Draft version April 25, 2019

Preprint typeset using LATEX style emulateapj v. 12/16/11

COLD, OLD AND METAL-POOR: NEW STELLAR SUBSTRUCTURES

IN THE MILKY WAY’S DWARF SPHEROIDALS
V. Lora1,2 , E. K. Grebel2 , S. Schmeja2,3 , and A. Koch2
1 Instituto de Radioastronomı́a y Astrofı́sica,
Antigua Carretera a Pátzcuaro 8701, 58089 Morelia, México
2 Astronomisches Rechen-Institut, Zentrum für Astronomie der Universität Heidelberg,
Mönchhofstr. 12-14, 69120 Heidelberg, Germany and
arXiv:1904.10560v1 [astro-ph.GA] 23 Apr 2019

3 Technische Informationsbibliothek,
Welfengarten 1b, 30167 Hannover, Germany
Draft version April 25, 2019

ABSTRACT
Dwarf spheroidal galaxies (dSph) orbiting the Milky Way are complex objects often with complicated
star formation histories and internal dynamics. In this work, we search for stellar substructures in
four of the classical dSph satellites of the Milky Way: Sextans, Carina, Leo I, and Leo II. We apply
two methods to search for stellar substructure: the minimum spanning tree method, which helps us to
find and quantify spatially connected structures, and the “brute-force” method, which is able to find
elongated stellar substructures. We detected the previously known substructure in Sextans, and also
found a new stellar substructure within Sextans. Furthermore, we identified a new stellar substructure
close to the core radius of the Carina dwarf galaxy. We report a detection of one substructure in Leo I
and two in Leo II, but we note that we are dealing with a low number of stars in the samples used.
Such old stellar substructures in dSph galaxies could help us to shed light on the nature of the dark
matter halos, within which such structures form, evolve, and survive.
Subject headings: galaxies: dwarf — galaxies: evolution — astronomical databases: miscellaneous —
methods: data analysis

1. INTRODUCTION sub-population, possibly a disrupted cluster, i.e., a dy-

Dwarf spheroidal galaxies (dSph) are the most com- namically cold stellar clump.
mon type of galaxies in the Universe and often consid- These clumpy substructures indicate that there is a
ered to be the building blocks of more massive galax- high degree of complexity already on small scales, which
ies in hierarchical formation scenarios (Dekel & Silk it is also witnessed in other systems (such as in the dis-
1986; Bullock & Johnston 2005; Cooper et al. 2013; tribution of the globular clusters in the Fornax dSph),
Pillepich et al. 2015). The dSph satellites of the Milky mergers between dwarf galaxies in the local Universe
Way (MW) are the best studied dwarf galaxies, since in- (Amorisco et al. 2014; del Pino et al. 2015) and a pos-
dividual stars can be resolved and evolutionary histories sible merger event in the Large Magellanic Cloud (LMC)
can be derived in great detail. A common property of all (Mackey et al. 2016).
dSph galaxies is the presence of an old population (stars Another example is found in Sextans. Kleyna et al.
with ages of 10 Gyr and more), which in many cases (2004) reported the line-of-sight radial velocity disper-
turns out to be the dominant population e.g. (Grebel sion profile of Sextans based on a sample of 88 stars
2000; Grebel & Gallagher 2004). There is growing evi- extending to about 1.5 kpc from the center of Sextans.
dence, that some dSphs (e.g., Carina, Fornax, Sculptor, They found that the dispersion at the center of Sextans
and Sextans) experienced extended star formation his- was close to zero, and that such a low value of the disper-
tories or multiple episodes of star formation The gener- sion was in agreement with significant radial gradients
ally spatially extended old, metal-poor population, and in the stellar populations. They suggested that this is
spatially concentrated young, metal-rich population in- caused by the sinking and gradual dissolution of a star
dicate extended star formation episodes in their centers cluster at the center of Sextans. On the other hand,
(Harbeck et al. 2001). Walker et al. (2006) presented 294 possible stellar Sex-
Some dSph galaxies also present dynamically cold stel- tans members. Their data did not confirm Kleyna et al.’s
lar substructures. For example, Ursa Minor (UMi) shows (2004) report of a kinematically distinct stellar popula-
two distinct density peaks (Kleyna et al. 1998). One rep- tion at the center of Sextans with their more complete
resents the underlying galaxy’s field population with a sample, but they obtained similar evidence if they re-
radial velocity dispersion of σ = 8.8 km s−1 . The second stricted their analysis to a similarly small number of stars
density peak appears with a radial velocity dispersion of as used by Kleyna et al. (2004). Instead, Walker et al.
σ = 0.5 km s−1 . This second peak is located on the (2006) detected a region near Sextans’ core radius, which
north-eastern side of the major axis of UMi at a distance is kinematically colder than the overall Sextans sample
of ∼ 0.4 kpc from UMi’s center. The stars in the vicin- with 95% confidence. Very recently, Kim et al. (2019) re-
ity of this peak comprise a kinematically distinct cold ported the “possible detection of a relic” globular cluster
in Sextans.
Another example is provided by the possibly distinct
v.lora@irya.unam.mx stellar populations in the Canes Venatici I (CVn I) dSph
2 Lora et al.

(Zucker et al. 2006), where Ibata et al. (2006) found two seriously considered.
stellar populations with different kinematics: an ex- The detection of stellar substructures in dSphs, and the
tended metal-poor ([Fe/H]< −2) component with a ve- implications of their survival in the core/cusp problem,
locity dispersion of ∼ 10 km s−1 , and a more concen- prompted us to test other dwarf galaxies, and to inves-
trated metal-rich component of extremely low velocity tigate if we could find stellar substructures within them.
dispersion. On the other hand, Simon & Geha (2007) From velocities, metallicities, and positions of red giant
obtained 214 CVn I member stars that did not reveal any stars in the Sextans, Carina, Leo I, and Leo II dSphs, we
trace of different populations as reported by Ibata et al. analyze whether or not these dSphs could host dynami-
(2006). Ural et al. (2010) obtained spectroscopic data cally cold debris.
for 26 stars in the dSph and investigated whether their This article is organized as follows: In §2 we describe
data exhibit any evidence of multiple populations as pro- some properties of Sextans, Carina, Leo I, and Leo II. We
posed by Ibata et al. (2006) (under the assumption that describe the methods used to search for substructures in
each population was Gaussian), but no clear signature of §3. The results are presented in §4. Finally, we present
distinct sub-populations was found. They argue that the our conclusions in §5.
possible detection of a sub-population rather depends on
the total number of stars in the data set, the fraction of 2. THE FOUR DWARF SPHEROIDAL GALAXIES
stars in the sub-population, the difference in velocity dis- 2.1. Sextans
persion between the populations, and the observational
errors. The Sextans dSph is located at a distance of 86 ± 6 kpc
A final example is the Andromeda II (And II) dSph. (McConnachie 2012). It is a diffuse and faint dSph, with
McConnachie et al. (2007) identified different stellar MV = −9.3 ± 0.5 mag and µV = 27.1 ± 0.5 mag/sq arc-
populations in And II for which they constructed ra- sec (McConnachie 2012). It has a position angle P A =
dial profiles and found that the horizontal branch has 57°and an ellipticity e = 0.30 (Muñoz et al. 2018).
a constant spatial distribution out to a large radius. Battaglia et al. (2011) report the systemic velocity in
In contrast, they found that the reddest stars on the the heliocentric system of Sextans to be vsys = 226 ±
red giant branch in And II are more centrally concen- 0.6 km s−1 .
trated. The latter stellar component has an average age Irwin & Hatzdimitriou (1995) computed a core radius
of ∼ 9 Gyr and is relative metal-rich ([Fe/H]∼ −1), lead- rc = 16.6±1.2 arcmin and a tidal radius rt = 160±50 ar-
ing to a prominent red clump. And, it has to be noted cmin. More recently, Roderick et al. (2016) fitted a King
that And II has been identified as having experienced a model to the radial distribution of Sextans computing a
merger (Amorisco et al. 2014). core radius rc = 26.8 ± 1.2 arcmin and a tidal radius
The stellar substructures found within dSphs are of rt = 83.2 ± 7.1 arcmin (2.08 ± 0.18 kpc) Muñoz et al.
great importance, since the survival of stellar substruc- (2018) reported a core radius rc = 20.1 ± 0.05 arcmin
tures within the dark matter (DM) halo of dSphs sug- and a tidal radius rt = 60.5 ± 0.6 arcmin (see also
gests that dSphs are more in agreement with a cored Cicuendez et al., 2018).
DM profile, rather than with the cuspy DM profile pre- The Sextans dSph is more than 12 Gyr old
dicted by the Λ-cold DM model (Lora et al. 2012, 2013; (Mateo et al. 1991) and it is considered as a metal poor
Amorisco 2017; Contenta et al. 2017). dwarf galaxy (Kirby et al. 2001) with a mean [Fe/H]=
Using N-body simulations, Lora et al. (2012) and −1.93 (Battaglia et al. 2011). Lokas (2009) computed a
Lora et al. (2013) studied the survival of cold kinematic high mass-to-light-ratio of M/L ≈ 97. Battaglia et al.
stellar substructures embedded in the DM halo of the (2011) computed a M/L ≈ 460 − 920. Such a high mass-
Fornax and Sextans dwarf galaxies -respectively-, against to-light-ratio indicates that Sextans is a DM-dominated
phase mixing. They compared the evolution of the stel- dwarf galaxy.
lar substructures when the dark halo has a core and The existence of stellar substructure in the Sextans
when it follows the Navarro-Frenk-White (NFW) pro- dwarf galaxy has been previously studied. For exam-
file (Navarro et al. 1997). The core in the DM halo mass ple, Kleyna et al. (2004) found a drop in the stellar ve-
density distributions is large enough to make the gravita- locity dispersion near the center of Sextans, which was
tional potential almost harmonic, guaranteeing the sur- interpreted as a dissolving cluster. Some years later,
vival of the stellar structures mentioned even if the stellar Walker et al. (2006) reported no detection of a kinemat-
substructure is initially very extended. On the contrary, ically distinct population at the center of Sextans. In-
the stellar substructures are very rapidly destroyed when stead, they detected a region at a distance r ∼ 16 arcmin
embedded in a NFW DM profile. (400 pc) from Sextans’ center, corresponding to the core
Still, one cannot rule out a scenario where the DM radius reported by Irwin & Hatzdimitriou (1995).
profile of dwarf galaxies was initially cuspy and evolved Later, Battaglia et al. (2011) reported the detection of
to a cored DM halo via energy feedback from super- a cold substructure consisting of nine very metal-poor
nova explosions, stellar winds, and/or star formation stars close to the center of Sextans.These stars have very
(Maschenko et al. 2006; Chen et al. 2016; Pasetto et al. similar distances, kinematics, and metallicities. The av-
2010). However, in faint galaxies like Eridanus II and erage metallicity of this 9-star group is [Fe/H]= −2.6 dex
Andromeda XXV (i.e., LV ∼ 5 × 105 L⊙ and LV ∼ with a 0.15 dex 1σ-scatter. This group of stars was taken
6 × 104 L⊙ , respectively) the formation of large cores from the most metal-poor stars, which have a low veloc-
via stellar feedback is not obviously expected (Amorisco ity dispersion (∼ 1.2 km s−1 ) and whose average radial
2017). If stellar feedback in faint galaxies is found to be velocity is 72.5 ± 1.3 km s−1 . Battaglia et al. (2011) as-
inefficient, then alternative candidates for DM should be sume that the ratio of the stars in the substructure is
representative with respect to their total number of Sex-
 Old stellar substructures in MW’s dSphs 3

