The Quasar Absorption Line Group at New Mexico State University

INTRODUCTION TO QSO ABSORPTION LINES IN SIMULATIONS


| Main | Past Students | Collaborators | Other Groups | Background/Tutorials | Useful Science Links |


Welcome to our journey and quest to understand how galaxies form and evolve! We will be using an observational technique known as "QSO Absorption Line" and the theoretical approach of N-body + hydrodynamic cosmological simulations. We'll get to the QSO absorption lines shortly, including how we cipher the "codes" of the features in QSO spectra.

N-body refers to treating objects as point masses and having them obey the laws of gravitational attraction. N refers to their being many of these bodies, sometimes as many an several hundred million. Hydrodynamics, which involves complex fluid flow physics and energy balancing, is more difficult to handle. Simulations have now become sophisticated enough that both N-body treatment of stars (and dark matter particles) AND hydrodynamics of gas flows can be included in our computations.

The above image is from a hydrodynamic cosmological simulation that uses an adapted refinement tree (ART) method to employ high resolution in regions where the gas and stars are changing rapidly, while allowing the regions that are evolving slowly to be modeled at lower resolution. The scale of the background image is roughly 5 Megaparsecs* (Mpc) across. The zoom in box shows the gas density (no stars are shown) centered on a galaxy; the density scale (cm^-3) is given in the legend. The scale of this image is given by the vertical bar across the top, which measures 400 kiloparsecs* (kpc) across. The size of the galaxy (not shown) is on the order of 10 kpc in diameter. Note the filamentary structure out of which the galaxy forms.

Though not shown, stars form in the highest density regions. Individual stars cannot be modeled, instead massive star "particles" represent a population of stars, including their distribution of masses and lifetimes. This still allows us to model the evolution of the star, including how they build up heavy chemical elements and how the supernovae explode (including the outward force and heating of their energy into the surrounding medium). The exploding stars pollute the surrounding gas with heavy elements like carbon, nitrogen, carbon, and oxygen. So, the complete star-gas cycles taking place in galaxies is modeled.

* A parsec is a unit of distance equal to 3.26 light years, where a light year is the unit of distance light travels in a year (~31,500,00 seconds at 300,00 km/s). A kiloparsec (kpc) is a thousand parsecs.A Megaparsec (Mpc) is a million parsecs.


Let's step back and look at the big picture first and inventory the universe. The important thing to take away here is that normal matter (called Baryonic Matter) comprises only 4% of the total mass-energy density of the universe.

As for Dark Matter, it is indirectly detected only through its gravitational influence on the normal matter. Dark matter does NOT interact with light. The normal matter DOES interact with light, and light is the only physical quantity that astronomers can measure. Thus, we can only directly "see" 4% of the universe, quite literally the tip of the cosmic iceberg! The normal matter is gravitationally attracted to the dominating dark matter, so it is the cosmological distribution of dark matter that governs the cosmological distribution of normal matter. Not only do we deduce dark matter resides in individual galaxies, but we also deduce that it is distributed throughout galaxy groups and galaxy clusters.

The Dark Energy is a relatively new discovery (or rediscovery!). Like dark matter, dark energy is also an indirectly deduced entity, which behaves as a ubiquitous negative pressure throughout the cosmos. Actually, Albert Einstein introduced this negative pressure component into his General Theory of Relativity as the Cosmological Constant. But he later retracted it and then called its introduction his biggest blunder. Once again, it appears as if Albert was on the money; his self-proclaimed blunder may have been the real blunder! Love you Albert.

Dark energy is deduce from observed brightness of supernovae (which are employed as standard candles) in very distant galaxies (up to 10 billion light years away). In the late 1990s, their brightness was observed to be lower than was predicted by a cosmological model of the expanding universe with a null cosmological constant. In such a model, the expansion of the universe is predicted to be decelerating (i.e., the univers is gravitationally bound). However, the supernovae results indicate that the expansion of the universe is accelerating.

Thus, we needed to reinsert the cosmological constant back into the mathematical model (Einstein's General Relativity Theory). Observational data from the Wilkinson Microwave Anisotropy Probe (WMAP) and the aforementioned measured brightness of the distant supernovae have been combined to constrain this so-called negative pressure (dark energy) to comprise 70% of the mass-energy density of the universe.


