JSTOR

1. INTRODUCTION

The Infrared Astronomical Satellite (IRAS) provided the first unbiased survey of the sky at mid and far-infrared wavelengths, giving us a comprehensive census of the infrared emission properties of galaxies in the local Universe. The high luminosity tail of the infrared luminosity function can be approximated by a power law (), which implies a space density for the most luminous infrared sources that is well in excess of what is found in the optical for local galaxies (e.g., Schechter 1976). At the highest luminosities, Ultraluminous Infrared Galaxies, or ULIRGs (those galaxies with L_IR > 10¹² L_⊙) have a space density that is a factor of 1.5–2 higher than that of optically selected QSOs, the only other known objects with comparable bolometric luminosities (Schmidt & Green 1983).

Multiwavelength imaging surveys have shown that nearly all ULIRGs are found in systems that have undergone strong tidal perturbations due to the merger of pairs of gas-rich disk galaxies (Armus et al. 1987; Sanders et al. 1988a, 1988b; Murphy et al. 1996, 2001a). They have enhanced star formation rates compared to noninteracting galaxies, and the fraction of sources with active galactic nuclei (AGN) increases as a function of increasing L_IR (Armus et al. 1989; Murphy et al. 1999, 2001b; Veilleux et al. 1995, 1997; Kim et al. 1998). ULIRGs may represent an important evolutionary stage in the formation of QSOs (e.g., Sanders et al. 1988a, 1988b) and perhaps powerful radio galaxies (Mazzarella et al. 1993; Evans et al. 2005). In fact, numerical simulations have shown that tidal dissipation in mergers can be very effective in driving material from a gas-rich galaxy disk toward the nucleus, fueling a starburst and/or a nascent AGN (e.g., Barnes & Hernquist 1992). Morphological and kinematic studies of ULIRGs also suggest that their stellar populations are evolving into relaxed, elliptical-like distributions (Wright et al. 1990; Genzel et al. 2001, Tacconi et al. 2002).

Luminous Infrared Galaxies, or LIRGs (those galaxies with L_IR≥10¹¹ L_⊙), are interesting phenomena in their own right, but they also may play a central role in our understanding of the general evolution of galaxies and black holes, as demonstrated by two key observational results. First, observations with ISO and Spitzer have shown that LIRGs comprise a significant fraction (≥50%) of the cosmic infrared background and dominate the star-formation activity at z ∼ 1 (Elbaz et al. 2002; Le Floch et al. 2005; Caputi et al. 2007, Magnelli et al. 2009). In comparison to the local universe where they are relatively rare, ULIRGs are about a thousand times more common at z ≳ 2 (Blain et al. 2002; Chapman et al. 2005). Second, the ubiquity of supermassive nuclear black holes in quiescent galaxies and the scaling of their masses with stellar bulge masses (Magorrian et al. 1998; Ferrarese & Merritt 2000; Gebhardt et al. 2000) suggest an intimate connection between the evolution of massive galaxies and their central black holes. One explanation for the relationship is that mass accretion onto the black hole occurs during episodes of nuclear star formation. The study of a large, complete sample of local LIRGs spanning all merger and interaction stages can shed critical light on the coevolution of black holes and stellar bulges in massive galaxies.

The Great Observatories All-Sky LIRG Survey (GOALS) combines data from NASA’s Spitzer, HST, Chandra, and GALEX observatories in a comprehensive imaging and spectroscopic survey of over 200 low-redshift (z < 0.088) LIRGs (see Table 1). The primary new data sets consist of Spitzer Infrared Array Camera (IRAC) and Multi-band Imaging Photometer for Spitzer (MIPS) imaging; Infrared Spectrograph (IRS) spectroscopy (Houck et al. 2004); HST Advanced Camera for Surveys (ACS), Near Infrared Camera and Multi-Object Spectrometer (NICMOS), and Wide Field and Planetary Camera 2 (WFPC2) imaging; Chandra AXAF CCD Imaging Spectrometer (ACIS) imaging; and GALEX far and near-UV observations. The majority of the new observations are being led by GOALS team members, but we also make use of the Spitzer, HST, Chandra, and GALEX archives to fill out the sample data (see Table 2). In addition, optical and K-band imaging (Ishida et al. 2004); optical spectra (Kim et al. 1995); J, H, and K_s-band near-infrared images from 2MASS (Skrutskie et al. 2006); submillimeter images (Dunne et al. 2000); CO and HCN (Sanders et al. 1991; Gao & Solomon 2004); and 20 cm VLA imaging (Condon et al. 1990) exist for various subsets of the LIRGs in GOALS, in addition to the IRAS data with which the sources were selected (Sanders et al. 2003). In this article we introduce the main components of the survey, present our primary science objectives, and discuss the multiwavelength results for one source, VV 340. The layout of the paper is as follows. The scientific objectives are discussed in § 2, followed by a definition of the sample in § 3. In § 4 we describe the observations, in § 5 we outline the Spitzer Legacy data products being delivered, and in § 6 we present results for the LIRG VV 340.

