Cosmic Dawn

by George Rhee

  The cosmic background measurements tell us that the ratio of baryonic (atomic) matter to dark matter in the universe is 1/5. The observed ratio, mass in stars to total mass, for our Milky Way galaxy is 1/20. In other words we see far fewer stars than we would expect based on cosmic measurements. This effect is more pronounced for dwarf galaxies which are even less efficient at forming stars out of their gas than Milky Way sized galaxies.

  This cosmic accounting relates the basic components of the mass in the universe (dark matter and atoms) to the observed properties of galaxies (dark matter halos, stellar disks, gas and dust). Our theories must explain how the dark matter clumped into structures, how the gas condensed in these dark matter clumps and how stars were formed. The star formation efficiency or fraction of gas that turned into stars is a number that depends on the mass of the dark matter halo within which a galaxy finds itself. Current theories cannot account for this number yet.

  The Milky Way according to the cold dark matter theory of structure formation should be surrounded by many hundreds of small galaxies but we only see about 50 in our observations. A challenge is to reconcile the large predicted number of low mass concentrations with the smaller number of observed dwarf galaxies. The theory also predicts that merging of smaller galaxies with the Milky Way should be taking place today. We do see evidence in the Milky Way halo of past mergers with dwarf galaxies. The visible consequences of such a merger are illustrated in Fig. 10.2.

  Fig. 10.2The distribution of stars in the shredded Sagittarius dwarf galaxy as revealed by observations. The image is based on the best model match to the map of 2MASS M-giant stars. The thin flat blue spiral represents the disk of our Milky Way galaxy. The yellow dot represents the position of the Sun. Sagittarius debris can be seen extending from the dense ‘core’ of the Sagittarius dwarf, wrapping around the galaxy, and descending through the Sun’s position (Credit: David Law/UCLA)

  The Andromeda Galaxy

  The most distant object that can be seen with the naked eye is the Andromeda Nebula also known as M31. It is the nearest large galaxy to our own at a distance of 2.5 million light years. The disk of the galaxy spans several moon diameters on the sky. The disk is round like a dinner plate but appears elliptical on the sky because it is not viewed directly face on but rather at an angle to our sight line.

  Like our own galaxy, M31 has been imaged from the ultraviolet to the radio part of the spectrum. Ultraviolet images (Fig. 10.3) reveal the presence of young hot stars. The near infrared enables us to map the old star population, the far infrared emission is a tracer of dust and the radio emission lets us map the neutral hydrogen gas. In the infrared, the dust reveals the presence of a ring of star formation with spiral structures extending inwards. The ring is also seen in the ultraviolet. Simulations suggest that the ring formed when a companion dwarf galaxy plunged through the center of M31 about 200 million years ago. In the simulations, the encounter destroys the spiral arms of M31 and brings about the formation of a ring-like structure. M31 is moving towards the Milky Way and the two galaxies will collide in about 3 billion years. Simulations of this event indicate a large probability that the Sun will be kicked out to distances a few times larger than its current distance from the Milky Way. In a few billion years the night sky seen from earth may look very different than it does today with the Milky Way being visible to anyone on earth as an external galaxy.

  Fig. 10.3Viewed in ultraviolet light, M31 looks more like a ring galaxy than a spiral. The ring is highlighted in this image taken by the GALaxy Evolution Explorer (GALEX) satellite. Ultraviolet colors have been digitally shifted to the visual. Young blue stars dominate the image, indicating the star forming ring as well as other star forming regions. The origin of the huge 150,000-light year ring is unknown but likely related to gravitational interactions with small satellite galaxies that orbit near the galactic giant (Credit: GALEX team, Caltech, NASA)

  The Local Group

  Swarms of dwarf galaxies surround the Milky Way and M31. They form a structure about 6 million light years in diameter known as the local group. It is a real structure in the sense that its member galaxies are participating in the expansion of the universe. The local group is shown schematically in Fig. 10.4. The local group is what we call a bound structure, held together by its own gravity.The next most massive galaxies after the Milky Way and M31 are the spiral galaxy M33 and the small and large Magellanic clouds. Most of the remaining galaxies are dwarf galaxies with masses of just a few percent of the Milky Way.

