WISE DISCOVERY OF LOW-METALLICITY BLUE COMPACT DWARF GALAXIES

In the local universe, we observe a class of galaxies known as blue compact dwarf galaxies (BCDs) which show signatures of strong, recent, and ongoing star formation activity in low-metallicity environments. These galaxies, which appear to be experiencing their first major burst of star formation, are characterized as generally having compact sizes (⩽1 kpc), low luminosities (M_B ⩾ −18), strong emission line spectra, blue optical colors (U − B ∼ −0.6), and their metal distribution peaks at one-tenth solar (Z = 1/10 Z_☉; Kunth & Östlin 2000). Typical Hβ equivalent widths range from 30 to 200 Å, implying a burst of star formation no older than a few Myr (Kunth 1999). BCDs offer pristine laboratories from which to study star formation processes, enrichment mechanisms for the interstellar medium, and the star formation history of galaxies in low metallicity and extremely dusty environments. In particular, BCDs are likely good analogs of the first generation of galaxies to form in the universe, but are at redshifts accessible to detailed study (e.g., Meier & Terlevich 1981). For a comprehensive review of BCDs and the most metal-poor galaxies, see Kunth & Östlin (2000).

Although discovered as early as the 1960s, BCDs and extremely low metallicity galaxies have remained an elusive population. The difficulty in finding low-metallicity BCDs can be viewed as stemming from several factors. First, they have low luminosities, generally much fainter than L*. Second, at the lowest metallicities, below 0.1 solar, the major coolant species shift from [O iii] to H and He (Kunth & Sargent 1986), resulting in lower [O iii] λ5007/Hβ ratios. This made such sources less readily identified in the previous generation of objective prism surveys (Kunth & Sargent 1986), though recent fiber spectroscopy surveys have less significant biases. Finally, they are rare. The most abundant samples of BCDs have been acquired by the Sloan Digital Sky Survey (SDSS) which has spectroscopically identified hundreds of BCDs at redshifts of a few tenths (e.g., Izotov et al. 2006), implying a surface density of 0.1 BCD deg⁻² to the limits of the SDSS. However, SDSS rarely finds the most extreme metal-poor systems, galaxies with Z ≲ 1/10 Z_☉. For example, Kniazev et al. (2003) estimate 0.004 deg⁻² for extremely metal-poor galaxies [12 + log (O/H) ⩽ 7.65] for r ⩽ 17.77 mag.

The two lowest metallicity BCDs known to date are SBS 0335-052E (Izotov et al. 1990; Melnick et al. 1992; Izotov et al. 1997; Thuan et al. 1997, 1999; Dale et al. 2001; Houck et al. 2004) and I Zw 18 (Zwicky 1966; Lauqué 1973; Thuan & Martin 1981; Davidson & Kinman 1985; Dufour & Hester 1990; Kunth et al. 1994; Hirashita & Hunt 2004; Wu et al. 2007), both of which have been extensively studied across the electromagnetic spectrum. Though the metallicities for these two systems are very similar, 12 + log (O/H) < 7.3 or Z < 1/40 Z_☉, these galaxies exhibit very different morphological and physical properties. Hirashita & Hunt (2004) investigated these differences and suggest that they are best understood if two different modes of star formation are considered, an "active mode" and a "passive mode." SBS 0335-052E is an active-mode BCD while I Zw 18 is a passive-mode BCD. In the Hirashita & Hunt (2004) framework, active BCDs host super star clusters (SSCs) and are characterized by compact star formation regions (radius ⩽50 pc), high gas densities (⩾500 cm⁻³), rich H₂ content, large dust optical depth, and high dust temperature. Passive BCDs, in contrast, have more diffuse star-forming regions (⩾100 pc), lower gas density (⩽100 cm⁻³), cooler dust, and lack SSCs and large amounts of H₂.

Hirashita & Hunt (2004) describe how the compactness of the star-forming regions can simultaneously explain most of the observed differences between BCDs. Compact star-forming regions will rapidly become dusty from Type II supernovae, and this dust will be effective at reprocessing photons into the infrared. In particular, the gas free-fall timescale in such systems is less than 5 Myr, which leads to run-away star formation and the efficient creation of large quantities of dust. In contrast, the dynamical time of diffuse star-forming regions, characteristic of passive BCDs, is longer than 10⁷ yr. Such regions thus have lower star formation rates (SFRs) and are less infrared luminous. Hirashita & Hunt (2004) suggest that the physical state of the gas might be driven by the size and spatial distribution of dust grains.

