HALO GAS AND GALAXY DISK KINEMATICS OF A VOLUME-LIMITED SAMPLE OF Mg ii ABSORPTION-SELECTED GALAXIES AT z ∼ 0.1

Absorption lines detected in the spectra of background quasars and gamma-ray bursts provide powerful probes of the intervening universe. In particular, it was predicted that discrete metal lines, produced in the clumpy metal-enriched gas distribution arising from intervening groups and isolated galaxies, should be detected as absorption in quasar spectra (Bahcall & Salpeter 1965, 1966; Bahcall & Spitzer 1969). Since the Mg ii λλ2796, 2803 doublet arises in metal-enriched low-ionization gas, with neutral hydrogen column densities of $10^{16}\,{\rm cm^{-2}} \lesssim \hbox{N(\hbox{{\rm H}\,{\sc i}})} \lesssim 10^{22}$ cm⁻² (Churchill et al. 2000; Rigby et al. 2002), it should be a dominant absorption feature detected in quasar spectra. In fact, Mg ii absorption was indeed detected along quasar sight lines within close proximity to, and with similar redshifts as, foreground galaxies (Bergeron 1988; Bergeron & Boissé 1991).

Since then, the association of Mg ii with normal, bright, field galaxies is now well established (see Churchill et al. 2005, and references therein). It has been demonstrated that galaxy Mg ii halos extend out to ∼120 kpc and have gas covering fractions of 50%–80% (Tripp & Bowen 2005; Kacprzak et al. 2008; Chen & Tinker 2008; Chen et al. 2010a). However, despite the recent progress on the galaxy–halo connection, the origins of this extended halo gas are still widely debated.

Host galaxy environment has been demonstrated to play some role in determining the gas covering fraction and extent of their halos from studies of galaxy close pairs, galaxy groups, and galaxy clusters associated with Mg ii absorption (Nestor et al. 2007; Lopez et al. 2008; Padilla et al. 2009; Chen et al. 2010a; Kacprzak et al. 2010a, 2010c). Although the bulk of the evidence supports that Mg ii absorption detected along quasar sight lines is a result of outflowing gas (e.g., Bond et al. 2001; Ellison et al. 2003; Bouché et al. 2006; Bouché 2008; Ménard et al. 2009; Nestor et al. 2010), several studies are now emerging that support infalling gas as the source of the absorption (Chen & Tinker 2008; Chen et al. 2010a; Kacprzak et al. 2010a, 2010b; Stocke et al. 2010) or a combination of both inflow and outflow (Chen et al. 2010b; Chelouche & Bowen 2010). In fact, Kacprzak et al. (2010b) have suggested that different mechanisms may be responsible for different equivalent width regimes: low equivalent width systems (≲ 1 Å) may be dominated by infall and the strong equivalent width systems may be dominated by outflows.

It is difficult to isolate separate dynamic processes, such as inflow and outflow, responsible for producing the Mg ii absorption by only studying statistically large samples of absorbers. However, it may be possible to further understand these individual dynamic processes by studying galaxy–absorber pairs on a case-by-case basis. Furthermore, studying a direct comparison of the galaxy disk kinematics and absorbing Mg ii halo gas kinematics may yield a more in depth understanding of these processes.

A direct comparison of the galaxy disk kinematics and absorbing Mg ii halo gas kinematics has been performed for 19 z ∼ 0.6 systems (Steidel et al. 2002; Ellison et al. 2003; Chen et al. 2005; Kacprzak et al. 2010a). The galaxy halos were probed over a range of impact parameters, 8 kpc ⩽ D ⩽ 110 kpc. Kacprzak et al. (2010a) studied 10 host galaxy/absorber pairs and found that the absorption was fully to one side of the galaxy systemic velocity and usually aligned with one arm of the rotation curve in most cases. These results are consistent with earlier studies of five galaxies by Steidel et al. (2002), one galaxy by Ellison et al. (2003), and three galaxies of Chen et al. (2005). In only 5/19 cases, the absorption velocities span both sides of the galaxy systemic velocity. These observations are highly suggestive that Mg ii halos obey "disk-like" rotation dynamics, given the alignment of the absorption and the velocity offsets of the absorbing gas relative to the galaxy.

Steidel et al. (2002) applied simple disk–halo models and concluded that a large fraction of the Mg ii halo gas velocities could be explained by an extension of the disk rotation with some lag in velocity. However, the models were not able to account for the full velocity spread of the gas. This was also confirmed by Kacprzak et al. (2010a) where a large fraction of the Mg ii could not be explained by disk–halo rotation alone and the observed additional velocities were consistent with infalling gas as demonstrated by their study of hydrodynamical galaxy simulations within a cosmological context.

With the increasing blue sensitivity of CCDs, a new redshift window has recently opened, enabling us to detect photons down to the atmospheric cutoff of around 3050 Å. This provides a lower Mg ii absorption detection redshift of z = 0.09. Only a few studies have taken advantage of this new regime to explore the evolution of covering fractions, halo sizes, and kinematics as a function of redshift (Barton & Cooke 2009; Chen et al. 2010a, 2010b).

We have obtained the rotation curves of 13 ∼L_⋆ galaxies, at 0.096 ⩽ z ⩽ 0.148, that also have Keck/LRIS quasar absorption profiles of Mg ii. In this paper, we perform a kinematic comparison of these galaxies and their associated halo Mg ii absorption. We compare our data with a simple rotating disk–halo model, implemented in Kacprzak et al. (2010a) and Steidel et al. (2002), and examine the maximum absorption velocities that could be attributed to the halo gas as disk rotation alone. We further examine the host galaxy environments and also study the intrinsic host galaxy properties, such as star formation rates (SFRs), Mg ib and Na iD line ratios, and Na iD and Hα velocity offsets, that are used to identify and quantify strong outflows. The paper is organized as follows. In Section 2, we present our sample and explain the data reduction and analysis. In Section 3, we present the results of our galaxy–Mg ii absorption kinematic observations, and in Section 4, we compare the observed Mg ii absorption velocities with a simple disk kinematic halo model. In Section 5, we evaluate and discuss the potential mechanisms that produce the observed Mg ii absorption detected near these galaxies. We end with our discussion and conclusions in Sections 6 and 7, respectively. Throughout we adopt an H₀ = 70 km s⁻¹ Mpc⁻¹, Ω_M = 0.3, Ω_Λ = 0.7 cosmology.

