THE STRUCTURE OF THE LOW SURFACE BRIGHTNESS
STELLAR HOST IN BLUE COMPACT DWARF GALAXIES
gotzon
18-1-95
N. Caon1, L.M. Cairós2, J.A.L. Aguerri1, C. Muñoz-Tuñón1, P. Papaderos2 , K. Noeske2
(1) Instituto de Astrofísica de Canarias
(2) Universitäts Sternwarte Göttingen
WHAT ARE BCD GALAXIES ?
THE STUDY OF THE LSB HOST IS FUNDAMENTAL...
Blue Compact Dwarf galaxies are compact and low luminosity
systems (starburst diameter < 1 kpc, MB > -18), which have
spectra typical of the HII regions in spiral galaxies [ref. TM81].
They also are low metallicity objects (Z < Z /2), which form
stars at high rates.
Provided that Dark Matter does not dominate within the optical extent
of a BCD, the LSB host forms, together with the HI gas, the
gravitational potential within which gas collapses to form stars. Its
study provides important diagnostics for the understanding of the
regulation of star-forming activity in BCDs [P96a].
Initially it was hypothesized that BCDs are truly young galaxies,
forming their first generation of stars, however...
The comparison of the LSB structural parameters (scale-length,
central surface brightness, total luminosity, average colors and color
gradients) with those of other dwarf classes (dEs, dIs) will help assess
the possible evolutionary links among dwarf galaxies [P96b, MMH99].
THEIR LOW SURFACE BRIGHTNESS
STELLAR HOST
...today it is well established that the vast majority of BCDs have
an extended stellar host, as revealed by deep photometric
studies [P96a, D97, C01a, C01b, BO02, N03, C03].
This low surface brightness (LSB) component can extend
several kpc from the central star-forming (SF) region, shows
regular, elliptical isophotes, and displays red colors, indicative
of an age of typically a few Gyr [P96b, BO02, C02]. See examples
in Figures 1 and 2.
The study of the structural parameters of the LSB host has
recently become a hot topic, and an increasing number of
works are now devoted to it .
Modeling and subtracting the line-of-sight contribution of the LSB
host within the SF region is indispensable to determine the true
broad-band colors and emission line equivalent widths of the
starburst knots [C02].
Figure1: Texto
Figure 1: B-band image of Mrk 35
Figure 2: B-band image of Mrk 86
Mrk 35 belongs to the most common
BCD type, iE [LT86]. It shows a central
star-forming bar, which extends to
southwest where two other faint knots
are visible. The starburst is about 25
arcsec (1.7 kpc) in radius; outside, a
regular outer envelope, well described
by elliptical isophotes, is clearly
visible. Its light profile can be traced
out to 60 arcsec (4 kpc).
Mrk 86 is the prototype of the iE BCD
galaxy class. It consists of a complex
ensemble of about 80 star-forming
knots, which occupy a region of about
45 arcsec (1.5 kpc) in radius. The light
profile of the outer stellar envelope can
be traced out to 130 arcsec (4.5 kpc).
Stellar and ionized gas emission from the SF component outshines
the LSB host in its inner and intermediate parts. The study of the LSB’s
structure relies entirely on the outer galaxy regions, so it is necessary
to map it down to very deep surface brightness levels (µB > 28 mag
arcsec-2 ) .
Whether the exponential model, widely used to describe the light
profile of the LSB host, is the most appropriate fitting function is
currently matter of debate. In fact, some BCDs show systematic
deviations which call for alternative laws [P96a, C01a, N03].
HOW CAN WE OVERCOME THESE
PROBLEMS?
OUR STUDY
From the 28 galaxies forming our BCD sample [C01a, b], we selected
for the present study those with the most regular and extended LSB
hosts. For most galaxies we have optical data in U, B, V, R, and I, and
for part of them, also near-infrared data in J, H and K.
In this initial stage, we limit our analysis to the B, V, R light profiles,
which are generally those with the best photometric quality.
There are three basic requirements for a meaningful study of
the LSB host structure:
1) Deep images, both in the optical and in the near-infrared.