Wu et al. (2008) show that the infrared luminosities of these prototypical BCDs differ by two orders of magnitude, with SBS 0335-052E being ∼100 times more infrared luminous than I Zw 18. The mid-infrared emission of similar low-metallicity BCDs originates mainly from hot (200–1500 K) dust, causing active BCDs to have significantly red colors at mid-infrared wavelengths. For example, Houck et al. (2004) find that the spectrum of SBS 0335-052E peaks at ∼28 μm. Dale et al. (2001) model the infrared emission of SBS 0335-052E and conclude that it can be characterized as having two dust components, a warm (∼80 K) and a hot (∼210 K) component. The prevalence of hot dust in active BCDs suggests that searching for the most extreme low metallicity, star-forming BCDs at mid-infrared wavelengths may be fruitful.

This Letter reports the discovery of two new low-metallicity BCDs discovered by the Wide-field Infrared Survey Explorer (WISE; Wright et al. 2010). Launched on 2009 December 14, WISE completed its first coverage of the entire sky on 2010 July 17, obtaining at least eight exposures per sky position in four passbands, 3.4, 4.6, 12, and 22 μm (W1, W2, W3, and W4, respectively). The corresponding 5σ point source sensitivities in unconfused regions are better than 0.08, 0.11, 1, and 6 mJy, respectively, and the point spread function FWHM values for the four WISE bands are 61, 64, 65, and 120, respectively (Wright et al. 2010). Focusing on two early examples of WISE-identified, low-metallicity BCDs, we discuss the physical properties of these two new BCDs and place them within the context of recent BCD literature. C.-W. Tsai et al. (2011, in preparation) present a larger sample of several dozen newly identified BCDs and discuss the selection technique in more detail. We assume a solar metallicity (Z_☉) of 12 + log (O/H) = 8.96 (Allende Prieto et al. 2001) in this Letter. Magnitudes are reported in the Vega system and, where necessary, we adopt the concordance cosmology with Ω_M = 0.3, $\Omega _\Lambda = 0.7$, and H₀ = 70  km s⁻¹ Mpc⁻¹.

We identified WISEP J080103.93+264053.9 (hereafter W0801+26) and WISEP J170233.53+180306.4 (hereafter W1702+18) from their extreme mid-infrared colors and relatively bright optical flux densities (B < 19.5, from the USNO-B1 catalog). The mid-infrared photometry was acquired from the WISE Preliminary Release Source Catalog (thus the "P" designation in the WISE nomenclature). The simultaneous multi-wavelength WISE observations of W0801+26 and W1702+18 were carried out on UT 2010 April 12 and UT 2010 March 2, respectively, and provided total exposure times of ∼100 and 120 s, respectively. For details on the WISE processing in the Preliminary Release, see the WISE Preliminary Release Explanatory Supplement.⁷ WISE images of the two new BCDs are presented in Figure 1. For comparison, we also show WISE images of the two prototypical low-metallcity BCDs, SBS 0335-052E and I Zw 18. Photometry for all four sources is presented in Table 1. The new BCDs have unique and easily identified mid-infrared colors, similar to SBS 0335-052E. In contrast, I Zw 18 is unremarkable in the WISE data. None of the sources are resolved by the 6''  WISE beam. The signal-to-noise ratios (S/N) of SBS 0335-052E and the new BCDs are >25 in all four WISE bands.

Zoom In Zoom Out Reset image size

Noting their unusual WISE colors and relatively bright optical magnitudes, we obtained optical spectroscopic follow-up observations of W0801+26 and W1702+18 during twilight using the dual-beam Low Resolution Imaging Spectrometer (LRIS; Oke et al. 1995) on the Keck I telescope. We observed W1702+18 on UT 2010 March 12 and W0801+26 on UT 2010 November 8. The conditions were photometric on both nights, and both observations used the 15 wide long slit, the 5600 Å dichroic, and the 400 ℓ mm⁻¹ grating on the red arm of the spectrograph (blazed at 8500 Å; resolving power R ≡ λ/Δλ ∼ 700 for objects filling the slit). The 2010 March observations used the 600 ℓ mm⁻¹ grism on the blue arm (blazed at 4000 Å; R ∼ 1000), while the 2010 November observations used the slightly lower resolution 400 ℓ mm⁻¹ grism (blazed at 3400 Å; R ∼ 600). A single 300 s exposure was obtained of each source during the observing runs. We processed the data using standard procedures, flux calibrated the spectra using observations of standard stars from Massey & Gronwall (1990), and corrected for Galactic extinction using the dust maps of Schlegel et al. (1998).