Our sample consists of 11 Mg ii absorption systems detected in the spectra of background quasars that are associated with 13 foreground galaxies at similar redshifts. Six of the 11 Mg ii absorption systems were selected from Barton & Cooke (2009) who performed a volume-limited survey targeting ∼L_⋆ galaxies at z ∼ 0.1 with M_r + logh ≳ −20.5 that were in close proximity to quasar lines of sight. They detected six Mg ii absorption systems that are at a similar redshift to seven foreground ∼L_⋆ galaxies. We are in the process of expanding this survey to a more luminous absolute magnitude limit (−21) and, hence, a somewhat higher redshift (E. J. Barton et al. 2011, in preparation). Ultimately, the study will involve approximately 45 galaxies. Here, we add an additional five Mg ii absorption systems, associated with six galaxies, discovered from the expanded survey. Here, we focus on the first kinematics study of the absorbing galaxies at z ∼ 0.1. We do not include the non-absorbing galaxies identified in this survey in this study. Thus, our sample is composed of 11 absorption systems that are associated with 13 galaxies, two of which are double galaxy systems. The galaxy–quasar impact parameters range from 12 kpc ⩽ D ⩽ 90 kpc. We discuss our data and the analysis in the next subsections.

2.1. Quasar Spectroscopy

2.2. Galaxy Spectroscopy

The majority of the galaxy spectra were obtained during nine nights between 2008 May and 2010 February using the double imaging spectrograph (DIS) at the Apache Point Observatory (APO) 3.5 m telescope in New Mexico. Details of the observations are presented in Table 2. The total exposure time per target ranges from 1800 to 5400 s. For each galaxy, the slit position angle was selected to lie along the galaxy major axis.

Table 2. APO/DIS and Keck II/ESI Galaxy Observations

Galaxy Name	R.A.	Decl.	Galaxy	Galaxy	Date (UT)	Instrument	Slit	Exposure
	(J2000)	(J2000)	mr	Mr		/Telescope	P.A.	(s)
J005244G1	00:52:43.92	−00:57:09.23	16.8	−20.8	2009 Nov 16	DIS/APO	35	3 × 1600
J005244G2	00:52:44.02	−00:56:46.41	19.5	−19.7	2009 Nov 15	DIS/APO	−60	3 × 1600
J081420G1	08:14:22.08	+38:33:49.32	16.7	−20.1	2008 Dec 1	DIS/APO	0	3 × 1700
J091119G1	09:11:16.80	+03:12:10.69	16.8	−20.0	2009 Feb 2	DIS/APO	50	4 × 1700
J092300G1	09:23:00.96	+07:51:05.11	16.4	−20.6	2009 Mar 2	DIS/APO	22	3 × 1800
J102847G1	10:28:46.56	+39:18:42.78	17.1	−20.1	2008 May 8	DIS/APO	94	3 × 600
J111850G1	11:18:49.68	−00:21:10.02	17.1	−20.4	2010 Jan 18	DIS/APO	126	3 × 1800
J114518G1	11:45:19.92	+45:16:10.56	17.0	−20.5	2010 Jan 18	DIS/APO	45	3 × 1800
J114803G1	11:48:03.84	+56:54:25.92	16.5	−20.5	2008 Dec 1	DIS/APO	30	3 × 1200
J144033G1	14:40:35.52	+04:48:50.43	16.9	−20.3	2008 May 8	DIS/APO	50	4 × 960
J144033G2	14:40:34.56	+04:48:25.10	17.2	−20.2	2010 Aug 7	ESI/Keck	110	2 × 500
J161940G1	16:19:39.36	+25:43:33.60	18.1	−19.0	2010 Apr 7	DIS/APO	120	4 × 1400
J225036G1	22:50:37.70	+00:07:45.02	15.9	−21.2	2009 Nov 16	DIS/APO	60	3 × 1600

Download table as:  ASCIITypeset image

The DIS spectrograph has separate red and blue channels that have plate scales of 040 pixel⁻¹ and 042 pixel⁻¹, respectively. The B1200 grating was used for the blue channel resulting in a spectral resolution of 0.62 Å pixel⁻¹ with a wavelength coverage of 1240 Å. The R1200 grating was used for the red channel resulting in a spectral resolution of 0.58 Å pixel⁻¹ with a wavelength coverage of 1160 Å. Wavelength centers for each grating were selected to target z ∼ 0.1 galaxies having either Hα plus [N ii] (red channel) and [O ii] (blue channel) in emission or [Na i] (red channel) and Ca ii H & K (blue channel) in absorption. We used a 15 wide by 6' long slit with no on-chip binning of the CCD. The spectral resolution is 1.76 Å (∼120 km s⁻¹) and 1.28 Å (∼53 km s⁻¹) FWHM in the blue and red channels, respectively. The observations were performed during good weather conditions with a typical seeing of 1''–2''.

The spectrum of galaxy J144033G2 was obtained using the Keck Echelle Spectrograph and Imager (ESI; Sheinis et al. 2002) on 2010 August 7. Details of the ESI/Keck observations are presented in Table 2. The slit length is 20'' and 1'' wide and we used 2 × 1 on-chip CCD binning in the spatial direction. Binning by two in the spatial directions results in pixel sizes of 027–034 over the echelle orders of interest. The mean seeing was 08 (FWHM) with clear skies. The total exposure time was 1000 s. The wavelength coverage of ESI is 4000–10000 Å, which allows us to obtain multiple emission lines (such as [O ii] doublet, Hβ, [O iii] doublet, Hα, [N ii] doublet, etc.) with a velocity dispersion of 11 km s⁻¹ pixel⁻¹ (FWHM ∼ 45 km s⁻¹).

All DIS and ESI galaxy spectra were reduced using IRAF. External quartz dome-illuminated flat fields were used to eliminate pixel-to-pixel sensitivity variations. Each DIS spectrum was wavelength calibrated using HeNeAr arc line lamps, and the ESI spectrum was calibrated using CuArXe arc line lamps. Again, galaxy spectra are both vacuum and heliocentric velocity corrected for comparison with the absorption line kinematics. The arc emission line vacuum wavelengths were obtained from the National Institute of Standards and Technology (NIST) database.

A Gaussian fitting algorithm (FITTER; see Churchill et al. 2000), which computes best fit Gaussian amplitudes, line centers, and widths, was used to obtain the galaxy redshifts from an emission or absorption line(s). The galaxy redshifts derived here are consistent with those derived by the Sloan Digital Sky Survey (SDSS). The galaxy redshifts are listed in Table 3; their accuracy ranges from 2 to 30 km s⁻¹.

