An Extensive Catalog of Early-type Dwarf Galaxies in the Local Universe: Morphology and Environment

Stated as a morphology–density relation, galaxy population depends on the local environment (e.g. Dressler, 1980; Whitmore et al., 1993). Massive galaxy clusters are dominated by red and dead early-type morphology galaxies, and low-dense groups and fields host a larger fraction of star-forming late-type galaxies (Oemler, 1974). In the dwarf galaxy regime (commonly defined as less massive than the Large Magellanic Cloud (LMC)), we can see an extreme form of the morphology–density relation, with the non-existence of early-type morphology, dwarfs outside the group and cluster environments (Binggeli et al., 1987; Geha et al., 2012).

A natural explanation for the existence of an extreme form of the morphology-density relation in the dwarf galaxy population is that the environment-related mechanisms mainly cause the quenching of dwarf galaxies and there are varieties of candidates (Boselli et al., 2008; Kormendy et al., 2009). Ram pressure stripping (RPS; high-speed interaction between hot cluster gas and reservoir of star-forming gas of in falling galaxy introduces a viscous drag, which eventually removes the star-forming gas (Gunn & Gott, 1972)), tidal harassment/stripping (removal of gas through tidal deformations (Moore et al., 1996; Mayer et al., 2001; Smith et al., 2010)), and starvation (a short supply of star-forming gas in the cluster environment (Larson et al., 1980)), are the prime candidates that are frequently discussed in the literature. Since the environments themselves vary from the cluster and group to the field, a longstanding and critical issue is which particular mechanism is the more efficient in which particular environment. The RPS is expected to be more efficient in the cluster environment. Evidence of ongoing RPS in the cluster environment, explicitly in the Virgo cluster, have been frequently presented in the recent literature (Kenney et al., 2004, 2014; Vollmer et al., 2001). The tidal stirring or harassment is more frequent in the small group environments (Mayer et al., 2001; Paudel & Ree, 2014). It is also important to note that the combination of time spent while undergoing a given mechanism as well as its strength, should be considered. It has been suggested that only a combination of several different processes can explain the diverse morphological properties of cluster early-type dwarf galaxies (dEs, see e.g., Lisker, 2009, for an overview). On the other hand, the importance of the environment relative to a small-scale merger in the evolution of dwarf galaxies remains poorly understood.

By definition, dEs are non-star forming and gas-poor objects, somewhat lower mass cousins of Elliptical galaxies (Ferguson & Binggeli, 1994). Therefore, a smooth appearance and red and dead (aka old stellar populations) are the prime characteristics of these galaxies, but more notably, their low-surface brightness nature makes them an outstanding class of object compared with massive early-type galaxies. Due to their low-surface brightness and scarcity in the nearby universe, particularly in low-dense environments like the field, a detailed study of environmental effects on their observational properties using a large sample of dEs has been lacking.

On the other hand, detailed studies on the morphological properties of cluster dEs have revealed that dEs are not a homogeneous class of objects. With the help of sophisticated image analysis techniques, a great deal of complexity in dEs structure has been discovered. Fine structures such as tidal tails, blue cores, spiral-arms or bars are frequent in the brighter dE population (Jerjen et al., 2000; Lisker et al., 2007a, b; Paudel & Ree, 2014) and in general, cluster dEs light profiles are not well fitted by a single Sérsic function (Janz et al., 2014; Su et al., 2021). A careful analysis of these fine structures has shown that they could contribute up to 10 percent of the total light of the dE main body (Smith et al., 2021; Paudel & Ree, 2014).

Very recent deep imaging surveys have also revealed that the dEs of stellar mass as low as the Fornax dwarf galaxy is hosting tidal features that are likely to have originated through mergers (Crnojević et al., 2014; Paudel et al., 2017).

