Solo dwarfs II: the stellar structure of isolated Local Group dwarf galaxies

ABSTRACT

The Solo (Solitary Local) Dwarf Galaxy survey is a volume-limited, wide-field g- and i-band survey of all known nearby (<3 Mpc) and isolated (>300 kpc from the Milky Way or M31) dwarf galaxies. This set of 44 dwarfs is homogeneously analysed for quantitative comparisons to the satellite dwarf populations of the Milky Way and M31. In this paper, an analysis of the 12 closest Solo dwarf galaxies accessible from the Northern hemisphere is presented, including derivation of their distances, spatial distributions, morphology, and extended structures, including their inner integrated light properties and their outer resolved star distributions. All 12 galaxies are found to be reasonably well described by two-dimensional Sérsic functions, although UGC 4879 in particular shows tentative evidence of two distinct components. No prominent extended stellar substructures, which could be signs of either faint satellites or recent mergers, are identified in the outer regions of any of the systems examined.

1 INTRODUCTION

Dwarf galaxies are a fascinating and challenging regime in which to study galaxy evolution. Both internal and external processes dramatically impact the morphology and evolution of dwarfs as a consequence of their low masses and shallow gravitational potential wells. Star formation and the resulting feedback (e.g. El-Badry et al. 2018), ram-pressure stripping (e.g. Gunn & Gott 1972; Grebel, Gallagher & Harbeck 2003), tidal stripping (e.g. Fattahi et al. 2018), tidal stirring (e.g. Mayer et al. 2001; Kazantzidis et al. 2011), reionization (e.g. Ordoñez & Sarajedini 2015; Wheeler et al. 2019), mergers, and interactions (e.g. Deason, Wetzel & Garrison-Kimmel 2014) are all processes that can significantly alter a dwarf’s structure, morphology, and stellar populations. The observational challenge then becomes understanding the complex interplay between initial formation processes and the internal and external processes that shape and reshape dwarfs.

Dwarf galaxies are the most numerous class of galaxy in the Universe and the population of known nearby dwarfs has increased dramatically in recent years (e.g. numerous dwarfs discovered in the Dark Energy Survey, Bechtol et al. 2015; Drlica-Wagner et al. 2015; Crnojević et al. 2019; Drlica-Wagner et al. 2020; Crater III and Bootes IV from the Hyper Suprime-Cam Subaru Strategic Program, Homma et al. 2018, 2019; Sagittarius II, Draco II, and Laevens 3 from the Pan-STARRS 1 3π Survey, Laevens et al. 2015, among others). The parameter space that dwarfs are known to inhabit has been broadened by these discoveries and others (e.g. Crater 2 and Antlia 2; Torrealba et al. 2016, 2019). New terminology such as ultra-faint and ultra-diffuse is now in common usage (Simon 2019). Many dwarf galaxies are so faint that the only detailed information obtainable for them comes through their resolved stars, with the result that the dwarfs in the nearby Universe are a critical observational sample for astronomers.

Within the Local Group (LG), the satellite dwarfs of the Milky Way (e.g. the MegaCam Survey of Outer Halo Satellites; Marchi-Lasch et al. 2019) and those around M31 [e.g. the Pan-Andromeda Archaeological Survey (PAndAS); McConnachie et al. 2009; Martin et al. 2016] are generally well studied. Typically, these ground-based, resolved stellar studies are limited to the LG and work in concert with the extensive number of surveys and studies at larger distances (e.g. M101: Müller et al. 2017; Karachentsev & Makarova 2019; Cen A: Crnojević et al. 2019; Müller et al. 2019; NGC 1291: Byun et al. 2020), in other environments (e.g. Next Generation Virgo Cluster Survery ‘NGVS’: Ferrarese et al. 2012, 2020 and Fornax: Venhola et al. 2018), and which question the ‘stereotypical’ nature of the LG (e.g. the Satellites Around Galactic Analogs ‘SAGA’ Survey: Geha et al. 2017; Carlsten et al. 2020; Mao et al. 2020).

The importance of environment on dwarf galaxy evolution is well known, most famously the presence of increased star formation with distance from a large companion (Einasto et al. 1974; also see McConnachie 2012, and references therein). However, while it is clear that dwarfs around a large host galaxy are affected by their companion, the interplay of these effects remains relatively poorly understood and quantified (Geha et al. 2012; Spekkens et al. 2014). The study of dwarfs well separated from large galaxies, where the current impact of the host is likely to be minimal, offers an interesting opportunity. Even here though, we must be aware that some of these dwarfs may be ‘backsplash’ dwarfs (Buck et al. 2019, and references therein), which could have had a close pericentric passage with a massive galaxy some point in their history. The isolated dwarfs in the LG present a unique opportunity to study the lowest mass galaxies in great detail, largely in the absence of strong environmental influences.

Of course, isolated LG galaxies have been studied for many years. This includes individual dedicated studies for some of the most prominent members, e.g. WLM (Wolf 1909; Melotte 1926; Leaman et al. 2012), IC 1613 (Hodge et al. 1991; Pucha et al. 2019), and DDO 210 (McConnachie et al. 2006), as part of catalogues like Mateo (1998), Karachentsev, Makarov & Kaisina (2013), van den Bergh (1999), and McConnachie (2012), as inclusion in H i Surveys like LITTLE THINGS (Hunter et al. 2012) or FIGGS2 (Patra et al. 2016), as part of LG-wide studies like the star formation history analysis of Weisz et al. (2014), and as targets to understand their stellar dynamics (e.g. Kirby et al. 2014; Kacharov et al. 2017, among numerous other works). However, homogeneous, systematic studies focused on this population are generally lacking but with exceptions, such as LG Galaxies Survey Project (Massey et al. 2006, 2007).

Given the limited sample of galaxies available, homogeneous analyses are important in order to minimize systematic errors in what is, fundamentally, a relatively small statistical sample. It is with this intent that we began Solo, the Solitary Local Dwarf Galaxy Survey.

Solo is a volume-limited sample of all known, nearby, and isolated dwarf galaxies, designed to study the secular evolution of dwarfs. A general introduction to the Solo Survey is given in Higgs et al. (2016). Briefly, all 44 dwarfs that are known within 3 Mpc, and which are located more than 300 kpc from the Milky Way or M31 are included. In this distance range, at least the outer parts of these dwarfs can usually be resolved into stars using ground-based facilities. While the adoption of 300 kpc as a threshold for isolation is somewhat arbitrary (although it does correspond to the approximate expected halo virial radius for these galaxies; e.g. Klypin, Zhao & Somerville 2002; Posti & Helmi 2019), work by Geha et al. (2012) and Spekkens et al. (2014) suggests that it is around this distance that there is a notable change in dwarf morphology (see also Einasto et al. 1974). While there is no doubt that backsplash galaxies will be in the sample (e.g. Buck et al. 2019 suggest both IC 1613 and And XXVIII are likely backsplash systems), it is clear that most of the dwarfs in the Solo sample have largely evolved in isolation, and are the best nearby systems to use to explore the faint end of the galaxy luminosity function away from the influence of massive galaxies.

