Project Topics
Get Complete Project Material File(s) Now! »
SDSS-III Baryon Oscillation Spectroscopic Survey
Survey Goals
A legacy project of the Sloan Digital Sky Survey (SDSS, York et al. 2000), the Baryon Oscillation Spectroscopic Survey (BOSS, Dawson et al. 2013) continued the SDSS tra-dition of investigating the underlying structure of the universe through observations of galaxies and quasars. As its name implies, the primary goal of BOSS was to measure the characteristic scale of baryon acoustic oscillations (BAOs) that arise from small varia-tions in the density of the primordial universe. An overdensity of photons at points that had a slightly higher matter density created a sound wave driven by radiation pressure. Highly-ionized baryons were tightly coupled to the photons and traveled outward from their initial location with the sound wave. When photons decoupled during recombina-tion, the baryons that were traveling with the acoustic oscillation remained at the final position of the sound wave. As gravity again became the dominant force, the baryon overdensity at this position skewed the dark matter distribution. When measuring the power spectrum of the baryon distribution, a peak should be seen at the typical scales corresponding to the sound waves. BOSS used these scales as standard rulers.
Eisenstein et al. (2005) first detected a BAO peak at the 100 h−1 Mpc (comoving) scale in the correlation function of z ≃ 0 35 luminous red galaxies from the SDSS-I (Figure 2.1). The original spectroscopic sample included 46 748 galaxies in the range 0 16 < z < 0 47. The shape of the galaxy correlation function provides a constraint on the matter density that is independent of the measurement from the CMB anisotropies. Furthermore, when measured from the three-dimensional clustering of matter tracers at redshift z, the transverse BAO scale constrains the angular diameter distance dA(z), and the line-of-sight scale constrains the Hubble parameter H(z) (Figure 2.2). Using the Hubble parameter and the comoving angular diameter distance evaluated at z = 0 35 with the LRG sample and at z = 1089 from WMAP results, Eisenstein et al. (2005) constructed a distance ratio for two extremely different redshifts. This distance ratio provided an argument for dark energy, based on a geometric comparison. The detection of BAOs in the galaxy correlation function laid the foundation for the SDSS-III BOSS project, which will expand the potential for the BAO measurement.
Over the course of five years, from Fall 2009 to Summer 2014, SDSS-III BOSS mapped the spatial distribution of luminous red galaxies and high-redshift quasars (Dawson et al. 2013; Eisenstein et al. 2011). The collaboration planned to obtain spectra of 1.5 million luminous galaxies extending to z = 0 7 and 160 000 quasars primarily in the range 2 15 ≤ z ≤ 3 5. With this data, the goal was to measure the location of the BAO peak at low redshift from the galaxy correlation function and at high redshift from correlated absorptions in the Lyα forest. Since the Lyα forest absorptions indicate the amount of hydrogen along the quasar line-of-sight, the three-dimensional correlation function of the transmitted flux can reveal the underlying dark matter distribution. If the sample of background quasars is sufficiently large and sufficiently dense, a peak in the distribution due to BAOs may be detected (McDonald & Eisenstein 2007). By measuring the BAO scale from many quasar and galaxy spectra at multiple epochs, BOSS intended to determine the distance scale and Hubble expansion rate at the one-percent level.
Obtaining a surface density of 15 quasars at z ≥ 2 15 per square degree with gP SF < 22 0 was crucial for adequately measuring the correlation function from transmitted flux in the Lyα forest. Although the surface density of z ≥ 2 15 quasars to the BOSS magnitude limit is approximately 28 per square degree, (Palanque-Delabrouille et al. 2013a), cor-rectly selecting targets was challenging. In ugriz color space, the quasar locus overlaps with the stellar locus in the range 2 2 < z < 3 5 (e.g., Ross et al. 2012). A uniform sample was not necessary for the Lyα forest BAO analysis, since quasars are probes of the IGM and do not affect its large-scale distribution; however, statistical studies related to the quasars themselves, like measuring the quasar luminosity function or clustering analyses, do require a uniform sample. BOSS therefore incorporated multiple target selection strategies optimized to produce (primarily) CORE and BONUS samples. Ad-ditional samples, including previously known quasars, radio-loud quasars, and ancillary programs made up ∼6% of targets. The CORE sample is uniform, whereas the BONUS sample was designed to maximize the spectroscopic quasar density. Initially, the CORE sample was targeted from a likelihood method (Kirkpatrick et al. 2011), which used the photometry and models for the stellar and quasar locus to determine the likelihood that each object is a quasar. After the first year of operations, an extreme deconvolution method (Bovy et al. 2011) that incorporates photometric uncertainties into a density estimation of stars and quasars took over the CORE sample targeting. The BONUS sample was based also on the likelihood and extreme deconvolution methods, as well as kernel density estimator (Richards et al. 2009) and neural network (Y`eche et al. 2010) techniques. The target selection strategy for the first two years of BOSS is presented in Ross et al. (2012).
The full SDSS-III collaboration included over 330 people with groups from France, Ger-many, Spain, and Brazil contributing to the projects. The French Participation Group, of which I am a member, focused on the quasar component of the BOSS project and was particularly invested in target selection, visual inspection, quasar catalogs, and the Lyα forest analysis.
