AMIGA: Isolation degree of galaxies in low density environments.
S. Verley1,2, S. Odewahn3, S. Leon2, L. Verdes-Montenegro2, D. Espada2, E. Garcia2, U. Lisenfeld2,4, J. Sabater2, J. Sulentic5
(1) LERMA, Observatoire de Paris, 61, Av. de l'Observatoire, 75014 Paris, France - simon.verley@obspm.fr
(2) Instituto de Astrofísica de Andalucía, CSIC, Apdo. 3004, 18080, Granada, Spain - simon, stephane, lourdes, daniel, garcia, ute, jsm@iaa.es
(3) McDonald Observatory, University of Texas at Austin, Austin, TX 78712, USA - sco@astro.as.utexas.edu
(4) Universidad de Granada, Granada, Spain - ute@ugr.es
(5) Department of Astronomy, University of Alabama, Tuscaloosa, USA - giacomo@merlot.astr.ua.edu
Abstract: The evolutionary history of galaxies is thought to be strongly conditioned by the environment. In order to quantify and set limits on the role of nurture one must identify and study an
isolated sample of galaxies. But it is not enough to identify a small number of the "most isolated" galaxies. Instead one wishes to evaluate the slope of the nurture function as it approaches zero. We
begin with 950 galaxies from the Catalogue of Isolated Galaxies (Karachentseva 1973) and reevaluate isolation using an automated star-galaxy classification procedure on large digitized POSS-I fields.
We define, compare and discuss various criteria to quantify the degree of isolation for these galaxies: Karachentseva's revised criteria, local surface density computations and an estimation of the external
tidal force affecting each isolated galaxy. Comparison of multiwavelength ISM measures for these subsamples will allow us to search for the slope of nurture induced effects as it approaches zero.
INTRODUCTION
THE ADVANTAGES
The role of the environment on galaxy evolution is still not fully understood. In order to quantify and set limits on the role
of nurture one must identify and study a sample of isolated galaxies. The AMIGA project, "Analyzing the Interstellar
Medium of Isolated Galaxies", is doing a multiwavelength study of a large sample of isolated galaxies in order to examine
their star formation activity and interstellar medium.
We define, compare and discuss various criteria to quantify the degree of isolation for these galaxies: e.g.
Karachentseva's revised criteria, local surface density computations, estimation of the external tidal force affecting
each isolated galaxy. We also apply our pipeline to a subsample of Abell Clusters and Hickson Compact Groups which
serve as control samples.
We begin with a large sample of reasonably isolated galaxies and end up with refined subsamples reflecting degrees of
isolation. Comparison of multiwavelength ISM measures (e.g. Halpha, FIR, radio continuum, HI: see, for instance, the
poster ”AMIGA: HI content of the most isolated galaxies.”, Espada et al. 2004 ) for these subsamples will allow us to
search for the slope of nurture induced effects as it approaches zero.
Our study carries out some improvements, compared with the original CIG study (Karachentseva et al. 1973, 1980, 1986):
(1) computer processing: systematic detection and classification of the sources;
(2) magnitude: we revised the catalogue up to a magnitude B = 17.5 while the previous catalogue was limited at B = 15.7;
(3) redshift: we use all the redshift informations currently available (cf. new surveys such as Sloan, and up to 42 references from
the literature) for the central isolated galaxy (we compiled velocities for 856 CIG galaxies in order to obtain reliable distance
estimation for the CIG galaxy, and hence to be able to evaluate a physical size around the CIG);
(4) isolation degree: we define homogeneous isolation degrees for the whole sample. These parameters allow to compare galaxy
properties as a function of isolation.
We use two different isolation parameters, estimations of the tidal forces and of the local density affecting the isolated galaxy:
(a) The tidal force per unit mass produced by a companion is proportional to McR-3, where Mc is the mass of the companion, Dc its
diameter, and R is its distance from the center of the primary. However, no information on either Mc or on the absolute R is
available in most cases. The dependence of Mc on the galaxy diameter Dc is uncertain ( M D ), we adopt γ = 1.5 (See Dahari 1984).
Since the redshifts of the companions are unknown, the diameters Dp of the primary galaxies are used as a scaling factors, i.e.,
D D D
, and R S D
(S is the projected separation between the primary galaxy and a given companion). Accordingly,
c
c
Our sample of isolated galaxies gathers 1051 objects, listed in the Catalogue of Isolated Galaxies (CIG - Karachentseva
1973). All of the CIG objects are found in the Catalogue of Galaxies and Clusters of Galaxies (Zwicky et al. 1960-1968;
CGCG) with mpg < 15.7 and δ > -30 (< 3% of the CGCG). The CIG sample was assembled with the requirement that no
similar sized galaxies with diameter d (between ¼ and 4 times diameter D of the CIG galaxy) lie within 20d.
c
c
M
p
p
D D
c
c
1.5
p
Q
R
3
S
3
where Q, defined by this equation, is a dimensionless estimation of the gravitational interaction strength.
