Detection of a Large Population of Ultra Diffuse Galaxies in Massive Galaxy Clusters: Abell S1063 and Abell 2744
aa r X i v : . [ a s t r o - ph . GA ] J un Draft version June 26, 2017
Preprint typeset using L A TEX style emulateapj v. 12/16/11
DETECTION OF A LARGE POPULATION OF ULTRA DIFFUSE GALAXIES IN MASSIVE GALAXYCLUSTERS: ABELL S1063 AND ABELL 2744
Myung Gyoon Lee , Jisu Kang , Jeong Hwan Lee , and In Sung Jang Astronomy Program, Department of Physics and Astronomy, Seoul National University, Gwanak-gu, Seoul 08826, Korea and Leibniz-Institut f¨ur Astrophysik Potsdam (AIP), An der Sternwarte 16, D-14482, Potsdam, Germany
Draft version June 26, 2017
ABSTRACTWe present the detection of a large population of ultra diffuse galaxies (UDGs) in two massive galaxyclusters, Abell S1063 at z = 0 .
348 and Abell 2744 at z = 0 . to 10 M ⊙ . Radial number density profiles of the UDGs show a turnoveror a flattening in the central region at r <
100 kpc. We estimate the total masses of the UDGs usingthe galaxy scaling relations. A majority of the UDGs have total masses, M = 10 to 10 M ⊙ , andonly a few of them have total masses, M = 10 to 10 M ⊙ . The total number of UDGs withinthe virial radius is estimated to be N(UDG)= 770 ±
114 for Abell S1063, and N(UDG)= 814 ± M > M ⊙ with a power law,N(UDG) = M . ± . . These results suggest that a majority of the UDGs have a dwarf galaxyorigin, while only a small number of the UDGs are massive L ∗ galaxies that failed to form a normalpopulation of stars. Subject headings: galaxies: clusters: individual (Abell S1063, Abell 2744) — galaxies: formation —galaxies: dwarf INTRODUCTIONUltra Diffuse Galaxies (UDGs) are a mysterious typeof galaxies with low surface brightness (LSB), whichhave larger sizes and fainter surface brightness thannormal galaxies with similar luminosity. Since the re-cent revival of the UDGs in Coma (van Dokkum et al.2015; Koda et al. 2015), the number of the known UDGskeeps increasing. The UDGs are found from the low-density regions to the high density environments suchas galaxy clusters (Mihos et al. 2015; Yagi et al. 2016;Smith Castelli et al. 2016; van der Burg et al. 2016;Martin et al. 2016; Merritt et al. 2016; Bellazzini et al.2017; Rom´an & Trujillo 2017a,b).The scenarios suggested to explain the origin ofthe UDGs can be divided roughly into two types.First, UDGs are massive L ∗ galaxies that failed toform a normal amount of stars given their dynam-ical mass (van Dokkum et al. 2015; Koda et al. 2015;van Dokkum et al. 2016). Second, they are dwarfgalaxies that were inflated due to some physical (dy-namical or thermal) processes (Yozin & Bekki 2015;Amorisco & Loeb 2016; Beasley et al. 2016; Peng & Lim2016; Di Cintio et al. 2017). While observational evi-dence is being accumulated, the nature of UDGs is notyet clear and whether the UDGs are failed L ∗ galaxies orinflated dwarf galaxies is still debated (Rom´an & Trujillo2017b; Zaritsky 2017).In this study, we search for UDGs in two massivegalaxy clusters, Abell S1063 and Abell 2744, which aremore massive than Coma. Because they are massive clus-ters, it is expected that they should contain a large num- [email protected] ber of UDGs and that these UDGs may be a good sam-ple to reveal the nature of the UDGs. They are part ofthe target galaxy clusters in the Hubble Frontier Fields(HFF) Program, for which deep HST images are avail-able (Lotz et al. 2017). Abell S1063 and Abell 2744 arelocated at the redshift, z = 0 .
348 and z = 0 . H = 73 km s − Mpc − , Ω M = 0 .
27, and Ω Λ = 0 .
73. Forthese parameters, luminosity distance moduli of AbellS1063 and Abell 2744 are ( m − M ) = 41 .
25 ( d = 1775Mpc) and 40.94 ( d = 1540 Mpc), and angular diame-ter distances are 978 Mpc and 901 Mpc, respectively.Corresponding image scales of the clusters are 4.744 kpcarcsec − and 4.370 kpc arcsec − , respectively.The virial radius and mass of Abell S1063 are R =8 ′ .
64 = 2 . M = 2 . +0 . − . × M ⊙ (Zenteno et al. 2016), respectively. Abell 2744 has acomplex structure so that it is not easy to determineits mass. We use the virial radius and mass of thecentral ‘ a ’ subcluster (called as the southern core) cov-ered by the HFF images given by Boschin et al. (2006): R = 9 ′ .
16 = 2 . M = 2 . +0 . − . × M ⊙ . Foreground reddening values toward these clus-ters (Schlafly & Finkbeiner 2011) are negligible: E ( B − V ) = 0 .
010 for Abell S1063 and E ( B − V ) = 0 .
012 forAbell 2744. Lee et al. DATA AND DATA REDUCTION2.1.
Data
We used ACS/F814W( I ) and WFC3/F105W( Y ) im-ages for Abell S1063 and Abell 2744 in the HFF(Lotz et al. 2017). The effective wavelengths of theF814W and F105W filters for the redshifts of Abell S1063and Abell 2744 (6220 ˚A and 8030 ˚A) correspond approx-imately to SDSS r ′ and Cousins I (or SDSS i ′ ) in therest-frame, respectively. This combination of filters is ef-ficient for the search of old stellar systems like UDGs andglobular clusters in these galaxy clusters (Lee & Jang2016). The HFF provides data for the central field ofeach cluster and the parallel fields at ∼ ′ . ∼ ′ . . ′′
03. The total exposuretimes are: T exp (F814W) = 116,169s and T exp (F105W) =67,341s for Abell S1063, T exp (F814W) = 106,998s and T exp (F105W) = 66,141s for the Abell S1063 parallelfield, T exp (F814W) = 104,270s and T exp (F105W) =68,952s for Abell 2744, and T exp (F814W) = 107,766s and T exp (F105W) = 67,329s for the Abell 2744 parallel field. Figure 1 (Upper panels) displays color images of theHST fields for Abell S1063 and Abell 2744.The full width at half-maximum (FWHM) values ofthe point sources in the images are ∼ . ′′ ∼
430 pc for AbellS1063 and ∼
390 pc for Abell 2744. Therefore the sourceslarger than these values can be detected as extendedsources in the images of the galaxy clusters. As a ref-erence for the background control, we used the data forthe Hubble Extreme Deep Field (HXDF) (RA(2000)=3 h m s .8, Dec(2000)= –27 ◦ ′ ′′ ) provided byIllingworth et al. (2013), as in the study of globular clus-ters, ultracompact dwarfs, and dwarf galaxies in Abell2744 by Lee & Jang (2016).2.2. Selection of UDGs
We searched for UDGs in the images of the targetfields, considering mainly the effective radii and sur-face brightness, and secondarily the colors, of the ex-tended sources. Our search procedure is similar tothose in Yagi et al. (2016); van der Burg et al. (2016),which consists mainly of two steps: the first basedon the application of SExtractor for source detection(Bertin & Arnouts 1996), and the second based on theapplication of GALFIT for parameter measurements(Peng et al. 2010).First, we ran SExtractor as a dual mode to detect ex-tended sources and derive their photometry in the im-ages. We used only the sources detected in both F814Wimages and F105W images. F814W images were usedas a reference image for the dual mode photometry sothat structural parameters of the detected sources arebased on the F814W images. We used SExtractor pa-rameter values: DETECT MINAREA = 20 pixels, SEE-ING FWHM = 0 . ′′
09, DEBLEND NTHRESH = 32, DE-BLEND MINCONT = 0.01, BACKPHOTO TYPE =LOCAL and BACK SIZE = 32 pixels. We chose top-hat filters which are optimized for the detection of ex-tended LSB objects, with a low detection threshold, DE-TECT THRESH = 0 . σ . For photometry of the sources we used PHOT AUTOPARAMS (Kron factor, minimumradius) = (2.5, 3.5) for magnitudes, and (2.5, 1.75) forcolors. We calibrated the instrumental magnitudes ofthese sources to the AB system, following the STScI web-page . Note that Lee & Jang (2016) adopted Vega mag-nitudes for their photometry of the sources in Abell 2744.AB magnitudes for F814W are 0.424 mag fainter thanVega magnitudes: F814W(AB) = F814W(Vega)+0 . < .