As an astronomer, you have two tools in your toolbox: (1) the knowledge of physics, and (2) photon* counts. From the pattern of light we observed, we apply the laws of physics to deduce the composition, temperature, velocities, etc of the matter that emitted (or absorbed) the light. So, in a very real sense, all astronomical knowledge rests on the laws of physics being unchanging throughout cosmic time and space( known as the universality of physical law). All our knowledge is based upon deductions; astronomers do not, they cannot, make direct measurements of the matter in the universe.

* photon is a name given to a single "particle" of light


That 4% presents to us a beauty and majesty beyond compare. Above are selected examples of various galaxies. Note the dark areas which can be seen in silhouette; they are cool gas clouds in the interstellar medium of the galaxies that absorb light. These dark clouds are where new stars will form. A typical spiral galaxy rotates about once every 250 million years. Our very own Milky Way is a spiral galaxy (see upper left galaxy) and the sun is located about 2/3rd the way out in the disk. The Milky Way may have a very small bar- see 2nd galaxy down on the left for an example of a large bar).

Also note that galaxies can appear pathological in their shapes or that their gas is distributed in odd patterns. This is due to interactions between galaxies; we sometimes are seeing the galaxy well after a collision, and pathological shapes suggest that an interaction happened in the past. If two galaxies pass close enough to one another, they can become gravitationally bound and eventually merge together. When galaxies merge, the stars actually never collide, only the gas collides. Thus, the gas usually falls to the center or the merging system while the stars stream in what are known as tidal tails. When the gas collides, it forms new stars (see bottom right, which shows the inner region of a merging system).


(Click on Slide to See Movie; Use the BACK BUTTON on your broswer to return)

This movie (top center panel) was motivated to help understand the merging system known as The Mice (bottom left image). To set up the simulation, the galaxy masses, rotation directions and speeds, and approach vectors had to be estimated. Also, the viewing angle had to be chosen. So, there is a bit of educated guess work, but if the simulation accurately models the merging system, then it gives us confidence that these input quantities correctly characterize the system. This helps us understand the physics of merging galaxies. In this case, not only do we learn the initial conditions of the merger, but we also deduce that The Mice are infalling and on their second passage with each other!

When playing the movie, try to estimate when the simulation matches The Mice. In practice, astronomers analyze the detailed motions of the stars in The Mice and quantitatively compare them to the motions of the stars in the simulations. It is very detailed business.

Movie Credit: John Dubinski


The large scale structure of the universe, meaning the relative positions of galaxies over all cosmic space, is literally a "frozen" fossil imprint of the distribution of dark matter in first fraction of a second after the Big Bang. When the universe was microscopic in size, matter was not "created" uniformly within it; it was created in small localized quantum fluctuations of higher and low density. As the universe expanded, the regions of relatively higher density gravitationally attract each other, and become more pronounced (even higher relative density localized regions) while regions of lower density became less and less relatively dense.

The normal matter, which formed into galaxies, followed along for the gravitational ride and so, overall, galaxies are concentrated in the regions of increasing dark matter density. This pattern of how galaxies are distributed throughout the universe is shown on the right panel. The initial pattern of these fluctuations were also "frozen" into the temperature variations of the cosmic microwave background (CMB) light (see image on the left, which is an all sky picture). The CMB light was emitted in a big flash when the universe was about 380,000 yrs old. The CMB was emitted in a very short burst at the moment in time when the universe cooled down enough for free electrons to become bound to protons, forming neutral hydrogen atoms for the first time. Regions of higher density are red and of lower density are blue (though we measure them as temperature variations).


As such, the initial fluctuations of matter in the Big Bang are critical to know because we must know them to have our cosmological simulations yield the correct observed cosmological distribution of galaxies (i.e., the large scale cosmic structure).


(Click on Slide to See Movie; Use the BACK BUTTON on your broswer to return)

A movie of a cosmological simulation (sorry, the first frame is blank!). First, the dark matter is shown (whitish). Note the cosmic web of filaments and the concentrations that are at the intersections of the filaments. When the simulation pauses and the image rotates the scale across the image is about 20 million light years! On this scale, individual galaxies are pin points, whereas galaxy clusters are the knots at the intersections of the filaments. Then the movie switches to green, the baryonic gas (normal matter) and evolves in time showing how the gas dynamically concentrates. Then the simulation shows the stars, including the eventual formation of a Milky Way like galaxy. Note how this teaches us how our own galaxy was formed.. and it is a violent process! Finally, we see the fully flowered model galaxy.