TABLE 1 The GOALS Sample

Open New Window

TABLE 2 GOALS Data Sets

Open New Window

2. SCIENTIFIC OBJECTIVES OF GOALS

While a great deal of effort has been devoted to the study of ULIRGs, LIRGs have, in comparison, suffered from a lack of attention. Ground-based optical and near-infrared imaging studies of low-redshift LIRGs have been performed (e.g., Ishida 2004), but multiwavelength imaging and spectroscopic surveys of a large sample of the nearest and brightest LIRGs have not been completed. This is undoubtedly due to the fact that LIRGs form a morphologically diverse group of galaxies, unlike ULIRGs which are nearly always involved in the final stages of a violent and spectacular merger. LIRGs are also much harder to detect and analyze at high redshift in many shallow surveys, which has traditionally removed much of the impetus behind establishing a robust library of low-redshift analogs. This has changed recently, with surveys such as COSMOS (Scoville et al. 2007) and GOODS (Dickinson et al. 2003), which include significant samples of LIRGs at cosmologically interesting epochs.

An important question regarding the nature of LIRGs is, what is the source of their power? What mechanism is responsible for generating energy at a rate that is tens to hundreds of times larger than that emitted by a typical galaxy? Interactions between large, late-type galaxies are evident in many systems, but a significant fraction of LIRGs show no morphological evidence (e.g., double nuclei, tidal tails, stellar bridges, etc.) for a major merger. Large molecular gas masses, large gas fractions, or unusual dynamics may all play a role. Regardless of the trigger, an AGN or an intense starburst is often invoked as the most likely source of the energy in LIRGs. The relative importance of AGN and starburst activity within individual systems, for the generation of far-infrared radiation from LIRGs as a class, and how the dominant energy source changes as a function of merger stage, are important questions that can only be answered by studying a large, unbiased sample at a wide variety of wavelengths. This is the primary motivation behind the GOALS project.

By their very nature, LIRGs are dusty galaxies, wherein a large fraction (over 90% for the most luminous systems) of the UV light emitted by stars and/or AGN is absorbed by grains and reradiated in the far-infrared. This makes traditonal UV and optical spectroscopic techniques for detecting and measuring starbursts and AGN (e.g., emission-line ratios and line widths) difficult to link quantitatively to the ultimate source of the IR emission. However, by combining these data with observations in the infrared, hard x-ray, and radio, it is possible to penetrate the dust and better understand the LIRG engine.

With the IRAS data we have been able to construct unbiased, flux-limited samples of LIRGs and measure their global far-infrared colors, albeit with little or no information about the distribution of the dust, due to the large IRAS beam sizes. The overall LIRG Spectral Energy Distributions (SEDs) are shaped by the relative locations of the dust and the young stars, as well as the presence of an active nucleus. At a basic level, the size of the warm dust-emitting region can give an indication of the power source, as a starburst will be spread out on kpc scales while the region heated by an AGN will be 2–3 orders of magnitude smaller. Spitzer imaging with IRAC and MIPS can give us our first estimates of the size of the IR-emitting regions in the nearest LIRGs, although higher spatial resolution is required for the vast majority of systems if we are to place meaningful constraints on, for example, the luminosity density of the nuclear starburst. Ground-based 10–20 μm imaging can provide a high spatial resolution view of the nucleus, but these wavelengths are most sensitive to the hottest dust, and extrapolations to the far-infrared are always required. The IRAC data provide the finest spatial resolution on Spitzer, and as the IRAC bands for low-redshift galaxies are dominated by stars, hot dust, and finally polycyclic aromatic hydrocarbon (PAH) emission as one progresses from 3.6 μm to 8 μm, the nuclear colors themselves can indicate the presence of an AGN (Lacy et al. 2004; Stern et al. 2005). Although the spatial resolution of the MIPS data is limited, the flux ratios give an indication of the dust temperature and hence, the importance of an AGN to the overall energy budget. When coupled with the radio continuum data, the MIPS fluxes can also be used to estimate the radio–far-infrared flux ratio, itself an indicator of the presence of a (radio loud) AGN.

Mid-infrared and X-ray spectroscopy provide powerful tools for not only uncovering buried AGN, but also measuring the physical properties of the gas and dust surrounding the central source. The low-resolution IRS spectra are ideal for measuring the broad PAH features at 6.2, 7.7, 8.6, 11.3, and 12.7 μm; the silicate absorption features at 9.7 and 18 μm; and water or hydrocarbon ices at 5–7 μm (e.g., Spoon et al. 2004; Armus et al. 2007). The PAH ratios indicate the size and ionization state of the small grains (e.g., Draine & Li 2001), while the silicate absorption gives a direct measure of the optical depth toward the nuclei and clues as to the geometry of the obscuring medium (e.g., Levenson et al. 2007, Sirocky et al. 2008). The ratio of the PAH emission to the underlying (hot dust) continuum can be used as a measure of the strength of an AGN, as Seyert galaxies and quasars typically have extremely low PAH equivalent widths (EQW) compared to starburst galaxies (Genzel et al. 1998; Sturm et al. 2002; Armus et al. 2004, 2006, 2007). Among ULIRGs, the PAH EQW have even been shown to decrease with increasing luminosity (Tran et al. 2001; Desai et al. 2007), suggesting an increasing importance of an AGN in the highest luminosity systems. Since the strengths of the PAH features can dominate broadband, mid-infrared filter photometry, it is critical to understand how the relative strengths of the PAH features vary within the LIRG population and how they affect the colors with redshift (see Armus et al. 2007). In addition, the shape of the silicate absorption can indicate the presence of crystalline silicates, themselves a clear sign of violent, and recent, grain processing (e.g., Spoon et al. 2006).