In comparison to LG Galaxies Survey Project, Massey et al. (2006, 2007) focused on detailed analysis of the colour–magnitude diagrams (CMDs) using UVBRI photometry of some of the more massive Local Volume dwarfs (M31, M33, NGC 6822, WLM, IC 10, Phoenix, Pegasus, Sextans A, and Sextans B). In contrast, Solo has a larger field of view by a factor of ∼2 and deeper photometry by ∼2–3 mag, but in only two filters. Our focus is largely on the extended structure of these galaxies rather than their detailed stellar populations, as in Massey et al. (2006, 2007).

The Solo Survey is designed to probe several questions pertaining to the role of environment in the evolution of the lowest mass galaxies. In particular, the intent is to provide observational data and parameters of a similar quality to those that exist for the Milky Way and M31 satellite populations. This particular contribution is focused towards obtaining updated wide-field structural parameters in a similar way to the corresponding M31 and Milky Way data for galaxies in a similar luminosity range, i.e. by analysis of the distribution of the oldest stellar populations, primarily the red giant stars. This is not an insignificant point, given that many of the isolated LG galaxies have their light dominated by young stellar populations, which are invariably more centrally concentrated than other stellar populations. Weisz et al. (2014) present star formation histories showing substantial recent star formation in many of these systems, and studies of radial gradients in the stellar populations of dwarf galaxies invariably show that the younger populations are more concentrated (e.g. Harbeck et al. 2001; Tolstoy et al. 2004; Bernard et al. 2008; Mercado et al. 2021, among others). Thus, in determining the physical extent of the galaxy, studies that are based solely on integrated light or young stars will generally find a smaller size for these galaxies than studies based on the distribution of older stellar populations (as is nearly always the case for studies of Milky Way and M31 satellites). Minimizing these systematic differences between observations of satellites and isolated galaxies is a primary purpose of this study.

This paper is structured as follows. Section 2 describes the survey, including the observations, data processing, calibrations, and a brief introduction to the dwarfs analysed. Section 3 analyses the integrated light in the central regions of the target galaxies. Section 4 describes the analysis of the resolved stellar components, including distance estimates and shape analysis. Section 5 derives radial profiles from a combined fit of the integrated light and resolved stars, and presents the associated parameters describing the dwarf galaxies. Section 6 compares our results to the literature, and Section 7 concludes.

2 OBSERVATIONS AND DATA PROCESSING

2.1 Observations

The Solo Survey is a volume-limited, wide-field imaging survey of all known dwarf galaxies that are within 3 Mpc and more than 300 kpc from either M31 or the Milky Way. Galaxies are observed with either Canada–France–Hawaii Telescope (CFHT)/Megacam in the Northern hemisphere or Magellan/Megacam or IMACS in the Southern hemisphere. Some targets are observed with multiple instruments for calibration purposes. The total survey area per galaxy is approximately 1 square deg, regardless of telescope/instrument (for Magellan, multiple pointings are used to cover this area, whereas only a single pointing is required for CFHT). All dwarfs were observed in the g and i bands. Additionally, almost half (21 out of 44 dwarfs) were also observed in the u band. Some of the u-band observations were taken as part of the Canada France Imaging Survey (CFIS; Ibata et al. 2017) or are from archival data.

2.2 The LG subset

While the full Solo Survey contains 44 dwarfs, this paper explores an LG subset, consisting of 12 dwarfs within the zero velocity surface of the LG (≈1 Mpc) and visible from the Northern hemisphere. Table 1 lists these 12 galaxies (hereafter the ‘LG subset’) and gives details of the observations relevant to this work. Other dwarfs within the zero velocity surface but observed for Solo only from the Southern hemisphere are Cetus, Eri II, Phoenix, and Tucana. A full list of the Solo targets and observational details can be found in Higgs et al. (2016) (hereafter Paper I).

Fig. 1 shows galactocentric radial velocity against galactocentric distance for dwarfs in and around the LG. Our transformation from the heliocentric values listed in McConnachie (2012) assumes (R_⊙, V_c) = (8.122 kpc, 229 km s⁻¹) and (U_⊙, V_⊙, W_⊙) = (11.1, 12.24, 7.25) km s⁻¹ (Schönrich, Binney & Dehnen 2010; Gravity Collaboration 2018). Within this plot, the LG subset can be defined as those whose motion is not affected by the Hubble flow. In contrast, those galaxies further away are clearly affected by the expansion of the Universe. The boundary, the so-called zero-velocity surface, is at approximately 1–1.3 Mpc from the Milky Way (about 1 Mpc from the centre of the LG; McConnachie 2012). In Fig. 1, the M31 satellite population stands out as the distinct cluster of galaxies around 800 kpc (McConnachie et al. 2005).

Fig. 2 shows colour composite images for six of the dwarfs in our sample, and qualitatively demonstrates the diversity of the LG subset. Striking differences in size, stellar density, number of bright stars, colour, and density of MW foreground can clearly be seen by eye. The dwarfs range from low-mass, faint systems like Perseus to large, obviously star-forming galaxies like WLM. Some dwarfs – like WLM – are well known and have been well studied (e.g. Leaman et al. 2012, 2013, among others), while others in our sample, like And XXVIII, are less so.

2.3 Data processing

An extensive and detailed description of the data processing is given in Paper I. Preliminary image processing is similar to Richardson et al. (2011). The pre-processing (de-biasing, flat-fielding, and fringe corrections) is completed using CFHT’s Elixir system prior to transferring the data to the Cambridge Astronomical Survey Unit. Here, the data are reduced using the adapted pipeline originally developed for the Wide Field Camera observations from the Isaac Newton Telescope (INT; see Irwin 1985, 1997; Irwin & Lewis 2001; Irwin et al. 2004, for details).

Catalogues of stellar sources, matched in each band (g and i with some additional u band), are generated for each dwarf, although in what follows, we focus exclusively on g and i observations. Before stacking all exposures, catalogues are generated for each science image, which determines the data quality and refines the astrometric calibration. For details, see Paper I and references therein. Each science exposure is assessed based on the average seeing, ellipticity of point sources, sky level, and sky noise. The images are stacked using the World Coordinate System (WCS), and the source catalogues generated for each exposure to align the images with a sub-pixel precision. Overlapping background regions are used to account for sky variations between exposures in the stack. Images in the final stack were weighted in accordance to the average seeing. Cosmic rays were removed. Source catalogues are generated from these full stacks and the WCS astrometry is re-derived. The sources are classified using their morphology as stellar or non-stellar and a final catalogue including both bands is produced. If two sources are identified in the different bands and lie within 1 arcsec of each other, they are considered the same object, and the position for the higher signal to noise object is retained. Any object that is detected in only one band is retained as the catalogues are merged. For some dwarfs, seeing for one band (generally i) was significantly better than the other band. In this situation, sources detected in the better band were solely used to generate the final catalogue using forced photometry for the lower quality band.

2.3.1 Photometric corrections and calibrations

Only stellar sources are used for the subsequent analysis, using the morphological classification procedures described in Irwin (1985, 1997). Sources within 2σ of the stellar locus in both bands are considered stellar. From the source catalogues, the positions of the sources are de-projected, converting right ascension (RA) and declination (Dec.) to standard plane coordinates (ξ, η). The standard plane is centred on the dwarf’s coordinates as listed in McConnachie (2012).