An observational image of the galaxy M81 showing the stars and gas. The neutral hydrogen gas (reddish) surrounds the galaxy. This is relatively nearby galaxy and the gas can be observed in weak emission. Even so, this is the tip of the gas iceberg around this galaxy.

On the other hand, very very diffuse gas can be sensitively observed due to its absorption of background light. We use bright background sources behind galaxies and measure the absorption due to the gas in the intervening galaxy. This works with equal sensitivity regardless of how far away the galaxy.


Need I really say it? ... a picture is worth a 1000 words ... a spectrum is worth a thousand pictures. Photos do not provide physical details such as motions and chemical composition. Spectra record an incredible amount of detail.


Each atom or ion interacts with light of specific energy (wavelength or color). Each atom has a unique series of exact wavelengths/energies with which it interacts. Thus, each atom has its own unique "spectral fingerprint" (pattern of absorption in the spectrum). In the spectrum shown on the right, the fingerprint pattern is from neutral hydrogen. As can be seen, often several absorption lines arise from a single type of atom or ion. As a spectroscopist, our job is to learn how to decode these patterns, much like a cipher expert.


Some people think that astronomers don't need to learn the Periodic Table... but they would be wrong. It turns out that certain ions exhibit so-called fine structure, which results in their fingerprint having two very close together absorption features; we call these doublets. In gas clouds in space, the atoms are often ionized, meaning that they have had one or more electrons stripped from them. The most obvious doublets that we see in astronomical spectra are once ionized Mg (MgII), three times ionized carbon (CIV), and five times ionized oxygen (OVI). These ions have strong doublet absorption because, after they are ionized, their electron configurations are similar to lithium or sodium atoms (the physics of which is beyond the scope of this presentation!). Also, hydrogen is very abundant and we almost always see the so-called Lyman-alpha transition. These absorption features are easy to spot in spectra and using them is how we find the otherwise invisible gas clouds in the universe.


Here is an example of a background QSO (or quasar) that lies behind some intervening gas clouds and galaxies. It was very nice of Mother Nature to provide such bright background light sources! Each time the light path along the line of sight to the QSO pierces through a gas structure, new absorption lines appear in the spectrum (see panels A-D).


In order to have a full understanding of the absorption lines, we also need to determine whether the line was formed in a distant gas structure (like system A in the previous slide) or a less distant gas structure (like system D in the previous slide). How to tell? Because of the expansion of the universe, the wavelength that is observed in the spectrum will be longer than the wavelength that was actually absorbed by the gas cloud. The further the gas cloud is from us, the longer amount of time has passed between when the light was absorbed and when the light was observed. Light wavelengths are "locked" into the expansion of space...so, the longer the photon has traveled in time, the more it is stretched with the continuously expanding space of the universe.

Thus, when an absorption line is found in a spectrum at wavelength λ, the trick is to be able to identify it in terms of the wavelength that was absorbed in the cloud (λ_0). This is can be a most challenging part of the analysis, but it is usually not really all that difficult. The amount of stretching that photon experienced can be then be determined; this stretching factor is called the redshift, and it is denoted by the letter "z" (see the equation for how it is solved). Redshift can be translated both into distance to the cloud and into how much time has passed since the absorption occurred in the cloud (though this requires a mathematical model of the universe, called a cosmological model). Thus, we can measure how distant and how far back in time we are looking at different gas structures in the same spectrum!


The image on the left illustrates the concept of the technique of QSO absorption lines. The observer is on the right and the QSO is on the left. The expanded view on the upper right illustrates how each pierced gas structure leaves an absorption line signature in the spectrum. Note how the relative transmitted intensity is reduced due to the absorption of the gas. The panel on the bottom right shows a schematic of how a single atom in the huge gas structure contributed to the absorption of the light. Each atom has orbiting electrons; when light passes through the atom, an electron will absorb the energy from the photon and, using the photon's energy, change its current orbit to a higher energy orbit. In a given gas cloud, billions of these atoms contribute to the overall absorption resulting in the spectrum as shown above.