4, circularized effective radius R e , c , SE (FLUX RADIUS) > µ , F814W , SE (MU MAX) > . − , elon-gation parameter ( q = b/a where a and b are major andminor axes of the sources) q > . − . < F814W–F105W < .
0, and FLAGS <
4. Here we adopted a min-imum value R e , c , SE = 1 kpc, which is smaller than theUDG limits used in the literature. The effective radii andcentral surface brightness values of the sources providedby SExtractor show good correlations with the valuesprovided by GALFIT, but they show some differences,especially for multiple sources. Thus SExtractor valuesare used only for the initial selection of the UDG candi-date. The numbers of these candidates are 521 and 304for Abell S1063 and its parallel field, and 352 and 295for Abell 2744 and its parallel field, respectively.Second, we ran GALFIT to the images of the sourcesin the initial list of UDG candidates, deriving effectiveradii and surface brightness at the effective radii of thesources. Before fitting of the UDG candidate images, wemasked out neighboring sources around the target UDGsusing the list of the sources detected with SExtractor andthe segmentation maps. We derived point spread func-tions (PSFs) of the point sources in the images usingPSFEx (Bertin 2011), which are used as an input forGALFIT measurements. We performed the fitting of thesurface brightness profiles with S´ersic index n free. Cir-cularized effective radii ( R e , c ) were calculated from themajor axis effective radii ( R e ) and elongation parameter,with R e,c = R e √ q . We used these GALFIT values as thefinal values for the UDGs.Finally we inspected the images of these galaxies, re-moving sources with artifacts, multiple sources, sourcesclose to the frame edges, gravitational lens arcs, andbright galaxies. Most of the finally selected UDGs showsmooth structures, but some of them show substructuressuch as nuclei and disk features.For the comparison of the UDGs at different redshifts,we transform the measured surface brightness at z tothe value at the rest frame ( z = 0). The mean effectivesurface brightness for the redshift z ( h µ i e , z ( λ )) is givenin terms of the evolutionary correction E ( z ) and the K -correction K ( z ), so that the absolute mean effective sur-face brightness h µ i e , abs ( λ ) can be derived as follows ((Eq) DGs in Abell S1063 & Abell 2744 3
Abell S1063 (z=0.348) Abell 2744 (z=0.308)
F814W image GALFIT model residual F814W image GALFIT model residual
AS1063_UDG05 R e,c = 3.38 kpc< µ > e,abs (r‘) = 24.29n = 1.74b/a = 0.82 AS1063_UDG15 R e,c = 1.86 kpc< µ > e,abs (r‘) = 24.45n = 0.59b/a = 0.84 AS1063_LSBdw026 R e,c = 1.22 kpc< µ > e,abs (r‘) = 24.68n = 0.62b/a = 0.52 1"4.74 kpc A2744_UDG27 R e,c = 4.97 kpc< µ > e,abs (r‘) = 25.43n = 3.84b/a = 0.75 A2744_UDG20 R e,c = 1.53 kpc< µ > e,abs (r‘) = 25.15n = 1.18b/a = 0.48 A2744_LSBdw078 R e,c = 1.06 kpc< µ > e,abs (r‘) = 24.66n = 0.54b/a = 0.80 1"4.37 kpc Fig. 1.— (Upper panels) Color images of Abell S1063 (left) and Abell 2744 (right) HST fields. Red and yellow squares represent theposition of example UDGs and LSB dwarfs, respectively. (Lower panels) 4 . ′′ × . ′′ zoom-in images of two UDGs and one LSB dwarfin each cluster. Left, middle, and right sections show F814W images, GALFIT models, and residual images after subtracting GALFITmodels, respectively.
13 in Graham & Driver (2005)): h µ i e , abs ( λ ) = h µ i e , z ( λ ) − z ) − E ( z ) − K ( z ) . (1)We derived E ( z ) and K ( z ) of the sources usingGALAXEV for simple stellar populations with ages of12 Gyrs (Bruzual & Charlot 2003). The adopted val-ues of the parameters for GALAXEV are as follows: theChabrier stellar initial mass function (Chabrier 2003),[Fe/H] = − .
6, and the age of 8.3 Gyrs for Abell S1063and 8.7 Gyrs for Abell 2744. The values of E ( z ) and K ( z ) in the F814W band are –0.36 and +0.11 for AbellS1063, and –0.32 and +0.09 for Abell 2744, respectively.If we adopt the age of 10 Gyrs for z = 0, the ages ofAbell S1063 and Abell 2744 will be 6.3 Gyrs and 6.7 Gyrs, respectively, and the values of K ( z ) will be +0.06for Abell S1063, and +0.05 for Abell 2744. We convertedmeasured F814W magnitudes to SDSS r ′ -band using thevalues given by GALAXEV. Figure 2 displays R e , c versus h µ i e , abs ( r ′ ) of the galax-ies in Abell S1063, Abell 2744, the parallel fields, and theHXDF. The parameters for the galaxies in the HXDFwere derived for the redshifts of Abell S1063 and Abell2744, respectively. Most of the galaxies detected in theHXDF are much more distant than the two clusters.Thus the parameters for the HXDF are useful only as areference. We also plotted the data for the Coma UDGs(Yagi et al. 2016) for comparison.As the final UDGs (plotted by large red starlets), weselected the galaxies with R e , c > . h µ i e , abs ( r ′ ) > Lee et al. R e , c [ kp c ] (a) AS1063 -2 (b) AS1063 Parallel (c) HXDF (z=0.348)
26 24 22 20 181 R e , c [ kp c ] (d) A2744 -2
26 24 22 20 18< µ > e,abs (r ’ ) [mag arcsec -2 ] (e) A2744 Parallel
26 24 22 20 18 (f) HXDF
UDGs LSB Dws Bright Gs Non-UDGs Coma UDGs (Yagi+2016) UDGs LSB Dws Bright Gs Non-UDGs Coma UDGs (Yagi+2016) (z=0.308)
Fig. 2.—
Scaling relations of the galaxies in Abell S1063, Abell 2744, the parallel fields, and the HXDF, derived with GALFIT.Circularized effective radii ( R e , c ) versus absolute mean effective surface brightness ( h µ i e , abs ( r ′ ) ) transformed to SDSS r ′ -band fromF814W band: large red starlets for UDGs, small red starlets for LSB dwarfs, blue circles for bright galaxies with high central surfacebrightness ( µ , F814W , SE < . − ), and yellow triangles for non-UDGs (which were in the initial list of UDGs but wereexcluded later). The parameters for the HXDF galaxies are calculated for the redshifts of Abell S1063 and Abell 2744, respectively. Graydots represent the UDGs in Coma (Yagi et al. 2016). . − , and q > .
