Introducing the observed I(r) into this expression, one obtains the actual luminosity distribution p(R). In the figure the solid curve shows the three-dimensional luminosity distribution obtained from the Vancouleurs' law (the dashed line).

Since z2 = R2 — r2, a change of the variable of integration yields

If the galaxy is not spherical, its three-dimensional shape can only be determined if its inclination with respect to the line of sight is known. Since galactic discs are thin and of constant thickness, the inclination i of a disc galaxy is obtained directly from the axis ratio of its projected image: sin i = b/a.

When the inclination is known, the real axis ratio of the bulge q0 can be determined from the projected value q. For a rotationally symmetric bulge the relation between q and q0 is

The flattenings of disc galaxy bulges obtained from this relation lie in the range q0 = 0.3-0.5. Since the inclinations of ellipticals are generally unknown, only the statistical distribution of q can be determined from that of q0.

18.4 Dynamics of Galaxies

18.4 Dynamics of Galaxies

We have seen how the masses of galaxies can be derived from observed velocities of stars and gas. The same observations can be used to study the internal distribution of mass in more detail.

Slowly Rotating Systems. The dynamics of elliptical galaxies and disc galaxy bulges are studied by means of the Doppler shifts and broadenings of stellar absorption lines. Since a given absorption line is the sum of contributions from many individual stars, its Doppler shift gives their mean velocity, while its broadening is increased by an amount depending on the dispersion of stellar velocities around the mean. By observing how the wavelengths and widths of spectral lines behave as functions of the radius, one can get some insight into the distribution of mass in the galaxy.

Examples of the observed radial dependence of the rotational velocity and velocity dispersion derived for some ellipticals were given in Fig. 18.8. The observed rotational velocities are often small (< 100kms-1), while the velocity dispersion may typically be about 200kms-1. If elliptical galaxies were in fact ellipsoids of revolution, there should be a statistical relation (when projection effects have been taken into account) between flatness, rotational velocity and velocity dispersion. Such a relationship has been observed for fainter ellipticals and for disc galaxy bulges. However, some of the brightest ellipticals rotate very slowly. Therefore their flattening cannot be due to rotation.

The radial dependence of the velocity dispersion gives information on the distribution of mass within the galaxy. Since it also depends on how the shapes of stellar orbits in the galaxy are distributed, its interpretation requires detailed dynamical models.

Rotation Curves. In spiral galaxies the distribution of mass can be studied directly using the observed rotational velocities of the interstellar gas. This can be observed either at optical wavelengths from the emission lines of ionised gas in H II regions or at radio wavelengths from the hydrogen 21 cm line. Typical galactic rotation curves were shown in Fig. 18.9.

The qualitative behaviour of the rotation curve in all spiral galaxies is similar to the rotation curve of the Milky Way: there is a central portion, where the

rotational velocity is directly proportional to the radius, corresponding to solid body rotation. At a few kpc radius the curve turns over and becomes flat, i. e. the rotational velocity does not depend on the radius. In early Hubble types, the rotation curve rises more steeply near the centre and reaches larger velocities in the flat region (Sa about 300 km s-1, Sc about 200 km s-1). A higher rotational velocity indicates a larger mass according to (18.7), and thus Sa types must have a larger mass density near the centre. This is not unexpected, since a more massive bulge is one of the defining properties of early type spirals.

A decrease of the rotational velocity at large radii would be an indication that most of the mass is inside that radius. In some galaxies such a decrease has been detected, in others the rotational velocity remains constant as far out as the observations can reach.

Spiral Structure. Spiral galaxies are relatively bright objects. Some have a well-defined, large-scale two-armed spiral pattern, whereas in others the spiral structure is made up of a large number of short filamentary arms. From galaxies where the pattern is seen in front of the central bulge, it has been deduced that the sense of winding of the spiral is trailing with respect to the rotation of the galaxy.

The spiral structure is most clearly seen in the interstellar dust, H II regions, and the OB associations formed by young stars. The dust often forms thin lanes along the inner edge of the spiral arms, with star forming regions on their outside. Enhanced synchrotron radio emission associated with spiral arms has also been detected.

The spiral pattern is generally thought to be a wave in the density of the stellar disc, as discussed in Sect. 17.4. As the interstellar gas streams through the density wave a shock, marked by the dust lanes, is formed as the interstellar gas is compressed, leading to the collapse of molecular clouds and the formation of stars. The density wave theory predicts characteristic streaming motions within the arm, which have been detected in some galaxies by observations of the H121 cm line.

