Michael Rauch's Homepage
 |
Michael Rauch
Astronomer
Mail:
Observatories of the
Carnegie Institution of Washington
813 Santa Barbara Street
Pasadena, CA 91101
e-mail:
mr(at)ociw.edu
phone: 626-304-0262
fax: 626-795-8136
|
|
Introductory literature on the Intergalactic Medium / Cosmic Web for Claremont Astrophysics Seminar
Introductory article on the Cosmic Web by Rob Simcoe (MIT)
Article on the Lyman alpha forest from the Encyclopedia of Astronomy etc
Annual Reviews article on Lyman alpha forest
NEWS
The paper on a "Population of Faint Extended Line Emitters..." is here.
My Research Interests
I am an observational astronomer, interested in various problems in
astrophysics and cosmology. Most of my current work is done with the
10m Keck and Carnegie's two Magellan 6.5m
telescopes.
The main observational technique I have
been using is the study of the spectra of distant quasars (or
Quasi-Stellar Objects, QSOs). My colleagues and I are using these
bright ancient objects as background light sources to illuminate matter
intervening at cosmological distances between the QSO and us. The whole
setup up works very much like a gigantic slide projector of cosmic
dimensions: gas clouds in galaxies or in the intergalactic medium (IGM)
absorb light in a characteristic pattern from the beam of the
background QSO, just as slides absorb certain colors from the white
light of a projection bulb. The particular absorption pattern tells the
astronomers how far the gas cloud is away from us, how much gas there
is, and how fast it is moving internally. We can also measure its
temperature and see whether it has already been enriched with chemical
elements by stars.
The picture above shows the spectrum of one such
QSO (1422+231) at redshift 3.62, when the universe was only 3
Giga-years old, about 20 % of its present age. The many ripples on top
of the smoothly varying continuum are individual absorption lines
caused by the neutral hydrogen Lyman alpha transition. There are
hundreds of individually distinguishable absorption lines, all caused
by the same transition at 1215.67 Angstrom. We are seeing this absorption
line repeated all over the place because the universe is expanding so the gas clouds are receding from us,
each cloud at a slightly different velocity, which causes a Doppler-shift
in the wavelength of the absorption line. The clouds recede the faster the further they are away from us, according to Hubble's Law.
This phenomenon is also known as the Lyman alpha forest. Below are a number of topics my collaborators and I have been working on over the
past few years.
How the Cosmic Web Moves
It is now generally believed that most of the ordinary matter in the
universe resides in the form of a connected network of galaxies and gas, the so-called cosmic web (see below).
Galaxies are embedded in a gaseous matrix (the intergalactic medium or IGM) consisting mostly of a hydrogen-helium plasma; they grow partly by accreting gas from the IGM and partly by merging with other galaxies.
The traditional picture of galaxies as "Island Universes", separated by vast
oceans of nothingness, has been superseded by a scenario where galaxies
are more like raisins in a cake, with most of the matter in the universe
being in the "dough", i.e., the intergalactic medium, as opposed to having fallen into the galaxies themselves.
Until recently, very
little was known about how the cosmic web at high redshift actually moves.
Does it strictly follow the general expansion of the universe, or can we
see the pull of gravity directing streams of gas into future galaxies and
galaxy clusters ?
My collaborators and I have been attempting to measure the kinematics of the gaseous cosmic web at redshifts (2-3.5) to answer these questions.
We observed hundreds of absorption systems that simultaneously appeared in the lines of sight to a number
of close pairs of QSOs and measured the velocity differences between the
lines of sight for each absorber, as a function of perpendicular distance across the sky.
The spatial separation between the points where the absorption was measured are large enough that these absorbing "clouds" generally must
be large scale filaments or sheets of gas, rather than galaxies.
The velocity differences (stricly speaking we only observe the one-dimensional velocities
projected along the lines of sight to the QSOs) allow us to estimate the gradients in
the velocity field between two points in space. On large enough scales
most of these gradients
are expected to be due to the stretching of the cosmic web, following the general expansion of the universe, but this cannot
be the whole story: without
local departures from a uniform Hubble flow (i.e., regions in the universe breaking away from
the expansion) there could never be any galaxies, for example.
Our velocity measurements showed that most of the IGM does indeed follow
the Hubble expansion quite closely, but there are important departures:
On scales of a few tens to a couple hundred kiloparsec the average expansion velocity of the IGM falls short of the general expansion by up to about 20 percent
(at the lowest redshifts (z~2) that we observed). This average is dominated
by a few absorbers that appear to be contracting (perhaps into galaxies).