3, following the se-lection criteria adopted for the nearby galaxy clus-ters in van der Burg et al. (2016). van der Burg et al.(2016) adopted a criterion, h µ i e , z=0 . ( r ′ ) > . − for the nearby galaxies at the mean redshift of z = 0 . h µ i e , abs ( r ′ ) > . − for z = 0. These galaxies are often called MilkyWay (MW)-sized UDGs (Note that a slightly smallerlimit, R e , c > .
25 kpc (or R e > . R e , c = 1 . − h µ i e , abs ( r ′ ) = 23 . − . − . In the figure, blue circles represent thebright normal galaxies with high central surface bright-ness µ , F814W , SE < . − .In addition, we selected the LSB dwarf galaxies withthe same surface brightness range as, but smaller than,the UDGs: 1 . < R e , c < . h µ i e , abs ( r ′ ) > . − (plotted by small red starlets). Similargalaxies were included in the sample of Coma UDGsin the studies of Koda et al. (2015); Yagi et al. (2016).Yagi et al. (2016) called these galaxies as Subaru UDGs to distinguish from the MW-sized UDGs. The sampleof LSB dwarf galaxies is used for the comparison withUDGs in our target clusters. The locations of the newUDGs and LSB dwarfs in Abell S1063 and Abell 2744 in Figure 2 are consistent with those of the Coma UDGswith higher surface brightness and larger sizes.We marked, with yellow triangles in the figure, thegalaxies that were included as the initial UDG candidatesbased on the SExtractor criteria, but were excluded lateraccording to the GALFIT criteria. They are mostly lo-cated between the domain of the UDG plus LSB dwarfsand the domain of the bright galaxies. We call them asnon-UDGs hereafter.
Figure 1 (Lower panels) displays zoom-in images oftwo UDGs and one LSB dwarf in each cluster. The left,middle, and right columns for each cluster show F814Wimages, GALFIT model images, and the residual im-ages after GALFIT model subtraction, respectively. Thenumbers in the middle panels show the values of the fit-ting parameters given by GALFIT. RESULTS3.1.
Color-Magnitude Diagrams of the UDGs
DGs in Abell S1063 & Abell 2744 5 F W (a) AS1063 (b) AS1063 Parallel (c) HXDF -0.5 0.0 0.526242220 F W (d) A2744 -0.5 0.0 0.5F814W - F105W (e) A2744 Parallel -0.5 0.0 0.5 (f) HXDF UDGs LSB Dws Bright Gs Non-UDGs Compact Gs UDGs LSB Dws Bright Gs Non-UDGs Compact Gs
Fig. 3.—
CMDs of the UDGs and other galaxies in Abell S1063, Abell 2744, the parallel fields, and the HXDF. Symbols are same as in
Figure 2 , except for black dots denoting the compact galaxies with R e , c , SE < In Figure 3 we display the color-magnitude diagrams(CMDs) of the UDG candidates as well as other galaxiesin Abell S1063, Abell 2744, the parallel fields, and theHXDF.Following features are noted in this figure. First, bothclusters show a strong red sequence, while the HXDFdoes not. Second, the red sequences in the parallel fieldare much weaker than those in the cluster field, showingthat the contributions due to background galaxies arelarger in the parallel fields. Third, no UDG candidatesare found in the HXDF, while a small number of UDGcandidates are seen in the parallel fields. This indicatesthat the UDG candicates in the parallel fields are clus-ter memebers. In addition, a much more number of LSBdwarf candidates are seen in the parallel fields than inthe HXDF, showing that most of them are cluster mem-bers. Fourth, the UDGs and LSB dwarfs are located atthe faint end of the red sequence, showing that they aremainly made of old stars. A small number of the faintUDGs and LSB dwarfs in both clusters are bluer thanthe red sequence, indicating that they have some youngstellar populations. Fifth, the UDGs are fainter thanthe bright normal galaxies in the red sequence, and theF814W magnitude of the UDGs ranges from 24.0 to 28.0mag. Thus most of the UDGs and LSB dwarfs are old stel-lar systems, while some show a recent activity of starformation. It is noted that a few UDG-like galaxies inthe cluster and parallel fields are redder than the red se-quence (larger gray starlets). Similarly a small number ofLSB dwarfs are also redder than the red sequence (smallgray starlets). They are probably background sources.Therefore we excluded these galaxies in the final list ofUDGs and LSB dwarfs.3.2.
The Census of the UDGs
Finally we select 47 UDGs in Abell S1063 (35 UDGsin the cluster and 12 UDGs in the parallel field) and40 UDGs in Abell 2744 (27 UDGs in the cluster and 13UDGs in the parallel field). If we adopt the age of 10Gyrs for z = 0, these numbers will be slightly changed:53 UDGs in Abell S1063 (40 UDGs in the cluster and 13UDGs in the parallel field) and 42 UDGs in Abell 2744(29 UDGs in the cluster and 13 UDGs in the parallelfield). Thus the numbers of UDGs in these two clus-ters are similar, which is consistent with the expectationbased on the similarity in the cluster mass. We use theseUDGs for the following analysis.Also we select 96 and 47 LSB dwarf candidates in AbellS1063 and its parallel field (143 in total), and 62 and 31 Lee et al. S e r s i c n (a) AS1063 UDGs LSB Dws Bright Gs Non-UDGs Coma UDGs (Yagi+2016) UDGs LSB Dws Bright Gs Non-UDGs Coma UDGs (Yagi+2016) (b) AS1063 Parallel ± ± (c)
26 24 22 20 180.11.0 S e r s i c n < µ > e,abs (r ’ ) [mag arcsec -2 ] (d) A2744
26 24 22 20 18 (e) A2744 Parallel ± ± (f) Fig. 4.— (Left and middle panels) S´ersic index n versus h µ i e , abs ( r ′ ) for the UDGs and other galaxies in Abell S1063, Abell 2744, andthe parallel fields. Blue circles denote only the bright galaxies in the red sequence. Gray dots denote the Coma UDGs (Yagi et al. 2016).(Right panels) Red line histograms are for the UDGs in Abell S1063 and Abell 2744, and gray line histograms are for the MW-sized UDGsin Coma. The numbers and dotted lines represent mean values of S´ersic index n for UDGs. such sources in Abell 2744 and its parallel field (93 intotal), respectively. On the other hand, we see 6 and 4LSB dwarf-like galaxies in the HXDF for the redshiftsof Abell S1063 and Abell 2744, respectively. The arearatios of the cluster and parallel fields with respect tothat of the HXDF are 0.990 for Abell S1063, 0.989 forits parallel field, 1.288 for Abell 2744, and 0.986 for itsparallel field, respectively. Thus the net numbers of theLSB dwarf galaxies in Abell S1063 and its parallel fieldare, respectively, 90 and 41 (131 in total), subtractingthe contribution of the background sources based on theHXDF. Similarly we derive the net numbers of the LSBdwarf galaxies in Abell 2744 and its parallel field are,respectively, 57 and 26 (83 in total).Tables 1 and 2 list the catalogs of UDGs and LSBdwarfs in Abell S1063, and Tables 3 and 4 list the cat-alogs of UDGs and LSB dwarfs in Abell 2744, respec-tively. In the tables we include the following informa-tion: IDs, RA(J2000), Dec.(J2000), major axis effectiveradii, effective surface brightness, F814W magnitudes,(F814W–F105W) colors, S´ersic index n , elongation pa-rameter, circularized effective radii, and absolute meaneffective surface brightness of the galaxies.3.3. Structural Parameters of the UDGs
We compare structural parameters of the UDGs inAbell S1063 and Abell 2744 with those in Coma(Yagi et al. 2016). In
Figure 4 we plotted the S´ersicindex n versus h µ i e , abs ( r ′ ) of the UDGs (large red star-lets) and LSB dwarfs (small red starlets) in Abell S1063and Abell 2744 and their parallel fields and the ComaUDGs (gray dots). We also plotted the bright normalgalaxies in the red sequence of Abell S1063 and Abell2744 (blue circles) and the non-UDGs (yellow triangles).The bright red sequence galaxies are mostly the mem-bers of each cluster and can be considered to be simplestellar populations of old age. In the right panels of thefigure we show the histograms of S´ersic index n for theUDGs in Abell S1063 and Abell 2744 in comparison withthose of the Coma UDGs. We selected only the MW-sized UDGs in the Coma UDG sample for comparison ofthe histograms.It is seen that the values of n for most UDGs in AbellS1063 and Abell 2744 are smaller than three, while someof the bright red sequence galaxies have n >
3. Thedistributions of n in these two clusters are similar tothose of the MW-sized UDGs in Coma. The mean values, h n i = 1 . ± .