It is not known how the spiral wave was produced. In multiarmed galaxies the spiral arms may be short-lived, constantly forming and disappearing, but extensive, regular, two-armed patterns have to be more long-lived. In barred spirals the bar can drive a spiral wave in the

Fig. 18.12. Above: A spiral galaxy from above: M51 (type Sc). The interacting companion is NGC5195 (type Irr II). (Lick Observatory). Below: A spiral galaxy from the side: the Sb spiral NGC4565. (NOAO/Kitt Peak National Observatory)

Fig. 18.12. Above: A spiral galaxy from above: M51 (type Sc). The interacting companion is NGC5195 (type Irr II). (Lick Observatory). Below: A spiral galaxy from the side: the Sb spiral NGC4565. (NOAO/Kitt Peak National Observatory)

gas. Some normal spirals may have been produced by the tidal force from another galaxy passing nearby. Finally, in some galaxies a two-armed spiral may have been spontaneously generated by an instability of the disc.

18.5 Stellar Ages and Element Abundances in Galaxies

From the Milky Way we know that stars of populations I and II are different not only in respect to their spatial distribution, but also in respect to their ages and heavy element abundances. This fact gives important evidence about the formation of the Milky Way, and it is therefore of interest if a similar connection can be found in other galaxies.

The indicators of composition most easily measured are the variations of colour indices inside galaxies and between different galaxies. Two regularities have been discovered in these variations: First, according to the colour-luminosity relation for elliptical and SO galaxies, brighter galaxies are redder. Secondly, there is a colour-aperture effect, so that the central parts of galaxies are redder. For spirals this relationship is due to the presence of young, massive stars in the disc, but it has also been observed for elliptical and SO galaxies.

Galactic spectra are composed of the spectra of all their stars added together. Thus the colours depend both on the ages of the stars (young stars are bluer) and on the heavy element abundance Z (stars with larger Z are redder). The interpretation of the observational results thus has to be based on detailed modelling of the stellar composition of galaxies or population synthesis.

Stars of different spectral classes contribute different characteristic absorption features to the galaxy spectrum. By observing the strength of various spectral features, one can find out about the masses, ages and chemical composition of the stars that make up the galaxy. For this purpose, a large number of characteristic properties of the spectrum, strengths of absorption lines and broad-band colours are measured. One then attempts to reproduce these data, using a representative collection of stellar spectra. If no satisfactory solution can be found, more stars have to be added to the model. The final result is a population model, giving the stellar composition of the galaxy. Combin-

ing this with theoretical stellar evolution calculations, the evolution of the light of the galaxy can also be computed.

Population synthesis of E galaxies show that practically all their stars were formed simultaneously about 15 x 109 years ago. Most of their light comes from red giants, whereas most of their mass resides in lower main sequence stars of less than one solar mass. Since all stars have roughly the same age, the colours of elliptical galaxies are directly related to their metallicities. Thus the colour-luminosity relation indicates that Z in giant ellipticals may be double that in the solar neighbourhood, while it may be smaller by a factor 100 in dwarfs. Similarly, the radial dependence of the colours can be explained if the value of Z at the centre is an order of magnitude larger than it is at larger radii.

The stellar composition of disc galaxy bulges is generally similar to that of ellipticals. The element abundances in the gas in spirals can be studied by means of the emission lines from HII regions ionised by newly formed stars. In this case too, the metallicity increases towards the centre, but the size of the variation varies in different galaxies, and is not yet well understood.

18.6 Systems of Galaxies

The galaxies are not smoothly distributed in space; rather, they form systems of all sizes: galaxy pairs, small groups, large clusters and superclusters formed from several groups and clusters. The larger a given system, the less its density exceeds the mean density of the Universe. On the average, the density is twice the background density for systems of radius 5 Mpc and 10% above the background at radius 20 Mpc.

Interactions of Galaxies. The frequency of different types of galaxies varies in the various kinds of groups. This could either be because certain types of galaxies are formed preferentially in certain environments, or because interactions between galaxies have changed their shapes. There are many observed interacting systems where strong tidal forces have produced striking distortions, "bridges" and "tails" in the member galaxies.

