The Sloan Digital Sky Survey (SDSS) is little known to the public, but represents one of the most-challenging efforts in observational cosmology ever attempted. The most recent phase, SDSS-III, began in 2008 and includes the Baryon Oscillation Spectroscopic Survey (BOSS), a part of SDSS-III aimed at mapping the cosmos. Its goal is to map the physical locations of all major galaxies back to seven billion years ago, and bright quasars back to 12 billion years ago – two billion years after the Big Bang. This is being done so we can gain a better understanding of dark matter and energy, and hopefully encounter a few surprises.

Measuring the position of an object projected on the sky is straightforward – what is difficult is measuring its distance. This is accomplished by measuring the redshift of the object, and converting the redshift to distance. Redshift can be thought of as being to light waves what the Doppler effect is to sound waves. Similar to the change in the pitch of a siren as an emergency vehicle passes, first approaching and then pulling away from our position, light waves from galaxies undergo a similar shift to the longer wavelengths (colors become redder) as they move away from us due to the expansion of the Universe.

Redshift is defined as z, which is equal to the ratio of the observed wavelength of a spectral line to the actual wavelength minus 1. The most distant galaxies observed by Hubble had a redshift of z~0.003, while the limit of the BOSS survey is z~3. The relation between redshift and distance is linear up to about z=0.3 (three billion light years distant), after which general relativistic effects force a nonlinear relationship, the details of which provide substantial clues on the matter, energy, and fundamental laws of the Universe.

Baryon acoustic oscillations appear as rings of larger galaxy density. Following these rings into the past revels a great deal about how the current configuration of the Universe came about (Image: SDSS-III)

Of particular interest is measuring the size of baryon acoustic oscillations as a function of time. Baryons are a type of elementary particle, including protons and neutrons. In the very early Universe, photons and baryons are so tightly coupled that they form a nearly incompressible liquid. This coupling also prevents photons from traveling, so that during this period the Universe is opaque. If you drop a rock into a pond, you will see rings move outward from the point of impact. Similarly, a local disturbance in the photon-baryon fluid propagates outward, but at about half the speed of light.

As the Universe expands, there is enough room for the photon-baryon liquid to break apart into separate particle fields. This is the era, about 400,000 years after the Big Bang, when the 3K cosmic background radiation was emitted into the Universe. When the particles break apart, the photon-baryon "sound waves" stop, leaving a spherical shell of large baryon density where the sound wave ended. Those denser regions then expand along with the Universe, but continue to have more matter than average. Currently, this "sound horizon" has expanded to about 500 million light years. As looking deeper into space is equal to looking into the past, it is possible to measure the diameter of the sound horizon as a function of time, and to extract therefrom the rate of expansion as a function of time.

For many decades, the rate of Universal expansion was thought to be constant, or possibly slowing, as there is no mechanism in general relativity or in the Standard Model of particle physics which could cause acceleration of expansion. However, detailed studies of Type IA supernovae at cosmological distances has recently provided clear evidence for such acceleration. This discovery, for which the 2011 Nobel Prize for Physics was awarded, is the reason you hear about the unknown "dark energy" – it is required to explain accelerating expansion of the Universe within our current cosmological models. Measuring the rate of expansion using the baryon acoustic oscillations is one of the most reliable methods we have to study this unexpected phenomenon.

A 2-D slice through our Universe showing galaxies up to two billion light years away. The filaments and clusters now known to be formed by the distribution of dark matter in the Universe can be clearly seen (Photo: SDSS-III)

BOSS is examining roughly a quarter of the sky, so the maps are also suitable for studying the filament-like structure of the Universe at large sizes, thought to be the result of dark matter filaments, which have recently been discovered. Now, midway through its search, SDSS-III has determined the position and distance of about half a million galaxies and one hundred thousand quasars. Images are being taken revealing objects down to magnitude 23 – the brightness of the Sun as seen from a distance of 140,000 light years.