63 for Abell S1063 and h n i = 0 . ± .
65 forAbell 2744, are similar to that of the Coma UDGs, h n i =1 . ± .
37. Thus the radial surface brightness profiles ofDGs in Abell S1063 & Abell 2744 7 q (a) AS1063 (b) AS1063 Parallel ± ± (c)
26 24 22 20 180.20.40.60.81.0 q < µ > e,abs (r ’ ) [mag arcsec -2 ] (d) A2744
26 24 22 20 18 (e) A2744 Parallel ± ± (f) Fig. 5.—
Elongation parameter ( q = b/a ) versus h µ i e , abs ( r ′ ) (left and middle panels) and histograms of elongation parameter (rightpanels) for the UDGs and other galaxies in Abell S1063, Abell 2744, and the parallel fields. Symbols are same as in Figure 4 . the UDGs in Abell S1063 and Abell 2744 are mostly fitby an exponential profile, similar to the case of UDGs inComa and other nearby galaxy clusters (Yagi et al. 2016;Rom´an & Trujillo 2017b; van der Burg et al. 2016). Figure 5 displays the elongation parameter q versus h µ i e , abs ( r ′ ) (left and middle panels), of the same galax-ies as before, and the histograms of elongation parame-ters for the same galaxies as in Figure 4 . Most of theUDGs in Abell S1063 and Abell 2744 have the elonga-tion parameters larger than 0.4. The mean values of theelongation parameters for the UDGs in Abell S1063 andAbell 2744 are h q i = 0 . ± .
16 and 0 . ± .
17, whichare similar to the value of the MW-sized Coma UDGs, h q i = 0 . ± .
15. Structural parameters ( n and q ) of theLSB dwarfs in Abell S1063 and Abell 2744 show similarfeatures to those of the UDGs.3.4. Stellar Masses of the UDGs In Figure 6 we plotted the effective radii versus M r ′ magnitudes of the galaxies in the same fields as before.We derived a relation between the stellar mass and M r ′ magnitudes for simple stellar populations, for the sameGALAXEV parameters as described in Section 2. Weoverlayed the values for corresponding stellar masses inthe upper axis of each panel. It shows that the absolutemagnitudes of most UDGs in Abell S1063 and Abell 2744range from M r ′ = − . − . M ∗ = 10 to 10 M ⊙ . TheLSB dwarfs in Abell S1063 and Abell 2744 cover lowerranges than the UDGs: M r ′ = − . − . M ∗ = 5 × to 5 × M ⊙ . The slanted solid linesrepresent the varying surface brightness, h µ i e , abs ( r ′ ) =23, 24, 25, 26, and 27 mag arcsec − (from left to right).For given surface brightness, the larger the UDGs are,the brighter (more massive) they are. The stellar massesof the UDGs are much lower than those of the bright redsequence galaxies.In the right panels of Figure 6 we plotted the his-tograms of effective radii for Abell S1063 and Abell 2744in comparison with that of the Coma UDGs (Yagi et al.2016). The mean values, h R e , c i = 1 . ± .
43 kpc forAbell S1063 and h R e , c i = 1 . ± .
64 kpc for Abell 2744,are similar to that of the MW-sized UDGs in Coma, h R e , c i = 2 . ± .
48 kpc. The histograms of the com-bined sample of UDGs and LSB dwarfs are similar tothose of the large UDGs in Coma.3.5.
Spatial Distributions of the UDGs
Figure 7 (left panels) displays the spatial distribu-tion of the UDGs, LSB dwarf galaxies, and bright redsequence galaxies in Abell S1063 and Abell 2744 fields.The circles in the left panels represent the boundaries ofthe radial bins used for deriving radial number density Lee et al. R e , c [ kp c ] (a) AS1063 M * [M O • ]
23 25 m a g a r c s ec - (b) AS1063 Parallel ± ± (c) -20 -151 R e , c [ kp c ] (d) A2744 -20 -15M r ’ [mag]
23 25 m a g a r c s ec - (e) A2744 Parallel ± ± (f) Fig. 6.— (Left and middle panels) Circularized effective radii versus r ′ -band absolute magnitudes (lower X-axes) and stellar masses(upper X-axes) of the UDGs and other galaxies in Abell S1063, Abell 2744, and the parallel fields. Symbols are same as in Figure 5 . Solidlines represent iso-surface brightness magnitudes of h µ i e , abs ( r ′ ) = 23, 24, 25, 26, and 27 mag arcsec − from left to right. (Right panels)Red line histograms are for the UDGs and LSB dwarfs in Abell S1063 and Abell 2744, and gray line histograms for the UDGs in Coma.The numbers and dotted lines represent mean values of S´ersic index n for UDGs. profiles. We excluded the central circular region in theUDG survey, because of high surface brightness of thecentral galaxy.We derived the radial number density profiles of thegalaxies, subtracting the background contribution usingthe data for the HXDF. As the center for the radial pro-files we adopted the center of the cD galaxy in AbellS1063, and the center of CN-1 in Abell 2744. Note thatLee & Jang (2016) adopted the center of CN-2 at thesouth-east of CN-1, which is the brightest galaxy in theHST field of Abell 2744, for deriving radial number den-sity profiles of the point sources (mostly globular clustersand ultra compact dwarfs). In this study of UDGs, wechose CN-1, which is closer to the center of the southerncore in Abell 2744. We plotted the results in the rightpanels of the figure.We estimated the completeness of our UDG and LSBdwarf detection using a mock galaxy experiment. Wegenerated images of mock galaxies with structural pa-rameters similar to those of the UDGs and LSB dwarfsonto the original images, using IRAF/ARTDATA. Thenwe repeated the same search procedure as used for UDGand LSB dwarf detection. Finally we derived a radialprofile of completeness (recovery fraction) as a func- tion of clustercentric distance for each cluster, as shownin Figure 8 for two magnitude ranges: h µ i e , abs ( r ′ ) =23 . − . . − . − . We correctedthe radial number density profiles of the sources for com-pleteness using this result.A few interesting features are distinguishable in Fig-ure 7 . First, the radial number density profiles of theUDGs, LSB dwarfs, and bright red sequence galaxies inboth clusters keep increasing as the clustercentric dis-tance decreases in the outer region at 100 kpc < r < r <
100 kpc, while thatof the bright galaxies keeps increasing in the central re-gion. Thus the relative number density of the UDGsplus LSB dwarfs with respect to that of the bright galax-ies is relatively lower in the central region than in theouter region. Third, spatial distributions of the UDGsare more inhomogeneous than those of the bright red se-quence galaxies. This is a very interesting feature, whichDGs in Abell S1063 & Abell 2744 9 -101 (a) ∆ D ec . [ a r c m i n ] AS1063 110 N / a r c m i n Bright Gs UDGs+LSB Dws UDGs LSB Dws ∆ R.A. [arcmin]-101 (c) ∆ D ec . [ a r c m i n ] A2744 110 N / a r c m i n Fig. 7.—
Spatial distributions (left panels) and radial number density profiles corrected for detection completeness (right panels) of theUDGs (large red starlets), LSB dwarfs (small red starlets), and bright red sequence galaxies (blue circles) in Abell S1063 and Abell 2744.Black dots denote the compact galaxies with R e , c , SE < may be related with the origin of the UDGs. Inhomo-geneous distributions of the UDGs indicate that UDGsare relatively new comers that came from outside thegalaxy clusters so their distribution is not yet dynami-cally virialized. Further studies with a larger sample ofgalaxy clusters or with simulations of UDGs are neededto investigate this issue. DISCUSSION4.1.