DATA REDUCTION
(b) An estimation of the local density can be found by considering the kth nearest neighbor (an unbiased estimator can be obtained
if neither the central point nor the kth neighbor are counted, see Casterano & Hut 1985); for this parameter, only the companions of
equivalent size (0.25 to 4 DCIG) are taken into account. k is equal to 5, or less if there were not enough companions in the field:
Starting from the complete Catalogue, we first remove from our
study all galaxies with V < 1500 km/s (101 CIGs), because the
area searched for potential companions of the very nearby CIGs
would be extremely large. We reevaluate isolation using an
automated star-galaxy classification procedure (to M B 17.5 )
on large digitized POSS-I fields surrounding each CIG galaxy.
The images are reduced using AIMTOOL in LMORPHO
(Odewahn 1995), and GUI-driven star-galaxy separation
procedure is used to classify detected sources (with SExtractor Bertin & Arnouts 1996) as: STAR, GALAXY or UNKNOWN.
As a final step, we archive our catalogues in the form of simple
ASCII files; a CIG database manager (CIGWORK) has been
developed under LMORPHO to manage and evaluate these
catalogues.
k
To perform the GUI-based star-galaxy separation, we have
decided to use the logAREA vs. MAG_ISO SExtractor
parameter space to separate star-galaxy images measured on
POSS I images (FIGURE 1).
k 1
V r
, where V(rk) = 4π(rk/rp)3/3
k
To illustrate our two estimators, we also applied our whole pipeline to two galaxies belonging to the Hickson Compact Groups
HCG 51 and HCG 57, and to 5 galaxies in Abell Clusters.
RESULTS
FIGURE 1: A typical star-galaxy separation parameter space from a POSS-I E
image (CIG0714). All the points that lie above the curve defined by the RED points
will be classified as GALAXY. The points below this curve are classed as STAR.
Points that lie outside the spline range (brighter or fainter in MAG_ISO than the
extent of the red points) are classified as UNKNOWN.
As a second check, the LMORPHO-classified
objects are inspected visually on the POSS I
images and reclassified accordingly. On the
figure the blue ellipses indicate the galaxies
detected, the red ones overplot the stars and the
green circles mark the sources that were not
classified (FIGURE 2).
There is a good correlation between our two parameters qualifying the isolation
(FIGURE 6).
The range covered by our two isolation parameters in our CIG sample is
extended and reached the values comparable to those corresponding to galaxies
in denser environments such as HCGs and ACOs (the scale is logarithmic on
both axes).
This plot also shows the consistence of our two parameters as being good
tracers of the isolation degrees.
FIGURE 6: We show in red all the CIG galaxies. It can be seen that the
galaxies in ACOs (green points) and HCGs (blue) show higher
interaction level than the isolated sample, as expected.
As an example, we show in FIGURE 7 and FIGURE 8 the
two Hickson Compact Groups that we use to compare
with the CIG sample. For HCG51, the central galaxy
used to compute the influence of its neighbor was the
galaxy HCG51A. For HCG57, the primary galaxy was
HCG57A.
For the Abell Clusters, we use:
- ACO 76, central galaxy: LEDA 073397;
- ACO 119, central galaxy: UGC 579;
- ACO 154, central galaxy: IC 1635;
- ACO 194, central galaxy: NGC 547;
FIGURE 8: Hickson Compact Group 57
FIGURE 7: Hickson Compact Group 51
- ACO 262, central galaxy: NGC 705.
For the CIG sample, the mean tidal force value is 1.28 (standard deviation = 0.76), while the mean local density equals to 1.39 (stand
deviation = 0.56), see FIGURE 9 and FIGURE 10. TABLE 1 gives the numbers of galaxies as a function of the standard deviation from the
mean values of the local density and tidal forces estimations.
We searched the companion galaxies within a
minimum physical distance of 0.5 Mpc (FIGURE
3), centered on each CIG. We have selected
three different sizes for the fields, depending on
the recession velocity:
673 fields of 55 * 55 arcmin2;
134 fields of 110 * 110 arcmin2;
49 fields of 165 * 165 arcmin2.
When the velocity of the CIG was not known
(94 galaxies), we used fixed size fields of 55
arcmin2.