The 2.5 meter (8.2 foot) SDSS telescope at Apache Point Observatory in southern New Mexico (Photo: SDSS-III)

The main tool supporting BOSS is a dedicated 2.5 meter (8.2 foot) wide-angle optical telescope at Apache Point Observatory in New Mexico. The location was chosen owing to its altitude of 2,788 meters (9,147 feet) and a semi-arid climate that provided 65 percent clear observing nights.

The telescope is equipped with a spectrograph capable of taking as many as 1,000 spectrograms simultaneously of objects across a three degree field of view. The spectrograms are recorded by two CCD arrays with quantum efficiency greater than 80 percent. The survey process generates a lot of data – about 200 GB per night.

Edwin Hubble's 1929 discovery of the relationship between cosmological distance and redshift was based on spectra of galaxies within about seven million light years of Earth, about 2,000 times closer than the quasars of interest to the BOSS study. By coincidence, Hubble also carried out his measurements of galactic distance using a 2.5 meter (8.3 foot) telescope. Why could he not carry out the BOSS survey, given that his telescope was of equal size? Limited technology forced Hubble to use photographic plates, take one spectrograph at a time, and record and analyze his data by hand.

Shown here are three different "plug plates", which hold the spectrograph's fiber optics in position to gather light from the desired objects. Below is a photo of the fiber optics feeding the light into the spectrograph itself (Photo: SDSS-III)

Clearly, taking one spectrogram at a time is much slower than taking 1,000 simultaneously. The BOSS spectrograph works with more than two thousand large metal plates that are placed at the focal plane of the telescope. These plates are drilled at the precise locations of nearly two million galaxies and quasars. Optical fibers are plugged into a thousand tiny holes in each of these "plug plates." Once the telescope is precisely aligned with the sky, the fibers carry light from each observed galaxy or quasar to BOSS's new spectrographs.

Another important factor in the speed of carrying out the survey is the improvement of CCD cameras over photographic plates. Modern CCD cameras have quantum efficiency for photon detection around 50 times larger than did Hubble's photographic plates. In addition, photographic emulsions become slower with long exposures. The three-night exposure required by Hubble to capture an extremely dim galactic spectrum could be made in about an hour if emulsions maintained their short-exposure speed.

The imaging camera on the Apache Point 2.5 meter telescope, which has a total of 226 megapixels and a resolution of 0.4 arc-seconds per pixel - sufficient to resolve the major topography of Titan (Photo: SDSS-III)

The emulsion slow-down factor of 25 is multiplied by the 50x CCD sensitivity to suggest that a CCD camera is more than a thousand times more sensitive than photographic plates for recording very dim images. As a thousand galactic spectra can now be captured at a time, modern astronomical surveys can be carried out about a million times faster than in the 1920s. An additional factor is that it would have been impossible even 15 years ago to record the observational data from the BOSS survey. Making sense of 200 GB of data each night is a daunting prospect without teraflop supercomputers and inexpensive disc storage.

At this point in the BOSS survey, nearly 100,000 new quasars have been discovered, including many of the most distant known quasars. The baryon acoustic oscillations have been analyzed, and the relation between redshift and distance refined to yield distances accurate to better than one percent throughout the sway of the BOSS survey. A large number of sub-stellar objects ("failed stars") have been discovered in the Milky Way galaxy – many more than expected. Dark matter halos of galaxies have been studied through their gravitational lensing effects. The dark matter structure of the early Universe has been studied using quasar absorption lines. The successes and surprises of the SDSS survey have been a great blessing to astronomers, astrophysicists, and cosmologists alike.

One result of these analyses is the movie below, a video fly-through of the SDSS-III galaxies. Each galaxy in the animation is placed at the location mapped by SDSS and is represented by an actual image of the galaxy. Galaxies are clearly concentrated into clusters and filaments with voids in between. The SDSS-III is exploring this structure to determine the nature of dark energy and the distribution of dark matter in the Universe. Those of us who sit by and marvel at such results are expecting future occasions for further sitting and marveling.

Source: Sloan Digital Sky Survey