The QSO spectrum on the right sports a clear absorption doublet from singly ionized Mg (known as MgII). MgII absorbs the wavelengths 2796 and 2803 angstroms. Note that that the doublet does not appear at those wavelengths; this is because the absorption takes place in a gas cloud at a redshift z=0.93. These doublets are easy to spot and easy to identify. The doublet absorption occurs because MgII has a small energy spitting (called fine structure) in the higher energy electron orbit; this energy splitting is schematically shown in the diagram on the left; but for a different ion). Ions that are singly ionized arise in cooler denser clouds; they have only one electron stripped from them and that happens in cooler environments. Ions that are missing more electrons, like CIV (three electrons stripped) or OVI (five electrons stripped) arise in hotter less dense gas. Thus, we can determine the temperature and density of the gas cloud (even when it is 11 billion light years away!)


Consider a QSO at z=3. It is surrounded by very hot material, which generates emission lines (the opposite process of absorption). By identifying the emission lines, we can determine the redshift of the QSO.


Consider clouds that have high concentrations of neutral hydrogen. The intergalactic filaments of the cosmic web are an example (shown as red clouds in diagram). Neutral hydrogen absorbs the Lyman alpha transition (rest frame wavelength 1215.7 angstroms). The observed wavelength of the absorption then depends upon the redshift at which the cloud absorbed the light. There are so many of these intergalactic hydrogen clouds that sometimes hundreds are seen in a single QSO spectrum. We actually call this "forest" of absorption lines in the spectrum the Lyman Alpha Forest. In this example, we show only a few lines.


Now, consider when the light path passes near a galaxy (shown in green at redshift z=1), metal lines like MgII and CIV are often seen. This is because the gas near galaxies is enriched with metals from the stars in the galaxy. Again, these lines are easy to identify in real QSO spectra.


This is a schematic of the twin Keck telescopes on Mauna Kea in Hawaii. The inset photo was taken by me on my first trip to the summit while driving (uhm, riding) in the observatory 4-wheel drive. It was an exciting day (July 4, 1994)!


These are also some pitcures I took on one of my runs at Keck. (upper left) The Keck I telescope primary mirror is 10 meters (30 feet) in diameter. It is a wondrous piece of engineering. The secondary mirror is to the left in the foreground. (lower right). The spectrograph is housed in its own clean room facility that moves with the telescope so that the spectrograph is always in the same location with respect to the telescope (for stability).


Some more personal photos- and yes I am dressed in one of these white clean suits! (top left) inside the HIRES clean room examining the spectrograph. Most of the components are covered; they are uncovered by computer from the control room when it is time to observe. In this photo: Steven S. Vogt, designer and builder of HIRES (and my Ph.D. advisor) and Fred Chaffee, then director of the Keck Observatory. (top right) The echelle grating, which disperses the light into the spectrum, with the viewing slit below it. Compare the photos to the schematic diagram showing the light path through the instrument. (lower right) an actual image of the spectrum; note all the absorption lines in this spectrum!


An expanded view of a HIRES/Keck spectrum. Now you can really see all the absorption lines. Our job is to objectively locate them (a statistically intensive task), identify them, and then analyze them. This is not a QSO spectrum, however. This is a spectrum of the sun! Note the very strong absorption in the very upper right in the red; this is from neutral hydrogen. Also note the two strong absorption lines in the upper left in the blue/orange. This is a doublet pair that is due to neutral sodium atoms in the atmosphere of the sun.


An actual HIRES/Keck QSO spectrum with most of the major features labeled. This QSO has a redshift z=2.406, as determined from the observed wavelengths of the emission lines. Note the forest of hydrogen Lyman-alpha absorption lines in the blue due to the cosmic web of hydrogen structures


To appreciate the level of detail, one must zoom into the spectrum. Shown here is the Lyman-alpha forest in the region of z~2.1. The center panel is an expanded view of the indicated redshift range in the top panel. The bottom panel is the expanded view of the indicated redshift range in the center panel. Note that each absorption line shows detailed structure, which provides detailed gas physics. As described in Slide 18, each line corresponds to gas cloud in the cosmic web, each pierced by the line of sight at different redshift. Thus, these lines are arise from cosmologically distributed gas.


