Galaxies

Hubble’s Classification Scheme

The true nature of the galaxies was first convincingly demonstrated by Edwin Hubble when he showed that they are far outside the Milky Way. On the basis of photographic images Hubble suggested there are three major categories of galaxies. An example of each is shown below.

elliptical (E)	spiral (S)	irregular (I)
M87	M51	Small Magellanic Cloud
All Images NOAO/AURA/NSF

Hubble expanded this preliminary classification system into what has now become known as the “tuning fork” diagram.

In this picture the ellipticals are distributed along the “handle” of the tuning fork, with decreasing ellipticity to the left. The spirals are distributed along two sequences. The splitting was necessary because of the presence of two apparently similar sequences of spirals. One contains the “normal” spirals, in which the spiral arms appear to come from a central galactic bulge. On the other, the “barred” spirals have spiral arms that appear to come from the ends of a linear bar-like structure that passes through the bulge.

The spirals can be placed in a sequence along which the arms become more open and the bulge becomes less prominent. This is indicated by the designations Sa through Sc (and SBa to SBc for barred spirals). Hubble thought this ordering might have been an evolutionary sequence, with objects evolving from tightly wound to a more open state. Thus, the Sa galaxies are still referred to as “early” galaxies while the Sc galaxies are referred to as “late” galaxies.

Additional Resources

Seeing Farther: Hubble Ultra Deep Field – An interactive site where you can see the evolution of galaxies through the Hubble Ultra Deep Field.

Galaxy Evolution and the Tuning Fork Diagram

Hubble expected there should be an S0 class which formed an evolutionary bridge between the ellipticals and the spirals. Though he never detected such an object, the S0’s were added later by Hubble’s student, Allan Sandage. Sandage greatly expanded Hubble’s original system. Additional enhancements were also suggested by Gerard de Vaucouleurs, who added the class Sd as a class later than the Scs. An Sd galaxy has extremely open arms and a negligible or even non-existent nucleus.

We now know that the tuning fork diagram bears no relation to the evolution of galaxies, at least, not in the sense that galaxies evolve from “early” to “late” as envisioned by Hubble. However, we do see collisions in which two spirals can merge to form an elliptical, counter to the evolutionary development assumed by Hubble. This is only one way to form an elliptical though, and many galaxies likely form as ellipticals from the very start. Also, in the early universe we see that large ellipticals and spirals seem to be forming from small irregular galaxies; there is even strong evidence that this process continues to the modern day.

NGC 2787, a lenticular galaxy. Image: NASA/ESA Hubble Space Telescope

Spirals like the Milky Way and Andromeda have cannibalized small galaxies in the past, and they continue to do so (see the image of M51 at the top of this page). The Milky Way is currently in the process of devouring the Large Magellanic Cloud, one of its small companion galaxies. The well-known “globular cluster” Omega Centauri seems in fact to be the remnant nucleus of a small elliptical galaxy that has been consumed by the Milky Way. Furthermore, in several billion years observational evidence suggests that the Milky Way and Andromeda will collide and merge, eventually settling down to form a large elliptical galaxy. The image at right shows two galaxies engaged in a similar encounter.

Evidence now suggests that the process of galaxy evolution is one of titanic mergers and cannibalism, in striking contrast to the slow migration envisioned by Edwin Hubble early in the Twentieth Century. So even though his tuning fork diagram is not at all related to the evolution of galaxies, the classification scheme originated by Hubble is still a useful way to describe the current appearance of a galaxy, if not where it came from or where it might be headed.