The high-resolution IRS spectra are ideally suited for measuring the equivalent widths and relative line strengths of the narrow, fine-structure atomic emission lines such as [S IV] 10.5 μm, [Ne II] 12.8 μm, [Ne III] 15.5 μm, [Ne V] 14.3 and 24.3 μm, [O IV] 25.9 μm, [S III] 18.7 and 33.5 μm, [Si II] 34.8 μm, [Fe II] 17.9 and 25.9 μm, as well as the pure rotational H₂ lines at 9.66 μm, 12.1 μm, 17.0 μm, and 28.2 μm. These lines provide a sensitive measure of the ionization state and density of the gas, and they can be also used to infer the basic properties (e.g., T_eff) of the young stellar population. The H₂ lines probe the warm (100–500 K) molecular gas and, when combined with measurements of the cold molecular gas via the mm CO line, can be used to infer the warm-to-cold molecular gas fraction as a function of evolutionary state (e.g., Higdon et al. 2007). The high-resolution spectra can also indicate the presence of a warm, interstellar medium, via absorption features of C₂H₂ and HCN at 13.7 μm and 14.0 μm, respectively (see Lahuis et al. 2007).

The low background of the ACIS detector, together with the excellent spatial resolution of Chandra, provide an unprecedented view of the 0.4–8 keV X-ray emission in LIRGs. X-ray binaries, an AGN, and hot gas associated with an outflowing wind can all contribute to the measured X-ray emission on different physical scales and at different characteristic energies. The X-ray images will allow us to quantify the amount and extent of resolved X-ray emission in LIRGs as an indication of extent of the starburst, and/or the presence of an outflowing wind. The winds are driven through the combined action of overlapping supernovae, producing both a hot central component and soft, extended emission, made up predominantly of shocked, swept-up gas in the ISM (e.g., Fabbiano et al. 1990; Heckman et al. 1990; Armus et al. 1995; Read et al. 1997; Dahlem et al. 1998; Strickland et al. 2004a, 2004b). An AGN, on the other hand, will produce unresolved hard X-ray emission, sometimes accompanied by a strong neutral or ionized iron line.

The penetrating power of the hard X-rays implies that even obscured AGN should be visible to Chandra (e.g., Komossa et al. 2003). Unresolved, hard X-ray (2–8 keV) flux with strong iron line emission is a clear indication of an AGN, and by fitting the spectrum we can derive both the intrinsic luminosity and the HI column toward the accretion disk(s). Of course, identifying the true nature of the nuclear emission from the Chandra spectra alone is difficult for sources where the HI column density exceeds ∼10²⁴ cm^-2, but the combination of X-ray spectral and spatial information can help disentangle the extended starburst emission from the unresolved, harder emission from an AGN in many LIRGs. Of particular interest among the LIRG sample is whether or not the X-ray luminosity can be used as a quantitative measure of the star-formation rate. In galaxies dominated by star formation, where the X-ray emission is dominated by high-mass X-ray binaries, there is a correlation between the X-ray luminosity and the star-formation rate as measured by the infrared or radio emission (Ranalli et al. 2003; Grimm et al. 2003). It will be interesting to determine whether or not this relation holds for the powerful starbursts found in LIRGs.

The radio emission from galaxies is immune to the effects of dust obscuration, and the far-infrared to radio flux ratio, q (Helou et al. 1985), shows a tight correlation for star-forming galaxies (Yun et al. 2001). In addition, the radio spectral index and the radio morphology can be used to identify buried AGN and/or jets in the nuclei of LIRGs at the high spatial resolution afforded with the VLA. The available GHz radio data for the GOALS sample (Condon et al. 1990) will therefore provide a critical baseline for understanding the power sources and the ISM properties of the GOALS sample. In particular, it will be valuable to compare the radio and far-infrared morphologies of the nearest LIRGs and explore the far-infrared–radio correlation in detail within luminous infrared galaxies, such has been done for other nearby samples (e.g., Murphy et al. 2006).

In addition to uncovering buried AGN and nuclear starbursts in LIRGs, GOALS provides an excellent opportunity to study star formation in LIRG disks as a function of merger stage. The wide wavelength coverage and sensitivity provide a detailed look at the old and young stars, the dust, and the gas across the merger sequence. GOALS will allow us to answer some important questions about galactic mergers, such as, what fraction of the star formation occurs in extranuclear (super) star clusters? Can the ages and locations of these clusters be used to reconstruct the merger history and refine detailed models of the merger process, and what are the physical processes that drive star-formation on different scales in the merger? A simple question that still needs to be addressed is the range in UV properties of LIRGs. Are LIRGs as a class, weak or strong UV emitters? How much of the ionizing flux emerges, and is the ratio of IR to UV emission related to any of the other properties, e.g., the merger state or the distribution of the dust?