Coming back to our z=0.93 MgII doublet (top panel), we have an example of a metal enriched gas structure that likely is associated with a galaxy. The top panel is a QSO spectrum obtained in 1989 with the 200-inch Hale telescope at the Palomar Observatory (courtesy Chuck Steidel). The bottom panel shows a small expanded region of the spectrum of the same QSO obtained with HIRES/Keck in 1995 (from my PhD thesis). There is a great deal of additional detail in the HIRES spectrum! The resolution of the Palomar spectrum is low enough that details lurking within the absorption lines get blurred out; in such spectra we can detect the MgII doublet clearly, but we cannot undertake detailed physical analysis of the gas. The resolution of the HIRES/Keck spectrum is very high, such that great detail is seen. In the HIRES spectrum, we now see that this MgII systems is in fact, three unique systems, A, B, and C. System A (z=0.9254) comprises five clouds; system B (z=0.9272) comprises a complex blending of roughly four clouds, and systems C (z=0.9343) comprises a single weakly absorbing cloud. What causes these systems to have clouds that come in bunches?


Here is a simple model to help with the interpretation (thanks to Jane Charlton).

(left top diagram) Consider a simple scenario of gas clouds falling radially toward the center of a galaxy and that every cloud has a velocity of 200 km/s. A QSO line of sight through this structure will undoubtedly pierce several individual clouds. Rule (1): if a given cloud is pierced near its center the absorption will be stronger, and it if is pierced near its edge the absorption will be weaker (in proportion to the path length of the light through the cloud). Rule (2): relative to the center of the galaxy, which we will assume is at rest (v=0) relative to us, the absorption line from a given cloud will be slightly Doppler shifted due to the cloud'ss motion in the galaxy halo. The red arrows on the pierced clouds all indicate 200 km/s velocity toward the galaxy center. But only the component of this velocity that is parallel to the line of sight to the QSO is measured (we called this measured velocity component the line of sight velocity). The spectrum for this infalling cloud scenario is shown and illustrates the varying absorption strengths and how the absorption from the individual clouds appear in the spectrum at slightly different line of sight velocities relative to the galaxy (v=0).

(right top diagram) This is the same idea but using a galaxy disk model that is rotating at 200 km/s. In such a geometric configuration, the line of sight velocities will all be systematically Doppler shifted to one side of the galaxy velocity. In this example, the line of sight passes through the part of the disk that is moving away from the observer, so all the absorption is at a positive velocity relative to the galaxy. Note how very different the absorption pattern looks compared to the radial infall model. Such very different patterns of individual and/or blended complex shape of the absorption profiles provide insight into the geometry and kinematics of the absorbing complex (galaxy halo or disk). As stated in the bottom panel, many absorption systems exhibit a mixture of both simple models. However, this simple picture is just that- WAY too simple.


Because of the Doppler shifting due to relative velocities of different clouds, we like to convert the wavelength scale to a velocity scale relative to the galaxy (called the rest-frame velocity). This is actually a very easy calculation, as shown. Note that for the MgII, CIV, and OVI absorption, we are showing only the blue member of the doublet pair. For example, what you are seeing for MgII is the velocity structure of the 2796 transition only. All further presentations of absorption will be on the velocity scale and usually show only the single member of the doublets. (If we were to show the second member of the double, it would have its own panel and the absorption would appear virtually identical to the first member of the doublet.)


From an observational standpoint, we ultimately we want to correlate the physical properties of the absorbing gas to the physical properties of the galaxies hosting the gas. This is slow and difficult work and it has taken the better part of a decade to acquire the necessary data. For this QSO, called Q0002+051 (the first four digits provide its right ascension, 00 hrs 92 minutes, the last three provide its declination, 5.1 degrees north), we observed three MgII systems, one at z=0.2981, one at z=0.5915, and one at z=0.8514. We then hunt for galaxies at these redshifts. The full image (upper left) is from the Hubble Space Telescope. Clearly some galaxy candidates are seen! Then, we take the spectrum of each of the galaxy candidates and determine the candidate galaxy's redshift. In this case, we found the galaxies associated with each MgII system. (in collaboration with Glenn Kacprzak and Chuck Steidel)


This is our sample of galaxy and absorption "pairs" as of 2005 (we am currently publishing the expanded sample and do not have it ready to showcase). The line up the center of the left hand panel is "impact parameter" ruler. The impact parameter is the physical sky projected separation in kiloparsecs* between the QSO line of sight and the center of the host galaxy (it requires a cosmological model and some math to compute it). Each galaxy/absorber pair is plotted and its impact parameter is indicated on the ruler (red points indicate that the absorption is very weak). There is a lot we can do with these data, but most of the analysis is beyond the scope of this presentation.