Total Masses of the UDGs
The dynamical mass of the UDGs provides a criticalinformation to understand the origin of the UDGs. How-ever, it is difficult to determine the dynamical mass of theLSB galaxies like UDGs so dynamical estimates of theUDG mass are available only for a few UDGs: Dragon-fly 44 in Coma (van Dokkum et al. 2016), VCC 1287 inVirgo (Beasley et al. 2016), and UGC 2162, the nearestUDG in the M77 group (Trujillo et al. 2017).Zaritsky et al. (2008) suggested a new method to es-timate the total mass of stellar systems without thekinematic measurements, which is based on the funda-mental manifold. A kinematic term of a stellar system( V = p σ v + v r / σ v is the line-of-sight velocitydispersion and v r is the rotational velocity) can be es- R ec ov e r y fr ac ti on (a) AS1063 < µ > e,abs (r ’ ) = 23.5-25.0< µ > e,abs (r ’ ) = 23.5-25.0 < µ > e,abs (r ’ ) = 23.5-25.0< µ > e,abs (r ’ ) = 23.5-25.0 (b) A2744 Fig. 8.—
Completeness of galaxy detection with respect to pro-jected clustercentric distance for Abell S1063 (a) and Abell 2744(b). Yellow and green lines are for h µ i e , abs ( r ′ ) = 23.5 – 25.0 magarcsec − , and h µ i e , abs ( r ′ ) = 25.0 – 26.5 mag arcsec − , respec-tively. timated, if the values of surface brightness and effectiveradii are known in the fundamental manifold.We estimate the total mass of the UDGs in AbellS1063 and Abell 2744, following this method thatZaritsky (2017) applied to the MW-sized UDGs in Coma(van Dokkum et al. 2015) and the dwarf galaxies in For-nax clusters (Mu˜noz et al. 2015). The galaxy scaling re-0 Lee et al.lations are given in terms of V , the mass-to-light ratio,Υ e , and the mean surface brightness, I e , within the ef-fective radius, R e , c (Zaritsky et al. 2008; Zaritsky 2017):logΥ e = 0 . V ) + 0 . I e ) − . V − . I e − . V I e + 1 .
49 (2)and log R e , c = 2log V − log I e − logΥ e − . . (3)Combining these two equations leads to log V = f ( I e , R e , c ) . Thus we can derive the values of V and Υ e from I e and R e , c . Here we assume v r = 0 and V = σ v for velocity dispersion-supported systems.Then we estimate the enclosed mass within the 3Dhalf-light radius R / , using the formula for veloc-ity dispersion-supported galaxies given by Wolf et al.(2010), M ( < R / ) = 4 σ v R e , c /G = 930( σ v / km s − ) ( R e , c / pc) , (4)where R / is related with the 2D effective radius by R / = 4 R e , c /
3. Finally the total mass (virial mass)of the galaxies can be estimated from a comparisonof the enclosed mass with the NFW mass profiles forgiven concentration parameters (Navarro et al. 1997;Ludlow et al. 2016).
Figure 9 displays the enclosed mass versus the 3Dhalf-light radii ( R / ) of the UDGs, LSB dwarfs, andbright red sequence galaxies in Abell S1063 and Abell2744. For comparison we also plotted the results for theUDGs in Coma (Yagi et al. 2016) and the dwarf galax-ies with h µ i e ( r ′ ) > . − in Fornax selectedfrom the catalog given by Mu˜noz et al. (2015). Most ofthe Fornax dwarf galaxies in Mu˜noz et al. (2015) haveeffective radii smaller than 1 kpc and only a small num-ber of them have effective radii larger than 1.5 kpc. Wealso plotted the data for Dragonfly 44, VCC 1287, andUGC 2162. Trujillo et al. (2017) provided an enclosedmass within 5 kpc for UGC 2162. We multiplied it by afactor of 0.3 (the ratio of the radii = 1 . /
5) to estimatethe enclosed mass for R e , c = 1 . M = 10 ,10 , 10 , and 10 M ⊙ (with concentration parameter c =12.5, 10.6, 8.7, and 6.9 in Ludlow et al. (2016)) aredisplayed by solid curved lines. The distributions of theenclosed mass of UDGs and LSB dwarfs in Abell S1063,Abell 2744, and Coma are shown in the right panel ofthe figure.A few distinguishable features are noted in this fig-ure. First, the distributions of the enclosed mass of theUDGs in Abell S1063 and Abell 2744 are similar, andthey are overlapped with the high mass part of the ComaUDGs. The enclosed masses of most UDGs range from M ( < R / ) = 6 × M ⊙ to 3 × M ⊙ , and three largestUDGs have much higher masses, 6 × M ⊙ to 10 M ⊙ .Second, the larger the UDGs are, the higher enclosedmass they have. Third, most of the UDGs in Abell S1063and Abell 2744 have total masses M = 10 − M ⊙ .The LSB dwarfs have slightly lower masses, M = 1 − × M ⊙ . Fourth, only nine of the UDGs in these clus-ters have total masses larger than M = 10 M ⊙ . Sixof them have total masses and sizes similar to VCC 1287, [kpc]10 M ( < R / ) [ M O • ] (a) (b)5 ~ Bright galaxies Fornax dws (Munoz+15)Coma UDGs (Yagi+16)AS1063 UDGsA2744 UDGsDF44 (van Dokkum+16)VCC 1287 (Beasley+16)UGC 2162 (Trujillo+17) M [ M O • ] = R e,c = 1.5 kpcR = 2.0 kpc Fig. 9.— (a) Enclosed mass M ( < R / ) versus 3D effective radii R / of the UDGs (large starlets) and LSB dwarfs (small starlets)in Abell S1063 (red) and Abell 2744 (violet) in comparison withFornax dwarfs (turquoise squares) and Coma UDGs (gray circles).Dragonfly 44 and VCC 1287 are marked by a green triangle and agreen pentagon, respectively (based on the observed values in thereferences (van Dokkum et al. 2016; Beasley et al. 2016)). UGC2162, the nearest UDG, (Trujillo et al. 2017) is also plotted by ayellow circle. Note that R / denotes 3D half-light radii, givenby R / = 4 R e , c /
3. Solid lines represent the NFW profiles forthe total mass M = 10 , , and 10 M ⊙ . Blue circlesdenote the bright red sequence galaxies in Abell S1063 and Abell2744. (b) Distributions of the enclosed mass of the UDGs and LSBdwarfs in Abell S1063 (red), Abell 2744 (violet), and Coma (gray). and three of them (AS1063 UDG05, A2744 UDG27, andA2744 UDG04 (located in the parallel field)) have totalmasses and sizes similar to those of Dragonfly 44 that isan proto-example of massive UDGs in Coma.4.2. Total Numbers of UDGs and the Masses of theirHost Systems
From the study of UDGs in nearby galaxy clusters,van der Burg et al. (2016) found that the total numberof UDGs in nearby galaxy clusters shows a strong corre-lation with the virial mass ( M ) of their host clusters.They fit the data with a power law, N(UDG) = M α ,obtaining α = 0 . ± .