GALEX has provided our first look at the properties of large numbers of UV-selected starburst and AGN as a function of cosmic epoch (e.g., Gil de Paz et al. 2007; Xu et al. 2007; Martin et al. 2007; Schiminovich et al. 2007). By combining the GALEX and Spitzer data, we will be able to make a full accounting of the energy balance throughout the GOALS sample, relating the UV emission to the infrared photometric and spectroscopic properties of a complete sample of low-redshift LIRGs. In the GOALS targets that are resolved by GALEX (a small but important subsample), we will be able to separate the nucleus and disk in single systems, and the individual galaxies in interacting systems, in order to measure the relative UV emission escaping from the systems and directly compare this to the far-infrared emission measured with Spitzer. This will provide a detailed look at the variation in the IR to UV ratio within interacting galaxies at all stages.

The HST imaging data in GOALS provides our sharpest look at the stellar clusters and detailed morphological features in LIRGS. The depth and high spatial resolution of the HST imaging with the Advanced Camera for Surveys (ACS) Wide Field Camera (WFC) allows for a sensitive search for faint remnants of past, or less-disruptive (minor) mergers (e.g., shells, fading tails, etc.), small scale nuclear bars, and even close, double nuclei, although the latter are notoriously difficult to measure at optical wavelengths in heavily obscured systems. These data will be used to fit models of galaxy surface brightness profiles with standard programs (e.g., GALFIT) and estimate basic structural parameters (e.g. bulge to disk ratio, half-light radius) for the LIRGs as a class, and as a function of nuclear properties (e.g., AGN or starburst-dominated spectra as determined from the Spitzer IRS and Chandra data).

The subarcsec resolution of the ACS also allows a sensitive search for and quantitative measurement of the young, star-forming clusters in LIRGs. The cluster populations can be placed on color magnitude diagrams in order to estimate their ages, and the number and luminosity of such systems can be studied as a function of the relative age of the merger. Particularly powerful in studying the cluster ages will be the combination of visual and UV imaging with ACS (the latter obtained with the Solar Blind Channel, SBC) as the colors will then allow us to break the age-reddening degeneracy over a large range in cluster ages (10⁶–10⁸ yrs). The cluster results can then be compared to models of the interaction, as has been done in some well-studied mergers (e.g., NGC 4038/9–Whitmore & Schweizer 1995; Whitmore et al. 2005, 2007; Mengel et al. 2001, 2005). For example, it has recently been shown that while the interacting LIRG NGC 2623 possesses a large number of young star clusters that probably formed in the interaction, they account for a negligible fraction (< 1%) of the bolometric luminosity (Evans et al. 2008).

Our understanding of the UV emission properties of LIRGs will be greatly enhanced by the addition of the high-resolution ACS far-UV imaging data. Only a handful of LIRGs have been imaged in the vacuum ultraviolet at high resolution (e.g., Goldader et al. 2002). Although the study of the cluster populations is the primary scientific goal for the ACS imaging, they also will be extremely valuable for understanding the physical conditions that drive luminous infrared galaxies away from the well-known IRX-β correlation—the correlation between infrared excess and UV spectral slope (see Meurer et al. 1999; Goldader et al. 2002). ULIRGs fall well off the correlation established for starburst galaxies, and this is usually interpreted simply as disconnect between the sources of the observed UV and far-infrared emission caused by dust obscuration. However, the physical conditions over which this occurs and when this sets in during the merger process is not known. By combining the ACS, GALEX, and Spitzer imaging data, we will be able to greatly extend the luminosity range over which the UV (and IR) emissions have been mapped, as well as to sample galaxies along the merger sequence.

To understand the true stellar distribution in the dusty circumnuclear regions in LIRGs, it is important to combine high-resolution and long-wavelength imaging. The subarcsecond HST NICMOS imaging in GOALS will provide us with our clearest picture of the stellar mass in the LIRG nuclei, and allow us to accurately deredden the stellar clusters seen in the ACS data. With these data we will be able to search for very small-scale, extremely red point sources that may be buried AGN, and find secondary nuclei hidden from even the deepest ACS optical observations. In addition, we will be able to find optically faint or even invisible clusters around the nuclei, and together with the ACS data, to assemble a true cluster luminosity function unaffected by dust. The H-band imaging provides us with the best (existing) near-infrared imaging resolution, and it samples the peak in the stellar light curve for the old stars that will dominate the mass in these galaxies. While the NICMOS F160W and the IRAC 3.6 μm data are both effective tracers of the old stars and stellar mass, the NICMOS data have a spatial resolution that is an order of magnitude higher–on the order of 100 pc or better for many of the LIRGs. By combining the IRAC imaging of the entire sample with the NICMOS imaging of the luminous GOALS systems, we will be able to trace the stellar masses from the largest to the smallest scales.

