The Nearest Ultra Diffuse Galaxy: UGC 2162

In the last few years, there has been a renewed interest in the study of extended low-surface brightness galaxies (Impey et al. 1988; Bothun et al. 1991; Dalcanton et al. 1997; Caldwell 2006). The discovery of dozens of these objects in the Coma Cluster (coined Ultra Diffuse Galaxies (UDGs) by van Dokkum et al. 2015) has been followed by a large number of detections in other clusters (Koda et al. 2015; Mihos et al. 2015; Muñoz et al. 2015; Román & Trujillo 2016a; van der Burg et al. 2016), groups (Merritt et al. 2016; Román & Trujillo 2016b; Smith Castelli et al. 2016), and in the field (Martínez-Delgado et al. 2016). The low stellar mass (10⁷–10⁸ ${M}_{\odot }$) of these objects together with their large size (${R}_{e}\,\gt \,$ 1.5 kpc) have opened a number of questions about the ultimate nature of these galaxies: are UDGs "failed" galaxies (i.e., do they inhabit dark matter halos larger than those expected according to their stellar mass content; van Dokkum et al. 2015; Beasley & Trujillo 2016)? What is the role of environment? Are the properties of UDGs produced by their interaction with dense environments (Yozin & Bekki 2015a)? Are they simply the high-spin tail of normal dwarf galaxies (Amorisco & Loeb 2016)? Are UDGs produced by feedback-driven gas outflows and subsequent dark matter and stellar expansion (Di Cintio et al. 2017)?

Observations indicate that UDGs are a heterogeneous population of dwarf galaxies. Some of them are relatively red ($g-i$ ∼ 0.8), have spheroidal shapes, and inhabit rich galaxy clusters (e.g., van Dokkum et al. 2015), whereas other UDGs are blue ($g-i$ ∼ 0.4), have irregular shapes, and are found in groups (e.g., Román & Trujillo 2016b). Are all these UDGs connected evolutively? Recently, Román & Trujillo (2016b) have suggested a scenario where all this diversity could be understood if UDG progenitors were born in the field, processed by groups, and ended their lives inhabiting clusters. To answer all the above questions and shed more light on the nature of UDGs, it would be extremely useful to have the opportunity to probe, in full detail, the properties of a close (D < 15 Mpc) UDG. This would give us the opportunity to explore its individual stars. In particular, it would be extremely useful to have some information about the gas content of one of these galaxies. In this work, we present the serendipitous discovery of a very nearby UDG: the galaxy UGC 2162. This galaxy is located in the M77 group (at only 12.3 Mpc distance from us) and has HI observations. This proximity allows us to have a superb spatial resolution of 60 pc arcsec⁻¹. In this work, we conduct a detailed analysis of the characteristics of this galaxy and confront the observational data with the theoretical expectations. As we will show, this galaxy is quite rich in HI gas and is currently forming stars at a rate of 0.01 ${M}_{\odot }$ yr⁻¹. If this galaxy were suddenly depleted of its gas, it would evolve into a red ($g-i$ ∼ 0.8) object with ${R}_{e}\,\sim$ 1.8 kpc and ${\mu }_{g}(0)\,\sim$ 27 mag arcsec⁻². All of these are characteristics of the population of the faintest UDGs currently found in rich clusters (Mihos et al. 2015; Beasley et al. 2016).

UGC 2162 (R.A. = 02^h40^m231 and decl. = +01^d13^m45^s) is located within the IAC Stripe82 Legacy Survey (Fliri & Trujillo 2016). The galaxy has a spectroscopic redshift of z = 0.00392. The IAC Stripe82 data set is a careful new co-addition of the SDSS Stripe82 data with the aim of preserving the faintest surface brightness structures. The pixel scale of these images is 0.396 arcsec and the average seeing is 1 arcsec. The following work is based on the rectified images of this data set (http://www.iac.es/proyecto/stripe82/). The mean limiting surface brightness of this data co-addition is 29.1, 28.6, and 28.1 mag arcsec⁻² in the g, r, and i bands respectively (3σ in boxes of 10 × 10 arcsec). To put this data into context, they are ∼1.2 mag deeper than the Dragonfly images used to explore UDGs in Coma (van Dokkum et al. 2015) and similar to Koda et al. (2015).

