star-forming galaxies as seen through emission lines Nearby and distant Maryam Shirazi

star-forming galaxies as seen through emission lines Nearby and distant Maryam Shirazi
Nearby and distant
star-forming galaxies
as seen through emission lines
Maryam Shirazi
ISBN 978–9–46–191906–9
Cover by Alireza Rahmati
Front: The star formation rate map of a nearby star-forming galaxy based on
the spectral energy distribution fitting. This image is based on Figure 5.8.
Back: Astrolabe: An ancient astronomical instrument that shows how the sky
looks at a specific place at a given time based on the projection of the celestial
sphere onto the plane of the equator. It has been used through the ages for finding
the local time and finding the time of celestial events such as sunrise or sunset.
The history of the astrolabe goes back to more than two thousand years ago and
was highly developed by Muslim astronomers.
Nearby and distant
star-forming galaxies
as seen through emission lines
Proefschrift
ter verkrijging van
de graad van Doctor aan de Universiteit Leiden,
op gezag van de Rector Magnificus Prof. mr. C. J. J. M. Stolker,
volgens besluit van het College voor Promoties
te verdedigen op dinsdag 15 oktober 2013
klokke 10:00 uur
door
Maryamosadat Sadatshirazi
geboren te Tehran
in 1977
Promotiecommissie
Promotor:
Co-promotor:
Prof. dr. M. Franx
Dr. J. Brinchmann
Overige leden:
Prof. dr. H. J. A. Röttgering
Prof. dr. A. G. G. M. Tielens
Prof. dr. P. van der Werf
Dr. S. Charlot (Institute d’Astrophysique, Paris, France)
Table of Contents
1 Introduction
1.1 Nebular physics . . . . . . . . . . . . . . . . . . . . . . . . . .
1.1.1 Nebular emission lines in star-forming galaxies . . . .
1.1.2 Measuring galaxy properties using emission line ratios
1.1.3 Emission lines as indirect tracers of massive stars . . .
1.1.4 Emission lines as indirect tracers of the ISM . . . . . .
1.2 Evolution in the properties of star-forming galaxies . . . . . .
1.2.1 Stellar mass evolution . . . . . . . . . . . . . . . . . .
1.2.2 Mass-metallicity evolution . . . . . . . . . . . . . . . .
1.3 Star and galaxy formation over cosmic times . . . . . . . . .
1.3.1 From clumps to bulges . . . . . . . . . . . . . . . . . .
1.4 Towards probing the small-scale properties of distant galaxies
1.5 This thesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.5.1 Observations . . . . . . . . . . . . . . . . . . . . . . .
1.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2 Strongly star-forming galaxies in the local universe with nebular
He IIλ4686 emission
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.2 Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.2.1 Sample selection and classification . . . . . . . . . . . . . .
2.2.2 AGN contamination estimation . . . . . . . . . . . . . . . .
2.3 Physical properties of the sample . . . . . . . . . . . . . . . . . . .
2.3.1 Mass measurements . . . . . . . . . . . . . . . . . . . . . .
2.3.2 Emission line derived parameters . . . . . . . . . . . . . . .
2.4 Model predictions . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.4.1 CL01 predictions for nebular He II emission . . . . . . . . .
2.4.2 Starburst99 predictions . . . . . . . . . . . . . . . . . . . .
2.4.3 The effect of binary evolution on the He II 4686 emission .
2.5 The origin of nebular He II 4686 emission . . . . . . . . . . . . . .
1
2
2
3
6
7
8
8
9
10
11
11
12
14
15
16
19
20
21
22
28
30
30
31
32
33
35
37
38
v
TABLE OF CONTENTS
2.6 Why are there galaxies with He II 4686 emission but no WR features?
2.7 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.8 Appendix A: Fitting models to the emission lines . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3 The physical nature of the 8 o’clock arc based on near-IR IFU
spectroscopy with SINFONI
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2 Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.1 Near-IR spectroscopy with SINFONI . . . . . . . . . . . . .
3.2.2 Data Reduction . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.3 HST Imaging . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.4 PSF Estimation . . . . . . . . . . . . . . . . . . . . . . . .
3.3 Analysis of the SINFONI data . . . . . . . . . . . . . . . . . . . .
3.3.1 Nebular emission lines . . . . . . . . . . . . . . . . . . . . .
3.3.2 The integrated Hβ profile . . . . . . . . . . . . . . . . . . .
3.3.3 Spatially-resolved emission-line properties of the 8 o’clock
arc in the image plane . . . . . . . . . . . . . . . . . . . . .
3.4 The physical properties of the 8 o’clock arc . . . . . . . . . . . . .
3.4.1 SED fitting . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.4.2 Parameters derived from emission line modeling . . . . . .
3.4.3 AGN contribution . . . . . . . . . . . . . . . . . . . . . . .
3.4.4 Star formation rate and dust extinction . . . . . . . . . . .
3.4.5 Metallicity and dust-to-gas ratio . . . . . . . . . . . . . . .
3.4.6 The gas surface mass density . . . . . . . . . . . . . . . . .
3.4.7 Electron density . . . . . . . . . . . . . . . . . . . . . . . .
3.5 Source Reconstruction . . . . . . . . . . . . . . . . . . . . . . . . .
3.5.1 Gravitational lens modeling . . . . . . . . . . . . . . . . . .
3.5.2 Reconstructed-Hβ and [O II] emission lines maps in the
source plane . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.5.3 Hβ profile of the reconstructed source . . . . . . . . . . . .
3.6 Dynamics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.6.1 Hβ Kinematics . . . . . . . . . . . . . . . . . . . . . . . . .
3.6.2 Dynamical mass . . . . . . . . . . . . . . . . . . . . . . . .
3.6.3 A massive outflow of gas? . . . . . . . . . . . . . . . . . . .
3.7 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.8 AppendixA: Gaussian decomposition . . . . . . . . . . . . . . . . .
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4 Stars were born in significantly denser regions in the
verse
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . .
4.2 Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.3 Methodology . . . . . . . . . . . . . . . . . . . . . . . .
4.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.5 Metallicity dependence . . . . . . . . . . . . . . . . . . .
vi
42
47
57
59
61
62
64
64
64
65
67
67
67
68
69
70
70
73
76
77
79
80
82
83
83
87
88
90
90
91
93
94
97
99
early Uni103
. . . . . . 104
. . . . . . 105
. . . . . . 106
. . . . . . 107
. . . . . . 109
4.6 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
4.7 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
5 On the spatial distribution of star formation in distant and nearby
galaxies
121
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
5.2 High-z sample . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
5.2.1 Resolved stellar population modeling . . . . . . . . . . . . . 124
5.2.2 The Voronoi binning of the UDF data . . . . . . . . . . . . 124
5.2.3 Fitting models to the color map of the UDF data . . . . . . 124
5.2.4 Integrated properties . . . . . . . . . . . . . . . . . . . . . . 125
5.3 Low-z sample . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
5.3.1 Resolved stellar population modeling of low-z sample . . . . 135
5.3.2 Fitting models to the color map of the SDSS data . . . . . 135
5.3.3 Integrated properties . . . . . . . . . . . . . . . . . . . . . . 136
5.4 Structural parameters . . . . . . . . . . . . . . . . . . . . . . . . . 136
5.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
5.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 148
Nederlandse samenvatting
151
Publications
155
Chapter
1
Introduction
Galaxies with all their varieties, have been home to billions of stars during their
life. It is because of the presence of these shining stars that we are able to observe
them through the cosmic time. Although we observe galaxies mostly through the
light emitted by their stars, we cannot resolve these stars individually unless they
are very close by. Because of this, the cumulative light from billions of stars in
every galaxy is analyzed using stellar population models to extract information
about the evolution of galaxies. Stellar light does not reach us without passing
through the interstellar medium (ISM) which contains clouds of gas and dust
particles. Gas and dust can absorb and re-emit the light from stars, or scatter it
towards us and make interpreting what we observe in galaxies very complicated.
Despite all these difficulties, just by analyzing the total light from galaxies, we
can constrain the global physical properties of galaxies such as stellar mass, star
formation rate and age, based on the stellar population models. By combining
stellar population models and photoionization models we can further analyze the
emission line spectrum of star-forming galaxies coming from ionized gas around
young stars which provide us with a wealth of information about the small-scale
properties of galaxies e.g., the ISM. This thesis is an attempt in understanding
the relation between these small-scale properties and global properties of starforming galaxies over cosmic time using stellar population synthesis models and
photoionization models.
Introduction
1.1
Nebular physics
Star formation in galaxies can be traced not only by stars directly, but also through
studying their impacts on their surrounding gas. The radiation from stars ionizes
the gas around them and produces nebular emission lines as a result of ionized gas
recombination. We can measure the star formation rate, along with other galaxy
properties, from emission lines coming from both nearby and distant galaxies.
In this thesis I extensively use nebular emission lines for measuring the intrinsic
properties of galaxies in the nearby and the distant Universe. In the following,
I briefly review the physics of ionized gas and mechanisms that produce strong
optical emission lines. I also discuss how we can use emission line ratios to constrain
various galaxy properties. I continue by discussing that how we can use emission
lines to indirectly trace massive stars and also to probe small-scale properties of
the ISM.
1.1.1
Nebular emission lines in star-forming galaxies
The optical emission line spectrum contains the Balmer series of Hydrogen (e.g.,
Hδ, Hγ, Hβ, Hα), together with lines of Helium (e.g., He ii λ4686) and numerous
other emission lines such as Oxygen (e.g., [O i] λ6300, [O ii] λ3726, [O iii] λ5007),
Nitrogen (e.g., [N ii]λ6584), Sulfur (e.g., [S ii]λ6716). Figure 3.7 shows an example
of the emission line spectrum of a nearby actively star-forming galaxy. The source
of radiation for producing these emissions in star-forming galaxies is mostly hot,
luminous stars with spectral types O and B. OB stars are very massive and shortlived stars with effective temperatures, T e , > 20000 K which enable them to emit
ultraviolet radiation at λ <912 Å. These ultraviolet photons can ionize neutral Hydrogen atoms (H i) around OB stars and produce ionized Hydrogen (H ii) regions.
The photoionization (the removal of a bound electron from the atom by a photon,
e.g., H i+hν → H ii + e− ) and recombination processes (combining free electron by
the ionized atom, e.g., H ii + e− → H i + hν) are two basic processes that derive
the physics of nebulae. The Balmer lines are produced by cascades of electrons in
Hydrogen atoms from energy levels above n = 2, following recombination of the
highly ionized gas (see Table 1.1 for Hydrogen Balmer recombination lines). The
energy of the photons created through the recombination process depends on the
kinetic energy of the free electrons and the binding energy of the bound-level into
which the recombination occurs (e.g., n = 2 for Balmer lines).
The other lines of importance are the collisionally-excited lines (e.g.,
[O iii] λ5007 Å). These lines are known also as forbidden lines because they are
forbidden by quantum selection rules. Excitation potential from sub-levels or fine
structure splitting of the ground level in elements that are heavier than Hydrogen
and Helium which are called metals (e.g., see fine structure splitting of [O iii] in
Figure 2.2) to upper energy level is ∼ 1 eV. This is approximately equal to the
typical thermal energy of the electrons (KT e ∼ 1 eV at T e = 104 K). Collisions with
free electrons excite bound electrons in the lower level of atoms to higher level
and take the energy of free electrons. Therefore, the formation process of these
forbidden lines by removing kinetic energy from the gas and transforming it to
2
Nebular physics
line
Hα
Hβ
Hγ
Hδ
λ (Å)
6563
4861
4340
4102
transition
n=3→n=2
n=4→n=2
n=5→n=2
n=6→n=2
Table 1.1 Balmer transitions in Hydrgen
photon energy which escapes from the gas can cool the nebula. The Differences
in the wavelengths of the lines lead to differences in their velocities and therefore
differences in their collisionally-excitation rates.
The low energy states of singly ionized Oxygen ([O ii] ) and doubly ionized
Oxygen ([O iii] ) are illustrated in Figure 2.2. The forbidden transitions amongst
these levels include the lines at 3726, 3729, 7319, 7320, 7329 and 7330 Å for [O ii]
and 5007, 4959 and 4363 Å for [O iii] .
1.1.2
Measuring galaxy properties using emission line ratios
Two fundamental measures of the physical properties in the ISM of galaxies are
the temperature and the density. Emission line intensities of forbidden lines can be
used for measuring the temperature of H ii regions. For instance, the line ratios of
([O iii] λ4959 + [O iii] λ5007)/[O iii] λ4363 can be used to measure the electron temperature. In a hot nebula, the [O iii]λ4363/[O iii]λ5007 or [O iii]λ4363/[O iii]λ4959
ratios are higher than a cooler nebula. This is because to produce [O iii] λ4363
an electron needs to be excited to 5.3 eV, which requires more energy than for
4959/5007 which result from decay from 1 D2 level, 2.5 eV above the ground-state
(see Figure 2.2). Therefore, these ratios can be used to constrain the electron
temperature in a nebula.
For measuring the electron densities, the [O ii] λ3726, 3729 doublet or the
[S ii] λ6716, 6731 doublet can be used (see Figure 5.8 in Osterbrock & Ferland
2006 for the density dependence of these line ratios for a given nebula temperature). For example, [O ii] λ3726 and [O ii] λ3729 are two lines of the same ion which
are emitted from different levels with nearly the same excitation energy. However,
because they have different transition probabilities and different collisional excitation rates, the relative population of the two levels or the ratio of their intensities
depend on electron densities.
Depending on the production rate of hydrogen ionizing photons, Q, produced
by stars and density of H i cloud surrounding stars, a specific radius is ionized by
them which is known as Strömgren radius (3Q/4πnH 2 αB ) where is the volume
filling factor of the ionized gas, which is defined as the ratio between the volumeweighted and mass-weighted average hydrogen densities (Charlot & Longhetti,
2001) and αB is the case-B Hydrogen recombination coefficient (Osterbrock & Ferland, 2006). The Strömgren sphere (H ii region) grows with time until equilibrium
between ionization and recombination is reached. After assuming that most of
ionizing photons are absorbed locally, the volume averaged ionization parameter
3
Introduction
Figure 1.1 A spectrum of a nearby strongly star-forming galaxy with strong emission lines indicated. The galaxy spectra (PlateID-MJD-FiberID: 752-52251-340) is
taken from SDSS DR7 (Abazajian et al., 2009) which covers a wavelength range of
3800-9200 Å. Bruzual & Charlot (2003) stellar population model is used to fit the
continuum which is shown by black solid line. The spectra were analyzed using
the methodology discussed in Tremonti et al. (2004, see also Brinchmann et al.
2004) to provide accurate continuum subtraction and were additionally analyzed
using the platefit pipeline discussed in Brinchmann et al. (2008b) to measure a
wider gamut of emission lines. The blue line shows the measured nebular emission
lines.
in a typical ionized region is:
< U >3 ≈
α2B 3Q(t) nH 2
(
),
4π
c3
(1.1)
The ionization parameter which is a measure of intensity of ionizing sources
and also Hydrogen number density can be estimated using emission line ratios of
high ionization lines to low ionization lines (e.g., [O iii] λ5007/[O ii] λ3727 ratio).
Emission line ratios are also used for classifying galaxies in terms of their
sources of ionization/excitation (see Figure 2.3). For instance, [O iii] λ5007/Hβ
is strongly correlated with hardness of source of ionization/excitation and temperature of ionized gas. Therefore, galaxies with different main ionizing sources
are distributed differently in diagnostic diagrams (e.g. the BPT diagram, Bald4
Nebular physics
[OII]
2
2
[OIII]
1
1/2
P3/2
7330
7329
7320
7319
3/2
D5/2
3729
4
S3/2
S0
4363
1
2321
D2
3726
5007
3
P
4959
2
1
0
Figure 1.2 Energy level diagram of the 2p3 ground configuration of singly ionized Oxygen ([O ii] ) and doubly ionized Oxygen ([O iii] ) are illustrated based on
Osterbrock & Ferland (2006). Splitting of the ground 3 p is exaggerated for the
[O iii] and energy levels in the two diagrams are not comparable. All emissions
are in the optical except lines 2321 Å which is in the ultraviolet. All wavelengths
shown are in Å.
win, Phillips & Terlevich, 1981) that are based on the line ratios that correlated with the metallicity (e.g., [N ii] λ6584/Hα) and ionization properties (e.g.,
[O iii] λ5007/Hβ) of galaxies. In star-forming galaxies where the main source of
ionization/excitation is star formation, electrons lose their energy more efficiently
through optical transitions (strong [O iii] λ5007 or high [O iii] /Hβ) at low metallicity. When the metallicity increases, metal line cooling gets stronger and the
electron temperature drops. Thus, electrons evacuate their energies through low
ionization lines in infrared (e.g., [O iii] λ88 µm) which makes high ionization lines
weaker (weak [O iii] λ5007 or low [O iii] /Hβ). Other sources of ionization such as
shock and active galactic nuclei (AGN) can be the source of strong [O iii] λ5007
or high [O iii] /Hβ, at high metallicities. Using these line ratio diagrams, galaxies
are generally classified as star-forming, AGN or composite galaxies (Kewley et al.,
2001; Kauffmann et al., 2003; Kewley et al., 2013), where the source of ionizing radiation for composite galaxies could be a combination of star formation and AGN
or shock.
5
Introduction
Figure 1.3 This plot shows the BPT diagnostic diagram that is used for classifying
galaxies in terms of their main source of ionization/excitation. The distribution
of emission line galaxies in the SDSS is shown by the colored 2D distribution
where the color-scale shows the logarithm of the number of galaxies in each bin.
The classification line presented by Kauffmann et al. (2003) is shown as a dashed
line, galaxies below this line are star-forming (SF) galaxies. Kewley et al. (2001)
classification line is shown as a solid line, galaxies between this line and dashed
line are composite galaxies (comp) and galaxies above this line are classified as
AGN.
1.1.3
Emission lines as indirect tracers of massive stars
As was mentioned earlier, the main source of ionization for producing strong emission lines in star-forming galaxies is very massive stars. However, our knowledge
about these massive stars is limited because direct observations of them often
cannot be carried out as these stars are often heavily enshrouded and at low
metallicities, they are only found outside the Milky Way. Despite the small fraction of these stars among billions of stars in star-forming galaxies, and their very
short life times (e.g., a few million years), they have a significant impact on the
galaxy evolution through their hard radiation, strong winds and their explosions
as supernovae.
At very high energies (e.g. λ < 228 Å) normal OB stars, emit a negligible
number of photons. In standard models of massive stars, stars in the Wolf-Rayet
(WR) phases have sufficiently hard spectra at these wavelengths. This is, however,
6
Nebular physics
a poorly tested assumption in general, and particularly at low metallicities. The
ionizing spectrum of WR stars is still subject to significant uncertainty and softer
spectra are being predicted in more recent models (Schaerer, 1996; Smith et al.,
2002). WR stars are very short lived (e.g., a few Myr) and typically have masses
of 10 − 25 M and they are descended from O-type stars (Meynet & Maeder, 2005;
Crowther, 2007).
In order to produce the observed number ratio of WR to O stars, rotation
and binary evolution should be considered in the models of massive stars (see
e.g, Brinchmann et al., 2008a). These two effects help removing the outermost
atmospheres of stars, thus encouraging the formation of hot WR stars. However,
despite an extensive effort in modeling these massive stars, the impact of rotation
on the models is still uncertain (Meynet & Maeder, 2005; Heger et al., 2005). The
role of binary evolution or single star evolution, especially at low metallicities is
also still open for discussion (Han et al., 2007; Eldridge et al., 2009).
Although we cannot observe the high energy part of the continuum coming
from massive stars due to interstellar absorption we can observe emission lines
that are produced by hard ionization from these stars. We can use these high
ionization emission lines such as the nebular He ii λ4686 emission line to indirectly
trace massive stars. This is a recombination line with ionization potentials of 54.4
eV, coming from ionized gas around very massive stars. Using this, we can probe
the high energy part of the spectral energy distribution of very massive stars.
1.1.4
Emission lines as indirect tracers of the ISM
Recent studies have shown that star formation conditions at high redshift (high-z)
are different from what we observe in the local Universe. For instance, it has been
shown that high-z emission line galaxies are systematically offset from low redshift
(low-z) trends in emission line ratio diagrams. This is seen particularly well in the
BPT diagram, log [N ii]/Hα vs. [O iii]/Hβ diagram (Brinchmann et al., 2008b;
Liu et al., 2008). In this diagram, high-z star-forming galaxies unlike low-z ones
are distributed in the regions that need other sources of ionization/excitation than
star formation (e.g., Shapley et al., 2005; Erb et al., 2006; Newman et al., 2013).
Kewley et al. (2013) recently studied the cosmic evolution of the BPT diagram and
showed that the extreme ISM conditions at high-z cause this offset between distant
and nearby galaxies in the BPT diagram. Therefore, we can use the emission line
intensities of distant galaxies to indirectly trace the ISM at high-z and to study
the evolution of intrinsic physical properties of star-forming galaxies from distant
to nearby Universe. However, studying the evolution of the physical conditions at
which stars are forming has proven very challenging and is hidden in the strong
evolution of global mean properties of galaxies such as stellar mass and SFR. In
the next section, I briefly summarize these evolutions in the global properties of
star-forming galaxies from distant to nearby Universe.
7
Introduction
1.2
Evolution in the properties of star-forming
galaxies
The average integrated properties of star-forming galaxies have evolved significantly during the last ∼ 12 Gyr. When the Universe was only 2-3 Gyr old (redshift,
z ∼ 3 − 2), star formation in typical galaxies was happening at a rate that today is
only found in the most extreme star-forming galaxies (Brinchmann et al., 2004).
The star formation rate of the Universe within a comoving volume element as a
function of redshift was first presented by the Madau plot (Madau et al., 1996) and
confirmed by further observations (see Figure 1.4 that shows the evolution of the
SFR density of Universe with redshift, data are taken from Hopkins & Beacom,
2006).
These very high star formation rates have been measured based on rest-frame
UV emission from young stars in high-z galaxies (e.g., Noeske et al., 2007; Daddi et
al., 2007) or infrared observations that determine the contribution of obscured light
to the SFR of high-z galaxies (e.g., Elbaz et al., 2007, see also Shapley, 2011 for
other methods used for estimating SFR of distant galaxies). A strong correlation
between SFR and stellar mass is observed at high-z known as the star-forming
main sequence (Noeske et al., 2007; Daddi et al., 2007). This tight main sequence
locus evolves smoothly with redshift showing that galaxies with the same stellar
mass at low-z and high-z have higher SFR at high-z (see Bouché et al., 2010). The
high SFR of these galaxies implies they are much more gas rich than local starforming galaxies. High gas fractions (several times higher than what we observe in
the local galaxies) also were observed for some of these high-z galaxies (e.g., Daddi
et al., 2008; Tacconi et al., 2010; Genzel et al., 2010; Tacconi et al., 2013). A more
clumpy morphology has been observed for many of high-z star-forming galaxies
(e.g., Elmegreen & Elmegreen, 2006; Genzel et al., 2011; Wuyts et al., 2012).
These clumpy structures can be caused from gravitational instability within these
very gas rich disks at high-z.
1.2.1
Stellar mass evolution
Stellar population synthesis models can be used to estimate the global physical
properties of galaxies such as stellar mass. Based on stellar population synthesis
models (e.g., Bruzual & Charlot, 2003), a combination of optical broadband photometry and spectral indices (the 4000Å spectral break and the strength of Balmer
absorption lines) can be modeled to measure stellar masses in nearby Universe
(e.g., Kauffmann et al., 2003). At higher redshifts, however, due to lower signalto-noise, stellar absorption features are difficult to measure and only broadband
photometry tends to be used for measuring stellar masses. The rest-frame near-IR
luminosity is more closely tied to stellar mass than the rest-frame optical luminosity (Bell & de Jong, 2001). The rest-frame UV emission from distant galaxies
can only probe their massive stars light. Therefore, these data should be combined with longer wavelengths observations to probe older stellar populations of
high-z galaxies. With availability of these observations for many of high-z galaxies
in recent years, the global evolution of the stellar content in galaxies from the
8
Evolution in the properties of star-forming galaxies
Figure 1.4 The SFR density of the Universe as a function of redshift. The data
have been taken from Hopkins & Beacom (2006).
distant to the nearby Universe has become possible to estimate with good accuracy. The evolution in stellar mass can be described by constructing the galaxy
stellar mass function at a range of redshifts. Based on recent observations in the
COSMOS/ULTRAVISTA survey which confirms previous measurments, a strong
evolution in stellar mass has been observed from z = 4 to z = 0.2 (e.g., Ilbert et al.,
2013; Muzzin et al., 2013). These studies show the mass density of star-forming
galaxies grows by a factor of 1.59 since z = 3.5 and the typical mass of a galaxy of
Log(M∗ /M ) = 10.5 at z = 0.3 would be Log(M∗ /M ) . 9.5 at z = 2.
1.2.2
Mass-metallicity evolution
As galaxies evolve, they form more stars and their gas content is converted into
stars and their metal content increases. Thus, stars made out of material that has
been enriched for many generations will be more metal-rich. Some of the metals
that are produced will be ejected out of the galaxy through outflows into the
intergalactic medium (IGM). Galaxies also accrete some gas from the IGM. Gas
metallicities are derived from emission line properties and stellar metallicities are
derived from Lick indices (Faber, 1973). Based on these metallicity measurements,
we can study the evolution in the metallicity of galaxies from distant to nearby
Universe.
There is an evolution in the relation between stellar mass and metallicity known
as mass-metallicity relation as we look back in time. This evolution shows that
galaxies with the same stellar mass have lower metallicities at high-z compared to
9
Introduction
similar galaxies in the local Universe. However, Mannucci et al. (2010); Lara-López
et al. (2010) found that when including the SFR, mass-metallicity-SFR relation
holds up to z ∼ 3 which means galaxies with the same stellar mass have higher
SFR when they show lower metallicity. There should be also a more fundamental
relation between atomic gas mass, SFR and metallicity as it is observed for local
galaxies (Bothwell et al., 2013). This suggests that the reason for having high
SFR at low metallicity is because of having more gas. However, because of the
lack of enough atomic gas data available for high-z galaxies (e.g., H i gas cannot
be observed at z > 0.4 with current instrumentation, and the molecular, i.e. H2 ,
contents of high-z galaxies are estimated from CO observations), this has not been
studied yet.
1.3
Star and galaxy formation over cosmic times
After the Big Bang, the Universe had no stars but was filled with only gas and dark
matter. Dark matter perturbations in the early universe grew gravitationally and
ended up as galaxy dark matter haloes. The gas which is bound to dark matter
haloes radiates its energy away and cools down. Because of the conservation of
angular momentum, this collapsing gas forms a rotating disk within which smallscale instabilities could grow to form molecular clouds. These molecular clouds
have been the birth place for most stars. Gravitational instability with the critical
density that is set by turbulence from stellar feedback (both negative and positive
through heating of gas, and compressing it) is believed to determine star formation
processes. Although we know the formation of stars to the first order, we do not
know well the physical processes that control the interplay between gas and stars.
Therefore, understanding the star formation history over cosmic time remains a
major theoretical and observational challenge.
Understanding how galaxies were assembled across the cosmic time also remains
a challenging question. For instance, how today’s Hubble Sequence with different
galaxy morphologies is shaped and which physical processes can constrain and
control the galaxy evolution, are important questions to be addressed. In the
standard scenario, galaxies are believed to form as disk galaxies, which can then
be transformed into ellipticals mainly due to major mergers. If new gas from the
merger remnants is able to cool then new disks can form and this process can make
disk-bulge systems (e.g., Kauffmann et al., 1993; Baugh et al., 1996). However,
galaxies at the peak of star formation in the Universe show very distinct features,
such as clumpy morphologies. The nature of these clumps and their evolution
determines whether host galaxies have inside-out growth and form bulges from
migration of these clumps towards the center.
Another important question in this regard is what fuels star formation. Hydrodynamical cosmological simulations predict that at high-z, gas accretion plays
a significant role for fueling star formation (Kereš et al., 2005). The existence
of disk-like kinematics in star-forming galaxies during the peak of star formation
suggests that gas accretion is the dominant process for growth of galaxies. In the
local Universe, however, mergers are believed to play a more dominant role.
10
Towards probing the small-scale properties of distant galaxies
1.3.1
From clumps to bulges
Small-scale instabilities in a rotationally supported gaseous disk are unstable
against gravitational collapse and can grow if the Toomre stability parameter
(Toomre 1964), Qgas < 1 where the Q parameter for stability of a disk is:
Qgas =
κσ
πGΣgas
(1.2)
where σ is velocity dispersion and Σgas is the gas mass density. κ = a v/R is the
epicyclic frequency where a is a dimensionless factor 1 < a < 2 depending on the
rotational structure of the disk, v is the circular velocity and R is the radius.
A clump of gas that is large enough, i.e. larger than the Jeans length
L J ' σ2 /GΣgas , can collapse under its self-gravity despite its velocity dispersion.
Because of its rotation within the disk this clump experiences an outward centrifugal acceleration ' L J κ2 ; if this acceleration is larger than the gravitational
acceleration, GΣgas , then the disk is stable.
It is widely accepted that the majority of gas clumps form from gravitational
instability with Jeans scale of ≈1 kpc. Clumpy galaxies at high-z usually show disklike kinematics with high turbulent Hα and CO velocity dispersions (e.g., Genzel
et al., 2006; Epinat et al., 2012; Tacconi et al., 2013). From a theoretical point of
view, violent disk instabilities and high velocity dispersions are required to regulate
disks with a Toomre parameter Qgas ≈ 1 (Dekel et al., 2009). Observationally, the
Qgas ≈ 1 instability limit has been estimated for gas within high-z galaxies (see
Genzel et al., 2011). However, local spiral disks tend to have Qgas ∼ 2 (van der
Kruit & Freeman, 1986) and they cannot form gas clumps.
The formation and evolution of clumps has an important impact on the formation of the central bulges in galaxies. However, it is not known yet whether they
can survive stellar feedback long enough and migrate inward to build the central
bulge, or whether these clumps disrupt like the molecular clouds.
Recent high-resolution optical and near-infrared images and spatially-resolved
observations allow us to study the stellar populations of these kpc-size clumps
within high-z galaxies.
1.4
Towards probing the small-scale properties of
distant galaxies
The last decade has seen a dramatic increase in our knowledge about galaxy population at z ∼ 1–3 (e.g., Shapley , 2011). This has been achieved mostly by studying
the integrated properties of galaxies. However, recently using near-IR integral field
unit (IFU) observations with adaptive optics (AO), many high-z galaxies have
been resolved spatially. The steadily growing effort to obtain resolved near-IR
spectra of high-z galaxies in a systematic manner, such as the SINS, MASSIV and
LSD/AMAZE surveys (Förster Schreiber et al., 2006; Contini et al., 2012; Maiolino
et al., 2010), is leading to large samples of spatially-resolved emission line maps
of distant star-forming galaxies. These maps provide us with spatially-resolved
11
Introduction
galaxy properties such as metallicity and SFR at high-z. Based on these studies, we know a large fraction (30% or larger) of these galaxies are dominated by
rotating disk kinematics with an increased fraction towards higher stellar masses
(e..g., Förster Schreiber et al., 2009, see their Figure 17 for kinematics of some
SINS galaxies) and they show high turbulent Hα velocity dispersions (e.g., Genzel et al., 2006). Metallicity gradients were measured within these galaxies based
on spatially-resolved emission line maps and it was discovered that some of these
galaxies show a lower metallicity in their centers (Cresci et al., 2010) as opposed to
what we observe in the local Universe for rotating disks systems, see Glazebrook
(2013) for a review on kinematic studies of star-forming galaxies across cosmic
time.
However, obtaining sub-kpc resolution, even with AO observations, is very
difficult because of the intrinsic faintness and the small sizes of high-z galaxies.
Our limitation to reach sub-kpc scale resolution for high-z galaxies using current
instruments can be resolved by observing high-z galaxies which are significantly
magnified due to gravitational lensing. This allows us to study the properties of
those high-z galaxies at a level similar to what is achieved at lower redshifts (e.g.,
Yee et al., 1996; Pettini et al., 2000, 2002; Teplitz et al., 2000; Savaglio et al., 2002;
Siana et al., 2008). However, even the ∼100 pc resolution achieved for some lensed
galaxies at high-z (e.g., Swinbank et al., 2009; Jones et al., 2010) is not enough
to study the small-scale properties of the ISM in high-z galaxies at the same level
that we can study those in the local Universe (e.g., Kennicutt & Evans, 2012).
The Atacama Large Millimeter/submillimeter Array (ALMA) observations will
soon provide us with insightful information about sub-kpc scale kinematics and
distribution of gas and star formation within distant galaxies. Afterwards, integral
field spectroscopic capability with the James Webb Space Telescope (JWST) will
allow us to accurately map distant galaxies in emission and absorption. This will
significantly change our view about the resolved small-scale properties of high-z
galaxies in the coming decade.