Fig. 3 shows photometric errors as a function of magnitude in g and i for a sample of dwarfs. These uncertainties are derived from the square root of the variance of the weighted flux measurements. They are generally less than 0.1 mag at (g ∼ 24.5, i ∼ 24.0), meaning red giant branch (RGB) stars in the dwarfs are easily detectable in all cases.

WLM was observed in non-photometric conditions for the i band. As such, a correction was required and was derived from INT Wide Field Camera observations of WLM, DDO 210, and Peg DIG (published in McConnachie et al. 2005). RGB stars in common between the two data sets for DDO 210 and Peg DIG were used to define the appropriate transformation between the two filter systems for photometric conditions. Applied to the INT data for WLM, this transformation then allowed us to determine the appropriate correction to convert the non-photometric MegaCam data to the correct photometric system.

2.4 Summary of target galaxies

Here, we briefly summarize some of the notable properties of each of our target dwarfs, and highlight the pertinent literature as it relates to their structural properties. This background is not intended to be an exhaustive summary of all previous work, but rather a useful and easy reference for the interested reader.

2.4.1 Andromeda XVIII

And XVIII is in the vicinity of M31, and was discovered by McConnachie et al. (2008) as part of precursor observations for the PAndAS (McConnachie et al. 2009). Unfortunately, in these discovery observations, the majority of the galaxy fell in one of the large chip gaps on the detector, obscuring the central region. The dwarf’s parameters were estimated under the assumption of circular symmetry to address the obscured portion. Martin et al. (2016) analysed the dwarf as part of a wider study of M31’s dwarfs; however, they also used the same data. They estimated its structural parameters, and included a more robust surface brightness estimate; however, its ellipticity was poorly constrained. Here, new observations, placing the dwarf well away from chip gaps, have been obtained that allow for a complete picture of its global structure.

Tollerud et al. (2012) obtained kinematics of And XVIII stars as part of the SPLASH survey. They concluded that it is a kinematically cold, dark matter-dominated dwarf. Given its distance from M31 (of the order of 600 kpc distant), it is unlikely to have interacted with this massive galaxy, but Tollerud et al. (2012) noted that there is a remarkable agreement between their systematic velocities.

2.4.2 Andromeda XXVIII

Slater, Bell & Martin (2011) determined the structural parameters for And XXVIII from Sloan Digital Sky Survey (SDSS) observations using the maximum likelihood methods described in Martin, de Jong & Rix (2008) (our new observations are approximately 3 mag deeper than the SDSS observations). Slater et al. (2015) followed up with Gemini/GMOS and Keck/DEIMOS observations, finding no signs of recent star formation but the presence of a red clump and other features suggested an extended period of star formation. Buck et al. (2019) suggest that this dwarf has a highly likelihood (91 per cent) of having interacted with the MW (i.e. it could be a backsplash galaxy). This estimate is based on the comparison of observed galaxy properties (distance, radial velocity, and velocity dispersion) with simulated satellite and field dwarfs in the NIHAO (Numerical Investigation of a Hundred Astrophysical Objects) simulations (Wang et al. 2015).

2.4.3 The Pegasus Dwarf Irregular Galaxy

Peg DIG (also known as DDO 216 and UGC 12613) is a transition-type dwarf that is host to a large, central globular cluster (Cole et al. 2017; Leaman et al. 2020), making it among the lowest mass dwarfs in the LG with a globular cluster (in addition to Eri II; see Zoutendijk et al. 2020, and references therein). A notable population of RR Lyrae stars in the cluster has been used to derive a distance of 24.77 ± 0.08 (Cole et al. 2017).

McConnachie, Peñarrubia & Navarro (2007) found an asymmetry between the gas and stellar components indicative of ram-pressure stripping removing the gas from the system. However, more recent H i observations by Kniazev et al. (2009) contradict this conclusion, finding minimal evidence of interactions. These authors concluded that this dwarf has only recently joined the LG. The structural parameters derived by Kniazev et al. (2009) provided a good comparison to this work; however, the SDSS observations used are significantly (>3 mag) shallower than the Solo data. They found the dwarf is more extended than previously thought and noted possible stellar extensions on the north-west end. In their H i observations, Kniazev et al. (2009) observed solid body-like rotation. Kirby et al. (2014) found rotation in the RGB stars as well.

2.4.4 Leo T

Leo T is one of lowest mass and closest dwarfs in Solo. It is notable for being one of the few very faint dwarfs with a young stellar population, and like Peg DIG, it is often described as a transition-type dwarf, given that its other characteristics are generally typical of dwarf spheroidal galaxies. de Jong et al. (2008) determined its structural parameters based on deep, wide-field Large Binocular Telescope data, and it is also included in the MegaCam Survey of Outer Halo Satellites (Muñoz et al. 2018a, b), using CFHT gr photometry. Its H i structure was most recently studied by Adams & Oosterloo (2018). Teyssier, Johnston & Kuhlen (2012) argued that both Leo T and NGC 6822 have likely had a close pericentric passage with the Milky Way, whereas Buck et al. (2019) suggested that there is only an approximately 50 per cent chance of Leo T being a backsplash galaxy.

2.4.5 Leo A

Leo A (DDO 69) shows delayed star formation, like DDO 210, and only minimal star formation happened before 10 Gyr ago. As such, it is sometimes argued as being a possible survivor of reionization (Cole et al. 2007, 2014). Stonkutė et al. (2014, 2018) presented detailed studies of Leo A based on the Subaru/SuprimeCam data, in addition to using HST ACS fields at large radius. Stonkutė et al. (2015) reported the discovery of a young star cluster in Leo A. van Zee, Skillman & Haynes (2006) studied the H ii regions in Leo A, confirming that its interstellar medium has one of the lowest metallicities of nearby galaxies [van Zee et al. (2006) find 12 + log(O/H) = 7.30 ± 0.05]. This low metallicity makes it a useful system for calibration (e.g. for Cepheid observations; Bernard et al. 2013). Leo A was also used by Ruiz-Lara et al. (2018) to test whether there are differences between star formation histories derived from CMDs compared to integrated light studies.

2.4.6 Perseus

Perseus I, also known as And XXXIII, was discovered by Martin et al. (2013) in the Pan-STARRS1 3π survey and followed up Martin et al. (2014) with spectroscopy. The structural parameters derived are based on the maximum likelihood method (Martin et al. 2008). The Solo observations are about 2 mag deeper than the PanSTARRS data. The location of this dwarf makes it an interesting object with respect to the ‘planes of satellites’ around the LG (see Pawlowski & McGaugh 2014, for a discussion of Perseus in this context).

2.4.7 UGC 4879

UGC 4879 (also known as VV 124) was first studied in detail by Kopylov et al. (2008). Bellazzini et al. (2011a, b) performed a comprehensive analysis using HST ACS, the Large Binocular Telescope, and H i observations, and identify two young star clusters. Kopylov et al. (2008) and Jacobs et al. (2011) noted that it is a very isolated dwarf, and so likely not a backsplash system. Its structure is somewhat unusual, in so far as there are two ‘wings’ on either side of an otherwise elliptical system, perhaps suggestive of a stellar disc. Kirby, Cohen & Bellazzini (2012) placed an upper limit on the possible stellar disc rotation of 8.6 km s⁻¹.