We took long exposures at 2-4m class telescopes, and took
great care to accurately perform flat-fielding correction and skybackground subtraction, which are crucial steps to reach
surface brightness limits of 28 mag arcsec-2 in B and of 23-24
mag arcsec-2 in K.
2) Proper determination of the extension of the SF region.
We identified the spatial region occupied by the starburst by
examining, for each object, its optical color and Ha emission
maps. See Figures 3 and 4.
We could thus define a radius Rtran outside which the starburst
emission is practically absent. Profile fitting must be done for
R > Rtran.
... BUT IT IS ALSO QUITE DIFFICULT.
After excluding two objects, the first because it sits on the psf wings
of a very bright star, the other because it has a very inhomogeneous
sky background, we ended up with eight BCD galaxies.
Figure 3: V-R color map of Mrk 35
Figure 4: continuum-subtracted Ha
map of Mrk 35
3) Choice of a suitable fitting function.
It is nowadays recognized that several LSB hosts have profiles
which show systematic deviations from the simple exponential
model which has been generally adopted so far. The Sérsic law
[S68], successfully used to parametrize the shape of light
profiles of elliptical galaxies (including dEs) and spiral bulges
[e.g. CCD93], is the most natural alternative. But ...
We corrected first for small residual gradients in the sky background.
Then we derived, for each galaxy, the light profiles of the LSB host by
fitting ellipses to its isophotes, after masking out foreground and
background objects, small isolated knots, and other disturbances.
The light profiles (expressed as function of the equivalent radius)
were fitted by a Sérsic law, taking care to exclude the SF region as
well as those outermost datapoints affected by large uncertainties.
The consistency checks explained above helped us determine the
best fitting radial range for each profile.
For each object we thus produced a final list of Sérsic parameters,
which is shown below in Table 1.
Table 1: Derived Sérsic parameters.
THE DRAWBACKS OF FITTING A
SÉRSIC LAW
m(R) = me+ cn .[1- (R/Re)1/n]
... the Sérsic law has important drawbacks when applied to the
light profiles of BCDs, drawbacks that must not be
underestimated, and require a thorough investigation before it
can be used for a systematic study of the structure of the LSB
host.
Galaxy Band Rtran” Rmax” µmin µmax
Figure 5: Fits done in three different
radial intervals show how much the
index n can depend on the choice of the
fitted radial range. Filled points mark the
fitted interval. For this galaxy, Rtran = 20
arcsec.
Figure 6: The green line represents the
best Sérsic fit to the observed light
profile. The red and blue solid lines are
the best fits obtained after over- and
under-subtracting the sky background,
respectively.
Figure 7: Plotting the values of n, Re and
µe as a function of Rtran, for three
different choices of Rmax, is a quick and
effective way of estimating the
uncertainties on the Sérsic parameters.
Figure 8: The Sérsic fits to the B, V and R
profiles of the LSB host in Mrk 370 give
values of n very close to 1. Filled points
mark the fitted radial interval.
1) Sensitivity to the radial range. For small radial (Rtran > 1-2
scale lengths) and surface brightness intervals (typically less
than 4 mag), the Sérsic law parameters become very sensitive to
the choice of the fitted radial range. This sensitivity is
exacerbated if the fitted interval includes part of the SF region,
which usually produces a higher value of n. See Figure 5.
2) Sensitivity to sky-subtraction errors. When working at such
low surface brightness levels, small errors on the determination
of the sky background level can alter significantly the derived
Sérsic parameters. See Figure 6.
3) The derived Sérsic parameters may also depend, in a lesser
degree, on whether the fit is done weighting the datapoints or
not, on whether the profile is resampled to a constant step, or,
for instance, whether the major axis or equivalent radius profile
is fitted. Moreover, at faint surface-brightness levels different
algorithms to extract 1-D profiles from a 2-D image may give
slightly different results.
CONSISTENCY CHECKS:
1) Plot the resulting values of n, Re and µe as function of
Rtran; check whether these values are stable, or show a strong
dependence on Rtran (also, different values for Rmax should be
explored). See Figure 7.