1.5
This thesis
In this thesis I analyze emission lines from gas ionized by very massive stars in
nearby and distant star-forming galaxies. In the local Universe, based on these
emission lines we can determine the source of ionization for producing them at
different environments. At high-z, these emission lines provide us with information
about the small-scale properties of the ISM. A brief summary of the contents of
this thesis is given below.
Chapter 2: Strongly star-forming galaxies in the local universe with
nebular He ii λ4686 emission
The evolution of massive stars is a complex and not fully understood process.
While we are limited by interstellar absorption in observing the stellar continuum
at λ < 228 Å, using the nebular He ii λ4686 emission line gives us valuable information about this high energy part of the stellar spectral energy distributions.
Only the most extreme star-forming galaxies show nebular He ii emission and it
12
This thesis
is generally believed that Wolf-Rayet (WR) stars provide the required ionizing radiation for them. In Chapter 2, we study the physical properties of emission line
galaxies in the SDSS showing He ii emission. Based on these data, we find that
the He ii is not associated with WR features in a large number of star-forming
galaxies with this emission at low metallicities. This lack of WR stars has important implications for the evolution of the most massive stars at low metallicity.
Non-homogenous stellar evolution models (e.g., Yoon et al., 2006) and spatial offset between the location of WR stars and He ii regions (e.g., Kehrig et. al, 2008)
might be two possible explanations for this discrepancy. We also show the current
stellar population models cannot produce observed He ii /Hβ ratios in low metallicity environments. This result has implications for interpreting observations of
high-z galaxies where the metallicity is expected to be typically lower. Another
key result from this study is to define a new diagnostic diagram using the He ii /Hβ
ratio, which can be used to constrain AGN contribution in star-forming galaxies
showing He ii emission.
Chapter 3: The physical nature of the 8 o’clock arc based on near-IR
IFU spectroscopy with SINFONI
The detailed analysis of distant galaxies is limited by their small angular sizes
and faint apparent magnitudes. Both of these limitations can be overcome by
observing gravitationally lensed galaxies. In Chapter 3, we analyze spatiallyresolved data of the 8 o’clock arc, a lensed Lyman break galaxy, in conjunction
with HST imaging of this galaxy, from which the lens model for the galaxy was
reconstructed. Based on this lens modeling, the de-lensed Hβ map, velocity and
velocity dispersion maps are reconstructed. We show a simple rotating disk model
is unable to fit the observed velocity field of the galaxy, and a more complex
velocity field is needed. The Hβ profile of the galaxy shows a broad blueshifted
wing, suggesting an outflow of 200 km/s. The estimated gas surface density and
gas mass of the 8 o’clock arc shows a factor of 2.5-7 higher gas content compared
to similar galaxies in the SDSS.
Chapter 4: Stars were born in significantly denser regions in the early
Universe
Most of stars that surround us today were formed several billion years ago,
around the peak of star formation activity in the Universe. The conditions under which these stars were born is of great interest but very difficult to study
due to limited observational resolution for distant objects. In Chapter 4, we
present a novel approach to directly compare the density in the star-forming regions of galaxies that are near the peak of star formation activity in the Universe
to those of nearby galaxies. To indirectly trace the ISM at high-z, we use the emission line intensities of distant galaxies. We calibrate a new relation between the
[O iii] λ5007/[O ii] λ3727 emission line ratio and ionization parameter to estimate
the difference between the ionization parameters in the high and low-z samples.
We analyze the ionization properties of a sample of high-z galaxies at redshift
2.6–3.4, including the 8 o’clock arc, and compare them with that of galaxies with
similar physical properties in the local Universe. We show that after accounting
13
Introduction
for all differences in large-scale properties, such as mass and specific star formation
rate, the density in the star-forming regions was eight times higher in the past.
This implies that the majority of stars in the Universe were formed in gas that
obeyed very different scaling relations than what we see in the present day Universe. This is a striking result that provides strong constraints on the conditions
of star formation in normal galaxies in the early Universe.
Chapter 5: On the spatial distribution of star formation in distant and
nearby galaxies
In Chapter 5, we study the differences between the spatial distribution of
star formation in the distant and the nearby Universe for galaxies with the similar
global properties (e.g., stellar mass and specific star formation rate). We use the
multi-band imaging data available in the HUDF and compare this quantitatively
with the low-z data from the SDSS. Based on this, we study the physical processes
that cause clumpy star formation distribution for galaxies with similar star formation activity at low-z and high-z. We compare the resolved stellar populations
of these galaxies by measuring the structural parameters of distant galaxies and
their nearby counterparts. We show galaxies at high-z have more concentrated
stellar content but their star formation is more extended compared to galaxies
with the same global properties at z∼0. We show high-z galaxies are more clumpy
in their star formation distributions than their local analogs. This clumpy morphology suggests that distant galaxies need to have more surface density of the
disk compared to their local analogs.
1.5.1
Observations
In this thesis I use emission line intensities of a sample of distant galaxies from
the literature (Maiolino et al., 2008; Mannucci et al., 2009; Richard et al., 2011;
Dessauges-Zavadsky et al., 2011). Analyzing near-IR IFU spectroscopy of the 8
o’clock arc, a lensed Lyman break galaxy at redshift 2.735 is also added to my
high-z studies. These data were taken with SINFONI on VLT covering λ = 2900
Å to 6500 Å in the rest-frame. The SINFONI data are analyzed in conjunction
with the HST images of the galaxy.
The low-z studies in this thesis are based on the Sloan Digital Sky Survey
(SDSS) (York et al., 2000) data. The MPA-JHU1 value added catalogues (Brinchmann et al., 2004; Tremonti et al., 2004) for SDSS DR7 (Abazajian et al., 2009)
are used and star-forming galaxies following Brinchmann et al. (2004) are selected.
Furthermore, SDSS DR8 (Aihara et al., 2011) photometry are used to estimate
stellar masses in Chapter 4.
In this thesis I also use the multi-band imaging data available in the Hubble
Ultra Deep Field (HUDF) to study the spatial distribution of star formation in
some HUDF galaxies that have confirmed spectroscopic redshifts2 (Coe et al.,
2006) .
1 http://www.mpa-garching.mpg.de/SDSS/DR7
2 http://adcam.pha.jhu.edu/~coe/UDF/paper/zspec.cat
14
Summary
1.6
Summary
This thesis aims at studying the physical properties of galaxies from the distant
to the nearby Universe based on their emission line observations. We study these
properties using Charlot & Longhetti (2001, CL01) models which combine Bruzual
& Charlot (2003) stellar population models with CLOUDY photoionization models
(Ferland et al., 1998). We use high ionization emission lines to probe the high
energy part of the stellar SEDs at low metallicities and constrain current stellar
populations. We show that stellar population models need to consider harder
spectra for O type stars in order to explain observations of high ionization emission
lines such as He ii λ4686. These results can be further confirmed by studying the
spatially-resolved analysis of He ii galaxies.
Recent observations have shown that high-z star-forming galaxies form a different population compared to star-forming galaxies in the local Universe. These
studies, however, cannot tell if the main difference between low-z and high-z starforming galaxies is related to their strongly evolving global properties (e.g., mass,
SFR) or their different intrinsic properties (e.g., the ISM). Based on the CL01
models and emission line ratios, we probe the intrinsic properties of high-z galaxies
that show no evolution in their global properties compared to a sample of nearby
star-forming galaxies. Using this, we show that star-forming regions are denser at
high-z and also there is an evolution in the relation between the surface density
of gas and the surface density of SFR known as the star-formation law towards
less efficient relation at high-z. Emission line observations for a larger sample of
star-forming galaxies at high-z and also follow up observations of emission line
ratios that are density tracers can confirm this higher density at high-z.
We compare also the distribution of star formation between distant and nearby
galaxies with similar physical properties based on their deep imaging. We show
that stellar content of distant star-forming galaxies is more compact than their
local analogs. The same result has been shown before for elliptical galaxies in the
local and high-z Universe. To do a more statistical analysis our study should be
done for a larger sample of star-forming galaxies at z > 1.5.
15
REFERENCES
References
Abazajian, K. N., Adelman-McCarthy, J. K., Agüeros, M. A., et al. 2009, ApJS,
182, 543
Aihara, H., Allende Prieto, C., An, D., et al. 2011, ApJS, 193, 29
Baldwin, J. A., Phillips, M. M., & Terlevich, R. 1981, PASP, 93, 5
Baugh, C. M., Cole, S., & Frenk, C. S. 1996, MNRAS, 283, 1361
Bell, E. F., & de Jong, R. S. 2001, ApJ, 550, 212
Bothwell, M. S., Maiolino, R., Kennicutt, R., et al. 2013, MNRAS, 433, 1425
Bouché, N., Dekel, A., Genzel, R., et al. 2010, ApJ, 718, 1001
Brinchmann, J., Charlot, S., White, S. D. M., et al. 2004, MNRAS, 351, 1151
Brinchmann, J. ,Kunth, D., Durret, F., 2008, A&A, 485, 657
Brinchmann, J., Pettini, M., & Charlot, S. 2008b, MNRAS, 385, 769
Bruzual, G. ,Charlot, S., 2003, MNRAS, 344, 1000
Charlot, S., & Longhetti, M. 2001, MNRAS, 323, 887
Coe, D., Benı́tez, N., Sánchez, S. F., et al. 2006, AJ, 132, 926
Contini T., et al., 2012, A&A, 539, 91
Cowie, L. L., Hu, E. M., & Songaila, A. 1995, AJ, 110, 1576
Cresci, G., Mannucci, F., Maiolino, R., et al. 2010, Nature, 467, 811
Crowther, P. A., 2007, ARA&A, 45, 177
Daddi, E., Dickinson, M., Morrison, G., et al. 2007, ApJ, 670, 156
Daddi, E., Dannerbauer, H., Elbaz, D., et al. 2008, ApJL, 673, L21
Dekel, A., Birnboim, Y., Engel, G., et al. 2009, Nature, 457, 451
Dessauges-Zavadsky, M., Christensen, L., D’Odorico, S., et al. 2011, A&A, 533,
A15
Elbaz, D., Daddi, E., Le Borgne, D., et al. 2007, A&A, 468, 33
Elbaz, D., Dickinson, M., Hwang, H. S., et al. 2011, A&A, 533, A119
Eldridge, J. J., Stanway, E. R., 2009, MNRAS, 400, 1019
Elmegreen, B. G., & Elmegreen, D. M. 2006, ApJ, 650, 644
Epinat, B., Tasca, L., Amram, P., et al. 2012, A&A, 539, A92
Erb, D. K., Shapley, A. E., Pettini, M., Steidel, C. C., et al. 2006, ApJ, 644, 813
Faber, 1973, ApJ, 179, 731
Ferland, G. J., Korista, K. T., et al., 1998, PASP, 110, 761
Förster Schreiber N., et al., 2006, Msngr, 125, 11
Förster Schreiber, N. M., Genzel, R., Bouché, N., et al. 2009, ApJ, 706, 1364
Förster Schreiber, N. M., Shapley, A. E., Genzel, R., et al. 2011, ApJ, 739, 45
Genzel, R., Newman, S., Jones, T., et al., 2011, ApJ, 733, 101
Genzel, R., Tacconi, L. J., Gracia-Carpio, J., et al. 2010, MNRAS, 407, 2091
Genzel, R., Tacconi, L. J., Eisenhauer, F., et al. 2006, Nature, 442, 786
Jones, T. A., Swinbank, A. M., Ellis, R. S., et al. 2010, MNRAS, 404, 1247
Han, Z., Podsiadlowski, P., & Lynas-Gray, A. E. 2007, MNRAS, 380, 1098
Heger, A., Woosley, S. E., & Spruit, H. C. 2005, ApJ, 626, 350
Hopkins, A. M., & Beacom, J. F. 2006, ApJ, 651, 142
Ilbert, O., McCracken, H. J., Le Fevre, O., et al. 2013, arXiv:1301.3157
Kauffmann, G., Heckman, T.M., et al. 2003, MNRAS, 346, 1055
Kauffmann, G., White, S. D. M., & Guiderdoni, B. 1993, MNRAS, 264, 201
16
REFERENCES
Kehrig, C., Vı́lchez, J. M., et al., D., 2008, A&A, 477, 813
Kennicutt, Jr., R. C. 1998, ApJ, 498, 541
Kennicutt, R. C., & Evans, N. J. 2012, ARA&A, 50, 531
Kewley, L. J., Maier, C., Yabe, K., et al. 2013, arXiv:1307.0514
Kewley, L. J., & Dopita, M. A. 2002, ApJS, 142, 35
Kewley, L. J., Dopita, M. A., et al., 2001, ApJ, 556, 121
Kereš, D., Katz, N., Weinberg, D. H., & Davé, R. 2005, MNRAS, 363, 2
Glazebrook, K. 2013, arXiv:1305.2469
Lara-López, M. A., Cepa, J., Bongiovanni, A., et al. 2010, A&A, 521, L53
Liu, X., Shapley, A. E., Coil, A. L., et al. 2008, ApJ, 678, 758
Madau et al. 1996, MNRAS, 283, 1388
Maiolino R., et al., 2010, Msngr, 142, 36
Maiolino, R., Nagao, T., Grazian, A., et al. 2008, A&A, 488, 463
Mannucci, F., Cresci, G., Maiolino, R., Marconi, A., & Gnerucci, A. 2010,
MNRAS, 408, 2115
Mannucci, F., Cresci, G., Maiolino, R., et al. 2009, MNRAS, 398, 1915
Meynet, G., & Maeder, A. 2005, A&A, 429, 581
Muzzin, A., Marchesini, D., Stefanon, M., et al. 2013, arXiv:1303.4409
Noeske, K. G., Weiner, B. J., Faber, S. M., et al. 2007, ApJL, 660, L43
Newman, S. F., Buschkamp, P., Genzel, R., et al. 2013, arXiv:1306.6676
Osterbrock, D. E., & Ferland, G. J. 2006, Astrophysics of gaseous nebulae and
active galactic nuclei, 2nd. ed. by D.E. Osterbrock and G.J. Ferland. Sausalito,
CA: University Science Books, 2006
Penston, M. V., Robinson, A., Alloin, D., et al. 1990, A&A, 236, 53
Pettini, M., Rix, S. A., Steidel, C. C., et al. 2002, ApJ, 569, 742
Pettini, M., Steidel, C. C., Adelberger, K. L., et al. 2000, ApJ, 528, 96
Richard, J., Jones, T., Ellis, R., et al. 2011, MNRAS, 413, 643
Savaglio, S., Panagia, N., & Padovani, P. 2002, ApJ, 567, 702
Schaerer, D., Vacca, W. D., 1998, ApJ, 497, 618
Schaerer, D. 1996, ApJ, 467, 17
Siana, B., Teplitz, H. I., Chary, R.-R., et al. 2008, APJ, 689, 59
Shapley A., 2011, ARA&A, 49, 525
Shapley, A. E., Steidel, C. C., Erb, D. K., et al. 2005, ApJ, 626, 698
Shirazi, M., Brinchmann, J., & Rahmati, A. 2013, arXiv:1307.4758
Smith, L. J., Norris, R. P. F., Crowther, P. A., 2002, MNRAS, 337, 1309
Swinbank, A. M., Webb, T. M., Richard, J., et al. 2009, MNRAS, 400, 1121
Tacconi, L. J., Genzel, R., Neri, R., et al. 2010, Nature, 463, 781
Tacconi, L. J., Neri, R., Genzel, R., et al. 2013, ApJ, 768, 74
Teplitz, H. I., McLean, I. S., Becklin, E. E., et al. 2000, ApJ, 533, L65
Tremonti, C. A., Heckman, T. M., Kauffmann, G., et al. 2004, ApJ, 613, 898
van der Kruit, P. C., & Freeman, K. C. 1986
Wuyts, S., Förster Schreiber, N. M., Genzel, R., et al. 2012, ApJ, 753, 114
Yee, H. K. C., Ellingson, E., Bechtold, J., et al. 1996, AJ, 111, 1783
Yoon, S.-C., Langer, N., Norman, C., 2006, A&A, 460, 199
York, D. G., Adelman, J., Anderson, Jr., J. E., et al. 2000, AJ, 120, 1579
17
Chapter
2
Strongly star-forming galaxies
in the local universe with
nebular He IIλ4686 emission
We present a sample of 2865 emission line galaxies with strong nebular
He ii λ4686 emissions in Sloan Digital Sky Survey Data Release 7 and use this
sample to investigate the origin of this line in star-forming galaxies. We show that
star-forming galaxies and galaxies dominated by an active galactic nucleus form
clearly separated branches in the He ii λ4686/Hβ versus [N ii] λ6584/Hα diagnostic
diagram and derive an empirical classification scheme which separates the two
classes. We also present an analysis of the physical properties of 189 star forming
galaxies with strong He ii λ4686 emissions. These star-forming galaxies provide
constraints on the hard ionizing continuum of massive stars. To make a quantitative comparison with observation we use photoionization models and examine how
different stellar population models affect the predicted He ii λ4686 emission. We
confirm previous findings that the models can predict He ii λ4686 emission only
for instantaneous bursts of 20% solar metallicity or higher, and only for ages of
∼ 4 − 5 Myr, the period when the extreme-ultraviolet continuum is dominated
by emission from Wolf-Rayet stars. We find however that 83 of the star-forming
galaxies (40%) in our sample do not have Wolf-Rayet features in their spectra
despite showing strong nebular He ii λ4686 emission. We discuss possible reasons
for this and possible mechanisms for the He ii λ4686 emission in these galaxies.
Maryam Shirazi, Jarle Brinchmann
Monthly Notices of the Royal Astronomical Society
Volume 421, Issue 3, pp. 1043-1063 (2012)
Star-forming galaxies with nebular He II 4686 emission
2.1
Introduction
The ionizing continuum of stars at λ < 912 Å is of major importance for interpreting emission line observations of galaxies because many of the strong lines observed
in the spectra of galaxies, such as [O iii] λ5007, [Ne iii] λ3869 and He ii λ4686, have
ionization potentials in excess of 13.6 eV. Despite this importance we are severely
limited by interstellar absorption in observing stellar spectra in this spectral window directly (e.g. Hoare et al., 1993). Although we can get direct information at
slightly longer wavelengths with space-based UV spectroscopy (e.g. Crowther et
al., 2002), most of our knowledge about the λ < 912Å region is based on indirect
evidence, even for solar metallicity.
A promising way to indirectly obtain information on the stellar ionizing continuum is to compare emission line properties (e.g. flux, equivalent width) to
predictions from photoionization codes such as CLOUDY (Ferland et al., 1998)
or MAPPINGS III (Allen et al., 2008). In practice these kinds of studies provide
modest constraints on stellar atmosphere models (e.g. Crowther et al., 1999). However where predictions of models differ significantly, this approach can yield useful
information. This is the approach we will adopt in this work, where we will make
use of the He ii λ4686 nebular emission line to place constraints on stellar models
and in particular on the ionization mechanism for this line.
The presence of a nebular He iiλ4686 line in the integrated spectrum of a galaxy
indicates the existence of sources of hard ionizing radiation as the ionization energy
for He+ is 54.4 eV (λ ≈ 228 Å). This hard radiation can of course be produced by
an active galactic nucleus (AGN), and most sources with luminous He ii emission,
in a flux limited sample, are indeed galaxies with an AGN1 . However the required
hard radiation can also be provided by stellar sources and He ii λ4686 emission
is frequently seen in H ii-galaxies. The line appears to be associated with young
stellar populations; for instance, Bergeron et al. (1997) proposed Of stars as the
sources of He iiλ4686 emission in dwarf galaxies. Subsequent discussion has mostly
focused on Wolf-Rayet (WR) stars, although the distinction between these two
classes is rather blurred (e.g. Gräfener et al., 2011). Schaerer (1996, see also
Schaerer & Vacca (1998)) showed that the hard radiation field of WR stars could
provide a good explanation of the nebular He ii λ4686 seen in H ii-galaxies. Guseva
et al. (2000) tried to test this in a careful study of H ii-galaxies with prominent
WR features. They were however unable to find WR features in 12 out of the
30 galaxies with nebular He ii λ4686 emission. The same lack of WR features in
metal poor Blue Compact Dwarf (BCD) galaxies was pointed out by Thuan &
Izotov (2005). They proposed that fast radiative shocks could be responsible for
this emission (see also Garnett et al. 1991).
Similar results were reported by Brinchmann et al. (2008, hereafter B08), who
analyzed a sample of strong emission line galaxies in the Sloan Digital Sky Survey
(SDSS, York et al., 2000) with He ii λ4686 emission. They showed that at least at
metallicities of 12 + log O/H > 8, there appeared to be a close correlation between
WR features in galaxies and the presence of He iiλ4686 emission, but this appeared
1 We will here not distinguish between the host galaxy and its nuclear power source so will
refer to these galaxies as AGNs
20
Data
not to be so clear-cut at lower metallicities.
This apparent lack of connection of He ii emission with the hard UV radiation from WR stars has also been seen in spatially resolved spectroscopic studies.
Kehrig et. al (2008) performed an integral field spectroscopy study for the H ii
galaxy II Zw 70 and found that the region associated with nebular He ii λ4686
emission was a few arcsec offset from the region with detected WR features. More
recently Kehrig et. al (2011) and Neugent & Massey (2011) have presented studies of He ii λ4686 emission in M33. Both studies find some regions with nebular
He ii λ4686 emission that are not associated with WR stars (see also Hadfield &
Crowther 2007, López-Sánchez & Esteban 2010 and Monreal-Ibero et al. 2010).
Thus a series of studies have shown that while He iiλ4686 emission frequently is
found in association with WR stars, it appears not to be so in all cases, particularly
at low metallicity. As mentioned above, possible additional sources of high energy
photons could be X-ray binaries (Garnett et al. 1991), strong shocks (Dopita
& Sutherland, 1996), low-level AGN activity and alternative models for stellar
evolution (Yoon & Langer, 2005). However the existing studies do not show clear
trends that allow us to distinguish between these scenarios.
Crucially the samples in most of the previous studies have not been selected
specifically to study He ii emission lines. To make progress in understanding this
puzzle it is important to have as large as possible sample of He ii emitting galaxies
to allow one to study the relationship between He ii emission and other physical
properties. To this end we present here an analysis of emission line galaxies with
strong He ii λ4686 emission in SDSS Data Release 7 (DR7, Abazajian et al., 2009).
In section 3.2 we discuss the sample selection and carefully account for AGN
emission. The physical properties of the He ii emitting galaxies are discussed in
section 2.3 and the observed He ii λ4686/Hβ ratios are compared to model predictions in section 2.4. In section 2.5 we test these model predictions and investigate
whether the presence of He ii λ4686 is associated with WR features. We find that
low metallicity systems frequently do not show signs of WR stars. We discuss
possible explanations for this finding in section 2.6 and conclude in section 5.6.
2.2
Data
Our sample is based on galaxy spectra from SDSS DR7 which cover a wavelength
range of 3800-9200 Å. The spectra were analyzed using the methodology discussed
in Tremonti et al. 2004 (see also Brinchmann et al. 2004) to provide accurate
continuum subtraction. All emission line sources were additionally analyzed using
the pipeline discussed in B08 to measure a wider gamut of emission lines. For each
galaxy we measure 40 lines, These lines and the number of spectra that show these
lines with S /N > 5.5 are summarized in Table 2.1.
Our concern in this paper is not to analyze a volume- or magnitude-limited
sample of galaxies, we therefore do not impose a redshift cut nor a magnitude
limit. Since the blue wavelength cut-off of the SDSS spectrograph is ∼ 3800Å, the
[O ii] λ3727, 3729 doublet falls outside the spectral range for z < 0.02. This is a
concern, because as we will see later 55% of our final sample fall in this region
21
Star-forming galaxies with nebular He II 4686 emission
Measured line
[O ii] λ3726, 3729
[Ne iii] λ3869
H8
[Ne iii] λ3967
H
He i λ4026
[S ii] λ4069
Hδ
Hγ
[O iii] λ4363
He i λ4472
[Fe iii] λ4658
He ii λ4685
[Ar iv] λ4711
[Ar iv] λ4740
Hβ
[O iii] λ4959
[O iii] λ5007
[N i] λ5197
[N i] λ5200
[Cl iii] λ5518
[Cl iii] λ5538
[N ii] λ5755
He i λ5876
[O i] λ6300
[S iii] λ6312
[O i] λ6363
[N ii] λ6548
Hα
[N ii] λ6584
He i λ6678
[S ii] λ6717
[S ii] λ6731
He i λ7065
[Ar iii] λ7135
[O ii] λ7318, 19, 29, 30
Number of spectra
243977
39143
92230
18367
64120
6368
1503
163534
295528
7886
15563
2626
4034
2893
1074
458324
147734
339212
1318
758
160
190
4888
75256
119328
5552
16100
361641
613338
562659
19084
410903
341386
4232
24034
659
Fraction
16.51%
2.65%
6.24%
1.24%
4.34%
0.43%
0.10%
11.07%
20.00%
0.53%
1.05%
0.18%
0.27%
0.20%
0.07%
31.02%
10.00%
22.96%
0.09%
0.05%
0.01%
0.01%
0.33%
5.09%
8.08%
0.38%
1.09%
24.48%
41.51%
38.08%
1.29%
27.81%
23.11%
0.29%
1.63%
0.04%
Table 2.1 The table shows number of spectra that have the indicated line detected
at S /N > 5.5. The total number of analyzed spectra is 1,477,411.
of redshift space and their oxygen abundances are therefore somewhat uncertain.
When possible we use the [O ii] λ7318 − 7330 quadruplet instead (Kniazev et al.,
2004), but as this is a fairly weak line and falls in a region with significant sky
emission, we cannot always make use of this line.
2.2.1
Sample selection and classification
We select our sample requiring signal to noise ratio > 5.5 in He ii λ4686, the resulting data set is given in Table 2.2 2 . When the width of the He ii line is consistent
with that of the strong forbidden lines, we make the assumption that it has a
nebular origin. Given that He ii lines from individual WR stars typically are considerably broader than the forbidden lines in galaxies (e.g. B08), we feel this is a
reasonable assumption. In addition we require a S/N> 3 in each of Hβ, [O iii]λ5007,
2 The
full table of 3292 spectra is available in electronic form in http://www.strw.leidenuniv.
nl/~shirazi/SB011/.
22
Data
Hα and [N ii] λ6584 emission lines to reliably classify our galaxies (Brinchmann et
al. 2004, hereafter B04). The resulting sample contains 2865 spectra with strong
nebular He ii λ4686 emission.
In parallel, the spectra of the sample galaxies are examined for WR signatures
using the approach discussed by B08. This resulted in a total sample of 385 spectra
with likely and secure WR features (Class 1, 2 and 3 from B08). While we do not
discuss the sample of all WR galaxies in DR7 with Class 1–3 in detail here, we
note that it intersects that of the He ii sample but is not a strict subset of it (see
Table 2.3).
Figure 3.7 shows the redshift distribution for the fraction of the He ii sample
in the SDSS as a shaded grey histogram. The cut-off at z ∼ 0.4 is due to Hα
falling outside the spectrograph range. The red histogram shows the redshift
distribution for the fraction of just the star-forming (SF) galaxies in the SDSS
showing He ii λ4686 emission (see below for a discussion of the classification). For
each class we have divided the number of He ii emitting galaxies in that redshift
bin by the number of similarly classified galaxies in the parent sample (SDSS
galaxies that have S/N> 3 in each of Hβ, [O iii] λ5007, Hα and [N ii] λ6584) in
that redshift bin. A constant value would therefore indicate a similar redshift
distribution of the He ii sample and the parent sample. It is clear from this that
full He ii sample closely follows the overall distribution of the SDSS, but that the
star-forming galaxies with He ii λ4686 emission are predominantly found at low
redshift. We can also see that less than 2% of all galaxies, and less than 1% of the
star-forming galaxies in the SDSS DR7 show He ii λ4686 emission in their spectra.
To classify the dominant ionization source in each galaxy we follow previous
studies in using the Baldwin, Phillips & Terlevich (1981, BPT) line ratio diagnostic diagram of [O iii] λ5007/Hβ versus [N ii] λ6584/Hα as our starting point
(Figure 2.2). As has been discussed extensively (e.g. Terlevich et al., 1991; Kewley
et al., 2001, hereafter Ke01; Kauffmann et al., 2003, hereafter Ka03; Kewley et al.,
2006; Stasińska et al., 2006) this diagram allows a separation of AGN and starforming galaxies because of their significantly different ionizing spectra, typically
leading to high [O iii] λ5007/Hβ and [N ii] λ6584/Hα when an AGN is dominating
the output of ionizing photons. Ke01 used a combination of stellar population
synthesis and photoionization models to compute a theoretical maximum starburst line that isolates objects whose emission line ratios can be accounted for by
photoionization by massive stars (below and to the left of the curve) from those
where some other source of ionization is required. Ka03 defined an empirical upper limit to the H ii region sequence of SDSS galaxies in the BPT diagram. The
region lying between these two lines represents objects more naturally explained
as having a composite spectrum combining H ii region emission with a harder ionizing source. As we will see later, this interpretation is further corroborated by
the He ii λ4686/Hβ ratios we find for our sample.
In the present study, we adopt a two-stage classification methodology. We
start out by classifying all galaxies using the BPT diagram, and we will then
refine our classifications for a subset of the galaxies using detailed inspection of
the spectra and the He ii λ4686 line properties. For the initial classification we
23
Star-forming galaxies with nebular He II 4686 emission
Figure 2.1 The redshift distribution of the full He ii sample, relative to that of
the parent sample is shown as a shaded grey histogram. The red histogram shows
the redshift distribution of the He ii SF sample relative to that of all star-forming
galaxies in the parent sample; the parent sample consisting of those galaxies in
the SDSS DR7 having Hβ, [O iii] λ5007, Hα and [N ii] λ6584 detected at S/N> 3.
In comparison with the SDSS, it is clear that the galaxies in our SF sample have
a redshift distribution strongly shifted toward low redshift.
adopt a similar methodology to that of B04 and use the separation criteria defined
by Ke01 and Ka03 to divide galaxies into different classes. Galaxies which are
distributed above the Ke01 dividing line are considered AGNs, galaxies between
the Ke01 and Ka03 limits have composite classification, which means their source
of ionizing radiation could be a combination of star formation and AGN activity.
Finally, galaxies below the Ka03 line have a SF classification, which as we will see
might be modified subsequently.
Figure 2.2 shows the BPT diagram for our sample, the Ke01 and Ka03 classification lines are shown as solid and dotted lines, respectively. The distribution of
all emission line galaxies in the SDSS with S /N > 3 in Hβ, [O iii] λ5007, Hα and
[N ii] λ6584 is shown as a grey-scale 2D distribution, where the grey-scale shows
the logarithm of the number of galaxies in each bin. Blue circles show star-forming
galaxies, triangles show composite galaxies and red squares mark AGNs. The grey
dashed-dotted line shows the N2 (Pettini & Pagel 2004, hereafter PP04) metallicity
calibration for 12 + log O/H = 8.2 ([N ii] λ6584/Hα = −1.2). These classifications
are the final ones and incorporate further information as described in the following.
While the BPT diagram is a useful classification diagram, it is not particularly
sensitive to low levels of AGN contamination and some progress can be made by
including lines originating in the mostly neutral ISM (e.g. Kewley et al., 2006).
For our purposes we however need to be very confident in the lack of AGNs in our
24
Data
Figure 2.2 This plot shows the BPT diagnostic diagram for the sample. The Ka03
classification line is shown as a dotted line and Ke01 classification line as a solid
line. The grey dashed-dotted line shows the N2 PP04 metallicity calibration for
12 + log O/H = 8.2. The distribution of emission line galaxies in the SDSS is shown
by the gray-scale 2D distribution where the grey-scale shows the logarithm of the
number of galaxies in each bin. Blue circles show star-forming galaxies, triangles
show composite galaxies and red squares mark AGNs. As discussed in the text,
for some galaxies the classification has been adjusted which is why some objects
in the star-forming region are classified as AGN. We mark these galaxies with a
blue plus over the red square.
sample and we will use the He ii λ4686/Hβ ratio for this purpose.
As remarked earlier, only photons with energy in excess of 54.4 eV can ionize
He+ and thence produce the He ii λ4686 recombination line. If we consider single
stellar population models from Starburst99 (Leitherer et al., 1999) at an age of 2
Myr (so that all stars are on the main sequence), we find that the integrated spectrum of this population typically contain 2-3 orders of magnitude fewer photons at
this energy than at the energy required to ionize O+ (35.5 eV), which is needed to
produce the [O iii] λ5007 line from collisionally excited O+2 . This line is therefore
a very sensitive probe of AGN activity, particularly if used in conjunction with
other line ratios. We therefore make use of the He ii λ4686/Hβ vs [N ii] λ6584/Hα
diagram to further refine the classification of our sources. We show this diagram
in Figure 2.3 where the symbols and colors are the same as in Figure 2.2.
