★ This work has been funded by #HST-HF2-51351
Rest-frame NUV spectroscopy with dark energy surveys
DATA: The composite spectra from the pilot eBOSS ELG survey are available here
☆ http://www.sdss.org/dr13/data_access/vac/ (search for composite spectra of emission-line galaxies)
The detection of the baryon acoustic oscillation (BAO) in the large-scale distribution of galaxies has shifted the focus of observational cosmology to massive redshift surveys over the entire observable sky. Among the BAO dark energy surveys are the Extended Baryon Oscillation Spectroscopic Survey (eBOSS) (Dawson & eBOSS 2016) within SDSS-IV, the Dark Energy Spectroscopic Instrument survey (DESI), and the Prime Focus Spectrograph survey (PFS). These programs will obtain optical spectra in the observer frame for tens of millions of galaxies at redshift 0.6<z<2.0, providing unprecedentedly rich spectroscopic data sets with the coverage in the rest-frame NUV.
Among the new BAO projects, the eBOSS survey started taking data in Fall 2014. One of the large-scale structure tracers in eBOSS is emission-line galaxies (ELGs). We have obtained about 12,000 ELG spectra across the redshift range 0 < z <1.5 in the pilot observations (Comparat & eBOSS 2016), conducted to optimize targeting strategies.
The image below is a composite spectral image of the ELGs from our pilot observations. Each row is a composite spectrum of about 50 ELGs at the same redshift and galaxies are from nearby to distant universe from bottom to top (from low to high redshift). The green color shows emission lines and relatively higher continuum level (e.g., in the FUV in the top-left corner). The blue color shows absorption lines and relatively lower continuum level (e.g., between the [O II] λλ3727, 3730 emission doublet and the Mg II λλ2796, 2803 absorption doublet). The whopping emission lines make redshift determination feasible even for faint distant galaxies. DESI and PFS, in particular, will target tens of millions of [O II] emitters at redshift z>1 with high-resolution spectroscopy (R>3000).
These ELGs are overwhelmingly star-forming galaxies (SFGs) based on their strong emission lines and line ratios. The spectrum above is the composite spectrum of all the ELGs from the eBOSS pilot observations, spanning wavelength range from 2000 Ã… to 7500 Ã….
The high redshift coverage of dark energy surveys has provided us with a rare opportunity to study the rest-frame NUV wavelength region of the SED of SFGs that otherwise requires space-based telescopes such as the great Hubble telescope. Below is the composite spectrum (continuum-normalized) in the wavelength region featuring Fe II λ2586, λ2600. The absorption lines are asymmetric and preferentially blue-shifted, signatures of ubiquitous star formation-driven outflows.
Another interesting feature is the non-resonant emission lines, e.g., Fe II* λ2613, λ2626, which otherwise are not seen in quasar absorption-line systems. In the outflow model described in Zhu & eBOSS (2015), these lines are fluorescent emission after the occurrence of absorption that is re-emitted into the line of sight. For a better physical understanding, it is easier to start from the energy-level diagram. Above, I'm showing the the energy-level diagram of the Fe II UV1 group between the ground term and the first excited term of the singly-ionized iron. Take the absorption line Fe II 2600 as an example. The absorption happens when Fe+ absorbs a photon and have the outer electron excited to the lowest level of the first excited term. After the occurrence of the absorption, the excited electron, however, can return to either the ground state (the original lowest level of the ground term), or the second lowest level of the ground term since the total angular momentum number difference is ΔJ = 9/2-7/2 = 1 and the transition is allowed.
We can learn a lot about the physics and the baryon processes in galaxy formation from these spectral features. Bipolar outflows driven by strong star formation have been known for about half a century (e.g., Lynds & Sandage 1963, Bland & Tully 1988, Heckman et al. 1990). An excellent example is the star-burst galaxy M82 (image to the right). Because we are working on composite spectra, we are averaging over all orientations and, in our paper Zhu & eBOSS (2015), we introduced a sperical outflow model (figure above), which can simultaneously explain the multiple observed properties.
For more information, please refer to our paper Zhu & eBOSS (2015) or feel free to send me a message.