Image: Hubble Heritage Site, NASA/ESA Hubble Space Telescope

Galaxies Observed in Multiple Wavelengths

	We customarily classify galaxies and attempt to describe their morphological characteristics primarily on the basis of visual light images. This is in part a historical accident in that we first examined galaxies (spiral nebulae or elliptical nebulae) 200 years ago using visual inspection of telescopic images. Beginning 100 years ago photographic emulsions began to be used to record astronomical images. While the spectral sensitivity of photographic emulsions is slightly different from the spectral sensitivity of the eye, the two are close enough so that they are both commonly referred to as “visual light” images. Of course, it turns out that the visual-light region is the most advantageous one for studying galaxies in general since this is the region where stars (and hence galaxies) usually emit the most radiation.     However, as technological advances have made it possible to image astronomical sources outside the visible band, it has become apparent that some morphological structures are more pronounced in the IR, radio or even x-ray than they are in the visible. Indeed, some relatively featureless galaxies have revealed vast amounts of structure in non visible bands. For example, in spiral galaxies the dust scatters and obscures visible light while allowing IR, radio and x-ray photons to pass through relatively unimpeded. Furthermore, the dust becomes a stronger emitter of IR photons at the temperatures of the interstellar medium. In addition, neutral atomic hydrogen can be observed in the radio (1.414 MHz/21 cm) and yields views of galaxy interactions that are not at all obvious in other bands. As this technology has improved, we have come to recognize that a far clearer understanding of galaxies is possible when we observe both visible and non-visible photons.
	Spitzer IR Data: NASA/JPL-Caltech/S. Willner (Harvard-Smithsonian Center for Astrophysics)

Galaxy Color

When we refer to galaxy color we are generally using it as a proxy for the stellar population. Early type galaxies like ellipticals do not contain cool gas or dust. As a result they do not form stars and are dominated by an older, redder stellar population. Spirals, on the other hand, do form stars and therefore have a younger, bluer stellar population mixed in with their underlying old disk population. This effect is not a large one, however, and most spirals are not appreciably bluer than elliptical galaxies.

Besides stellar content there are other properties that affect the color of a galaxy. One is the presence of dust. Dust will scatter short wavelength light while letting the longer red wavelengths through. Therefore, extremely dusty galaxies will appear red regardless of their underlying stellar population (this is the same process that causes the Sun to be redder at sunrise and sunset than at midday). Indeed, such dusty galaxies will often be undergoing extremely vigorous star formation, but their thick dust content blocks a direct view of their young, bright stars. Instead we may see that they have inordinate emission in the infrared, the result of the massive stars warming the dust and causing it to glow in the IR.

Sometimes a galaxy can be bluer than we expect given its morphology or stellar content. This will be the case for galaxies with extremely old stellar populations. Generally we think of old stars as being red, but if the stars are extremely old they would have formed before the galaxy had produced many metals (in astronomy, “metals” are all elements heavier than hydrogen and helium, and all those heavier than boron are formed inside stars). They will therefore be metal-deficient. The preponderance of atomic transitions for many of these elements is in the UV part of the spectrum. In stars with many metals the absorption in this region somewhat blocks emission there; much of the flux that would normally be emitted in the UV must come out at longer wavelengths, making these stars redder than they would otherwise be. On the other hand, low-metal stars are able to emit more light in the blue and UV parts of the spectrum, which somewhat diminishes their emission at longer wavelengths. While this is a small effect, it can subtly change the color of the stars in a galaxy, making it slightly bluer would be expected (actually, slightly less red is a better way to think about it).

So color is not always a simple thing. Generally a red galaxy will contain older, cooler stars. But in some cases there is more to it, either because of the presence of dust in the galaxy or because of low metal content.

Comparison of galaxies of different ages.

Galaxy Spectra

Galaxies are studied mostly through the light they emit. When astronomers study an object, they often spread its light out into the constituent wavelengths or energies that the light contains. For visible light these are seen as different colors. The light is often displayed on a graph, with brightness plotted against wavelength. The brightness can also be plotted against energy or frequency… these are all basically equivalent. Such a plot is called a spectrum. Studying the spectra of astronomical objects is how many astronomers spend their time. This is because understanding an object’s spectrum is the key to knowing many things about it. Below we discuss some general features of galaxy spectra.