Note that there is generally a very significant offset between the star-forming
galaxies and AGNs in this diagram, in contrast to the gradual transition in much
of the BPT diagram. To make a quantitative separation, we draw a random set of
star-forming galaxies from the SDSS and gradually add their emission line fluxes
to that of a random set of AGNs. We quantify this by the variable f which is
defined to be the fraction of the total Hβ flux comes from the AGN. The total flux
25
Star-forming galaxies with nebular He II 4686 emission
Figure 2.3 This plot shows our sample in the He ii λ4686/Hβ versus [N ii] λ6584/Hα
diagnostic diagram. The dotted line shows an empirical line separating AGN
and composite objects from star-forming galaxies. The grey dashed-dotted line
shows the N2 PP04 metallicity calibration for 12 + log O/H = 8.2. Symbols are the
same as the BPT diagram. As He ii λ4686 have a higher ionization potential in
comparison to [O iii] λ5007 and is much less sensitive to the electron temperature,
we can clearly see separation between the classes in this diagram. The lines with
color gradients in the figure are simulated fluxes drawn from adding a random set
of star-forming galaxy from the SDSS with emission line fluxes of a random set
of AGN spectrum. The coloring of the lines corresponds to the fraction of the
spectrum contributed by the star-forming galaxy, 1 − f , as indicated by the color
bar on the side. The solid line shows the theoretical upper limit for He ii λ4686/Hβ
ratio.
is therefore f × AGNFlux + (1 − f ) × S FFlux . Changing f will trace out a path in the
diagnostic diagram in Figure 2.3 as illustrated by the lines with color gradients
in the figure. The coloring of the lines corresponds to 1 − f , as indicated by the
color bar on the side. We repeat this process for a thousand AGN and SF objects
located in different bins in the He ii /Hβ diagram.
Based on this analysis we find that there is a well defined locus where 10%
percent of the He ii λ4686 flux comes from an AGN (in this case ≈ 1% of the Hβ
flux comes from the AGN, see section 2.2.2 for further details). A good fit to this
relation is given by:
log(
FHe ii λ4686
1
) = −1.22 +
,
F[N ii] λ6584
FHβ
8.92 log( FHα ) + 1.32
(2.1)
which is shown as a dotted line in Figure 2.3. The solid line shows a theoretical
maximum starburst line, similar to the Ke01 line in the BPT diagram — we discuss
this further below.
26
Data
Sample
Total
AGN
Star forming
Composite
He ii
2865
2474
199
179
He ii + WR
385
234
116
35
Table 2.3 An overall numerical summary of the He ii sample. See section 3.2 for
a summary of the selection and classification details. See section 2.5 for details on
the WR classification.
Finally, we look at the spectra of galaxies that we would classify as star-forming
on the basis of their location in the BPT diagram, but that are offset from the
rest of the star-forming galaxies in the He ii /Hβ diagram and check whether they
show AGN features such as broad Balmer lines, strong Nevλ3426, Fe ii emission
or if they show a similar [O iii] λ4363/Hγ ratio to that of AGNs. If some of these
features are present, we change the classification from star-forming to AGN (these
objects are plotted as red squares containing a blue cross in Figures 2.2 and 2.3).
This happens for 127 (39%) galaxies. This is a conservative approach as we would
exclude star-forming galaxies with strong outflows and hence a broad base to the
Balmer lines for instance.
To estimate the maximum starburst line in Figure 2.3, we adopted the Charlot
& Longhetti (2001, hereafter CL01) models. These combine evolving stellar populations models from Bruzual & Charlot (unpublished BC00 models using Padova
1994 tracks, see Bruzual & Charlot 2003 (BC03) for the current models) with the
photoionization code Cloudy (Ferland et al., 1998) and adopt the simple dust attenuation prescription of Charlot & Fall (2000). The main model parameters for
our calculations are the metallicity Z, the ionization parameter, U, the dust attenuation τV and the dust-to-metal ratio, ξ, the model parameters used are given in
Table 2.4, see B04 for a more detailed discussion. We use the CL01 Single Stellar
Population (SSP) models since these achieve the highest possible He ii λ4686/Hβ
values. We then identify the maximum ratio reached by the different models and
use this upper envelope to define the maximum starburst line shown as a solid line
in Figure 2.3.
This combination of classification methods means that there is not a one-to-one
mapping between the location of an object in a diagnostic diagram and its final
classification, as is clear from Figure 2.2 and 2.3. We also mention that we do
not change classes for galaxies classified as AGNs or composites using the BPT
diagram, but which fall within the star-forming region in the He ii /Hβ diagram and
this is the reason why there are 7 AGN below our star-formation–AGN dividing
line.
By contrasting Figure 2.2 and 2.3 we can make a couple of interesting observations. The first is that while in the BPT diagram we see a steady increase in
[O iii] λ5007/Hβ with decreasing [N ii] λ6584/Hα, in Figure 2.3 we see no major
change in He ii λ4686/Hβ with [N ii] λ6584/Hα. This indicates that He ii λ4686/Hβ
and consequently the ionizing spectrum of stars at λ < 228 Å vary only weakly
with metallicity. The second difference between the two plots is that the He ii /Hβ
27
Star-forming galaxies with nebular He II 4686 emission
diagram can separate star-forming galaxies from composite galaxies better than
BPT diagram since the He ii λ4686/Hβ ratio is more sensitive to the hardness of
the ionizing source than [O iii] λ5007/Hβ. We can use this to further support our
supposition that the gas in the galaxies falling between the Ka03 and Ke01 lines
in the BPT diagrams is ionized by a combination of stars and an AGN — while
they are adjacent to the star-forming sequence in the BPT diagram they nearly
all clearly separate from the star-forming sequence in the He ii /Hβ diagram, corresponding to an AGN contribution to He ii λ4686 > 50%. Thus referring to these
as composite objects appear to be justified.
2.2.2
AGN contamination estimation
In view of the clear separation of the AGN and star-formation branches in Figure 2.3, and our sensitivity to low-level AGN contamination, it is beneficial to
study the impact of an AGN on the line ratios in more detail. This has been discussed in previous studies (e.g. B04, Stasińska et al. (2006)) but here we extend
those efforts to include He ii λ4686.
We focus our attention on the BPT diagram as it is most widely used for
emission line classification. We follow the same approach as in the previous section
of adding gradually more of an AGN emission line spectrum with He ii λ4686 in
emission at S/N> 3, to a star-forming one, and find where it intersects the Ka03
line (see bottom panel of Figure 4). At this point the galaxy would cease to be
classified as a star-forming galaxy. We repeat this for a total of 10,000 random
combinations of spectra. The median AGN contribution to Hβ at the point where
a galaxy ceases to be classified as star-forming is ≈ 10%.
Figure 4 shows the region where a median galaxy would be classified as starforming as the gray shaded region. On top of this we show the median trend for
the fraction of flux in the indicated line as a function of the fraction of the Hβ flux
coming from an AGN (shown for reference as the dashed diagonal line). We note
that the exact shape and location of the [N ii] λ6584 and [O iii] λ5007 lines does
depend somewhat on the AGN sample chosen but the qualitative trend remains
the same.
What this figure shows, is firstly the well-known result that some lines are
more sensitive to the presence of an AGN than others. As mentioned before,
the Balmer lines are expected to have less than 10% contribution from an AGN,
while the [N ii] λ6584 line can have more than 30% of its flux coming from an
AGN, putting in question its use as an abundance indicator on its own (see also
Stasińska et al. (2006)). But for our purposes, it is more important to note that
by adopting a classification based on the BPT diagram, we would classify a galaxy
as star-forming even when ∼ 65% of its He ii λ4686 flux would come from an AGN.
We can now combine this with our previous result in Figure 2.3, where we
found that only 10% of the He ii λ4686 emission comes from an AGN in our refined
star-forming sample. Applying this to Figure 2.4, we conclude that less than 1% of
the Hβ flux comes from an AGN and the other lines will also only have very small
contributions from an AGN implying we have a quite pure star-forming sample.
Our final sample of He ii star-forming galaxies consists of 189 star-forming
28
Data
Figure 2.4 The upper panel shows the sensitivity of different lines to the presence
of an AGN as a function of the fraction of the Hβ flux originating from an AGN.
The grey shading indicates the region where a typical galaxy would be classified
as star-forming in the BPT diagram (see text for details). We see that we might
classify galaxies as star-forming even if 10% of the Hβ flux comes from AGN. The
bottom panel illustrates the method used and shows the path traced by a galaxy in
the BPT diagram as an increasing amount of AGN light is added to a star-forming
galaxy spectrum. The coloring of the line corresponds to the fraction of the Hβ
line flux contributed by the star-forming galaxy, 1 − f , as indicated by the color
bar on the side. The dashed line is the Ka03 classification line, the intersection of
this line with the trajectory of the simulated flux is at f ≈ 0.1.
29
Star-forming galaxies with nebular He II 4686 emission
Parameter
Z, The metallicity
U, The ionization parameter
τV , The total dust attenuation
ξ, The dust-to-metal ratio
Range
−1 < logZ/Z < 0.6, 24 steps
−4.0 < logU < −2.0, 33 steps
0.01 < τV < 4.0, 24 steps
0.1 < ξ < 0.5, 9 steps
Table 2.4 The model grid used for the present work. We calculate this both for a
constant star formation history at t = 108 yrs as well as for an SSP.
galaxies (199 spectra which have been summarized in Table 2.5). As mentioned
above we have also checked these spectra for the presence of WR signatures. We
will return to a detailed discussion of this in section 2.5 but will use the result of
this classification in the following plots.
2.3
Physical properties of the sample
While the majority of the galaxies in our sample have some physical parameters in
the MPA-JHU value added catalogues3 , our galaxies are sufficiently extreme that
we need to rederive some properties and add some physical parameters to what is
in the MPA-JHU catalogue.
For the calculation of physical parameters we will adopt the Bayesian methodology outlined by Ka03 and B04. For each model we calculate the probability
of that model given the data assuming Gaussian noise and obtain the Probability Distribution Function (PDF) of every parameter of interest by marginalisation
over all other parameters (see Appendix A). We take the median value of each
PDF to be the best estimate of a given parameter.
2.3.1
Mass measurements
The standard SDSS pipeline often segment nearby actively star-forming galaxies
incorrectly, thus we need to redo the photometry of our galaxies. We do this
using the Graphical Astronomy and Image Analysis Tools (GAIA4 ). Most of our
galaxies have strong emission lines within some of the broad-band filters. Prior
to fitting we therefore correct the magnitudes for emission line contributions by
assuming that the relative contribution of the lines found in the SDSS fiber spectra
is applicable to the galaxy as a whole. As most galaxies in our sample appear to
have a uniformly blue color, presumably due to active star formation, we expect
this to be a reasonable assumption.
Stellar masses are calculated as outlined above, by fitting a large grid of
stochastic models to the SDSS u, g, r, i, z band photometry. The grid contains
pre-calculated spectra for a set of 100,000 different star formation histories using
the BC03 population synthesis models, following the precepts of (Gallazzi et al.,
2005, 2008).
3 http://www.mpa-garching.mpg.de/SDSS/DR7
4 http://astro.dur.ac.uk/%7Epdraper/gaia/gaia.html
30
Physical properties of the sample
Figure 2.5 This plot shows He ii λ4686/Hβ as a function of oxygen abundance (see
text for details on the calculations). Oxygen abundances derived using the direct
method are shown by black circles, we can not calculate oxygen abundances for
13 objects with this method as no [O ii] lines are available for them. Oxygen
abundances derived from fits to the CL01 models are shown by red circles while
the blue squares show the O3N2 oxygen abundance estimates. Note that the
qualitative trends are similar, but the O3N2 estimator does not reach as high O/H
values as the other two models. At low metallicity the three methods are in good
agreement.
2.3.2
Emission line derived parameters
We use the CL01 model to analyze the emission lines in our sample. We adopt
a constant star formation history (SFH) and use the same grid used by B04 (see
Appendix A and B04 for further details). In total the model grid used for the fits
have 2 × 105 different models.
The main quantity of interest for the present discussion is the oxygen abundance, quantified as 12 + log O/H. As there are significant differences between
methods for estimating oxygen abundance (Kewley & Ellison, 2008), we have
complemented the estimate from the CL01 method with two independent
methods: Firstly, we estimate gas-phase oxygen abundances with the empirically calibrated estimators proposed by PP04. They used the line ratios of
[O iii] λ5007/Hβ/[N ii] λ6584/Hα, the O3N2 method, and [N ii] λ6584/Hα, the N2
method, as abundance indicators. For all objects with detected [O iii] λ4363 with
S /N > 3, we also use the T e method, or direct method, using the fitting formulae
provided by Izotov et al. (2006) to estimate the oxygen abundances. For those
objects without [O iii] λ4363, we adopt the electron temperature estimates from
the CL01 models fits for the direct method calculation. Whenever we do not have
[O ii] λ3727, 3729, we use the [O ii] λ7320, 7330 lines to calculate abundances. However, for 13 objects we are unable to use the direct method for estimating oxygen
31
Star-forming galaxies with nebular He II 4686 emission
Figure 2.6 The contours show the mass-metallicity relation for SDSS galaxies
(Tremonti et al. 2004). The present sample of star-forming galaxies with WR
features is shown by blue circles, while the red triangles show the locations of
those that do not show WR features. At low mass we see that our sample is offset
from the rest bulk of the of SDSS but overall they sample much the same region.
abundance as none of the [O ii] lines are available.
Figure 2.5 compares different abundance indicators in the He ii λ4686/Hβ flux
ratio versus oxygen abundance plane. The oxygen abundances derived using the
direct method are shown by black circles while those derived from the fit to the
CL01 model are shown by red circles, the blue squares show O3N2 oxygen abundances. The main conclusion we can draw from this comparison is that all methods
agree well at low metallicity while at high metallicity the trends are similar but the
O3N2 estimator reaches a lower maximum O/H. For concreteness we will adopt
the CL01 estimates for the remainder of the paper but as our main focus will be
on the low metallicity region, our results are robust to the estimator chosen.
Figure 2.6 shows the mass-metallicity relation for the sample compared to the
mass-metallicity relation for all star-forming SDSS galaxies (Tremonti et al. 2004)
shown as a contour. The sample galaxies with WR features in the spectra are
shown as blue circles, while those that do not show WR features are plotted as
red triangles, we will use the same symbols in the following. Overall there is a
reasonable agreement with the main SDSS sample except for an offset towards
slightly lower metallicity at a fixed mass at low masses.
2.4
Model predictions
In the previous section we carried out an empirical analysis of the properties of
star-forming galaxies in the SDSS which show strong nebular He ii λ4686 emission
in their integrated spectra. Now we will build on the preceding to explore whether
32
Model predictions
Figure 2.7 This figure shows the logarithm of the ratio of the model prediction
for He ii λ4686/Hβ to the observed ratio as a function of oxygen abundance. Red
triangles show the ratio for objects without WR features. It is clear that there is
good agreement in the range 8.4 < 12 + log O/H < 8.8, but at lower metallicity, the
discrepancy between model and observations can be up to and order of magnitude.
At higher than solar metallicity where an AGN contribution to the He ii λ4686 flux
is more likely we see that model also fail to predict the same ratio as the observed
value.
current stellar models can be used to explain the He ii λ4686 emission seen in the
spectra of these galaxies. We start by predicting nebular He ii λ4686 emission for
galaxies in our sample with the CL01 model. Then we change the stellar population
model and explore the effect of changing the model on the predicted He ii λ4686
line flux.
2.4.1
CL01 predictions for nebular He II emission
We follow the same procedure as in the calculation of PDFs for the galaxy parameters in the previous section and calculate the likelihood of the model for each
object in our sample by fitting the CL01 grid of models to the five important
[O ii] λ3727, 3729, Hβ, [O iii] λ5007, Hα, [N ii] λ6584 emission lines.
We now want to see whether the models that reproduce the main strong lines
in the optical spectrum also reproduce the He ii λ4686 emission line strength. We
build the likelihood distribution of He ii λ4686 flux for each galaxy in the same
way as before by weighting the He ii λ4686 flux in each model by the probability of
that model. We take the median of the likelihood distribution as a prediction for
nebular He ii λ4686 emission and the associated confidence interval to be the 16th84th percentile range. We follow the same approach to estimate the He iiλ4686/Hβ
ratio.
33
Star-forming galaxies with nebular He II 4686 emission
Figure 2.8 This figure shows the spectral energy distribution (SED) of an instantaneous burst calculated with Starburst99. Each panel corresponds to one metallicity
and shows the SED for a range of ages as indicated. Note that the appearance of
WR stars 4 Myr after the burst results in a much harder UV continuum. After
5-6 Myr the WR stars disappear and the UV continuum rapidly fades. Also note
that in these models, the lowest metallicity SED does not show a significant WR
phase.
In Figure 2.7 we show the logarithm of the ratio of this model prediction
for He ii λ4686/Hβ to the observed He ii λ4686/Hβ ratio as a function of oxygen
abundance. It is clear that there is acceptable agreement between the model
predictions and the observations in the range 8.4 < 12 + log O/H < 8.8, but a model
that can reproduce most the strong lines in the spectrum well, predicts up to one
order of magnitude lower He ii λ4686/Hβ ratio than the observed value for some
objects at lower metallicities. We also see a deviation at high metallicity — in this
regime an AGN contribution to the He ii line flux is more likely, both because the
galaxies are more massive, and also because the star formation-AGN separation is
more gradual in this regime. We will not discuss this mismatch further here.
The population synthesis model used in the CL01 models approximate the WR
emission as black bodies at their effective temperature, note in passing that this
is not the case in the current BC03 models. This will overestimate the hardness
of the ionizing spectra compared to models that consider more sophisticated WR
34
Model predictions
atmosphere models such as e.g. Starburst99 (Leitherer et al., 1999) and BPASS
(Eldridge et al., 2008). Given the relatively simple treatment of the WR phases
in the CL01 models, one might be concerned that the failure to match the data is
due to an inherent weakness of the models. In the following we therefore look at
the effect of different stellar evolution and atmosphere models on the prediction of
He ii λ4686 emission line strengths.
Figure 2.9 Each panel compare the calculated SED of two SFHs with starburst99
for a different metallicity, solid lines show SEDs of instantaneous burst and dotted
lines show SEDs with constant star formation rate and different colors show different burst ages. Fluxes have been normalized to the flux at 912 Å. We can see
although models with constant star formation form WR stars continuously after
4 Myr, the overall shape of the UV continuum is softer than in the instantaneous
burst models because the continuous formation of luminous O stars softens the
extreme UV spectrum for a fixed rate of hydrogen ionizing photons.
2.4.2
Starburst99 predictions
To better understand the origin of the He ii λ4686 emission and its dependence on
the stellar models adopted, we use the latest version (6.0) of the spectral synthesis
code Starburst99 (Stb99, Leitherer et al., 1999, 2010) to calculate spectral energy
distributions (SEDs) predictions for a range of ages and metal abundances. We
35
Star-forming galaxies with nebular He II 4686 emission
calculate models with an instantaneous burst and a constant star formation history with a Kroupa IMF. We have explored a range of stellar evolution models
but the differences are small so we only show the results of one model. For this
we adopt the Padova AGB evolutionary tracks combined with Pauldrach/Hillier
atmospheres (Smith et al. 2002), Stb99 uses O star model spectra from Pauldrach
et al. (2001) and WR model spectra from the code of Hillier & Miller (1998). The
models include stellar and nebular continuum. We create models with different
metallicities (0.0004, 0.004, 0.008, 0.02, 0.05), where the reference solar metallicity
is Z = 0.02 (Anders & Grevesse 1989). We do not consider dust and run the
models up to 100 Myr with time-steps of 105 years. Figure 2.8 shows the resulting
SEDs for an instantaneous burst with a range of metallicities. Each panel corresponds to one metallicity as indicated and shows the time of the SED for six burst
ages. The plots show clearly that the appearance of WR stars, 4 Myr after the
burst, results in a harder UV continuum shortwards of the He+ ionizing edge at
228 Å. After ∼ 5 Myr the WR stars disappear and the UV continuum becomes
softer. This is in good agreement with the discussion in Schaerer & Vacca (1998,
see their Figure 9), which is natural as those models lie at core of the WR modeling in Stb99. Figure 2.9 compares the SEDs of an instantaneous burst (solid
line) with that of a continuous star formation model (dotted line). The SEDs have
been normalized at 912Å, so have the same amount of hydrogen ionizing photons.
The continuous star formation models form WR stars continuously after 3 Myr,
but the overall shape of the UV continuum is softer than for instantaneous burst
models because of the continuous formation of luminous O stars which dilute the
SED for a given total mass.
To calculate the emission lines, we use the UV continuum generated by the
Stb99 models as an input to the photoionization code Cloudy (version c08, Ferland et al., 1998). For each time step, ionization bounded models are calculated by
varying the ionization parameter log U = −2, −3, −4, where for consistency with the
CL01 models we calculate the ionization parameter at the edge of the Strömgren
sphere, and a constant hydrogen density of log nH /cm−3 = 2.5. This range in ionization parameter spans the range found by Stasińska & Leitherer (1996) in their
analysis of intensely star-forming galaxies (approximately −3.5 < log U < −2.5).
Figure 2.10 shows the calculated He ii λ4686/Hβ ratio versus [N ii] λ6584/Hα for
different metallicities and different ionization parameters in the left panel and different burst ages in the right panel. The triangles along each model line correspond
to the metallicities (0.02, 0.2, 0.4, 1) Z in the left panel and to ages of 3, 4, 5,
and 6 Myr in the right panel. In the left panel we fix the the age to 4 Myr, and in
the right we set log U = −2. The galaxies in our sample are shown as filled purple
circles. From these plots we can see how the He ii λ4686/Hβ ratio depends on age,
metallicity and ionization parameter. The lowest metallicity considered in this
work is Z = 0.02 Z and for this metallicity there is a strong discrepancy between
the model predictions and the observed data but predictions for other metallicities at 4 Myr agree well with the observed He ii λ4686/Hβ ratios. The models can
predict the He ii λ4686 emission line ratio, but only for instantaneous bursts with
metallicity of 20% solar and above, and only for ages of ∼ 4 − 5 Myr, the period
when the extreme-ultraviolet continuum is dominated by emission from WR stars.
36
Model predictions
Figure 2.10 The left panel shows the starburst99 instantaneous burst model prediction for the He ii λ4686/Hβ ratio for different ionization parameters at a fixed
age of 4 Myr. The triangles along each line correspond to the metallicities (0.02,
0.2, 0.4, 1) Z and our sample galaxies are shown as purple circles. The right hand
panel shows the same for a range of different metallicities and ages with log U set
to −2. The triangles along each model line correspond to the burst ages (3, 4, 5, 6)
Myr. Note that the lowest metallicity model is unable to cover the observational
data. The grey dashed-dotted lines in each panel show the N2 PP04 metallicity
calibration for 12 + log O/H = 8.2
For burst ages younger than 4 Myr and older than 6 Myr, and for models with a
continuous star formation (not shown here), the softer ionizing continuum results
in an emission spectrum that has too weak He ii lines to be consistent with the
observational data.
2.4.3
The effect of binary evolution on the He II 4686 emission
The Stb99 models consider single-star evolution only, but it is well-known that
massive stars are frequently found in binaries and higher order systems which can
have a major effect on the evolution of massive stars. To explore this possibility
we compare the observed He ii λ4686/Hβ ratio to the prediction of the Binary
Population and Spectral Synthesis (BPASS, Eldridge et al 2008, 2009, 2011) model.
The BPASS code includes a careful treatment of the effect of binary evolution on
massive short lived stars, and Eldridge et al found that including massive binary
evolution in the stellar population leads to WR stars forming over a wider range
of ages up to 10 Myr which increases the UV flux at later times. In Figure 2.11
and Figure 2.12 we show the observed He ii λ4686/Hβ ratio in comparison with
their instantaneous burst binary and single-star population models, respectively.
Each panel shows four different metallicities (0.05, 0.2, 0.4,1) Z . The grey dasheddotted line shows the N2 PP04 metallicity calibration for 12 + log O/H = 8.2. We
should note that the lowest metallicity in these plots is a factor of 2.5 higher than
that of Stb99 (0.02 Z ).
37
Star-forming galaxies with nebular He II 4686 emission
Figure 2.11 The plots show the prediction of instantaneous burst binary model
(BPASS code) for He ii λ4686/Hβ ratio for different metallicities (0.05, 0.2, 0.4,1)
z and burst ages. The grey dashed-dotted line shows the N2 PP04 metallicity
calibration for 12 + log O/H = 8.2.
We get the highest value for the He ii λ4686/Hβ ratio at 3 Myr and the period
with elevated He ii λ4686/Hβ lasts longer. Comparing Figure 2.11 and Figure 2.12,
there is not a striking difference in the predicted peak ratio for He ii λ4686/Hβ
between single-star and binary population models but clearly the binary model
predict an elevated ratio for a longer period of time than the single-star models.
2.5
The origin of nebular He II 4686 emission
The preceding discussion makes it clear that in standard models for stellar evolution, only during the phases where the ionizing spectrum is dominated by WR
stars will we see strong He ii λ4686 emission. This was already pointed out by
Schaerer (1996) and our results are in good agreement with that work as well as
a number of other previous works (Schaerer & Vacca, 1998; Guseva et al., 2000;
Thuan & Izotov, 2005).
What has been less studied is the direct test of this prediction, namely to
ask whether WR features are seen whenever He ii λ4686 is observed. Kehrig et. al
38
The origin of nebular He II 4686 emission
Figure 2.12 The plots show the prediction of instantaneous burst single-star model
using the BPASS code for the He ii λ4686/Hβ ratio for different metallicities, (0.05,
0.2, 0.4, 1) Z , and burst ages. The grey dashed-dotted line shows the N2 PP04
metallicity calibration for 12 + log O/H = 8.2.
(2011) and Neugent & Massey (2011) have recently studied individual star forming
regions with He ii λ4686 emission in the Local Group and have shown that while
most He ii -emitting regions also show evidence of WR stars, not all do. In more
distant galaxies, previous efforts have primarily looked at He ii emission in samples
selected for other purposes. As an example, B08 looked at He ii galaxies in SDSS
DR6 and studied whether they showed WR features, but their sample selection
was not optimised for finding galaxies with He ii λ4686 emission. The excellent
analysis of Thuan & Izotov (2005) is another example, where the focus was on
BCD galaxies. Since the present sample is selected purely on the presence of
He ii λ4686, we can carry out this study with more reliability but we need to check
for WR features in the spectra of our galaxies.
The presence of WR stars can be recognised via the WR bumps around λ4650
Å(blue bump) and λ5808 Å(red bump). The former is a blend of He ii λ4686
and several metal lines while the latter is caused by Civλ5801 − 12 (see Crowther
ARA&A review (2007), B08 and references therein). We have therefore inspected
the spectra in our star-forming sample following the methodology described in B08
(see Table 2.3). Each spectrum is assigned a classification of 0 (no WR), 1 (possible
39
Star-forming galaxies with nebular He II 4686 emission
Figure 2.13 This plot shows a comparison between two stacked spectra of galaxies
with He ii emission. The black spectrum shows the result of stacking the spectra
of all galaxies in our sample that show WR features. In this case we can clearly see
the blue WR bump (gray shaded region). The red spectrum shows the result of
stacking the spectra of all galaxies that show no WR features. Despite the increase
in S/N we see no sign of WR features, strengthening our claim that this class of
galaxies show no signs of WR features.
WR), 2 (very probable WR), or 3 (certain WR). We will consider any spectrum
with class 1, 2 or 3 to show WR features. To check the reliability of these classifications we can check whether duplicate observations of the same object are given the
same classification. There are four sets of duplicate observations (0417-51821-513
Class 2, 0418-51817-302 Class 3, 0418-51884-319 Class 2), (0455-51909-073 Class
0, 0456-51910-306 Class 0), (0266-51602-089 Class 1, 0266-51630-100 Class 0) and
(0308-51662-081 Class 3, 0920-52411-575 Class 3). Only for 0266-51602-089 and
0266-51630-100 is there uncertainty whether spectrum show evidence of WR features or not — we choose to keep the original classifications. This is in agreement
with, but somewhat better than what we find for the full sample of WR galaxies
from the B08 sample where the RMS classification uncertainty from duplicates is
0.4 classes without any apparent dependence on the median S/N of the spectra
down to S/N∼ 10; at lower S/N we do not have sufficient numbers of duplicate
observations to make a statement. In total we find that 116 of objects show WR
features.
40
The origin of nebular He II 4686 emission
To check whether the non-detection of WR features is due to a low S/N, we
have created stacked spectra of galaxies with and without WR features. The result
of this exercise is shown in in Figure 2.13. The black line is the stack of spectra
that show WR features while in red we show the stack of non-WR galaxies. We
see no sign of WR features in the stack, further supporting the notion that this
class of objects show no signs of WR features.
We now turn to explore whether there are physical differences in the galaxies showing WR features or not. Figure 2.14 shows the ratio of He ii λ4686/Hβ
versus oxygen abundance for galaxies with (blue circles) and without WR features (triangles). The symbols for the non-WR galaxies, here and in the following,
are colored red for 12 + log O/H < 8.2, and orange otherwise. At oxygen abundances lower than 8.2 we see there is a trend of increasing He ii λ4686/Hβ ratio
towards lower metallicities. We interpret this as being due to a harder ionizing continuum at lower metallicities (e.g. Thuan & Izotov, 2005, and references
therein). We also see a trend of increasing He ii λ4686/Hβ towards higher metallicity for 12 + log O/H > 8.7. It is less clear what causes this, but since these are
more massive systems with higher star formation rates and with stronger stellar
winds due to their higher metallicity, it is likely that what we are seeing is due to
an increased contribution of shocks and/or a low-level AGN contamination.
Another striking result is that for 12 + log O/H > 8.2, essentially all He ii emitting galaxies show WR features, in agreement with what one would expect
from the models discussed in the previous section. At high metallicity there appear
to be some systems that show a high He iiλ4686/Hβ ratios but no sign of WR stars.
Since these are more massive systems, and fall intermediate between the SF and
AGN groups in the He ii λ4686/Hβ versus [N ii] λ6584/Hα diagram, we interpret
this as a likely sign of a low-level AGN contribution.
In Figure 2.15 we show the fraction of galaxies that show WR features as a
function of metallicity. To calculate this we draw a number of random realisations
of the data using the uncertainty estimates on the oxygen abundance to draw a
random realisation. We also draw a random realisation of the WR classification
assuming a random uncertainty in the classification of 0.4 as determined from
the analysis of duplicate spectra as discussed above. We carry out 101 random
realisations for each of the 101 bootstrap repetitions and calculate the median
fraction of galaxies showing WR features in each metallicity bin as well as the 16%–
84% scatter around the median which is shown by the error bars in Figure 2.15.
What is clear is that there is a transition at around an oxygen abundance of
12 + log O/H ≈ 8.2 ± 0.1. The uncertainty of 0.1 dex encapsulates the fact that
the exact value of this transition abundance depends somewhat on the metallicity
calibration adopted and 0.1 dex corresponds to the scatter found when using the
different calibrations discussed earlier.
As we showed in earlier, e.g. Figure 2.10, the He ii λ4686/Hβ ratio depends
strongly on the age of the starbursts so it is reasonable to ask whether the systems
without WR features are systematically younger or older than the systems that
show WR features. In Figure 2.16 we test this by plotting the He ii λ4686/Hβ ratio
versus EW(Hβ) for the sample, where we take the EW(Hβ) as a proxy for starburst
age. We show galaxies with WR features by black and blue circles at high and low
41
Star-forming galaxies with nebular He II 4686 emission
Figure 2.14 The He iiλ4686/Hβ ratio versus oxygen abundance for the sample. The
blue solid circles show the location of galaxies with WR features. Galaxies which
do not show WR features are indicated by red and orange triangles at lower and
higher oxygen abundances than 12 + log O/H = 8.2, respectively.
metallicities, respectively. Galaxies without WR features are as before shown by
orange and red triangles at high and low metallicities. We see the He ii λ4686/Hβ
ratios decrease with EW(Hβ) for both WR and non-WR objects. However, galaxies
without WR features show higher ratios than galaxies with WR feature for the
same EW(Hβ), especially at lower EW(Hβ). Alternatively one might say that at
a fixed He ii λ4686/Hβ the systems without WR features have a higher EW(Hβ),
or with our assumption, a younger starburst. It is not possible with our data to
disentangle these two possibilities.
2.6
Why are there galaxies with He II 4686 emission but no WR features?
The models that we discussed previously agree that WR stars are the source of the
hard ionizing photons necessary to produce the He iiλ4686 emission. It is therefore
a puzzle why some of the galaxies with nebular He ii λ4686 emission do not show
stellar WR features. This lack of WR features has been pointed out before (e.g
Thuan & Izotov, 2005), but this is the first time a clear trend with metallicity
appears. In this section we therefore turn to discuss some possible reasons for this
lack of WR features.