The final, reduced spectra show a multitude of extremely high equivalent width, narrow emission lines at low redshift and are presented in Figure 2. We have used the ratio of Hα to Hβ and Hγ to Hδ emission line strengths to derive and correct for intrinsic dust extinction assuming Case B recombination and the extinction law of Cardelli et al. (1989) with R_V = 3.1. For W0801+26, the Balmer line ratios are close to the extinction-free ratios of Case B, implying no significant extinction is observed. For W1702+18, we find an intrinsic extinction of A_V = 0.18. The derived redshifts are presented in Table 1, while Table 2 presents the extinction-corrected line flux measurements and errors obtained using the SPECFIT package within IRAF. Equivalent widths from Hβ are also given in Table 2.

Zoom In Zoom Out Reset image size

3.1. Mid-infrared Properties

3.2. Spectroscopic Properties and Metallicity

The low-metallicity and mid-infrared-dominated spectral energy distributions of these newly identified, WISE-selected galaxies place them in a rare class of active, low-metallicity BCDs. We compare the infrared properties of these BCDs with the infrared properties of known BCDs and find that the majority of low-metallicity BCDs in the literature are several magnitudes fainter at 12 μm (W3) than the two new sources. The properties of the two new BCDs are strikingly similar to one of the most well-known BCD, SBS 0335-052E, suggesting that the mechanisms responsible for SBS 0335-052E might be the same mechanism at work in W0801+26 and W1702+18.

In Figure 3 we show WISE W1 − W2 color versus metallicity, 12 + log (O/H), for different samples of BCDs. After visual inspections to remove galaxies that clearly are not BCDs, we include 14 BCDs from Wu et al. (2008), 173 from Izotov et al. (2006), three from Kniazev et al. (2003), SBS 0335-052E, I Zw 18, HS 0822+3542 (Izotov et al. 2006), HS 2236+1344 (Izotov & Thuan 2007), and our two newly discovered BCDs. With the exception of I Zw 18 and HS 0822+3542, we observe a slight correlation between W1 − W2 and metallicity for these samples, with W1 − W2 color becoming increasingly redder with decreasing metallicity. We apply the Spearman's rank correlation coefficient test on the Izotov et al. (2006) sample and recover a Spearman coefficient of ρ = −0.36. For a sample of that size, this indicates a noticeable correlation, deviating from the null-hypothesis at the 4σ level.

Zoom In Zoom Out Reset image size

Since polycyclic aromatic hydrocarbon emission is largely absent in low-metallicity environments (Engelbracht et al. 2008), the mid-infrared spectra of BCDs are dominated by thermal emission with a minor contribution from nebula continuum (Smith & Hancock 2009). Thus, the 3.4 μm to 4.6 μm flux ratio is a relatively clean temperature indicator for the hot dust heated by young stars. At the lowest metallicities, 12 + log (O/H) < 7.8, BCDs seem to group into two different color regions, W1 − W2 ∼ 2.0 and ∼0.5. This suggests two distinct star formation processes can be dominant, supporting the passive and active BCD modes suggested by Hirashita & Hunt (2004).

We are currently conducting follow-up observations of many more promising candidates discovered using the extreme infrared properties exhibited by these two sources and SBS 0335-052E. These results will be presented in a follow-up paper (C.-W. Tsai et al. 2011, in preparation). We expect that WISE will significantly increase the number of low-metallicity BCDs known, perhaps finding the most pristine example in the local universe. Such studies will allow us to construct a statistical sample from which to study and characterize the BCD population. The WISE BCDs will be a unique sample of mid-infrared, bright, low-metallicity galaxies for studying dust grain formation in low-metallicity environments, and will help us to understand the thermal dust emission of high-redshift, starburst galaxies.

The authors thank the anonymous referee for timely and beneficial comments that have improved the manuscript. This publication makes use of data products from the Wide-field Infrared Survey Explorer, which is a joint project of the University of California, Los Angeles, and the Jet Propulsion Laboratory, California Institute of Technology, funded by the National Aeronautics and Space Administration. Some of the data presented herein were 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. The authors also recognize and acknowledge the very significant cultural role and reverence that the summit of Mauna Kea has always had within the indigenous Hawaiian community; we are most fortunate to have the opportunity to conduct observations from this mountain.