In addition, a compact star cluster is also found in the center of many dEs, called the nucleus. Many detailed observational studies have shown that the central nucleus is almost universal in the brighter dE population of the cluster environment (Côté et al., 2006) and the nucleated fraction declines to as low as 30% among the fainter ($M_{g}$ $>$ $-$12 mag) dEs (Sánchez-Janssen et al., 2019). Interestingly, a few dedicated studies on these sub-structural properties of dEs beyond the cluster environment, particularly outside of the Virgo and Fornax cluster, are available to date and it is not well understood that the cluster environment is necessary for forming these sub-structural properties.

There is a recent upsurge in the exploration of a diffuse and extended dE on the name of Ultra Diffuse galaxies (UDGs, van Dokkum et al., 2015; Marleau et al., 2021), although their examples have been known for decades (Thompson & Gregory, 1993; Caldwell & Bothun, 1987; Impey et al., 1988; Jerjen et al., 2000; Conselice et al., 2002, 2003; Penny et al., 2011). A multitude of recent literature has presented the UDGs, as if a newly discovered class of dwarf galaxies (Mihos et al., 2015; Martínez-Delgado et al., 2016; Merritt et al., 2016; Román & Trujillo, 2017; Zaritsky et al., 2021) and there are reports of discovering them in all environments -cluster, group, and field, with a diverse color, morphology, and globular cluster population (Beasley & Trujillo, 2016; Lim et al., 2020; Müller et al., 2020). The diverse and complex UDG’s properties essentially suggest that UGDs could also form a mixed bag of populations, and there is no single origin for them, with multiple possible evolutionary tracks at play (Duc et al., 2014; Amorisco & Loeb, 2016; Bennet et al., 2018; Chan et al., 2018; Carleton et al., 2019; Jones et al., 2021).

With the advent of sophisticated observing techniques and wide-aperture telescopes, there has been a growing interest in low-surface brightness galaxies like dEs. Some groups have focused on building catalogs of dwarf galaxies with different morphology, and others have focused on a more specific type, like dEs or UDGs in the cluster or group environment. For example, Habas et al. (2020) investigated the overall dwarf galaxy population of different morphologies around massive early-type galaxies, and Ferrarese et al. (2012); Venhola et al. (2017); Eigenthaler et al. (2018); Ferrarese et al. (2020); La Marca et al. (2022) are more focused on the cluster environment. As a part of the Satellites Around Galactic Analogs (SAGA) survey Geha et al. (2012) searched dwarf galaxy populations around Milky Way analogs and Carlsten et al. (2022) explored satellite populations in local volume ($D$ $<$ 12 Mpc) host galaxies. In addition, Chiboucas et al. (2013); Crnojević et al. (2016); Müller et al. (2017); Bennet et al. (2019) searched member dwarf satellite galaxies in nearby groups.

In this paper, we investigate the morphological properties of dEs sampled from diverse environments trying to understand their relationship with the environment where they are located and provide statistically homogeneous data sets of a large sample of dEs and their morphological properties. This paper is organized as follows: Section 2 briefly describes the dE candidate identification procedure. In section 3, we present a measurement of photometric properties. Section 4 presents the results analysis of photometric and morphological properties. Finally, How our dEs are distributed in the cluster and group environment and their association with massive galaxies is described in section 5. Our main findings are concluded in section 6.

A primary motivation of this work is to create a large sample of dEs located in various environments, i.e., cluster, group, and field, of the local universe $z$ $<$ 0.01. We first used the publicly available imaging and spectroscopic data of the Sloan Digital Sky Survey (SDSS Aihara et al., 2011) to identify pure red and dead dwarf galaxies, selecting all dwarf galaxies of $M_{g}$ $>$ -18 mag and visually classified them as dE. We classify a galaxy as dE if the SDSS color image shows a red (color index $g-r$ $>$ 0.5 mag; Lisker et al. (2006)) and smooth appearance and the presence of no significant H_α emission in the SDSS optical spectrum. It is customary to use the first criteria to classify galaxies as dE in case the optical spectrum is not available. Since the SDSS also provides the optical spectrum, we have added the second constraint to confirm that we are selecting purely non-star-forming galaxies.