1. Differences in the S/N of the spectra
To detect the WR features we need fairly high S/N in the continuum as they are
broad and weak features. In the top panel in Figure 2.17 we show the relationship
between the equivalent width of the WR blue bump and the median S/N in the
42
Why are there galaxies with He II 4686 emission but no WR features?
Figure 2.15 The fraction of objects with detected WR features in the He ii sample
as a function of gas-phase oxygen abundance. The points show the median fraction
in each abundance bin and the error bars the 16%-84% scatter around the median
(see text for details). While essentially all high-metallicity star-forming galaxies
with He ii λ4686 nebular emission show WR features, this fraction drops rapidly
at metallicities below 12 + log O/H ≈ 8.2.
continuum for WR galaxies in SDSS DR7. We see that a S/N of 10 is sufficient to
detect even very weak features. The distribution of the S/N for WR and non-WR
objects in the sample versus their metallicities in the bottom panel shows that 20
out of 70 of the non-WR galaxies at 12 + log O/H < 8.2 have S/N less than 10.
Since the total number of galaxies at 12+log O/H < 8.2 is 115, if all the 20 low S/N
galaxies are assumed to have WR features this brings the fraction of galaxies with
WR features in this metallicity range from 30% to 43%. Thus low S/N could cause
us to underestimate the WR fraction by at most ∼ 15%. But the fact that the
co-added spectrum in Figure 2.13 which has a higher S/N shows no WR features
is suggestive that S/N might not be the problem for non-detection of WR features.
Furthermore even adding 15% would still mean that less than half He ii emitters at low metallicity would show WR features. We also studied the difference
in WR classification of duplicate observations as a function S/N and as remarked
earlier, we saw no trend with S/N for more uncertain classification down to a S/N
of 10. The dataset is insufficient to test this at lower S/N. Thus our conclusion is
that it is unlikely that the systematic absence of WR features in low metallicity
objects is due to low S/N in the spectra.
2. Weak lined WN stars
One possible reason for the lack of WR features in the most metal poor galaxies,
is that the WR lines are too weak to be seen. It is well known (e.g. Conti et al.,
1989; Crowther & Hadfield, 2006) that WR stars in the SMC have narrower and
less luminous lines than equivalent stars in the Milky Way. Crowther & Hadfield
43
Star-forming galaxies with nebular He II 4686 emission
Figure 2.16 This plot shows the dependence of the He ii λ4686/Hβ ratio on the
EW(Hβ) and metallicity. We show galaxies with WR features by black and blue
circles at high and low metallicities, respectively. Galaxies without WR feature
are shown by orange and red triangles at high and low metallicities. We see the
He ii λ4686/Hβ ratios decrease with EW(Hβ) for both WR and non-WR objects.
However, galaxies without WR features show higher ratios than galaxies with
WR feature for the same EW(Hβ), especially at lower EW(Hβ)s, alternatively one
could read this to say that they have a high EW(Hβ) (young age) for a given
He ii λ4686/Hβ ratio.
(2006), for instance, find that He II lines in WR stars in the SMC are typically
a factor of 4-5 weaker than the Milky Way and hence one would expect that in
galaxies at the same distance, it would be harder to see WR features in lower
metallicity systems. This is however countered by the fact that low metallicity
systems on average are closer, and hence the SDSS fibre subtends a smaller physical
size. Since WR emitting regions typically are small, they do not fill the 3” aperture
in more distant galaxies, which means that the contrast of the WR features is being
enhanced in low redshift systems. This is reflected in the fact that in the WR
survey by B08, the equivalent width of the WR features in low metallicity systems
(12 + Log O/H < 8.25) is slightly higher than that in metal rich objects (12 +
Log O/H > 8.5), a mean of 5.1 Å vs 4.1 Å. Thus the increased contrast appears
to approximately cancel out the decrease in line luminosity leading to a fairly
constant detection potential with redshift. Thus we do not believe that this is the
cause of the dearth of WR features in the low metallicity He ii emitting galaxies,
at least down to 12 + log O/H ∼ 8. At very low metallicities, 12 + log O/H ∼ 7.5,
Eldridge & Stanway (2009) found that in their models the WR features become
very weak and if those results are correct it should be extremely hard to detect WR
features in those galaxies. We do however caution that I Zw 18 has very prominent
WR features (e.g. Legrand et al., 1997), so at least some extremely metal poor
44
Why are there galaxies with He II 4686 emission but no WR features?
galaxies do show clear WR features. Furthermore, even if we are unable to detect
WR features at the very lowest metallicities, the same models predict strong WR
features at 12 + log O/H > 8, thus this is not a sufficient explanation for the result
in Figure 2.15.
3. Shocks
Thuan & Izotov (2005) studied the hard ionizing radiation in very metal poor
BCD galaxies in the local Universe and concluded that fast radiative shocks could
be responsible for the nebular He ii λ4686 emission. Therefore, another possibility
is that there is a contribution to the He ii λ4686 from shocks in the ISM (Dopita
& Sutherland, 1996). That He ii λ4686 can have some contribution from shocks is
plausible, but whether it can explain the systematic lack of WR features at low
metallicity is less clear.
One question is whether shock models can reproduce the line luminosity in He ii
at low metallicities. Using the predictions of the Dopita et al. (2005) and Allen
et al. (2008) shock models for the He ii λ4686/Hβ ratio versus [N ii] λ6584/Hα, we
found that we can only obtain a ratio comparable to the observed one for objects
at high metallicity. A second issue centers on the observation that shocks would
most likely come from supernovae and stellar winds, but the latter are though
to be weaker at low metallicities. Shocks can also be induced by outflows from
starbursts (e.g. galactic winds) and mergers but only two galaxies in our sample
are interacting and show a perturbed morphology and we see no clear difference
between the low- and high-metallicity subsamples.
4. X-ray binaries
Another candidate source for He ii ionization that has been discussed in previous studies of He ii λ4686 emitting nebulae are massive X-ray binaries (Garnett
et al. 1991). While these are likely present in many active star forming regions,
the question here is why massive X-ray binaries should be more common at low
metallicity. If the He+ ionizing photons at low metallicities come from X-ray binaries, we would expect an increasing binary fraction with decreasing metallicity.
Without a theoretical justification for this, we consider an increased abundance of
X-ray binaries at low metallicity to be an unlikely explanation for the trend seen
in Figure 2.15.
5. post-AGB stars
Binette et al. (1994) demonstrated that photoionization by post-AGB stars can
produce nebular He iiλ4686 emission. After about a few 107 years (i.e, after massive
stars disappear), the ionizing radiation comes from post-AGB stars. So, the ionization from post-AGB stars become more important in more evolved systems and
this is not the case for our objects especially not for objects at low metallicities
(c.f. Figure 2.16).
6. Spatial offset
One possible explanation for non detection of WR features is that there could
be a significant spatial separation between the WR stars and the region emitting
He ii . Kehrig et. al (2008) saw indeed such a spatial separation based on integral
field spectra of II Zw 70. The found that the location of the WR stars and the
He ii λ4686 emission appear to be separated by ∼ 80pc.
Similarly, Izotov et al. (2006b) studied two-dimensional spectra of an extremely
45
Star-forming galaxies with nebular He II 4686 emission
Figure 2.17 Top panel: The relationship between the equivalent width of the WR
blue bump and the median S/N in the continuum for WR galaxies in SDSS DR7.
We see that a S/N of 10 is sufficient to detect even very weak features. Bottom
panel: The distribution of the median S/N in the continuum for WR and non-WR
objects in the sample versus their oxygen abundance, we see that 20 out of 70 of
the non-WR galaxies at 12 + log O/H < 8.2 have S/N less than 10. See text for
more discussion.
46
Conclusion
metal-deficient BCD galaxy SBS 0335-052E and showed that the He ii λ4686 emission line was also offset from the near evolved star clusters but in their case, by
studying the kinematical properties of the ionized gas from the different emission
lines they suggested that the hard ionizing radiation responsible for the He ii λ4686
emission was not related to the most massive youngest stars, but rather was related
to fast radiative shocks.
If this offset between the WR stars and the region emitting He ii might be
an explanation for our non-detections of WR features, it would mean that such a
spatial separation is much more common at low metallicity which is rather surprising, since stellar winds are thought to be considerably weaker there, however
low metallicity galaxies are also on average closer so the SDSS fibre subtends a
smaller physical scale and hence a smaller volume would need to be blown out by
the wind (see Figure 2.18). To get a more quantitative estimate we calculate the
gravitational binding energy of a cloud with radius 1.5” (SDSS aperture radius)
at the redshift of each non-WR object. We assume a hydrogen density of ∼ 50
cm−3 . The median energy require to excavate a hole of this size, is of the order of
1055 erg for our full sample. For the lowest redshift objects the energy requirement
is a much more manageable 1049 erg and thus in these cases the absence of WR
features in the spectrum could be due to a spatial offset from the He ii emitting
region. At higher redshift the energetics makes this a much less likely explanation.
The possibility does however warrant further examination and to test this we are
undertaking a spectroscopic follow-up of a subsample of these sources.
7. Chemically homogeneous stellar evolution
A final explanation could be that the stellar populations at very low metallicities
can have much higher temperatures than is currently expected in models. This
would be the case if some stars rotated fast enough to evolve homogeneously
(Maeder, 1987; Meynet & Maeder, 2007; Yoon & Langer, 2005; Yoon et al., 2006;
Cantiello et al., 2007). In that case we can get a higher continuum at 228 Å and correspondingly a higher He ii λ4686/Hβ ratio in comparison with non-homogeneous
stellar evolution models. An appealing aspect of this, speculative, explanation is
that homogeneous evolution is predicted to be more common at low metallicity.
There are however currently no studies of the nebular He ii λ4686 line in the literature. Eldridge & Stanway (2011b) looked at the effect of quasi-homogeneous
evolution in their binary models and showed that including it led to strengthened
WR features at low redshift, in contrast to what we need. However that still leaves
the possibility open that there is a period where strong nebular He ii λ4686 is seen
but no WR features as that has not yet been tested.
2.7
Conclusion
We have presented a sample of rare star-forming galaxies with strong nebular
He ii λ4686 emission spanning a wide range in metallicity. We have derived physical parameters for these galaxies and showed that emission line models that can
reproduce the strong lines in the galaxy spectra are not able to predict the observed
ratio of He ii λ4686/Hβ at low metallicities. In agreement with previous studies
47
Star-forming galaxies with nebular He II 4686 emission
we found that current models for single massive stars are able to reproduce the
He ii λ4686/Hβ ratio in galaxies in our sample, but only for instantaneous bursts of
20% solar metallicity or higher, and only for ages of ∼ 4 − 5 Myr, the period when
the extreme-ultraviolet continuum is dominated by emission from WR stars. For
stars younger than 4 Myr or older than 5 Myr, and for models with a constant
star-formation rate, the softer ionizing continuum results in He ii λ4686/Hβ ratios
typically too low to explain our data. Including massive binary evolution in the
stellar population analysis leads to WR stars occurring over a wider range in age
which leads to acceptable agreement with the data at all metallicities sampled as
long as WR stars are present.
However, the most notable result of our studies is that a large fraction of the
galaxies in our sample do not show WR features and this fraction increases systematically with decreasing metallicity. We find that 70% of galaxies at oxygen
abundances lower than 8.2 do not show WR features in their spectra. We discussed a range of different mechanism responsible for producing He ii λ4686 line
apart from WR stars in these galaxies and conclude that spatial separation between
WR stars and the region emitting He ii emission can be a possible explanation for
non-detection of WR features in these galaxies. Moreover, if the stellar population
models at very low metallicities can have much higher temperatures than is currently expected in models, as would for instance be the case if some stars rotate
fast enough to evolve homogeneously, then such models might explain the origin
of the He ii λ4686 line and also the metallicity trend of the He ii sample better.
We will explore these possibilities in a future paper.
Acknowledgements
First of all we would like to thank the anonymous referee for insightful comments
and valuable suggestions which improved the paper significantly.
We also would like to thank Marijn Franx, Norbert Langer, S.-C. Yoon and
Brent Groves for useful discussion. Finally, we would like to express our appreciation to Daniel Schaerer, Johan Eldridge, Carolina Kehrig and Alireza Rahmati
for their kind comments on this paper.
Funding for the Sloan Digital Sky Survey (SDSS) and SDSS-II has been
provided by the Alfred P. Sloan Foundation, the Participating Institutions, the
National Science Foundation, the U.S. Department of Energy, the National Aeronautics and Space Administration, the Japanese Monbukagakusho, and the Max
Planck Society, and the Higher Education Funding Council for England. The
SDSS Web site is http://www.sdss.org/.
The SDSS is managed by the Astrophysical Research Consortium (ARC) for
the Participating Institutions. The Participating Institutions are the American
Museum of Natural History, Astrophysical Institute Potsdam, University of Basel,
University of Cambridge, Case Western Reserve University, The University of
Chicago, Drexel University, Fermilab, the Institute for Advanced Study, the Japan
Participation Group, The Johns Hopkins University, the Joint Institute for Nuclear Astrophysics, the Kavli Institute for Particle Astrophysics and Cosmology, the
48
Conclusion
Korean Scientist Group, the Chinese Academy of Sciences (LAMOST), Los Alamos
National Laboratory, the Max-Planck-Institute for Astronomy (MPIA), the MaxPlanck-Institute for Astrophysics (MPA), New Mexico State University, Ohio State
University, University of Pittsburgh, University of Portsmouth, Princeton University, the United States Naval Observatory, and the University of Washington.
Many thanks go to Allan Brighton, Thomas Herlin, Miguel Albrecht, Daniel
Durand and Peter Biereichel, who are responsible for the SkyCat developments at
ESO, in particular for making their software free for general use. GAIA and SkyCat
are both based on the scripting language Tcl/Tk developed by John Ousterhout
and the [incr Tcl] object oriented extensions developed by Michael McLennan.
They also make use of many other extensions and scripts developed by the Tcl
community. Thanks are also due to the many people who helped test out GAIA
and iron out minor and major problems (in particular Tim Jenness, Tim Gledhill
and Nigel Metcalfe) and all the users who have reported bugs and sent support
since the early releases and continue to do so. The 3D facilities of GAIA make
extensive use of the VTK library. Also a free library.
GAIA was created by the now closed Starlink UK project, funded by the
Particle Physics and Astronomy Research Council (PPARC) and has been more
recently supported by the Joint Astronomy Centre Hawaii funded again by PPARC
and more recently by its successor organisation the Science and Technology Facilities Council (STFC).
This research has made use of the Perl Data Language (PDL, http://pdl.
perl.org) and the Interactive Data Language (IDL).
49
Star-forming galaxies with nebular He II 4686 emission
616-52442-364
(# 40) z= 0.0022
0667-52163-634
(# 45) z= 0.0049
1 kpc
0947-52411-569
(# 62) z= 0.0075
0752-52251-340
(# 51) z= 0.0029
1 kpc
1 kpc
1446-53080-134
(#103) z= 0.0034
1 kpc
2016-53799-185
(#151) z= 0.0162
1 kpc
1646-53498-616
(#115) z= 0.0027
1 kpc
2211-53786-486
(#163) z= 0.0021
1 kpc
2329-53725-536
(#169) z= 0.0097
1 kpc
1 kpc
Figure 2.18 Images of nine galaxies which do not show WR features in their spectra
while having strong He ii λ4686 emission. The sizes of boxes are 50 ” × 50 ”. In
these images north is up, east to the left. Red circles show the size of SDSS fibre
aperture (3”) and yellow boxes show 100 pc around the center of the SDSS fibre.
50
+00
+00
+00
+00
+00
δ
00
00
01
01
01
06.78
10.81
24.49
25.06
33.00
Plate–MJD–FiberID
0285-51930-485
0363-51989-400
0470-51929-309
0294-51986-438
0694-52209-113
log He ii /Hβ
-0.93
-0.79
-0.84
-1.12
-0.88
log[N ii] /Hα
-0.14
-0.31
-0.01
-0.07
-0.36
Class
AGN
AGN
AGN
AGN
AGN
Class(BPT)
AGN
AGN
AGN
AGN
AGN
SN
OKSN
OKSN
OKSN
OKSN
OKSN
WR Features
Non-WR
Non-WR
WR
Non-WR
Non-WR
Class(WR)
—
—
1
0
—
Plate–MJD–FiberID
0752-52251-340
0390-51900-291
0390-51900-445
0753-52233-094
0418-51884-319
0691-52199-389
1905-53706-628
0394-51913-075
0695-52202-137
2329-53725-536
0429-51820-495
0666-52149-331
1073-52649-409
0667-52163-634
0456-51910-306
1070-52591-072
0456-51910-076
1666-52991-310
0413-51821-480
1733-53047-528
1922-53315-588
0761-52266-361
1267-52932-384
0548-51986-503
Z
0.00291
0.09437
0.09840
0.01424
0.01791
0.19554
0.15928
0.16742
0.00556
0.00975
0.05662
0.15155
0.03994
0.00492
0.08222
0.00421
0.00456
0.03992
0.06868
0.07881
0.06978
0.04592
0.04722
0.00706
α(J2000)
+00 09 53.09
+00 17 28.29
+00 21 01.03
+00 24 25.95
+00 32 18.59
+00 42 52.31
+00 49 13.88
+00 55 27.46
+01 15 33.82
+01 25 34.19
+01 47 07.04
+01 59 53.07
+02 13 06.62
+02 15 13.98
+02 40 52.20
+02 42 39.86
+02 48 15.94
+03 14 31.71
+03 17 43.12
+07 29 30.29
+08 06 19.50
+08 22 27.44
+08 23 54.97
+08 26 04.80
δ
+15 44 04.80
-00 56 24.98
+00 52 48.08
+14 04 10.65
+15 00 14.17
+00 27 30.09
+00 24 01.99
-00 21 48.77
-00 51 31.17
+07 59 24.40
+13 56 29.29
-08 13 48.99
+00 56 12.44
-08 46 24.39
-08 28 27.43
-00 00 58.64
-08 17 16.51
+41 05 25.95
+00 19 36.84
+39 49 41.62
+19 49 27.31
+42 23 31.14
+28 06 21.75
+45 58 07.36
ID
51
13
14
52
19
48
138
15
49
169
20
44
73
45
24
72
23
117
16
121
139
53
78
30
log He ii /Hβ
-1.35
-1.57
-1.96
-1.98
-1.97
-0.75
-1.64
-1.76
-2.32
-2.02
-1.95
-1.73
-1.97
-1.12
-1.60
-1.41
-2.11
-1.95
-0.77
-1.97
-2.05
-1.09
-2.16
-2.07
log[N ii] /Hα
-1.67
-0.56
-1.10
-1.36
-1.52
-0.82
-2.05
-0.89
-1.27
-2.22
-1.52
-1.43
-1.58
-1.81
-1.63
-0.33
-1.90
-1.40
-0.54
-0.93
-1.19
-0.37
-0.95
-0.76
12 + log O/H
7.92±0.09
9.27±0.05
8.83±0.04
8.10±0.12
8.01±0.07
8.88±0.05
7.85±0.06
8.90±0.04
8.31±0.35
7.80±0.04
8.02±0.05
8.04±0.06
8.09±0.05
7.74±0.09
8.02±0.04
8.93±0.25
7.87±0.08
8.08±0.19
9.07±0.11
8.90±0.04
8.83±0.04
9.25±0.07
8.88±0.04
8.64±0.26
Features
Non-WR
WR
WR
WR
WR
WR
Non-WR
WR
WR
Non-WR
Non-WR
Non-WR
Non-WR
Non-WR
Non-WR
Non-WR
Non-WR
WR
WR
Non-WR
WR
WR
WR
WR
Class
0
2
2
3
2
2
0
2
3
0
0
0
0
0
0
0
0
1
2
0
2
1
1
1
Other Names
N/A
N/A
UM 228
N/A
SHOC 022
N/A
N/A
N/A
NGC 0450
N/A
N/A
SHOC 099
UM 411
SHOC 111
SHOC 133
M077
N/A
N/A
N/A
N/A
N/A
N/A
N/A
UGC 04393
Table 2.5: The positions and identifications of the sample of star-forming galaxies with nebular He ii λ4686 emission.
α(J2000)
+11 59 09.71
+15 30 26.32
+08 58 28.60
+13 06 00.68
+01 10 06.09
Table 2.2: The positions and the identifications of galaxies with nebular He ii λ4686 emission. OKSN is for galaxies having S N > 3
in their [O iii] λ5007, Hβ, [N ii] λ6584 and Hα lines. See B04 for the BPT classification. The full table is available in electronic form in
http://www.strw.leidenuniv.nl/∼shirazi/SB011/.
Conclusion
51
Plate–MJD–FiberID
2425-54139-372
0445-51873-404
0828-52317-148
2278-53711-411
0564-52224-216
1875-54453-549
1785-54439-201
2430-53815-117
0551-51993-279
2086-53401-458
0566-52238-497
1194-52703-397
0899-52620-594
0554-52000-190
0553-51999-602
0485-51909-550
2580-54092-470
1594-52992-563
1305-52757-269
0266-51630-100
1947-53431-448
0769-54530-086
1306-52996-005
2583-54095-062
2364-53737-618
1745-53061-475
2588-54174-369
1745-53061-196
1955-53442-354
1427-52996-221
2366-53741-124
0575-52319-521
2591-54140-222
0999-52636-517
0947-52411-569
0875-52354-226
1998-53433-304
0274-51913-187
2478-54097-370
0578-52339-060
Z
0.01974
0.00245
0.14746
0.07219
0.09109
0.07498
0.09186
0.07592
0.00961
0.00981
0.03909
0.00513
0.02727
0.00767
0.00772
0.01366
0.10060
0.01486
0.01085
0.00478
0.01730
0.04626
0.00486
0.01522
0.00423
0.06131
0.05563
0.00957
0.03762
0.00388
0.01897
0.03319
0.04432
0.04453
0.00747
0.02850
0.00489
0.01856
0.00398
0.01287
α(J2000)
+08 29 32.66
+08 37 43.48
+08 38 43.64
+08 40 00.37
+08 44 14.24
+08 51 03.67
+08 51 15.65
+08 52 21.72
+08 52 58.21
+09 05 26.34
+09 05 31.08
+09 10 28.78
+09 14 34.95
+09 20 55.92
+09 20 56.07
+09 30 06.43
+09 38 01.64
+09 42 52.78
+09 42 56.74
+09 44 01.87
+09 50 00.77
+09 51 31.77
+09 54 49.56
+09 56 42.49
+10 10 32.81
+10 10 42.54
+10 10 59.30
+10 12 27.02
+10 14 10.58
+10 16 24.52
+10 24 02.75
+10 24 29.25
+10 26 23.65
+10 33 28.53
+10 34 10.15
+10 35 08.88
+10 36 13.22
+10 39 24.38
+10 41 09.60
+10 44 57.79
δ
+14 27 06.92
+51 38 30.26
+38 53 50.50
+18 05 31.01
+02 26 21.10
+62 13 26.93
+58 40 55.02
+12 16 51.76
+49 27 33.91
+25 33 02.57
+03 35 30.38
+07 11 17.97
+47 02 07.24
+52 34 07.34
+52 34 04.32
+60 26 53.40
+13 53 17.07
+35 47 25.98
+09 28 16.26
-00 38 32.18
+30 03 41.04
+52 59 36.05
+09 16 15.94
+15 38 11.34
+22 00 39.63
+12 55 16.81
+15 42 23.53
+12 20 37.50
+34 20 34.79
+37 54 45.97
+21 04 50.04
+05 24 51.02
+17 10 14.36
+07 08 01.76
+58 03 49.06
+49 21 42.47
+37 19 27.57
-00 23 21.44
+21 21 42.80
+03 53 13.15
52
ID
173
21
56
168
34
136
130
174
31
154
35
76
60
33
32
25
186
110
80
2
140
54
81
187
171
123
188
122
141
101
172
36
189
69
62
58
146
3
176
37
log He ii /Hβ
-2.00
-2.38
-1.88
-2.02
-2.13
-2.06
-2.01
-2.06
-1.83
-1.93
-1.93
-2.02
-2.13
-2.10
-1.71
-2.17
-1.18
-1.92
-2.38
-1.88
-1.97
-2.05
-1.90
-2.02
-2.12
-1.92
-1.92
-2.24
-1.66
-1.70
-1.88
-2.07
-1.12
-1.99
-1.60
-1.79
-1.93
-2.16
-2.13
-1.80
log[N ii] /Hα
-1.35
-1.71
-1.32
-1.53
-1.20
-1.39
-1.83
-1.45
-0.62
-1.12
-1.78
-1.77
-1.55
-1.83
-1.63
-1.59
-0.43
-1.44
-1.51
-1.98
-1.51
-1.64
-0.42
-0.91
-1.91
-1.40
-1.70
-1.77
-0.54
-2.21
-1.56
-1.72
-0.64
-0.77
-1.97
-0.76
-0.53
-0.85
-1.81
-2.55
Table 2.5: continued.
12 + log O/H
8.13±0.20
7.95±0.08
8.69±0.05
7.98±0.07
8.75±0.13
8.59±0.10
7.84±0.08
8.23±0.08
8.80±0.24
8.40±0.45
7.79±0.08
7.91±0.09
7.97±0.07
7.85±0.08
7.88±0.07
7.98±0.06
9.28±0.06
8.19±0.18
8.14±0.16
7.77±0.06
8.04±0.10
7.93±0.06
9.03±0.21
8.56±0.28
7.82±0.09
8.40±0.04
7.91±0.04
8.04±0.10
9.13±0.05
7.80±0.05
8.08±0.09
7.84±0.07
9.12±0.05
8.91±0.04
7.78±0.06
8.92±0.04
9.09±0.26
8.67±0.26
7.83±0.09
7.80±0.00
Features
Non-WR
Non-WR
WR
Non-WR
WR
Non-WR
Non-WR
Non-WR
WR
WR
Non-WR
WR
Non-WR
Non-WR
WR
Non-WR
WR
Non-WR
WR
Non-WR
Non-WR
WR
WR
WR
WR
WR
Non-WR
WR
WR
Non-WR
WR
Non-WR
Non-WR
WR
Non-WR
WR
WR
WR
WR
Non-WR
Class
0
0
1
0
1
0
0
0
3
3
0
1
0
0
1
0
2
0
3
0
0
2
3
3
1
3
0
2
1
0
2
0
0
2
0
2
3
1
3
0
Other Names
N/A
MRK 0094
N/A
N/A
N/A
N/A
N/A
N/A
SBS 0849+496
UGC 04764
N/A
N/A
SBS 0911+472
N/A
N/A
SBS 0926+606A
N/A
N/A
UGC 05189
CGCG 007-025
N/A
SBS 0948+532
NGC 3049
UGC 05342
WAS 05
N/A
N/A
N/A
KUG 1011+345
N/A
LSBC D568-03
N/A
N/A
CGCG 037-076
MRK 1434
SBS 1032+496
NGC 3294
IC 0633
MRK 0724
N/A
Star-forming galaxies with nebular He II 4686 emission
Plate–MJD–FiberID
2147-53491-514
0275-51910-445
1749-53357-499
1981-53463-438
2483-53852-254
0275-51910-622
0876-52669-175
2359-53826-205
1362-53050-617
2213-53792-359
2211-53786-486
1363-53053-510
2494-54174-361
1223-52781-128
1014-52707-254
2500-54178-084
1754-53385-151
0967-52636-302
0967-52636-339
1442-53050-599
2012-53493-407
2506-54179-357
2008-53473-467
0967-52636-540
2513-54141-309
2510-53877-560
2508-53875-615
1761-53376-636
0330-52370-471
1446-53080-134
1313-52790-423
0516-52017-315
1991-53446-584
2226-53819-157
1763-53463-094
2227-53820-389
0517-52024-504
2004-53737-439
2644-54210-188
2610-54476-421
Z
0.05487
0.02620
0.01061
0.02945
0.08443
0.05061
0.00435
0.00452
0.03745
0.00214
0.00214
0.02154
0.00492
0.00345
0.00988
0.00470
0.01764
0.00084
0.02601
0.01018
0.00483
0.02082
0.00601
0.00558
0.02676
0.04512
0.07911
0.00245
0.00353
0.00337
0.01726
0.05811
0.01097
0.08188
0.06675
0.05588
0.00429
0.04894
0.02435
0.00277
α(J2000)
+10 45 20.42
+10 45 54.78
+10 46 53.99
+10 47 23.61
+10 50 32.51
+10 50 46.59
+10 53 10.82
+10 54 21.87
+11 00 24.90
+11 04 58.30
+11 04 58.54
+11 05 08.12
+11 17 46.30
+11 27 10.93
+11 27 32.67
+11 29 14.15
+11 32 35.35
+11 33 28.95
+11 34 45.72
+11 36 23.82
+11 36 39.57
+11 36 54.01
+11 41 07.49
+11 45 06.26
+11 48 05.45
+11 48 27.34
+11 48 40.87
+11 50 02.73
+11 52 37.68
+11 54 41.22
+11 55 28.34
+11 57 12.45
+11 57 31.73
+12 00 16.49
+12 00 33.42
+12 01 49.90
+12 08 11.11
+12 09 24.64
+12 09 27.95
+12 15 18.60
δ
+09 23 49.10
+01 04 05.84
+13 46 45.77
+30 21 44.29
+15 38 06.31
+00 36 40.11
+50 16 53.21
+27 14 22.16
+43 01 11.93
+29 08 16.55
+29 08 15.72
+44 44 47.24
+17 44 24.69
+08 43 51.70
+53 54 54.47
+20 34 52.01
+14 11 29.83
+49 14 13.01
+50 06 03.33
+47 09 29.08
+36 23 42.89
+19 55 34.80
+32 25 37.22
+50 18 02.44
+21 49 45.35
+25 46 11.77
+17 56 33.02
+15 01 23.48
-02 28 06.39
+46 36 36.35
+57 39 51.97
+02 28 27.88
+32 20 30.17
+27 19 59.01
+13 43 07.99
+28 06 10.67
+02 52 41.82
+32 44 02.05
+22 06 16.69
+20 38 26.72
ID
160
4
124
143
177
5
59
170
91
164
163
92
178
77
70
179
125
64
65
102
150
180
149
66
183
182
181
126
8
103
82
27
145
165
127
166
28
148
193
191
log He ii /Hβ
-2.18
-2.28
-2.23
-2.18
-1.79
-1.44
-1.70
-1.93
-2.11
-1.73
-2.18
-2.05
-2.04
-1.71
-1.98
-2.32
-2.19
-1.22
-2.25
-2.07
-1.82
-1.77
-1.82
-1.55
-1.95
-2.10
-2.00
-2.23
-2.13
-1.81
-2.20
-1.99
-2.14
-1.93
-2.05
-1.92
-1.84
-1.44
-1.91
-2.59
log[N ii] /Hα
-1.30
-1.34
-1.80
-1.15
-1.76
-0.46
-1.75
-0.74
-1.29
-1.84
-1.83
-1.28
-1.69
-1.27
-1.62
-1.44
-1.48
-1.76
-1.41
-1.92
-1.52
-0.77
-1.82
-2.04
-1.27
-1.70
-1.78
-1.59
-1.75
-1.68
-1.77
-1.15
-1.13
-1.86
-1.08
-1.71
-0.27
-0.43
-0.64
-1.80
Table 2.5: continued.