* a parsec is equivalent to 3.26 light years distance, a kiloparsec [kpc] is 1000 parsecs


For example, we are very interested in examining whether the gas is moving around the galaxy in the same sense that the galaxy is moving, or if the gas motions are decoupled from those of the galaxy. To do this, we measure the rotation speed of the galaxy; we take a spectrum of the galaxy and use the Doppler shifts of its emission lines along the spectrograph slit to obtain the so-called rotation curve (panel b; upper right). This galaxy has a projected rotation of ~120 km/s and lower portion of the galaxy is moving toward us while the upper portion is moving away from us. But, the actual rotation is larger, because the galaxy is slightly inclined, and we need to correct for the inclination.

(panel b; lower right) The MgII 2796 absorption lie and the CIV 1548 absorption line shown in relative velocity. It is interesting that the absorption gas is also moving at about -110 km/s, similarly to the lower portion of the galaxy. But alas, this does not indicate that the gas and the galaxy are both rotating around the galaxy. Recall, that the absorption is from otherwise invisible gas that is located directly in front of the QSO, so it is about 107 kpc projected separation from the galaxy center! When we model a simple rotation scenario, we predict that the absorption from a rotating disk/halo model would give rise to absorption only at the velocities indicated by the blue curves superimposed on the spectra. Note that much of the absorption is actually not consistent with rotating with the galaxy.


This is a second example of such an analysis. However, for this example, note that the velocity range over which the low ionization gas traced by MgII (velocities range from -200 to 0 km/s) and the high ionization gas traced by CIV (velocity range +20 to +100 km/s) are not the same. This would suggest different physical origins and kinematics for the low ionization gas and the high ionization gas. So, things can get complicated quickly. The trouble is, for all the power of the data have to offer, we cannot actually determine where in 3D the absorbing gas is relative to the galaxy- we only know is projected separation on the sky from the galaxy and that it is at a similar redshift as the galaxy. We need more power! Eaaaauuhhh!? Uhmmmmmm.


We need simulations to guide our understanding of the spatial and dynamic properties of the gas relative to the galaxies. These simulations provide our only window on such knowledge. How are the gas structures spatially distributed? What are the motions of the gas structures? What is the distribution of metals ions in these gas structures? And, how did they form and how will they evolve with time? The three panels show an integrated (over depth) slice of a simulated galaxy at three different cosmological redshifts (top: z=2.3, furthest in the past; center: z=1.3; and bottom: z=0.2, closer to the present epoch z=0). The bar across the top of each panel is 1000 kpc = 1 Megaparsec (Mpc). For comparison, The Milky Way and Andromeda galaxy are presently separated by 700 kpc. Cosmic time is increasing downward. From left to right, stars, gas density (cm^-3), gas temperature (Kelvin), and metal enrichment in units of solar metallicity. Though hard to read, legends for the color scale are provided in the vertical bar in the right hand side of each panel. As time passes, more stars form and the two galaxies merge (happens at z=0.6, not shown), Note that the galaxies are connected via the gas filament between them (z=2.3 and z=1.3), but then comprise single gas structure by z=0.2. Note that the gas temperature increases with time; the filaments at higher redshift are relatively cool. Also note that the gas becomes very metal enriched over a 1 Mpc volume by z=0.2. We run "mock" QSO sightlines through the simulations and generate "mock" QSO spectra and study them. In tandem with the direct knowledge from the simulations, we learn a great deal about observed absorption line systems I the real word.


This is a similar simulation (not the same one as the previous panel). Here, we show the gas temperature; stars are not shown. The galaxy itself is located in the very center and is about 30 kpc across (about the size of the red blob near the top of the z=3 panel), The bar across the top is 400 kpc, about half the distance between the Milky Way and Andromeda. Note that the temperature structure of this galaxy changes dramatically with cosmic time.


The distribution of metals ions for the same simulation as the precious slide. Note that as cosmic time passes, metals are highly spread out into the medium well outside the galaxy itself.