A basic result from the Spitzer imaging in GOALS will be a much better understanding of the distribution of the dust in mergers. Since many of the LIRGs are interacting, yet not fully merged, the IRAC and MIPS images allow us to separate the contribution from each galaxy to the total far-infrared flux as measured by IRAS. They also allow us to separate nuclear from disk emission in many systems, pinpointing the location of the enhanced infrared luminosity as a function of merger state. The MIPS images provide a clean separation between the warm dust (T ∼ 50–100 K) at 24 μm and 70 μm, and the cold dust, which contributes mostly at 160 μm. One of the most remarkable results from ISO was the discovery that most of the far-infrared emission in the merging galaxy NGC 4038/9 (the “Antennae”) originates from the disk overlap region, which is extremely faint in the UV and optical (Mirabel et al. 1998). The Spitzer GOALS images will allow us to find other systems like the Antennae, and determine the frequency of this completely dust-enshrouded starburst state among the LIRG population as a whole.

Taken together, the imaging and spectroscopic data in GOALS will provide the most comprehensive look at the LIRG population in the local Universe. By studying LIRGs from the X-ray through the radio and far-infrared, we can piece together an understanding of the stars, central black holes, and the interstellar medium (ISM) in some of the most actively evolving galaxies today. These observations will be invaluable as a local library with which to interpret the limited photometric and spectroscopic information garnered from deep surveys of LIRGs at high redshift, providing an in-depth glimpse of the coevolution of starbursts and black holes at late times, and the role of feedback on the ISM of merging galaxies.

3. SAMPLE DEFINITION AND PROPERTIES

The IRAS Revised Bright Galaxy Sample (RBGS; Sanders et al. 2003) is a complete sample of extragalactic objects with IRAS S₆₀ > 5.24 Jy, covering the full sky above a Galactic latitude of |b| > 5 degrees. The RBGS objects are the brightest 60 μm sources in the extragalactic sky, and as such they are the best sources for studying the far-infrared emission processes in galaxies and for comparing them with observations at other wavelengths. The 629 objects in the RBGS all have z < 0.088, and near-infrared properties spanning a wide range from normal, isolated gas-rich spirals at low luminosities (L_IR < 10^10.5 L_⊙) through an increasing fraction of interacting galaxy pairs and ongoing mergers among the more luminous LIRGs and ULIRGs. The sample includes numerous galaxies with optical nuclear spectra classified as starbursts, Type 1 and 2 Seyfert nuclei, and LINERS. The 21 ULIRGs (3%) and 181 LIRGs (29%) in the RBGS form a large, statistically complete sample of 202 infrared-luminous, local galaxies which are excellent analogs for comparisons with infrared and submm selected galaxies at high redshift. These infrared-luminous sources define a set of galaxies sufficiently large to sample each stage of interaction and provide a temporal picture of the merger process and its link to the generation of far-infrared radiation. The full sample of LIRGs, along with their basic properties, is listed in Table 1. Note that 77 of the LIRG systems contain multiple galaxies. From this point on we refer to the 202 LIRGs as “systems,” comprising approximately 291 individual galaxies. The median distance to the LIRGs in the GOALS sample is 94.8 Mpc. Throughout this article we adopt H₀ = 70 km s^-1 Mpc, Ω_vacuum = 0.72, and Ω_matter = 0.28.

Note that an update to the cosmological parameters, primarily H₀ = 70 km s^-1 Mpc (Hinshaw et al. 2009) instead of H₀ = 75 km s^-1 Mpc, which was used when the IRAS Revised Bright Galaxy Sample was compiled (RBGS; Sanders et al. 2003), results in 17 additional RBGS sources classified as LIRGs, namely: IRAS F01556 + 2507 (UGC 01451, log(L_IR/L_⊙) = 11.00); F02072 - 1025 (NGC 0839, log(L_IR/L_⊙) = 11.01); F04296 + 2923 (log(L_IR/L_⊙) = 11.04); F04461 - 0624 (NGC 1667, log(L_IR/L_⊙) = 11.01); F06142 - 2121 (IC 2163, log(L_IR/L_⊙) = 11.03); NGC 2341, log(L_IR/L_⊙) = 11.17); F10221 - 2318 (ESO 500-G034, log(L_IR/L_⊙) = 11.01); F11122 - 2327 (NGC 3597, log(L_IR/L_⊙) = 11.05); F12112 - 4659 (ESO 267-G029, log(L_IR/L_⊙) = 11.11); F12351 - 4015 (NGC 4575, log(L_IR/L_⊙) = 11.02); F14004 + 3244 (NGC 5433, log(L_IR/L_⊙) = 11.02); F14430 - 3728 (ESO 386-G019, log(L_IR/L_⊙) = 11.01); F15467 - 2914 (NGC 6000, log(L_IR/L_⊙) = 11.07); F17468 + 1320 (CGCG 083-025; log(L_IR/L_⊙) = 11.05); F19000 + 4040 (NGC 6745, log(L_IR/L_⊙) = 11.04); F22025 + 4204 (UGC 11898, log(L_IR/L_⊙) = 11.02); F22171 + 2908 (Arp 278, log(L_IR/L_⊙) = 11.02). Because these galaxies were not part of the original LIRG sample drawn from the RBGS, they are not included in GOALS.