The interactions between galaxies are not always dramatic. For example, the Milky Way has two satellites,

Fig. 18.13. Above: The irregular Virgo Cluster of galaxies. Below: The regular Coma Cluster (ESO and Karl-Schwarzschild-


the Large and Small Magellanic Clouds (see Fig. 18.5), which are Irr I type dwarf galaxies at about 60 kpc distance. It is thought that approximately 5 x 108 years ago, these passed the Milky Way at a distance of about 10-15 kpc, leaving behind the Magellanic Stream, a 180° long thin stream of neutral hydrogen clouds. Systems of this type, where a giant galaxy is surrounded by a few small companions, are quite common. Computations show that in many such cases the tidal interactions are so strong that the companions will merge with the parent galaxy at the next close approach. This is likely to happen to the Magellanic Clouds.

During earlier epochs in the Universe, when the density was larger, interactions between galaxies must have been much more common than at present. Thus it has been proposed that a large fraction of bright galaxies have undergone major mergers at some stage in their history. In particular, there are good reasons to believe that the slowly rotating, non-axisymmetric giant ellipticals may have formed by the merger of disc galaxies.

Groups. The most common type of galaxy systems are small, irregular groups of a few tens of galaxies. A typical example is the Local Group, which contains two larger galaxies in addition to the Milky Way - the Andromeda Galaxy M31, an Sb spiral of about the same size as the Milky Way with two dwarf companions, and the smaller Sc spiral M33. The rest of the about 35 members of the Local Group are dwarfs; about 20 are of type dE and 10 of type Irr I. The diameter of the Local Group is about 1.2 Mpc.

Clusters. A system of galaxies may be defined to be a cluster if it contains a larger number (at least 50) of bright galaxies. The number of members and the size of a cluster depend on how they are defined. One way of doing this is to fit the observed distribution of galaxies within a cluster with an expression of the form (18.8). In this way a characteristic cluster radius of about 2-5 Mpc is obtained. The number of members depends both on the cluster radius and on the limiting magnitude. A large cluster may contain several hundred galaxies that are less than two magnitudes fainter than the characteristic luminosity L * of (18.2).

Clusters of galaxies can be ordered in a sequence from extended, low-density, irregular systems (some-

times called clouds of galaxies) to denser and more regular structures (Fig. 18.13). The galaxy type composition also varies along this sequence in the sense that in the loose irregular clusters, the bright galaxies are predominantly spirals, whereas the members of dense clusters are almost exclusively type E and S0. The nearest cluster of galaxies is the Virgo Cluster at a distance of about 15 Mpc. It is a relatively irregular cluster, where a denser central region containing early galaxy types is surrounded by a more extended distribution of mainly spiral galaxies. The nearest regular cluster is the Coma Cluster, roughly 90 Mpc away. In the Coma Cluster a central pair of giant ellipticals is surrounded by a flattened (axis ratio about 2: 1) system of early type galaxies.

X-ray emission from hot gas has been detected in many clusters. In irregular clusters the gas temperature is about 107 K and the emission is generally concentrated near individual galaxies; in the regular ones the gas is hotter, 108 K, and the emission tends to be more evenly distributed over the whole cluster area. By and large, the X-ray emission follows the distribution of the galaxies. The amount of gas needed to explain this emission is about equal to the mass of the galaxies - thus it cannot solve the missing mass problem. X-ray emission lines from multiply ionised iron have also been observed. On basis of these lines, it has been concluded that the metal abundance of intergalactic gas is roughly equal to that of the Sun. For this reason it is most likely that the gas has been ejected from the galaxies in the cluster.

Superclusters. Groups and clusters of galaxies may form even larger systems, superclusters. For example, the Local Group belongs to the Local Supercluster, a flattened system whose centre is the Virgo Cluster, containing tens of smaller groups and clouds of galaxies. The Coma Cluster is part of another supercluster. The diameters of superclusters are 10-20 Mpc. However, on this scale, it is no longer clear whether one can reasonably speak of individual systems. Perhaps it would be more accurate to think of the distribution of galaxies as a continuous network, where the large clusters are connected by walls and strings formed by smaller systems. Between these there remain empty regions containing very few galaxies, which can be up to 50 Mpc in diameter (Fig. 18.14).

Fig. 18.14. The large-scale space distribution of 245591 galaxies. The radial co-ordinate is the red-shift, which can be translated into a distance using a value for the Hubble constant. The thickness of the slices is about 10°. (2dFGRS Team, http://www2.aao.gov.au/ TDFgg/)

Fig. 18.14. The large-scale space distribution of 245591 galaxies. The radial co-ordinate is the red-shift, which can be translated into a distance using a value for the Hubble constant. The thickness of the slices is about 10°. (2dFGRS Team, http://www2.aao.gov.au/ TDFgg/)

18.7 Active Galaxies and Quasars

So far in this chapter we have been concerned with the properties of normal galaxies. In some galaxies, however, the normal galaxy is overshadowed by violent activity. This activity is produced in the nucleus, which is then called an active galactic nucleus (AGN).