In addition to morphology, as described above, the spectra of galaxies can also be used as an aid in classifying them. The spectrum of a galaxy is produced mostly by the stars it contains, and so galaxies generally show the same spectral lines commonly seen in stars. Of course, the lines are broadened due to the galaxy’s rotation since, at any given moment, some stars are moving toward us, while others are moving away. These motions shift the position of the lines according to the Doppler shift formula. So one thing we can measure from the spectrum of a galaxy is the speed at which its stars move within it (which in turn allows us to measure its mass). But we can learn many other things about a galaxy from its spectrum. In particular, the presence of strong emission lines in a galaxy’s spectrum is a sign of star formation.

The spectra below at right are both taken from Parisi et al., 2009 (arXiv:09095345v1). Each is a plot of flux (in erg per square centimeter per second per angstrom) vs wavelength, here plotted in Ansgstrom units (1 angstrom is 0.1 nanometers, or 10^-10 meters). To break this down, the plots show how much energy (in ergs) from the galaxy passes through each square centimeter of the detector during each second at each wavelength plotted (in angstroms). For reference, the bluest light visible to humans has a wavelength just short of 4000 angstroms, and the reddest light visible is around 7000 angstroms. So this spectrum runs from violet to just beyond the red part of the spectrum in the near infrared. Of course, the electromagnetic spectrum continues to shorter wavelengths (ultraviolet) and longer (infrared), but our eyes cannot see this radiation. Our detectors can, but those wavelengths have not been measured for these spectra.

You should keep in mind that these galaxies are at great distances, so their spectra have both been redshifted by the expansion of the universe. As a result, the lines shown are not at the wavelengths they would have in the lab. Despite these shifts, the patterns of the lines are the same as they would be in a nearby galaxy; the redshift effect only shifts the pattern. Redshift acts on each wavelength the same, shifting it by the same percentage as all other wavelengths.

Redshifted absorption lines on a spectrum

https://tl.wikipedia.org/wiki/Edwin_Hubble

Galaxy Size

Sizes of galaxies are determined by measuring their angular extent on the sky and determining their distance.

Galaxy Luminosity, aka Absolute Magnitudes

Luminosities or absolute magnitudes can be determined for a galaxy by measuring its apparent magnitude and combining that with its distance. Here is it assumed that the flux of the galaxy diminishes as the inverse of the square of its distance from us (strictly, this is true only for point sources, but the methods used to measure galactic distances generally employ such sources). Just as for size, apparent magnitude determinations are difficult for galaxies since it is difficult to define the precise location of the “edge” for a nebulous object. In addition, corrections must be applied because of the attenuation by dust in our own galaxy, the attenuation due to dust in the external galaxy, the orientation of the galaxy in space (edge-on or face-on), and the K-correction due to the redshift of the galaxy.

The following results are believed to be typical absolute magnitudes.

Antennae Galaxies seen by NASA Hubble

Elliptical Galaxies

Elliptical galaxies appear in the sky as luminous elliptical disks. Elliptical galaxies have more stars concentrated at their center than at their outer edges, but otherwise have little or no structure. The light distribution is smooth with the surface brightness decreasing smoothly outward from the center.

M87

Image Provided By Anglo-Australian Observatory

The actual light distribution of ellipticals is well represented as

cD Galaxies

This type of galaxy appears similar to the ellipticals, and they are sometimes referred to as giant ellipticals. The type was introduced by W. W. Morgan. The “c” refers to an old spectroscopic designation for supergiants, while the “D” simply stands for diffuse. The objects appear to have greatly extended envelopes and frequently exhibit structure and multiple nuclei near their centers. They may have sizes which are ten times larger than a normal elliptical.

ABELL 779

Image Provided by John Gretchen

cD galaxies are probably the result of two or more objects merging together, or of a larger object “swallowing” or cannibalizing smaller galaxies.

Irregular Galaxies

Irregular galaxies show no symmetrical or regular structure. Galaxies of type Irr I have resolved OB stars and HII regions. Irr II do not resolve into stars and are generally amorphous. Irregulars seem to contain a higher than usual concentration of dust and gas. Because of the dependence on resolution, it is impossible to distinguish between Irr I and Irr II for more distant objects.