Table 3. Galaxy Redshifts and Mg ii Absorption Properties

Galaxy Name	D	zgal	zabs	Δvra	Wr(2796)	Wr(2803)	DRb	ΔV−c	ΔV+c
	(kpc)			(km s−1)	(Å)	(Å)		(km s−1)	(km s−1)
J005244G1	86.1 ± 1.2	0.13429 ± 0.00005	0.13460 ± 0.00002	−14.7	1.46 ± 0.04	1.23 ± 0.04	1.19 ± 0.05	−353.8	241.9
J005244G2	32.4 ± 0.2	0.13465 ± 0.00002		+82.2
J081420G1	51.1 ± 0.3	0.09801 ± 0.00004	0.09833 ± 0.00001	+105.2	0.57 ± 0.05	0.28 ± 0.05	2.04 ± 0.37	21.7	178.5
J091119G1	72.1 ± 0.4	0.09616 ± 0.00004	0.09636 ± 0.00009	+66.3	0.82 ± 0.10	0.34 ± 0.07	2.41 ± 0.59	−336.6	454.0
J092300G1	11.9 ± 0.3	0.10385 ± 0.00005	0.10423 ± 0.00004	+127.1	2.25 ± 0.14	1.40 ± 0.12	1.61 ± 0.17	−186.3	418.2
J102847G1	89.8 ± 0.4	0.11348 ± 0.00002	0.11411 ± 0.00004	+168.3	0.30 ± 0.02	0.13 ± 0.02	2.23 ± 0.36	−54.9	329.5
J111850G1	25.1 ± 0.3	0.13159 ± 0.00001	0.13158 ± 0.00002	−4.8	1.93 ± 0.08	1.82 ± 0.07	1.06 ± 0.06	−253.2	227.7
J114518G1	39.4 ± 0.8	0.13389 ± 0.00004	0.13402 ± 0.00002	+46.0	1.06 ± 0.06	1.07 ± 0.05	0.99 ± 0.07	−188.8	247.8
J114803G1	29.1 ± 0.5	0.10451 ± 0.00011	0.10433 ± 0.00002	−62.1	1.59 ± 0.06	1.25 ± 0.05	1.27 ± 0.07	−341.0	217.3
J144033G1	67.1 ± 0.1	0.11277 ± 0.00005	0.11304 ± 0.00001	+90.7	1.18 ± 0.04	0.93 ± 0.03	1.28 ± 0.06	−115.0	293.4
J144033G2	24.9 ± 0.2	0.11271 ± 0.00001		+88.8
J161940G1	45.7 ± 0.7	0.12438 ± 0.00006	0.12501 ± 0.00003	+211.8	0.32 ± 0.03	0.28 ± 0.03	1.12 ± 0.18	72.5	370.9
J225036G1	53.9 ± 0.7	0.14826 ± 0.00002	0.14837 ± 0.00002	+38.0	1.08 ± 0.07	1.11 ± 0.07	0.97 ± 0.09	−140.8	224.4

Notes. ^aΔv_r is the rest-frame velocity offset between the mean Mg ii λ2976 absorption line and the galaxy where, Δv_r = c(z_abs − z_gal)/(1 + z_gal) km s⁻¹. ^bDoublet ratio, DR ≡ W_r(2796)/W_r(2803). ^cBlue and red Mg ii absorption velocity edges.

Download table as:  ASCIITypeset image

The rotation curve extraction was performed following the methods of Kacprzak et al. (2010a; also see Vogt et al. 1996; Steidel et al. 2002). We extracted individual spectra by summing three-pixel-wide apertures (corresponding to approximately one resolution element of 120–126 for DIS and 081–102 for ESI) at 1 pixel spatial increments along the slit. To obtain accurate wavelength calibrations, we extract spectra of the arc line lamps at the same spatial pixels as the extracted galaxy spectra. Fitted arc lamp exposures provided a dispersion solution accurate to ∼0.15 Å (6 km s⁻¹) and ∼0.025 Å (1 km s⁻¹) for DIS and ESI, respectively. Each galaxy emission line (or absorption line in some cases) was fit with a single Gaussian in order to extract the wavelength centroid for each line. The velocity offsets for each emission line in each extraction were computed with respect to the redshift zero point determined for the galaxy (see Table 3). The rotation curves for the 13 galaxies are presented in Figures 1–7.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

2.3. Galaxy Images and Models

Here we discuss the halo gas and galaxy kinematics of the 13 absorption-selected galaxies along 11 different quasar sight lines shown in Figures 1–7. In Table 3, we list all the galaxies in each field that have spectroscopically confirmed redshifts. The table columns are (1) galaxy name, (2) the quasar–galaxy impact parameter, (3) the galaxy redshift, (4) the Mg ii absorption redshift, (5) the Mg ii absorption and galaxy velocity offset, (6) the rest-frame Mg ii λ2796 and (7) the Mg ii λ2803 equivalent width, (8) the doublet ratio, DR, and the blue (9) and red (10) velocity limits of the Mg ii λ2796 absorption profiles. The galaxy velocity offsets from the optical-depth-weighted mean Mg ii absorption range from −14 to +212 km s⁻¹. Galaxy redshifts will only be quoted to four significant figures from here on for simplicity. We will later discuss kinematic halo models in Section 4.

3.1. Mg ii Absorption from Isolated Galaxies

3.2. Mg ii Absorption from Galaxy Pairs/Groups

3.3. Summary I: Observational Kinematic Comparisons

We now apply the simple monolithic halo model of Steidel et al. (2002) to determine whether an extended disk-like rotating halo is able to account for all or most of the observed Mg ii absorption velocity spread measured in our galaxy/absorber systems (see Steidel et al. 2002 for a detailed description of the model). Here we briefly describe the model, which treats the halo gas as a co-rotating thick disk with decreasing velocity as a function of scale height.