The luminosities of active galactic nuclei may be extremely large, sometimes much larger than that of the rest of the galaxy. It seems unlikely that a galaxy could maintain such a large power output for long. For this reason it is thought that active galaxies do not form a separate class of galaxy, but rather represent a passing stage in the evolution of normal galaxies.

Activity appears in many different forms. Some galaxies have an exceptionally bright nucleus similar to a large region of ionised hydrogen. These may be young galaxies, where near the centre large numbers of stars are forming and evolving into supernovae (star-burst nuclei). In other nuclei the radiation cannot have been produced by stars, and the most plausible source of energy in these nuclei is the gravitational energy of a supermassive black hole (mass > 108 M0). In some galaxies, the spectral lines are unusually broad, indicating large internal velocities. These may be either rotational velocities near a black hole or due to explosive events in the nucleus. In some galaxies, jets are seen coming out of the nucleus. Many active galaxies radiate a nonthermal spectrum, apparently synchrotron radiation produced by fast electrons in a magnetic field.

The classification of active galaxies has been developed rather unsystematically, since many of them have been discovered only recently, and have not been completely studied. For example, the Markarian galaxies catalogued by Benyamin Yerishevich Markarian in the early 1970's are defined by strong ultraviolet emission. Many Markarian galaxies are Seyfert galaxies; others are galaxies undergoing a burst of star formation. The N galaxies form another class closely similar to the Seyfert galaxies.

Two natural basic classes of active galaxies are the Seyfert galaxies and the radio galaxies. The former are spirals; the latter are ellipticals. Some astronomers think that the Seyfert galaxies represent the active stage of normal spiral galaxies and the radio galaxies that of ellipticals.

Seyfert Galaxies. The Seyfert galaxies are named after Carl Seyfert, who discovered them in 1943. Their most important characteristics are a bright, pointlike central nucleus and a spectrum showing broad emission lines. The continuous spectrum has a nonthermal component, which is most prominent in the ultraviolet. The emission lines are thought to be produced in gas clouds moving close to the nucleus with large velocities.

On the basis of the spectrum, Seyfert galaxies are classified as type 1 or 2. In a type 1 spectrum, the allowed lines are broad (corresponding to a velocity of 104 km s-1), much broader than the forbidden lines. In type 2, all lines are similar and narrower (< 103 km s-1). Transitions between these types and intermediate cases have sometimes been observed. The reason for the difference is thought to be that the allowed lines are formed in denser gas near the nucleus, and the forbidden lines in

more diffuse gas further out. In type 2 Seyfert galaxies, the denser gas is missing or obscured.

Almost all Seyfert galaxies with known Hubble types are spirals; the possible exceptions are of type 2. They are strong infrared sources. Type 1 galaxies often show strong X-ray emission.

The true Seyfert galaxies are relatively weak radio sources. However, there are compact radio galaxies with an optical spectrum that is essentially the same as for Seyfert galaxies. These should probably be classified with the Seyfert galaxies. In general, the stronger radio emission seems to come with a type 2 spectrum.

It is estimated that about 1% of all bright spiral galaxies are Seyfert galaxies. The luminosities of their nuclei are about 1036-1038 W, of the same order as all the rest of the galaxy. Brightness variations are common.

Radio Galaxies. By definition, radio galaxies are galaxies that are powerful radio sources. The radio emission of a radio galaxy is non-thermal synchrotron radiation. The radio luminosity of radio galaxies is typically 1033-1038 W, and may thus be as large as the total luminosity of a normal galaxy. The main problem in explaining radio emission is to understand how the electrons and magnetic fields are produced, and above all, where the electrons get their energy.

The forms and sizes of the radio emitting regions of radio galaxies have been studied ever since the 1950's, when radio interferometers achieved the resolution of optical telescopes. The characteristic feature of a strong radio galaxy is a double structure: there are two large radio emitting regions on opposite sides of the observed galaxy. The radio emitting regions of some radio galaxies are as far apart as 6 Mpc, almost ten times the distance between the Milky Way and Andromeda galaxies. One of the smallest double radio sources is the galaxy M87 (Fig. 18.15), whose two components are only a few kpc distant from each other.