4. GOALS OBSERVATIONS

5. GOALS DATA PRODUCTS

On regular, approximately six-month, intervals, Spitzer images and spectra will be delivered to the Spitzer Science Center and made public through their Legacy program web pages. The first delivery is now available through the Spitzer Legacy web pages at the Spitzer Science Center (http://ssc.spitzer.caltech.edu/legacy/all.html), and the Infrared Science Archive (IRSA) at the Infrared Processing and Analysis Center (http://irsa.ipac.caltech.edu). For GOALS, the delivered products include: (1) IRAC image mosaics in all four IRAC bands. Images are single-extension FITS files with a pixel scale of 0.6^″pixel^-1. (2) MIPS image mosaics in all three MIPS bands. Images are single-extension FITS files, with wavelength-dependent pixel scales–1.8^″pixel^-1 at 24 μm, 4.0^″pixel^-1 at 70 μm, and 8.0^″pixel^-1 at 160 μm. (3) IRS nuclear spectra in all four low and high-resolution modules. Spectra are delivered in ASCII (IPAC table) format, similar to those produced by the IRS pipelines; (4) Spatial profiles at two positions along the IRS Short-Low slit (the two nod positions), in the 8.6 μm PAH, 10 μm continuum, and 12.8 μm [Ne II] fine-structure emission line. Profiles are provided in ASCII format, in units of e^- s^-1 for each pixel. Also included in each file are the profiles at the same wavelengths for an unresolved star. The IRAC images are corrected for “banding,” “muxbleed,” and “column pulldown” effects, as described in the IRAC Data Handbook. The MIPS images are corrected for latents, weak jailbars, and saturation effects, where possible. The IRAC and MIPS mosaics were constructed using MOPEX to align, resample, and combine the data. The IRS spectra are extracted from the two-dimensional BCD pipeline products using the SPICE package. A standard “point source” extraction has been used for all sources. The two nod positions and multiple exposures (“cycles”) are used to remove residual cosmic rays and/or bad pixels in the spectra.

6. THE LUMINOUS INFRARED GALAXY VV 340

As an example of the utility of the multiwavelength GOALS data, we present images and spectra of the LIRG, VV 340 (IRAS F14547+2449; Arp 302). VV 340 consists of two large, spiral galaxies, one face-on and one edge-on, separated by approximately 40″ (27.3 kpc). The two spirals are apparently in the early stages of a merger (Bushouse & Stanford 1992, Lo et al.1997). This system has an infrared luminosity of 5 × 10¹¹ L_⊙ (Sanders et al. 2003).

6.2. Results & Discussion

The Chandra, GALEX, HST, and Spitzer images of VV 340 are shown in Figure 1. A false-color image made from the ACS F435W and F814W data is shown in Figure 2. The thick, flaring dust lane makes VV 340 N much fainter in the UV than VV 340 S, but the disk and bulge are prominent in the IRAC images. Using the ACS images, we have detected 173 unresolved “clusters” in the VV 340 system, with most of these (159, or 92%) being in the spiral arms of VV 340 south. The apparent magnitudes range from 21–27 mag in B, with absolute magnitudes ranging from 9 to -13 mag, and B - I colors of 0.5–2 mag. These colors are typical for the GOALS LIRGs as a class (Vavilkin et al. 2009, in preparation). Assuming a typical cluster mass of 10⁶ M_⊙, the ages of these clusters are ≤ 10⁸ yr, uncorrected for reddening. As a whole, the observed clusters in VV 340 south account for 3% of the total B-band light.

(52 KB)

Fig. 1.— Multiwavelength images of VV 340. From upper left to lower right, these are (a) Chandra 0.4–8 keV, (b) GALEX FUV, (c) GALEX NUV, (d) HST ACS F435W, (e) HST ACS F814W), (f) Spitzer IRAC 3.6 μm, (g) Spitzer IRAC 4.5 μm, (h) Spitzer IRAC 5.8 μm, (i) Spitzer IRAC 8 μm., (j) Spitzer MIPS 24 μm, (k) Spitzer MIPS 70 μm, and (l) Spitzer MIPS 160 μm. In all cases a projected linear scale of 5 kpc at the distance of VV 340 is indicated by a bar in the lower left.

Open New Window

(17 KB)

Fig. 2.— False color image of VV 340 (UGC 9618) made from the HST ACS F435W and F814W data. The image is 100″ (68 kpc) on a side. North is up and east is to the left.

Open New Window

The total far-infrared flux densities for the VV 340 system, as measured with MIPS, are 0.43, 9.38, and 15.73 Jy at 24, 70, and 160 μm, respectively. The northern galaxy dominates at 24, 70, and 160 μm, accounting for approximately 80%, 82%, and 95%, respectively, of the total emission from the system (the 160 μm ratio is uncertain because of the large size of the point-spread function, PSF). For reference, the IRAS 25, 60, and 100 μm flux densities for the entire VV 340 system are 0.41, 6.95, and 15.16 Jy (Sanders et al. 2003). It is interesting to note that the ratio of far-infrared fluxes implies that VV 340 south, by itself, is not a LIRG (log L_IR ∼ 10.71 L_⊙).