2.4.8 DDO 210

DDO 210 (also known as Aquarius) is classified as a transition type dwarf in that there is no current star formation observed but, prior to the discovery of Leo T, it was the faintest known dwarf with a significant gaseous component. Along with Leo A, DDO 210 has a delayed star formation history (Cole et al. 2014), meaning that most of the stars are of intermediate age. Cole et al. (2014) noted a strong blue horizontal branch population that is unusual relative to similar dwarfs (like Leo A), as well as the presence of a red clump and a red bump. Its free-fall time is longer than a Hubble time; hence, it is likely on its first infall into the LG (McConnachie 2012). It is well separated from neighbouring galaxies.

The H i and stellar components show some offset and differences in spatial structure (McConnachie et al. 2006). Hermosa Muñoz et al. (2020) looked at the metallicity and kinematics of stars, finding significant misalignment of the H i and stellar components. They suggested that this may be due to recent H i accretion or re-accretion of gas after feedback from star formation.

2.4.9 The Sagittarius Dwarf Irregular Galaxy

Sag DIG was studied as the first Solo dwarf in Paper I. We previously noted differences between the H i structure and its stellar structure, similar to DDO 210 (Aquarius). There are faint extended RGB stars regions along the semimajor axis (SMA) of this dwarf. Qualitatively, these extensions are not found to have an H i counterpart in LITTLE THINGS observations (Hunter et al. 2012). Sag DIG lies at a low galactic latitude, so the foreground contamination from the MW in the field is significant.

2.4.10 Wolf–Lundmark–Melotte

WLM, also known as DDO 221, is one of the more massive dwarfs in the Solo Survey. It has been extensively studied, from its discovery over 100 yr ago (Wolf 1909) to recent work presented in Albers et al. (2019). Several spectroscopic studies have been conducted. For example, RGB stars have been targeted by Leaman et al. (2009, 2012) and blue supergiants have been targeted by Venn et al. (2003), Bresolin et al. (2006), and Urbaneja et al. (2008). Leaman et al. (2012) concluded that this dwarf is a highly oblate spheroid. They found good agreement, kinematically and photometrically, between the position angle (PA) and ellipticity of the RGB stars, young carbon stars, and H i gas. They concluded that external environmental effects have not had a dominant role in shaping the structure of WLM. Using HST observations, Albers et al. (2019) derived a star formation history and found relatively continuous star formation, with no dominant early episode (Gallart et al. 2015 classified the star formation history of WLM as ‘slow’). This dwarf is one of the few with a known globular cluster, in this case a single large globular cluster located just outside the central region.

2.4.11 NGC 6822

NGC 6822 is one of the more massive dwarfs in Solo and has been extensively studied. Demers, Battinelli & Kunkel (2006) identified this dwarf as a polar ring galaxy, based on kinematics of H i compared to asymptotic giant branch (AGB) stars. This claim is disputed by Thompson, Ryan & Sibbons (2016), who found that the axis of rotation for the stars and H i are largely aligned. Battinelli, Demers & Kunkel (2006) derived the stellar structure from wide-field CFHT observations, and noted that it has a bright central bar. NGC 6822 hosts a population of globular clusters (Huxor et al. 2013; Veljanoski et al. 2015). Hwang et al. (2014) used extended star clusters as kinematic tracers and derived a mass-to-light ratio of |$75^{+45}_{-1}(M/L)_{\odot }$|⁠, making NGC 6822 dark matter dominated. Teyssier et al. (2012) suggested that NGC 6822 is possibly a backsplash galaxy, based on its current observed position and radial velocity. In contrast, McConnachie et al. (2021) find that NGC 6822 is highly likely to be isolated and on its first infall, based on orbital properties derived using Gaia Data Release 2 proper motions of the brightest stars (Gaia Collaboration 2018).

2.4.12 IC 1613

IC 1613 is another massive dwarf in Solo and has been well studied, including by Bernard et al. (2007) and recently by Pucha et al. (2019). Pucha et al. (2019) found strong age gradients, and the young stellar population has a considerably smaller scale radius than the older stars. The star formation history they determined is consistent with this picture, with a decreasing star formation rate with increasing radius. They found that the young population is off centre, and the star-forming regions are distributed in a ring, coincident with areas of high H i column densities. Buck et al. (2019) suggest that this dwarf, along with And XXVIII, has a highly likelihood of having previously interacted with the Milky Way.

3 CHARACTERIZING SOLO DWARFS USING INTEGRATED LIGHT

The Solo Survey selects dwarfs such that at least their outer regions can be resolved into stars from the ground in reasonable seeing. In the central parts of most of these dwarfs, crowding becomes a significant issue and the stellar catalogues become increasingly incomplete. As a result, these central regions are better analysed using the integrated light rather than resolved stars, particularly with respect to their radial profiles in both bands (and which allows identification of any colour trends present).

The integrated light profile is determined in both the g and i bands. Here, elliptical annuli with the median PA and ellipticity as derived from the resolved RGB stars (see the next section) are used. All bright sources and their surrounding pixels are masked to reduce contamination from other sources, including saturated objects and diffraction spikes. Fig. 4 shows an example of an i-band image and the applied mask for DDO 210. Note that the central regions of each galaxy fit within a single CCD, and so we only do the integrated light analysis for a single CCD per galaxy. This helps us to minimize systematic uncertainties due to detector responses and other instrumental effects.

To estimate the background, the median luminosity is found within annuli that are well separated from the galaxy, i.e. which are outside of a large ellipse centred on the galaxy. The size of this ellipse varies from galaxy to galaxy from 2 to 4R_s. The inner bound of this background region is shown in Fig. 4 for DDO 210 as the blue ellipse. To determine the uncertainty on the background estimation, we look at the variation between annuli. This technique accounts for the non-uniformity of the background and acknowledges that systematic errors are dominant compared to Poisson uncertainties. These systematic errors have spatial scales associated with them, depending on the source, and so it is important that the annuli we use to measure these variations are the same width as are used for the main profile. For example, some galaxies like Leo A have a bright star in the image. Consequently, internal reflections within the MegaCam instrument can create large haloes; hence, the background across the image is not expected to be uniform. The radial profile of DDO 210 is shown in ADUs in Fig. 5, along with the background estimate and the adopted uncertainty.

The background subtracted profiles for all Solo dwarfs are shown in Fig. 6. For some of the faintest systems, there is no discernible i-band integrated light component above the background. We show the colour profiles as sub-panels to the main panels for those dwarfs where both g- and i-band profiles were determined.

For all nine galaxies with observable colour profiles, the colour profiles are largely flat or slightly rising, indicating blue centres and increasingly red towards larger radii. This structure is consistent with the fact that the Solo dwarfs are dwarf irregulars, with recent or ongoing star formation.

4 CHARACTERIZING SOLO DWARFS USING RESOLVED STARS

4.1 CMDs and distance estimates

Fig. 7 shows the CMDs of two Solo galaxies to illustrate the different degrees of contamination from the Milky Way foreground. Perseus is relatively low apparent luminosity (m_V = 14.1 ± 0.7) at a relatively low Galactic latitude (b = 147.8^○), whereas Peg DIG is brighter (m_V = 12.6 ± 0.2) at a higher latitude (b = 94.8^○). Note that, in what follows, the term ‘background’ is used to describe foreground contamination arising from the Milky Way as well as from true background sources (generally distant, compact galaxies misidentified as stars at faint magnitudes). In both cases in Fig. 7, the stellar populations of the dwarf are most easily discerned by selecting only stars near the centre of the galaxy.