Within the area covered by the SDSS, there are 786 dEs in the redshift range of less than 0.01. These spectroscopically selected dEs served as a training sample to further extend the dEs sample from a wider sky coverage imaging database provided by the Legacy survey (Dey et al., 2019). This training is crucial, particularly for students who are starting to learn galaxy evolution research.

Since the Legacy survey does not have spectroscopic observation, we only used visual inspection to identify the dEs using $g-r-z$ tri-color images provided by the Legacy survey visual tool.

Contrary to the common practices, which first determine the sources/objects from a pre-built catalog that is primarily produced by an algorithm and then select the object of interest using specific criteria, we identified the dEs directly in the field, taking advantage of experienced knowledge of visual analysis of dE selected from the SDSS. Our method avoids the pre-selection of the sources detected by some automated software like source extractor (Bertin & Arnouts, 1996). Given the heterogeneous morphology of dEs, including extremely diffuse UDG, the automated source detection may miss the faint low-surface brightness objects due to their low signal. Even detected, specific criteria may not cover all dEs-like objects.

To minimize the contamination from the compact, high-surface brightness interlopers (such as foreground stars, globular clusters, and high-redshift galaxies), we explicitly select low-surface brightness and extended galaxies, the prime characteristics of dE. This selection may be biased against the compact early-type dwarf galaxies, such as M32. Nevertheless, they are several rare kinds of objects, e.g., there are a hand-full of cE like compact non-star forming dwarf galaxies while there is more than 500 normal dE in the Virgo Cluster Catalog (VCC) of the Virgo cluster (Janz & Lisker, 2008; Chilingarian et al., 2009).

To select a variety of environments that host dE, we have chosen all three essential aspects, i.e., cluster, group, and field. For the cluster, we have chosen two local clusters, Virgo and Fornax, and for the group, we have selected groups from a group catalog provided by (Makarov & Karachentsev, 2011). For the field, we define an area around bright galaxies of magnitude $M_{k}$ $<$ -21 mag, which are not a member of any cluster or group according to Makarov & Karachentsev (2011) group catalog. We select the host environment to avoid distance uncertainty of newly detected dE. Since, by design, most newly identified will not have radial velocity information, we assume they are located at line-of-sight distances of the hosts. Note that the definition used here for the field dE could be inconsistent with that commonly used in the literature. Since we deliberately searched dEs around giant galaxies, they are indeed members of the satellite system of that host galaxy, assuming that their distances are similar to the host. The selected environment represents the environment of host galaxies, not those detected dEs. However, we use the terminology “field dE” for those detected around the field host to simplify environment classification.

In total, we selected two clusters, 265 groups and 586 fields within a redshift range between $-$0.0013 and $<$ 0.01.

Indeed, there is a large body of dE surveys in cluster environments, particularly for Virgo and Fornax (Binggeli et al., 1985; Venhola et al., 2017; Ordenes-Briceño et al., 2018; Ferrarese et al., 2020), which can be found in the literature. However, this study aims to provide a more homogeneous sampling of dEs in diverse environments ranging from cluster to field.

Finally, we visually inspected a 1 Mpc area around the group center, typically a 2.5-degree sky-projected radius for a group located at a distance of 25 Mpc. This distance corresponds to the median distance of our selected groups. We inspected a 15-degree radius around the Virgo and a 10-degree radius around the Fornax cluster. We inspected a 450 kpc area from the host center for the field; this corresponds to a one-degree angular radius for the field located at 25 Mpc, which is the median distance of our field sample.

The motivation for selecting areas around the cluster, group, and area around bright Field galaxies is the finding of Geha et al. (2012), there are none or rare existence of non-star forming dwarf outside of group or cluster environment. In addition, from our spectroscopy sample, we noticed that very few dEs were located beyond 500 kpc of giant galaxies with a relative velocity larger than 500 km s^-1.