(Click on Slide to See Movie; Use the BACK BUTTON on your broswer to return)

This is a movie of how the metals (created in stars and then distributed back into and beyond the galaxies) evolve with cosmic time. The left panel is the total density of metals (cm^-3) and the right panel shows how the low ionization (traced by CII ions), intermediate ionization (traced by CIV ions), and high ionization (traced by OVI ions) evolve around galaxies and in the intergalactic medium. The panels are roughly 20 Mpc on a side. On this scale, the galaxies themselves are pinpoints, so you are looking at a huge volume of the universe (about 20 Mpc across). The simulation begins at z=20 and evolves to z=0 (present time). Note that the low ionization gas (red) is located in small halos coinciding with the galaxies and that the intermediate ionization gas (green) is in larger more extended halos surrounding the low ionization halos. The highest ionization gas (blue) grows dramatically throughout the intergalactic medium around redshift 3 to 2. This is because (and this is included in the simulation), QSOs were very abundant during this brief period of the universe and the QSOs send out a powerful background of high energy photons which highly ionizes the diffuse low density intergalactic gas.


(Click on Slide to See Movie; Use the BACK BUTTON on your broswer to return)

Here is a simulation/movie showing how the absorption lines through a fixed line of sight can change as the structure evolves over cosmic time. Of course, our real observations are only a single freeze-frame; but if we could come back and take a spectrum of the same QSO every 1000 years for the next billion years or so and made a movie of the absorption lines, the movie might look something like this. Bottom left: the line of sight velocity at each position along the line of sight. This information provides where in velocity the absorption lines will be seen (for the most dense gas). Top left: CII and CIV absorption lines as a function velocity. Right: the simulation box. The solid vertical line up the center is the position of the QSO line of sight.


In collaboration with my ex-graduate student (Glenn Kacprzak, now a postdoc at Swinburne University in Australia), we perform the same experiments in the simulations as we do in real life. That is, we measure all the absorption properties of the gas and then measure all the luminous properties of the galaxies, both in the real world and in simulations! Then we compare using rigorous statistical testing to quantify the similarities and differences. Similarities teach us that we are modeling the galaxies well; differences tell us that we need to reconsider how we are modeling the galaxies, the gas hydrodynamics, and the stellar feedback processes. Either way, we learn!


As an example of putting it altogether, we show (right panels) the LOS 1 absorption lines from various ions (this time including both members of the doublets). Blue curves superimposed provide the expectation for the absorption velocities if the galaxy were a simple planar disk geometry and the gas was co-rotating with it. It appears that the MgII gas could be explained in this way, but not the CIV nor the OVI absorption. The simulations reveal that the MgII is actually arising from infalling filaments around the galaxy and these filaments are not co-rotating with the galaxy. This is a powerful lesson; in the 1990s and early 2000s, astronomers really favored the co-rotating halo model. The bottom left panels are bit more complicated to explain. Each panel shows how a different gas properties behaves with line of sight position relative to the galaxy. For example, N(MgII) is the column density of MgII, Z is the metallicity, T is temperature, n is the density, and v is the line of sight velocity. Note how the density spikes as the line of sight passes the galaxy at its closest approach and that this higher density gas is much cooler (the downward temperature spike). This is the region from which the MgII and HI (Lyman beta) absorption arises.


There is much work to do, but so much progress has been made!


I would like to thank the Rose City Astronomers and the OMSI, the Oregon Museum of Science and Industry, for their kind hospitality and for inspiring the construction of this presentation. The Powerpoint files and the movies from which this page was derived are available for download. (right click the link and use "Save Link As")

Powerpoint:
| PPTX format | PPT format |
Movies (these may not run in the PPT unless re-inserted):
| Simulation of The Mice Collision | (source: John Dubinski) (also see: GRAVITAS)
| Cosmological Evolution & Galaxy Formation | (credit: Nichols, M.M. & Carney, K.E)
| Cosmological Distribution and Evolution of Metals | (source: Ben Oppenheimer)
| Time Sequence of QSO Absorption Lines | (source: Matthias Steinmetz)
You may also enjoy:
The Millennium Simulation Project
University of Washington N-Body Shop
Google Inventory
Public Outreach via Cosmological Simulations (PDF file)

This page created and maintained by Chris Churchill; last update: June 2011