12 + log O/H
8.70±0.06
8.74±0.04
7.92±0.09
8.85±0.04
7.92±0.04
9.16±0.26
7.94±0.08
8.67±0.41
8.77±0.08
7.79±0.05
7.81±0.06
8.78±0.04
7.90±0.09
8.09±0.11
8.04±0.11
8.03±0.10
8.20±0.09
7.89±0.08
8.21±0.54
7.83±0.09
8.07±0.10
8.61±0.40
7.79±0.05
7.77±0.09
8.35±0.11
7.95±0.07
7.91±0.04
7.99±0.05
7.84±0.09
7.80±0.05
8.03±0.10
8.85±0.06
8.36±0.49
7.94±0.07
8.72±0.04
7.92±0.05
9.04±0.17
9.27±0.04
9.13±0.08
7.88±0.08
Features
WR
Non-WR
Non-WR
WR
Non-WR
WR
Non-WR
WR
Non-WR
Non-WR
Non-WR
WR
Non-WR
WR
Non-WR
WR
WR
WR
WR
Non-WR
WR
WR
Non-WR
Non-WR
WR
Non-WR
Non-WR
WR
Non-WR
Non-WR
Non-WR
WR
WR
Non-WR
WR
Non-WR
WR
WR
WR
WR
Class
2
0
0
3
0
2
0
3
0
0
0
2
0
3
0
2
2
3
2
0
1
3
0
0
2
0
0
3
0
0
0
1
3
0
2
0
2
1
2
2
Other Names
SCHG 1042+097
SHOC 308
N/A
TON 0542
N/A
N/A
N/A
NGC 3451
N/A
N/A
N/A
MRK 0162
N/A
IC 2828
MRK 1446
IC 0700
N/A
Mrk 0178
MRK 1448
N/A
NGC 3755
MRK 0182
KUG 1138+327
N/A
MRK 1459
N/A
N/A
MRK 0750
N/A
N/A
MRK 0193
UM 469
NGC 3991N
N/A
N/A
N/A
NGC 4123
N/A
UGC 07137
MRK 1315
Conclusion
53
Plate–MJD–FiberID
1625-53140-386
2001-53493-146
2880-54509-277
0955-52409-608
2880-54509-095
1453-53084-322
1371-52821-053
1371-52821-059
1452-53112-011
1452-53112-016
1615-53166-120
1768-53442-476
2613-54481-507
0494-51915-007
1372-53062-072
1975-53734-498
2236-53729-038
1455-53089-556
1989-53772-089
2461-54570-089
0602-52072-369
2018-53800-096
0339-51692-083
0782-52320-022
2016-53799-185
0602-52072-019
2023-53851-263
0526-52312-097
0341-51690-606
1282-52759-057
2112-53534-557
1376-53089-637
2606-54154-474
2110-53467-499
1464-53091-370
1801-54156-583
2094-53851-487
1043-52465-308
1803-54152-448
0854-52373-514
Z
0.00864
0.00061
0.00428
0.00234
0.09423
0.00100
0.00071
0.00072
0.00157
0.00192
0.00418
0.00697
0.04855
0.08778
0.04190
0.00198
0.00350
0.02377
0.02782
0.00797
0.02766
0.03595
0.00450
0.11183
0.01620
0.03678
0.00288
0.18400
0.02246
0.01637
0.01462
0.02797
0.09425
0.01613
0.01170
0.14730
0.00320
0.00594
0.05450
0.03041
α(J2000)
+12 16 47.89
+12 17 49.31
+12 22 25.79
+12 25 05.41
+12 26 11.90
+12 26 15.69
+12 28 09.26
+12 28 13.86
+12 30 28.33
+12 30 38.45
+12 30 48.60
+12 31 54.67
+12 38 29.93
+12 40 49.89
+12 41 34.25
+12 43 56.70
+12 45 16.87
+12 48 46.36
+12 53 06.56
+12 54 23.74
+13 02 49.20
+13 03 54.44
+13 04 32.27
+13 04 45.63
+13 06 24.19
+13 14 26.56
+13 14 47.37
+13 22 11.96
+13 23 47.46
+13 25 19.89
+13 25 49.42
+13 28 44.05
+13 29 16.56
+13 30 17.38
+13 31 26.91
+13 39 24.24
+13 41 56.48
+13 42 51.85
+13 44 24.06
+13 45 31.50
δ
+08 02 56.28
+37 51 55.50
+04 34 04.77
+61 09 11.29
+04 15 36.06
+48 29 38.43
+44 05 08.02
+44 07 10.43
+41 41 22.07
+41 39 11.31
+12 02 42.82
+15 07 36.48
+19 59 21.36
+66 24 20.17
+44 26 39.24
+32 10 14.67
+27 07 30.78
+47 42 53.45
+36 49 11.41
+58 53 40.67
+65 34 49.27
+37 14 01.88
-03 33 22.12
+62 24 20.88
+35 13 43.04
+63 33 11.37
+34 52 59.81
+01 30 34.39
-01 32 51.95
+48 02 26.15
+33 03 54.38
+43 55 50.51
+17 00 21.00
+31 19 58.02
+41 51 48.29
+07 39 27.61
+30 31 09.62
+52 42 30.57
+07 45 00.08
+04 42 32.71
54
ID
112
147
198
63
197
106
93
94
104
105
111
128
192
26
95
142
167
107
144
175
39
152
9
55
151
38
153
29
10
79
157
96
190
156
108
131
155
71
132
57
log He ii /Hβ
-1.81
-2.28
-2.28
-1.97
-2.00
-2.27
-1.67
-2.50
-2.33
-2.14
-1.44
-1.82
-1.90
-1.88
-0.19
-1.77
-1.83
-2.14
-1.85
-2.03
-1.92
-2.11
-2.26
-1.84
-2.08
-1.62
-2.37
-2.04
-1.88
-1.93
-2.05
-2.16
-1.92
-2.14
-2.13
-1.80
-2.19
-1.79
-1.78
-1.97
log[N ii] /Hα
-1.44
-1.72
-1.33
-1.96
-1.66
-2.11
-1.09
-1.95
-1.23
-1.31
-2.04
-0.68
-0.99
-1.49
-0.93
-1.88
-1.15
-1.48
-0.52
-1.15
-1.57
-1.39
-1.36
-0.95
-1.71
-1.15
-1.45
-1.00
-2.48
-1.64
-1.57
-1.28
-1.13
-1.41
-2.12
-0.82
-1.73
-1.53
-1.00
-1.55
Table 2.5: continued.
12 + log O/H
7.99±0.05
7.92±0.09
8.10±0.11
7.96±0.07
7.96±0.06
7.80±0.04
8.22±0.75
7.63±0.06
8.29±0.43
8.07±0.08
7.73±0.07
9.18±0.10
8.93±0.04
7.91±0.07
8.80±0.08
7.93±0.08
8.15±0.19
8.13±0.06
9.27±0.07
8.23±0.62
8.04±0.06
8.54±0.11
8.10±0.09
8.94±0.04
7.87±0.08
8.87±0.08
8.22±0.20
8.89±0.05
7.80±0.03
7.94±0.08
8.15±0.16
8.84±0.06
8.80±0.09
8.11±0.14
7.79±0.04
8.91±0.04
7.98±0.07
7.99±0.06
8.93±0.04
7.92±0.04
Features
Non-WR
Non-WR
Non-WR
WR
Non-WR
Non-WR
WR
WR
WR
WR
WR
WR
WR
Non-WR
Non-WR
WR
WR
WR
WR
WR
WR
Non-WR
WR
WR
Non-WR
WR
WR
WR
Non-WR
Non-WR
WR
WR
WR
Non-WR
Non-WR
WR
WR
WR
WR
WR
Class
0
0
0
3
0
0
1
3
3
3
1
3
2
0
0
3
3
2
3
3
1
0
3
2
0
3
3
2
0
0
2
2
3
0
0
1
2
1
2
1
Other Names
VCC 0207
N/A
N/A
SBS 1222+614
N/A
UGCA 281
NGC 4449
NGC 4449
NGC 4485
NGC 4490
N/A
IC 0797
N/A
SHOC 379
N/A
NGC 4656
NGC 4670
N/A
NGC 4774
N/A
N/A
N/A
UGCA 322
N/A
N/A
N/A
UGC 08323
F13196+0146
UM 570
SBS 1323+483
WAS 69
MRK 0259
N/A
UGC 08496
N/A
N/A
Mrk 0067c3
MRK 1480
N/A
TOLOLO 1343+049
Star-forming galaxies with nebular He II 4686 emission
Plate–MJD–FiberID
1776-53858-632
1158-52668-062
2770-54510-583
1324-53088-271
1378-53061-023
1323-52797-002
1323-52797-014
1325-52762-356
1325-52762-350
1642-53115-155
1324-53088-234
1325-52762-412
2786-54540-084
2746-54232-104
1381-53089-470
1644-53144-564
1827-53531-503
0305-51613-604
2137-54206-310
1709-53533-215
0920-52411-575
1646-53498-616
1383-53116-110
2145-54212-388
1843-53816-087
1844-54138-311
1399-53172-299
2911-54631-344
1651-53442-255
1679-53149-384
2163-53823-546
0616-52442-364
1725-54266-068
2167-53889-071
2524-54568-146
2527-54569-147
0624-52377-361
0364-52000-187
0624-52377-092
1570-53149-021
Z
0.02156
0.03383
0.02775
0.00058
0.00442
0.00056
0.00087
0.00085
0.00093
0.01194
0.00097
0.07731
0.00855
0.00771
0.02230
0.08624
0.17350
0.01341
0.01502
0.00466
0.02740
0.00267
0.00396
0.08403
0.00942
0.00610
0.03255
0.04684
0.06816
0.00826
0.03395
0.00225
0.03772
0.01098
0.10704
0.01217
0.00237
0.03133
0.02993
0.00910
α(J2000)
+13 46 49.45
+13 59 50.92
+14 01 07.12
+14 02 28.23
+14 02 36.07
+14 03 01.17
+14 03 34.06
+14 03 39.84
+14 04 11.24
+14 04 14.87
+14 04 28.63
+14 09 56.76
+14 18 51.13
+14 23 48.53
+14 26 28.17
+14 28 05.51
+14 29 47.01
+14 30 53.51
+14 31 08.88
+14 32 48.36
+14 48 05.38
+14 48 52.02
+14 50 56.56
+14 51 33.55
+14 54 12.15
+14 56 36.63
+15 09 34.18
+15 19 47.15
+15 23 32.19
+15 26 30.31
+15 34 56.40
+15 37 04.18
+15 45 43.55
+15 46 58.88
+16 06 27.54
+16 15 17.02
+16 16 23.54
+16 24 10.11
+16 26 04.26
+16 47 10.66
δ
+14 24 01.68
+57 26 22.98
+21 14 34.60
+54 16 33.08
+39 13 13.28
+54 14 29.40
+54 18 36.91
+54 18 56.87
+54 25 18.67
+36 43 32.67
+54 23 52.80
+54 56 48.89
+21 02 39.74
+14 38 16.54
+38 22 58.67
+36 27 10.40
+06 43 34.97
+00 27 46.35
+27 14 12.29
+09 52 57.15
-01 10 57.72
+34 42 42.99
+35 34 19.59
+26 46 03.56
+30 12 36.25
+30 13 52.36
+37 31 46.11
+39 45 37.85
+29 31 12.08
+41 17 22.34
+24 51 39.24
+55 15 50.62
+08 58 01.35
+17 53 03.07
+13 55 47.88
+13 01 33.08
+47 02 02.32
-00 22 02.58
+46 22 05.79
+21 05 14.51
ID
129
75
195
86
97
83
84
88
87
113
85
89
196
194
98
114
133
6
158
119
61
115
99
159
134
135
100
199
116
118
161
40
120
162
184
185
42
12
41
109
log He ii /Hβ
-1.75
-1.89
-2.14
-2.12
-1.92
-2.59
-2.25
-2.29
-2.40
-2.14
-2.41
-2.00
-1.97
-2.04
-1.87
-1.93
-1.80
-2.28
-2.15
-2.31
-2.02
-1.48
-2.24
-1.51
-2.38
-2.31
-2.01
-1.98
-1.96
-1.83
-1.68
-2.46
-1.69
-1.72
-1.99
-1.96
-1.56
-2.11
-1.36
-1.83
log[N ii] /Hα
-0.42
-1.42
-1.20
-0.89
-1.46
-1.75
-0.86
-0.67
-1.48
-0.88
-1.30
-1.32
-2.24
-0.48
-1.96
-1.60
-1.17
-1.67
-1.20
-1.42
-1.73
-1.66
-1.41
-0.67
-1.52
-1.50
-2.09
-0.63
-1.79
-1.04
-0.64
-1.33
-2.01
-0.40
-1.28
-1.57
-1.91
-1.58
-0.39
-2.06
Table 2.5: continued.
12 + log O/H
8.92±0.32
8.03±0.08
8.86±0.04
8.58±0.24
8.17±0.16
7.78±0.07
8.57±0.30
8.75±0.23
7.99±0.06
8.54±0.29
8.33±0.32
8.34±0.05
7.76±0.04
9.04±0.33
7.77±0.07
7.88±0.04
8.79±0.04
8.02±0.11
8.44±0.31
8.06±0.10
7.98±0.07
7.91±0.09
8.05±0.10
9.00±0.04
8.14±0.16
8.01±0.08
7.81±0.04
9.08±0.12
7.88±0.04
8.42±0.38
9.05±0.10
8.54±0.04
7.72±0.05
9.08±0.14
8.76±0.06
7.96±0.07
7.87±0.08
7.99±0.05
9.25±0.03
7.74±0.07
Features
WR
Non-WR
WR
WR
WR
WR
WR
WR
Non-WR
WR
Non-WR
WR
Non-WR
WR
Non-WR
Non-WR
Non-WR
WR
Non-WR
WR
WR
Non-WR
WR
WR
Non-WR
WR
Non-WR
WR
Non-WR
WR
WR
Non-WR
Non-WR
WR
WR
Non-WR
WR
WR
WR
Non-WR
Class
2
0
1
3
2
3
3
3
0
3
0
3
0
3
0
0
0
1
0
2
3
0
1
1
0
2
0
2
0
3
2
0
0
3
1
0
3
3
2
0
Other Names
MRK 0796
MRK 1486
UGC 08929
NGC 5447
N/A
N/A
NGC 5461
NGC 5461
N/A
MRK 1369
N/A
SBS 1408+551A
N/A
N/A
N/A
N/A
N/A
N/A
MRK 0685
NGC 5669
SHOC 486
UGC 09540
N/A
N/A
N/A
N/A
N/A
N/A
N/A
UGC 09856
N/A
N/A
N/A
NGC 5996
N/A
N/A
Arp 2
SHOC 536
N/A
N/A
Conclusion
55
Plate–MJD–FiberID
1342-52793-112
0976-52413-600
0978-52441-118
0358-51818-504
1115-52914-309
0673-52162-312
1893-53239-476
0742-52263-179
0677-52606-533
0650-52143-330
Z
0.03205
0.01195
0.01483
0.04723
0.01381
0.06669
0.02061
0.03029
0.03312
0.03592
α(J2000)
+16 49 05.27
+17 12 36.63
+17 18 53.45
+17 35 01.25
+20 47 59.21
+22 25 10.13
+22 38 31.12
+23 01 23.59
+23 02 10.00
+23 56 21.96
δ
+29 45 31.61
+32 16 33.42
+30 11 36.20
+57 03 08.55
-00 10 53.98
-00 11 52.84
+14 00 29.78
+13 33 14.79
+00 49 38.84
-09 04 07.42
56
ID
90
67
68
11
74
46
137
50
47
43
log He ii /Hβ
-1.52
-1.99
-1.92
-2.17
-2.18
-1.85
-2.16
-2.06
-1.83
-1.51
log[N ii] /Hα
-0.39
-1.96
-0.84
-1.31
-1.17
-1.81
-2.22
-1.68
-2.21
-1.43
Table 2.5: continued.
12 + log O/H
9.30±0.03
7.79±0.06
8.60±0.27
8.31±0.04
8.44±0.34
7.90±0.04
7.79±0.03
7.90±0.04
7.75±0.04
8.00±0.07
Features
WR
WR
WR
WR
WR
Non-WR
Non-WR
Non-WR
Non-WR
WR
Class
1
1
2
2
1
0
0
0
0
1
Other Names
KUG 1647+298
N/A
IRAS 17169+3014
SHOC 579
N/A
N/A
N/A
N/A
N/A
N/A
Star-forming galaxies with nebular He II 4686 emission
Appendix A: Fitting models to the emission lines
2.8
Appendix A: Fitting models to the emission
lines
The CL01 model grid is calculated by varying the model parameters, U, ionization
parameter at the edge of the Strömgren sphere, τV , total V–band optical depth,
ξ, the dust-to-metal ratio of ionized gas, µ, the fraction of the total optical depth
in the neutral ISM contributed by the ambient ISM, and Z, the metallicity, over
a certain range for 221 unequally spaced time steps from t = 0 to t = 20Gyr (see
Table 2.4).
In this paper we use the interpolated model grid of various luminosities for 50
time-steps from B04. This includes a total of 2 × 105 different models. To fit to the
data we adopt the Bayesian methodology described by Ka03. We obtain the PDF
of every parameter of interest by marginalisation over all other parameters. The
resulting PDF is used to estimate confidence intervals for each estimated physical
parameter. We need to fit to at least five strong emission lines, [O ii] λ3727, 3729,
Hβ, [O iii] λ5007, Hα, [N ii] λ6584 to get a good constraint on the parameters.
We take the median value of each parameter to be the best estimate of a given
parameter.
In Figure 2.19 we illustrate our technique by showing the effect of adding lines
on the PDFs of parameters when we fit a model to the data. We start with
[O ii] λ3727, 3729 and show how we get more well defined PDFs for the indicated
parameters as we add the emission lines indicated on the left. We show the PDFs
for dust attenuation parameter in V -band, gas phase oxygen abundance, ionization
parameter, dust-to-metal ratio of ionized gas and the conversion factor from Hα
and [O ii] luminosity to star formation rate (see CL01 for further details), for one
object in our sample. The dust-to-metal ratio, ξ, is hard to constrain, except at
high metallicity.
57
Star-forming galaxies with nebular He II 4686 emission
Figure 2.19 We show how the PDFs for dust attenuation parameter in V-band,
gas phase oxygen abundance, ionization parameter of ionized gas, dust-to-metal
ratio and Hα and [O ii] efficiency factors change when we add more emission lines
to the fit. ξ is hard to constrain, except at high metallicity.
58
REFERENCES
References
Abazajian K. N., et al., 2009, ApJS, 182, 543
Allen, M. G., Groves, B. A., et al., 2008, ApJS, 178, 20
Anders, E., Grevesse, N., 1989, GeCoA, 53, 197
Baldwin, J. A., Phillips, M. M., & Terlevich, R. 1981, PASP, 93, 5
Bergeron, P., Ruiz, M. T., Leggett, S. K., 1997, ApJS, 108, 339
Binette, L., Magris, C., et al., 1994, A&A, 292, 13
Brinchmann, J., Charlot, S., et al. 2004, MNRAS, 351, 1151
Brinchmann, J. ,Kunth, D., Durret, F., 2008, A&A, 485, 657
Bruzual, G. ,Charlot, S., 2003, MNRAS, 344, 1000
Cantiello, M., Yoon, S.-C., et al., 2007, A&A, 465, 29
Charlot, S., Fall, S. M., 2000, ApJ, 539, 718
Charlot, S., Longhetti, M., 2001, MNRAS, 323, 887
Conti, P. S., Garmany, C. D., Massey, P., 1989, Apj, 341, 113
Crowther, P. A., Pasquali, A., et al., 1999, A&A, 350, 1007
Crowther, P. A., Dessart, L., Hillier, D. J., et al. 2002, A&A, 392, 653
Crowther, P. A., Hadfield, L. J.,2006, A&A, 449, 711
Crowther, P. A., 2007, ARA&A, 45, 177
Dopita, M. A. & Sutherland, R. S., 1996, ApJS, 102, 161
Dopita, M. A., Groves, B. A., et al., 2005, ApJ, 619, 755
Eldridge, J. J., Izzard, R. G., Tout, C. A., 2008, MNRAS, 384, 1109
Eldridge, J. J., Stanway, E. R., 2009, MNRAS, 400, 1019
Eldridge, J. J., Langer N., Tout, C. A., 2011, MNRAS, 414, 3501
Eldridge, J. J., & Stanway, E. R. 2011, MNRAS, 1595 , in press
Ferland, G. J., Korista, K. T., et al., 1998, PASP, 110, 761
Gallazzi, A., Charlot, S., et al., 2005, MNRAS, 362, 41
Gallazzi, A., Brinchmann, J., et al., 2008, MNRAS, 383, 1439
Garnett, D. R., Kennicutt, R. C., et al., 1991, ApJ, 373, 458
Gräfener G., Vink J. S., de Koter A., Langer N., 2011, A&A, 535, A56
Guseva, N. G., Izotov, Y. I., Thuan, T. X., 2000, ApJ, 531, 776
Hadfield, L. J., Crowther, P. A., 2007, MNRAS, 381, 418
Hillier, D. J. & Miller, D. L., 1998, ApJ, 496, 407
Hoare, M. G., Drew, J. E., Denby, M., 1993, MNRAS, 262, 19
Izotov, Y. I., StasiÅĎska, G., et al., 2006, A&A, 448, 955
Izotov, Y. I., Schaerer, D., Blecha, A., et al. 2006, A&A, 459, 71
Kauffmann, G., Heckman, T.M., et al. 2003, MNRAS, 346, 1055
Kehrig, C., Vı́lchez, J. M., et al., D., 2008, A&A, 477, 813
Kehrig, C., Oey, M. S., et al., 2011, A&A, 526, 128
Kewley, L. J., Dopita, M. A., et al., 2001, ApJ, 556, 121
Kewley, L. J., Groves, B., et al., 2006, MNRAS, 372, 961
Kewley, L. J. & Ellison, S. L. 2008, ApJ, 681, 1183
Kniazev, A. Y., Pustilnik, S. A., et al., 2004, ApJS, 153, 429
Legrand, F., Kunth, D., Roy, et al. 1997, A&A, 326, L17
Leitherer, C., Schaerer, D., et al. 1999, ApJS, 123, 3
Leitherer, C., Ortiz Otálvaro, P. A., et al. 2010, APJS, 189, 309
59
REFERENCES
López-Sánchez, Á. R., & Esteban, C. 2010, A&A, 516, A104
Maeder, A., 1987, A&A, 178, 159
Meynet, G., Maeder, A., 2007, A&A, 464, 11
Monreal-Ibero, A., Vı́lchez, J. M., et al., 2010, A&A, 517, A27
Neugent, K. F., Massey, P., 2011, ApJ, 733, 123
Pauldrach, A. W. A., Hoffmann, T. L., Lennon, M., 2001, A&A, 375, 161
Pettini, M. & Pagel, B. E. J. 2004, MNRAS, 348, 59
Schaerer, D. 1996, ApJ, 467, 17
Schaerer, D., Vacca, W. D., 1998, ApJ, 497, 618
Smith, L. J., Norris, R. P. F., Crowther, P. A., 2002, MNRAS, 337, 1309
Stasińska, G., & Leitherer, C., 1996, ApJ, 107, 661
Stasińska, G., Cid Fernandes, R., et al., 2006, MNRAS, 371, 972
Terlevich, R., Melnick, J., et al., 1991, A&AS, 91, 285
Thuan, T. X., Izotov, Y. I., 2005, ApJS, 161, 240
Tremonti, C. A., Heckman, T. M., et al. 2004, ApJ, 613, 898
Yoon, S.-C., Langer, N., 2005, A&A, 443, 643
Yoon, S.-C., Langer, N., Norman, C., 2006, A&A, 460, 199
York, D. G. et al. 2000, AJ, 120, 1579
60
Chapter
3
The physical nature of the 8
o’clock arc based on near-IR
IFU spectroscopy with SINFONI
We present an analysis of near-infrared integral field unit spectroscopy for the
gravitationally lensed Lyman break galaxy, the 8 o’clock arc, taken with SINFONI
on the Very Large Telescope. We explore the shape of the spatially-resolved Hβ
profile and demonstrate that we can decompose it into three main components
that partially overlap (spatially) but are distinguishable when we include the dynamical information. To study the de-lensed morphology of the galaxy we make
use of existing B & H imaging from the Hubble Space Telescope and construct
a rigorous lens model using a Bayesian grid based lens modeling technique. We
apply this lens model to the SINFONI data cube to construct the de-lensed Hβ
line map and the velocity and velocity dispersion maps of the galaxy. We explore
the dynamical state of the galaxy and find that the 8 o’clock arc has a complex
velocity field that is not simply explained by a single rotating disk. The Hβ profile
of the galaxy shows a blue-shifted wing suggesting gas outflows of ∼ 200 km s−1 .
We confirm previous findings that the 8 o’clock arc lies on the stellar mass–oxygen
abundance–star formation rate plane found locally, but it has nevertheless significantly different interstellar medium properties. We show that the gas surface
density of the 8 o’clock arc is a factor of 2–4 higher than similar low-redshift
galaxies selected from the Sloan Digital Sky Survey. We also find that the electron
density in the ionized gas is approximately five times higher than in the comparison sample, which implies a higher H ii-region pressure as well as likely a higher
density interstellar medium than in similar nearby galaxies.
Maryam Shirazi, Simona Vegetti, Nicole Nesvadba, Sahar Allam, Jarle
Brinchmann, Douglas Tucker
Monthly Notices of the Royal Astronomical Society
2013, submitted
SINFONI observations of the 8 o’clock arc
3.1
Introduction
The last decade has seen a dramatic increase in our knowledge of the galaxy
population at redshift z > 2. In particular, the large samples of high redshift
(high-z) galaxies identified by the Lyman-break dropout technique (Steidel et al.,
2003, and references therein) have allowed detailed statistical analysis of the
physical properties of these galaxies (Shapley, 2011). While early studies made
use of long-slit near-IR spectroscopy (Erb et al., 2006a,b,c) to study the physical
properties of these galaxies, more recent studies have focused on near-IR integral
field units (IFUs) (Förster Schreiber et al., 2006, 2009; Genzel et al., 2008, 2010).
The steadily growing effort to obtain resolved near-IR spectra of high-z galaxies in a systematic manner as in the MASSIV, SINS, SINS/zC-SINF and LSD/
AMAZE surveys is leading to samples of spatially-resolved emission line maps of
distant (z ∼ 1 − 3.8) star-forming galaxies. Studying these maps has provided us
with spatially-resolved physical properties, metallicity gradients and kinematics of
high-z star-forming galaxies (e.g., Contini et al., 2012; Epinat et al., 2012; Förster
Schreiber et al., 2009; Cresci et al., 2009; Genzel et al., 2011; Förster Schreiber et
al., 2011a,b; Newman et al., 2012b; Maiolino et al., 2008; Mannucci et al., 2009;
Gnerucci et al., 2011).
A particularly important question for these studies is whether the observed dynamics are due to, or significantly influenced by major mergers. While this is generally difficult to establish, Genzel et al. (2006) have shown that with sufficiently
high resolution integral field unit (IFU) spectroscopy, it is possible to distinguish
between rotation and merging. However, variations in spatial resolution still cause
inconclusive interpretations. As an example, using SINFONI observations of 14
LBGs Förster Schreiber et al. (2006) argued for rotationally supported dynamics in many LBGs (7 out of 9 resolved velocity fields). In contrast, by studying
spatially resolved spectra of 3 galaxies at redshift z ∼ 2 − 3, using the OSIRIS in
combination with adaptive optics (AO), Law et al. (2007) showed that the ionized
gas kinematics of those galaxies were inconsistent with simple rotational support.
Analysis of the SINS Hα sample studied by Förster Schreiber et al. (2009)
showed that about one-third of 62 galaxies in their sample show rotation dominated
kinematics, another one-third are dispersion-dominated objects, and the remaining
galaxies are interacting or merging systems. However, more recent AO data have
shown that many of these dispersion-dominated sources are in fact rotating and
follow the same scaling relations as more massive galaxies (Newman et al., 2013).
They also show that the ratio of rotation to random motions (V/σ) increases with
stellar mass. This result shows the importance of spatial resolution for studying
high-z galaxies.
While we are essentially limited by the intrinsic faintness of these objects,
gravitational lensing can significantly magnify these galaxies and allow us to study
their properties at a level similar to what is achieved at lower redshifts (e.g., MS
1512-cB58; see Yee et al., 1996; Pettini et al., 2000, 2002; Teplitz et al., 2000;
Savaglio et al., 2002; Siana et al., 2008). Although near-IR IFU observations with
AO have been able to spatially resolve high-z galaxies (Förster Schreiber et al.,
2006, 2009), obtaining a resolution better than 0.2” even with AO is very difficult
62
Introduction
and lensing is the only way to obtain sub-kpc scale resolution for high-z galaxies
using current instruments. Studies of this nature will truly come into their own in
the future with 30m-class telescopes.
Given a sufficiently strongly lensed Lyman break galaxy (LBG), we might be
able to study its dynamical state, the influence of any potential non-thermal ionizing source, such as a faint active galactic nucleus (AGN), and the physical properties of the interstellar medium.
Spatially-resolved studies of six strongly lensed star-forming galaxies at z ∼
1.7 − 3.1 using the Keck laser guide star AO system and the OSIRIS IFU spectrograph enabled Jones et al. (2010) to resolve the kinematics of these galaxies on
sub-kpc scales. Four of these six galaxies display coherent velocity fields consistent
with a simple rotating disk model. Using the same instrument, Jones et al. (2010)
also studied spatially-resolved spectroscopy of the Clone arc in detail. Deriving a
steep metallicity gradient for this lensed galaxy at z = 2, they suggested an insideout assembly history with radial mixing and enrichment from star formation. A
detailed study of the spatially-resolved kinematics for a highly amplified galaxy
at z = 4.92 by Swinbank et al. (2009) suggests that this young galaxy is undergoing its first major epoch of mass assembly. Furthermore, analyzing near-infrared
spectroscopy for a sample of 28 gravitationally-lensed star-forming galaxies in the
redshift range 1.5 < z < 5, observed mostly with the Keck II telescope, Richard et
al. (2011) provided us with properties of a representative sample of low luminosity
galaxies at high-z.
The small number of bright z ∼ 2 lensed galaxies has recently been increased
by a spectroscopic campaign following-up galaxy-galaxy lens candidates within
the Sloan Digital Sky Survey (SDSS) (Stark et al., 2013). These high spatial and
spectral resolution data, will provide us with constraints on the outflow, metallicity
gradients, and stellar populations in high-z galaxies.
Given its interesting configuration and brightness, the 8 o’clock arc (Allam et
al., 2007) is of major interest for the detailed investigation of the physical and
kinematical properties of LBGs. Indeed, there has been a vigorous campaign to
obtain a significant collection of data for this object. In particular, the following
observations have been made: 5-band HST imaging covering F450W to F160W,
a Keck LRIS spectrum of the rest-frame UV, near-IR H- and K-band long-slit
spectroscopy with the Near InfraRed Imager and Spectrometer (NIRI) on the
Gemini North 8m telescope (Finkelstein et al., 2009) and X-shooter observations
with the UV-B, VIS-R and NIR spectrograph arms (Dessauges-Zavadsky et al.,
2010; Dessauges-Zavadsky et al., 2011, hereafter DZ10 and DZ11).
Measuring the differences between the redshift of stellar photospheric lines and
ISM absorption lines, Finkelstein et al. (2009) suggested gas outflows on the order
of 160 km s−1 for this galaxy. DZ10 also showed that the ISM lines are extended
over a large velocity range up to ∼ 800 km s−1 relative to the systematic redshift.
They showed that the peak optical depth of the ISM lines is blue-shifted relative
to the stellar photospheric lines, implying gas outflows of 120 km s−1 .
Studying the rest frame UV, DZ10 showed that the Lyα line is dominated
by a damped absorption profile with a weak emission profile redshifted relative
to the ISM lines by about +690 km s−1 on top of the absorption profile. They
63
SINFONI observations of the 8 o’clock arc
suggested that this results from backscattered Lyα photons emitted in the HII
region surrounded by the cold, expanding ISM shell.
DZ11 argued that the 8 o’clock arc is formed of two major parts, the main
galaxy component and a smaller clump which is rotating around the main core of
the galaxy and separated by 1.2 kpc in projected distance. They found that the
properties of the clump resembles those of the high-z clumps studied by Swinbank
et al. (2009), Jones et al. (2010), and Genzel et al. (2011). They also suggested
that the fundamental relation between mass, SFR, and metallicity (Mannucci et
al., 2010; Lara-López et al., 2010) may hold up to and even beyond z = 2.5, as
also supported by two other lensed LBGs at 2.5 < z < 3.5 studied by Richard et
al. (2011).
In this work, we use near-IR IFU spectroscopy of the 8 o’clock arc with SINFONI to spatially resolve the emission line maps and the kinematics of this galaxy.
In Section 3.2, we introduce the observed data. In this section, we also discuss
the data reduction procedure and the PSF estimation as well as the SED fitting
procedure. The analysis of the IFU data is covered in Section 3.3. The physical
properties of the 8 o’clock arc are discussed in Section 3.4. In Section 3.5, we
introduce our lens modeling technique and also our source reconstruction procedure. In this section, we also present the emission line maps and the Hβ profile in
the source plane. The kinematics of the galaxy are discussed in Section 3.6. We
present our conclusions in Section 5.6.
3.2
3.2.1
Data
Near-IR spectroscopy with SINFONI
We obtained J, H and K band spectroscopy of the 8 o’clock arc (α(J2000) :
00 22 40.91 δ(J2000) : 14 31 10.40) using the integral-field spectrograph SINFONI
(Eisenhauer et al., 2003; Bonnet et al., 2004) on the VLT in September 2009 (Program ID: 83.A-0879 A). The observation was done in seeing-limited mode with the
0.125” pixel−1 scale, for which the total field of view (FOV) is 8” × 8”. The total
observing time was 4h for J, 5h for H and 3.5h for K with individual exposure
times of 600s.
3.2.2
Data Reduction
The SINFONI data were not reduced with the standard ESO pipeline, but with
a custom set of routines written by N. Nesvadba, which are optimized to observe
faint emission lines from high-z galaxies. These routines are very well tested on
SINFONI data cubes for more than 100 high-redshift galaxies, and have been used
to reduce the data presented, e.g., in Lehnert et al. (2009) and Nesvadba et al.
(2006a,b, 2007a,b).