The CMDs for all of our sample are shown in Figs 8 and 9. The main feature visible in all CMDs is RGB stars. Young, blue stars are also present (for those with recent star formation), and can be seen as vertical sequences around (g − i)_o ≃ −0.5. Sag DIG, Leo A, WLM, IC 1613, and NGC 6822 show a clear blue sequence, whereas DDO 210 and Peg DIG (and tentatively Leo T and Perseus) show a few stars that may be young, blue stars. Foreground contamination makes distinguishing individual stars associated with the dwarfs not possible in this region without additional observations. AGB stars may also be present for some of the galaxies, although these are hard to distinguish from RGB stars, except at brighter magnitudes, where they are brighter than the tip of the RGB (e.g. WLM). On each panel, a Dartmount isochrone (Dotter et al. 2008) is shown, adjusted to the correct distance using our tip of the RGB measurements (see below) and with a metallicity selected to match the average colour of the RGB. The metallicities are listed in the upper right of each plot, ranging from [Fe/H] = −0.9 (UGC 4879) to −1.7 (Sag DIG). This range is broadly similar to the satellites of the Milky Way and M31 with the satellites tending to have additional lower metallicity dwarfs (from values listed in the updated tables of McConnachie 2012, and respective references). The isochrones plotted are selected to be old (∼12 Gyr) and intended to help visualize the RGB. As these dwarfs are irregulars, they are not necessarily simple, old stellar populations. For instance, the Perseus is not ideally matched; however, this does not impact the distance estimate.

The distance to each dwarf is found using the tip of the RGB (TRGB) as a standard candle. The TRGB is ideal as its luminosity is insensitive to small variations in age or metallicity given an older, metal-poor population (Lee, Freedman & Madore 1993). Indeed, using the Darthmouth isochrones (Dotter et al. 2008), the difference between a 10 Gyr, 11 Gyr, or 12 Gyr stellar population is of the order of Δi = 0.002. This uncertainty is not significant given that the uncertainties in identifying the TRGB from the CMD are of the order of Δi = 0.1.

First, the RGB stars are selected from a CMD using parallel ‘tram-lines’. These tram-lines are visually selected and generously encompass the RGB branch. If the RGB branch is well populated, the TRGB is apparent by a significant ‘break’ in the RGB luminosity function. The location of this break is identified using a five-point Sobel filter, similar to Sakai, Madore & Freedman (1996). As the luminosity function is binned, we remove the dependence of the choice of bin locations and widths by repeating the analysis with different binning choices, similar to the method used by Battaglia et al. (2012). Note that we adopt upper and lower bounds on the region in which the TRGB may be found, to avoid spurious signals due to structures like the red clump or random stars clearly unassociated with the TRGB. These limits are chosen visually and generously encompass the whole region in which the TRGB is located. For low-mass dwarfs with poorly populated RGBs, the limits are selected cautiously as to avoid biasing results. The distribution of resulting TRGB values (10 000 combinations of bin width and location in total) is shown with the median values and 1σ interquartile range highlighted.

In the above analysis, we do not find it necessary to subtract a reference luminosity function (due to the Milky Way contamination), except in the case of Sag DIG. Typically, the contamination is fairly uniform near the TRGB and due to the low number of stars in some galaxies, subtracting a background did not help in identifying the TRGB, but rather just added more noise. However, for Sag DIG, which has a reasonably well-populated RGB branch and substantial contamination, we found it was necessary to statistically remove this background. As detailed in Paper I, a portion of the foreground structure in this direction is likely due to the Sagittarius stream.

The distance modulus, (m − M)_o, is computed from the TRGB using an absolute magnitude of i_o,MP9703 = −3.49. This value is derived from the Darthmouth isochrones (Dotter et al. 2008) using the median value for old (12 Gyr) stellar populations with [Fe/H] < 0.0 (consistent with the isochrones matched previously). The TRGB value is converted to the new CFHT filters using the previous equations (equations 1 and 2).

4.2 Structural analysis

Once the distance has been calculated for each dwarf galaxy, we examine their spatial distribution based on the RGB stars. These stars are selected from the full stellar catalogue (no spatial constraints) as within the tram-lines previously identified and fainter than the TRGB. Figs 10 and 11 show the distribution of these stars in the entire field of view for each dwarf in the subset. The fields of view were positioned so that the galaxies were slightly off-centre to avoid the small detector gaps between CCDs. The two large gaps in the detector between CCD rows 1 and 2 and rows 2 and 3 are clearly visible. Note that those observations taken with the newer filter set at CFHT have a different footprint. The new filters are physically larger, and the ‘ears’ of the detector can now be used whereas previously they were not (increasing the total number of science CCDs from 36 to 40).

The main panels of Figs 12 and 13 show the corresponding density maps for each of our dwarfs, where contour levels are set at a minimum of |$2 \sigma$| above the background, and pixels are 1 arcmin in each dimension. To estimate the background, we examine the density of stars outside of an ellipse corresponding to 5 × R_s centred on the galaxy (where R_s is the Sérsic scale radius fitted to the radial profile as described in Section 5). This criterion is modified for the largest three galaxies as 5 × R_s does not lie within 1^○ field of view. For WLM and IC 1613, we estimate the background at 3 and 2 × R_s, respectively. For NGC 6822, no background is estimated; RGB stars belonging to the dwarf dominate over the entire full field of view.

Starting with a circular aperture, the PA, e, and centre are determined as a function of radius, in 0.5 arcmin steps, using these relationships. The analysis is then repeated, but this time only using those pixels that are contained within an ellipse with the newly derived PA and e. This process is repeated until the values converge.

The resulting PA and ellipticity profiles are shown as a function of SMA radius in the side panels of Figs 12 and 13. These profiles are fairly regularly behaved in the uncrowded regions of the galaxy (indicated by the unshaded regions in Figs 12 and 13) for all but the faintest galaxies. A single PA and e are adopted for each galaxy by determining the values between R_s and 3 × R_s for use in the following analysis. We include the standard deviation within this range as uncertainties on these quantities.

Radial profiles based on star counts alone are determined using the median values for PA, ellipticity, and centre. When computing the radial profile, a mask is used to ignore areas of our field with chip gaps, edge regions, and areas where detection of stars was not possible due to the presence of highly saturated foreground stars.

In all cases except NGC 6822, the profile is background subtracted using the background values estimated previously, generally using the area beyond 5R_s. The appropriate multiplicative factor for R_s, however, is estimated in the next section. In practice, this resulted in an iterative process, whereby we adopted an initial value for R_s to estimate the background, and then derived a new value of R_s based on the resulting profile. The entire process was repeated and continued until convergence. The final radial profiles based on star counts alone are shown in Fig. 14. These radial profiles clearly extend to larger radii that the integrated light profiles shown in Fig. 6. The impacts of crowding and the resulting incompleteness of the stellar catalogues are apparent at small radii, most strongly for the more massive dwarfs (for example, WLM very clearly shows this feature). In the case of NG 6822, the dwarf dominates over background sources across the entire image, so no background is subtracted. The resulting R_s is still robustly measured.