Standard Deviation
Note: The fields larger than 55'*55' are
composed by various 55'*55' fields, with
little overlaps between two adjacent
fields. We developed a tool to keep only
one source when an object was detected
more than once on various fields.
TABLE 1
-3 -2.5
-2 -1.5
-1
-0.5
0
0.5
1
1.5
2
2.5
3
3.5
Local density
2
6
19
39
78
124
157 217 143
71
29
22
3
1
1
Tidal forces
0
0
10
34
74
174
227 179 112
69
32
15
10
8
6
FIGURE 2: The distribution of galaxies around CIG0714 (the bottom-right galaxy).
Finally, we will also use POSS 2 red plates of all our
GALAXY objects to perform a final check of our final
catalogues of companions. The choice of POSS 2 instead of
POSS I for this second check will remove the detected plate
defects of the POSS I survey that could have passed
through our first revision and will provide a better spatial
resolution to distinguish compact galaxies from stars.
FIGURE 10: Distribution of the values for the tidal forces parameter. The values of
the HCGs and the mean of ACOs are also marked
FIGURE 9: Distribution of the values for the local density parameter. The values of
the HCGs and the mean of ACOs are also marked
FIGURE 3:Physical radius of the fields inspected for our CIG sample.
KARACHENTSEVA'S CRITERIA
Originally, the CIG sample was assembled with the requirement that no similar
sized galaxies with diameter d (between ¼ and 4 times diameter D of the CIG
galaxy) lie within 20d.
Adams et al. 1980 and Karachentseva 1986, refined isolation assigning codes:
0 = isolated galaxies;
1 = marginally isolated;
2 = definitely not isolated according to the criteria.
We show in FIGURE 4 the distributions of these isolation
codes.
A raw detection of background galaxies or over density of small
companions in the CIG fields can be seen in FIGURE 11. All the fields that
are located above the 2 sigma level show a very high level of companions
compared with the mean density found in the CIG fields.
Finally, in the following table, we give a first classification of the CIG
galaxies as a function of the tidal force estimation, ranging from 'very
isolated' to 'interacting', that will allow us to discuss the ISM properties of
the CIG galaxies as a function of the isolation degree (AMIGA project).
Very isolated
TABLE 2
Q
Number of galaxies
FIGURE 4: The distribution of the isolation codes
according to Karachentseva. The X axis represents our tidal
force distribution (see "The advantages" section for the
definition.)
Although the criteria used in our study are not equivalent to the
Karachentseva's selection, they have allowed us to find some of the CIGs
that failed her criteria (FIGURE 5). According to Karachentseva, a
perturbative companion can be 4 times bigger and 20 d away from the CIG
galaxy (this is a huge distance: 20d = 20*4DCIG = 80 DCIG!).
We could only cover this area for 67 fields (among them, we can attest
than 54 CIGs are isolated following Karachentseva's criteria). For the
remaining fields, we have found 299 CIG galaxies violating
Karachentseva's isolation definition (although we were not able to check
on the whole 80 times DCIG). Still, 651 CIG galaxies remain isolated
accordingly to Karachentseva, taking into account that we cannot assert
that some of these latter galaxies will not move from the 'isolated' to the
FIGURE 5: Green: the galaxies that are violating Karachentseva's
'not isolated' sample, with a more exhaustive study.
criteria. Red: still isolated galaxies.
isolated
0
25
Quite isolated Poorly isolated Interacting
1
318
2
460
3
120
27
FIGURE 11: The Y axis shows the (density of companions - mean density),
computed with all the companions found within 25 arcmin from the CIG.
SUMMARY AND FUTURE WORK
SUMMARY:
We have revised the isolation of 950 CIG galaxies, discussed Karachentseva's criteria, but also refined this complete sample in a
homogeneous way, by giving continuous parameters of isolation, based on local density and tidal forces estimations.
FUTURE WORK:
(1) We will perform a second visual check for our catalogues of companion galaxies on POSS 2 plates.
(2) In order to eliminate the contamination due to background galaxies, we will test the possibility to use the POSS 2 surveys (red
and blue) to reject the farthest galaxies on the base of their color, as well as the use of all the redshift informations available in the
literature/archive for the companions.
(3) We will also use nearby fields to compare with each CIG field and find any excess of companions (see Morgan et al 1998).
(4) One third of the CIGs in our sample belongs to SDSS fields: this will provide redshift for the candidate companions as well as
photometric information.
The AMIGA project will use the information on the isolation degree of the CIG galaxies in order to perform multiwavelength
studies of the ISM components as a function of the degree of isolation. This will also allow to study the effect of dwarf companions
on its properties.