TABLE 1
Parameters of Sextans, Carina, Leo I and Leo II dSphs.

Parameter Sextans Carina Leo I Leo II Ref.a

′ ′ ′ ′
(αJ 2000 , δJ 2000 ) 10h 13m 03s , -01°13 03′′ 6h 40m 24.3s , -50°58 0′′ 10h 8m 27.0s , 12°18 30′′ 11h 13m 29s , 22°9 12′′ 1,1,1,1
+16
Distance 86 ± 6 kpc 105 ± 6 kpc 254−19 kpc 233 ± 15 kpc 2,2,2,2
MV −9.3 ± 0.5 mag −9.3 ± 0.5 mag -11.9 mag -9.6 mag 2,2,3,2
PA 57 60 78 12 11,11,11,5
e 0.30 0.38 0.31 0.07 11,11,11,11
Rcore 20.1±0.05 arcmin 7.97±0.16 arcmin 3.6±0.1 arcmin 2.25±0.1 arcmin 11,11,11,11
Rtidal 60.5±0.6 arcmin 58.4±0.98 arcmin 13.5±0.3 arcmin 9.82±0.4 arcmin 11,11,11,11
vsys 226 ±0.6 km s−1 223.9 km s−1 284.2 km s−1 79.1 km s−1 6,7,8,9
M/L 97 66 24 25-50 10,10,8,9
〈[Fe/H]〉 -1.93±0.1 -1.72±0.1 -1.43±0.1 -1.62±0.1 6,7,8,9

a References: (1) Mateo (1998), (2) McConnachie (2012), (3) Bellazzini et al. (2004), (4) Bellazzini et al. (2005) (5) Irwin & Hatzdimitriou
(1995), (6) Battaglia et al. (2011), (7) Koch et al. (2006), (8) Koch et al. (2007a), (9) Koch et al. (2007b), Lokas (2009) & (11) Muñoz et al.
(2018).

tans members (174 stars). Thus the substructure would lations corresponding to 11, 5, and 0.6 Gyr (Monelli
account for 5% of Sextans stellar population. 2003; Bono et al. 2010) and shows a slight population
Finally, Roderick et al. (2016) reported on photomet- gradient where the more metal rich population is more
ric evidence of stellar substructures associated with Sex- centrally concentrated in the galaxy (Koch et al. 2006).
tans. The stellar substructures that they find extend out de Boer et al. (2014) confirmed that the star formation
to a distance of 82 arcmin (2 kpc) from Sextans’ centre. in the Carina dSph is episodic. There are two main
The existence of such stellar substructures in the outer episodes of star formation that occurred at > 8 Gyr and
regions of Sextans might indicate that Sextans is under- 2 − 8 Gyr (or at ∼ 12 Gyr and 4 − 8 Gyr according
going tidal disruption. However, Roderick et al. (2016) to Monelli, 2014). Monelli (2014) argue that these two
found that the substructures surrounding Sextans appear episodes of star formation are inconsistent with a simple
to be both aligned with, and perpendicular to Sextans’ evolution of an isolated system.
major axis. The latter suggests that Sextans is not neces- Kordopatis et al. (2016) found that the youngest
sarily undergoing a strong tidal disruption. Muñoz et al. metal-rich population in Carina is more extended than
(2018) reported a fairly regular morphology in Sextans, the intermediate-metallicity population, while generally
with no obvious signs of tidal features. the metal-rich stellar populations are more spatially con-
In contrast, recently Cicuendez et al. (2018) reported centrated and kinematically colder. To explain this,
signs of past accretion/merger events in Sextans: a ‘ring- Sales et al. (2010) argue that strong tidal interaction
like’ feature. They claim that the kinematically detected may play a big role on Carina’s peculiar configuration. In
ring in Sextans bears a morphological resemblance to the addition to this scenario, Muñoz et al. (2006) detected an
stellar stream in the And II dSph (Amorisco et al. 2014), extended power-law component in the density distribu-
which probably merged with another dwarf galaxy. tion of Carina. Such a distribution is a typical signature
of disrupted galaxy satellites (e.g. the Sagittarius dSph,
2.2. Carina Majewski et al. 2003). Moreover, Muñoz et al. (2006)
Carina is located at a heliocentric distance of 105 ± found that the extra-tidal stars around Carina could be
6 kpc, with MV = −9.3 ± 0.5 mag and µV = 25.5 ± related to the LMC, indicating complex interactions.
0.5 mag/sq arcsec (McConnachie 2012). It has a position Fabrizio et al. (2011) presented radial velocity mea-
angle P A = 60°, a core radius rc = 7.97 ± 0.16 arcmin, a surements of Carina’s stars and collected spectra of old,
tidal radius rt = 58.4±0.98 arcmin, an ellipticity e = 0.38 intermediate, and young stellar tracers. They detected
(Muñoz et al. 2018), and a heliocentric systemic velocity a maximum in the radial velocity, which they suggest
vsys = 223.9 km s−1 (Koch et al. 2006). might be reminiscent of a stellar structure located at
The large velocity dispersion of Carina is often in- ∼ 200 pc of Carina’s center. Fabrizio et al. (2016) com-
terpreted as evidence of a large mass-to-light ratio pared observations of the radial velocity distribution of
(Mateo et al. 1993). Lokas (2009) found a mass-to-light old and intermediate-age stars in the Carina dSph with
ratio for Carina of 66, from where they inferred that Ca- N-body simulations. They found a good agreement with
rina’s kinematics is dominated by its DM halo. the Vrot /σ ratio in the central regions of the dwarf. The
Spectroscopic observations of Carina indicate that it latter indicates that Carina might have been a disky
has two dominant stellar populations that should be in dwarf galaxy that experienced several strong tidal in-
equilibrium in the same DM halo, and that such a DM teractions with the MW.
halo has a less cuspy inner density profile than previously
thought (Hayashi et al. 2018). 2.3. Leo I
The Carina dSph has experienced a complex star The Leo I dSph is one of the most remote dSphs asso-
formation history. It is the only dSph to exhibit ciated with the MW (e.g. Grebel et al. 2003). It has
clearly episodic star formation interrupted by long qui- a Galactocentric distance of 254 kpc (Bellazzini et al.
escent periods (de Boer et al. 2014; Santana et al. 2016; 2004), a position angle P A = 78°, a core radius rc =
Hurley-Keller et al. 1998; Monelli 2003). Its color- 3.6 ± 0.1 arcmin, a tidal radius rt = 13.5 ± 0.3 arcmin, an
magnitude diagram shows three different stellar popu- ellipticity e = 0.31 (Muñoz et al. 2018), and a systemic
4 Lora et al.