We show an all-sky view of the inspected area covered by the Legacy survey DR9¹¹1https://www.legacysurvey.org/dr9 in Figure 1, where each circle represents a group area with a 2.5^° radius. The two nearby clusters, Virgo and Fornax, are defined by the red circle of 10 and 7-degree radii, respectively. As we can see, the inspected areas frequently overlap with one another; in total, we have inspected 7643 square degrees of the Legacy survey images.

For visual inspection purposes, we extensively use tri-color cutout images provided by sky-viewer of the Legacy survey²²2https://www.legacysurvey.org/viewer. By design, we will not have radial velocity information for our dEs. Instead, based on their morphology and surface brightness, we assessed their heliocentric distance, less than 40 Mpc for $z$ = 0.01. The basic idea is our dEs have lower surface brightness than background sources for a given apparent magnitude. Furthermore, morphologically our dEs are smooth and regular and they follow a well-defined relation between absolute magnitude versus radius and surface brightness, which can also be used to constrain their heliocentric distances.

The visual catalog is created by a simple visual inspection of tri-color cutout images provided by the sky-viewer. At least two persons visually check each field, and we find that there are quite good agreements between each other ($>85\%$). Indeed such a visual scanning procedure is not 100% complete, and there remains the possibility that some dEs in the field of view has been missed. In an effort to quantify the missing frequency, we rescanned the field after placing the previously identified dE in a randomly chosen position. The mock dEs were selected based on binning of magnitude and mean surface brightness, and we repeated this exercise 100 times for each designated bin. In Table 1, we list the recovery fraction of dEs in our identification method, where the results are binned with the width of 2 mags in both surface brightness (y-axis) and magnitude (x-axis). We also list the total number of dEs in each bin and note that we ran our experiment only for those bins that have more than 100 dEs. As we expected, the recovery fraction depends on the magnitude and the mean surface brightness. We find that the recovery fraction is 99 percent for the brightest dE sample and decreases to 61 percent for the faintest sample. The median $g$-band magnitude of our sample dE is 19 mag and the magnitude range 18$-$20 covers two-thirds of our dEs sample (see section 3); thus, we can consider that our recovery fraction is over 80% on average. Our detection might also be influenced by the fact that we have used the SDSS confirmed dE (mostly brighter one) for the training purposes, and that may lead to a memory bias toward these bright dEs compared to the faint dEs, which is seen for the first time by the inspector.

An all-sky distribution of our dE sample is shown in Figure 3. To assign a host for a dE, we used the minimum radial separation between the host center and the dE; this is particularly important for the overlapping groups. For the Virgo Cluster member dEs, we used a circle of a 10-degree radius centering on M87 and for the Fornax Cluster, we used a circle of a 7-degree radius centering on NGC 3407. As we can see clearly, a large majority of dEs are concentrated around the clusters and large groups. The two clusters alone contribute nearly half (2437 out of 5405) of the total dE sample, and 265 groups contribute 2103 dEs. We have found 864 dEs in 586 fields.

We have identified a total of 5405 dEs in the clusters, groups and fields; among them, 1324 dEs have radial velocity information (hereafter spectroscopic sample) obtained from the NED and the SDSS. We show the redshift distribution of the spectroscopic sample in Figure 2, and the peak of the distribution matched with the mean radial velocity of the Fornax and Virgo Clusters, and indeed, they dominate the spectroscopic sample dEs. The two clusters constitute more than two-thirds of spectroscopy sample dEs.

We list this main sample of dEs in Table 2. We consider distance to the dE similar to their host cluster, group or field. We consider 16.5 and 17.5 Mpc distance for the Virgo and Fornax Clusters and for the group we obtained it from Makarov & Karachentsev (2011). For the field galaxies that do not belong to any group or cluster we use Hubble flow distance, using a standard cosmology with $H_{0}$ = 71 km s^-1 Mpc^-1, $\Omega$_m = 0.3, and $\Omega$_Λ = 0.7.