UGC 2162 is located in the vicinity of M77 (R.A. = 02^h42^m407, decl. = 00^d00^m48^s; z = 0.00379). Its projected radial separation to this galaxy is 13684. A redshift independent measurement of the distance to M77 (Tully et al. 2009) locates this galaxy at a distance of D = 12.3(±1.7) Mpc. Because this is the most massive galaxy of the group, we use its distance as a reliable measurement for the distance of UGC 2162.³ At that distance, the projected radial separation from UGC 2162 to M77 is 293.8(±40.6) kpc and 1 arcsec corresponds to 60(±8) pc. M77 is the central member of the M77 Group. This is a small group of galaxies that also harbors NGC 1055, NGC 1073, UGC 2275, UGC 2302, UGCA 44, and Markarian 600.

Figure 1 shows a color image of UGC 2162 as seen in the IAC Stripe82 images. UGC 2162 appears to be an irregular galaxy, in fact, it has been morphologically classified as Im (de Vaucouleurs et al. 1991). The depth of the Stripe82 image allows us to see that the inner star-forming region of the galaxy is surrounded by an extended disk-like structure.

Zoom In Zoom Out Reset image size

UGC 2162 has been observed by HIPASS (HI Parkes All Sky Survey; Meyer et al. 2004) with a spectral resolution of 18 km s⁻¹. The HIPASS survey detects at the position of the galaxy a HI line with a radial velocity peak of 1171.9 km s⁻¹ (in agreement with the velocity recession of its optical counterpart: 1175 ± 3 km s⁻¹). The HI line flux peaks at 0.089 Jy and has an integrated HI flux density (${{\rm{S}}}_{\mathrm{HI}}$) of 5.4 Jy km s⁻¹. The integrated HI flux density corresponds to a HI mass ${M}_{\mathrm{HI}}$ = 2.36 × 10⁵ × D² × S_HI = 1.9(±0.6) × 10⁸ M_⊙ (see, e.g., Filho et al. 2013). One can also use the HI line width ${W}_{20}$ = 89.7 km s⁻¹ to infer a dynamical mass ${M}_{\mathrm{dyn}}$ = 2.326 × 10⁵ × (W/2)² × ${r}_{\mathrm{HI}}$ within the HI radius (${r}_{\mathrm{HI}}$). To have an estimation of the dynamical mass, it is necessary to correct the line width ${W}_{20}$ for the inclination i of the galaxy, i.e., W = ${W}_{20}$/$\sin i$. We estimate the inclination using the axis ratio of the 27 mag arcsec⁻² isophote (g band). This isophote is still bright enough, but sufficiently far away from the central part of the galaxy, to produce a reliable estimation of the shape of its outer disk. We obtain an axis ratio b/a = 0.7. This translates into an inclination (under the assumption of a thin disk) of $i\,\sim 45^\circ$, and consequently, W = 128 km s⁻¹. We assume ${r}_{\mathrm{HI}}$ to be three times the optical ${R}_{25}$ radius (see, e.g., Filho et al. 2013). For our galaxy, ${R}_{25}$ = 26 arcsec = 1.6(±0.3) kpc (measured in g band). With these values, we estimate ${M}_{\mathrm{dyn}}$= 4.6(±0.8) × 10⁹ ${M}_{\odot }$ within the inner R ∼ 5 kpc.

Once a dynamical mass in the inner region of the galaxy has been estimated, it is possible, using the expected cumulative mass curve, to have an estimation of the total virial mass of the dark matter halo. We follow the same approach as in Beasley et al. (2016). In that work, the authors compare the cumulative mass distribution from the EAGLE simulation (Schaller et al. 2015) with the observed dynamical mass of their galaxy within a given radius (see their Figure 4, right panel). From that comparison they infer a virial mass for the dark matter halo. Using our measurement of the enclosed mass ∼4.6(±0.8) × 10⁹ ${M}_{\odot }$ within the inner R ∼ 5 kpc, we estimate a virial ${M}_{200}$ mass similar to that found by Beasley et al. (2016) for their galaxy (i.e., ∼8(±2) × 10¹⁰ M_⊙).

To obtain the structural properties of UGC 2162 we have used the code IMFIT (Erwin 2015). The surface brightness distribution of the galaxy in each band was modeled using a single Sérsic component. The Sérsic model was convolved with the PSF of the image. The IAC Stripe82 Legacy Survey provides, for each band, a PSF representative of the local (0.5 × 05) conditions of the image. To have a first estimate of the spatial coordinates of the source, the position angle and the effective radius, we use SExtractor. These values are used later as input parameters for IMFIT. In addition, we mask the closest sources surrounding our galaxy (see Figure 1).