The IRAC 3.6–4.5 and 5.8–8 colors, as measured in 10 kpc radius apertures centered on the two nuclei, are 0.19 and 1.97 mag for VV 340 north, and 0.04 and 1.96 mag for VV 340 south, respectively. The nuclear (2 kpc radius aperture) colors are slightly redder, being 0.29 and 2.06 mag for VV 340 north, and 0.06 and 2.16 mag for VV 340 south, respectively. Both the large and small aperture measurements place both galaxies in VV 340 outside of the AGN “wedge” in the IRAC color-color diagram of Stern et al. (2005).

The Spitzer IRS low-resolution spectra of the northern and southern nuclei are shown in Figure 3. VV 340 N has a steeper spectrum and a deeper silicate absorption (τ_9.7 ∼ 1.3 for VV 340 N). The PAH emission features are prominent in both galaxies, although the 17 μm feature appears much stronger, with respect to the continuum and the fine structure lines, in VV 340 S. Both [Ne V] 14.3 and 24.3 μm lines are seen in the high-resolution spectrum of VV 340 N, as is the [O IV] 25.9 μm line. The line fluxes are 1.2 × 10^-17 W m^-2, 2.3 × 10^-17 W m^-2, and 12.3 × 10^-17 W m^-2, repectively, for the [Ne V] 14.3, 24.3 μm, and [O IV] 25.9 μm lines in VV 340 N. A portion of the Short-High spectra of VV 340 N and S surrounding the location of the [Ne V] 14.3 μm emission line is shown in Figure 4. The [Ne II] 12.8 μm line has a flux of 23.4 × 10^-17 W m^-2 in VV 340 N, and 6.1 × 10^-17 W m^-2 in VV 340 S. In VV 340 N, where the 6.2 μm PAH EQW is 0.48 μm, the [Ne V]/[Ne II] and [O IV]/[Ne II] line flux ratios are consistent with the presence of a weak AGN, contributing less than 10–15% of the IR emission, based on scaling from IRS spectra of local AGN and starburst nuclei (see Armus et al. 2007 and references therein). As the diagnostic ratios are calculated from the nuclear spectrum, this is an upper limit to the contribution from an AGN to the global IR emission in this source. Alternatively, the coronal-line region could be sitting behind at least 50 mag of visual extinction, assuming an intrinsic [Ne V]/[Ne II] flux ratio of unity. This is much larger than that implied by the optical depth at 9.7 μm (A_V ∼ 22 mag) for VV 340 N. There is no [Ne V] detected from VV 340 S ([Ne V]/[Ne II] < 0.04), and the 6.2 μm PAH EQW is 0.53 μm, comparable to pure starburst galaxies (Brandl et al. 2006). The lack of [Ne V] emission and the strong PAH EQW both suggest that the southern nucleus has no detectable AGN in the mid-infrared.

(21 KB)

Fig. 3.— IRS low-resolution spectra of the nucleus of VV 340 north (above), and south (below). Prominent emission features are marked. The offset at 14 μm in the spectrum of VV 340 S is real, and represents a flux difference in the SL and LL spectra for this source, most likely due to significant extended emission in the LL slit.

Open New Window

(20 KB)

Fig. 4.— IRS Short-High spectra of the nucleus of VV 340 north (above), and south (below) from 14–14.7 μm in the rest frame. Fits to the PAH 14.22 μm emission feature, as well as the [Ne V] 14.322 μm, and [Cl II] 14.368 μm emission lines indicated in green and red, respectively in VV 340 N, as is the underlying local continuum (blue dashed line). The lines are not detected in VV 340 S.

Open New Window

The H₂ S(2), S(1), and S(0) rotational lines at 12.28, 17.03, and 28.22 μm are detected in VV 340 N with line fluxes of 5.67 × 10^-17 W m^-2, 4.56 × 10^-17 W m^-2, and 1.3 × 10^-18 W m^-2, respectively. In VV 340 S, only the S(2) and S(1) lines are seen, with fluxes of 4.4 × 10^-18 W m^-2, and 1.0 × 10^-17 W m^-2, respectively. Fits to the emission lines imply masses and temperatures for the warm gas components of 1.73 × 10⁷ M_⊙ at 530 K in VV 340 N, and 7.8 × 10⁶ M_⊙ at 310 K in VV 340 S. The nuclear warm gas fractions (warm/cold) are then ∼3 × 10^-4 and ∼7 × 10^-4 in VV 340 N and VV 340 S, respectively, using the masses of cold H₂ derived by Lo et al. (1997). However, the large extent of the CO emission, 23 kpc in VV 340 N and 10 kpc in VV 340 S (much larger than the IRS slit widths) suggests that these ratios are probably lower limits for the nuclei. The large extent of the CO, the relatively low infrared luminosity to H₂ mass ratio in both galaxies, and the regular kinematics led Lo et al. (1997) to suggest that the VV 340 system is in an early interaction, prestarburst, phase.