(a) (b)

(c) (d)
Fig. 1.— Brute-force method: 2D Histograms of metallicity and radial velocity per elliptical annulus bin. As an example, using
Battaglia et al.’s 2011 data, we show the 2D histograms for four elliptical-radial bins for the Sextans dwarf; (a) [0-10] arcmin, (b) (10-
20] arcmin, (c) (20,30] arcmin, and (d) (30,40] arcmin. The typical size of the uncertainty in metallicity is −0.2 dex, and the typical size
of the uncertainty in the radial velocity is 2.1 km s−1 .
velocity vsys = 284.2 km s−1 (Koch et al. 2006). Leo I was formed and evolved in isolation, and it is now
Because of Leo I’s high radial velocity and its approaching its first encounter with the Galactic halo.
large Galactocentric distance, Koch et al. (2007a) ar- Gaia Collaboration, Helmi, et al. (2018) estimated
gue that Leo I might be an isolated system, which that Leo I has a period greater than 5 Gyr, and they
is currently not affected by Galactic tides. However, predict that its last pericentric passage (at a distance
Boylan-Kolchin et al. (2013) argue that it is very unlikely of ∼ 100 kpc) took place around 1 Gyr ago. The Leo I
that Leo I is not bound to the MW galaxy, under the dSph has been used to derive a limit on the mass of the
premise that Galactic satellites are associated with DM MW DM halo. If Leo I is indeed bound to the MW, it
subhalos. They used high resolution numerical simula- sets constrains on the MW mass (Boylan-Kolchin et al.
tions of a MW like DM halo and found that 99.9% of the 2013). Gaia Collaboration, Helmi, et al. (2018) set
subhalos in the simulations are bound to their host halo. a lower limit on the mass of the MW DM halo of
On the other hand, if Leo I passed very close to the 9.1+6.2
−2.6 × 10
11
M⊙ , based on the assumption that Leo I
Galactic center (around 1 Gyr ago), then the observed is indeed bound to the MW. Recentely, Fritz et al.
kinematics and population segregation in Leo I, along (2018) found that Leo I is “back-splashing” if one
with its distorted structural parameters, its age, its last considers the (preferred) heavy MW DM halo model
prominent burst of star formation, and its large radial ve- (MMW = 1.6 × 1012 M⊙ ).
locity relative to the Galactic Center, can be explained Interestingly, stellar substructure has been detected in
(Mateo et al. 2008). Because Leo I presents no tidal Leo I. Mateo et al. (2008) reported six stars uniformly
arms, the latter scenario could result from the interac- distributed and kinematically distinct from the main
tion with a third body, which placed Leo I into its present Leo I stellar component. They claim that these stars
highly-elliptical orbit. might represent a tidal feature, but they warn that the
Sohn et al. (2013) studied the detailed orbital history statistics are too poor and that further members of this
of Leo I. They found that Leo I entered the MW virial kinematic structure would need to be identified in order
radius 2.3 Gyr ago, and that (confirming Mateo et al. to conclude whether the substructure is real or not.
2008) it had a pericentic approach (at a Galactocentric
distance DGC = 91 kpc) around one Gyr ago. 2.4. Leo II
However, Koch et al. (2012b) measured the proper mo-
tion of Leo I and found that Leo I might not be bound The Leo II dSph is located at a Galactocentric distance
to the MW. Furthermore, they say that it is likely that of 233 kpc (Bellazzini et al. 2005). It has a position an-
gle P A = 12°, a core radius rc = 2.25 ± 0.1 arcmin, a
 Old stellar substructures in MW’s dSphs 5

tidal radius rt = 9.82 ± 0.4 arcmin, an ellipticity e = 0.07 tributions for both metallicity and velocity (in the same
(Muñoz et al. 2018), and a systemic velocity in the helio- elliptical annulus bin where the peak was found) and find
centric system vsys = 223.9 km s−1 (Koch et al. 2006). the average and standard deviation (σ) of these distribu-
Leo II is believed not to be experiencing strong Galac- tions. If the number of counts in the peak (which is found
tic tides, and its proper motion indicates that it might in a defined metallicity-velocity cell) is ≥ 1.96σ, or in
not even be a bound satellite to the MW (Lépine et al. other words, the probability of finding a peak > 1.96σ is
2011). Piatek et al. (2016) measured the proper mo- ≤ 5% we take such a count peak to be significant.
tion of Leo II. They found that the motions they mea-
sured support the idea that Leo II fell into the MW 3.2. Minimum spanning tree method
DM halo as a part of a group. On the other hand, Another method to investigate whether stars of a cer-
Gaia Collaboration, Helmi, et al. (2018) concluded that tain metallicity range are in some way kinematically
the infall of Leo II as a group is unlikely. clumped (i.e., spatially more concentrated) we make use
Vogt et al. (1995) calculated a M/L ≈ 7, which indi- of a minimum spanning tree (MST, e.g., Schmeja 2011).
cates that Leo II must be embedded in a DM halo, but The MST is the unique set of straight lines (“edges”) con-
it is not an extreme case. Koch et al. (2007b) obtained necting a given set of points (“vortices”) without closed
a large data set of radial velocity measurements out to loops, such that the sum of the edge lengths is mini-
Leo II’s limiting radius. They found (depending on the mum. This construct from graph theory has been widely
total luminosity adopted) a M/L = 25 − 50. They con- applied in astronomy to cluster and structure analysis,
cluded that this M/L ratio together with the flatness from the large-scale distribution of galaxies to the inter-
of its dispersion profile indicate that Leo II is a DM- nal structure of star clusters (Schmeja 2011, and refer-
dominated system. ences therein). We apply an approach similar to the one
Komiyama et al. (2007) reported the existence of a introduced by Allison et al. (2009) to identify and quan-
small stellar substructure beyond Leo II’s tidal radius. tify mass segregation in star clusters. We construct the
The substructure’s luminosity compares to that of a glob- MST for the stars of a given metallicity range and deter-
ular cluster (MV < −2.8). The substructure might be a mine the mean edge length γmp for those stars. We use
disrupted globular cluster or a group of stars that were the geometric mean rather than the arithmetic mean in
stripped away from the galaxy. order to minimize the influence of outliers (Olczak et al.
3. METHODS 2011). Then we construct the MST of the same number
of randomly selected stars from the entire sample and
3.1. “Brute-force” method
determine the mean edge length γrand . This is done 200
We search for stellar substructures in the four selected times in order to obtain the mean value hγrand i. The
dSph galaxies. For this purpose, we first constructed ratio
2-dimensional histograms on a grid of metallicity vs. ve- hγrand i
locity (see Figure 1) for every elliptical annuli of constant R= , (3)
γmp
ellipticity and position PA in the selected galaxy.
We adopt an initial metallicity meti (the minimum is a measure for the concentration of the stars of the sub-
metallicity value of the studied data) and a final metallic- sample relative to the entire stellar population. A value
ity metf (the maximum metallicity in the data), with an of R ≈ 1 implies that both samples are distributed in
interval in metallicity ∆met and an interval in velocity a similar manner, while R ≫ 1 indicates that the sub-
of ∆v, see Figure 1. sample is more concentrated than the sample as a whole.
Subsequently, with both metallicity (met) and velocity
(v) fixed, we count the stars in our data, that satisfy 4. RESULTS
being in the interval 4.1. The case of Sextans
met < met∗ 6 met + ∆met , (1) For the case of Sextans, we used Battaglia et al.’s
(2011) data. They obtained VLT/FLAMES
where met∗ is the metallicity of the star that is being intermediate-resolution spectroscopic observations
analyzed. We repeat the procedure with the velocity, in the near-infrared CaII triplet (CaT) region for 1036
such that distinct targets along the line-of-sight to Sextans. The
v < v∗ 6 v + ∆v , (2) magnitudes and colors of those targets are broadly
where v∗ corresponds to the velocity of the star that consistent with red giant branch (RGB) stars. From
is being analyzed. With the resulting number of stars that sample they obtained 789 stars with S/N and error
that satisfy the conditions 1 and 2 (per elliptical annu- in velocity that produce reliable line-of-sight velocities
lus width), we build histograms of the number of stars to and CaT equivalent widths. A subset of 174 stars
find peaks in the counts of stars that could be interpreted from those are RGB stars that are probable members
as stellar substructures. with line-of-sight velocities accurate to ±2 km/s and
The peaks in the counts of stars that are interesting CaT [Fe/H] measurements accurate to ±0.2 dex. The
to us are the ones with low metallicities, and that are majority of the Galactic contaminants were eliminated
dynamically cold, meaning that the velocity dispersion from Battaglia et al.’s sample using a 3σ kinematic cut.
of such group of stars should be significantly lower that Finally, in order to refine their membership criteria, they
that of the complete stellar sample (per elliptical annulus used the Mg I line at 8806.8 Å as an empirical indicator
bin). of stellar surface gravity, so that the probable members
In order to declare whether a peak in counts is signifi- (RGB stars) of Sextans and the Galactic contaminants
cant or not, we realize Monte Carlo tests of normal dis- could be distinguished.
6 Lora et al.

