3.5. Redshift Surveys: Defining Physical Structures
The previous manifestations of structure in the Universe were all
based on the distribution of galaxies as projected onto the plane
of the sky. This was the state of the field as of 1980. To
make further progress it was necessary to move beyond projected
galaxy distributions and into the realm of three-dimensional
data sets. To do this necessitated the acquisition of galaxy
redshifts. This is a somewhat daunting observational challenge.
If the galaxies have lots of neutral hydrogen, then the Arecibo
radio telescope will be able to detect that galaxy and get a
redshift in approximately 10 minutes of observing time, for galaxies
with velocities 
10,000 km
s-1. Unfortunately, the Arecibo
telescope can only observe galaxies that are located between
declinations of 0 and 40°. Nonetheless, throughout the 1980s,
Arecibo staff astronomers, specifically Martha Haynes and Riccardo
Giovanelli, measured redshifts for thousands of galaxies that
are accessible in the Arecibo declination strip. Those observations
were the first to define a large scale feature called the Pisces-
Perseus supercluster (see Giovanelli and Haynes 1985; Giovanelli
et al. 1986), which is a large,
"sheet-like" structure with mean redshift 5500 km s-1.
Optical redshifts of galaxies can be much harder to obtain. If
the galaxy has active star formation then it will be rich in
emission lines from H II regions. Emission line redshifts require
little integration time as only the emission line needs to be
detected instead of the continuum light from the galaxy. In this
case, using a telescope of diameter
1.5-m, a redshift can
be obtained in about 5 minutes of integration time. Approximately
20% of all galaxies have sufficient emission line strength to
allow for this rapid determination of redshift. For most
galaxies redshifts
must be determined from the absorption lines which come from the
integrated contribution of all the stellar atmospheres in the
galaxy. The principle absorption lines that can be detected are
the H and K lines of Calcium at 3933 and 3963 Å, the
Magnesium I line at 5175 Å and the Sodium D lines at 5800 Å.
A typical absorption line spectrum of a galaxy is shown in
Figure 3-9.
On a 1.5 telescope an exposure time of 15-60 minutes is required
to get sufficient signal-to-noise to determine an absorption line
redshift. To get redshifts of thousands of galaxies in this manner
required a dedicated telescope, patience, and an institutional
commitment to performing this kind of service for the community.
|
Figure 3-9: A typical absorption line spectrum of a galaxy showing the many absorption lines in the spectrum. The principal lines used for redshift determination are the Calcium H and K lines at wavelength about 3900 angstroms, the Magnesium I complext at 5175 angstroms, and the sodium D lines at 5800 angstroms. Image courtesy of Jeff Willick. |
These qualities were found in the form of Marc Davis, Steve Shectman
and John Huchra
at the Smithsonian Astrophysical Observatory. In the late 70's,
Davis and Shectman designed a redshift machine for the 60-inch telescope at Mt.
Hopkins. This spectrograph was optimized to detect the Mg I and Sodium
D lines for purposes of determining a redshift and was known as the
Z-machine. In 1979
this telescope plus detector was commissioned and the first redshift
surveys were made. The pursuit of galaxy redshifts requires dedicated,
patient and persistent observers. All of these qualities were found in
Huchra and together with Davis, the performed the first redshift
survey which contained 2500
galaxies brighter than b = 14.5.
After this survey was completed, Margaret Geller joined the project
and brought a fresh theoretical insight that allowed this rich data
set to blossom into the first clear view and characterization of
the 3D galaxy distribution. Much of this was based on a redshift
survey of 20,000 galaxies brighter than b = 15.5. This basic dataset,
known as the CFA Redshift Survey has been second only to
the observations of the CMB in terms of its historic observational value to
the field of cosmology. It is also safe to say that when this
survey started, no astronomer could have predicted the complexity of
the galaxy distribution that would be revealed by an obscure 1.5-m
telescope located above the desert in Arizona.
The disadvantage of the CFA redshift survey instrumentation is that only one galaxy at a time could be observed. A revolution in galaxy redshift survey efficiency occurred in the mid 80's with the development of multi-fiber spectrography. In essence, galaxy coordinates are machined into a "plug-board" in which optical fibers are placed. Each optical fiber is fed to the spectrograph and the CCD detector. This allows for multiple objects to be observed in a single exposure. Multiple object spectrography can generally measure between 20 and 200 objects simultaneously which represents an enormous gain over the one galaxy at a time approach. Such an instrument is most valuable when measuring redshifts in a field where there are lots of galaxies. This could either be in a galaxy cluster or in some deep field where there are lots of faint galaxies. Initial work by Couch et al. (1985), Ellis et al. (1985), and Broadhurst et al. (1988) directly showed the tremendous advantage afforded by multiple object spectroscopy.
In tooling up to do the Southern Hemisphere equivalent of the CFA redshift survey, Steve Shectman took full advantage of multiple object spectroscopy and was able to measure as many redshifts that are in the CFA survey in a much shorter period of time. However, there are two potential problems with fiber-based spectroscopy:
Since the fibers can not be
put arbitrarily close together, there
is a selection effect against detecting close pairs of galaxies.
The S/N of the spectrum is
extremely dependent upon the central
surface brightness of the galaxy. A typical fiber has an angular extent
of 1-5 arcseconds, depending upon the camera optics. Galaxies of low
surface brightness are therefore greatly underrepresented in fiber surveys.
However, this two minor problems only impact studies of the small-scale clustering of galaxies. On a large scale, their impact is negligible.
There are now about
100,000 redshifts measured between the CFA survey and a complimentary
survey initiated in the southern hemisphere a few years ago by
Steve Shectman. In a few years, the Sloan Digital Sky Survey hopes
to have measured 106 galaxy redshifts although its hard to believe
this increase will significantly change the view of LSS that is presently
defined by 100,000 redshifts.