16. This correlation is also seen inthe expanded sample including other galaxy clusters andgroups (Rom´an & Trujillo 2017a; Janssens et al. 2017).From the numbers of the detected UDGs, we estimatethe total numbers of UDGs in Abell S1063 and Abell2744. We counted the number of UDGs within the virialradius of each galaxy cluster. Considering the radialnumber density profiles corrected for completeness foreach galaxy cluster, we derive N(UDG)= 770 ±
114 forAbell S1063, and N(UDG)= 814 ±
122 for Abell 2744.In
Figure 10 we plotted these results in compari-son with the previous results in the literature compiledby Rom´an & Trujillo (2017a); Janssens et al. (2017):nearby galaxy clusters (van der Burg et al. 2016), Coma(Yagi et al. 2016), Fornax (Mu˜noz et al. 2015), Abell 168DGs in Abell S1063 & Abell 2744 11 M [M O • ]1101001000 N ( UDG ) N(UDG) ∝ M ± N(UDG) ∝ M ± AS1063 (This study)A2744 (This study)A2744 (Janssens+17)Nearby clusters (vdB+16)Coma (Yagi+16)Fornax (Munoz+15)A168 and UGC842 (RT17b)Nearby groups (RT17a) ~ Fig. 10.—
Total numbers of UDGs vs total mass of their parentgalaxy clusters for Abell S1063 (red starlet), Abell 2744 (violetstarlet) and other clusters in the literature (Mu˜noz et al. 2015;Yagi et al. 2016; van der Burg et al. 2016; Janssens et al. 2017;Rom´an & Trujillo 2017a,b). Dashed line and solid line representthe power law fitting for all data and for massive systems with M > M ⊙ , respectively. For fitting we used the values inthis study in the case of Abell 2744. and UGC 842 (Rom´an & Trujillo 2017b), three Hicksoncompact groups (Rom´an & Trujillo 2017a), and Abell2744 (Janssens et al. 2017). Note that we adopted themass of Abell 2744, M = 2 . +0 . − . × M ⊙ , as de-scribed in Section 1, which is smaller than the value M = 5 × M ⊙ used in Janssens et al. (2017). AbellS1063 and Abell 2744 are the most massive clusters inthis sample so that they are precious targets to extendthe range of virial mass in the study of this relation.In the figure, the data for the UDGs in Abell S1063and Abell 2744 in this study appear to be located atthe upper end of the previous data. Fitting the datafor all UDGs with a power law (we use the values inthis study in the case of Abell 2744), we obtain α =0 . ± .
07. This value for the power law index is verysimilar to that given by Rom´an & Trujillo (2017a), α =0 . ± .
05, and is consistent with the value presentedby van der Burg et al. (2016), α = 0 . ± .
16. Thus thetwo clusters in this study follow well the power law givenby lower mass systems. It is interesting that the datafor the galaxy group mass of M = 10 to the massivecluster mass of 3 × M ⊙ are represented remarkablywell by the power law.The derived value of the power law index, 0 . ± . α = 0 . ± .
05 derivedin their study, Rom´an & Trujillo (2017a) suggested a sce-nario that UDGs are dwarf galaxies and that the progen-itors of today’s UDGs are formed in the low density field,are processed in galaxy groups, and then some of themare disrupted during the infall to galaxy clusters. However, the data for the low mass end have muchlarger errors than those for the high mass systems, be-cause the numbers of UDGs in the Hickson compactgroups presented by Rom´an & Trujillo (2017a) are small.If we fit the data for massive systems with M > M ⊙ in the sample, we derive α = 1 . ± .
09. Thisvalue is close to one, implying that the formation (orsurvival) efficiency of UDGs depends little on the theirhost mass for M > M ⊙ . It is needed to studymore UDGs in the galaxy groups with M < M ⊙ to investigate the flattening of the slope in the low massrange.It is noted that the data for Abell 2744 given byJanssens et al. (2017) shows a slightly larger deviationfrom the power law fit than our result, if the same valuefor the cluste mass is adopted. If it is real, Abell 2744 hasan excess of UDGs compared with lower mass systems.However, the total number of UDGs in Abell 2744 de-rived in this study, N = 814 ± N = 1961 ± R e , c > h µ i e , abs ( r ′ ) = 26.3mag arcsec − . We had also similar sources in our initiallist of UDG candiates, but most of them were removedin the visual inspection step. Therefore about a half ofthe difference between the two studies may be due to thedifference in the detected numbers, and another half maybe due to the difference in the total number estimationprocedure. 4.3. Origin of UDGs
Here we discuss the primary results of the UDGs inAbell S1063 and Abell 2744 derived in the previous sec-tions, in relation to the scenarios for the origin of theUDGs: an inflated dwarf galaxy scenario (Yozin & Bekki2015; Amorisco & Loeb 2016; Beasley et al. 2016;Rom´an & Trujillo 2017b; Di Cintio et al. 2017) anda failed L ∗ galaxy scenario(van Dokkum et al. 2015;Koda et al. 2015; van Dokkum et al. 2016). Radial Number Density Profiles of the UDGs : Wefound two interesting results on the radial distributions ofthe UDGs in Abell S1063 and Abell 2744: a) a similaritybetween the radial number density profiles of the UDGs,the LSB dwarfs, and the bright red sequence galaxies inthe outer region at 100 kpc < r < r <
100 kpc) than in the outer region (at r >
100 kpc)of the galaxy clusters. These results imply the followingpoints. First, the UDGs and the LSB dwarfs are lessmassive than the bright red sequence galaxies so thatthey cannot survive as long as the bright galaxies in thecentral region of the cluster. Second, the UDGs and theLSB dwarfs are vulnerable to the harsh environments inthe galaxy clusters so a significant fraction of them aredisrupted in the central region of the clusters.2 Lee et al.
Total Masses of the UDGs:
A small number of theUDGs in Abell S1063 and Abell 2744 have high totalmasses exceeding M = 10 M ⊙ . Six of them havetotal masses and sizes similar to those of VCC 1287,but three of them have total masses and sizes similarto those of Dragonfly 44 in Coma. Dragonfly 44 (with M = 8 × M ⊙ van Dokkum et al. (2016)) is knownto be one of the most massive UDGs in the local uni-verse, being used as an example for UDGs with the L ∗ galaxy origin. Thus the most massive UDGs in thesegalaxy clusters can be failed L ∗ galaxies. On the otherhand, a majority of the UDGs in Abell S1063 and Abell2744 have total masses smaller than M = 10 M ⊙ .However, they are more massive than M = 10 M ⊙ .Thus they correspond to the upper end in the mass dis-tribution of Coma UDGs (see Figure 9 ). This result isconsistent with the inference based on the radial numberdensity distribution in the previous section.From these results, we conclude that a majority of theUDGs in Abell S1063 and Abell 2744 are relatively mas-sive dwarf galaxies, supporting the dwarf galaxy originhypothesis. A small number of the UDGs in these galaxyclusters have masses and sizes similar to Dragonfly 44,which is consistent with the failed L ∗ scenario. SUMMARY AND CONCLUSIONAnalysing deep HST F814W and F105W images in theHFF, we discovered a large population of the UDGs intwo massive galaxy clusters, Abell S1063 and Abell 2744.We adopted the UDG selection criteria consistent withthose used for nearby galaxy clusters: R e , c > . h µ i e , abs ( r ′ ) > . − , q > .
3, and (F814W–F105W) colors bluer than the red boundary of the redsequence. For comparison, we also selected LSB dwarfswith 1 kpc < R e , c < . h n i =1 . ± .
63 for Abell S1063 and h n i = 0 . ± .
65 for Abell 2744. These are similar to those of ComaUDGs, h n i = 1 . ± . h q i = 0 . ± .
16 for Abell S1063and h q i = 0 . ± .