(a) (b)

Fig. 2.— (a) The Sextans dSph members reported by Battaglia et al. (2011) are shown as filled-black circles. The white small circle (in
the center) represents the center of Sextans reported by Mateo (1998). The stellar substructure previously found by Battaglia et al. (2011)
(called sxt1 in the text), is shown with filled blue squares. The new stellar substructure, reported in this work (sxt2) is shown with filled
light blue squares (and two extra violet squares, see text in Section 2.1). The solid black stars, in both substructures (sxt1 and sxt2)
correspond to the centroid of each stellar substructure. The solid ellipse correspond to the core radius reported by Irwin & Hatzdimitriou
(1995). The dotted ellipse is placed at the center of sxt2 substructure, which corresponds to 43.5 arcmin (1 kpc) from the center of Sextans.
The region marked in white shows the kinematically colder region detected by Walker et al. (2006).
(b) In this panel we show contours of star counts similar to Roderick et al.’s (2016) Figure 13. We over plotted their core (26.8 arcmin,
see white dotted ellipse) and their tidal radius (83.2 arcmin, see black ellipse), which they found by fitting a King model to the radial
distribution of Sextans. The substructure named sxt2 is shown as black filled squares.
Our first target of study is the stellar clump found in the velocities range from 3.7 ≤ v ≤ 10.08 km s−1 . We
Sextans by Battaglia et al. (2011). We will refer to this computed the centroid (relative to the center of Sextans,
clump of stars as sxt1. With our brute-force method see Table 2) of sxt2 to be located at (4.537,−29.265) ar-
(see Section 3.1), we should be able to reproduce previ- cmin (relative to the center of Sextans).
ous findings. Using the Battaglia et al. (2011) data, we In Figure 2a, we show the stars of the new clump
follow the method of Section 3.1. We set meti = −3 dex, (sxt2) as filled cyan squares, and the respective centroid
metf = −0.28 dex and ∆met = 0.5 dex. For the veloc- plotted as a solid black star. It has to be noted that if one
ities, we set vi = −20 km s−1 , vf = 20 km s−1 , and a relaxes the velocity constraint and lets the velocity cover
∆v = 10 km s−1 . 3.7 ≤ v . 16 km s−1 , we can add two new metal-poor
Then we construct histograms of star counts per el- stars. Therefore, we end with a 10-star substructure with
liptical radius. These histograms help us to identify the an average metallicity of −2.76. If we include these two
accumulation of stars in a particular radius bin. The el- stars, the centroid of the 10-star substructure changes to
liptical radial bins in the case of Sextans have a size of (4.334, −29.700) arcmin. The two extra stars are shown
10 arcmin. As expected, we find sxt1. as two filled magenta squares in Figure 2a.
The metallicity of this group of stars (sxt1) is in the The MST analysis also detects sxt1 and sxt2 and
range −2.84 ≤ met ≤ −2.43 dex, with an average metal- confirms them as highly concentrated clumps with R =
licity [Fe/H]= −2.64, and the velocity is in the range 3.28 ± 0.20 and R = 3.53 ± 0.31, respectively (see Equa-
−2 ≤ v ≤ −12 km s−1 . The fact that we are able to tion 3).
recover the sxt1 clump is an indicator that our method In order to demonstrate that random groupings of stars
is working well. with matching metallicities and velocities are in fact very
In Figure 2a we plot Sextans’ member stars (black cir- rare, we realize Monte Carlo tests of normal distributions
cles) taken from the Battaglia et al. (2011) data set. The for both metallicity and velocity (see Section 3.1). We
center of the Sextans dSph (Mateo 1998) is shown as a find that the probability of finding sxt1 is ∼ 1% (i.e.,
white circle. The sxt1 clump stars are shown as filled ∼ 2.5σ). We also find that the probability of finding
blue squares, and the solid black star shows the centroid sxt2 is ∼ 5% (i.e., 2.1σ).
of sxt1. The latter supports the suggestion that in fact the new
We did not only recover the sxt1 clump reported by clump sxt2 reported in this work is likely physical.
Battaglia et al. (2011), but we also found a second old Interestingly, Roderick et al. (2016) reported clear ev-
cold-stellar substructure. We will refer to this new sub- idence of a stellar substructure distributed evenly about
structure as sxt2. The sxt2 stellar substructure consists the center of Sextans. The substructure extends up
of eight stars with metallicities ranging from −2.99 to to a distance of 2 kpc. In order to compare our find-
≤ −2.63. The average metallicity is [Fe/H]= -2.78, and ings to Roderick et al. (2016), we built contours of star
counts using Roderick et al.’s (2016) substructure data,
 Old stellar substructures in MW’s dSphs 7

(a) (b)
Fig. 3.— (a) In this panel we show the Carina dSph members reported by Koch et al. (2008) as filled black circles. The center of Carina
is represented with a filled red circle. The new stellar substructure found in this work (car1) is shown with filled yellow squares.The white
star corresponds to the centroid of the stellar substructure (relative to Carina’s center, see Table 2). The two solid black ellipses show the
core radius at 8.8 arcmin (240 pc) and the tidal radius at 28.8 acrmin (787 pc).
(b) Fabrizio et al.’s (2011) Figure 13: sky distribution of the stars from Fabrizio et al.’s (2011) photometric catalog. Isodensity contours
are shown ranging from 5 to 95% of the total number of stars. The blue star shows the position of the star HD48652. The solid (vertical
and horizontal) black lines mark the two secondary peaks in RV. The nine stars making up car1 are overplotted as filled yellow squares.
The black solid ellipse is located at the distance of the core radius (8.8 arcsec).

shown in Figure 2b. We over-plotted their core (white 2.76. The velocity dispersion of the 10-star clump is a
dashed-line) and tidal radius (black line) (26.8 arcmin bit higher than the one computed with eight stars, but
and 83.2 arcmin, respectively). It is very encouraging it remains colder than the velocity dispersion of sxt2’s
to observe that our sxt2 substructure lies in a dense associated elliptical annulus (9 km s−1 ).
substructure region, giving a hint that sxt2 might be a As done by Battaglia et al. (2011), if the 174 stars in
member of the annular substructure within Sextans. the Sextans sample are representative, then sxt2 would
Battaglia et al. (2011) transformed the heliocentric account for 4.6% (or ∼ 6% if one includes the extra
line-of-sight velocities into line-of-sight velocities in a two stars) of Sextans’ stellar population. If we take the
frame at rest with respect to the Galactic center, value for the luminosity of Sextans of 4.37 × 105 L⊙
vGSR (where GSR stands for Galactic standard of rest, (Irwin & Hatzdimitriou 1995; Lokas 2009) and a typi-
vsys,GSR = 78.4 ± 0.6 km s−1 ). They reported a velocity cal value for the mass-to-light ratio for globular clusters,
dispersion for the six innermost stars making up their M/L ∼ 2 (McLaughlin & van der Marel 2005); then zero
clump of 1.4 ± 1.2 km s−1 , and an average GSR velocity order estimates for the mass and luminosity of sxt2 are
of 72.5 ± 1.3 km s−1 . They also computed an average 4 × 104 M⊙ and 2×104 L⊙ (5×104 M⊙ and 2.5×104 L⊙ ,
metallicity for their clump of [Fe/H]= -2.6. if we include the two extra stars).
We follow Battaglia et al. (2011) and compute a veloc- The fact that the MST analysis confirmed that sxt2
ity dispersion and a mean GSR velocity for eight stars of is concentrated with respect to the overall Sextans sam-
sxt2 (see the cyan squares in Figure 2a) of 2.2 km s−1 ple, along with its metal-poor, dynamically cold nature,
and 87.1 km s−1 , respectively. reasserts the idea that Sextans has not experienced a
Since the velocity dispersion could be a strong function strong tidal encounter, indicating that it could have a
of the galactocentric radius, we compute the velocity dis- more circular orbit around the Galactic Center. The low
persion of all the stars in Battaglia et al.’s (2011) data metallicity of sxt2 suggests that it might be very old
belonging to the same elliptical annulus as sxt2. We (i.e., > 10 Gyr) and its progenitor (as well as the progen-
found that the velocity dispersion of the elliptical annu- itor of sxt1, Battaglia et al. 2011) would be most likely
lus associated with sxt2 is 9 km s−1 . a globular cluster, and such a globular cluster would then
The velocity dispersion of sxt2 is very cold (compared be among the most metal-poor globular clusters known.
to the velocity dispersion of all the stars belonging to The detection of cold old stellar substructures in
the same elliptical annulus) and its average metallicity dwarf spheroidal galaxies has been helpful to shed some
([Fe/H]= -2.78) is even lower than that of sxt1. The low light on the core/cusp DM problem at galactic scales.
metallicity of sxt2, together with its cold kinematics, For example, a cored DM halo profile is preferred when
suggests that the stars belonged to a stellar (globular) explaining the existence of the stellar substructures in
cluster. If we add the two extra stars shown as magenta UMi and Sextans (Kleyna et al. 2003; Lora et al. 2009,
squares in Figure 2a (where we have relaxed the velocity 2013). The stellar substructure sxt2 gives a new test to
condition, see text above), we obtain a velocity dispersion corroborate if (in particular) Sextans is embedded in a
of 4.2 km s−1 and an average metallicity of [Fe/H]=- cored DM halo.
8 Lora et al.