We derived the structural parameters of the galaxy in the g, r, and i bands. In all of these bands, the structural parameters of the galaxy were very similar. We obtained ${R}_{e}$ = 28 arcsec (which is equivalent to 1.7(±0.2) kpc). The central surface brightnesses were ${\mu }_{g}(0)$ = 24.4 ± 0.1 mag arcsec⁻², ${\mu }_{r}(0)$ = 24.2 ± 0.1 mag arcsec⁻², and ${\mu }_{i}(0)$ = 24.1 ± 0.1 mag arcsec⁻². These values have been corrected for Galactic reddening (0.117, 0.081, and 0.060 in the g, r, and i bands respectively; Schlafly & Finkbeiner 2011). The Sérsic index in all the bands was around n = 0.9. The total apparent magnitudes were g = 16.1 mag, r = 15.9 mag, and i = 15.8 mag.

Using the global color of the galaxy and its absolute magnitude, we can have a rough estimate of its stellar mass. We follow the recipe by Roediger & Courteau (2015; assuming a Chabrier IMF), using the $g-i$ color and the absolute magnitude in the r band (${M}_{r}$ = −14.6(±0.3) mag). We obtain a stellar mass of $\sim 2{(}_{-1}^{+2})$× 10⁷ M_⊙.

UGC 2162 has an SDSS spectrum (Plate = 1070; Fiber = 450; MJD = 52591) located at the coordinates: R.A. = 4009751 and decl. = 122476 (see Figure 2). The spatial location of the SDSS spectrum is indicated in Figure 1. This region corresponds to the brightest knot of star formation of the galaxy. This knot has a radius of 1.2 arcsec (Petrosian mag at 90%) and a magnitude in the r band of only 21.9 mag. Using the ratio N2 ≡ [N ii]λ6583/Hα from the SDSS spectrum and the calibration by Pettini & Pagel (2004), we have estimated an oxygen abundance for the star-forming ionized gas 12+log(O/H) = 8.22 ± 0.07, which corresponds to one-third of the solar abundance. The gas of UGC 2162 is fairly metallic for its mass and magnitude, since the galaxy is a high-metallicity outlier of the mass–metallicity relation and the magnitude–metallicity relation worked out by Berg et al. (2012). We also checked the spectrum for the presence of [O iii]λ4363, which appears in low metallicity objects (e.g., Sánchez Almeida et al. 2017). The line is not in the spectrum, which is consistent with the moderate metallicity inferred from N2. Since the SDSS spectrum is quite noisy, the estimated O abundance should be regarded as an upper limit.

Zoom In Zoom Out Reset image size

Using the Hα flux (uncorrected for extinction since Hβ does not seem to be reddened), the distance to the source, and the recipe of Kennicutt (1998), we have estimated the SFR and the surface SFR of the bright knot, which turn out to be SFR = 8.7(±1.2) × 10⁻⁵ M_⊙ yr⁻¹ and ${{\rm{\Sigma }}}_{\mathrm{SFR}}$ = 3.4(±0.5) × 10⁻³ M_⊙ yr⁻¹ kpc⁻². Assuming that the galaxy has 100 such star-forming knots (reasonable in view of the shape and size of the galaxy) the total SFR of the galaxy would be SFR = 8.7(±1.2) × 10⁻³ M_⊙ yr⁻¹. This value is consistent with the value around 10⁻² M_⊙ yr⁻¹ worked out by Hunter & Elmegreen (2004) for this object using Hα imaging. Using the above value, we can derive a specific SFR for UGC 2162: sSFR ∼ 5(±0.7) × 10⁻¹⁰ yr⁻¹. If the SFR of UGC 2162 were constant, the galaxy would double its stellar mass in ∼2 Gyr.