Telescope and Spectrograph
The 2.5 m SDSS-III telescope and 1,000-fiber BOSS twin spectrographs are located at Apache Point Observatory in New Mexico, where the median seeing is 0.9′′. Aluminum plates with a 3◦ diameter were drilled with the position of target galaxies, stars, quasars, or blank sky areas to a accuracy of 9 µm (Figure 2.3). About 10% of holes on a plate were dedicated to sky calibration. In preparation for observing, plates were mounted on interchangeable cartridges that attach to the telescope, and the spectroscopic fibers, which have a 2′′ diameter, were plugged by hand (Figure 2.4). The observing time for one BOSS plate was about an hour under excellent conditions, which included three fifteen minute science exposures, calibration observations, and the time required to swap one plate cartridge for another on the back of the telescope. On exceptional nights, up to nine plates could be observed. However, under typical conditions, usually four to six science exposures were required to obtain sufficient S/N per pixel at a fiducial magnitude, and the observing time per plate was longer. More than 2,000 unique plates were observed over the course of the survey.
The spectra have a wavelength range of 3600 A to 10 000 A, and the spectral resolution varies from ∼1500 – 3000, such that the instrument velocity dispersion is consistently ∼150 km s−1. The two BOSS spectrographs each received 500 fibers, which were treated identically, regardless of which spectrograph they fed. To achieve a consistent spectral resolution across the entire wavelength range, a beamsplitter in each spectrograph sent incident light to two CCDs optimized separately for the red and blue ends of the spec-trum. The blue CCD covered ∼3600 − 6350 A, while the red CCD detected light at∼5650 − 10 000 A. The two ends of the spectrum then had to be reconciled by calibrat-ing the flux at the overlapping wavelengths. Smee et al. (2013) fully characterize the two double spectrographs.
Data Processing and Quasar Redshift Estimation
Each BOSS spectrum has a unique identifier based on the PLATE, MJD, and FIBERID. The PLATE number specifies the plate drilled for a particular target. Since the same plate may have been plugged and observed on multiple occasions, with different mappings between fibers and target holes, MJD indicates the modified Julian date for a unique fiber configuration. Additionally, the spectroscopic pipeline (Bolton et al. 2012) numbered each fiber arriving at the slit-head with the sequential index FIBERID, 1 to 500 for one side and 501 to 1000 for the other. During the spectroscopic data reduction process, all good data from a unique plugging were co-added together.
Table of contents :
1 Introduction
1.1 A Brief Cosmic History
1.1.1 Growth of Galaxies
1.2 Absorption Systems
1.2.1 N(Hi) Frequency Distribution
1.2.2 Characterizing Strong Absorbers
1.2.2.1 Neutral Hydrogen Mass Density
1.2.2.2 Metallicity
1.3 Relating Absorption Line Systems and Galaxies
1.3.1 Mg ii Systems
1.3.2 CGM
1.3.3 DLA Systems
1.3.4 DLA Galaxies in Simulations
1.3.5 Observing DLA Galaxies
1.4 Thesis Work
2 SDSS-III Baryon Oscillation Spectroscopic Survey
2.1 Survey Goals
2.2 Telescope and Spectrograph
2.3 Data Processing and Quasar Redshift Estimation
2.4 Data Releases
2.5 Value-Added Catalogs
2.6 Quasar and Absorption System Science
2.7 Survey Results
2.8 Future Prospects
3 A glance at the host galaxy of high-redshift quasars using strong damped Lyman-alpha systems as coronagraphs
3.1 Introduction
3.1.1 Observing Quasar Host Galaxies
3.1.2 Associated DLAs
3.2 Sample Definition
3.2.1 Strong associated DLAs
3.2.2 Measuring DLA column densities and emissions
3.2.3 The Statistical Sample
3.2.4 The Emission Properties Sample
3.2.5 The Redshift Distribution
3.3 Anticipated Number of Intervening DLA Systems within 1 500 km s−1 of zQSO
3.3.1 Anticipated Number in DR9
3.3.2 The Effect of Clustering near Quasars
3.4 Characterizing the Statistical Sample
3.4.1 Scenario
3.4.2 Kinematics
3.4.3 Metals
3.4.4 Reddening
3.5 Characterizing the Narrow Lyα Emission
3.5.1 Correlation with other properties
3.5.2 Position and profile of the emission line
3.5.3 Comparison with Lyman Break Galaxies
3.5.4 Comparison with Lyman-Alpha Emitters
3.5.5 Comparison with Radio Galaxies
3.6 DLAs with partial coverage
3.7 Discussion and Conclusions
3.8 Follow-Up Projects
3.8.1 Observations with Magellan/MagE
3.8.2 Observations with VLT/X-shooter
3.8.3 Observations with HST/WFC3
4 Close Line-of-Sight Pairs
4.1 Introduction
4.2 Close Line-of-Sight Pairs in SDSS-III BOSS
4.2.1 The Transverse Correlation Function for Lyα Forest Absorptions
4.2.2 Quasar Host Galaxy Environments
4.3 VLT/X-shooter Follow-Up Observations
4.3.1 Pair SDSS J0239-0106
4.3.2 Pair SDSS J2338-0003
4.3.3 Pair SDSS J0913-0107
4.4 Preliminary Conclusions
5 A ∼6 Mpc overdensity at z ≃ 2.7 detected along a pair of quasar sight lines: filament or protocluster?
5.1 Introduction
5.2 Data
5.3 Absorption Systems
5.3.1 Background Quasar Line-of-Sight
5.3.1.1 Hi Absorption Systems
5.3.1.2 Abundances
5.3.2 Foreground Quasar Line-of-Sight
5.3.2.1 Hi Absorption Systems
5.3.2.2 Abundances
5.4 Discussion and Conclusions
6 Conclusions and Prospects
6.1 Conclusions
6.2 Prospects
6.2.1 Follow-Up Projects
6.2.2 Future Work