the mass estimate of sxt1 and sxt2.

As we mentioned before, Fabrizio et al. (2011) found
4.2. The case of Carina evidence of a secondary maximum in radial velocity at a
For the case of Carina, we used Koch et al.’s (2006) distance ∼ 200 pc from Carina’s center, which they in-
data. Their targets in Carina were chosen from pho- terpreted as reminiscent of a substructure with transition
tometry and astrometry obtained by the ESO Imaging properties (i.e., a transition between a bulge-like and/or
Survey (Nonino et al. 1999). Koch et al. (2006) selected disk-like structure). In Figure 3b we took the sky distri-
their targets to cover magnitudes from the tip of the RGB bution of the stars in Fabrizio et al.’s (2011) photometric
down to 3 mag below the RGB tip (20.3 mag in apparent catalog (their Figure 13), where they plot isodensity con-
V-band magnitude). They chose their RGB magnitude tour levels from 5% to 95%. We over plotted the car1
range such that even for the faintest stars they would clump of stars. All the (nine) stars lie inside the 45%
be able to achieve high signal-to-noise ratios. They ob- isodensity contour level, which coincides with the core
tained VLT/FLAMES spectroscopic observations in the radius of Carina (see the black ellipse in Figure 3b), and
near-infrared CaT region, using the GIRAFFE spectro- it also coincides with the location where Fabrizio et al.
graph with both MEDUSA fiber slits in “low-resolution” (2011) reported their substructure.
mode. Their observed fields covered most of Carina’s
area, and extended even beyond its nominal tidal radius. 4.3. The case of Leo I
The membership probabilities were calculated using an For the case of Leo I, we used Koch et al.’s (2007a)
error-weighted maximum-likelihood fit of a Gaussian ve- (hereafter K07) data together with Bosler et al.’s (2007)
locity distribution (in the 150 − 300 km s−1 range), and (hereafter B07) data.
then rejecting 3σ outliers. Koch et al. (2006) determined The targets in Koch et al. (2007a) were selected from
the metallicities of the RGB stars sample through the photometry obtained with the framework of the Cam-
equivalent widths-CaT method. bridge Astronomical Survey Unit (Irwin & Lewis 2001)
We apply the same methods described in Section 3 to at the 2.5 m Isaac Newton Telescope. They selected
the Carina dSph. We set the metallicity limits meti = red giant stars in order to cover magnitudes from the
−3 dex, metf = −0.3 dex and ∆met = 0.5 dex. For the tip of the RGB (I ∼ 18 mag) down to 1.6 mag below
velocities we set vi = −20 km s−1 and vf = 20 km s−1 the RGB tip (I . 19.6 mag). Their observations were
with a ∆v = 10 km s−1 . We find a group of nine stars carried out with the Gemini Multiobject Spectrograph
with metallicities in the range −2.89 ≤ met ≤ −2.44 (GMOS) and with the Deep Imaging Multi-Object Spec-
and velocities 2.93 ≤ v ≤ 13.31 km s−1 , within elliptical trograph (DEIMOS). Their low-resolution spectra have
annuli widths of 5 arcmin. We refer to this group of a signal-to-noise of ∼ 5, sufficient to obtain accurate ra-
stars as car1. We plot car1 in Figure 3a as filled yellow dial velocity measurements. They derived radial veloci-
squares. car1’s centroid is shown with a white star. We ties from their final reduced spectra by cross-correlating
notice that the centroid of car1(−0.504, 0.153 arcmin) the three strong Ca lines (8498, 8542, and 8662) with a
is quite close to the center of Carina (just ∼ 14 pc away), synthetic template spectrum of the CaT region. They
represented with a filled red circle. estimated the metallicities of their RGB stars sample,
Applying the MST analysis, car1 is found as a promi- from the widely used method of relating the equivalent
nent clump with a ratio R = 3.51±0.30 (see Equation 3). widths of the CaT to metallicity.
We demonstrate that random groupings of stars with On the other hand, Bosler et al. (2007) obtained low-
matching metallicities and velocities in Carina’s data dispersion spectra of red giants in Leo I (and Leo II)
are very rare. We realize Monte Carlo tests of normal using the Keck I 10-m telescope and LRIS. They ob-
distributions for both metallicity and velocity (see Sec- tained a mean S/N of 18 (for Leo I stars). They verified
tion 3.1). We find that the probability of finding car1 is the membership of their RGB stars by deriving their ra-
∼ 4.5% (i.e., 2.15σ) supporting the suggestion that the dial velocities. From their 121 stars observed in Leo I,
substructure CAR1 is physical. The stellar substructure a number of 90 have heliocentric velocities within 3σ of
found in Carina could be an unbound object in the throes the average velocity, and thus were selected as members
of destruction, which could have started out as a typical of Leo I. They estimated the metallicities of their sample
cluster with a radius of ∼ 3 − 5 pc, very similar to the relating the equivalent widths of the CaT (as done by
stellar substructure sxt1. Koch et al. 2007a).
For the nine stars group (car1) we computed a It has to be noted that the K07 and B07 data sets do
velocity dispersion of 4 km s−1 , a mean velocity of not overlap.
230.2 km s−1 , and an average metallicity of [Fe/H]∼ We took both data sets, and applied the brute force
−2.6. method (Section 3.1) to scrutinize the possible existence
We compute the velocity dispersion of all the stars in of stellar substructure. We set the ranges for the metal-
Koch et al.’s (2006) data belonging to the same annulus- licity meti = −2.7 dex, metf = −0.3 dex and ∆met =
bin as car1. We found that the velocity dispersion of 0.5 dex. For the velocities we set vi = −20 km s−1 and
the elliptical-annuli associated to CAR1 is 7.2 km s−1 . vf = 20 km s−1 with ∆v = 10 km s−1 .
If the 437 star-sample is representative of Carina’ s We only found one group of stars that was relevant,
stellar component, then car1 would account for 2% using elliptical annuli widths of 4 arcmin. This group of
of Carina’s stellar population. If we adopt a mass-to- stars consists of six stars (hereafter, li1). The metallicity
luminosity-ratio of 2 (typical for globular clusters) and a of li1 is in the range −1.78 ≤ met ≤ −1.52, with an
luminosity of 3.5×105 (Lokas 2009), then we can roughly average metallicity [Fe/H]= −1.63. The velocities are in
estimate that car1’s mass is ∼ 1.4 × 104, very similar to the range 288 ≤ v ≤ 284.3 km s−1 , the average velocity
 Old stellar substructures in MW’s dSphs 9

(a) (b)

Fig. 4.— (a) Leo I dSph members reported by Koch et al. (2007a) are shown as black open circles. The red giant stars reported by
Bosler et al. (2007) are represented with black open triangles. The white plus symbols represent Mateo et al. (2008)’s data. Overplotted
are the group of stars reported in this work li1 (-filled-magenta squares); the group of stars reported by Koch et al. (2007a) (filled green
squares), and the group of stars reported by Mateo et al. (2008) (filled blue squares). The inner solid black ellipse corresponds to the core
radius and the outer solid black ellipse corresponds to the tidal radius (Muñoz et al. 2018).
(b) Heliocentric velocities as a function of the elliptical radius. The dotted vertical lines show the core (3.3′ ) and tidal (12.6′ ) radius. The
horizontal dotted line shows vHelio ≈ 96 km s−1 , the average velocity of Mateo et al. (2008)’s sample.