  Fig. 10.4The location of the Milky Way galaxy in the local group. The two dominant galaxies are surrounded by many smaller galaxies a few of which are shown in this figure (Credit: NASA/CXC/SAO)

  The faintest known dwarf galaxies have luminosities ten million times smaller than that of the Milky Way. These galaxies are located in the lowest mass dark matter clumps that are known to host stars. These galaxies are so small that we can ask whether they should not be referred to as star clusters. However, if we define galaxies as consisting mostly of dark matter, then these objects are dwarf galaxies. These faint objects are very difficult to detect against the bright Milky Way. Large digital surveys of the sky in several colors have made this possible. Techniques have been developed (see Fig. 10.7) that have detected 14 new dwarf galaxies in the local group, taking the census up to 50. Local group studies have revealed facts of great interest to cosmology. These findings include the number of dwarf galaxies, the stellar population (color and brightness) in local group dwarf galaxies, and the structure of the halos of the two large galaxies M31 and the Milky Way.

  The Missing Satellite Problem

  The currently favored model of cosmology is very successful at accounting for observations on the large scales of tens of millions of light years and larger. The model has encountered a number of problems on smaller scales. One such problem concerns the number of satellite galaxies of the Milky Way and Andromeda.

  We use the motions of stars in dwarf galaxies to estimate the masses of these galaxies and thus the mass of their dark matter halos. We can then compare the observed number of halos of given mass with the model predictions. For the most massive dwarf galaxies the agreement is fairly good. Galaxies in the models with the mass of the Milky Way should have one or two companions as massive as the Magellanic Clouds. As we go to less massive halos a discrepancy emerges. There are many more halos in the dark matter simulations than there are dwarf galaxy satellites surrounding the Milky Way. In fact, our local group is expected to have about 1,000 small mass concentrations but we only observe 50 or so galaxies. The challenge is thus to reconcile the large number of low mass dark matter concentrations with the smaller number of observed dwarf galaxies. How then are we to solve this so called “missing satellite problem”?

  We could modify the dark matter distribution so that gravity does not bring about the formation of low mass halos. This is not an option since independent observations of gas clouds constrain the dark matter distribution. If we accept that the predicted number of dark matter halos is correct, we can argue that not all halos form stars. Supernova explosions might prevent low mass halos from being visible to us. The first generation of stars are believed to be massive, they eventually explode as supernovas and the energy released in these explosions is sufficient to blow the gas out of low mass halos and prevent further star formation. The details of star formation and gas cooling suggest that most of the dark matter halos with mass less than 1% of the Milky Way halo should be completely dark. In fact, we may then encounter the opposite problem of explaining how the faintest detected dwarf galaxies could even exist. To solve this problem we need to know how efficient halos of given mass are at turning their gas into stars.

  Star Formation in Dwarf Galaxies

  Figure 10.5 shows the dwarf irregular galaxy NGC 1705. Some dwarf galaxies may be the first galaxies to have formed after the Big Bang. If the bigger galaxies formed by mergers of smaller galaxies, then the dwarf galaxies we see today may re
semble the building blocks that assembled to form larger galaxies like

  Fig. 10.5The dwarf irregular NGC 1705 is a small galaxy lacking regular structure. Young, blue, hot stars are strongly concentrated toward the galaxy’s center. Older, red, cooler stars are more spread out. This galaxy has been forming new stars throughout its lifetime, but a burst of star-formation activity occurred as recently as 26–31 million years ago (Credit: NASA, ESA and the Hubble Heritage Team STScI/AURA)

  our Milky Way. With the technological leap provided by the Hubble Space Telescope it became possible to image individual stars in several local group dwarf galaxies. This made it possible to measure the history of star formation in these nearby dwarf galaxies by looking at their individual stars. This is done by plotting each star as a point on a plot of star brightness versus star color.