The line-of-sight disk–halo velocity, v_los, is dependent on four measurable quantities, D, i, P.A., and v_max ≡ maximum projected galaxy rotation velocity, such that

$$\begin{eqnarray} v_{{\rm los}}(y)&=&\frac{-v_{{\rm max}}}{\sqrt{1+\left(\displaystyle \frac{y}{p}\right)^2}}\mbox{ } \exp \left\lbrace -\frac{\left|y-y_{\circ }\right|}{h_{v}\tan i}\right\rbrace,\mbox{ }\mbox{where} \\ &&\nonumber \\ [2.0ex] y_{\circ }&=&\frac{D\sin {\rm P.A.}}{\cos i} \quad \mbox{and}\quad p=D\cos {\rm P.A.},\nonumber \end{eqnarray} \tag{ 1 }$$

where the free parameter, h_v, is the lagging halo gas velocity scale height. The line-of-sight velocity is a function of y, which is the projected line-of-sight position above the disk plane, and the parameter y_○ represents the position at the projected mid-plane of the disk. The range of y is constrained by the model halo thickness, H_eff, such that y_○ − H_efftan i ⩽ y ⩽ y_○ + H_efftan i. The distance along the line of sight relative to the point where it intersects the projection of the disk mid-plane is then D_los = (y − y_○)/sin i, thus, D_los ≡ 0 at the disk mid-plane. There are no assumptions about the spatial density distribution of Mg ii absorbing gas, except that H_eff is the effective gas layer thickness capable of giving rise to absorption.

Here we set h_v = 1000 kpc in order to maximize the rotational velocity predicted by the model, which effectively removes the lagging halo velocity component (such that the exponential in Equation (1) is roughly equal to unity).

4.1. Dealing with Spectral Resolution

In order to compare our present low-redshift results directly to the kinematic studies of Kacprzak et al. (2010a) and Steidel et al. (2002) performed at intermediate redshift, we must account for the difference in the spectral resolution of the two samples. Here, our LRIS spectra have a velocity resolution of v ≃ 155 km s⁻¹, whereas the Kacprzak et al. (2010a) HIRES/Keck and UVES/Very Large Telescope (VLT) spectra have a velocity resolution of v ≃ 6 km s⁻¹.

To compute the velocity broadening of the LRIS Mg ii absorption profiles due to spectral resolution, we degraded 26 HIRES/Keck and UVES/VLT W_r(2796) > 0.2 Å systems from Kacprzak et al. (2010b) and Kacprzak et al. (2010c) to have identical resolution of the LRIS data (these include the systems of the kinematics studies of Kacprzak et al. 2010a and Steidel et al. 2002). We utilized the Voigt profile parameters (column densities, b parameters, and velocities) from the Voigt profile fits to the Kacprzak et al. spectra and generated synthetic LRIS spectra of the profiles. We convolved the smooth Voigt profile model with a Gaussian instrumental spread of LRIS (R = 3500) and sampled the spectra with a pixel size of 0.18 Å. We introduced noise using a signal-to-noise ratio (S/N) of 15 per pixel (using Gaussian deviates). An uncertainty spectrum, σ(λ), is generated that accounts for the LRIS read noise (which affects the noise in the line cores), using a Poisson plus read-noise, model (see Churchill 1997)

$$\begin{equation} \sigma (\lambda) = I_c^{-1} [ F(\lambda)I_c + {\rm RN}^2] ^{1/2}, \end{equation} \tag{ 2 }$$

where F(λ) is relative counts in the synthesized spectra at wavelength λ and the continuum I_c is

$$\begin{equation} I_c = \frac{{\rm ({\rm S/N})}^2}{2} \left\lbrace 1 + \left[ 1 + \left(\frac{2{\rm RN}}{{{\rm S/N}}} \right) ^2 \right] ^{1/2} \right\rbrace, \end{equation} \tag{ 3 }$$

where S/N is the signal-to-noise ratio and RN is the read noise.

We generated synthetic LRIS spectra of both members of the Mg ii doublet and then fully analyze these spectra in an identical fashion performed for the LRIS data used in this work. We thus obtained the equivalent widths, double ratios, velocity moments, and velocity spreads (and all uncertainties) for these synthetic LRIS spectra. In Figure 8(a), we show an example of the degraded spectra, noting the symmetric broadening of the absorption profile.

Zoom In Zoom Out Reset image size

In Figure 8(b), we show the distribution of absorption velocity widths for the degraded and non-degraded systems. We find that, using the identical measurement standards for both the synthetic LRIS spectra and the observed LRIS spectra, the resolution, pixelization, and noise in the observed spectra introduce an average apparent velocity spread increase of ∼170 km s⁻¹ with a scatter of ∼60 km s⁻¹. Furthermore, Figure 8(c) shows that the average apparent increase of velocity is symmetric for both the blue and red wings (edges) of the absorption profiles. We find an average velocity increase of ±85 km s⁻¹ (to the red and to the blue) relative to the Voigt profile models of the absorbers observed at HIRES and UVES resolution. There is a scatter of approximately 25 km s⁻¹ for the 26 systems we examined from the Kacprzak et al. spectra.

We have now applied a resolution correction such that the LRIS Mg ii absorption velocity widths are translated into observed HIRES or UVES velocity widths.

4.2. Results of the Halo–Disk Models

In Figures 9(a)–(l), we show the Mg ii absorption profiles for each galaxy, where the shaded regions indicate detected absorption. We have applied a resolution correction such that the Mg ii velocity widths (shaded region) are translated into "observed" HIRES or UVES velocity widths as indicated by the vertical dashed lines (red).

Zoom In Zoom Out Reset image size

Below each absorption profile are the thick-disk–halo model velocities as a function of D_los derived for each galaxy (solid line) using Equation (1) and parameters in Table 4. Recall that, at D_los = 0 kpc, the model line of sight intersects the projected mid-plane of the galaxy. The dashed curves represent the range of disk–halo model velocities derived from the combination of the minimum and maximum uncertainties in the P.A. and i. In some cases, the values of the P.A. and i are well determined such that the dashed curves lie on the solid curves (see Figure 9(b)). The model also predicts the line-of-sight position, D_los, of the halo gas at each velocity, v_los.

The thick-disk–halo model is successful at predicting the observed Mg ii absorption velocity distribution when the solid (or dashed) curves span the same velocity spread as that of the Mg ii absorption gas defined by the shaded region between the vertical dashed (red) lines. If this is not the case, one can conclude that disk-like halo rotation is not the only dynamic mechanism responsible for the Mg ii kinematics. In the following subsections, we discuss the halo models of the individual galaxies.

4.2.1. J005244G1 & J005244G2

Galaxies G1 and G2 are potentially interacting galaxies, given their close angular proximity and redshifts. G1 is the brightest galaxy of the two and is also the closest to the quasar line of sight. In Figure 9(a), we plot the disk–halo models for G1 and G2. The G1 model has velocities that are consistent with up to 150 km s⁻¹ of the Mg ii blueward of its systemic velocity. There is a 100 km s⁻¹ gap between the models of G1 and G2. G2 is also counter-rotating with respect to G1 as viewed from the quasar sight line. The G2 model velocities are consistent with the remaining absorption redward of its systemic velocity. Although the disk-like halo model is mostly successful at accounting for some of the absorption velocity, it does not reproduce all of the Mg ii absorption velocities.