is 286 km s−1 , and the velocity dispersion is 1.6 km s−1 . group of five stars reported by K07 (filled green squares).
We computed the velocity dispersion of all the stars in We perform a statistical study to show whether ran-
the K07 and B07 data belonging to the same elliptical dom groupings of stars in K07 and B07 data are rare or
annulus as LI1. We found that the velocity dispersion of not. We realize Monte Carlo tests of normal distribu-
the elliptical annulus associated with LI1 is 13 km s−1 . tions for both metallicity and velocity (see Section 3.1).
We show the stars belonging to li1 as filled magenta We find that the probability of finding li1 is ∼ 75% (i.e.,
squares in Figure 4. The open black circles are the red 0.33σ). Such a finding reflects the fact that, in the case
giants reported by K07 and the open black triangles cor- of Leo I, we are dealing with low number statistics, and
respond to B07’s stars. thus li1 is not statistically significant.
The MST method only finds few rather weak concen- The velocity dispersion of li1 is a factor of ∼ 3 smaller
trations with R . 1.5 and a low number of members, (σv,LI1 = 1.6 km s−1 ) than the dispersion found in
therefore it did not help us to corroborate if li1 is a real M08’s group of stars. This, together with the low av-
substructure. erage metallicity found in li1, indicates that li1 (if real)
K07 detected a minor significant rise in the radial dis- might be a dynamically cold stellar substructure.
persion profile, around 220 pc from the center of Leo I Five of the six stars in li1 are contained in the B08
(corresponding to the core radius), which they not find data. Then, if we suppose that the B07’s data are repre-
to be associated with any real localized kinematical sub- sentative of the Leo I main stellar population, li1 would
structure (we will refer to this group of stars as k1). We account for ∼ 5% of it. Depending on the luminosity
plot this group of five stars as filled green squares in Fig- adopted (McConnachie 2012), a zero order mass estimate
ure 4. We observe that the clump reported in this work for li1 would be MLI1 ≈ (3.4 − 5) × 105 M⊙ . The calcu-
and k1 only have one star in common. Adding to K07’s lated mass for li1 is higher than for the stellar substruc-
and B07’s data, additional data of Leo I star members, tures in Sextans and Carina (see Section 4.1 and 4.2, re-
Mateo et al. (2008) (hereafter M08) presented kinematic spectively), but it is comparable to the mass of the most
results of stars in the Leo I dSph (see white plus sym- massive globular cluster in the Fornax dSph and Sagit-
bols in Figure 4). They found a group of six stars (see tarius dSph, ∼ (2 − 4.5) × 105 M⊙ (Mackey & Gilmore
filled blue squares in Figure 4) with velocities in a nar- 2003).
row range, with a dispersion of ∼ 5 km s−1 and a mean
velocity of ∼ 96 km s−1 (see horizontal dotted line in 4.4. The case of Leo II
Figure 4b). They argue that this group of stars might
be a kinematically cold group, but they warn that the For the case of Leo II, we used Koch et al.’s (2007b)
statistics are poor. They found this group of stars by (hereafter K207) data, together with B07 data.
plotting the heliocentric velocity as a function of the el- The targets in Koch et al. (2007b) were selected from
liptical radius (see our Figure 4b and Figure 9 of M08). photometry that was obtained by the Cambridge Astro-
For comparison, we over plotted the clump of stars re- nomical Survey Unit (Irwin & Lewis 2001) at the 2.5 m
ported in this work, li1 (magenta filled squares); and the Isaac Newton Telescope. The selection criteria of the tar-
gets involved colors and luminosities consistent with be-
10 Lora et al.

Fig. 5.— We show the sky position of K207 and B07’s Leo II data as open circles, and S17 data as plus symbols. The two substructures
found in the S17 data are shown as filled cyan squares, and filled purple squares. The (cyan/purple) stars show the position of the centroid
of each of the substructures. The two sold ellipses show the core (2.9’) and tidal (8.7’) radius. The black star shows the position of the
center of the substructure reported by Komiyama et al. (2007), the white region shows the extent of the substructure (4 × 2.5 arcmin2 ).

ing members of Leo II’s RGB stars. Koch et al. (2007b) 3σ were taken to be Galactic foreground stars. After per-
obtained VLT/FLAMES spectroscopic observations in forming the kinematic cut, they still had 11 stars with
the near-infrared CaT region, using the GIRAFFE mul- velocities and positions similar to those of Leo II stars.
tifiber spectrograph in low-resolution mode, centered at Therefore, they applied an extra cut in the data based
the near-infrared CaT at 8550 Å. To eliminate Galac- on stellar surface gravities. From both criteria, they end
tic contaminants, Koch et al. (2007b) determined indi- with a total of 175 Leo II members.
vidual radial velocities by means of cross-correlation of We look for stellar substructures in the Leo II dSph,
the three CaT lines against synthetic Gaussian template combining the RGB star data of K207 and B07. We
spectra. The templates were synthesized adopting rep- applied the brute-force method setting the ranges for
resentative equivalent widths of the CaT in RGB stars. the metallicity meti = −3.4 dex, metf = −1. dex
The spectroscopic metallicity and age distributions were and ∆met = 0.3 dex. For the velocities we set vi =
derived using the near-infrared CaT calibration method −20 km s−1 and vf = 20 km s−1 , with ∆v = 4 km s−1 .
(Koch et al. 2007c). The typical measurement errors for the metallicity and
On the other hand, B07 obtained CaT abundances and the velocity (as in Leo I) are 0.11 dex and 5.5 km s−1 ,
radial velocities for 74 RGB stars in Leo II, using the respectively.
low-resolution spectrograph LRIS. They obtained a mean Using the brute-force algorithm and the MST method,
S/N of 23. They verified the membership of their RGB we did not find any significant stellar substructure in
stars by deriving their radial velocities. From their 90 the K207 + B07 data set. The non-detection of stellar
stars observed in Leo II, 83 have heliocentric velocities substructures in these data might be related to the small
within 3σ of the average velocity, and thus are designated sample size (128 with K207 and B07 combined), and/or
as members of Leo II. B07 estimated the metallicities of to the large Galactocentric distance of Leo II.
their sample relating the equivalent widths of the CaT. On the other hand, analyzing the S17 data with the
It has to be noted that K207 and B07 data sets do not brute-force algorithm we found two considerable groups
overlap. of stars. In Figure 5 we show K207 combined with B07
In addition, Spencer et al. (2017a) (hereafter S17) ob- data as open circles; and K17 data as plus symbols. The
tained a large data set of RGB member candidates of two groups of stars are shown as filled cyan squares (here-
Leo II, which were separately analyzed. S17 performed after lii1), and filled purple squares (hereafter lii2).
spectroscopic observations with the Multiple Mirror Tele- The substructure lii1 is the most metal-poor group
scope using Hectochelle, a multifiber, single-order echelle that we find. Five stars make up lii1. It has a mean
spectrograph. They obtained simultaneous estimates metallicity of −2.35, a mean velocity of 72.4 km s−1 , and
of radial velocity, effective temperature, surface gravity, a velocity dispersion of 1.4 km s−1 , in a elliptical annulus
and metallicity by fitting a library of smoothed, syn- width of 5 arcmin. The substructure lii2 is constituted
thetic stellar spectra to each Hectochelle spectrum in of seven stars. The mean metallicity is −2, the mean
pixel space (Walker et al. 2015). In order to separate velocity is 78.8 km s−1 , and it has a velocity dispersion
stellar members from nonmembers they employed a ve- of 1.7 km s−1 , in a elliptical annuli width of 6 arcmin.
locity cut. Again, stars with radial velocities larger than We compute the velocity dispersion of all the stars in
 Old stellar substructures in MW’s dSphs 11