5 DWARF GALAXY PROFILES

5.1 Profile fitting

Extended radial profiles are now generated using the RGB stellar density profiles as tracers of the outskirts of the galaxies and the integrated light profiles as tracers of the inner regions of the galaxy. A significant assumption behind this approach is that the integrated light and the RGB stars trace the same populations. This is clearly incorrect in detail. However, it is a reasonable and necessary assumption to connect the outer stellar profile to the inner integrated light region. The systematic effects are limited by matching the i-band profile to the RGB stars, since younger stars contribute more heavily to bluer bands. For some of the smaller dwarfs – And XXVIII, Leo T, and Perseus – the dwarf does not appear in the i band via integrated light and we can only detect the integrated light in the g band. For these dwarfs, we instead match the stellar density profiles in the g band and adjust them to the i band using the median colour of the RGB from the CMD for these dwarfs.

5.2 Two-component models

Each radial profile is also fitted with the sum of two Sérsic profiles (again including γ_rgb as a free parameter) to explore any evidence for more than one component in any of these galaxies. Both Bayesian and Akaike Information Criteria (BIC and AIC, respectively; Kass & Raftery 1995) are computed to determine whether the extra parameters in the multicomponent model are statistically warranted. Specifically, we compute these parameters for both the single- and double-component profiles, and look at the difference. Models with lower BIC and AIC values are formally preferred, and the specific difference between the two sets of values indicates to what degree the model is preferred. These two criteria differ in how goodness of fit is weighted versus the number of parameters. Broadly speaking, the AIC is likely to favour models that fit well even if there are more parameters, and the BIC is likely to favour simpler models that fit less well.

Comparison of the AIC and BIC values for each dwarf for the single- and double-component models indicates that most dwarfs are clearly more suitably described by single component models. UGC 4879, DDO 210, and WLM are exceptions for which the preferred model is less clear. Fig. 16 shows the two-component Sérsic fits for each of these systems. Also shown are the values of Δ(BIC) = BIC₁ − BIC₂ and Δ(AIC) = AIC₁ − AIC₂ for each galaxy, where the subscript refers to the number of components fitted.

For the BIC statistic, a difference in values less than around 2 does not imply any significant preference between models, whereas differences of the order of 2–5 are indicative of moderate preference for the model with the lower BIC value. For the AIC statistic, the exponential of half of the difference in values is broadly proportional to the probability that the model with the lower AIC value minimizes the information loss. In our case, this implies that, for UGC 4879, the one-component model is a factor of exp (−6.3/2) = 0.04 more likely than the two-component model to minimize the information loss (i.e. the two-component model is significantly preferred); for DDO 210 and WLM, the corresponding numbers are 5.5 × 10⁻³ and 1.7 × 10⁻³.

Interpretation of these results requires some caution. In particular, stellar population gradients could have a very strong effect on these profiles, because of our necessary reliance on integrated light studies for the central regions. Inspection of the CMDs of these galaxies in Figs 8 and 9 shows that young stars are present in these galaxies, and appear especially strong in the magnitude range of our observations in WLM and DDO 210. Indeed, the differences in the radial gradients of young blue stars and RGB stars have been measured by McConnachie et al. (2006), Leaman et al. (2012), and Jacobs et al. (2011) for DDO 210, WLM, and UGC 4879, respectively. Thus, it is quite possible that the preference for two components in our analysis for these galaxies is a result of these differences in the spatial distribution of the stellar populations, although we note that the very few bright blue stars in our CMD for UGC 4879 suggest other factors potentially at play. Indeed, it has previously been argued (Bellazzini et al. 2011a) that UGC 4879 may host a disc-like component because of the ‘wings’ that are visible in the 2D surface brightness map (see Fig. 12). We will discuss this feature in more detail in Section 6.

5.3 Derived structural parameters

For each galaxy, Table 2 lists our derived positional and shape parameters, specifically RA, Dec. (from McConnachie 2012), distance modulus, distance, PA, ellipticity, and offsets of the centre of the RGB stellar distribution from the adopted centres.

Table 3 lists the fitted Sérsic parameters for each dwarf, specifically μ_e, R_s, n, and the adopted background density of RGB stars. Also included are the best-fitting parameters for the two-component Sérsic profiles for DDO 210, UGC 4879, and WLM.

Table 4 shows derived parameters from the Sérsic fits following Graham & Driver (2005): μ_o is the central surface brightness, 〈μ〉_e is the mean surface brightness inside the effective radius, m_tot is the apparent total magnitude, and M_tot is the absolute magnitude using the distance modulus given in Table 2.

6 DISCUSSION

6.1 Structures and substructures

Together, the 2D and 1D surface brightness profiles presented in Figs 8, 9, 12, and 13 provide considerable information on the extended structures of all these isolated galaxies. Figs 12 and 13 show that nearly all of the galaxies have a relatively regular appearance, essentially appearing as large spheroids with no evidence of large-scale asymmetries like tidal tails. For almost all of the dwarfs, the best-fitting Sérsic profile to the RGB distribution yields n ∼ 0.7; i.e. the stellar distribution falls off more slowly than an exponential. This is fairly typical of many dwarf galaxies (e.g. Ferrarese et al. 2006; Gerbrandt, McConnachie & Irwin 2015; Muñoz et al. 2018a).

An exception to both of the above generalizations is UGC 4879. Here, the stellar distribution falls off faster than an exponential (n ≃ 1.3), and the galaxy does present evidence of extended ‘wings’ along its major axis. This latter feature has previously been interpreted as an edge-on disc. The tentative evidence of two components in Fig. 16 also suggests that a multicomponent interpretation could be correct for this galaxy, where both components are dominated by old stars (potentially in contrast to DDO 210 and WLM; see discussion for WLM below). The extreme isolation of this system makes it extremely unlikely that such signatures are the results of a tidal interaction with a more massive galaxy (M31 or the MW). However, it is possible that this type of structure – especially the surface brightness map – could also be the result of a merger in this galaxy. Numerous observational studies in recent years (e.g. Martínez-Delgado et al. 2012; Paudel et al. 2017; Zhang et al. 2020) – including in the LG (e.g. Nidever et al. 2013; Amorisco, Evans & van de Ven 2014) – suggest that dwarf–dwarf mergers are not uncommon. In addition to UGC 4879, Sag DIG also has some evidence of a weak extension in its outer isophotes on the eastern end of its major axis (also discussed in Paper I).

Beyond the more distinct ‘wings’ of UGC 4879 and perhaps Sag DIG, the outskirts of several dwarfs are not ideally characterized by the Sersic fit (e.g. DDO 210, Leo A, and IC 1613). In these cases, an excess of stars is found in the outskirts, rather than a deficit. Irwin & Hatzidimitriou (1995) discuss these ‘extra-tidal’ stars extensively in the context of Milky Way dwarf spheroidals, whether these stars are being stripped or are evidence that the system is not in dynamic equilibrium. For Solo dwarfs, the likelihood of tidal interactions is reduced (though not eliminated) by their isolation. Interestingly, the excess of stars found in UGC 4879 at large radius persists in both the single and two component fits.