4.2.2. J081420G1

Galaxy G1 is a low-inclination galaxy with the absorption lining up exactly with one side of the galaxy rotation curve. In Figure 9(b), we show the halo model velocities and the Mg ii absorption profile. Note that there is only a single dashed line (red) indicating the corrected velocity width since the profile velocity width is less than the ±85 km s⁻¹ velocity correction. As we previously mentioned, there is a velocity correction scatter of ±25 km s⁻¹ and by taking this into account, the Mg ii absorption line is likely a very narrow component centered on the dashed line. Regardless, we see that although the absorbing gas is aligned with the rotation curve, it is in the opposite direction expected for disk rotation, i.e., the galaxy is "counter-rotating" with respect to the Mg ii absorption.

4.2.3. J091119G1

Galaxy G1 is an edge-on spiral with the quasar line-of-sight probing roughly perpendicular to the galaxy major axis. The Mg ii λ2796 spans both sides of the galaxy systemic velocity. In Figure 9(c), the halo model can account for the Mg ii absorption blueward of the galaxy systemic velocity. However, the bulk of the Mg ii (specifically the λ2803 transition) is redward of the galaxy systemic velocity and is "counter-rotating" with respect to the galaxy's direction of rotation.

4.2.4. J092300G1

Galaxy G1 is an S0-like galaxy that does not have emission lines but exhibits rotation as measured from the absorption lines. Since the quasar line of sight probes along the minor axis of the galaxy, if halo rotation was responsible for the absorption kinematics, then one would expect the absorbing gas to have velocities more consistent with the galaxy systemic velocity. In Figure 9(d), we find that the halo model can adequately account for the total velocity spread blueward of the galaxy systemic velocity. However, the model cannot explain ∼300 km s⁻¹ of Mg ii absorption detected redward of the galaxy systemic velocity. These high velocities detected along the quasar line sight only 12 kpc away only the major axis may be signatures of outflow or infall. The galaxy does appear to have a separate optical clump/component or satellite seen below the galaxy (along the slit) which may be interacting and causing an infall of metal-enriched gas. We do not have a spectrum of this second object.

4.2.5. J102847G1

The strongest component of the Mg ii absorption associated with G1 aligns exactly with its redward maximum rotation. In Figure 9(e), we find that the galaxy is "counter-rotating" with respect to the strongest Mg ii component. In fact, the halo model cannot explain the large velocity width of the absorption. The halo model does not well represent the kinematics observed along this quasar line of sight.

4.2.6. J111850G1

The quasar line of sight probes the minor axis of G1 and the absorption spans the entire rotation velocity. In Figure 9(f), we see that disk rotation can account for the bulk of the Mg ii that is redward of the galaxy systemic velocity. However, it cannot account for the absorption blueward of the galaxy systemic velocity. Thus, this gas is also "counter-rotating" with respect to the galaxy and is inconsistent with disk rotation.

4.2.7. J114518G1

The quasar line of sight near G1 probes the galaxy major axis with the bulk of the Mg ii residing to one side of the galaxy systemic velocity. In Figure 9(g), we find that the disk model can adequately account for the Mg ii absorption that has velocities redward of the galaxy systemic velocity along the line of sight. However, the model does not account for a small fraction of the Mg ii absorption blueward of the galaxy systemic velocity. Thus, the model can account for the majority of the absorption but cannot explain some of the weaker absorption 100 km s⁻¹ blueward of the galaxy systemic velocity.

4.2.8. J114803G1

The galaxy G1 exhibits a low-level velocity shear. Given the velocity spread of the Mg ii absorption, it is impossible for the bulk of the absorbing gas to be consistent with the observed velocities of G1. In Figure 9(h), we see that the galaxy-disk–halo model is counter-rotating with respect to the bulk of the absorbing gas. There is little overlap between the predicted halo model velocities with those of the Mg ii absorption. Even if the galaxy had a highly significant velocity shear, the bulk of the Mg ii would not be consistent in velocity space.

4.2.9. J144033G1 and J144033G2

Galaxies G1 and G2 are potentially interacting galaxies, given their close angular proximity and redshifts. G1 is the brightest galaxy of the two, however G2 is much closer to the quasar sight line and is clearly forming stars. In Figure 9(j), we plot the disk–halo models for G1 and G2. The halo models for G1 and G2 cover roughly the same velocity range of 100 km s⁻¹ and are rotating in the same direction. Although the model is mostly successful, it fails to predict the absorption at higher velocities between 100 and 200 km s⁻¹.

4.2.10. J161940G1

The galaxy G1 exhibits a low-level velocity shear. In Figure 9(k), we see that the galaxy disk–halo model is "counter-rotating" with respect to the dominant saturated Mg ii component. There is no overlap between the predicted halo model velocities with those of the Mg ii absorption. Even if the galaxy had a highly significant velocity shear, the bulk of the Mg ii clouds would not be consistent in velocity space.

4.2.11. J225036G1

The almost edge-on galaxy G1 has the bulk of the Mg ii residing to the redward side of the galaxy systemic velocity. In Figure 9(l), we see that the disk–halo model can account for almost all of the Mg ii absorbing gas velocity spread. It does not quite account for the gas blueward of the galaxy systemic velocity, however with the ±25 km s⁻¹ errors on the corrected absorption line widths (red lines), we can say that the model is likely consistent with the absorption velocities. Thus, a disk-like halo model well represents the absorption kinematics.

4.3. Summary II: Disk–Halo Model

There are three likely scenarios that could help produce extended metal-enriched gaseous halos around galaxies: (1) galaxy group environments producing tidal streams and stripped gas from galaxies, (2) outflowing gas from star-forming regions and/or supernovae, and/or active galactic nuclei, and (3) infalling gas from streams, filaments, high-velocity clouds, satellites, and previously ejected gas from outflows. In this section, we attempt to determine if the Mg ii halo gas detected in absorption is produced via environmental effects, outflow, or infall.

5.1. Environment

5.2. Outflows

Here we explore evidence of outflows using two techniques: we first compute the host galaxy SFRs and compare them to the halo gas absorption strength. We then compute Mg ib (stellar) and Na iD (stellar+ISM) line ratios in order to identify possible outflows.