Our data are not yet good enough to draw further conclusions, but a look at
a computer simulation of the IGM reveals that the picture is likely to
be more complex yet: the average expansion velocity in the simulation is indeed
somewhat smaller than the Hubble flow, but
the majority of the clouds appear to be expanding even faster (by 5 - 20 percent)
than the general universe. This is probably because the absorption systems we observed
most often are likely to arise in gaseous filaments that are stretching under the pull
of gravity. We may be seeing the gas streaming into galaxies or galactic clusters at either end of the absorbing filaments.
Publications:
Rauch, Michael; Becker, George D.; Viel, Matteo; Sargent, Wallace L. W.; Smette, Alain; Simcoe, Robert A.; Barlow, Thomas A.; Haehnelt, Martin G: Expansion and Collapse in the Cosmic Web, Astrophysical Journal, 632:58, 2005
Small Scale Structure in the Intergalactic
and Interstellar Medium
Occasionally distant QSOs are gravitationally lensed by foreground
galaxies and the lensing effect produces multiple images, which are sometimes
separated by several arseconds on the sky. Taking spectra of these
slightly displaced images enables us to probe the intervening intergalactic
and interstellar matter at high redshift on very small (parsec to kiloparsec) scales . With Wal Sargent and Tom Barlow (Caltech) I have thus attempted to measure
the texture and the turbulence of a variety of gaseous environments
at high redshift. It appears that most of the IGM (by volume) has little
velocity or spatial structure on scales smaller than a kiloparsec.
In contrast the ubiquitous, somewhat denser, high ionization gas producing the CIV absorption
systems in QSO spectra does show some residual turbulence on similar
scales. We interpret this tentatively as evidence for galactic winds
or gas stirred by ram pressure from moving galaxies. Going to even denser,
low ionization gas (commonly seen in absorption by singly ionized magnesium, "MgII") the gas becomes much more turbulent. For MgII systems, unlikely the other two classes of absorbers we have not found yet a minimum size
below the clouds may be featureless. In one case of a very low ionization
system at redshift 3.5 we observe column density
differences by an order of magnitude over a distance of only 26 parsecs.
The object is consistent with an individual old supernova remnant (but other
interpretations may be possible). Thus, with the highest density, lowest ionization
absorbers we seem to probe conditions directly in the interstellar medium
of high redshift galaxies.
Publications:
Rauch, M., Sargent, W.L.W., Barlow, T.A,, Simcoe, R.A.: Small-Scale Structure at High Redshift. IV.
Low-Ionization Gas Intersecting Three Lines of Sight to
Q2237+0305, Astrophysical Journal, 576:45-60, 2002
Rauch, M., Sargent, W.L.W., Barlow, T.A, Carswell, R.F.:Small-Scale Structure at High Redshift. III. The
Clumpiness of the Intergalactic Medium on Subkiloparsec
Scales, Astrophysical Journal, 562:76-87, 2001
Rauch, M., Sargent, W.L.W., Barlow, T.A.: Small-Scale Structure at High Redshift. II. Physical
Properties of the C IV Absorbing Clouds, Astrophysical Journal, 554:823-840, 2001
Rauch, M., Sargent, W.L.W., Barlow, T.A.: Small-Scale Structure at High Redshift. I. Glimpses of the Interstellar Medium at Redshift 3.5, Astrophysical Journal, 515:500-505, 1999
Galaxy Formation from Gas
Going far enough back in time there must have been a period when most
galaxies had not been in place yet, and the universe consisted mostly
of gas. When and how did galaxies form from the gas ? The currently
most popular theoretical model of galaxy formation, the Cold Dark
Matter (CDM) scenario and its variants envisage galaxies growing
gradually over an extended period of time, rather than by a sudden
collapse at high redshift. In this picture galaxies grow both by
continuing accretion of gas and by merging of smaller galactic building
blocks which may already have formed stars by the time they collide. Up
to a few years ago, virtually all observational efforts in the study of
galaxy formation went into observing the stellar populations and the
merger process. This was because the primary process, the collapse of
galaxies from gas is hard to observe as the gas is very tenuous and
does not produce enough radiation to be seen in its own light, unlike
stars.
It had been speculated for quite some time that QSO absorption systems
may hold the key to observing the primary, gaseous part of galaxy formation, when finally
advances in computer hardware and software development led to a minor revolution
in numerical astrophysics.