the S17 data belonging to the same annulus as LII1 and Moreover, if we relax the constraint on the velocity
LII2. We found that the velocity dispersion of the ellip- (3.7 − 16 km s−1 ) we can add two more stars to the
tical annulus associated with LII1 is 7 km s−1 , and the clump maintaining the same low metallicity range. The
velocity dispersion associated with LII2 is 7.6 km s−1 . distance from the center of Sextans to the center of the
In Figure 6a we show the systemic velocity as a func- ten-star clump is ∼ 751 pc, and the velocity dispersion
tion of the elliptical radius, where we can clearly observe is cold (σ ≃ 2.01 km s−1 ). If the stars of the sxt2
that the substructures appear elongated (about 4 to 6 clump belong to a disrupted globular cluster then these
arcmin). The average velocity of lii1 is very similar to low metallicities would suggest that it would be one of
the systemic velocity of Leo II (δ = 0.3 km s−1 ). Fig- the most metal-poor globular clusters known. It is very
ure 6b shows the metallicity as a function of the elliptical encouraging to see that sxt2 lies in the densest region
radius. 25% of the stars in the S17 data have metallici- reported by Roderick et al. (2016).
ties lower than ∼ −2. Moreover, only 8.6% of the stars We also find a cold stellar substructure close to the
in the S17 sample have metallicities lower than ∼ −2.3, core (240 pc) of the Carina dSph. This substructure
and one third of those stars are contained in lii1. car1, consists of nine stars with metallicities ranging
The velocity as a function of metallicity is shown in from −2.89 to −2.44 dex and a velocity dispersion of
Figure 6c. In the velocity-metallicity space, one can σ ≃ 1.88 km s−1 . Such a substructure resembles a dis-
clearly observe both substructures clumped together. rupted globular cluster, very similar to that found by
The mean velocity of lii1 is 1σ away from Leo II’s aver- Battaglia et al. (2011) in Sextans (sxt1). It has to be
age velocity (shown as a shaded region in Figure 6c). The noted that the distance from the center of Carina to the
low metallicity of lii1 together with its velocity makes center of Carina’s substructure is only ∼ 14.4 pc, i.e.,
it a thought-provoking stellar substructure. very close to the center of Carina. Interestingly, the
It is worthwhile mentioning that Komiyama et al. car1 substructure could possibly be related to the sub-
(2007) carried out wide-field V, I imaging of Leo II ex- structure reported by Fabrizio et al. (2011), since both
tending far beyond Leo II’s tidal radius. They reported lie close to the core of Carina.
the existence of a substructure in the eastern part of the Analyzing Leo I, we find a new substructure, besides
galaxy (containing four bright RGB stars) with a physi- the one reported by K07 and M08. The li1 substructure
cal size of 270 × 170 pc2 located beyond the tidal radius has six stars with metallicities ranging from −1.78 to
(8.63′ ), with a luminosity close to that of a globular clus- −1.52 dex, an average velocity of 286 km s−1 , and a ve-
ter (see black star in Figure 5). They suggest that this locity dispersion of σ ≃ 1.6 km s−1 . After the statistical
substructure is a disrupted globular cluster that is merg- analysis, we concluded that li1 is not significant.
ing with the main stellar component of Leo II. In the dSph galaxy Leo II, we found two significant
Interestingly, two of the stars in lii1 are located be- groups of stars, lii1 and lii2. From those two groups,
yond the tidal radius, and the interpretation could be we found that only lii1 is statistically significant. Even
similar to that of the knotty stellar structure reported if the probability of finding lii1 is very low (0.5%), we
by Komiyama et al. (2007). We computed a first order have to mention that we are dealing with low number
approximation on the stellar mass of lii1 making the as- statistics.
sumption that the number of RGB stars in the S17 data The finding of new stellar substructures is of major
is representative of Leo II. Then, lii1 would account for importance, since it leads to the idea that the early evo-
∼ 3% of Leo II’s total stellar mass. Taking a luminosity lutionary histories of dSph are as complicated as the ones
of LV = 7.4 × 105 L⊙ (Coleman et al. 2007) and a typi- of massive galaxies. In particular, stellar substructures
cal value (for globular clusters) of the mass-to-light ratio are of great use when investigating the DM profiles of
equal to two, then lii1’s mass would be ∼ 4.4 × 104 M⊙ . dwarf galaxies. For example, using N -body simulations,
We perform a statistical study to investigate whether one can study the survival of old cold substructures in
random groupings of stars in the S17 data are rare. We the DM halo of dSphs against phase mixing, and com-
carried out Monte Carlo tests of normal distributions for pare the evolution of the stellar substructures when the
both metallicity and velocity (see Section 3.1). We find dark halo has a core and when the dark halo has a cuspy
that the probability of finding LII1 is ∼ 5% (i.e., 2.1σ). profile (Kleyna et al. 2003; Lora et al. 2009, 2012, 2013;
On the other hand, the probability of finding LII2 is Contenta et al. 2017; Amorisco 2017). The existence of
only 1.27σ. Therefore, we conclude that LII2 is not sta- the old stellar clump in UMi (Kleyna et al. 1998) is in
tistically significant. Indeed, LII1 could be real stellar agreement with a cored DM profile rather than a cuspy
debris, but we have to keep in mind that we are dealing NFW one. If the old stellar clump in UMi is dropped in
with a low number of Leo II members. a NFW cuspy profile, the stellar clump gets disrupted in
One must keep in mind that a high fraction of binary the first Gyr (Kleyna et al. 2003; Lora et al. 2009, 2012,
stars could falsify radial velocity measurements. Re- 2013).
cently, Spencer et al. (2017b) determined a binary frac- Lora et al. (2013) also found that a cored DM profile
tion of 0.3 − 0.34 for the Leo II dSph galaxy. is needed in order to guarantee the survival of the old
stellar substructure found in Sextans by Battaglia et al.
5. CONCLUSIONS (2011), and the one found by Walker et al. (2006). Then,
In this paper we searched for stellar substructures in the new stellar substructures found in this work will be
four dSph galaxies that are satellites of the MW. We crucial to further investigate the core/cusp problem in
were able to find the stellar substructure reported by other dSphs. Therefore, we will perform N-body simula-
Battaglia et al. (2011), and a new substructure in Sex- tions of the four dSph studied in this paper and the new
tans, sxt2. The latter stellar substructure consists of stellar substructures to investigate their evolution within
eight stars with metallicities from −2.99 to −2.63 dex. their parent DM halo.
12 Lora et al.

Fig. 6.— Combined Leo II data of K07 and B07 as black plus symbols. The data of S17 shown as open black circles. The two substructures
found are plotted as filled cyan squares (lii1), and filled purple squares (lii2). In panel (a) we show the radial velocity as a function of
the elliptical radius. In panel (b) we show the metallicity as a function of the elliptical radius. The vertical dashed lines in panels (a)
and (b) indicate the core and tidal radius of Leo II. In panel (c), we show the radial velocity as a function of the metallicity. The two
vertical/horizontal solid lines represent the mean metallicity/velocity for each of the substructures. The dashed vertical and horizontal
lines indicate the mean value of the metallicity and velocity of S17 sample. The shaded zone indicates the 1σ velocity region.
Given the very small number of stars found to be as- structures reported in this work might shed light on the
sociated with the substructures in the current study and way galaxies assemble through mergers at very small
in others, and given how difficult it is to destroy GCs in scales.
the absence of strong tidal fields, one might also suggest
that we are looking here at the possible remnants of very
old, metal-poor, low-mass clusters akin to open clusters. V.L. gratefully acknowledges support from the
Either way, another interesting question is how such clus- CONACyT Research Fellowships program. V.L. thanks
ters were able to form to begin with in systems that we Giuseppina Battaglia, Mathew Walker and Yutaka
see nowadays as very low stellar density objects. More- Komiyama for making their data available. V.L. thanks
over, one could argue that these discoveries may add to Alejandro Raga, Gustavo Bruzual and Sundar Srinivasan
the recent discoveries of old clusters in low-mass dSphs, for very helpful comments, suggestions and discussions
making them a more common occurrence than previously which resulted in an improved version of this paper.
thought. S.S. was supported by Sonderforschungsbereich SFB
It is worthwhile to mention that Amorisco et al. (2014) 881 “The Milky Way System” (subproject B5, funding
point out that mergers of low-mass galaxies are expected period 2011-2014) of the German Research Foundation
within the hierarchical model of galaxy formation. More- (DFG) during part of this work.
over, they report the kinematic detection of a stellar EKG and AK were supported by Sonderforschungs-
stream in the dSph Andromeda II, which they suggest bereich SFB 881 ”The Milky Way System” (subpro-
could be the remnant of a merger between two dwarf jects A02 and A08) of the German Research Foundation
galaxies (see also Koch et al. 2012a for low-mass range). (DFG).
Thus, further study on the properties of the stellar sub- We thank the anonymous referee for very kind and
useful comments that improved the presentation of this
paper.