NGC 55

Image Provided By Anglo-Australian Observatory

Hubble described this class as lacking both dominating nuclei and rotational symmetry.

In general, this class displays a lack of any organized structure.

The suffix p, for peculiar, should generally be reserved for galaxies with obvious structure that has been distorted by the tidal interaction with close companions.

Lenticular Galaxies

Lenticular galaxies (Type S0) are intermediate between the E7 ellipticals and the Sa spirals. They are flatter than E7s and have a thin disk as well as a spheroidal nuclear bulge. When they are seen edge on they sometimes have the shape of a convex lens, so they are also called lenticulars. The disk components have a light distribution that falls off more slowly than in the ellipticals.

NGC 936

Image Provided By Sloan Digital Sky Survey

S0 galaxies are disk galaxies, like spirals, but they have far less dust and gas than a normal spiral. It is possible that S0s are early type spirals that have lost their dust and gas as a result of tidal interactions with other galaxies. It is even possible that S0s are early type spirals which have suffered collisions. One clue to the nature of lenticular galaxies is that they are preferentially found near the centers of rich clusters of galaxies, where the intracluster gas would strip the interstellar dust and gas from a galaxy as it falls through the cluster.

Ring Galaxies

There are galaxies known to have very large ring-shaped structures beyond the nucleus. Some of these galaxies look almost like the planet Saturn, with their large prominent ring surrounding the nucleus. Sometimes the nucleus is nearly absent, so the only structure apparent is a large ring. In either case, these galaxies are termed “ring galaxies” and have been classified as R (for “Ring”). If an underlying spiral structure is discernible, the galaxy might be classified as S(R) or SB(R). Note that the upper case R is used for large ring structures beyond the nucleus, while the lower case r is used for ring structures features within or near the nucleus.

Hoag’s Object

This image was produced by the Hubble Heritage Team (STScI/AURA) using data collected by the Hubble Heritage team (STScI/AURA) and Ray Lucas (STScI).

It has been suggested by some that ring galaxies are produced as a consequence of a galaxy-galaxy collision.

Dwarf Galaxies

Most of the galaxies with which we are familiar are bright and easily detected. However, the vast majority of galaxies in the universe are actually small and faint. They tend not to fit into the classification scheme as we have outlined thus far. The most common form for these galaxies appears to be elliptical, and since they generally contain little or no gas, they are as a class referred to as dwarf ellipticals, dE. In addition to being smaller than the bright elliptical galaxies, the dEs do not exhibit a bright nuclear region. Another classification used for dwarf galaxies is dIrr (for dwarf irregular).

Leo I

Image Provided by Anglo-Australian Observatory

De Vaucleures has proposed the term “compact” to designate the dwarf galaxies. Thus, one can encounter the designations cE, for compact ellipticals, and cI, for compact irregulars.Since these objects are faint, they can only be detected if they are relatively nearby. This in itself can present a problem because some dwarf galaxies are so sparsely populated with stars that they are difficult to distinguish from Milky Way foreground stars. Nonetheless, given the large numbers of dwarf galaxies that we find in the local group and in other nearby groups, these are likely the most common types of galaxies in the universe.

Spiral Galaxies

Spiral galaxies are either ordinary (S or SA) or barred (SB). Both types have spiral shaped arms, with two arms generally placed symmetrically about the center of an axis of rotation. In the ordinary spirals the arms emerge from the nucleus, while in the barred spirals the arms emerge near the ends of a bar-like structure which passes through the center of the nucleus. Both types are classified according to how tightly the arms are wound, how patchy they are, and the relative size of the nucleus.