5.2.1. Star Formation Rates

We have computed SFRs for the host galaxies where possible. We note that the SDSS spectra are obtained using fibers that have an aperture radius of 15, which translates into 2.77 kpc at z = 0.10. Our galaxies have an average half-light radius, as measured from GIM2D (see Table 4), of 〈r_h〉 = 5.6 ± 1.3 kpc, so the SDSS fibers cover only the inner regions of the galaxies. Therefore, SFRs computed here are only within the fiber. This still provides useful information since all galaxies are roughly at the same redshift and therefore probing the same physical scales, and strong winds are expected to originate within the central regions.

For the fiber SFRs we have applied Galactic extinction correction obtained from NED.⁵ We are not able to apply Balmer decrement corrections since only two galaxies have detected Hβ emission. The Hα luminosities were measured from the SDSS spectra and Hα-derived SFRs were computed using the formalism of Kewley et al. (2002). We also performed aperture loss corrections to the SFRs. The applied scaling factor was determined from the ratio of the r-band galaxy total counts to those within the SDSS fiber.

In Figure 10(a), we show the inner galaxy (fiber) SFRs, and corrected SFRs, as a function of the Mg ii absorption strength. We find no correlation between the SFRs and the halo gas absorption strength. A similar distribution arises if one plots Hα against W_r(2796). Note that some galaxies have only low SFR limits yet are still associated with strong absorption. Given that Barton & Cooke (2009) noted that red galaxies appear to be closer to the quasar line of sight than blue star-forming galaxies, we normalized out the impact parameter in Figure 10(b): we find no statistically significant trend here.

Zoom In Zoom Out Reset image size

Another indicator of galaxy outflows is the star formation per unit area. Heckman (2002, 2003) demonstrated that outflows are ubiquitous in galaxies where the global SFR per unit area exceeds Σ = 0.1 M_☉ yr⁻¹ kpc² (the area is defined by the half-light radius). The interstellar medium (ISM) entrained in these winds has outflow speeds of ∼100 to ∼1000 km s⁻¹. Although the half-light radii of our galaxies are larger than the SDSS fibers, we can compute the surface star formation density within the SDSS fiber. For our sample, we find that the star formation per unit area of $0.03\,M_{\odot }\,{\rm yr^{-1}\,{\rm kpc^{2}}} \le \Sigma \le 0.0002\,M_{\odot }$ yr⁻¹ kpc², which is well below what is expected for strong winds.

These results possibly suggest that star formation-driven winds, at least in the galaxy central regions, are not producing the observed kinematics and absorption strength of the metal-enriched halo gas. It is possible that the metal-enriched gas detected along the quasar sight lines are potential reservoirs for future star formation.

5.2.2. Na iD and Mg ib Line Ratios

It has been demonstrated that Na iD and Mg ib absorption line ratios are good tracers of outflows. Both Na iD and Mg ib have similar ionization potentials (5.14 eV for Na iD, 7.65 eV for Mg ib). Although both Na iD and Mg ib appear in stellar spectra, where the absorption strength peaks in spectra of cool K–M stars (see Jacoby et al. 1984), the Mg ib band is a highly excited transition making it purely stellar in origin. On the other hand, the Na iD resonance line can also be absorbed by the ISM and has been detected as entrained gas within galactic scale winds (e.g., Martin 2005; Rupke et al. 2005a, 2005b; Heckman et al. 2000). Thus, this line ratio can be used to successfully separate starburst outflowing galaxies from quiescent galaxies with little to no winds (e.g., Heckman et al. 2000).

We have used the Na iD and Mg ib equivalent widths computed by SDSS, as observed in galaxy spectra to study the wind properties of our galaxy sample. Again, we note that the SDSS spectral fibers have an aperture radius of 15, which translates into 2.77 kpc at z = 0.10, thus the fibers only cover the central regions of the galaxies. The line ratio is only computed within the fiber, however, galactic winds are expected to originate from the centers of galaxies, which is where the wind signature in the line ratio is expected. As one would include more of the galaxy light, from regions where no winds are found, we would expect the wind line ratios to be more consistent with stellar origins.

In Figure 11, we show the Na iD and Mg ib equivalent width distribution. Heckman et al. (2000) estimated the expected stellar contribution to the Na iD, by scaling the equivalent width of Mg ib, and found the relation W_r(Na iD) = 0.75W_r(Mg ib) represented by the dotted line in Figure 11. Galaxies that reside to the right of the solid line are most likely to have contributions from the ISM to the Na iD absorption. The solid line, which is expressed as W_r(Na iD) = 3W_r(Mg ib), is the approximate location of starbursting galaxies with strong Na iD winds observed by Rupke et al. (2005a, 2005b). They found that 80% of the galaxies below this line have winds, while only a small fraction above the solid line have winds (25%).

Zoom In Zoom Out Reset image size

All of our galaxies reside far from the relation where winds are expected and reside tightly near the expected stellar contribution line, suggesting that our galaxy sample has little to no strong winds. Scatter about that line is most likely due to interstellar Na iD absorption.

Another way to detect outflows is to observe velocity offsets between galaxy nebular emission lines and absorption lines. This technique has been applied in previous studies on a range of absorption lines and has been demonstrated to detect strong winds (see Heckman et al. 2000; Rupke et al. 2005b; Weiner et al. 2009; Steidel et al. 2010; Rubin et al. 2010,, etc.). Eight of our absorbing galaxies have measurable Hα and Na iD lines. Although we have discussed above that Na iD absorption is contaminated by stellar absorption, it can still be used to trace winds from velocity offsets from the emission lines (e.g., Heckman et al. 2000; Rupke et al. 2005b). Furthermore, Heckman et al. (2000) found that galaxies with Na iD absorption residing within ±70 km s⁻¹ of the systematic galaxy redshift were consistent with a predominantly stellar origin. Na iD absorption blueshifted with velocities greater than 100 km s⁻¹ was associated with outflows ∼70% of the time. They also found that galaxies with strong outflows were viewed at low inclination angles.

For the eight galaxies in our sample, we find a mean velocity intrinsic Na iD absorption offset of Δv = −71 ± 26 km s⁻¹, consistent with a predominantly stellar origin. However, all galaxies have a negative velocity offset from the Hα emission line. This offset may suggest low-level winds, although we find no correlation between the velocity offset and the galaxy SFR, or Σ, as would be expected (Weiner et al. 2009).