The double structure of radio galaxies appears to be produced by ejections from the nucleus. However, the electrons in the radio lobes cannot be coming from the centre of the galaxy, because they would lose all their energy during such a long transit. Therefore electrons have to be continuously accelerated within the radio-emitting regions. Within the radio lobes there are almost point-like regions, hot spots. These are generally sym-

Fig. 18.15. Above: The active galaxy M87. In the lower right-hand corner a short exposure of the core region has been inserted (same scale as in the main photograph). One sees a blue jet coming out of the nucleus of a normal E0 galaxy. (NOAO/Kitt Peak National Observatory). Below: In the radio map made using the VLA the jet is observed to be two-sided. The area shown is much smaller than in the upper picture. (Owen, F.N., Hardee, P.E., Bignell, R.C. (1980): Astrophys. J. (Lett.) 239, L11)

metrically placed with respect to the nucleus, and are apparently consequences of nuclear ejections.

"Tailed" radio sources also exist. Their radio emission mainly comes from one side of the galaxy, forming a curved tail, which is often tens of times longer than the diameter of the galaxy. The best examples are NGC1265 in the Perseus cluster of galaxies and 3C129, which appears to be in an elliptical orbit around a com panion galaxy. The tail is interpreted as the trail left by the radio galaxy in intergalactic space.

Another special feature revealed by the radio maps is the presence of jets, narrow lines of radio emission, usually starting in the nucleus and stretching far outside the galaxy. The best known may be the M87 jet, which has also been observed as an optical and X-ray jet. The optically observed jet is surrounded by a radio source. A similar radio source is seen on the opposite side of the nucleus, where no optical jet is seen. Our nearest radio galaxy Centaurus A also has a jet extending from the nucleus to near the edge of the galaxy. VLBI observations of radio jets have also revealed superluminal motions: in many compact sources the components appear to be separating faster than the speed of light. Since such velocities are impossible according to the theory of relativity, the observed velocities can only be apparent, and several models have been proposed to account for them.

Quasars. The first quasar was discovered in 1963, when Maarten Schmidt interpreted the optical emission lines of the known radio source 3C273 as hydrogen Balmer lines redshifted by 16%. Such large redshifts are the most remarkable characteristics of the quasars. Properly speaking, the word quasar is an abbreviation for quasistellar radio source, and some astronomers prefer to use the designation QSO (quasistellar object), since not all quasars emit radio radiation.

Optically the quasars appear almost as point sources, although improved observational techniques have revealed an increasing number of quasars located inside more or less normal galaxies (Fig. 18.16). Although the first quasars where discovered by radio observations, only a small fraction of all optically identified quasars are bright radio sources. Most radio quasars are point sources, but some have a double structure like the radio galaxies. Satellite X-ray pictures also show the quasars to be pointlike.

In the visible region the quasar spectra are dominated by spectral lines with rest wavelengths in the ultraviolet. The first observed quasar redshifts were z = 0.16 and 0.37, and later searches have continued to turn up ever larger redshifts. The present record is 6.3. The light left the quasar when the Universe was less than one-tenth of its present age. The large inferred distances of the quasars mean that their luminosities have to be extremely large. Typical values lie in the range

of 1038-1041 W. The brightness of quasars may vary rapidly, within a few days or less. Thus the emitting region can be no larger than a few light-days, i. e. about 100 AU.

The quasars often have both emission and absorption lines in their spectra. The emission lines are very broad and are probably produced in the quasar itself. Much of the absorption spectrum consists of densely distributed narrow lines that are thought to be hydrogen Lyman a lines formed in gas clouds along the line of sight to the quasar. The clouds producing this "Lyman a forest" are young galaxies or protogalaxies, and they therefore provide important evidence about the formation of galaxies.

Some astronomers have questioned the cosmolog-ical interpretation of the redshift. Thus Halton Arp has discovered small systems of quasars and galaxies where some of the components have widely discrepant redshifts. For this reason, Arp thinks that the quasar redshifts are produced by some unknown process. This claim is highly controversial.

Unified Models. Although the forms of galactic activity may at first sight appear diverse, they can be unified within a fairly widely accepted schematic model. According to this model, most galaxies contain a compact central nucleus, which is a supermassive black hole, with mass 107-109 Me, surrounded by a disc or ring of gas. The source of energy is the gravitational energy released as gas is accreted into the black hole. The disc may also give rise to a jet, where some of the energy is converted into perpendicular motions along the rotational axis. Thus active galactic nuclei are similar to the nucleus of the Milky Way, although the masses of both the black hole and the gas disc may be much larger.