2) As we expect that the LSB host has flat color profiles, the
Sérsic parameters n and Re should be the same, within the
uncertainties, in all passbands. See Figure 8.
Table 3
REFERENCES
[BO02]
[C01a]
[C01b]
[C02]
[C03]
[CCD93]
[D97]
[LT86]
[MMH99]
[N03]
[P96a]
[P96b]
[S68]
[TM81]
Bergvall N. & Östlin G. 2002, A&A 390, 891
Cairós L.M. et al 2001, ApJS 133, 221
Cairós L.M. et al 2001, ApJS 136, 2
Cairós L.M. et al 2002, ApJ 577, 164
Cairós L.M. et al 2003, submitted to ApJS
Caon N., Capaccioli M. & D’Onofrio M. 1993, MNRAS 265, 1013
Doublier V. et al 1997, A&AS 124, 405
Loose H.-H. & Thuan T.X. 1986 in “Star-forming Dwarf Galaxies and Related Objects”
Marlowe A.T., Meurer G.R. & Heckman T.M. 1999, ApJ 522, 183
Noeske K. et al 2003, submitted to A&A
Papaderos P. et al 1996, A&AS 120, 207
Papaderos P. et al 1996, A&A 314, 59
Sérsic J.L. 1968, “Atlas de galaxias australes”
Thuan T.X. & Martin G.E. 1981, ApJ 247, 823
n
Re” µe
mtot
Mrk 5 B 13.78 34.58 23.99 27.63 2.83 3.35 20.23 14.39
V 13.67 34.00 23.56 26.92 2.46 4.89 20.97 14.38
R 13.64 35.65 23.15 26.76 2.32 5.57 20.96 14.12
Mrk 35 B 24.57 59.24 23.86 28.33 1.11 12.90 22.26 13.96
V 24.37 63.82 23.09 28.15 0.93 14.98 21.99 13.45
R 24.08 63.10 22.73 27.67 1.30 12.10 21.01 12.77
Mrk 36 B 12.18 23.92 24.93 27.60 0.95 8.24 24.05 16.80
V 12.39 33.03 24.53 28.89 1.11 7.36 23.27 16.18
R 12.07 26.44 24.04 27.09 1.21 7.48 22.92 15.76
Mrk 86 B 45.97 97.93 24.15 27.24 0.87 32.69 23.42 13.21
V 45.52 122.35 23.29 26.81 1.46 30.78 22.44 12.11
R 46.33 128.71 22.89 26.52 1.81 27.84 21.74 11.53
Mrk 314 B 16.85 39.40 23.69 26.93 1.42 9.98 22.49 14.62
V 16.15 38.10 23.50 26.71 1.21 11.03 22.69 14.69
Mrk 370 B 35.65 84.77 24.72 28.36 0.97 24.70 23.89 14.25
V 35.62 75.03 23.95 26.85 1.01 24.16 23.09 13.47
R 35.50 91.22 23.42 27.65 0.98 24.12 22.56 12.96
Izw 123 B 9.17 25.49 24.03 28.48 2.61 2.07 19.95 15.19
V 9.07 24.07 23.29 26.82 3.01 2.74 20.26 14.84
Tol 127 B 7.22 19.74 24.22 29.06 0.95 4.86 23.32 17.21
V 7.16 17.50 23.54 27.42 0.83 5.27 22.88 16.66
R 7.20 19.87 23.15 28.09 0.77 5.54 22.61 16.32
RESULTS
Inspection of Table 1 shows that, for the majority of the galaxies
in the present sample, Sérsic fits to the light profiles of the LSB
host give indexes n close to 1, and effective radii which are the
same, within the uncertainties, in the different filters. We can
then affirm that most LSB hosts have exponential profiles.
In two galaxies, however, n is about 2.5-3.0, which indicates
that not all the LSB hosts share the same structure.
An investigation of the possible correlations among the LSB
host structural parameters, as well as the comparison with
those of other dwarf types, based on a larger galaxy sample,
will be the next step along this research line.
Visit our BCD atlas: http://www.iac.es/proyect/GEFE/BCDs/