  We can plot star color versus star brightness which allows us to estimate the relative numbers of older and younger stars.

  Using Hubble Space Telescope data, star formation histories have been constructed for these galaxies and the results are quite

  Fig. 10.6 Top Row: Three color magnitude diagrams for three dwarf galaxies. The horizontal axis plots the color, with blue on the left and redder towards the right. The vertical axis is magnitude with the brightness increasing along the vertical axis. Each point in these diagrams is a star, observed with the Hubble Space Telescope Advanced Camera for Surveys. On the bottom row we see the star formation rate (derived from the upper figure) plotted against time in billions of years (Gyr) with redshift labeled on the upper horizontal axis. Each column corresponds to a different galaxy’s star formation history and color-magnitude diagram. As can be seen, the star formation histories can differ substantially from one galaxy to the next (Credit: E. Tolstoy, V. Hill and M. Tosi, Annual Review of Astronomy and Astrophysics, Volume 47, page 387, 2009, Star-Formation Histories, Abundances, and Kinematics of Dwarf Galaxies in the Local Group)

  surprising (Fig. 10.6). Each column in the figure corresponds to a galaxy. The top row shows the star observations, and the bottom row shows the rate at which the galaxies formed stars over the history of the universe. Some galaxies like the Carina dwarf spheroidal galaxy show a number of distinct episodes of star formation with the star formation going to zero in between. Other galaxies such as Cetus show a long period of star formation lasting about 5 billion years after the Big Bang. In the dwarf galaxy Leo A, star formation happened over the last 8 billion years. It is unlikely that these dwarf galaxies in the local group are the building blocks of the Milky Way. True fossils are expected to have a single old stellar population or at least a temporary suppression of star formation after reionization. We also see that the element abundances in the stars in the dwarf galaxies do not resemble the abundances seen in stars in the Milky Way halo. The Milky Way and M31 could certainly have formed by merging of dwarf galaxies in the early universe, but the dwarf galaxies we see around today are quite different to those early building blocks. What then did the first galaxies look like?

  Are the First Galaxy Fossils Lurking in the Local Group?

  After the first atoms form in the universe the gas starts to fall into the dark matter halos. The gas reacts by increasing its temperature and pressure but keeps falling inwards if the time to build pressure is too slow compared to the time to fall into the clump. Stars from 20 to 100 solar masses could form in halos having a mass of one million solar masses. These massive stars produce enough ultraviolet radiation to ionize the gas in the halo and prevents further star formation. Eventually the first massive star explodes as a supernova and the gas can cool thanks to the presence of molecular hydrogen and the presence of heavy elements created by the first stars. This early period of star formation reionized the universe making it harder for low mass halos to form stars.

  The first galaxies to form are believed to have masses between 100 million (108) and 1 billion (109) solar masses. These first galaxies will become dwarf galaxies. Between the formation of the first stars and the end of reionization it is possible that some galaxies form in halos less massive than 100 million solar masses. If these halos survive they will be galaxy relics from the pre-reionization era.

  Could we observe these objects? The oldest known galaxies would be expected to be assemblages of a few hundred stars moving at speeds of about 20 km s − 1. These galaxy fossils as they are called could in fact resemble the least massive dwarf spheroidal galaxies that we see around the Milky Way. These faint ghostly objects are very hard to detect. The galaxy Leo T contains stars moving at about 7 km s − 1 in a dark matter halo of mass ten million (107) solar masses. There are several recently discovered candidate fossil galaxies. They are so faint that they can only be detected within 150,000 light-years of the Milky Way. More work is needed to confirm that these newly discovered objects are indeed fossils from the early stages of galaxy formation.