Once we finalized the source catalog, we performed the image analysis and measured the photometric and structural parameters. For this purpose, we exclusively used the $g$-band stacked images provided by the Legacy survey. We retrieved stacked images from the Legacy survey archive. While the Legacy image processing pipeline is optimized for relatively small objects, the photometric measurement of the large extended object is often inaccurate owing to imperfect sky subtraction and deblending of multiple sources. Galaxies with large angular sizes and irregular morphologies are often shredded into multiple separate objects, and thus the total derived fluxes for these galaxies are often unreliable. Furthermore, the Legacy survey photometric pipeline systematically underestimates the luminosities of galaxies of large sizes due to the overestimation of the sky background. In order to overcome these problems, we performed photometric measurements for all dE. We retrieved 3′$\times$3′  fits image cutout. We subtracted the sky background preparing a background map for each object. The background maps were constructed after masking out all identified sources in the image, which were defined by source-extractor segmentation maps, and the segmentation images were filled by the median values calculated from surrounding all pixels. This method allows us to eliminate any contribution of light from stars and background galaxies. Finally, the background map was subtracted from the original fits file to remove the sky background contribution to the observed flux.

Further image analysis was done largely in the same way as in Paudel et al. (2015). All unrelated foreground and background objects were masked, and the centers of galaxies were calculated as the centroid of the flux distribution in the masked image. The total magnitudes were calculated using the Petrosian method (Petrosian, 1976), for this, we derived an azimuthally averaged light profile using a circular aperture. The Petrosian radius is defined as the galactocentric radial position where the ratio of surface brightness at the radius $R$ to the average surface brightness within the $R$ reaches a certain value, denoted by $n$ i.e.,

where $\mu(R)$ is surface brightness at radius $R$ , and $\langle\mu(R)\rangle$ is the average surface brightness. We determine the Petrosian radii, $R_{p,n=2}$ and corresponding magnitudes, $m_{p,n=0.2}$, by adopting the most commonly used parameter $n_{R}=0.2.$

We finally calculated the total brightness (m_g), summing up all pixel flux within the two $a_{p}$. The half-light radius is then calculated where the total flux becomes half and uncommon to the usual practice our derived half-light radius is circularized, and it is not necessary to correct it by multiplying a scaling factor$(b/a)^{0.5}$.

If the Petrosian radius did not converge, due to the fact that these galaxies sit within the light of nearby bright sources, or that are highly deblended/overlapped with the nearby foreground/background bright object, we performed manual aperture photometry in $g$-band images only, after manually masking those unrelated objects. In that case, we will not have structural parameters derived for these galaxies.

The mean surface brightness ($<\!\!\mu_{g}\!\!>$) was calculated using an equation

The derived magnitudes were corrected for Galactic extinction using Schlafly & Finkbeiner (2011), but not $K$-correction. As expected, however, the $K$-corrections factors are negligible for the distance range of our target galaxies. Therefore, no $K$-correction has been performed in the derived magnitudes. Using the $g$-band as a reference, we followed the same procedure to measure photometric parameters in the $r$-band images.

We show the distribution of $g$-band magnitude of our dE sample in Figure 4, which has a peak of around 19 mag. The median magnitude of our sample is 18.68 mag, immediately reflecting that the majority of our dE sample galaxies are fainter than the spectroscopic target selection limit of the SDSS, i.e., 18 mag, as shown by a vertical line. We also overlay the distribution of our spectroscopic sample dE (the blue histogram), and as expected, it mainly overlaps in the brightest part.