17 for Abell 2744, which aresimilar to those of Coma UDGs, h q i = 0 . ± . to 10 M ⊙ .6. The radial number density profiles of the UDGsand LSB dwarfs in both clusters show a drop ora flattening in the central region at r <
100 kpc,while that of the bright galaxies keeps increasingin the central region.7. We estimate the enclosed masses within effectiveradius of the UDGs using the galaxy scaling re-lations, finding that the enclosed masses of mostUDGs range from M ( < R / ) = 6 × M ⊙ to3 × M ⊙ , and three largest UDGs have muchhigher masses, 6 × M ⊙ to 10 M ⊙ . From thiswe find a majority of the UDGs have total masses, M = 10 to 10 M ⊙ , and only a few of themhave total masses, M = 10 to 10 M ⊙ .8. The total number of UDGs within the virial ra-dius of each cluster is estimated to be N(UDG)=770 ±
114 for Abell S1063, and N(UDG)= 814 ± M . ± . .However, if we fit the data for massive systems with M > M ⊙ , we obtain N(UDG) = M . ± . .This value of the power law index is close to one,implying that the efficiency of UDGs depends littleon the mass of their host systems.9. We conclude on the origin of the UDGs: A majorityof the UDGs in Abell S1063 and Abell 2744 haverelatively massive dwarf galaxies. Only a smallnumber of the UDGs can be massive enough tobe failed L ∗ galaxies.This work was supported by the National ResearchFoundation of Korea (NRF) grant funded by the KoreanGovernment (MSIP) (No. 2012R1A4A1028713). J.K.was supported by the Global Ph.D. Fellowship Program(NRF-2016H1A2A1907015). REFERENCESAmorisco, N. C., & Loeb, A. 2016, MNRAS, 459, L51Beasley, M. A., Romanowsky, A. J., Pota, V., et al. 2016, ApJ,819, L20Bellazzini, M., Belokurov, V., Magrini, L., et al. 2017, MNRAS,tmp238BBertin, E., & Arnouts, S. 1996, A&AS, 117, 393Bertin, E. 2011, Astronomical Data Analysis Software andSystems XX, 442, 435Boschin, W., Girardi, M., Spolaor, M., & Barrena, R. 2006, A&A,449, 461Bruzual, G., & Charlot, S. 2003, MNRAS, 344, 1000Chabrier, G. 2003, PASP, 115, 763Di Cintio, A., Brook, C. B., Dutton, A. A., et al. 2017, MNRAS,466, L1Graham, A. W., & Driver, S. P. 2005, PASA, 22, 118 Illingworth, G. D., Magee, D., Oesch, P. A., et al. 2013, ApJS,209, 6Janssens, S., Abraham, R., Brodie, J., et al. 2017, ApJ, 839, L17Koda, J., Yagi, M., Yamanoi, H., & Komiyama, Y. 2015, ApJ,807, L2Lee, M. G., & Jang, I. S. 2016, ApJ, 831, 108Lotz, J. M., Koekemoer, A., Coe, D., et al. 2017, ApJ, 837, 97Ludlow, A. D., Bose, S., Angulo, R. E., et al. 2016, MNRAS, 460,1214Martin, N. F., Ibata, R. A., Lewis, G. F., et al. 2016, ApJ, 833,167Merritt, A., van Dokkum, P., Danieli, S., et al. 2016, ApJ, 833,168Mihos, J. C., Durrell, P. R., Ferrarese, L., et al. 2015, ApJ, 809,L21
DGs in Abell S1063 & Abell 2744 13
Mu˜noz, R. P., Eigenthaler, P., Puzia, T. H., et al. 2015, ApJ, 813,L15Navarro J. F., Frenk C. S., White S. D. M. 1997, ApJ, 490, 493Peng, C. Y., Ho, L. C., Impey, C. D., & Rix, H.-W. 2010, AJ,139, 2097Peng, E. W., & Lim, S. 2016, ApJ, 822, L31Rom´an, J., & Trujillo, I. 2017a, MNRAS, 468, 4039Rom´an, J., & Trujillo, I. 2017b, MNRAS, 468, 703Schlafly, E. F., & Finkbeiner, D. P. 2011, ApJ, 737, 103S´ersic, J. L. 1963, Boletin de la Asociacion Argentina deAstronomia La Plata Argentina, 6, 41Smith Castelli, A. V., Faifer, F. R., & Escudero, C. G. 2016,A&A, 596, A23Trujillo, I., Roman, J., Filho, M., & S´anchez Almeida, J. 2017,ApJ, 836, 191 van der Burg, R. F. J., Muzzin, A., & Hoekstra, H. 2016, A&A,590, A20van Dokkum, P. G., Abraham, R., Merritt, A., et al. 2015, ApJ,798, L45van Dokkum, P., Abraham, R., Brodie, J., et al. 2016, ApJ, 828,L6Wolf, J., Martinez, G. D., Bullock, J. S., et al. 2010, MNRAS,406, 1220Yagi, M., Koda, J., Komiyama, Y., & Yamanoi, H. 2016, ApJS,225, 11Yozin, C., & Bekki, K. 2015, MNRAS, 452, 937Zaritsky, D., Zabludoff, A. I., & Gonzalez, A. H. 2008, ApJ, 682,68-80Zaritsky, D. 2017, MNRAS, 464, L110Zenteno, A., Mohr, J. J., Desai, S., et al. 2016, MNRAS, 462, 830 L eee t a l. TABLE 1A Catalog of UDGs in Abell S1063
ID R.A. Dec. R e a µ e , F814W F W F W − F W n b/a R e , c b h µ i e , abs ( r ′ ) c (J2000) (J2000) [kpc] [mag arcsec − ] [mag] [mag] [kpc] [mag arcsec − ]AS1063 UDG001 342.16211 -44.54686 2 . ± .
09 25 . ± .
06 25 . ± .
02 0 . ± .
02 0 . ± .
07 0 . ± .
02 1 . ± .
07 24 . ± . . ± .
14 26 . ± .
12 26 . ± .
03 0 . ± .
02 1 . ± .
21 0 . ± .
06 1 . ± .
14 24 . ± . . ± .
16 25 . ± .
14 25 . ± .
02 0 . ± .
01 1 . ± .
21 0 . ± .
02 1 . ± .
13 24 . ± . . ± .
21 25 . ± .
32 25 . ± .
01 0 . ± .
01 1 . ± .
33 0 . ± .
07 1 . ± .
19 24 . ± . . ± .
28 25 . ± .
15 24 . ± .
01 0 . ± .
01 1 . ± .
15 0 . ± .
02 3 . ± .
26 24 . ± . . ± .
15 25 . ± .
08 26 . ± .
03 0 . ± .
02 0 . ± .
14 0 . ± .
02 1 . ± .
10 24 . ± . . ± .
08 25 . ± .
06 25 . ± .
02 0 . ± .
01 0 . ± .
07 0 . ± .
02 1 . ± .
15 24 . ± . . ± .
80 27 . ± .
64 26 . ± .
05 0 . ± .
03 2 . ± .
82 0 . ± .
08 2 . ± .
07 25 . ± . . ± .
32 26 . ± .
32 26 . ± .
02 0 . ± .
02 2 . ± .
45 0 . ± .
04 1 . ± .
70 24 . ± . . ± .
12 25 . ± .
09 26 . ± .
03 0 . ± .
02 0 . ± .
15 0 . ± .
04 1 . ± .
24 24 . ± . a Assuming a distance modulus of ( m − M ) = 41 . b R e,c = R e p b/a c h µ i e , abs ( r ′ ) = h µ i e , z ( r ′ ) − z ) − E ( z ) − K ( z ) assuming a redshift z = 0 ..