The reduction package uses IRAF (Tody, 1993) standard tools for the reduction of long-slit spectra, modified to meet the special requirements of integral-field
spectroscopy, and is complemented by a dedicated set of IDL routines. Data
are dark frame subtracted and flat-fielded. The position of each slitlet is measured
64
Data
Filter
F450W
F814W
F160W
Band
B
I
H
A1
AB magnitude
21.89 ± 1.61
20.98 ± 1.03
19.35 ± 0.53
A2
AB magnitude
21.76 ± 1.55
21.06 ± 1.07
19.23 ± 0.5
A3
AB magnitude
21.02 ± 1.07
20.17 ± 0.7
18.46 ± 0.34
A4
AB magnitude
22.65 ± 1.27
21.71 ± 1.44
20.35 ± 0.79
Table 3.1 HST photometry of the 8 o’clock arc images A1-A4. The AB magnitudes
correspond to the total photometry of all components in each image.
from a set of standard SINFONI calibration data which measure the position of an
artificial point source. Rectification along the spectral dimension and wavelength
calibration are done before night sky subtraction to account for some spectral flexure between the frames. Curvature is measured and removed using an arc lamp,
before shifting the spectra to an absolute (vacuum) wavelength scale with reference to the OH lines in the data. To account for the variation of sky emission,
we masked the source in all frames and normalized the sky frames to the average of empty regions in the object frame separately for each wavelength before
sky subtraction. We corrected for residuals of the background subtraction and
uncertainties in the flux calibration by subsequently subtracting the (empty sky)
background separately from each wavelength plane.
The three-dimensional data are then reconstructed and spatially aligned using
the telescope offsets as recorded in the header within the same sequence of 6
dithered exposures (about one hour of exposure), and by cross-correlating the
line images from the combined data in each sequence, to eliminate relative offsets
between different sequences. A correction for telluric absorption is applied to each
individual cube before the cube combination. Flux calibration is carried out using
standard star observations taken every hour at a position and air mass similar to
those of the source.
3.2.3
HST Imaging
Optical and NIR imaging data of the 8 o’clock arc were taken with
the Wide Field Planetary Camera 2 (WFPC2) and the Near Infrared Camera and Multi-Object Spectrometer (NICMOS) instruments
on the Hubble Space Telescope (Proposal No.
11167, PI: Sahar Allam).
The 8 o’clock arc is clearly resolved, and was observed in the
five filters WFPC2/F450W, WFPC2/F606W, WFPC2/F814W, NIC2/
F110W, and NIC2/F160W, which we will refer to as B, V, I, J and H in
the following. Total exposure times of 4 × 1100 s per BV I band, 5120 s in the J
band, and 4 × 1280 s in the H band were obtained. The BV I frames, with a pixel
scale of 0.1”, were arranged in a four-point dither pattern, with random dithered
offsets between individual exposures of 1” in right ascension and declination. The
JH frames, with a pixel scale of 0.075”, were also arranged in a four-point dither
pattern, but with offsets between individual exposures of 2.5”. In order to resolve
the 8 o’clock arc better, the HS T images were drizzled to obtain a pixel scale of
0.05”. Figure 3.1 shows the B band HS T image of the 8 o’clock arc and defines
the images A1 through A4 as indicated.
65
SINFONI observations of the 8 o’clock arc
Figure 3.1 The B band HS T image (in counts per second) of the 8 o’clock arc
is shown. Three images A1-A3 form an arc and A4 is the counter image. The
foreground galaxy (lens) has been removed from this image. The scale-bar in this
and all following images are at the redshift of the source.
Figure 3.2 The I − H (rest frame NUV − B) color image of the arc is shown. We can
see that the substructures of the arc are better resolved in this image; for instance,
we can resolve better two lensed images of the same clump lying between the A3
and A2 images (see Figure 3.13, which marks the de-lensed image of this clump
with a purple ellipse). We can also resolve two images of another clump which are
between the A2 and A1 images (see Figure 3.13, which notes the de-lensed image
of this clump by a red ellipse).
We performed photometry using the Graphical Astronomy and Image Analysis
66
Analysis of the SINFONI data
Tools (GAIA1 ). Table 3.1 summarises the HS T photometry of the images A1-A4.
As an illustration of the power of the multi-wavelength HS T data set, we show
the I-H (rest frame NUV −B) color image of the arc in Figure 3.2. To construct this
we convolved the WFPC2/F814W image to the same PSF as the NICMOS/F160W
band before creating the color image. We can see that the substructures of the arc
are better resolved in this image; for instance, we can resolve two individual images
of the same clump that lie between the A3 and A2 images (see the de-lensed image
of the clump shown by a purple dashed ellipse in Figure 3.13).
3.2.4
PSF Estimation
We created model PSFs for the HS T images using the TinyTim package 2 (Krist
et al., 2011). A measure of the PSF was also obtained using a star in the field.
This estimate is consistent with TinyTim PSFs; however, because the star is significantly offset from the arc, in the rest of the paper we use only the the TinyTim
PSFs when analyzing the HS T data.
For the SINFONI data we use the standard star observations to estimate the
PSF. The standard star was observed at the end of each Observing Block (OB) at
a air mass similar to that of the data, in a fairly similar direction, and with the
same setup. We integrate the standard star cubes in each band along the spectral
axis to obtain the 2-dimensional images of the star.
We measure the FWHM size of the star along the x and y axes of the SINFONI
field-of-view by fitting a 2-dimensional Gaussian to the resulting image. We then
average the individual measurements of the standard star images in each direction
to determine the PSF for the corresponding band. The spatial resolutions in right
ascension and declination are always somewhat different for SINFONI data due to
the different projected size of a slitlet (0.25”) and a pixel (0.125”). The PSFs in
the J, H and K bands are [0.99”,0.7”], [0.8”,0.66”], [0.69”,0.51”], respectively.
3.3
3.3.1
Analysis of the SINFONI data
Nebular emission lines
The spectrum is first analysed using the platefit pipeline, initially developed for
the analysis of SDSS spectra (Tremonti et al 2004, Brinchmann et al., 2004, 2008)
and subsequently modified for high-z galaxies (e.g. Lamareille et al., 2006). The
nebular emission lines identified in the 8 o’clock arc images A2-A3 are summarized
in Table 3.3 . Specifically, the emission lines that we can detect in the spectra of
the galaxy are [O ii] λ3727, 29, Hδ, Hγ, He ii λ4686 and Hβ. Hβ is the strongest
detected emission line. In the following, we therefore concentrate on this line to
further study the dynamical properties of the galaxy.
Due to the redshift of the 8 o’clock arc, we can not study the [O iii]λ4959, 5007,
Hα and [N ii] λ6548, 6584 emission lines because they fall outside of the spectral
1 http://astro.dur.ac.uk/%7Epdraper/gaia/gaia.html
2 http://tinytim.stsci.edu/cgi-bin/tinytimweb.cgi
67
SINFONI observations of the 8 o’clock arc
range of the SINFONI bands. This means that we can not place strong constraints
on the ionisation parameter or the metallicity of the galaxy using the IFU data.
3.3.2
The integrated Hβ profile
The Balmer lines show asymmetric profiles, especially Hβ, as was also observed by
DZ11. We therefore start with a look at the integrated Hβ profile. This offers us,
among other things, the possibility of testing our reduction techniques because in
the absence of significant small-scale structure, profiles are expected to be similar
in shape in the different sub-images. We focus here on the spectra of the highest
magnification images, A2 and A3 (see Figure 3.1), and we only integrate over
the main galaxy structure, excluding the clump identified by DZ11. The counter
image (A4) is complete but is not resolved; the A1 image is only partially resolved
and is located near the edge of the data cube. Figure 3.3 shows the integrated Hβ
profiles for the images A2 and A3, respectively. We can see that the two images
show the same profile, which is what one expects as they are two images of the
same galaxy. This result differs from that of DZ11, who found different profiles
using their long-slit data. They suggested that this might be either due to the slit
orientation not optimally covering the lensed image A3, or alternatively, due to
the presence of substructure perturbing the surface brightness of the A2 image.
Since the IFU data show the same profile for both images, we can rule out the
possibility that substructure might have caused the differences.
We can see that the integrated Hβ profiles of both images show one main
component with a broad blue wing; thus, the full profile requires a second Gaussian
to be well fitted. The width of the Gaussian components for both images is 1.7±0.7
Å, which gives a velocity dispersion of 104 ± 42 km s−1 . The velocity offset between
the two fitted Gaussian components which are shown by the red dashed curves is
278 ± 63.5 km s−1 for the A3 image and 191 ± 63 km s−1 for the A2 image which
are consistent within the errors. We can clearly see this blue-shifted component in
both images in Figure 3.5 and Figure 3.6. DZ11 fitted two individual Gaussians to
the main component of the galaxy and concluded that these fits are related to the
two components (main and clump) with velocity offset of ∼ 61 ± 8 km s−1 . Since we
have resolved the clump using our IFU data and did not include it when integrating
the Hβ profile in Figure 3.3, the spectra of the A2 and A3 images plotted in Figure
3.3 do not contain any contribution from the clump. To illustrate the Hβ profile
of the clump, we add the spectra of the two images of the clump and show the
total profile with an orange line in both panels in Figure 3.3. We measure a
velocity offset of 126 ± 42 km s−1 between the clump and the main component of
the galaxy. The second component seen for both images (the left Gaussian fits in
Figure 3.3) is coming from the part of the arc that was not covered by the slit
used by DZ11. From the lens modeling described in Section 3.5, we know that the
spatially separated blue-shifted component in the A2 image is coming from the
north-east part of the galaxy (see Figure 3.13). However, we see from Figure 3.6
that this blue-shifted component of the A3 image is not separated spatially from
the main component of the galaxy. The difference between the two images might
be due to the fact that the data have insufficient spatial resolution to resolve the
68
Analysis of the SINFONI data
Figure 3.3 Left: integrated Hβ profile of A2, the highest magnification image,
showing two components that we can also resolve in individual spaxels. Middle:
integrated Hβ profile of the A3 image. Both images show the same Hβ profile (see
right panel). The widths of the Gaussian components for both images (red dashed
lines) are 1.7 Å, which gives velocity dispersions of 104 km s−1 . The velocity offset
between the two fitted Gaussian components is 278 km s−1 for the A3 image and
191 km s−1 for the A2 image. To illustrate the Hβ profile of the clump, we add the
spectra of two images of the clump and show this profile in orange in both panels
(note that the A2 and A3 profiles shown in this figure do not contain the clump
profile). The velocity offset between the clump Hβ profile and the main galaxy Hβ
profile is 126 km s−1 . The bottom panels show the residuals if we fit the Hβ profile
of each image with a single Gaussian. Right: Hβ profiles of the A2 and A3 images
and the clump are shown; the profiles are normalised to have the same peak.
components in the A3 image.
3.3.3
Spatially-resolved emission-line properties of the 8
o’clock arc in the image plane
As we saw above, the integrated Hβ profile is not well fitted by a single Gaussian,
and this is also true for Hγ and can also be seen in individual spatial pixels (spaxels)
for Hβ. We therefore fit these lines with two or three Gaussian components when
necessary. We carry out these fits to the Balmer lines using the MPFIT package
in IDL3 . During the fitting, we require the lines to have the same velocity widths.
This could be an incorrect approximation in detail but it leads to good fits to
the line profiles; the S/N and spectral resolution of the data are not sufficient
to leave the widths freely variable. The spatially-resolved Hβ profiles generally
show a main component, which we place at a systemic redshift of 2.7363 ± 0.0004
(rest-frame wavelength, λair = 4861.325) and an additional component that is blue3 http://cow.physics.wisc.edu/~craigm/idl/mpfittut.html
69
SINFONI observations of the 8 o’clock arc
shifted relative to the main component by 120-300 km s−1 . The best-fit Gaussian
intensity map of these blue-shifted and main components of the galaxy are shown in
Appendix A for the A2 and A3 images. There is also a redshifted component that
is detectable close to the clumps between the A3 and A2 images (see Figure 3.1).
This component is spatially separated from the main component by 1” (mentioned
also by DZ11). The velocity difference between this component and the main
component is ∼ 120 km s−1 .
The central map in Figure 3.4 shows the spatial distribution of Hβ line flux
across the main components of the arc, where we have integrated the line flux
between λ = 4855Å and 4867Å. The small panels around the Hβ line map show
the Hβ profiles of different spatial pixels as indicated. These individual panels
clearly show that the Hβ line shows different profiles in different regions across the
lensed images.
We can show these components in an alternative way, using the positionvelocity diagrams in Figure 3.5 and Figure 3.6 for the A2 and A3 images, respectively. Figure 3.5 clearly shows two spatially separated components corresponding
to the A2 image. The peak of one component is blue-shifted by ∼130 km s−1 and
spatially separated by ∼ 1” relative to the peak of the other. DZ11 identified
these two components with the main galaxy and the clump because they could
not separate the clump from the rest of the galaxy using long-slit observations.
Here, using IFU data, we have excluded the clump from these position-velocity
diagrams. The two retained components are associated with the galaxy and the
red (in the spectral direction) component that DZ11 identified as the clump is
part of the main galaxy. From the lens modeling that we describe in Section 3.5,
we will see that the blue-shifted component comes from the eastern part of the
galaxy (see Figure 3.13). The A3 image in Figure 3.6 also shows this blue-shifted
component but not as spatially separated. We will argue below (see Section 3.6.3)
that a reasonable interpretation of this component is that it corresponds to an
outflow from the galaxy.
3.4
3.4.1
The physical properties of the 8 o’clock arc
SED fitting
To determine the physical parameters of the 8 o’clock arc, we fit a large grid of
stochastic models to the HS T BV I JH photometry to constrain the spectral energy
distribution (SED). The grid contains pre-calculated spectra for a set of 100,000
different star formation histories using the Bruzual & Charlot (2003, BC03) population synthesis models, following the precepts of Gallazzi et al. (2005, 2008). Figure 3.7 shows the best-fit SED. We corrected the observed magnitudes for galactic
reddening. We corrected the photometry for Galactic foreground dust extinction
using E(B − V)Gal = 0.056 (Schlegel et al., 1998).
We follow the Bayesian approach presented by Kauffmann et al. (2003) to
calculate the likelihood of the physical parameters. We take the median values
of the Probability Distribution Functions (PDFs) as our best estimated values.
In particular, the parameters we extract are the stellar mass, M? , the current
70
The physical properties of the 8 o’clock arc
Figure 3.4 Middle panel shows Hβ line map of the arc. Small panels around the Hβ
line map show the Hβ profiles of different spatial pixels as indicated. The Hβ line
map was integrated over 4855Å< λrest <4867Å. We can see that Hβ shows different
profiles at different pixels, which are composed of multiple components.
star-formation rate, SFRSED , the dust attenuation, τV and the r-band luminosity
weighted age. The physical parameters from the SED fitting are summarised in
Table 3.2.
DZ11 also carried out SED fitting to the photometric data for the 8 o’clock arc.
They explored cases with and without nebular emission (Schaerer & de Barros,
2009, 2010). Since their results do not change significantly, we do not consider
the effect of nebular emission in our study. They also included photometry from
IRAC, which in principle should improve constraints on stellar masses. We have
opted against using these data, keeping the higher spatial resolution of HS T +
SINFONI, as the stellar mass from our fits is only slightly higher, but consistent
with their results within the errors, and this is the quantity of most interest to
this paper.
To compare our SED fit to the observed continuum spectrum, we estimate the
continuum in the SINFONI spectra by taking the median of the spectrum in bins
of 10 Å. Figure 3.8 shows the H-band median continuum, summed over all images
in the arc, in comparison to the estimated model continuum from the SED fitting.
The agreement is satisfactory, although the S/N of the continuum precludes a
detailed comparison.
71
SINFONI observations of the 8 o’clock arc
Figure 3.5 Position-velocity diagram of the A2 image, with negative values corresponding to blue shift. There are two components that are clearly offset both
spatially and in the velocity direction. The velocity offset between the two components is ∼ 130 km s−1 , and the spatial separation between them is ∼ 1”. Position
is relative to the center of the A2 image along the length of the arc.
Figure 3.6 Position-velocity diagram of the A3 image. This image shows a blueshifted component that is not spatially separated, in contrast to Fig. 3.5. Position
is relative to the center of the A3 image along the length of the arc.
72
The physical properties of the 8 o’clock arc
Figure 3.7 HS T photometry for the 8 o’clock arc in the rest-frame (cyan filled
circles) with the best-fit SED (solid curve) over plotted. The horizontal error bars
show the wavelength coverage of the HS T filters.
Image
log (M? /M )
log(S FR/M yr−1 )
log(sS FR/yr−1 )
log(age/yr)
τV
log(Z/Z )
A2
10.249.99
10.69
1.861.71
1.96
−8.47−8.98
−8.17
8.327.68
8.93
0.170.07
0.3
0.08−0.19
0.2
A3
10.3210.08
10.77
1.961.82
2.07
−8.44−8.93
−8.15
8.287.67
8.90
0.170.07
0.3
0.07−0.2
0.2
Table 3.2 Physical parameters derived from SED fitting. Parameters from left to
right are stellar mass, M? , current star-formation rate, SFR, specific star formation
rate (sSFR), r-band luminosity weighted age, dust attenuation τV and metallicity.
3.4.2
Parameters derived from emission line modeling
To derive physical parameters for the ionised gas in the 8 o’clock arc, we make use of
a grid of Charlot & Longhetti (2001, hereafter CL01) models. We adopt a constant
star formation history (SFH) and adopt the same grid used by Brinchmann et al.
(2004, hereafter B04; see Appendix A in Shirazi & Brinchmann (2012) and B04
for further details). In total, the model grid used for the fits has 2 × 105 different
models. The model grids and corresponding model parameters are summarised
in Table 3.4. Our goal here is to derive representative overall parameters for
the galaxy, and since the fitting methodology outlined in B04 works best with
[O ii] λ3727, Hβ, [O iii] λ4959, Hα, and [N ii] λ6584 all available, we here take the
emission line measurements from DZ11 since the last three lines fall outside the
spectral range of our SINFONI data cube. For the quantities that only depend on
line ratios, i.e., all but the star formation rates, this is appropriate, but for the star
formation rate we need to correct for light missed by the long slit observations of
DZ11, and we do this by normalising to the Hβ line flux from the SINFONI data.
73
SINFONI observations of the 8 o’clock arc
Figure 3.8 Median continuum in the H-band calculated in bins of 10Å (red filled
circles). Error bars show 16th and 84th percentiles of the distribution around the
median. Estimated model continuum from SED fitting is shown by the blue solid
curve.
We use the same Bayesian methodology as for the SED fit and again take the
median value of each PDF to be the best estimate of a given parameter and the
associated ±1 σ confidence interval to be spanned by the 16th–84th percentiles of
the PDF.
In Figure 3.9 we illustrate our technique by showing the effect on the PDFs of
parameters, when we fit a model to an increasing number of the emission lines. We
start with [O ii] λ3727 and show how we get more well defined PDFs for the indicated parameters as we add the emission lines indicated on the left. We show the
PDFs for the dust optical depth in V-band, the gas phase oxygen abundance, the
ionisation parameter, the conversion factor from Hα luminosity to star formation
rate (see CL01 for further details), the gas mass surface density, the dust-to-gas
ratio and the metal-to-dust ratio of the ionised gas. The latter three quantities are
discussed in some detail in Brinchmann et al. (2013, hereafter B13) and we discuss
them in more detail below. The resulting PDFs are shown for the A3 image and
the best-fit parameters derived from the final PDFs are summarised in Table 3.5.
The oxygen abundance reported by DZ10 and DZ11 is lower that what we find.
This is not entirely surprising for two reasons. First, it is well-known (e.g. Kewley
& Ellison, 2008) that metal abundance estimators show significant offsets, so even
when converted to a solar scale, one has to accept a systematic uncertainty in
any comparisons that use different methods for metallicity estimates. Second, the
estimates in DZ10 and DZ11 primarily rely on the calibration relationships (N2
calibration) from Pettini & Pagel (2004), which are based on local HII regions and
an extrapolation to higher metallicity. The use of these calibrated relationships
implicitly assumes that the relationship between ionisation parameter and metal74
The physical properties of the 8 o’clock arc
Line
[O ii]
[O ii]
Hδ
Hγ
Hβ
λair (Å)
3726.032
3728.815
4101.734
4340.464
4861.325
A2
65.8 ± 1.9
58.2 ± 1.8
25.6 ± 0.6
45.3 ± 0.6
102.6 ± 1.4
A3
36.2 ± 1.6
29.8 ± 1.5
12.9 ± 0.3
26.3 ± 0.3
75.8 ± 1
Table 3.3 Nebular emission lines identified in the 8 o’clock arc images A2 and A3
and their fluxes. Fluxes are not corrected for lens magnification. Integrated line
fluxes are in units of 10−17 erg s−1 cm−2 .
Parameter
Z, metallicity
U, ionization parameter
τV , total dust attenuation
ξ, dust-to-metal ratio
Range
−1 < log(Z/Z ) < 0.6, 24 steps
−4.0 < log U < −2.0, 33 steps
0.01 < τV < 4.0, 24 steps
0.1 < ξ < 0.5, 9 steps
Table 3.4 The CL01 model grid used in the present work.
licity is the same at high and low redshifts. This is a questionable assumption;
indeed the electron density we find for the 8 o’clock arc is considerably higher
than seen on similar scales in similar galaxies at low redshift (see Figure 3.10),
indicating that the U-Z relationship is different at high redshift, and thus that the
N2 calibration is problematic. In our modeling we leave U and Z as free variables;
thus, we are not limited by this. It is difficult to ascertain which approach is better
but the advantage of our approach for the 8 o’clock arc is that it uses exactly the
same models which used to fit local SDSS samples.
It is well-known that the estimation of ISM parameters from strong emission
lines is subject to systematic uncertainties (see however B13 for an updated discussion). To reduce the effect of these uncertainties, we have also assembled a
comparison sample of star-forming galaxies at z ∼ 0.1 from the SDSS. We used the
MPA-JHU value added catalogues (B04, Tremonti et al. (2004)) for SDSS DR74
as our parent sample. We define a star-forming galaxy sample on the basis of the
[N ii] 6584/Hα versus [O iii] 5007/Hβ diagnostic diagram, often referred to as the
BPT diagram (Baldwin, Phillips & Terlevich, 1981). For this we used the procedure detailed in B04 with the adjustments of the line flux uncertainties given in
B13. From this parent sample, we selecte all galaxies that have stellar mass within
0.3 dex of the value determined for the 8 o’clock arc and whose star formation rate
is within 0.5 dex of the 8 o’clock arc, based on the parameters determined from
SED fitting to the A2 image (Table 3.2). This resulted in a final sample of 329
galaxies, which we compare to the 8 o’clock arc below.
4 http://www.mpa-garching.mpg.de/SDSS/DR7
75
SINFONI observations of the 8 o’clock arc
log(Z/Z )d
log Ue
τV g
SFR (M yr−1 )h
log(Σgas /M pc−2 )i
12+Log O/H
16tha
median
0.046
-2.6
1.1
157
1.46
8.86
0.130
-2.2
1.6
165
1.60
8.93
b
84th
c
0.196
...f
1.8
173
1.87
9.02
a
The 16th percentile of the PDF of the given
quantity.
b
The 50th percentile, or median, of the PDF
of the given quantity.
c
The 84th percentile of the PDF of the given
quantity.
d
The log of the total gas-phase metallicity relative to solar.
e
The log of the ionization parameter evaluated at the edge of the Strömgren sphere (see
CL01 for details).
f
The electron density in the 8 o’clock arc is
higher than that assumed in the CL01 models, and the ionization parameter is therefore
close to the edge of the model grid, to we do
not quote an upper limit.
g
The dust attenuation in the V-band assuming an attenuation law τ(λ) ∝ λ−1.3 .
h
The star formation rate.
i
The log of the total gas mass surface density.
Table 3.5 Physical parameters of the ISM derived from spectrum of the A3 image.
3.4.3
AGN contribution
The preceding modeling assumes that the ionizing radiation in the 8 o’clock arc
is dominated by stellar sources. The position of the 8 o’clock arc in the BPT
diagnostic diagram (see Figure 6 in Finkelstein et al., 2009), which has been widely
used for classifying galaxies, does, however suggest that the emission lines from
this galaxy might be contaminated by an AGN. However, there is some evidence
indicating that the AGN contribution for this galaxy is negligible.
First, high-resolution VLA imaging at 1.4 and 5 GHz show that, although there
is a radio-loud AGN associated with the lensing galaxy and the arc is partially
covered by the radio-jet from this AGN, there is no detectable radio emission from
the unobscured region of the arc down to a 3σ flux-density limit of 108 µJy beam−1
(Volino et al., 2010). Second, we can detect He ii λ4686 for this galaxy, a highionization line that is very sensitive to the AGN contribution. Therefore, we can
use this line as a probe to estimate the AGN contribution to the spectrum of this
galaxy. We use a new diagnostic digram of He ii /Hβ versus [N ii] /Hα introduced
by Shirazi & Brinchmann (2012) to calculate this. As we do not have [N ii] and Hα
from the SINFONI observation, we use the DZ11 estimates for these emission lines.
The He ii λ4686/Hβ is not very sensitive to electron temperature and metallicity.
Therefore using the DZ11 estimate for [N ii] /Hα is sufficient for us to locate the
position of this galaxy in the He ii /Hβ diagram. Shirazi & Brinchmann (2012)
derive an almost constant line vs. metallicity at which the contribution of an
AGN to the He ii emission amounts to about 10%. They showed if the He ii is
76
The physical properties of the 8 o’clock arc
τV
12+Log O/H
Log U
η[Hα]
-3.5 -3.0 -2.5
6.0 6.5 7.0 7.5
Σgas
Log Dust/Gas
ξ
+[O II]
+Hγ
+Hβ
+[O III]
+Hα
+[N II]
0
1
2
3
7.5 8.0 8.5 9.0
1.5 2.0 2.5 -3.5 -3.0 -2.5 -2.0
0.2 0.3 0.4
Figure 3.9 The PDFs for optical depth in the V-band, the gas phase oxygen abundance, the logarithm of the ionization parameter, log10 of the conversion factor from
observed Hα luminosity to star formation rate, the logarithm of the gas surface
density in M /pc2 , the log of the dust-to-gas ratio and the metal-to-gas ratio. From
one row to the next we include the line indicated on the left side in the fitting in
addition to the previous emission lines. For the observational data in this figure
we make use of emission line fluxes measured by DZ11, from their Table 2, for
image A3 (see Figure 3.1). Our best-fit parameters and the associated 16th-84th
percentiles are summarized in Table 3.5. The abundance of oxygen reported by
DZ11 is lower than what we find.
contaminated by this amount, the total AGN contribution to other emission lines in
the spectrum of the galaxy is less than 1% (see Figure 3 in Shirazi & Brinchmann,
2012). As the position of the 8 o’clock arc in this digram (log He ii /Hβ = −1.4)
is below the above mentioned line, we can conclude that the contribution of AGN
to the optical emissions is negligible.
We note that the broad He ii λ1640 emission found by DZ10 can be affiliated
to the presence of Wolf-Rayet stars (see also Eldridge & Stanway, 2012).
3.4.4
Star formation rate and dust extinction
We have two main methods available to determine the star formation rate of the
8 o’clock arc from its emission line properties. We can use the SFR derived from
the emission line fits described above, but to provide spatially resolved SFR maps
77
SINFONI observations of the 8 o’clock arc
we need to turn to the lines available in the SINFONI data cube. The Hαλ6563
emission line is commonly used as a star formation rate indicator at low redshift
(Kennicutt, 1998). Unfortunately, for the redshift of the 8 o’clock arc, Hα falls
outside of the spectral range of the K band of SINFONI, and we can not use
this indicator to derive the spatially resolved SFR. We are therefore limited to
using Hβλ4861 as a tracer of the spatially resolved SFR. The advantages and
disadvantages of using this indicator to measure the SFR were originally discussed
by Kennicutt (1992) and were studied in detail by Moustakas et al. (2006). In
comparison to Hα, Hβ is more affected by interstellar dust and is more sensitive
to the underlying stellar absorption (see Section 3.3 and Figure 7 in Moustakas
et al., 2006).
We use the empirical SFR calibrations from Moustakas et al. (2006, Table 1),
parametrized in terms of the B-band luminosity, to calculate the SFR from Hβ
luminosity:
L(Hβ)
.
(3.1)
S FR (M yr−1 ) = 10−0.24 × 100.943 × 10−42
erg s−1
Moustakas et al. (2006) derived the SFR calibration assuming a Salpeter initial
mass function (IMF, Salpeter, 1955) over 0.1-100 M . The correction factor of
10−0.24 in Equation (3.1) is used to correct to a Chabrier IMF (Chabrier, 2003).
We interpolate between bins of L(B) to obtain the relevant conversion factor from
Hβ luminosity to SFR.
We use a dust extinction E(B − V) = 0.3 ± 0.1 derived from the Hγ/Hβ Balmer
line ratio with an intrinsic Hγ/Hβ = 0.468 to correct both SFRs for dust extinction. This is consistent with the estimate of DZ11 within the errors. We used
magnification factors of µA2 = 6.3 and µA3 = 4.9 to correct the SFR estimates
for the effect of gravitational lensing. The magnification calculated using the lens
modeling described in Section 3.5. We use 0.012 contour level (in count per second
unit) in the B HS T image for detecting individual images.
We measure Hβ = (102.6 ± 1.4) × 10−17 erg s−1 cm−2 for the A2 image corresponding to an observed SFR of 228±10.9 M yr−1 and Hβ = (75.8±1.)×10−17 erg s−1 cm−2
for the A3 image, corresponding to an observed SFR of 227±10.5 M yr−1 (corrected
for gravitational lensing magnification and dust extinction).
We can contrast this result to the integrated SFRs derived for the A2 and A3
images by fitting the CL01 models, after scaling the DZ11 line fluxes to match
the Hβ flux from the SINFONI cube. These are 160 ± 12 M yr−1 and 165 ± 10.5
M yr−1 , respectively. These values are somewhat discrepant but we note that
systematic uncertainties are not taken into account in the calculation here.
More importantly, the Hβ calibration allows us to calculate maps of the spatial
distribution of the optically visible star formation in the 8 o’clock arc. Furthermore, we can make use of our decomposition of the Hβ profile to calculate maps
for each component. This is shown in Figure 3.11 which shows these three calculated SFR maps for the A2 image. These were derived by integrating over the
blue (λ(4855 − 4859)Å), green (λ(4859 − 4863)Å), and red (λ(4863 − 4867)Å) parts
of the Hβ profile. We can see that the blue and red maps peak at different part of
the image, which suggests that they represent different components of the galaxy.
78
Number
The physical properties of the 8 o’clock arc
50
40
30
20
10
0
8.0
8.5
9.0
9.5
Number
Number
Number
12 + Log O/H
50
40
30
20
10
0
-3.5
-3.0
-2.5
Log dust/gas
-2.0
-1.5
50
40
30
20
10
0
0.5
1.0
1.5
Log Σgas [Msun/pc2]
2.0
2.5
50
40
30
20
10
0
0.5
1.0
1.5
2.0
2.5
Log ne [cm-3]
3.0
3.5
Figure 3.10 Top panel: distribution of oxygen abundance for the local comparison
sample. The solid red line shows the median value derived for image A3 with the
dashed vertical lines showing the 16% and 84% percentiles. The dash-dotted line
shows the median estimate for image A2 and is consistent with that derived for A3.
Second panel: same for the dust-to-gas ratio. Third panel: same for gas surface
mass density. Bottom panel: distribution of electron densities in the comparison
sample. These were derived using the ratio of the [S ii] 6716,6731 lines assuming
an electron temperature of T e = 104 K. The estimated electron density for the 8
o’clock arc, derived from the ratio of the [O ii] 3726,3729 lines, is shown by the
solid red line.
3.4.5
Metallicity and dust-to-gas ratio
The modeling described above gives an estimate of the oxygen abundance and the
dust-to-gas ratio of the 8 o’clock arc. Our best-fit oxygen abundance, 12+log O/H =
+0.09
8.93−0.07
(random errors only), is consistent with the determination by DZ11 within
the error. In the top panel of Figure 3.10 we compare this value to that determined
for our local comparison sample. The values for A3 are shown as the solid red
lines, with the 1σ confidence interval indicated by the dashed red lines. We also
calculate values from the fluxes provided for image A2, and these are shown as
the dot-dashed blue lines. We suppress the uncertainty estimate for this latter
estimate but it is similar to that of A3; thus, the two measurements are consistent
given the error bars.
79
SINFONI observations of the 8 o’clock arc
Two points are immediately noticeable from this plot. First, the oxygen abundance is mildly super-solar, (the solar oxygen abundance in the CL01 models is
8.82), and second, the value for the 8 o’clock arc is the same as for the local comparison sample. Since the local sample was selected to have approximately the
same stellar mass and star formation rate as the 8 o’clock arc, we must conclude
that the 8 o’clock arc lies on the stellar mass–oxygen abundance–star formation
rate manifold found locally (Mannucci et al., 2010). A similar conclusion was also
found by DZ11.