UGC 2162 is currently located at (a projected separation of) ∼300 kpc from M77. Due to its large amount of HI gas (∼10 times larger than its stellar mass), we can speculate that UGC 2162 is undergoing its first infall to the M77 galaxy group. When a galaxy like UGC 2162 falls into a group environment, it suffers a number of physical mechanisms that eventually will quench its star formation. These mechanisms can be either slow (of the order of a few gigayears) due to gas strangulation (see, e.g., McCarthy et al. 2008) or they can be relatively rapid if they are produced by ram pressure stripping (e.g., Wang et al. 2007). Following recent simulations (Yozin & Bekki 2015b), one can assume that gas rich dwarf galaxies will be depleted of gas 6 Gyr after their first infall into typical groups of galaxies (10^13–13.5 M_⊙). Motivated by this number, we simulate how our galaxy would look in 6 Gyr time if the object were to stop forming stars and followed a passive evolution. Naturally, this is an oversimplification of the actual evolution of the galaxy, but it can be an interesting exercise to understand how our object would look in the future.

To model the color and structural evolution of UGC 2162, we have used its present-day $g-r$ color map and its current g, r, and i surface brightness distributions. Then, we have estimated how every pixel of the images would look if we make their colors evolve passively for 6 Gyr. To quantify the color change and the dimming in surface brightness of every pixel, we have used the Vazdekis et al. (2015) models assuming a Kroupa IMF. Due to this passive evolution, the galaxy would not only change its global color (becoming $g-i$ = 0.77) but would also get dimmer (by ∼2.6 mag arcsec⁻²). The result of this evolution is illustrated in Figure 3. After 6 Gyr of passive evolution, the galaxy would have ${\mu }_{g}(0)$ = 27 ± 0.1 mag arcsec⁻² and ${R}_{e}$ = 1.8(±0.2) kpc. Its profile shape would not change dramatically, and its passively evolved Sérsic index $n\,\sim 0.8$ would be similar to its original value. With these characteristics, the galaxy would resemble closely the faintest UDG galaxies discovered so far in the Virgo cluster (Mihos et al. 2015). In fact, considering the virial mass of UGC 2162, in an eventual future, this galaxy could look very similar to VCC 1287, a very low-surface brightness UDG galaxy (${\mu }_{g}(0)$ = 26.7 mag arcsec⁻², ${R}_{e}$ = 2.4 kpc, $g-i$ = 0.83, ${M}_{\star }\,\sim$ 3 × 10⁷ ${M}_{\odot }$) inhabiting the Virgo cluster (Beasley et al. 2016).

Zoom In Zoom Out Reset image size

The structural and stellar population properties of UGC 2162 seem to fit very well within a scheme where the UDGs found both outside and inside clusters are just different evolutionary stages of the same type of objects (Román & Trujillo 2016b). In this view, UDGs outside clusters would be simply the progenitors of the redder UDGs found in rich clusters. The link between both types of UDGs would be an evolution due to the removal of their gas produced by the infall of these galaxies in rich environments. If this picture is correct UDGs outside dense environments should be dwarf galaxies with a large HI content. In fact, low-surface brightness dwarf galaxies with a large amount of HI seem to be fairly common in the field (e.g., Skillman et al. 2013; James et al. 2015; Hirschauer et al. 2016; Sanchez Almeida et al. 2017). On the contrary, those UDGs found in the richest environments should be depleted of HI gas due to the removal of this component.

UGC 2162 also teaches us an important lesson about UDGs and how these objects are observationally selected. This galaxy has a relatively low stellar mass (∼10⁷ ${M}_{\odot }$) compared to the general population of UDGs, which peaks at ∼10⁸ ${M}_{\odot }$ (e.g., Román & Trujillo 2016a). This issue is connected to the way UDGs are defined. UDGs are observationally selected using their surface brightness ($\mu (0)\,\gt$ 24 mag arcsec⁻²) and size (${R}_{e}\,\gt$ 1.5 kpc). This observational definition immediately biases the selection of the galaxies depending on their stellar populations. The redder UDGs will be more massive and older, whereas the bluer UDGs will be younger and with lower stellar mass. This implies that if one wants to connect the observed UDGs with their progenitors or with their descendants, it is important to take this into account. For instance, most of the progenitors of massive and red UDGs found in rich clusters would not satisfy the observational criteria to be classified as UDGs. These progenitors would have central surface brightnesses brighter than $\mu (0)=24$ mag arcsec⁻² and would have been classified as regular (blue) dwarf galaxies. On the other hand, those blue UDGs that have been discovered outside clusters (as is the case of UGC 2162) would evolve into the less massive (and red) UDGs found in rich clusters (as is the case of VCC 1287). Accounting for this selection effect in selecting UDGs is key to have a comprehensive picture about the nature of these objects and how to connect their different evolutionary stages.