Deducing the mass of the central black hole is difficult and uncertain. However, using a variety of methods involving the motions of stars and gas at the centre of nearby galaxies black hole masses for about 30 galaxies have been determined. The most important result of these studies is that there is a close relationship between the black hole mass and the central velocity dispersion of the galaxy. According to the virial theorem the velocity dispersion is a measure of the bulge mass, and therefore the conclusion is that there is a close relationship between the mass of the bulge and the mass of the central black hole.

Fig. 18.16. One of the nearest quasars, 3C 273, photographed with two cameras aboard the Hubble Space Telescope. On the left, the Wide Field Planetary Camera sees a bright pointlike source, with a jet blasted out from the quasar (towards 5 o'clock). On the right, a coronagraph in the Advanced Camera for Surveys blocks out the brightest parts of the quasar. Spiral arms in the host galaxy can be seen, with dark dust lanes, as well as new details in the path of the jet. (Photos Hubble/NASA/ESA)

Fig. 18.16. One of the nearest quasars, 3C 273, photographed with two cameras aboard the Hubble Space Telescope. On the left, the Wide Field Planetary Camera sees a bright pointlike source, with a jet blasted out from the quasar (towards 5 o'clock). On the right, a coronagraph in the Advanced Camera for Surveys blocks out the brightest parts of the quasar. Spiral arms in the host galaxy can be seen, with dark dust lanes, as well as new details in the path of the jet. (Photos Hubble/NASA/ESA)

A first characteristic parameter of the unified model is obviously the total luminosity. For example, the only essential difference between Seyfert 1 galaxies and radio-quiet quasars is the larger luminosity of quasars. A second basic parameter is the radio brightness, which may be related to the strength of a jet. On the basis of their radio luminosity one can connect Seyfert galaxies and radio-quiet quasars on one hand, and radio galaxies and radio quasars on the other.

The third important parameter of unified models is the angle from which we happen to view the nuclear disc. For example, if the disc is seen edge-on, the actual nucleus is obscured by the disc. This could explain the difference between Seyfert types 1 and 2: in type 2 we do not see the broad emission lines formed near the black hole, but only the narrower lines from the disc. Similarly a galaxy that looks like a double radio source when seen edge-on, would look like a radio quasar if the disc were seen face-on. In the latter case there is a possibility that we may be seeing an object directly along the jet. It will then appear as a blazar, an object with rapid and violent variations in brightness and polarization, and very weak or invisible emission lines. If the jet is almost relativistic, its transverse velocity may appear larger than the speed of light, and thus superluminal motions can also be understood.

One prediction of the unified model is that there should be a large number of quasars where the nucleus is obscured by the disc as in the Seyfert 2 galaxies. In analogy with the Seyfert galaxies these objects are referred to as type 2 AGN or quasars. Because of the obscuration such sources would not be included in surveys at optical, UV, or soft X-ray wavelengths. In hard X-rays the obscuration is less, and in the far infrared the absorbed energy is re-radiated. Searches for type 2 quasars have

Fig. 18.17. (a) The components of the Einstein cross are grav-itationally lensed images of the same quasar. (ESA/NASA) (b) The massive galaxy cluster Abell 2218 deflects light rays passing through it and acts as a giant gravitational lens. In

this Hubble picture from January 2000, dozens of arc-shaped images of distant galaxies can be seen. (Photo A. Fruchter, S. Baggett, R. Hook and Z. Levay, NASA/STScI)

this Hubble picture from January 2000, dozens of arc-shaped images of distant galaxies can be seen. (Photo A. Fruchter, S. Baggett, R. Hook and Z. Levay, NASA/STScI)

18.8 The Origin and Evolution of Galaxies been made at these wavelengths with the Chandra X-ray satellite and with the Spitzer Space Telescope. The indications from these searches are that at least 3/4 of all supermassive black holes are heavily obscured.

Gravitational Lenses. An interesting phenomenon first discovered in connection with quasars are gravitational lenses. Since light rays are bent by gravitational fields, a mass (e. g. a galaxy) placed between a distant quasar and the observer will distort the image of the quasar. The first example of this effect was discovered in 1979, when it was found that two quasars, 5.7" apart in the sky, had essentially identical spectra. It was concluded that the "pair" was really a double image of a single quasar. Since then several other gravitationally lensed quasars have been discovered (Fig. 18.17).

Gravitational lenses have also been discovered in clusters of galaxies. Here the gravitational field of the cluster distorts the images of distant galaxies into arcs around the cluster centre. In addition in 1993 microlens-ing was observed in the Milky Way, where the brightness of a star is momentarily increased by the lensing effect from a mass passing in front of the star. Thus the study of gravitational lens effects offers a new promising method of obtaining information on the distribution of mass in the Universe.