The second panel in Figure 3.10 shows the dust-to-gas ratio (DGR) of the
SDSS comparison sample as a histogram and the values for the 8 o’clock arc as
the vertical lines as in the top panel. The uncertainty in log DGR is fairly large,
but there is no evidence that the 8 o’clock arc differs significantly from similar
galaxies locally. As we will see next, this does not necessarily imply that the ISM
has all the same properties.
3.4.6
The gas surface mass density
It is of great interest to try to estimate the gas content of high redshift galaxies in
general, and we have two methods to do this for the 8 o’clock arc. The first is to
use the Kennicutt–Schmidt relation (Schmidt, 1959; Kennicutt, 1998) between the
gas surface density and star formation rate per unit area to convert our spatially
resolved SFRs to gas surface densities. We can then use this estimated gas surface
density to calculate the gas mass. The estimated gas surface density is plotted
in Figure 3.12 and is simply a transformation of the SFR map shown earlier. We
can then integrate this surface density of gas to get an estimate of the total gas
mass. With the canonical calibration of Kennicutt (1998) this gives an estimate
of log(Mgas /M ) = 10.43 ± 1.18. If instead we use the calibration used by DZ11
(Bouché et al., 2007), we find a total gas mass of 10.30 ± 1.20. In either case, the
gas content is comparable to the total stellar mass of the galaxy.
We can also estimate the gas content in a way independent of the star formation
rate by exploiting a new technique presented in B13, which exploits the temperature sensitivity of emission lines to provide constraints on the dust-to-metal ratio;
together with an estimate of the dust optical depth (see B13 for detail), one can
derive a constraint on the surface mass density of gas (molecular plus atomic).
B13 show that this technique can give gas surface densities in agreement with H
i+H2 maps to within a factor of 2. The results of our fits are shown in Fig. 3.9. In
passing, we note that this technique, in contrast to the Kennicutt-Schmidt method,
only relies on line ratios and these lines all originate in much the same regions;
thus, lensing amplification should be unimportant here.
80
The physical properties of the 8 o’clock arc
Figure 3.11 Top: Hβ line map of the arc. This map is not corrected for magnification. Lower panels: three different calculated SFR (corrected for lensing
magnification and dust extinction) maps for the A2 (middle) image. SFRs are
derived by integrating over the blue, green and red parts of the Hβ profile, respectively from top to bottom. The wavelength ranges covered in the indicated
above the SFR maps. The color scale shows the SFR per kpc2 as indicated.
Figure 3.12 The resolved gas surface density for the A2 image estimated using the
Kennicutt-Schmidt relation.
81
SINFONI observations of the 8 o’clock arc
The median estimate is log(Σgas /M pc−2 ) = 1.60, which is consistent with the
value that derived using the Kennicutt-Schmidt relation but derived in an independent manner, and crucially using a method that is formally independent of the
star formation rate. We contrast the estimate of log Σgas for the 8 o’clock arc to
the SDSS comparison sample in the third panel of Figure 3.10. The median value
for the local sample is Σgas ≈ 17 M pc−2 , while the median estimate for the 8
o’clock arc is Σgas ≈ 40 M pc−2 . Since this comparison is differential and does not
depend on scaling relations calibrated on local samples, the conclusion that the
surface density of gas in the 8 o’clock arc is more than twice that of similar z ∼ 0.1
galaxies should be fairly robust. Thus, while the 8 o’clock arc does lie on the M∗ –
O/H–SFR relation, it has a significantly higher gas surface density than galaxies
lying on the same relation locally. This highlights the fact that even though a
particular scaling relation is established at high-z, it does not imply that galaxies
lying on the relation are necessarily similar to local galaxies.
3.4.7
Electron density
We can estimate the electron density using the [O ii]λ3729/[O ii]λ3726 ratio. Both
the X-shooter and the SINFONI observations of the 8 o’clock arc give similar
ratios for the [O ii] λ3729, 3726 lines. The SINFONI observations give a value of
[O ii] λ3729/[O ii] λ3726 = 0.88 which corresponds to a high electron density of
ne ∼ 600 cm−3 for this galaxy. High electron densities have been found in many
other high-z galaxies (see e.g. Lehnert et al., 2009; Wuyts et al., 2012; Lehnert et
al., 2013).
For the local comparison sample we are unable to reliably estimate ne from
the [O ii] line ratio, so we use the [S ii] 6717,6731 lines instead. These values are
compared to that for the 8 o’clock arc in the bottom panel of Figure 3.10 and as
that figure makes clear, the 8 o’clock arc has significantly higher electron density
than similar SDSS galaxies. Note that the [S ii] ratio is not very sensitive to
electron density variations at ne < 100 cm−3 , hence the somewhat truncated shape
of the distribution there.
The high electron density is likely to lead to a high ionization parameter, and
this is born out by our fits (Figure 3.9). Indeed, the 8 o’clock arc lies close to
the maximum ionization parameter in the CL01 models, and its electron density
is well above the value assumed (ne = 100 cm−3 ) in the CL01 model calculations.
For this reason, we prefer to focus on the ne determination which is independent
of this and which implies a higher ionization parameter for this galaxy relative to
nearby objects.
More immediately, the electron density is related to the pressure in the H ii
region through P = ne kT e , and since the electron temperature in the H ii region
almost certainly will be very similar in the low-z sample and the 8 o’clock arc,
given that their metallicities are similar, we conclude that the pressure in the H ii
regions in the 8 o’clock arc is ∼ 5 times higher than in the typical SDSS comparison
galaxy.
The reason for this elevated H ii region pressure is less clear. Dopita et al.
(2006) showed that for expanding H ii regions the ionization parameter depends
82
Source Reconstruction
on a number of parameters. A particularly strong dependence was seen with
metallicity, but as our comparison sample has similar metallicity to that of the
8 o’clock arc we can ignore this. The two other major effects on the ionization
parameter come from the age of the H ii region and the pressure of the surrounding
ISM. It is of course possible that we are seeing the 8 o’clock arc at a time when its
H ii regions have very young ages, and hence high ionization parameter, relative
to the local comparison sample. Since we are considering very similar galaxies in
terms of star formation activity and probe a fairly large scale this seems to be
a fairly unlikely possibility, but it cannot be excluded for a single object. The
pressure in the surrounding ISM in Dopita et al (2006) models has a fairly modest
effect on the ionization parameter, U ∝ P−1/5
ISM . Thus we would expect the ISM
density in the 8 o’clock arc to also be higher than in the comparison sample by a
factor of ∼ 5. We already saw that the gas surface density is higher than in the
comparison sample by a factor > 2 thus this is not an unreasonable result and it
does not seem to be an uncommon result for high-z galaxies (Shirazi et al., 2013,
submitted).
3.5
Source Reconstruction
In order to study emission line maps and the kinematics of the galaxy in the source
plane, we need to reconstruct the morphology of the 8 o’clock arc using gravitational lens modeling. The lens modeling also allows us to derive the magnification
factors of the multiple-lensed images which were used to estimate the corrected
SFR and the stellar mass of the galaxy in the previous section. In the following,
we describe our lens modeling procedure.
3.5.1
Gravitational lens modeling
To reconstruct the lens model for this system, we make use of the Bayesian grid
based lens modeling technique presented by Vegetti & Koopmans (2009), which is
optimized for pixelized source surface brightness reconstructions.
In order to obtain a robust lens model, we first consider the high resolution
and high signal-to-noise ratio B band HS T image. We assume the lens mass
distribution to follow a power-law elliptical profile with surface mass density, in
units of the critical density, defined as follows
k(x, y) =
1−γ
γ
(y − y0 )2
k0
+ rc 2 )( 2 )
√ (2 − ) ((x − x0 )2 +
2
2 q
2
q
(3.2)
We also include a contribution from external shear. In particular, the free
parameters of the model are the mass density normalization k0 , the position angle
θ, the mass density slope γ (γ = 2.0 for an isothermal mass distribution), the axis
ratio q, the centre coordinates x0 and y0 , the external shear strength Γ, the external
shear position angle Γθ , and the source regularization level (i.e., a measure of the
level of smoothness of the source surface brightness distribution), while the core
radius is kept fixed to the negligible value of rc ≡ 0. The most probable parameters
83
SINFONI observations of the 8 o’clock arc
Figure 3.13 Top-left panel: the arc and the counter image in the B band HS T
image. The foreground galaxy (lens) has been removed from this image. Topright panel: the best-fit model. Lower-left panel: the residuals after subtracting
this model from the data. The reconstructed B band HS T image is shown in
the lower-right panel. From this image we see that the source in the rest frame
UV consists of at least three components; the main galaxy component, a clump
separated by 0.15”, which is shown by the purple dashed ellipse and another clump
separated by 0.15”, which is shown by the red dashed ellipse.
of the model are k0 = 3.367, θ = 14.54, q = 0.618, γ = 2.009, Γ = 0.062, Γθ = 10.597.
Using the same mass model for the lens galaxy, we also model the NICMOS data.
While the B band HS T data probe the rest frame UV and have a higher resolution
in comparison to the NICMOS data, the latter have the advantage of providing us
with information about the continuum in the J and H bands, where Hβ and [O ii]
emission lines are located in the spectra. The most probable mass model for the
NICMOS data has the following parameter values: k0 = 3.328, θ = 14.30, q = 0.672,
γ = 2.020, Γ = 0.077 Γθ = 13.37. Both results are consistent with DZ11 best-fit
parameters, within the error bars. However, unlike DZ11, we do not optimize for
the core radius because Einstein rings only make is possible to constrain the mass
distribution at the Einstein radius.
84
Source Reconstruction
Figure 3.14 The reconstructed H band HS T image is shown in the plot. From this
image we can see that the source in the rest frame optical is formed of multiple
components, two main galaxy components. The clump in the reconstructed UV
image is partially resolved in this image and it shows a slightly different morphology. The contour shows the HS T B-band reconstructed image. The yellow ellipse
in the corner of the image shows the spatial resolution in the source plane.
Figure 3.15 The reconstructed Hβ image is shown in the plot. From this image we
can see that the source in the rest frame optical is formed of multiple components,
two main galaxy components. The clump in the reconstructed UV image is not
resolved in this image. The contour shows the reconstructed HS T B-band image
from Figure 3.13. The yellow ellipse in the corner of the image shows the spatial
resolution in the source plane.
85
SINFONI observations of the 8 o’clock arc
Figure 3.16 From left to right this figure shows reconstructed Hβ maps from the
blue, green and red components of the spectral line. The contour line shows the
reconstructed HS T B-band image from Figure 3.13. It can be seen that the blue
map predominantly contributes to the eastern part — from a detailed inspection
of the lens model we find that in the image plane this is predominantly seen in
the A1 and A2 images. Note also that the red component of the spectral line
predominantly originates in the west.
Figure 3.17 The reconstructed [O ii] image. Here we are unable to separate two
components; this might be due to a higher [O ii] /Hβ ratio in the left component.
The red contour shows the reconstructed HS T B-band image from Figure 3.13.
These two most probable models are used to map the image plane into the
source plane and reconstruct the original morphology of the 8 o’clock arc in the B
and H bands, respectively. Thanks to the Bayesian modeling technique, the most
86
Source Reconstruction
probable source surface brightness distribution for a given set of lens parameters
are automatically provided. The reconstructed B band HS T image is shown in
the lower-right panel of Figure 3.13. From this image we can see that the source
in the rest frame UV consists of multiple components, including the main galaxy
component and two clumps separated by 0.15 arcsec (i.e., 1.2 kpc in projected
distance) indicated by the purple and red ellipses. The reconstructed H band
HS T image is shown in Figure 3.14. This image shows that the source in the
rest-frame optical also consists of multiple components.
3.5.2
Reconstructed-Hβ and [O II] emission lines maps in the
source plane
Here we make use of the B band HS T data modeling to reconstruct the Hβ and
[O ii] emission line maps of the galaxy. These lines have the highest signal-to-noise
that we obtain from the SINFONI data.
For each spectral pixel image (frame) of the SINFONI data cubes, we derive
the most probable source surface brightness distribution by keeping the lens parameters fixed at the best values recovered from the B band HS T data modeling
(after taking into account the rotation of the image), while re-optimizing for the
source regularization level. Because of the relatively low signal-to-noise SINFONI
data and non-homogenous sky background, we can not use all the lensed images.
We focus instead on the highest magnification image, which is the A2 image.
Before reconstructing the Hβ map in the source plane, we first bin in the
spectral direction by a factor of 4. This corresponds to the spectral resolution
(FWHM= 7.9 Å) that we measure from the line widths of the night sky lines around
the Hβ line. This provides us with higher signal-to-noise image plane frames. We
finally make a Hβ source cube from these reconstructed source frames and use that
to derive the kinematics of the galaxy. A reconstructed Hβ map is shown in Figure
3.15. This image also shows two galaxy components. In order to better understand
the morphology of the Hβ image, we divide the Hβ spectral range into three equal
spectral bins defined as blue (λ(4855 − 4859)Å), green (λ(4859 − 4863)Å) and red
(λ(4863 − 4867)Å) intervals, corresponding to three SFR maps shown in Figure
3.11. We then reconstruct the source surface brightness distribution for each of
these images, using the same method that was used for the whole Hβ image. The
three panels in Figure 3.16 show the reconstructed sources for these images. We
can see that the west part of the Hβ line map is very weak and only dominates
in the red image (right panel); on the other hand, the eastern part is dominant
in the blue image (left panel). Figure 3.17 shows a reconstructed [O ii] image of
the galaxy. We see that the eastern component is dominant in this image. Here
we are unable to separate the two components; this might be due to a higher
[O ii] /Hβ ratios in the eastern component, but could also be caused by the lower
spatial resolution at these wavelengths. However, we rule out the later explanation
by convolving the Hβ map with a Gaussian PSF matching the slightly different
[O ii] map (J-band) PSF.
87
SINFONI observations of the 8 o’clock arc
Figure 3.18 Hβ profile derived from the reconstructed Hβ source. The widths
of both Gaussian components are 1.59 Å, which gives a velocity dispersion of
98 km s−1 . The velocity offset between the two components is 246 km s−1 . In the
source plane, we have fewer bins than Figure 3.3, since we bin in spectral resolution
by a factor 4 before reconstructing the Hβ map in the source plane. To make this
profile, we interpolate between those bins to have the same binning as Figure 3.3.
3.5.3
Hβ profile of the reconstructed source
We use the same fitting method that we used in Section 3.3 to fit a two component
Gaussian to the Hβ profile for every pixel. We also integrate over the total flux of
the galaxy and fit a two component Gaussian to it to compare it to our study in the
observed plane (see Section 3.3.2). Figure 3.18 shows the Hβ profile derived from
the reconstructed Hβ source. We see that this profile also retains two components.
The width of the Gaussian for both components is 1.59 Å, which gives a velocity
dispersion of 98±44 km s−1 . This is consistent, within the errors, with our estimated
velocity dispersion for the A2 and A3 images in the observed plane (i.e., 104 ±
42 km s−1 ). The velocity offset between the two components is 246 ± 46 km s−1 and
matches the offset that we derive for the A2 and A3 images in the image plane.
Figure 3.19 shows the velocity and velocity dispersion maps derived from the
reconstructed Hβ source. From these maps, we see that the east galaxy component
has a lower velocity and velocity dispersion relative to the western component. The
western component also shows a smoother velocity gradient.
The Hβ line flux divided by the H-band continuum flux is shown as a proxy
for EW(Hβ) in Figure 3.20. Here, the H-band continuum map is convolved to the
same PSF as the Hβ map. From this we see that the outskirts of the galaxy show
a clumpy and higher EW(Hβ). The eastern component of the galaxy also shows a
higher EW(Hβ), which might be interpreted as a younger age relative to that of
the main component.
88
Dynamics
Figure 3.19 The top and middle panels show the velocity and velocity dispersion
maps, respectively. The velocity map is derived using single Gaussian fits but
velocity dispersion map represents both components. These are derived from the
reconstructed Hβ source. The bottom panel shows the velocity dispersion map
derived using single Gaussian fits. The contours show the reconstructed HS T
B-band image from Figure 3.13.
89
SINFONI observations of the 8 o’clock arc
Figure 3.20 Hβ line flux divided by the H-band continuum flux shown as a proxy
for EW(Hβ). We see that the outskirts of the galaxy show a clumpy and higher
EW(Hβ). The eastern component of the galaxy shows a higher EW(Hβ), i.e., a
younger age, relative to that of the main component. The red contour shows the
reconstructed HS T B-band image.
3.6
3.6.1
Dynamics
Hβ Kinematics
To test whether the kinematics of the galaxy are consistent with those of a rotating disk, we compare the velocity and velocity dispersion maps derived from the
reconstructed Hβ map with an exponential disk model. Given the low resolution
and the low S/N of our data, we simulate a very simple system. The disk models
is created using the DYSMAL IDL code (Davies et al., 2011, see also Cresci et
al. 2009 for description of the code). The code was used extensively to derive
intrinsic properties of disk galaxies (e.g., for estimating the dynamical mass of
high-z galaxies; see Cresci et al., 2009). The code uses a set of input parameters
which constrain the radial mass profile as well as the position angle and systemic
velocity offset, in order to derive a 3D data cube with one spectral (i.e., velocity)
and two spatial axes. This can be further used to extract kinematics. The best-fit
disk parameters are derived using an optimized χ2 minimization routine and the
observed velocity and velocity dispersion. The mass extracted from DYSMAL is
that of a thin disk model assuming supported only by orbits in ordered circular
rotation.
We do not have any constraints on the inclination of our system. Therefore,
we use a nominal inclination of 20 degrees. We account for spatial beam smearing
from the PSF and velocity broadening due to the finite spectral resolution of
the instrument and also rebin by a factor of 4 in the spectral direction in our
modeling. We then compare this spatially and spectrally convolved disk model to
90
Dynamics
Figure 3.21 The left panel shows the observed velocity derived from the reconstructed Hβ map. The right panel shows the best velocity fit.
the observations.
We focus on the western component in the velocity map shown in Figure 3.19
because from the lens modeling we know that this part contains the main component of the galaxy and also shows a smoother observed gradient. The best-fitting
exponential disk model for this component is shown in Figure 3.21. While the disk
model can reproduce some large-scale features of the velocity field, the residuals
are substantial. We can therefore rule out a single rotating disk as a reasonable
description of this system. We conclude that the 8 o’clock arc has a complex
velocity field that cannot be explained by a simple rotating disk.
Furthermore, there appears to be a second component from a clump (see red
ellipse in Figure 3.13) that partially overlaps with this component. Whether this is
a sign of an on-going merger is difficult to ascertain with the present data. Indeed,
the S/N in Hβ does not warrant a much more complex model to be fitted.
3.6.2
Dynamical mass
DZ11 estimated the dynamical mass of the 8 o’clock arc from the line widths via
the relation presented by Erb et al. (2006b). We use the same method to estimate
the dynamical mass using our estimated velocity dispersion (σ) and half-light
radius. For rotation-dominated disks, DZ11 assumed that the enclosed dynamical
91
SINFONI observations of the 8 o’clock arc
Figure 3.22 The observed velocity derived from the reconstructed Hβ map along
the slit (shown by dashed lines in Figure 3.21) is shown by the solid curve. The
dashed line here shows the best-fit velocity model.
mass within the half-light radius, r1/2 , is Mdyn,rot (r < r1/2 ) = (2.25σ2 r1/2 )/G and
multiply this resulting mass by two to obtain the total dynamical mass, where G =
4.3×10−6 kpc (km s−1 )2 M−1
is the gravitational constant. For dispersion-dominated
objects, they applied the isotropic virial estimator with Mdyn,disp = (6.7σ2 r1/2 )/G,
appropriate for a variety of galactic mass distributions (Binney & Tremaine, 2008).
In this case, Mdyn,disp represents the total dynamical mass.
For estimating the half-light radius we run galfit on the reconstructed B and H
band images. This gives us r1/2 = 2.8±0.2 kpc. We measure σ = 104±42 km s−1 and
a rotation-dominated dynamical mass log (Mdyn /M ) = 10.2 ± 0.3 and a dispersiondominated dynamical mass log (Mdyn /M ) = 10.7 ± 0.27 (these values are corrected
for instrumental broadening). Using the de-lensed spectra, we also estimate a
σ = 98 ± 44 km s−1 , which give us 10.1 ± 0.6, and 10.6 ± 0.6 for the rotationdominated and dispersion-dominated dynamical masses, respectively. The disk
model fit can also provide a dynamical mass estimate, log (Mdyn /M ) = 9.5, but
we do not use this here because it only accounts for the west component of the
velocity map.
We note that the idea of using a single line width to estimate dynamical mass is
not convincing (this should be done for blue, green and red components individually). In that case the velocity field is clearly more like a merger, so neither of these
dynamical mass indicators are reliable. Therefore, obtaining a robust dynamical
mass estimate would require considerably more sophisticated models. The simple
models are not physically constraining.
92
Dynamics
3.6.3
A massive outflow of gas?
It has been shown that many of high-z star-forming galaxies show evidence for
powerful galactic outflows, indicated by studying UV absorption spectroscopy
(Pettini et al., 2000; Shapley et al., 2003; Steidel et al., 2010; Weiner et al., 2009;
Kornei et al., 2012) and broad Hα emission-line profiles (Shapiro et al., 2009; Genzel et al., 2011; Newman et al., 2012a). Recently Newman et al. (2012b) showed
how galaxy parameters (e.g., mass, size, SFR) determine the strength of these
outflows. They decomposed the emission line profiles into broad and narrow components and found that the broad emission is spatially extended over ∼ a few
kpc. Newman et al. (2012b) showed that star formation surface density enforces
a threshold for strong outflows occurring at 1 M yr−1 kpc−2 . The threshold necessary for driving an outflow in local starbursts is 0.1 M yr−1 kpc−2 (Heckman,
2002).
The 8 o’clock arc with integrated star formation surface density of
9.2 M yr−1 kpc−2 is certainly in the regime of strong outflow. If we consider
the ratio of the Gaussian flux in the blue-shifted component to that of the main
component (∼ 0.5) as the Fbroad /Fnarrow , then our result is consistent with what
Newman et al. (2012b) show in their Figure 2. However, we note that this definition is not exactly what Newman et al. (2012b) introduced as Fbroad /Fnarrow as
we do not fit broad and narrow components but two component with the same
width. This provides us with a lower limit on the line width of the blue shifted
component.
In our data we find a blue-shifted component to Hβ as discussed in Section 3.3.2.
As we mentioned earlier we used the same Gaussian width for both main and
blue-shifted components. Given the low SN, a unique broad fit with a physical
meaning can not be found considering the fact that residual from sky lines might
create broad line widths. The velocity offset between this component and the main
component of the Hβ line is ≈ 190 km s−1 for the A2 image and ≈ 280 km s−1 for
the A3 image. This blue-shifted component could be due to an outflow of gas or
a minor merger.
In support of the outflow picture, Finkelstein et al. (2009) and DZ10 both
observed that ISM lines in the rest-UV spectrum of the 8 o’clock arc are blueshifted relative to the stellar photospheric lines. They also argued that this was a
sign of an outflow of gas from the galaxy and taken together with the SINFONI
results this strengthens the outflow picture. A further argument for an outflow
is the fact that as we saw in section 3.4.7, the H ii regions have an elevated
internal pressure, at least compared to similar galaxies locally, and it is reasonable
to assume that this aids in driving an outflow (Heckman et al., 1990). In support
of the merger picture, Figure 3.16 indicates that much of the blue component flux
is extended and arguably galaxy shaped.
Taking the evidence from the UV ISM lines, the Hβ profile, the lens model
and the high ISM pressure together, it is likely that there is a significant outflow
component to the blue-shifted wing, but the question here is whether these observations correspond to a reasonable amount of gas in the outflow. To answer this
we also estimate the mass of ionized hydrogen from the luminosity of Hβ, LHβ , of
93
SINFONI observations of the 8 o’clock arc
the blue-shifted component and main component of the 8 o’clock arc using (e.g.,
Dopita & Sutherland, 2003):
Mionised =
mH × LHβ
1.235 × 10−25 T 4−0.86 ne
,
(3.3)
where mH is the mass of the hydrogen atom, T 4 the electron temperature in units
of 10,000 K and ne the electron density. The mass of ionized hydrogen for the
blue-shifted component with LHβ of 45 × 1041 erg s−1 is 107.7±1.3 M assuming ne =
600 cm−3 , which can be contrasted with the mass of ionized hydrogen in the main
component which is 108.2±1.3 M (with LHβ = 127×1041 erg s−1 ). The estimate for the
blue-shifted component is clearly an upper limit because from the argument above
it seems likely that the blue component is not purely an outflow. Thus, taking this
at face value, we would find that with an outflow rate of 10% of the star formation
rate, a conservative value given local observations (e.g., Martin, 1999, 2006), we
need the current star formation activity to have lasted < 3 × 108 yr, which is not
unreasonable (to estimate the star formation time scale, the mass is divided by
10% of the SFR).
In order to use the outflow mass quantitatively, we need to have an estimate
of outflow mass in neutral hydrogen. However, we note that we can not estimate
this for the blue-shifted component since NHI estimated in DZ10 is given for the
whole galaxy. To estimate the total neutral hydrogen mass in the outflow and the
mass outflow rate (Pettini et al., 2000), we use the following formulae given in
Verhamme et al. (2008)
MHI ≈ 107
ṀHI = 6.
r 2 NHI M ,
1 kpc 1020 cm−2
Vexp
r NHI M yr−1 ,
20
−2
−1
1 kpc 10 cm
200 km s
(3.4)
(3.5)
where the first equation relate H I mass in the shell to its column density NHI
and the second equation assumes that the mechanical energy deposited by the
starburst has produced a shell of swept-up interstellar matter that is expanding
with a velocity of Vexp .
Assuming our estimated half-light radius from the rest-frame UV reconstructed
image and our assumed outflow velocity (r = 2.8 kpc, Vexp = 200 km s−1 ) and taking
NHI = 1020.57 cm−2 from DZ10, we find neutral gas masses of MHI = 2.9 × 108 M
and an outflow rate of ṀHI = 62.4 M yr−1 . This gives us a mass-loading factor of
η = ṀHI /SFR = 0.27.
3.7
Conclusions
We present a spatially-resolved analysis of the 8 o’clock arc using near-IR IFU
data. From this we recovere the Hβ map and the spatially-resolved Hβ profile. We
showed that Hβ has different profiles at different spatial pixels and is composed
of multiple components. We carefully modeled the strong emission lines in the
94
Conclusions
galaxy and compared the results to a local comparison sample. This allowed us to
conclude that
• The 8 o’clock arc lies on the same M∗ -O/H-SFR manifold as similar starforming galaxies locally (Mannucci et al. (2010); Lara-López et al. (2010),
DZ11).
• The gas surface density in the 8 o’clock arc log(Σgas / M pc−2 ) =
1.87
is more than twice (×2.164.01
1.61.46
1.55 ) that of similar galaxies locally
log(Σgas(analogs) / M pc−2 ) = 1.271.48
.
Comparing
this with other high-z results
1.02
(e.g., Mannucci et al., 2009, who measure gas surface densities in the range of
2.5-3.3 M pc−2 ), the gas surface density for the 8 o’clock arc is lower. Note
that as mentioned by Mannucci et al. (2009), they are sampling the central,
most active parts of the galaxies, so those values should be considered as the
maximum gas surface densities.
• The electron density, and thus the H ii region pressure, in the 8 o’clock arc
is ∼ 5 times that of similar galaxies locally. As (Wuyts et al., 2012) pointed
out, the electron density measurements for high-z galaxies range from the
low density limit to ne > 104 cm−3 . Although these differences depend on the
method for measuring the electron density, these also imply a huge difference
in the physical properties of star-forming regions in star-forming galaxies at
z ∼ 2. The difference between electron density at low-z and high-z have been
studied recently by Shirazi et al. (2013, submitted) who compared a sample
of 14 high-z galaxies with their low-z counterparts in the SDSS and showed
that high-z star-forming galaxies that have the same mass and sSFR as low-z
galaxies have a median of 8 times higher electron densities.
Taken together these results imply that although the 8 o’clock arc seems superficially similar to local galaxies with similar mass and star formation activity, the
properties of the ISM in the galaxy are nonetheless noticeably different.
We showed that the two images A2 and A3 have the same Hβ profiles, which
of course is to be expected because they are two images of the same galaxy. But
this contrasts with the results from long-slit observations of the object by DZ11
who found different profiles. The similarity of the profiles from the IFU data has
allowed us to rule out a significant contribution of substructures to the surface
brightness of the A2 image.
The integrated Hβ profile of both images show a main component with a blue
wing which can be fitted by another Gaussian profile with the same width. The
width of the Gaussian components for both images are 1.7±0.7 Å, which gives velocity dispersion ∼ 104 ± 42 km s−1 . The velocity offset between the two components
is 278 ± 63.5 km s−1 for the A3 image and 191 ± 63 km s−1 for the A2 image which
are consistent within the errors. Since both DZ11 and Finkelstein et al. (2009)
showed ISM lines are blue-shifted relative to the stellar photospheric lines, suggesting gas outflows with 120-160 km s−1 , and find a comparatively high pressure
in the H ii regions of the 8 o’clock arc, we interpret this blue-shifted component as
an outflow. However, we cannot rule out that the blue-shifted component might
represent a minor merger.
95
SINFONI observations of the 8 o’clock arc
To study the de-lensed morphology of the galaxy, we used existing B and H
band HS T images. Based on this, we constructed a rigorous lens model for the
system using the Bayesian grid based lens modeling technique. In order to obtain
a robust lens model, we used the lens modeling of the B band HS T image to
reconstruct the Hβ line map of the galaxy. We then presented the de-lensed Hβ
line map, velocity and velocity dispersion maps of the galaxy. As an example
application we derived the Hβ profile of the reconstructed source and showed that
this also requires two Gaussian components with a width of 98 ± 44 km s−1 and
velocity separation of 246 ± 46 km s−1 .
By fitting an exponential disk model to the observed velocity field, we showed
that a simple rotating disk cannot fit the velocity field on its own. Thus, a more
complex velocity field is needed, but the S/N of the present data does not allow a
good constraint to be had. This also implies that obtaining an accurate dynamical
mass for the 8 o’clock system is not possible at present.
Similar to some of clumpy galaxies studied by Genzel et al. (2011), the 8 o’clock
arc shows a blue-shifted wing but with a less broad profile. We note that as can
be seen for example from Figure 3.13, the galaxy has a very clumpy nature in the
source plane, but because of the lack of spatial resolution we are not able to study
these clumps in more detail.
Acknowledgements
We are very thankful for useful comments and suggestions of the anonymous referee. We would like to thank also Ali Rahmati for his useful comments on this paper, Raymond Oonk and Benoit Epinat for useful discussion about SINFONI data
reduction, Richard Davies for providing us with DYSMAL code, Johan Richard for
his help on the lens modeling and also Max Pettini, Alicia Berciano Alba,Thomas
Martinsson and Joanna Holt for useful discussions.
We would like also to express our appreciation to Huan Lin, Michael Strauss,
Chris Kochanek, Alice Shapley, Dieter Lutz, Chuck Steidel, and Christy Tremonti
for their help on the HS T proposal along with our spacial thanks to Andrew Baker.
M. Sh., S. A. and D. T. acknowledge the support of Mel Ulmer at North
Western University for providing them a meeting room and working place in MayJune 2011.
S.V. is grateful to John McKean for useful comments and discussions on the
lens modeling
During part of this work S.V. was supported by a Pappalardo Fellowship at
MIT.
This research has made use of the Interactive Data Language (IDL) and QFitsView5 .
5 www.mpe.mpg.de/~ott/QFitsView
96
AppendixA: Gaussian decomposition
3.8
AppendixA: Gaussian decomposition
As we have shown in Section 3.3.3, the resolved Hβ profiles are not well fitted
by a single Gaussian. Here we show the best-fit Gaussian intensity maps of the
main and blue-shifted components of the galaxy for the image A2 and A3 in
Figure 3.23 and Figure 3.24, respectively. As we mentioned earlier, during the
fitting we require the lines to have the same velocity widths. In both figures,
the upper panel shows the Gaussian intensity map of the main component and
the lower panel shows the Gaussian intensity map of the blue-shifted component.
Intensities are in unit of 10−17 erg cm−2 s−1 . We see that the main and the blueshifted components of the galaxy are offset spatially (∼ 1”) in the A2 image as
we see also in Figure 3.11 showing three calculated SFR maps of the A2 image
in blue, green and red channels. However, for the A3 image it is not possible to
decompose these components spatially.
97
SINFONI observations of the 8 o’clock arc
Figure 3.23 Lensed image A2: upper panel shows the Gaussian intensity map of the
main component; lower panel shows the Gaussian intensity map of the blue-shifted
component. Intensities are in unit of 10−17 erg cm−2 s−1 .