In Figure 6, we show the $g-r$ color distribution of our dE sample, and all three subsample dEs show a single peak Gaussian distribution, with a peak around 0.6 mag. The overall median value of the $g-r$ color of our sample dE is 0.58 mag, with a standard deviation of 0.14 mag.

To compare our measured $g$-band magnitudes with that of the Legacy survey pipeline values, we obtained the Legacy survey magnitudes using the SQL queries within a 6$\arcsec$ radius of the dE center. We used the primary source catalog provided by the tractor. In Figure 5, we compare our $g$-band magnitudes and the Legacy survey catalog values. We find that these two magnitudes agree well with each other for the bright galaxies, and for the fainter galaxies, it seems that the Legacy tractor pipeline underestimates the flux.

We show the distribution of $<\!\!\mu_{g}\!\!>$ in Figure 7. It is very clear that most dEs have mean surface brightness higher than 24.5 mag arcsec^-2, which has been frequently used as a cutoff magnitude to define UDGs, see section 4.2. Our dE sample has a median value of $<\!\!\mu_{g}\!\!>$ = 25.97 mag arcsec^-2. Although we do not use any magnitude or surface brightness limit for the visual selection while preparing the final catalog, we select dEs brighter than m_g = 22.5 mag and $<\!\!\mu_{g}\!\!>$ $<$ 28 mag arcsec^-2. We introduced these limits based on the measurement error.

As per our selection procedure, our dEs are either inside the cluster/group or near bright field galaxies (hereafter the host). In this subsection, we explore the dE positions with respect to the host and morphology types of their companion bright galaxies (BGs, $M_{k}$ $<$ $-$21 mag).

In the bottom panel of Figure 15, we show a phase-space position of our dEs within their respective host environment. In the top panel, we show the distribution of sky-projected separation of the dEs from the host center, where the host-centric distance is normalized by viral radii of their respective cluster, group, and field. We used 1.7 and 1.2 Mpc for Virgo and Fornax Clusters, respectively. For the group, we used 800 pc, and for the field we used 300 pc.

Almost all dEs are located well within the 1.5 virial radii; however, a few field dEs are significantly away from the host. Although they are few, we noticed that these field outliers are a spectroscopically selected sample of dEs from the SDSS. We have used an all-sky catalog from the SDSS rather than searching around predefined areas of the cluster, group, or filed.

In terms of number count, the two clusters contribute nearly half, 2437 (1837 and 864 in Virgo and Fornax, respectively) out of 5405. We find that there are 7.9 dEs per group and 1.5 dEs per field, on average. We noticed that there are many fields which contain no dEs. Out of total of 586 fields we have inspected, only 233 fields host one or more dEs. We further classify the field host as an early-type (E) and star-forming (Sf) and find that there are 437 Sps and 153 Es field hosts. We find that out of 437 fields with a Sfs host, only 154 have one or more dEs, and out of 153 field with an Es host, only 79 have one or more dEs, which indicates a higher chance of hosting a dE by an Es than Sfs. On average, we find that early-type fields host a significantly higher (2.8) dEs per field compared to star-forming field hosts, which have 1.0 dE per field, on average.

Further, we explore the relationship between the location of dEs and morphological types of neighbor BGs, regardless of their environment. We counted BG frequency of each type around each dE within a 300 kpc radius. The result of this exercise is shown in Figure 16. We find a higher frequency of dEs around Es than Sfs. On average dEs have a 2.5 times higher chance to have an Es as a neighbor than Sfs.

Indeed, this result might have been heavily influenced by the cluster environment, where E dominates by number. We performed this exercise again, separating the cluster, group, and field. We find a similar trend in the cluster and group dEs, but in the field environment, where Sf dominates, we find an equal chance to have both E and Sf as a neighbor of dE on average. The results are listed in Figure 16.