NGC 5236

Image Provided by Anglo-Australian Observatory

Ordinary spirals of type Sa have smooth ill-defined arms that are tightly wound around a large prominent nucleus. The arms are wound so tightly they are nearly circular. The intermediate Sb galaxies have more open arms which are often partly resolved into HII regions and stellar associations. The nuclei of Sc galaxies are usually quite small and the arms are well extended and resolved into HII regions and clumps of stars. An Sd galaxy displays virtually no nucleus at all.

Hubble described this sequence in the following manner:

“The arms appear to build up at the expense of the nuclear regions and unwind as they grow; in the end the arms are wide open and the nuclei inconspicuous. Early in the series the arms begin to break up into condensations, the [breakup] commencing in the outer regions and working inwards until in the final stages it reaches the nucleus itself. The “breakup” referred to is the appearance of HII regions and bright blue supergiants.”

Click Here to go to ViewSpace’s light spectrum activity with two colliding spiral galaxies.

Active Galactic Nuclei (AGN) Galaxies

Messier 100, Credit: NASA/STScI

Active Galactic Nuclei (AGN) Spectra

This is the spectrum of a spiral galaxy. Parisi et al., arXiv:0909.5345v1, 2009.

This first spectrum is from a spiral galaxy. Notice the strong strong emission lines of oxygen (near 5000 angstrom) and hydrogen (just short of 7000 angstrom). Other emission lines are seen as well. Those from sulphur (S), nitrogen (N) and helium (He) are labeled. These lines are produced by hot, ionized gas clouds which are often seen in regions where stars are forming. As a result, the presence of these emission lines in a galaxy means it is forming stars. Only spirals and some irregular galaxies form stars, so just from looking at this spectrum you can tell that the galaxy is not an elliptical. The Roman numerals after the chemical symbols indicate the ionization state of that atom. OI refers to neutral oxygen, OII is singly ionized oxygen (loss of one electron) and OIII is doubly ionized oxygen (loss of two electrons). The presence of the helium emission lines indicate that this is no ordinary spiral galaxy. In fact, it harbors an AGN, and is therefore an active galaxy. This is also indicated by the broadening of the lines of hydrogen (especially H-alpha), nitrogen and sulphur. In addition to the emission lines, several different absorption lines are seen in this galaxy. The absorption lines will be discussed below in the description of the elliptical galaxy spectrum. You can click on these images to view a larger version if you wish.

This spectrum is from an elliptical galaxy. Notice the lack of most of the emission lines seen in the spiral galaxy, above. The only emission line seen is OII, which in this galaxy has been shifted to almost 5000 angstroms (it has a rest wavelength of 3727 angstroms). Like the example above, this is an active galaxy. If it were a normal elliptical there would be no emission lines at all. From the previous discussion of spirals you will deduce that the lack of emission lines in elliptical galaxies results from them having no star formation, and thus no ionized gas clouds. However, this last statement is somewhat misleading. Elliptical galaxies do contain ionized gas, but it is at temperatures higher than a million kelvin. At these temperatures the gas is so hot and highly ionized that it emits x-rays. We do not see any sign of this hot gas in the optical spectra shown. The gas in spirals, on the other hand, has a range of temperatures from only a few kelvin in the cores of molecular clouds, to about ten thousand kelvin in the ionized clouds responsible for the emission lines. You may notice that the absorption lines in this elliptical galaxy seem much more pronounced than the ones seen in the spiral spectrum. This is not really the case. The presence of the strong emission lines in the spiral spectrum makes the vertical scale somewhat compressed. The continuum (the continuum is the average value of the spectrum, ignoring emission and absorption lines) appears flatter in the spiral, and the lines seem to be not as deep as in the elliptical. In some spiral galaxies the emission lines are so strong that the continuum appears completely flat, but this is just an artifact of how we have chosen to plot the graph, with a linear vertical scale. If we blew up the vertical scale we would see that the continuum in both galaxies is about the same, with the same rough shape and similar absorption features. In both galaxies the continuum is produced by cool giant stars for the most part. So, aside from the emission lines from the star forming regions in the spirals, their spectra are similar to the spectra of ellipticals.