  We have discussed in the previous chapter the search for the first galaxies at high redshift and in this chapter the search for nearby fossils of galaxy formation. The two methods produce quite different results. The distant searches reveal intrinsically bright massive objects that are forming stars at a very high rate. The redshift eight galaxies have stellar masses of one billion (109) solar masses in contrast with the ultra faint dwarf stellar masses of a few hundred solar masses. The nearby fossils may be in fact the building blocks of the bright massive galaxies we see in the distant universe, which themselves may be the ancestors of galaxies such as the Milky Way.

  The Search for the Faint Galaxy Fossils

  Large sky coverage digital surveys can be used to conduct computer searches for faint assemblages of stars that may be invisible to the eye. Prior to 1994, 10 galaxies were classified as satellites of the Milky Way, since then another 14 have been discovered. Based on these discoveries it is estimated that the actual total number of Milky Way satellites lies between 200 and 500. Note the large uncertainty!

  The trick to finding these ultra faint companions of the Milky Way lies in finding an excess or overdensity of stars in a part of the sky. This is very difficult because the dwarf galaxies lie behind the Milky Way, and in any image of the sky the foreground Milky Way stars will greatly outnumber the stars in the dwarf galaxy. One has to apply a filter to eliminate foreground Milky Way stars and

  Fig. 10.7 Left Panel: Map of all the stars in the field around the Ursa Major I dwarf satellite. The field of view is about two moon diameters on the sky. The center panel shows the same field of view after removal of all stars except those that could plausibly be associated with a dwarf galaxy based on their color and brightness. The right panel shows a smoothed map of the star distribution in the center panel. The faintly visible dwarf galaxy is now clearly visible as the dark blob. The data are from the Sloan Digital Sky Survey (Credit: Beth Wilman, In Pursuit of the Least Luminous Galaxies, Advances in Astronomy, Volume 2010, page 4, 2010)

  examine the remaining handful of stars. We make a guess as to what the colors and brightnesses of the dwarf galaxy stars would be and eliminate stars outside this range from the image. If the remaining stars are concentrated into a clump then one has found a candidate dwarf galaxy (Fig. 10.7). One then obtains spectra of the candidate stars to confirm whether they are members of a separate galaxy. The ultra faint galaxies found in this manner are as faint as star clusters in the Milky Way but larger in size. As we go fainter, dwarf galaxies become increasingly dark matter dominated. The spectra of the stars in these galaxies enable us to estimated abundances of elements such as iron relative to hydrogen. These dwarf galaxies have iron abundances much lower than that of our sun. Since iron can only be created inside stars, the iron abundance measures to what extent the gas in a star was processed in a previous generation of stars. This lack of iron suggests we are looking at very old stars that formed from gas that was not “polluted” by the gas ejected during supernova explosions.

  The Large Scale Synoptic Telescope survey will make it possible to detect dwarf galaxi
es about eight times further than is currently possible. We should be able to find satellite galaxies right at the edge of the Milky Way dark matter halo. New radio telescopes such as the Square Kilometer Array telescope will search for dark matter halos that may contain only hydrogen gas.

  Tidal Streams

  The final steps on our archaeological tour of the local group are the halos of the two dominant spiral galaxies, our own Milky Way and M31, the Andromeda galaxy. Galaxies grow by accreting smaller galaxies, the small galaxies get disrupted as they approach the larger galaxy halo. The details depend on the halo mass and the speed and angle with which the victim approaches the large halo. We expect small satellites to be ripped apart by tidal forces as they approach the larger halo because the force of gravity due to the large halo is not the same at each location in the small halo. Different parts of the small halo experience different forces with the result that the small halo get ‘pulled apart’ as it spirals into the large one. We have observed this effect close to home when the comet Shoemaker-Levy collided with the planet Jupiter. As it approached Jupiter the comet was ripped into 16 different pieces by tidal forces. Figure 10.8 shows observations of the results of mergers between spiral and dwarf galaxies.