18.8 The Origin and Evolution of Galaxies

Because the speed of light is finite we see distant galaxies at earlier stages in their life. In the next chapter we show how the age of a galaxy with given redshift can be calculated based on the rate of expansion of the Universe. However, this relationship will depend on the cosmological model, which therefore always has to be specified when studying the evolution of galaxies.

The beginning of galaxy evolution was marked by the formation of the first stars at a redshift around 10-20. The ultraviolet radiation of the earliest objets reionised the intergalactic gas, which made the Universe transparent to radiation, and thus made the distant galaxies and quasars visible in principle. The largest observed redshifts known at present are about 6.5.

According to the currently widely accepted cosmo-logical models most of the matter in the Universe is

in a form that emits no radiation, and is only observable from its gravitational effects. In this Cold Dark Matter (CDM) theory (see Sect. 19.7) the first systems to collapse and start forming stars were small, with masses like those of dwarf galaxies. Larger galaxies were formed later as these smaller fragments collected into larger clumps. This model, where most stars are formed in small galaxies is usually described as the hierarchical model. Before the introduction of the CDM model, the dominant model was one where the most massive systems are the first to start forming stars, which is often called the monolithic model. We have already seen in Sect. 17.5 that the formation of the Milky Way shows aspects of both types of models.

A large part of our theoretical ideas on galaxy evolution is based on numerical simulations of the collapse of gas clouds and star formation in them. Using some prescription for star formation one can try to compute the evolution of the spectral energy distribution and the chemical abundances in the resulting galaxies. The results of the models can be compared with the observational data presented in the previous sections of this chapter.

The density distribution of dark matter is expected to be very irregular, containing numerous small-scale clumps. The collapse will therefore be highly in-homogeneous, both in the hierarchical and in the monolithic picture, and subsequent mergers between smaller systems should be common. There are additional complicating factors. Gas may be expelled from the galaxy, or there may be an influx of fresh gas. Interactions with the surroundings may radically alter the course of evolution - in dense systems they may lead to the complete merging of the individual galaxies into one giant elliptical. Much remains to be learned about how the formation of stars is affected by the general dynamical state of the galaxy and of how an active nucleus may influence the formation process.

Our observational knowledge of galaxy evolution is advancing very rapidly. Essentially all the relationships described earlier in this chapter have been studied as functions of time. Still, a complete generally accepted description of the way the Universe reached its present state has not yet been established. Here we can only mention a few of the most central aspects of the processes leading to the galaxies we observe to-day.

Log flux density (W m"2 Hz"1) -29 -30 -31 -32 -33 -C

Log flux density (W m"2 Hz"1) -29 -30 -31 -32 -33 -C

20 25

AB Magnitude

Fig. 18.18. Galaxy counts in the U, B, I, and K wavelength bands. The counts are compared to those in a cosmological model without evolution. The cosmological parameters correspond to the currently preferred "concordance" model (see Sect. 19.5). (H.C. Ferguson et al. 2000, ARAA38, 667, Fig. 4)

20 25

AB Magnitude

Fig. 18.18. Galaxy counts in the U, B, I, and K wavelength bands. The counts are compared to those in a cosmological model without evolution. The cosmological parameters correspond to the currently preferred "concordance" model (see Sect. 19.5). (H.C. Ferguson et al. 2000, ARAA38, 667, Fig. 4)

Density and Luminosity Evolution The the most basic way of studying the formation and evolution of galaxies is by counting their numbers, either the number brighter than some given magnitude limit (as was already done by Hubble in the 1930's, see Sect. 19.1) or else the number density as a function of redshift. The counts can be compared with the numbers expected if there is no evolution, which depend on the cosmological model. They can therefore be used either as a cosmological test, or as a test of evolutionary models. However, the present situation is that more reliable cosmological tests are available, and the number counts are mainly used to study the evolution of galaxies.

There are two ways in which the number counts are affected by galaxy evolution. In density evolution the actual number of galaxies is changing, whereas in luminosity evolution only the luminosity of individual galaxies is evolving. The simplest form of luminosity evolution is called passive luminosity evolution, and is due to the changing luminosity of stars during normal stellar evolution. Pure luminosity evolution is expected to be predominant in the monolithic picture. In the hierarchical picture density evolution will be more prominent, since in this picture smaller galaxies are to a greater extent being destroyed to produce larger more luminous ones.

Figure 18.18 gives an example of the results of number counts. A model without evolution cannot explain these observed counts, and various models incorporating evolutionary effects have to be introduced. However, a unique model cannot be determined using just the number counts.