By the early
1990s gas-dynamical computer simulations could be performed with a spatial
and mass
resolution sufficient to study the formation of individual galaxies,
starting from the initial cosmological density fluctuations. My
colleagues Haehnelt, Steinmetz and I realized that the gaseous
environment of these simulated galaxies was conducive to producing QSO metal
absorption lines, and we investigated the possibility that at least
part of the metal absorption systems in the spectra of QSOs are caused
directly by the infalling gas feeding the growing galaxies, and by the
gaseous matrix in which the young galaxies are embedded. Assuming that
an early phase of metal enrichment pre-polluted the gas uniformly with
a small amount of metals (so as to produce atomic transitions of metal
ions in the gas) we could show that gas falling into galaxies can
produce remarkably realistic QSO metal absorption systems (in
particular, CIV and damped Lyman alpha absorbers). The distributions of
temperature, density and ionization state of the gas came out quite
well. The large scale structure of the gas and galaxy distribution
peculiar to the CDM model
(knotty filaments of gas with many small protogalactic clumps
embedded) also nicely explained at least some of the often large velocity widths
of metal absorbers as being due to peculiar motions among the proto-galaxies
and streaming gas within the filaments constituting the gaseous reservoir
from which the galactic subunits form.
Publications:
Rauch, M., Haehnelt, M.G., Steinmetz, M.: QSO Metal Absorption Systems at High Redshift and the Signature of Hierarchical
Galaxy Formation, Astrophysical Journal, 481:601-624, 1997 (3.8Mb(!) PDF)
The Large Scale Distribution of Baryons in the Universe
One of the fundamental questions in astronomy is : "where are the
baryons ?" Big Bang nucleosynthesis calculations predict that there
should be about ten times as much ordinary matter in the universe than
we can account for locally by looking at the matter contents of our
solar system, Galaxy etc. (this is different from the so-called "dark
matter" problem). Searches for the missing baryonic matter have usually assumed that the baryons are in compact
stellar objects (from planets to supermassive black holes), but none of the different candidate classes has proven
to contain more than a small fraction of all matter.
The possibility
that most of the ordinary matter may not be in galaxies at all but in
the form of gas in intergalactic space has received serious attention
only in the early 1990s. This was partly because the most sensitive
method for detecting baryons, i.e., looking for absorption of the light
of background QSOs by the intergalactic matter (the Lyman alpha
forest, see above) shows only the neutral fraction of the gas (ionized
hydrogen does not absorb light significantly). Unfortunately, only a
tiny fraction of all baryons are expected to be neutral under the
physical conditions prevailing in most of the cosmic volume. Most of
the gas is likely to be at a very low density, and is being kept
ionized by the pervasive UV background radiation field. Thus, to
estimate the total amount of gas that corresponds to the observed
neutral gas one needs to have independent estimates of the density
and the intensity of the ionizing radiation field. There are various
ways for estimating the radiation intensity to probably within a factor
of five so the main uncertainty has always been the density of the gas. If
most of the gas seen from its neutral hydrogen absorption in the Lyman
alpha forest were very tenuous the ionization could be extremely
high, and an enormous amount of matter could be hidden in the form of
ionizing gas. Conversely, if the intergalactic medium were quite
clumpy most of the gas could be neutral and what you observe is really all
there is.
A breakthrough in this subject occurred when several groups
(e.g., Smette et al; Bechtold et al, Dinshaw et al) began to measure
the lateral extent of the Lyman alpha clouds in the intergalactic
medium, from observations of multiple lines of sight to QSO pairs and
lensed QSOs. The cloud sizes (at least at high redshift) turned out to
be on the order of hundreds of kiloparsecs up to a Megaparsec.
That makes the gas densities in these clouds quite low (within a
factors of ten of the mean density of the universe) and the total
baryon contents of the intergalactic medium becomes very substantial.
In fact, if the clouds were spherical (i.e., as extended along the line
of sight as across the sky - the above size estimates measure only the
latter) then the clouds would contain more baryons than the universe
as a whole. This paradox can be resolved by assuming that the gas
clouds typically cannot be spherical but must be highly flattened
filamentary or sheet-like structures (Rauch & Haehnelt 1995). This
qualitative observational picture is in essential agreement with the
results from numerical cosmological simulations. Such models predict
that the whole universe is permeated by a "cosmic web" of gas in which
the galaxies are embedded. The simulations predict the density
distribution in the universe, so the only free parameter left when
determining the total baryon contents is the strength of the ionizing
radiation field. At those redshifts were the Lyman alpha forest is
most easily observed (z ~ 3) it is thought that QSOs make the dominant
contribution to the ionizing background. Adding a radiation field of
the requisite strength to the cosmological simulations one can predict
the amount of baryons necessary to produce a Lyman alpha forest
absorption pattern as strong as the one actually observed. Comparisons
of such simulations (performed by a number of groups: Cen et al;
Petitjean et al; Hernquist et al; Miralda-Escude et al; Zhang et al) to
actual data (Rauch et al 1997) show that the intergalactic medium may
indeed still contain on the order of 90 percent of all matter at
redshift 3.