The average galaxy inclination angle is 46° which is less than the average expected for a random distribution of galaxies. However, we do not find a significant correlation between the velocity offset and the galaxy inclination or absorption strength.

These results suggest that active outflows are not responsible for the dominant component of the Na iD absorption and are consistent with a stellar component plus some contribution from dynamically cold interstellar gas. The lack of strong outflows close to the galaxy suggests alternate origins of the Mg ii halo gas (0.3 Å ⩽ W_r(2796) ⩽ 2.3 Å).

Kacprzak et al. (2010a) compared Mg ii absorption and galaxy rotation kinematics of 10 W_r(2796) < 1.4 Å systems and found that, in most cases, the absorption was fully to one side of the galaxy systemic velocity and usually aligned with one arm of the rotation curve. These results are consistent with earlier studies of five galaxies by Steidel et al. (2002), one galaxy by Ellison et al. (2003), and three galaxies of Chen et al. (2005). In only 5/19 cases, the absorption velocities span both sides of the galaxy systemic velocity. Three of those have W_r(2796) > 1 Å and their absorption kinematics displayed possible signatures of winds or superbubbles (Bond et al. 2001; Ellison et al. 2003); two of these galaxies have SFRs and SFRs per unit area consistent with wind-dominated galaxies (Kacprzak et al. 2010a).

For our z ∼ 0.1 sample, we find that for only 5/13 galaxies the Mg ii absorption resides to one side of the galaxy systemic velocity and aligns with one side of the rotation curve. For the remaining 8/13 galaxies, the absorption spans both sides, although the bulk of the Mg ii resides mostly to one side of the galaxy systemic velocity.

In comparing the results from z ∼ 0.5 to our z ∼ 0.1 sample, we find a factor of three increase in the fraction of systems where the Mg ii absorption resides on both sides of the galaxy systemic velocity at lower redshift. These results may suggest an evolution in the halo gas kinematics as a function of redshift. Both samples span a similar range of impact parameters and Mg ii equivalent width. Hints that halo gas properties may evolve with redshift have already been observed. Barton & Cooke (2009) found that gas covering fractions may decrease by a factor of 2–3 by z = 0.1. It is also important to note that the samples of Kacprzak et al. (2010a) and Steidel et al. (2002) have an average 〈L_B〉 = 0.6L*_B whereas our sample has a 〈L_B〉 = 1.9L*_B: in this study, we have probed more massive galaxies at lower redshift. This leads to the possibility that there could be an evolution as a function of host galaxy mass. This is consistent with the cosmological smoothed particle hydrodynamic simulations of Stewart et al. (2010) who predict that the halo gas covering fraction exhibits a sharp decrease when the galaxy mass exceeds a critical minimum mass to form stable shocks which results in a transition from cold mode to hot mode gas accretion. This can reduce the covering fraction, over the same redshifts observed here, by factors of 6–10. It has also been demonstrated that for massive galaxy halos of ≳ 10¹³ h⁻¹ M_☉ at z ∼ 0.5, the covering fractions decrease by a factor of 7–15 (Gauthier et al. 2010; Bowen & Chelouche 2011). A more uniform sample is required to explore the possibility of an evolution as a function of mass and/or redshift.

Since the halo gas velocities at intermediate redshift were found to align in the same sense and as velocities expected for co-rotation, it strongly suggests "disk-like" rotation of the halo gas. Both Kacprzak et al. (2010a) and Steidel et al. (2002) applied simple co-rotating disk–halo models and concluded that an extension of the disk rotation was able to explain some of the gas kinematics. However, the models were not able to account for the full absorption velocity spreads.

Here we obtain a similar conclusion, except for one case, all of the observed kinematics velocity spread of the halo gas cannot be explained with a simple rotating disk model. However, contrary to previous studies, for 55% of our sample, the halo model is "counter-rotating" with respect to the bulk of the Mg ii absorption, and in two cases, there is zero overlap between the model and the absorption velocities. This suggests that at least some gaseous halos at z ∼ 0.1 are not co-rotating with their host galaxies and to a lesser extent than what was found at z ∼ 0.5. Again this implies that other mechanisms must be invoked to account for the full velocity spreads.

In an effort to identify the origins of the Mg ii absorption, Kacprzak et al. (2010a) used hydrodynamical cosmological simulations, combined with the quasar absorption line methods, to demonstrate that the majority of the Mg ii absorption arises in an array of cosmological structures, such as filaments and tidal streams. They showed that metal-enriched gas was infalling toward the galaxy along these structures with velocities in the range of the rotation velocity of the simulated galaxy and consistent with the observed galaxy halo gas kinematics. In this paper, we have not gone to these efforts to explore the origins of the halo gas, however this will be part of our future work. We have instead chosen to explore the galaxy environment and also physical properties of the host galaxies that are indicative of strong outflows.

The z ∼ 0.1 Mg ii galaxies appear to be isolated, aside from the two double galaxy systems identified in our own survey. Again, these two pairs are consistent with one dominate host galaxy and a smaller satellite galaxy. Only three host Mg ii absorbing galaxies have other nearby galaxies, however they reside far from the quasar line of sight making it unlikely that they contribute to the absorbing gas. Thus, for our sample, interactions may not play a crucial role in producing the halo gas absorption.

It is well known that highly star-forming galaxies tend to have strong outflows that are also detected in absorption (e.g., Weiner et al. 2009; Martin & Bouché 2009; Nestor et al. 2010). However, for our sample the host galaxy SFRs computed within the SDSS fiber, representing the galaxy central regions, are all less than 1 M_☉ yr⁻¹. In addition, we find no correlation between the SFRs and the W_r(2796) even when normalized by the impact parameter. Heckman (2002, 2003) demonstrated that outflows with speeds of ∼100 to ∼1000 km s⁻¹ are ubiquitous in galaxies where the global SFR per unit area exceeds Σ = 0.1 M_☉ yr⁻¹ kpc². For our sample we find Σ ⩽ 0.03 M_☉ yr⁻¹ kpc², which is well below what is expected for strong winds.

We also find that the Na iD (stellar+ISM) and Mg ib (stellar) absorption line ratios are consistent with being predominately stellar in origin and having kinematically cool ISM. The velocity offsets between the Na iD line and the nebular Hα emission line are on average −71 ± 26 km s⁻¹. Although the shift is in the negative direction, which is associated with outflows, the velocity shifts are small and consistent with little-to-no outflows (Heckman et al. 2000). In addition, the velocity offsets between Na iD and Hα do not correlate with SFR or Σ.