Distant Galaxies. A more direct approach to galaxy formation is the direct search for the most distant objects visible. In Fig. 18.19 we show the Hubble Deep Field (HDF), an area of the sky observed at several wavelengths with the Hubble Space Telescope using a very long exposure. These images show large numbers of galaxies with redshifts about 2.5-3.5, when the Universe was about 2 Ga old. The galaxies visible in the HDF appear to be bluer and more irregular than the

Fig. 18.19. The Hubble Deep Field South in the direction of the Tucana constellation (NASA; a colour version of the same picture is seen as Plate 33 in the colout supplement)

galaxies in our present neighbourhood, suggesting that they in the process of formation.

One striking result that the search for very distant galaxies has revealed is that some galaxies with stellar masses 1011-1012 Me had already formed around a redshift of 1-2, and with masses 1010 Me at the redshift 3, or even 5-6, when the Universe was only 1 Ga old. This rapid onset of star formation in massive galaxies is against the expectations that the first stars should be found in dwarf galaxies. Even for nearby systems there is a scarcity of dwarfs compared to the most immediate estimates of the CDM theory. The observational evidence seems to be pointing in the direction of downsizing: star formation is initially largely confined to massive systems, and starts to shift to smaller systems at a redshift about 2-1. The processes invoked in the hierarchical scheme may then take over as the most important force driving galaxy evolution.

The unexpected discovery of old galaxies at large redshifts has led to a classification of distant galaxies into a red and a blue class depending on whether star formation is still going on. These two classes will have different evolutionary histories, and by charting these histories we may learn better to understand the whole process of galactic evolution. The distant red objects have also inspired searches for their progenitors. Some such progenitors have been identified as ultralu-minous infrared galaxies and as submillimeter galaxies, but general agreement about their status does not yet exist.

Evolution of AGN The first clear indication of cosmic evolution was in the numbers of radio galaxies and quasars. Already in the late 1960's it was becoming clear that the density of quasars was increasing dramatically towards higher redshifts. Roughly, the number density of quasars increases relative to the present density by a factor 100 out to a broad maximum at redshift about 2. The observed behaviour may be due to either density or luminosity evolution. The density of radio galaxies also has a maximum at redshifts about 1.5-3, which is sometimes referred to as the quasar era.

The Star Formation History of the Universe. Since the Universe contained only neutral gas as the first stars began to form, the most general description of how galaxies came to be is in terms of the rate at which

Fig. 18.20. Star formation history of the Universe. The rate at which gas is being turned into stars is fairly well-determined up to a redshift of about 6. Estimates for even more distant epochs are still uncertain. The symbols refer to different searches for distant galaxies. The shaded and hatched regions indicate theoretical fits using various models for the initial mass function of the stars. The dashed line corresponds to the level of star formation required to reionise the Universe at that redshift. (A.M. Hopkins 2006, astro-ph/0611283, Fig. 3)

Fig. 18.20. Star formation history of the Universe. The rate at which gas is being turned into stars is fairly well-determined up to a redshift of about 6. Estimates for even more distant epochs are still uncertain. The symbols refer to different searches for distant galaxies. The shaded and hatched regions indicate theoretical fits using various models for the initial mass function of the stars. The dashed line corresponds to the level of star formation required to reionise the Universe at that redshift. (A.M. Hopkins 2006, astro-ph/0611283, Fig. 3)

the gas is being turned into stars. The star formation history going back to a redshift about 6 is shown in Fig. 18.20. The star formation rate was about an order of magnitude larger than its current value at redshift 1-2. For even larger redshifts it seems to have remained nearly constant or decreased slowly.

18.9 Exercises

Exercise 18.1 The galaxy NGC 772 is an Sb spiral, similar to M31. Its angular diameter is 7' and apparent magnitude 12.0. The corresponding values of M31 are 3.0° and 5.0. Find the ratio of the distances of the galaxies a) assuming their sizes are equal, b) assuming they are equally bright.

Exercise 18.2 The brightness of the quasar 3C279 has shown changes with the time scale of one week. Estimate the size of the region producing the radiation. The apparent magnitude is 18. If the distance of the quasar is 2000 Mpc, what is its absolute magnitude and luminosity? How much energy is produced per AU3?

Telescopes Mastery

Telescopes Mastery

Through this ebook, you are going to learn what you will need to know all about the telescopes that can provide a fun and rewarding hobby for you and your family!

Get My Free Ebook

Post a comment