UGC 2162 has many of the structural and stellar population properties expected if the large size of this galaxy is the result of feedback-driven gas outflows (Di Cintio et al. 2017). For instance, UGC 2162 has a large amount of HI gas and it is currently forming stars as the cosmological simulations predicted. In addition, Di Cintio et al. (2017) simulations are able to predict the dwarf-like halo mass of this galaxy, as well as its stellar mass, gas mass, Sérsic index, effective radius, absolute magnitude, SFR, irregular appearance, and off-center-star-formation episodes. According to these simulations, these galaxies would not be at all rare and they will be found in abundance outside clusters. In fact, many of these have already been detected (Román & Trujillo 2016a) in low density environments. If the picture sketched by the cosmological simulations is correct, a large number of the descendants of UDGs found in low density environments would be found in rich clusters having the following characteristics: ${M}_{\star }\,\gtrsim$ 10⁷ ${M}_{\odot }$, ${\mu }_{g}(0)\,\gtrsim$ 27 mag arcsec⁻², ${R}_{e}\,\gt$ 1.5 kpc, $n\,\lesssim$ 1, $g-i$ ∼ 0.8 and low HI gas content. These objects would be hard to find even for current deep surveys. In fact, UDGs with ${\mu }_{g}(0)\,\gtrsim$ 27 mag arcsec⁻² have only be reported by Mihos et al. (2015), while the remaining UDGs found in rich clusters have all been found with ${\mu }_{g}(0)\,\lt$ 27 mag arcsec⁻² (e.g., Koda et al. 2015; Muñoz et al. 2015; van Dokkum et al. 2015; Román & Trujillo 2016a; van der Burg et al. 2016). The existence of a large number of "dark" ${M}_{\star }\,\sim$ 10⁷ ${M}_{\odot }$ extended galaxies in rich clusters is a natural prediction of cosmological simulations if the above evolutionary picture for the UDGs is correct.

Finally, it is worth noting that there are a number of extremely low-surface brightness galaxies at a distance closer than UGC 2162 that technically satisfy the criteria to be considered UDGs (i.e., ${R}_{e}\,\gt$ 1.5 kpc and $\mu (0)\,\gt$ 24 mag arcsec⁻²). These objects are: (a) a satellite of M31, Andromeda XIX (${R}_{e}\,\gt$ 1.7 kpc and $\mu (0)$ = 29.3 mag arcsec⁻²; McConnachie et al. 2008; Martin et al. 2016); (b) a satellite of our own galaxy, the Sagittarius dwarf (${R}_{e}$ = 1.6 kpc and $\mu (0)$ = 25.2 mag arcsec⁻²; Ibata et al. 1994; Majewski et al. 2003); (c) and a satelitte of the galaxy NGC 4449 located at 3.8 Mpc, NGC 4449B (${R}_{e}$ = 2.7 kpc and $\mu (0)$ = 25.5 mag arcsec⁻²; Martínez-Delgado et al. 2012; Rich et al. 2012). Andromeda XIX has a very low mass (${M}_{V}$ = −9.3 and σ = 4.7 km s⁻¹; Collins et al. 2013) and it is significantly less massive than the population of UDGs that has been discussed in the literature (which are a factor of ∼100 times more massive). For this reason, And XIX cannot be considered  to be representative of the population of UDGs originally discovered in the Coma Cluster. In fact, at that distance, the object would appear invisible in present-day surveys. Martin et al. (2016) discuss the possibility that the And XIX large extention could be produced by the gravitational tides of M31. The other two galaxies (though with stellar masses around 10⁷ ${M}_{\odot }$ or larger) are being tidally disrupted. In this sense, their large effective radii are a consequence of the ongoing disruption process. Compared to the previous objects, the fact that UGC 2162 has a gas reservoir is strong evidence that it is not diffuse and extended because it is being tidally disrupted. For this reason, UGC 2162 is currently the nearest not tidally disrupted UDG known, whose large size is probably due to an internal origin alone.

We thank the referee for a report. We would also like to thank Michelle Collins for interesting insights into the population of extremely diffuse Local Group galaxies. We thank Michael Beasley and Chris Brook for their useful comments during the development of this work. The authors of this paper acknowledge support from grant AYA2013-48226-C3-1-P from the Spanish Ministry of Economy and Competitiveness (MINECO). J.R. thanks the Spanish Ministry of Economy and Competitiveness (MINECO) for financing his PhD through an FPI grant.