VV 340 has a total (north plus south) UV flux of 17.55 mag and 16.64 mag in the far and near-UV GALEX filters, respectively. The pair has an apparent infrared to UV flux ratio (the infrared excess, or IRX), of 81.3. The measured UV slope, β, is -0.38 in the GALEX system (see Kong et al. 2004), placing VV340 well above the fit to local starburst galaxies (Meurer et al. 1999; Goldader et al. 2002) by nearly an order of magnitude. As is obvious from Figure 1, most of the UV emission comes from VV 340 south, while most of the IR emission comes from VV 340 north. If we were to place the galaxies on an IRX-β plot individually, the northern galaxy would have an IRX ∼398 and a β ∼ -0.39, while the southern galaxy would have an IRX ∼17 and a β ∼ -0.44. This would place the southern source close to the starburst correlation, but the northern source much farther off the correlation than the system as a whole. A full discussion of where the LIRGs fall on the the IRX-β plot is given in Howell et al. (2009).

VV 340 is detected at both 1.49 and 4.85 GHz with the VLA (Condon et al. 1990; 1995). While both VV 340 N and S are resolved in the VLA C-array data, the A-array data also reveal a nuclear component and two “hotspots” at radii of about 5″, oriented N-S along the edge-on disk of VV 340 N. VV 340 N and VV 340 S have 1.49 GHz flux densities of 86, and 12 mJy, respectively, and 4.85 GHz flux densities of 30 and 3.5 mJy, respectively. The system has a logarithmic far-infrared to radio flux ratio, q (Helou et al. 1985) of q = 2.02 dex. This is within 1.5 σ of the standard value for star-forming galaxies (q = 2.34 ± 0.26 dex; Yun et al. 2001). VV 340 N and VV 340 S have q values of 2.02 and 2.00 dex, respectively. VV 340 S has a slightly steeper radio spectral index (α = -1.0) than does VV 340 N (α = -0.8), the latter being comparable to the standard value seen in star-forming galaxies (Condon 1992).

Chandra X-ray imaging of VV 340 is shown in Figures 5 and 6. While both galaxies are detected, the emission from the southern galaxy is much weaker, and more diffuse. The northern galaxy is clearly resolved in the soft bands (0.5–2 keV), with a size of nearly 25″ (17 kpc), but shows a hard (2–7 keV) X-ray point source coincident with the nucleus. The hard X-ray image also shows two faint blobs to the north and south, aligned with the radio structure (Condon et al. 1990).

(42 KB)

Fig. 5.— Soft (0.2–5 keV) X-ray image of VV 340 taken with the Chandra X-Ray Observatory (contours), overlaid on the HST ACS F435W image (greyscale). North is up, and east is to the left.

Open New Window

(27 KB)

Fig. 6.— The soft (left) and hard (right) X-ray images of VV 340 north taken with the Chandra X-Ray Observatory. The images have been smoothed with a Gaussian kernel with σ = 2 pixels. The orientation of the image is north up and east to the left.

Open New Window

The X-ray spectrum of VV 340 N (Fig. 7) is dominated by emission-line rich, extended soft X-ray emission but an excess at high energies above 4 keV is evident. In Figure 7 we compare the Chandra spectrum of VV 340 N to that of the starburst-dominated LIRG VII Zw 31 (Iwasawa et al. 2009b, in preparation). VII Zw 31 shows no hard-band excess above 4 keV. The hard-band excess in VV 340 N peaks at approximately 6.4 keV in the rest frame, suggesting a pronounced Fe K line. With the limited detected counts, the exact spectral shape cannot be constrained. However, the most likely interpretation for this hard excess is the emission from a heavily absorbed active nucleus with an absorbing column density (N_H) of 10²⁴ cm^-2 or larger. Note, this is significantly larger than that implied by the silicate optical depth at 9.7 μm. If the observed emission is reflected light from a Compton thick AGN, the intrinsic luminosity could be as high as 2 × 10⁴¹ erg s^-1 in the 2–7 keV band. There is also evidence for the 1.8 keV Si line, which could have a contribution from AGN photoionization, although it could also signal the presence of a large number of core-collapse SNe.

(25 KB)

Fig. 7.— The X-ray spectrum of VV 340 north obtained from the ACIS-S. For display purposes, the original data have been rebinned so that each spectral bin contains at least five counts. The vertical axis is in units of 10^-14 erg cm^-2 s^-1 keV^-1. Also included for comparison is the ACIS-S spectrum of VII Zw 31 from Iwasawa et al. (2009b, in preparation), scaled by 0.02, which is an example of a starburst-dominated LIRG. The hard X-ray excess in VV 340 north is evident at energies above about 4 keV.

Open New Window

Taken together, the Spitzer, Chandra, GALEX, HST, and VLA data suggest that the interacting system VV 340 is composed of two very different galaxies. The edge-on disk of VV 340 north hides a buried AGN seen in the mid-infrared and X-rays, although the apparent contribution of this AGN to the total energy of the galaxy is low. In contrast, VV 340 south is a face-on starburst galaxy that generates an order of magnitude less far-infrared flux, but dominates the short-wavelength UV emission from the system. VV 340 is a LIRG because of the enhanced emission coming from VV 340 N alone. VV 340 seems to be an excellent example of a pair of interacting galaxies evolving along different paths, or at different rates, implying that the details of the interaction can produce LIRGs whose global properties mask the true nature of the emission. This is particularly relevant for merging galaxies viewed at high redshift.