Wheeler et al. (2015) suggest that dwarf galaxies such as the Solo dwarfs could well host satellites within around 50 kpc of the main dwarfs. Dooley et al. (2017) take this further and explicitly calculate the probability of there being a satellite found around each of the Solo dwarfs within the approximate observational footprint for each of our fields (see their table  3). For reference, our MegaCam field reaches out to about 0.5 deg from the centres of each galaxy; for the closest galaxy (Leo T), this corresponds to a maximum (projected) radius of around 3.5 kpc; for the most distant galaxy (UGC 4879), this corresponds to around 10 kpc. Dooley et al. (2017) predict that And XVII, And XXVIII, Sag DIG, Peg DIG, and Leo A all have around a 10 per cent chance of having a satellite in this region, and WLM, IC 1613, and UGC 4879 all have around a 10–20 per cent chance of having a satellite in this region. Notably, NGC 6822, as one of the most massive dwarfs, has a relatively low probability due to the relatively small radius out to which we probe it.

No evidence of satellites or other concentrated substructures is found around any of the galaxies to our surface brightness limits (around 30–32 mag sq. arcsec), as inspection of Figs 12 and 13 makes clear. Candidate features [e.g. the feature in the isophotes around Perseus at (ξ ≃ −5 arcmin, η ≃ −2.5 arcmin)] were examined and found to be consistent with fluctuations (statistical or due to irregular features, e.g. holes due to bright foreground stars) in the background. Further, the OPTICS clustering algorithm (Ankerst et al. 1999; see its application to a similar data set in McConnachie et al. 2018) was also applied to the RGB data set, and no additional significant features were found. We note that this is relatively unsurprising, given that the probabilities given by Dooley et al. (2017) are not large for any single object. Further, we note that these probabilities correspond to satellites with stellar masses as low as |$10^4\, \mathrm{M}_\odot$|⁠. At this limit, detection would be very hard given the data in hand; the lowest stellar mass satellites identified in PAndAS (which used the same instrument, same bands, and broadly similar exposure times) were of the order of |$10^{4.2}\, \mathrm{M}_\odot$|⁠, although most were |$\gt 10^{4.5}\, \mathrm{M}_\odot$| (see discussion in Martin et al. 2016). Examination of the numbers in table 3 of Dooley et al. (2017) suggests that the probability of discovering new satellites around some of the more distant, more massive galaxies in the rest of the Solo data set remains good.

6.2 Specific case studies

Here, we present a closer look at our results for three galaxies included in our sample – Sag DIG, WLM, and DDO 210 – and use these as case studies to highlight aspects of our approach, interpretation of results, and comparison to previous studies. In particular, Sag DIG was already analysed in Paper I; WLM is a large and very well studied galaxy; DDO 210 has previously been studied with similar observations and methods and is at the faint end of our sample.

6.2.1 Sag DIG

While Sag DIG was extensively studied in Paper I, there are a few important differences between the methodology applied in Paper I and in this work. The integrated light and resolved RGB profiles were matched in Paper I by matching them ‘by eye’ in the overlapping regions. Archival HST observations were used in order to straddle the region between the resolved stars and integrated light from the CFHT data. However, appropriate HST observations are not available for all targets, and so this method was not used in this paper. Instead, we make the scaling parameter a free parameter and employ emcee to obtain the best fits and explore correlations between parameters.

Critically, this work takes a more rigorous approach to dealing with the background in the integrated light analysis of the dwarf. In Paper I, masking was applied to all pixels that fell above a certain threshold of counts. In this work, a threshold is still applied, but also pixels in the neighbourhood of these ‘hot’ pixels are also masked, in order to suppress the effect of haloes around bright stars and similar effects. This change in strategy has a significant effect for a galaxy like Sag DIG that is situated at relatively low Galactic latitude, and hence is impacted heavily by (bright) Milky Way foreground stars. In the CMD in Fig. 8, this foreground is prominent, and even Galactic substructure due to the Sagittarius Stream can be seen.

Fig. 17 shows the difference in the resulting integrated light profiles for Sag DIG. In particular, the more complete masking of bright sources in this paper means that the overall background that is estimated for Sag DIG is smaller (by around 20 ADU, or just less than 1 per cent of the previous background value). This changes the shape of the outer profile considerably, such that the new analysis clearly favours a more extended surface brightness profile. Note also that the uncertainties are much larger in this new analysis, since we explicitly take into account the systematic variation in the background due to the irregular distribution of foreground sources, whereas this was not the case in Paper I. The net flow-down effect of this change in our analysis is considerable for Sag DIG (R_s = 0.95 ± 0.01 in Paper I compared to R_s = 1.43 ± 0.08 arcmin in the current analysis). This emphasizes the challenge of integrated light studies of these faint systems, and the need for a homogeneous approach to the analysis.

Beccari et al. (2014) also present a wide and deep study of the resolved stellar and H i structures of Sag DIG. The Sérsic radius derived from Beccari et al. (2014) is approximately 1.1 arcmin; no error bars are given on this measurement but it appears to be in reasonable agreement to that derived here. The TRGB distance estimate (m − M)_o = 25.39 ± 0.08 agrees well with (m − M)_o = 25.16 ± 0.11 from Beccari et al. (2014). Beccari et al. (2014) and Lee & Kim (2000) estimate the PA to be 90^○, which is in agreement with the value (measured East from North) found in this work (87^○ ± 3^○). Similarly the ellipticity from Beccari et al. (2014) and Lee & Kim (2000) is e = 0.5, which agrees with 0.56 ± 0.18 estimated here.

6.2.2 WLM

WLM (also known as DDO 221) is one of the more massive dwarfs in the Solo Survey. For comparison of the structural parameters, we focus on the detailed analysis from Leaman et al. (2012, 2013). This deep, wide-field analysis also uses resolved RGB stars to study the older stellar populations in addition to including spectroscopy of 180 RGB stars. Leaman et al. (2012) derive the ellipticity and PA of this dwarf by fitting ellipses using MPFITELLIPSE (Markwardt 2009), finding e = 0.55 and PA = 179^○. These values are in good agreement with e = 0.54 ± 0.06 and PA = 177^○ ± 0.5^○ derived here. The ellipticity and PA of WLM are well defined and do not vary significantly with radius.

The half-light radius given for an exponential fit to the overall population in Leaman et al. (2012) is approximately 5.8 ± 0.2 arcmin, whereas the RGB-only profile has a corresponding radius of approximately 4.7 arcmin; here, we find a slightly smaller half-light radius (4.10 ± 0.13 arcmin). As discussed earlier, there is some evidence that the overall profile of WLM is best described by a two-component Sérsic model. We urge caution here, since it is unclear if this is due to the significant number of young stars in WLM, which certainly have a different spatial distribution to the RGB stars used to create the resolved star profile (see Leaman et al. 2012).

6.2.3 DDO 210

DDO 210 (Aquarius) is another well-studied dwarf and, prior to the discovery of Leo T, was the smallest known galaxy with a significant gaseous component. We measure the TRGB at (m − M)_o = 24.97 ± 0.09, in excellent agreement with Cole et al. (2014) who measured (m − M)_o = 24.95 ± 0.10 and comparable to McConnachie et al. (2006) with (m − M)_o = 25.15 ± 0.08. The PA and ellipticity are derived using the same method in McConnachie et al. (2006) and agree closely. McConnachie et al. (2006) found that the ellipticity changed from 0.6 in the inner regions to 0.4 in the outskirts; again, this agrees closely with this work where we find a median ellipticity of e = 0.53 ± 0.05 and a very modest radial gradient.