REFERENCES
Amorisco, N. C., Evans, N. W. & van de Venn, G., 2014, Nature, Contenta, F., Gieles, M., Balbinot, E. & Collins, M. L. M., 2017,
Volume 507, Issue 7492, 335 MNRAS, 466, 1741
Amorisco, N. C., 2017, ApJ, 844, 64 Cooper, A. P., D’Souza, R., Kauffmann, G., et al. 2013, MNRAS,
Allison, R. J., Goodwin, S. P., Parker, R. J., et al. 2009, MNRAS, 434, 3348
395, 1449 de Boer, T. J. L., Tolstoy, E., Lemasle, B., Saha, A., Olszewski,
Battaglia, G., Tolstoy, E., Helmi, A., Irwin, M., Parisi, P., Hill, V. E. W., Mateo, M., Irwin, M. J. & Battaglia, G. 2014, A&A
& Jablonka, P., 2011, MNRAS, 411, 1013 572, A10
Bellazzini, M., Gennari, N., Ferraro, F. R. & Sollima, A., 2004, Dekel, A. & Silk, J., 1986, ApJ, 303, 39
MNRAS, 354, 708 del Pino, A., Aparicio, A. & Hidalgo, S. L., 2015, MNRAS, 454,
Bellazzini, M., Gennari, N. & Ferraro, F. R., 2005, MNRAS, 360, 3996
185 Fabrizio M., Nonino, M., Bono, G., et al. 2011, PASP, 123, 384
Bono, G., Stetson, P. B., Walker, A. R., et al. 2010, PASP, 122, Fabrizio M. et al. 2016, ApJ, 830, 126
651 Fritz, T. K., Battaglia, G., Pawlowsky, M. S., Kallivayalil, N., van
Bosler, T. L., Smecker-Hane, T. A. & Stetson, 2007, MNRAS, der Marel, R., Sohn, S. T., Brook, C. & Besla, G. 2018, A&A
378, 318 619, A103
Boylan-Kolchin, M., Bullock, J. S., Sohn, S. T., et al. 2013, ApJ, Gaia Collaboration, Helmi, A. et al. 2018, A&A, 616, A12
768, 140 Grebel, E. K., 2000, in Star Formation from the Small to the
Bullock, J. S. & Johnston, K. V., 2005, ApJ, 635, 931 Large Scale, ed. F. Favata, A. A. Kaas, & A. Wilson (ESA
Chen, J., Bryan, G. L. & Salem, M., 2016, MNRAS, 460, 3335 SP-445; Noordwijk: ESA), 87
Cicuendez, L. et al. 2018, A&A, 609, 53 Grebel, E. K., Gallagher, J. S. & Harbeck, D., 2003, AJ, 125,1924
Coleman, M. G., Jordi, K., Rix, H. W., Grebel, E. K. & Koch, A. Grebel, E. K. & Gallagher, J. S., 2004, ApJ, 610L, 89
2007, AJ, 134, 1938 Harbeck, D., et al. 2001, AJ, 122, 3092
 Old stellar substructures in MW’s dSphs 13

Hayashi, K., Fabrizio, M.,Lokas E. L., Bono, G., Monelli, M., Mateo, M., Olszewski, E. W., Pryor, C., Welch, D. L. & Fischer,
Dall’Ora, M. & Stetson, P. B. 2018, MNRAS, 481, 250 P. 1993, AJ, 105, 510
Hurley-Keller, D., Mateo, M. & Nemec, J., 1998, AJ, 115, 184 Mateo M. L., 1998, ARA&A, 36, 435
Ibata, R., Chapman S., Irwin M. et al., 2006, MNRAS, 373, L70 Mateo M., Olszewski, E. W. & Walker, M. G. 2008, ApJ, 675, 201
Irwin, M., & Lewis, J. 2001, NewA Rev., 45, 105 McConnachie, A. W., Arimoto, N. & Irwin, M., 2007, MNRAS,
Irwin, M. & Hatzdimitriou, D., 1995, MNRAS, 277, 1354 379, 379
Kim, Hak-Sub, Han, Sang-Il, Joo, Seok-Joo, Jeong, Hyunjin, McConnachie, A. W. 2012, AJ, 144, 4
Yoon & Suk-Jin. 2019, ApJ, 870 McLaughlin, D.E. & van der Marel, R. P., 2015, ApJS, 161, 304
Kirby E. N., Lanfranchi G. A., Simon J. D., Cohen J. G. & Monelli, M., Pulone, L., Corsi, C. E., et al., 2003, AJ, 126, 218
Guhathakurta P., 2011, ApJ, 727, 78 Monelli, M. et al. 2014, ApJ, 796, 90
Kleyna, J. T., Geller, M. J., Kenyon, S. J., Kutz, M. J. & Muñoz, R. R., Majewski, S. R., Zaggia, S., et al., 2006, ApJ, 649,
Thorstensen, J. R., 1998, ApJ, 115, 2359 201
Kleyna, J. T., Wilkinson, M. I., Gilmore, G. & Evans, N. W., Muñoz, R. R. et al. 2018, 860, 66
2003, ApJ, 588, L21 Navarro, J. F., Frenk, C. S., & White, S. D. M. 1997, ApJ, 490,
Kleyna, J. T., Wilkinson, M. I., Evans, N. W. & Gilmore G., 493
2004, MNRAS, 354, L66 Nonino, M., et al. 1999, A&AS, 137, 51
Koch, A., Grebel, E. K., Wyse, R. F. G. et al. 2006, AJ, 131, 895 Olczak, C., Spurzem, R., & Henning, T., 2011, A&A, 532, A119
Koch, A., Wilkinson, M. I., Kleyna, J. T., et al. 2007a, ApJ, 657, Pasetto, S., Grebel, E. K., Berczik, P., Spurzem, R., & Dehnen,
241 W. 2010, A&A, 514, 47
Koch, A., Kleyna, J. T., Wilkinson, M. I., et al. 2007b, AJ, 134, Piatek, S., Pryor, C. & Olszewski, E. W. 2016, AJ, 152, 166
566 Pillepich, A., Madau, P. & Mayer, L., 2015, ApJ, 799, 184
Koch, A. et al. 2007c, AJ, 133, 270 Roderick, T. A., Jerjen, H., Da Costa, G. S. & Mackey, A. D.,
Koch, A., Grebel, E. K., Gilmore, G. F. et al., 2008, AJ, 135, 1580 2016, MNRAS 460, 30
Koch, A., Burkert, A., Rich, R. M. et al. 2012, ApJ, 755, 13 Sales L. V., Helmi A., Battaglia G., 2010, Advances in
Koch, A., Lépine, S. & Çalişkan, Ş., 2012, EPJWC, 19, 3002 Astronomy, 2010, 194345
Komiyama, Y. et al., 2007, AJ, 134, 835 Santana, A. S. et al. 2016, ApJ, 829, 86
Kordopatis, G., Amorisco, N. C., Evans, N. W., Gilmore, G. & Schmeja, S., 2011, A.N., 332, 172
Koposov, S. E., 2016, MNRAS, 457, 1299 Simon, J. D. & Geha, M., 2007, ApJ, 670, 313
Lépine S., Koch A., Rich R. M. & Kuijken, K., 2011, ApJ, 741, Spencer, M. E., Mateo, M., Walker, M. G. et al. 2017, AJ, 153,
100 254
Lokas E. L., 2009, MNRAS, 394, L102 Spencer, M. E., Mateo, M., Walker, M. G. & Edward W.
Lora, V., Grebel, E. K., Sánchez-Salcedo, F. J. & Just, A., 2013, Olszewski, E. w., 2017, ApJ, 836, 202
ApJ, 777,65 Sohn, S. T. et al., 2013, ApJ, 768, 139
Lora, V., Just, A., Sánchez-Salcedo, F. J. & Grebel, E. K., 2012, Ural, U., Wilkinson, M. I., Koch, A. et al., 2010, MNRAS, 402,
ApJ, 757, 87 1357
Lora, V., Sánchez-Salcedo, F. J., Raga, A. C. & Esquivel, A., Vogt, S. S., Mateo, M., Olszewski, E. W. & Keane, M. J., 1995,
2009, ApJ, 699, L113 AJ, 109, 151
MacKey A. D. & Gilmore G. F., 2003, MNRAS, 340, 175 Walker, M. G., Olszewski, E. W., & Mateo, M. 2015, MNRAS,
Mackey, A. D., Koposov, S. E., Erkal, D. et al. 2016, MNRAS, 448, 2717
459, 239 Walker, M. G., Mateo, M., Olszewski, E. W., Pal, J. K., Sen, B.
Majewski, S. R., Skrutskie, M. F., Weinberg, M. D., & & Woodroofe, M., 2006, ApJ 642, L41
Ostheimer, J. C. 2003, ApJ, 599, 1082 Zucker, D. B., Belokurov, V., Evans, N. W., et al. 2006, ApJ, 643,
Mashchenko, S., Couchman, H. M. P., & Wadsley, J. 2006, Natur, L103
442, 539
Mateo M., Nemec J., Irwin M. & McMahon R., 1991, Astron. J.,
101, 892