, however, none seems to satisfy all the requirements.


Copyright 1995 J. C. Evans
Physics & Astronomy Department, George Mason University
Maintained by J. C. Evans; jevans@gmu.edu