Publications:
Rauch, M., Miralda-Escude, J. Sargent, W.L.W., Barlow, T.A.,
Weinberg, D.H., Hernquist, L., Katz, N. Cen, R., Ostriker, J.P.: The Opacity of the Lyman Alpha Forest and Implications for Omega_baryon and the Ionizing Background, Astrophysical Journal, 489:7-20, 1997
Rauch, M.; Weymann, R. J.;
Morris, S. L.: Are Lyman-Alpha Clouds Associated with Low Surface Brightness Galaxies?, Astrophysical Journal, 458:518-523, 1996
Rauch, M., Haehnelt, M.G.: Omega_baryon and the Geometry of Intermediate-Redshift Lyman Alpha Absorption Systems, MNRAS, 275:76-78, 1995
The Nature of Damped Lyman Alpha Systems
Damped Lyman alpha absorptions systems (DLAS) in QSO spectra get
their name from the (Lorentzian) damping wings of the Lyman alpha line
of neutral hydrogen visible in some very high column density QSO absorption systems.
Pioneering work by Art Wolfe and his collaborators has elucidated the
basic properties of DLAS (column densities, neutral gas content,
overall contents of neutral hydrogen similar to the amount of matter
nowadays in stars), which are consistent with DLAS arising when a QSO
line of sight intersects the disk of an early massive disk galaxy (the
disk of the Milky Way also produces a DLAS against background QSOs).
This work revealed that DLAS, just like disk galaxies nowadays, contain indeed most of the neutral
gas in the universe (but not most of the gas; that appears to be in a
highly ionized form in the low density IGM).
However, there always was the question whether such massive galaxies
really exist at high redshift or whether young, primordial galaxies
perhaps look fundamentally different from present ones. For
example, the metal abundances of DLAS are lower than those of present
day disk galaxies, as Pettini and Lu and their collaborators have
demonstrated. Moreover, many of the other properties of matter at high
redshift (as observed, e.g., in the Lyman alpha forest or the Cosmic
Microwave Background) agree better with a picture of hierarchical
structure formation, where most high redshift galaxies are small and
each modern galaxy is an agglomeration of multiple predecessors at high redshift.
Our simulations of gaseous collapse in the vicinity of newly forming
galaxies (with Haehnelt and Steinmetz; see previous
paragraph) have shown that DLAS automatically form as a
by-product of hierarchical structure formation. In contrast to the
older standard picture (massive disks produced by monolithic collapse
at high redshift) the simulations revealed DLAS to be mostly small galaxies
with immature disks (the disks do not have much time to grow because of
the large merger rates at high redshift).
Although initially it was claimed that the smaller CDM protogalaxies
cannot produce the relatively large velocity widths of DLAS (Prochaska
and Wolfe, 1997) we could show that they indeed can, with the largest
velocity widths seen in DLAS being due to chance alignments of
multiple small protogalactic clumps along the line of sight to
the background QSO as opposed to the rotation of an individual
monolithic disk galaxy. This picture of DLAS as small building blocks
seems to have found general acceptance now (even among its erstwhile
opponents), especially after imaging of DLA absorbers at low redshift
with the Hubble Space Telescope (by Bergeron and collaborators)
revealed that not even at the present epoch is there a unique
correspondence between large disk galaxies and DLAS.
Publications:
Haehnelt, M.G., Steinmetz, M., Rauch, M.: Damped Ly Absorber and the Faint End of the Galaxy
Luminosity Function at High Redshift, Astrophysical Journal, Astrophysical Journal 534:594-597, 2000
Haehnelt, M.G., Steinmetz, M., Rauch, M.: Damped Lyman Alpha Absorber at High Redshift: Large Disks or Galactic Building Blocks ?, Astrophysical Journal, 495:647-658, 1998
Haehnelt, M.G., Steinmetz, M., Rauch, M.:C IV Absorption from Galaxies in the Process of Formation, Astrophysical Journal, 465:L95-L98, 1996
Personal Interests
somewhen to come....
last updated November 12, 2002