Our sample of galaxies appear to be isolated, and undergoing some star formation, but too little to be producing strong outflows. We find it is unlikely that the Mg ii gas originates from either environmental effects, such as galaxy–galaxy interactions or mergers, and/or outflowing gas; although accretion of cold gas ejected from previous star formation events is possible. We favor a scenario where the metal-enriched halo gas is infalling onto the host galaxy with velocities comparable to the dynamics of their host.

We have examined and compared the detailed galaxy and Mg ii absorption kinematics for a sample of 13 intermediate redshift, ∼L_⋆ galaxies along 11 quasar sight lines. The galaxy–quasar impact parameters range from 12 kpc ⩽ D ⩽ 90 kpc. The galaxy rotation curves were obtained from DIS/APO and ESI/Keck spectra and the Mg ii absorption profiles were obtained from LRIS/Keck quasar spectra. In an effort to compare the relative kinematics, we used a disk–halo model to compute the expected absorption velocities through a monolithic gaseous halo model. We further examined the host galaxy environments and also studied the intrinsic host galaxy properties, using them to quantify and identify strong outflows.

Our mains results can be summarized as follows.

• 1.  
  For all 13 galaxies, the velocity of the strongest Mg ii absorption component lies in the range of the observed galaxy rotation curve. We find that for the nine isolated galaxy/absorber pairs, seven galaxies have well-defined rotation curves while two galaxies display only shear. In 3/9 cases, the absorption resides to one side of the galaxy systemic velocity and the absorption redshift tends to align with one side of the rotation curve. In the remaining 6/9 cases, the absorption spans both sides of the galaxy systemic velocity, although the bulk of the Mg ii resides mostly to one side of the galaxy systemic velocity.For our double galaxy/absorber pairs, we find that all four galaxies exhibit well-defined rotation curves. In one case, the absorbing gas spans both sides of both host galaxy systemic velocities. In the other case, the absorbing gas resides to one side of the systemic velocity of both absorbing galaxies.
• 2.  
  In all cases, the thick disk rotating halo models are unable to reproduce the full spread of observed Mg ii absorption velocities. Contrary to previous studies at higher redshifts, for 55% of our sample the halo model is "counter-rotating" with respect to the bulk of the Mg ii absorption and in two cases there is zero overlap between the model and the absorption velocities. This potentially suggests that z ∼ 0.1 gaseous halos are not co-rotating with their host galaxies and to a lesser extent than what was found at z ∼ 0.5. In this simple scenario, even if some of the absorbing gas arises in a thick disk, there must be dynamical processes (such as infall, outflow, supernovae winds, etc.) that give rise to the remaining Mg ii absorption.
• 3.  
  Host galaxy SFRs are all ≲ 1 M_☉ yr⁻¹ and we find no correlation between the SFRs and the W_r(2796) even when normalized by the impact parameter. Their SFRs per unit area are ⩽0.03 M_☉ yr⁻¹ kpc², which is well below the lower limit of 0.1 M_☉ yr⁻¹ kpc² expected for strong winds.
• 4.  
  We find our absorbing galaxies tend to be isolated, or at least in low-density environments. This is further supported by an analysis of cosmological simulations at z ∼ 0.1 performed by Barton & Cooke (2009).
• 5.  
  The Na iD (stellar+ISM) and Mg ib (stellar) absorption line ratios are consistent with being predominately stellar in origin and having kinematically cool ISM. The velocity offsets between the Na iD line and the nebular Hα emission line are on average −71 km s⁻¹. Although the shift is in the negative direction, the velocity shifts are small and are not correlated with SFR or Σ. These results are consistent with our Mg ii absorption-selected galaxies having little-to-no outflows.

In our detailed study of 13 z = 0.1 absorbers, we find it unlikely that the Mg ii gas originates from either outflowing gas and/or environmental effects. These results are consistent with the hydrodynamical simulations of Kacprzak et al. (2010a) where the Mg ii gas is inflowing along the streams and filaments with kinematics comparable to the host galaxy. Thus, we favor a scenario of infalling gas that provides a gas reservoir for star formation at these low redshifts.

We thank Michael Murphy for advice and for carefully reading the manuscript and providing useful comments. C.W.C. was supported from NSF grant AST-0708210. E.J.B. gratefully acknowledges support from NSF grant AST-1009999 and the UC Irvine Center for Cosmology. J.C. gratefully acknowledges generous support by the Gary McCue Postdoctoral Fellowship. This work is based on observations obtained with the Apache Point Observatory 3.5 m telescope, which is owned and operated by the Astrophysical Research Consortium. Observations were also obtained at the W.M. Keck Observatory, which is operated as a scientific partnership among the California Institute of Technology, the University of California, and the National Aeronautics and Space Administration. The Observatory was made possible by the generous financial support of the W.M. Keck Foundation. This paper would not have been possible without data from the Sloan Digital Sky Survey (SDSS). Funding for the SDSS and SDSS-II has been provided by the Alfred P. Sloan Foundation, the Participating Institutions, the National Science Foundation, the U.S. Department of Energy, the National Aeronautics and Space Administration, the Japanese Monbukagakusho, the Max Planck Society, and the Higher Education Funding Council for England. The SDSS Web site is http://www.sdss.org/. The SDSS is managed by the Astrophysical Research Consortium for the Participating Institutions. The Participating Institutions are the American Museum of Natural History, Astrophysical Institute Potsdam, University of Basel, University of Cambridge, Case Western Reserve University, University of Chicago, Drexel University, Fermilab, the Institute for Advanced Study, the Japan Participation Group, Johns Hopkins University, the Joint Institute for Nuclear Astrophysics, the Kavli Institute for Particle Astrophysics and Cosmology, the Korean Scientist Group, the Chinese Academy of Sciences (LAMOST), Los Alamos National Laboratory, the Max-Planck-Institute for Astronomy (MPIA), the Max-Planck-Institute for Astrophysics (MPA), New Mexico State University, Ohio State University, University of Pittsburgh, University of Portsmouth, Princeton University, the United States Naval Observatory, and the University of Washington.

Facilities: ARC (DIS) - Astrophysical Research Consortium 3.5m Telescope at Apache Point Observatory, Keck:I (LRIS) - KECK I Telescope, Keck:II (ESI) - KECK II Telescope, Sloan (SDSS) - Sloan Digital Sky Survey Telescope