McConnachie et al. (2006) demonstrate the complex structure of this galaxy in terms of the differences in the radial gradients of its different stellar populations, and the apparent disconnect between the spatial properties of the stars and H i distribution. Hermosa Muñoz et al. (2020) obtained spectroscopic measurements for 53 RGB stars and determine the axis of rotation for the stellar component to be |$139^{\circ } {^{+17}_{-27}} ^{\circ }$| in contrast with the H i rotation measured by Iorio et al. (2017) (77.3^○ ± 15.2^○). In our analysis, we find that this galaxy is possibly better described with a two-component profile, but again it is hard to determine whether this represents two ‘old’ components or whether it reflects the age segregation in the stellar populations noted by McConnachie et al. (2006).

6.3 Comparing the suite of parameters to the literature

We now compare the global set of parameters derived in this work to previously derived values, particularly the updated compilation of parameters from McConnachie (2012). Fig. 18 shows the comparison for distance modulus [(m − M)_o], central surface brightness (μ_o), apparent magnitude (μ_tot), ellipticity (e), PA, and half-light radius (R_s). For the central surface brightness and apparent magnitude, μ_o and m_tot values in Table 4 measured in the i band were converted to their corresponding g-band values given the median (g − i) values given in the same table. These values were then transformed to the V band using V = g − 0.098 − 0.544(g − i) (Thomas et al., in preparation). Points with an ‘X’ have literature values that are highly uncertain, and do not have quoted uncertainties. And XVIII does not have any previous value for its ellipticity or PA (previous parameters were based on an image where the bulk of the object fell on a chip gap).

It is worth emphasizing that, to the best of our knowledge, some of these galaxies – especially NGC 6822, IC 1613, and Leo A – have not had detailed structural analyses conducted based on resolved stars in the modern era of CCD photometry. The integrated magnitudes of several of these galaxies date back to The Third Reference Catalogue of Bright Galaxies (de Vaucouleurs et al. 1991), and in fact the half-light radii quoted in McConnachie (2012) for these three galaxies are estimated by scaling other scale radii that are measured by de Vaucouleurs et al. (1991). It is clear from the lower right panel of Fig. 18 that this scaling is distinctly inconsistent for NGC 6822.

Other notably deviant measurements include the central surface brightness of And XVIII, but this is relatively easily understood given that the earlier measurement was affected by an inconvenient chip gap. Much more difficult to understand is the difference in ellipticity between our measurement of Perseus, which places the galaxy as nearly circular, compared to the measurement by Martin et al. (2013), which records a notable ellipticity. Inspection of the raw star count map in Fig. 11 does not suggest that the galaxy is notably elliptical, consistent with our findings, and we speculate that the higher quality photometry for the MegaCam data compared to the discovery PS1 photometry has allowed a cleaner sample of member stars to be selected (this galaxy is at b ≃ −15 deg). Obviously, the poor agreement with the ellipticity measurement correlates with the poor agreement with the PA measurement.

Generally, we see reasonable agreement between the bulk of our measurements and those from the literature. The only systematic difference that is obvious is that almost all of the effective radii that we derive are systematically larger than previous estimates. We suggest there are two primary factors that contribute. First, some previous work has used observations with a smaller field of view. There is a non-trivial trend in the literature such that when viewing a galaxy out to larger radius, the galaxy is found to be larger than previously estimated. Secondly, and likely more important for the specific systems under study, there is a difference between the half-light radius of a galaxy and a half-mass radius for any given tracer. Most Solo dwarfs are star forming, and so the central regions (and the light) are dominated by the younger, bluer brighter stars. However, our structural parameters are based on RGB stars, belonging to older stellar populations. It is often the case that older stellar populations are more extended than younger populations, and we stress that our results should be viewed as estimates of the older, halo-like populations of these dwarfs. The relative importance of these two factors (field of view versus half-light/half-mass) likely varies between dwarfs and the estimates used for comparisons. For example, Peg DIG and IC 1613 have a young stellar component (Weisz et al. 2014) and R_s in Fig. 18 is from de Vaucouleurs et al. (1991) that is based on the integrated light. As a result, the difference between the mass-weighted and light-weighted profiles is likely the dominant factor. This could also be a factor for some of the central surface brightness measurements (where integrated light studies will measure the flux density of all the stellar populations in the central regions, whereas our estimates are weighted towards the flux density due to older stellar populations). In contrast, structural parameters for And XVIII were previously based on observations in which the galaxy partially lies over a large chip gap in the MegaCam detector. Here, the differences seen in Fig. 18 are likely largely due to the wider, more complete Solo observations. In addition, And XVIII has a predominately older star formation history (as seen in Weisz et al. 2014), meaning that the discrepancy between the mass- and lighted-weighted profiles is reduced.

7 CONCLUSION

The purpose of this paper has been to determine a homogeneous set of structural parameters of isolated LG galaxies that is suitable for comparison with the well-studied Milky Way and M31 satellite populations. The diversity of this population of dwarf galaxies leads to considerable complexity in generating a homogeneous data set, but in general we find good agreement with previous studies. We have focused on their extended stellar structure and morphology and combined information from integrated light analysis of their inner regions with the resolved stellar populations of their outskirts, to parametrize their global structure to very faint surface brightness limits. Both the resolved stars and the integrated light require careful processing due the intrinsically small and faint nature of these dwarfs. Careful consideration of background subtraction is necessary due to the significant impacts on the resulting structure, particularly on the extended structure. The differences found in the structure of Sag DIG between Paper I and this work highlight these effects. All 12 galaxies are reasonably well described by 1D Sérsic functions, and no prominent stellar substructures, which could be signs of either faint satellites or recent mergers, are identified in the outer regions of any of systems examined.

Future contributions will use these results in comparison to surveys of LG satellite populations to examine differences between the populations of nearby dwarfs that may be attributed to the role of environment.

ACKNOWLEDGEMENTS

Based on observations obtained with MegaPrime/MegaCam, a joint project of CFHT and CEA/DAPNIA, at the CFHT that is operated by the National Research Council (NRC) of Canada, the Institut National des Sciences de l’Univers of the Centre National de la Recherche Scientifique (CNRS) of France, and the University of Hawaii. The observations at the CFHT were performed with care and respect from the summit of Maunakea that is a significant cultural and historic site. We respectfully acknowledge that UVic, where this work was conducted, is on the ancestral territories of Indigenous Peoples. As we explore the shared sky, we acknowledge our responsibilities to honour their past and enduring relationships with these lands. We thank Todd Burdulis for fantastic support throughout the Solo observational campaign. CH kindly thanks the IoA for its support and hospitality while visiting during the completion of this work. GB acknowledges financial support through the grant (AEI/FEDER, UE) AYA2017-89076-P, as well as by the Ministerio de Ciencia, Innovación y Universidades (MCIU), through the State Budget and by the Consejería de Economía, Industria, Comercio y Conocimiento of the Canary Islands Autonomous Community, through the Regional Budget.

DATA AVAILABILITY

The raw data on which this analysis is based are publicly available in the CFHT archive (accessed via the Canadian Astronomical Data Center at http://www.cadc-ccda.hia-iha.nrc-cnrc.gc.ca/en/cfht/).

Processed data may be shared on reasonable request to the corresponding author.

REFERENCES