24 24 . ± . a Assuming a distance modulus of ( m − M ) = 41 . b R e,c = R e p b/a c h µ i e , abs ( r ′ ) = h µ i e , z ( r ′ ) − z ) − E ( z ) − K ( z ) assuming a redshift z = 0 .. D G s i n A b e ll S & A b e ll TABLE 2A Catalog of LSB dwarfs in Abell S1063
ID R.A. Dec. R e a µ e , F814W F W F W − F W n b/a R e , c b h µ i e , abs ( r ′ ) c (J2000) (J2000) [kpc] [mag arcsec − ] [mag] [mag] [kpc] [mag arcsec − ]AS1063 LSBdw001 342.15747 -44.54634 1 . ± .
06 25 . ± .
07 26 . ± .
03 0 . ± .
02 0 . ± .
07 0 . ± .
02 1 . ± .
05 24 . ± . . ± .
06 25 . ± .
08 25 . ± .
02 0 . ± .
01 0 . ± .
04 0 . ± .
02 1 . ± .
05 24 . ± . . ± .
14 26 . ± .
11 26 . ± .
04 0 . ± .
03 0 . ± .
22 0 . ± .
06 1 . ± .
13 24 . ± . . ± .
13 25 . ± .
12 25 . ± .
02 0 . ± .
02 1 . ± .
21 0 . ± .
04 1 . ± .
12 24 . ± . . ± .
08 25 . ± .
07 25 . ± .
02 0 . ± .
01 0 . ± .
11 0 . ± .
03 1 . ± .
07 23 . ± . . ± .
14 25 . ± .
13 26 . ± .
03 0 . ± .
02 1 . ± .
29 0 . ± .
05 1 . ± .
11 24 . ± . . ± .
11 25 . ± .
11 26 . ± .
04 0 . ± .
03 0 . ± .
14 0 . ± .
04 1 . ± .
09 24 . ± . . ± .
12 25 . ± .
13 25 . ± .
02 0 . ± .
02 1 . ± .
23 0 . ± .
05 1 . ± .
12 24 . ± . . ± .
12 25 . ± .
10 26 . ± .
03 0 . ± .
03 0 . ± .
18 0 . ± .
08 1 . ± .
12 24 . ± . . ± .
07 25 . ± .
07 25 . ± .
02 0 . ± .
01 0 . ± .
11 0 . ± .
05 1 . ± .
07 24 . ± . a Assuming a distance modulus of ( m − M ) = 41 . b R e,c = R e p b/a c h µ i e , abs ( r ′ ) = h µ i e , z ( r ′ ) − z ) − E ( z ) − K ( z ) assuming a redshift z = 0 ..
07 24 . ± . a Assuming a distance modulus of ( m − M ) = 41 . b R e,c = R e p b/a c h µ i e , abs ( r ′ ) = h µ i e , z ( r ′ ) − z ) − E ( z ) − K ( z ) assuming a redshift z = 0 .. L eee t a l. TABLE 3A Catalog of UDGs in Abell 2744
ID R.A. Dec. R e a µ e , F814W F W F W − F W n b/a R e , c b h µ i e , abs ( r ′ ) c (J2000) (J2000) [kpc] [mag arcsec − ] [mag] [mag] [kpc] [mag arcsec − ]A2744 UDG001 3.46229 -30.36563 2 . ± .
08 25 . ± .
06 25 . ± .
02 0 . ± .
01 0 . ± .
03 0 . ± .
01 1 . ± .
05 24 . ± . . ± .
13 26 . ± .
07 25 . ± .
02 0 . ± .
02 0 . ± .
10 0 . ± .
03 1 . ± .
11 25 . ± . . ± .
08 25 . ± .
05 24 . ± .
01 0 . ± .
01 0 . ± .
07 0 . ± .
04 1 . ± .
09 24 . ± . . ± .
18 25 . ± .
08 23 . ± .
00 0 . ± .
00 1 . ± .
08 0 . ± .
01 3 . ± .
16 24 . ± . . ± .
07 25 . ± .
04 24 . ± .
01 0 . ± .
01 0 . ± .
05 0 . ± .
01 1 . ± .
05 23 . ± . . ± .
31 27 . ± .
21 26 . ± .
04 0 . ± .
03 1 . ± .
29 0 . ± .
03 1 . ± .
21 25 . ± . . ± .
06 25 . ± .
07 25 . ± .
02 0 . ± .
01 0 . ± .
04 0 . ± .
02 1 . ± .
05 24 . ± . . ± .
09 25 . ± .
08 26 . ± .
03 0 . ± .
02 0 . ± .
09 0 . ± .
04 1 . ± .
09 24 . ± . . ± .
11 25 . ± .
07 24 . ± .
01 0 . ± .
01 1 . ± .
11 0 . ± .
02 1 . ± .
09 23 . ± . . ± .
13 26 . ± .
06 25 . ± .
03 0 . ± .
02 0 . ± .
09 0 . ± .
02 2 . ± .
10 25 . ± . a Assuming a distance modulus of ( m − M ) = 40 . b R e,c = R e p b/a c h µ i e , abs ( r ′ ) = h µ i e , z ( r ′ ) − z ) − E ( z ) − K ( z ) assuming a redshift z = 0 ..
10 25 . ± . a Assuming a distance modulus of ( m − M ) = 40 . b R e,c = R e p b/a c h µ i e , abs ( r ′ ) = h µ i e , z ( r ′ ) − z ) − E ( z ) − K ( z ) assuming a redshift z = 0 .. D G s i n A b e ll S & A b e ll TABLE 4A Catalog of LSB dwarfs in Abell 2744
ID R.A. Dec. R e a µ e , F814W F W F W − F W n b/a R e , c b h µ i e , abs ( r ′ ) c (J2000) (J2000) [kpc] [mag arcsec − ] [mag] [mag] [kpc] [mag arcsec − ]A2744 LSBdw001 3.45478 -30.37182 1 . ± .
06 25 . ± .
07 25 . ± .
02 0 . ± .
01 0 . ± .
04 0 . ± .
03 1 . ± .
37 24 . ± . . ± .
10 25 . ± .
09 26 . ± .
02 0 . ± .
02 0 . ± .
19 0 . ± .
04 1 . ± .
06 24 . ± . . ± .
08 26 . ± .
13 27 . ± .
06 0 . ± .
04 0 . ± .
08 0 . ± .
06 1 . ± .
09 25 . ± . . ± .
06 25 . ± .
07 25 . ± .
02 0 . ± .
01 0 . ± .
10 0 . ± .
03 1 . ± .
09 23 . ± . . ± .
06 25 . ± .
10 26 . ± .
03 0 . ± .
02 0 . ± .
07 0 . ± .
04 1 . ± .
06 24 . ± . . ± .
09 25 . ± .
10 26 . ± .
03 0 . ± .
02 0 . ± .
06 0 . ± .
02 1 . ± .
06 24 . ± . . ± .
09 25 . ± .
07 25 . ± .
02 0 . ± .
01 1 . ± .
14 0 . ± .
04 1 . ± .
06 24 . ± . . ± .
08 25 . ± .
08 25 . ± .
02 0 . ± .
01 1 . ± .
17 0 . ± .
04 1 . ± .
08 23 . ± . . ± .
15 25 . ± .
14 25 . ± .
02 0 . ± .
01 1 . ± .
23 0 . ± .
03 1 . ± .
07 24 . ± . . ± .
14 26 . ± .
12 27 . ± .
06 0 . ± .
04 0 . ± .
24 0 . ± .
05 1 . ± .
11 25 . ± . a Assuming a distance modulus of ( m − M ) = 40 . b R e,c = R e p b/a c h µ i e , abs ( r ′ ) = h µ i e , z ( r ′ ) − z ) − E ( z ) − K ( z ) assuming a redshift z = 0 ..