We have identified 5405 dEs located in the various environments, i.e., cluster, group, and field, spanning the luminosity range $-$18 $<$ $M_{g}$ $<$ $-$8 mag. They are selected via visual inspection of 7643 deg² area of the Legacy survey $g-r-z$ combined tri-color images, which include two clusters (Virgo and Fornax), 265 groups, and around 586 field galaxies. Our dE catalog provides several different metrics by which we quantify morphological and environmental properties. The salient feature of this catalog lies in their morphological properties, as we identified morphological characters such as nucleated, tidal, and UDGs.

Given the nature of the selection procedure, it is not trivial to calculate the completeness limit of our sample. Although we aim to search dEs of up to $z$ = 0.01, we should emphasize that this is not a volume-limited sample in many regards. Indeed, we have not explored all sky areas of this volume. In addition, one of the main caveats of this work is the distance of candidate dEs, and in most cases we do not have that information. There may be a faction of dEs that are foreground or background galaxies or at least not directly related to the host. To confirm their distance and association with the host environment, we need radial velocity information, which we do not have in most cases.

This work is explicitly designed to find dEs in an extensive database, and there are not many surveys of this kind for which we can make a direct comparison. We found that Carlsten et al. (2022, hereafter C22) searched satellite dwarf galaxies (of all types) in the Local Volume ($D$ $<$ 12 Mpc) in the Legacy survey imaging.

Unfortunately, we found that only two areas around NGC 3379 group and NGC 3627 group broadly overlap with our searched sky regions. In the case of NGC 3379, they have found 36 dwarf galaxies, and among them, 33 can be classified as dE from our visual inspection; see Figure 17. Our catalog recovers 26 (78%) out of 33. Similarly, we recovered 16 (64%) out of 25 dEs in the case of NGC 3627. We noticed that those unrecovered 16 (9+7) dEs are mostly fainter dEs of $m_{g}$ $\gtrsim$ 20 mag. In the C22 search area, we identified 8 additional dEs, bringing the total number of identified dEs in two systems to 50. Calculating the recovery fraction of the C22 dE sample, we have recovered 42 (72%) dEs, which is smaller compared to the overall recovery fraction for $m_{g}$ $<$ 20 mag (i.e., 87%). Multiple factors may have contributed to this discrepancy, including environmental differences. The cluster environment plays a crucial role in shaping the brightness distribution of dEs, where brighter dEs are more likely to be formed (Geha et al., 2012) and they have higher recovery fractions. In this sample, a substantial portion of dEs are found within the cluster environment. Also, a potential memory bias in our calculation may have contributed to the high recovery fraction.

Based on this extensive data set, our results are summarized as follows:

1. 1.

   Systematic analysis of the derived structure parameters of dEs reveals that the dEs generally follow a universal relation between magnitude, size, and mean surface brightness regardless of their host environment.

2. 2.

   Using a standard definition, we identify 100 dEs in our sample that can be classified as UDGs. However, we find that UDGs do not occupy a special position in the parameter space of dEs to define them as a special class of objects. In addition, their structural parameters are not uniquely different from the overall dE population, being only one sigma outlying from the Log($R_{h}$) and $M_{g}$ relationship.

3. 3.

   We identify that only 40% of our sample dEs host a central nucleus, and among the UDG population, the vast majority, 79 out of 100, are non-nucleated. We find a mild radial dependence of nucleated fraction in both the Virgo and Fornax Clusters, and overall, the nucleated fraction declines as we go from the higher-density environment cluster to the low-density environment.

4. 4.

   We find that about 1.3%of dEs suffer a tidal disturbance from nearby massive galaxies and only 0.03%suffer recent dwarf-dwarf merging.

5. 5.

   We find that dEs are found more frequently in environments where massive galaxies are already quenched.

We hope that this catalog will serve as a valuable resource for follow-up studies on dEs. In particular, considering the laborious nature of visual inspection and classification, this data set will be useful for training and testing machine learning approaches. In future work, we plan to test the ability of convolutional neural networks to detect and characterize dEs with a view to applying them to a larger set of imaging data.