Figure 3.24 Lensed image A3: upper panel shows the Gaussian intensity map of the
main component; lower panel shows the Gaussian intensity map of the blue-shifted
component. Intensities are in unit of 10−17 erg cm−2 s−1 .
98
REFERENCES
References
Allam, S. S., Tucker, D. L., Lin, H., Diehl, H. T., Annis, J., Buckley-Geer, E. J.,
& Friemam, J. A. 2007, ApJ, 662, L51
Baldwin, J. A., Phillips, M. M., & Terlevich, R. 1981, PASP, 93, 5
Binney, J., & Tremaine, S. 2008, Galactic Dynamics, 2nd Edition, Princeton
University Press, Princeton, NJ USA
Bonnet, H., Conzelmann, R., Delabre, B., et al. 2004, Proc. SPIE, 5490, 130
Bouché, N., Cresci, G., Davies, R., et al. 2007, ApJ, 671, 303
Brinchmann, J., Charlot, S., et al. 2004, MNRAS, 351, 1151
Brinchmann, J., Charlot, S., Kauffmann, G., et al. 2013, MNRAS, 1276
Bruzual, G., & Charlot, S. 2003, MNRAS, 344, 1000
Chabrier, G. 2003, PASP, 115, 763
Charlot, S., Longhetti, M., 2001, MNRAS, 323, 887
Contini, T., Garilli, B., Le Fèvre, O., et al. 2012, A & A, 539, A91
Cresci, G., Hicks, E. K. S., Genzel, R., et al. 2009, APJ, 697, 115
Davies, R., Förster Schreiber, N. M., Cresci, G., et al. 2011, APJ, 741, 69
de Barros, S., Schaerer, D., & Stark, D., 2012, arXiv:1207.3663
Dessauges-Zavadsky, M., D’Odorico, S., Schaerer, D., Modigliani, A., Tapken, C.,
& Vernet, J. 2010, A&A, 510, 26
Dessauges-Zavadsky, M., Christensen, L., D’Odorico, S., Schaerer, D., & Richard,
J. 2011, A & A, 533, A15
Dopita, M. A. & Sutherland, R. S. 2003, Astrophysics of the diffuse universe (Astrophysics of the diffuse universe, Berlin, New York: Springer, 2003. Astronomy
and astrophysics library, ISBN 3540433627)
Dopita, M. A., Fischera, J., Sutherland, R. S., et al. 2006, APJS, 167, 177
Dopita, M. A., Fischera, J., Sutherland, R. S., et al. 2006, APJ, 647, 244
Eisenhauer, F., Abuter, R., Bickert, K., et al. 2003, Proc. SPIE, 4841, 1548
Eldridge, J. J., & Stanway, E. R. 2012, MNRAS, 419, 479
Erb, D. K., Shapley, A. E., Pettini, M., Steidel, C. C., Reddy, N. A., & Adelberger, K. L. 2006a, ApJ, 644, 813
Erb, D. K., Steidel, C. C., Shapley, A. E., Pettini, M., Reddy, N. A., & Adelberger, K. L. 2006b, ApJ, 646, 107
Erb, D. K., Steidel, C. C., Shapley, A. E., Pettini, M., Reddy, N. A., & Adelberger, K. L. 2006c, ApJ, 647, 128
Epinat, B., Tasca, L., Amram, P., et al. 2012, A & A, 539, A92
Finkelstein, S. L., Papovich, C., Rudnick, G., Egami, E., Le Floc’h, E., Rieke,
M. J., Rigby, J. R, & Willmer, C. N. A. 2009, ApJ, 700, 376
Fr̈ster Schreiber, N. M., Genzel, R., Lehnert, M. D., et al. 2006, ApJ, 645, 1062
Förster Schreiber, N. M., Genzel, R., Bouché, N., et al. 2009, ApJ, 706, 1364
Förster Schreiber, N. M., Shapley, A. E., Erb, D. K., et al. 2011, APJ, 731, 65
Förster Schreiber, N. M., Shapley, A. E., Genzel, R., et al. 2011, APJ, 739, 45
Gallazzi, A., Charlot, S., et al., 2005, MNRAS, 362, 41
Gallazzi, A., Brinchmann, J., et al., 2008, MNRAS, 383, 1439
Genzel, R., Tacconi, L. J., Eisenhauer, F., et al. 2006, Nature, 442, 786
Genzel, R., Burkert, A., Bouché, N., et al. 2008, ApJ, 687, 59
99
REFERENCES
Genzel, R., Tacconi, L. J., Gracia-Carpio, J., et al. 2010, MNRAS, 407, 2091
Genzel, R., Newman, S., Jones, T., et al., 2011, ApJ, 733, 101
Gnerucci, A., Marconi, A., Cresci, G., et al. 2011, A & A, 528, A88
Heckman, T. M. 2002, Extragalactic Gas at Low Redshift, ed. J. S. Mulchaey &
J. Stocke (San Francisco, AC: ASP), 254, 292
Jones, T. A., Swinbank, A. M., Ellis, R. S., Richard, J., & Stark, D. P. 2010,
MNRAS, 404, 1247
Jones, T., Ellis, R., Jullo, E., & Richard, J. 2010, APJL, 725, L176
Heckman, T. M., Armus, L., & Miley, G. K. 1990, APJS, 74, 833
Kauffmann, G., Heckman, T.M., et al. 2003, MNRAS, 346, 1055
Kennicutt, R. C., Jr. 1992, APJ, 388, 310
Kennicutt, R. C., Jr. 1998, ARA&A, 36, 189
Kewley, L. J. & Ellison, S. L. 2008, ApJ, 681, 1183
Kornei, K. A., Shapley, A. E., Martin, C. L., et al. 2012, APJ, 758, 135
Krist, J. E., Hook, R. N., & Stoehr, F. 2011, Proc. SPIE, 8127
Lara-López, M. A., Cepa, J., Bongiovanni, A., et al. 2010, A & A, 521, L53
Lamareille, F., Contini, T., Le Borgne, J.-F., et al. 2006, A & A, 448, 893
Law, D. R., Steidel, C. C., Erb, D. K., Larkin, J. E., Pettini, M., Shapley, A. E.,
& Wright, S. A. 2007, ApJ, 669, 929
Lehnert, M. D., Nesvadba, N. P. H., Le Tiran, L., et al. 2009, APJ, 699, 1660
Lehnert, M. D., Le Tiran, L., Nesvadba, N. P. H., et al. 2013, arXiv:1304.7734
Le Tiran, L., Lehnert, M. D., van Driel, W., Nesvadba, N. P. H., & Di Matteo,
P. 2011, A & A, 534, L4
Maiolino, R., Nagao, T., Grazian, A., et al. 2008, A&A, 488, 463
Mannucci, F., Cresci, G., Maiolino, R., et al. 2009, MNRAS, 398, 1915
Mannucci, F., Cresci, G., Maiolino, R., Marconi, A., & Gnerucci, A. 2010,
MNRAS, 408, 2115
Martin, C. L. 1999, APJ, 513, 156
Martin, C. L. 2005, APJ, 621, 227
Martin, C. L. 2006, APJ, 647, 222
Martin, C. L., Shapley, A. E., Coil, A. L., et al. 2012, APJ, 760, 127
Moustakas, J., Kennicutt, R. C., Jr., & Tremonti, C. A. 2006, APJ, 642, 775
Nesvadba, N. P. H., Lehnert, M. D., Eisenhauer, F., et al. 2006, APJ, 650, 661
Nesvadba, N. P. H., Lehnert, M. D., Eisenhauer, F., et al. 2006, APJ, 650, 693
Nesvadba, N. P. H., Lehnert, M. D., Genzel, R., et al. 2007, APJ, 657, 725
Nesvadba, N. P. H., Lehnert, M. D., De Breuck, C., Gilbert, A., & van Breugel,
W. 2007, A & A, 475, 145
Newman, S. F., Shapiro Griffin, K., Genzel, R., et al. 2012, APJ, 752, 111
Newman, S. F., Genzel, R., Förster-Schreiber, N. M., et al. 2012, APJ, 761, 43
Newman, S. F., Genzel, R., Förster Schreiber, N. M., et al. 2013, APJ, 767, 104
Osterbrock, D. E. 1989, Astrophysics of gaseous nebulae and active galactic nuclei, University Science Books, Mill Valley, CA USA
Pettini, M., & Pagel, B. E. J. 2004, MNRAS, 348, 59
Pettini, M., Rix, S. A., Steidel, C. C., Adelberger, K. L., Hunt, M. P., & Shapley,
A. E. 2002, ApJ, 569, 742
Pettini, M., Steidel, C. C., Adelberger, K. L., Dickinson, M., & Giavalisco, M.
100
REFERENCES
2000, ApJ, 528, 96
Richard, J., Jones, T., Ellis, R., Stark, D. P., Livermore, R., & Swinbank, M.
2011, MNRAS, 413, 643
Schaerer, D., & de Barros, S. 2009, A&A, 502, 423
Schaerer, D., & de Barros, S. 2010, A&A, 515, 73
Salpeter, E. E. 1955, ApJ, 121, 161
Savaglio, S., Glazebrook, K., Le Borgne, D., et al. 2005, ApJ, 635, 260
Savaglio, S., Panagia, N., & Padovani, P. 2002, ApJ, 567, 702
Shapiro, K. L., Genzel, R., Quataert, E., et al. 2009, APJ, 701, 955
Shapley, A. E., Steidel, C. C., Pettini, M., & Adelberger, K. L. 2003, ApJ, 588,
65
Shapley, A. E., Steidel, C. C., Pettini, M., Adelberger, K. L., & Erb, D. K. 2006,
APJ, 651, 688
Shapley A., 2011, ARA&A, 49, 525
Shapley, A. E., Steidel, C. C., Erb, D. K., et al. 2005, APJ, 626, 698
Schlegel, D. J., Finkbeiner, D. P., & Davis, M. 1998, ApJ, 500, 525
Shirazi, M., & Brinchmann, J. 2012, MNRAS, 421, 1043
Shirazi, M., & Brinchmann, J., Rahmati, A., 2013, arXiv:1307.4758
Schmidt M., 1959, APJ, 129, 243
Siana, B., Teplitz, H. I., Chary, R.-R., Colbert, J., & Frayer, D. T. 2008, APJ,
689, 59
Stark, D. P., Auger, M., Belokurov, V., et al. 2013, arXiv:1302.2663
Steidel, C. C., Adelberger, K. L., Shapley, A. E., et al. 2003, APJ, 592, 728
Steidel, C. C., Erb, D. K., Shapley, A. E., et al. 2010, APJ, 717, 289
Swinbank, A. M., Webb, T. M., Richard, J., et al. 2009, MNRAS, 400, 1121
Teplitz, H. I., McLean, I. S., Becklin, E. E., et al. 2000, ApJ, 533, L65
Tody, D. 1993, Astronomical Data Analysis Software and Systems II, 52, 173
Tremonti, C. A., Heckman, T. M., Kauffmann, G., et al. 2004, APJ, 613, 898
Vegetti, S., & Koopmans, L. V. E. 2009, MNRAS, 392, 945
Verhamme, A., Schaerer, D., Atek, H., & Tapken, C. 2008, A & A, 491, 89
Volino, F., Wucknitz, O., McKean, J. P., & Garrett, M. A. 2010, A & A, 524,
A79
Weiner, B. J., Coil, A. L., Prochaska, J. X., et al. 2009, APJ, 692, 187
Wuyts, E., Rigby, J. R., Sharon, K., & Gladders, M. D. 2012, APJ, 755, 73
Yee, H. K. C., Ellingson, E., Bechtold, J., Carlberg, R. G., & Cuillandre, J.-C.
1996, AJ, 111, 1783
101
Chapter
4
Stars were born in significantly
denser regions in the early
Universe
The density of the warm ionized gas in high-redshift galaxies is known to be higher
than what is typical in local galaxies on similar scales. At the same time, the mean
global properties of the high- and low-redshift galaxies are quite different. Here,
we present a detailed differential analysis of the ionization parameters of 14 star
forming galaxies at redshift 2.6–3.4, compiled from the literature. For each of those
high-redshift galaxies, we construct a comparison sample of low-redshift galaxies
closely matched in specific star formation rate and stellar mass, thus ensuring that
their global physical conditions are similar to the high-redshift galaxy. We find
that the median log[O iii] 5007/[O ii] 3727 line ratio of the high-redshift galaxies
is 0.5 dex higher than their local counterparts. We construct a new calibration
between the [O iii] 5007/[O ii] 3727 emission line ratio and ionization parameter
to estimate the difference between the ionization parameters in the high and lowredshift samples. Using this, we show that the typical density of the warm ionized
gas in star-forming regions decreases by a median factor of 8 from z ∼ 3.3 to
z ∼ 0 at fixed mass and specific star formation rate. We show that metallicity
differences can not explain the observed density differences. Because the highand low-redshift samples are comparable in size, we infer that the relationship
between star formation rate density and gas density must have been significantly
less efficient at z ∼ 2 − 3 than what is observed in nearby galaxies with similar
levels of star formation activity.
Maryam Shirazi, Jarle Brinchmann and Alireza Rahmati
Astrophysical Journal
2013, submitted
Denser star-forming regions in the early Universe
4.1
Introduction
The cosmic star-formation rate, averaged over all observed galaxies in the Universe, has dropped by a factor of > 10 during the last ∼ 10 Gyr (e.g., Hopkins
& Beacom, 2006). In addition to the increasing fraction of actively star-forming
galaxies with increasing look-back time, the star-formation rates of typical galaxies increases rapidly towards the earlier stages of galaxy formation (Noeske et al.,
2007; Daddi et al., 2007; Elbaz et al., 2007, 2011). Several studies also provide
hints that star formation conditions in distant galaxies (i.e., z ∼ 2 − 3) are significantly different from the nearby Universe: emission lines from ionized gas in and
around star-forming regions show different characteristics in distant and nearby
galaxies (Brinchmann et al., 2008b; Liu et al., 2008; Newman et al., 2013), actively star-forming galaxies show higher gas fractions at higher redshifts (Tacconi
et al., 2010; Genzel et al., 2010) and clumpy star-forming disks become increasingly
more prevalent at higher redshifts (Cowie et al., 1995; Elmegreen & Elmegreen,
2006; Genzel et al., 2011). The average density of the warm ionized gas in typical high-redshift (high-z) galaxies is also known to be significantly higher than
in typical low-redshift (low-z) galaxies on similar scales (Elmegreen et al., 2009;
Lehnert et al., 2009; Le Tiran et al., 2011; Newman et al., 2012; Tacconi et al.,
2013; Lehnert et al., 2013).
These studies have revealed that distant star-forming galaxies form a population of objects that are distinct from their nearby analogs. However, it is unclear
whether the main difference between low-z and high-z star-forming galaxies is related to their strongly evolving global properties, such as stellar mass (e.g., Ilbert
et al., 2013; Muzzin et al., 2013), star formation rate (e.g., Noeske et al., 2007;
Daddi et al., 2007; Elbaz et al., 2007, 2011) or metallicity (e.g., Mannucci et al.,
2010; Lara-López et al., 2010), or that the interstellar medium (ISM) conditions
were significantly different in similar galaxies at high-z. Comparing representative
samples of high-z and low-z star-forming galaxies (e.g. Rigby et al., 2011) cannot
disentangle the evolution in global characteristics from the possibly evolving starformation conditions. We address this issue by selecting a comparison ensemble
of low-z galaxies for each high-z star-forming galaxy in our sample, ensuring that
the stellar mass and star-formation activities are similar in our high-z galaxies
and their low-z comparison samples. This allows us to evaluate the differences in
star-formation conditions between the high-z star-forming galaxies and their local
analogs.
Although observations of some lensed galaxies at high-z reach spatial resolutions of ∼100 pc (e.g., Swinbank et al., 2009; Jones et al., 2010), even this spatial
resolution is insufficient to directly compare the small-scale properties of the ISM
in high-z and low-z star-forming galaxies. However, these properties can be constrained through their impact on the emission line spectra of galaxies (e.g., Yeh &
Matzner, 2012). Here we use emission line ratios to derive the average ionization
parameter of star-forming regions. Since the ionization parameter is a measure of
ionizing radiation intensity per unit density, we can use it to constrain the density
of star-forming regions in distant galaxies and compare it with that of their nearby
counterparts.
104
Data
The structure of the paper is as follows. In Section 4.2 we introduce our highz sample and explain how we select their low-z counterparts. In Section 4.3 we
introduce our new calibration for calculating the ionization parameter using the
emission line ratios. We present our main results in Section 4.4 and compare the
density of ionized gas in high-z and nearby galaxies. In Section 4.5 we investigate
the impact of metallicity variations between the high-z and local galaxy samples
on our results. We discuss the implications of our finding in Section 5.5 and end
the paper with concluding remarks in Section 4.7.
4.2
Data
We have assembled a sample of 14 high-z star-forming galaxies from the literature for which published [O ii] λ3727, [O iii] λ5007 and Hβ emission line fluxes
are available (they have [O iii] λ5007/Hβ > 0). This sample consists of 2 galaxies
(RXJ1053, Cl0949) from Richard et al. (2011, R11); 7 galaxies from the AMAZE
sample (Maiolino et al., 2008, M08); 4 galaxies from the LSD sample (Mannucci
et al., 2009, M09), and the 8 o’clock arc (Dessauges-Zavadsky et al., 2011; Shirazi
et al., 2013, arc). These galaxies span redshifts between z = 2.39 and z = 3.69 with
a median redshift of z = 3.39. All these galaxies also have gas metallicity, stellar
mass and star-formation rate estimates. To test our results further, we also use
a sample of 3 galaxies in the SINS survey (Förster Schreiber et al., 2009, 2011)
that have directly measured electron densities using [S ii] doublet (Lehnert et al.,
2009). The physical properties of our high-z sample are summarized in Table 4.1.
We compare these galaxies to matched samples of low-z galaxies from the Sloan
Digital Sky Survey (SDSS) (York et al., 2000). We used the MPA-JHU1 value added catalogues (Brinchmann et al., 2004; Tremonti et al., 2004) for SDSS DR7
(Abazajian et al., 2009) as our parent sample and selected star-forming galaxies
following Brinchmann et al. (2004), with the adjustments of the line flux uncertainties given in Brinchmann et al. (2013). Furthermore, we used SDSS DR8 (Aihara
et al., 2011) photometry to estimate stellar masses. The median and 1-σ scatter of
the physical properties of the low-z sample of each high-z galaxy are summarized
in Table 4.2.
As argued above, it is essential to take out correlations with global properties
of galaxies when comparing their ISM conditions. To achieve this we select, for
each high-z galaxy, all star-forming galaxies in the SDSS DR7 that have log M∗
and log SFR/M∗ within 0.3 dex of that of the high-z galaxy. We require that the
SDSS galaxies to have z > 0.02 so that [O ii] λ3727, 29 are measured, they also
have [O iii] λ5007/Hβ > 0. We note that for two galaxies in our high-z sample
that have very high sSFR (∼ 0.3 Gyr−1 ), we had to include low-z galaxies whose
log SFR/M∗ differ by up to 1 dex to find a local analog sample. By default, we do
not explicitly constrain the low-z samples to match the metallicity and/or size of
their high-z counterparts as this would reduce the size of our sample and in the
case of metallicity is subject to systematic uncertainties (e.g. Kewley & Ellison,
2008). However, as we show below, matching metallicities and/or sizes does not
1 http://www.mpa-garching.mpg.de/SDSS/DR7
105
Denser star-forming regions in the early Universe
affect our results significantly.
Any significant contribution of ionizing radiation from an Active Galactic Nucleus (AGN) could bias our estimates of the ionization parameter. For the low-z
sample we can exclude strong AGN using the BPT diagram (Baldwin, Phillips
& Terlevich, 1981). At high-z, the galaxies from M08 and R11 do not show any
evidence indicating the presence of AGNs in their rest frame UV spectra (i.e.,
[N v] ,[C iv] ,He ii or broad Lyα), X-ray and 24 µm Spitzer-MIPS observations
(Maiolino et al., 2008; Richard et al., 2011; Shirazi et al., 2013). The LSD galaxies
also show no evidence of AGN activity in X-ray observations (Mannucci et al.,
2009). While the aforementioned arguments do not rule out the presence of some
AGN activity that is optically thick for X-rays, this is unlikely to significantly influence the optical emission lines which originate in only moderately obscured regions.
One galaxy from the SINS sample (Q2343-BX610) that we use in this study has
an indication of possible AGN from mid-IR observations (Förster Schreiber et al.,
2011; Hainline et al., 2012) and from an analysis of resolved spectroscopy presented
by Newman et al. (2013). However, we note that we do not use our calibration
to infer ionization parameter for the SINS galaxies. Thus, we conclude that AGN
activity is unlikely to bias our results at high-z.
4.3
Methodology
The high-z galaxies all have measured [O iii]λ5007 and [O ii]λ3727 line fluxes. This
allows us to use the strong sensitivity of the [O iii] λ5007/[O ii] λ3727 (hereafter
O32) ratio to the ionization parameter (Penston et al., 1990) to estimate this.
Kewley & Dopita (2002) derived an estimator for the ionization parameter using
the dereddened O32 ratio. Since this can not easily be applied to our high-z
sample in the absence of reliable reddening estimates, we here calibrate a new
relation between the ionization parameter and the observed O32 ratio using the
Charlot & Longhetti (2001, hereafter CL01) models that account for variations
in dust properties and metallicities (see Table 4 in Shirazi & Brinchmann (2012)
for the CL01 model grid used in our study). The effective ionization parameter
in these models is taken to be the volume average over the Strömgren sphere (see
equation 9 and 10 in CL01).
We wish to construct a calibration between the ionization parameter and O32
ratio that treats the metallicity as a free parameter. Based on this approach, as
long as our high-z and low-z samples do not differ greatly in metallicity we do not
need to know this exactly. We discuss this assumption further below, but given
that we still have several possible ways to construct the calibration from the CL01
models:
a- Leaving all parameters in the CL01 models as free parameters in the fitting
procedure (including all dust attenuation parameters, 0.01 < τV < 4 ). This is
likely to give a large amount of scatter in the relationship.
b- Using only models with τV ∼ 0.2 and leaving all other model parameters
free. This fit is appropriate if line ratios are corrected for dust attenuation but
there is no constrain on the dust-to-metal ratio (ξ).
106
Results
c- Using only models with τV ∼ 0.2 and ξ ∼ 0.3 (i.e., the Galactic dust-to-metal
ratio) and leaving all other model parameters free. Since ξ is expected to evolve
weakly with time (Calura et al., 2008), it is reasonable to fix its value.
d- Using ξ ∼ 0.3 and leaving all other model parameters free. Since ξ is likely
not to differ strongly from this value, this is the best choice for a calibration when
the amount of dust attenuation is unknown.
These fits are plotted in Figure 4.1 from the top-left to the bottom-right,
respectively. The best fits for the relation between ionization parameter and
Log [O iii] λ5007/[O ii] λ3727 (Log O32) are summarised as equation 1 to 4, respectively. We use option d, Equation 4.4, as our reference in this study because
in general we do not have enough information to accurately constrain the dust
attenuation for the high-z galaxy sample. To derive our reference relation we fix
ξ = 0.3, which is the Galactic value (see Brinchmann et al., 2013, for a discussion),
and allow all other parameters to vary. We use the same fit for estimating the
ionization parameter for low-z counterparts of high-z galaxies.
Log U = −3.300 ± 0.017 + (0.481 ± 0.019) Log O32
(4.1)
Log U = −3.109 ± 0.039 + (0.586 ± 0.039) Log O32
(4.2)
Log U = −3.119 ± 0.027 + (0.804 ± 0.035) Log O32
(4.3)
Log U = −3.363 ± 0.011 + (0.593 ± 0.012) Log O32
(4.4)
We are primarily focused on relative statements in this work so the most important aspect of these calibrations is how they convert relative statements in O32
to relative statements about log U. Since the slope in equation 1, 2 & 4 is similar
they will result in similar relative statements about log U, while that in equation 3
is even steeper and would lead to an even stronger result than that outlined below.
4.4
Results
The left and middle panels of Figure 4.2 compare the [O iii] λ5007/[O ii] λ3727
ratios and corresponding ionization parameters (from Equation 4.4) for our highz sample (colored symbols), and the median values of their low-z analogs (black
circles). Error bars shown on the black circles indicate 1-σ scatter in the low-z
sample of each high-z galaxy. It is evident that the high-z star-forming galaxies
show significantly higher [O iii] λ5007/[O ii] λ3727 ratios (up to ≈ 0.8 dex higher)
compared to their low-z analogs. This translates into significantly higher ionization
parameters (up to ∼ 0.5 dex) in the high-z galaxies relative to low-z even though
their star formation rates and masses are constrained to be the same.
For a given production rate of hydrogen ionizing photons, Q, and after assuming
that most of ionizing photons are absorbed locally, the ionization parameter in a
typical ionized region can be related to the hydrogen number density, nH :
U3 ∝ Q(t) nH 2 ,
(4.5)
where is the volume filling factor of the ionized gas, which is defined as the
ratio between the volume-weighted and mass-weighted average hydrogen densities
107
Denser star-forming regions in the early Universe
-2.0
a
-2.0
-2.5
Log U
Log U
-2.5
-3.0
-3.5
-4.0
-4.0
-2
-1
0
1
2
Log [O III]5007/[O II]3727
-2
-1
0
1
2
Log [O III]5007/[O II]3727
c
-2.0
-3.0
-3.0
-3.5
-3.5
-4.0
-4.0
-2
-1
0
1
2
Log [O III]5007/[O II]3727
d
-2.5
Log U
-2.5
Log U
-3.0
-3.5
-2.0
b
-2
-1
0
1
2
Log [O III]5007/[O II]3727
Figure 4.1 Best-fit relations between ionization parameters and the
[O iii] λ5007/[O ii] λ3727 (O32) ratios are shown by blue dashed lines. The
top-left panel shows the best-fit using all CL01 models (0.01 < τV < 4), on the
top-right we show the best-fit using only models with τV ∼ 0.2, the bottom-left
panel shows the best-fit using only models with τV ∼ 0.2 and ξ ∼ 0.3 (i.e., Galactic
dust-to-metal ratio), and in the bottom-right panel we show the best-fit to all
models with ξ ∼ 0.3. The results in the paper are presented for the fit shown in
the lower-right panel.
(Charlot & Longhetti, 2001). This allows us to constrain the densities of starforming regions, by measuring their ionization parameters.
Assuming that the production rate of hydrogen ionizing photons and volume
filling factors of the ionized gas are similar in typical star-forming regions in high-z
galaxies and their low-z analogs, one can translate the ratio between the ionization
parameters of the high-z galaxies and their low-z counterparts into the ratio of
their ionized gas densities. The difference between the density of the ionized gas
in star-forming regions in our high-z galaxies and their low-z analogs is shown in
the right panel of Figure 4.2. This shows up to ≈ 25 times higher densities in
high-z star-forming galaxies.
To derive physical densities for our high-z galaxies from the relative density
differences shown in Figure 4.2, we exploit the fact that for the nearby galaxies
we can estimate the electron density from the [S ii] λ6716, 6731 ratio and thus get
an estimate of the electron density in the high-z galaxies. The resulting absolute
densities for the ionized gas in our high-z star-forming galaxies are shown in Figure 4.3. The median values of the electron densities of the low-z samples, inferred
108
Metallicity dependence
from the [S ii] 6716, 6731 doublet, are shown by the black filled circles in the figure
where error bars show 1-σ scatter. The median values of the redshifts of the low-z
samples and the number of low-z analogs in each sample are indicated with n in
the figure. Colored symbols show our high-z sample with their redshifts indicated.
The high-z values are inferred from the low-z values multiplied by ne (z)/ne (z = 0)
ratios shown in Figure 4.2, and their error bars show propagation of uncertainties
based on Equation 4.4. The grey small dashed and long dashed lines show the
median values for the electron density at low-z and high-z, respectively.
Besides the sensitivity of the ionization parameter to the density of the emitting gas, it also depends on the production rate of ionizing photons and the volume
filling factor of the ionized gas (Charlot & Longhetti, 2001). Therefore, our density
estimates might also be sensitive to the possible differences in the ionizing photons
production rate and the volume filling factor of the ionized gas between high-z and
nearby galaxies. To address this concern, in Figure 4.3 we show electron densities for a sample of five high-z star-forming galaxies in the SINS survey (Förster
Schreiber et al., 2009; Lehnert et al., 2009) as purple diamonds. The electron
density for these galaxies has been measured directly using the [S ii] λ6716, 6731
doublet and is in a good agreement with our inferred evolution in density estimated
from the ionization parameter. For three of these five objects that have available
stellar masses and specific star formation rates (Förster Schreiber et al., 2011), we
constructed low-z analog samples. The comparison between the electron density of
these three objects and their low-z analogs also shows good agreement (evolution
in density with a median factor of 8.4) with the density ratios we obtained for
our high-z star-forming galaxies using their ionization parameters (an evolution in
density with a median factor of 7.9). This further strengthens our argument that
an elevated density of star-forming regions in high-z galaxies is the main reason
for their higher ionization parameter.
4.5
Metallicity dependence
A key result in this work is that high-z galaxies have a typically 0.5 dex higher
Log O32 than low-z galaxies with the same mass and sSFR. We interpret this
as primarily being due to a difference in ionization parameter but O32 is also
sensitive to metallicity. Ideally we would select our high-z and low-z samples to
have the same metallicity but to do this we require a metallicity estimator that
can be applied equally at low-z and high-z allowing for a variation in ionization
parameter. With the current data available for high-z galaxies this is not possible,
thus we need to assess whether metallicity differences between the samples could
be the reason for the observed offset.
Mannucci et al. (2010) and Lara-López et al. (2010) showed that there is a
relationship between stellar mass, metallicity and star formation rate (SFR) that
appear to hold to high-z (z < 2.5 for Mannucci et al. and z < 3.5 for Lara-Lopez et
al.). Therefore, if this holds for our galaxies, a selection on stellar mass and SFR
should ensure that the metallicity difference between the high- and low-z sample is
small. Given our small sample and considering that Mannucci et al. (2010) argued
109
Denser star-forming regions in the early Universe
Figure 4.2 A comparison between [O iii] λ5007/[O ii] λ3727 ratio, ionization parameter and electron density at low-z and high-z. The x-axis on the left panel shows
the [O iii]λ5007/[O ii]λ3727 ratio, the middle panel shows the ionization parameter
and the right panel shows the electron density at high-z relative to that of lowz. Colored symbols show high-z galaxies with their redshift indicated and black
circles show the median values for the low-z sample of each high-z galaxy. Error
bars span from the 16% to the 84% confidence level. We see that high-z galaxies
show higher [O iii] λ5007/[O ii] λ3727 ratios than their low-z analogs (up to ≈ 0.8
dex higher), even though their masses and sSFR are the same. The middle panel
shows the ionization parameters derived using our new calibration between the
[O iii] λ5007/[O ii] λ3727 ratio and the ionization parameter. We see that high-z
galaxies show up to ∼ 0.5 dex higher (median ∼ 0.3 dex) ionization parameters
than their low-z analogs. This translates to up to 25 times higher electron density
for high-z galaxies relative to their low-z analogs.
that the multi-parameter relationship was not well established at z > 2.5, where
most of our high-z galaxies lie, it is necessary to examine this assumption more
carefully. It is useful to start this by asking what metallicity difference would give
a O32 difference similar to what is observed. From Brinchmann et al. (2008b, their
Figure 8), or directly using the CL01 models, we find that a change in metallicity
from 1 Z to 0.1 Z leads to a change in Log O32 of 0.40 ± 0.07 dex. Thus we need
a major difference in metallicity to explain the results.
We can test for a large offset in metal content by calculating the metallicities
of the high- and low-z samples in a consistent way. To do this we adopt the
methodology used for AMAZE and LSD described in (Maiolino et al., 2008) for
both high- and low-z galaxies. Note that, by construction, this method assumes
that all variation in O32 is due to metallicity. Therefore, by using it, we will
maximize the contribution of metallicity to the change in O32 and hence derive a
minimum difference in ionization parameter between the low- and high-z objects.
110
Metallicity dependence
Figure 4.3 The median value of the electron density for the low-z samples inferred
from the [S ii] 6716, 6731 doublet is shown by the black filled circles. The highz values are inferred from the low-z values multiplying by ne (z)/ne (z = 0) ratios
shown in Figure 4.2. Colored symbols show our high-z sample sorted based on
their redshifts from down to top as indicated on the figure. Five galaxies from the
SINS survey that have directly measured electron densities are shown by purple
diamonds. The median values of the redshifts of low-z samples are shown in black
and the number of low-z analogs in each sample are indicated with n. Error bars
span from the 16% to the 84% confidence level (low-z data: they show scatter in
the sample, high-z data: they show propagation of uncertainties through Equation
4.4). The grey small dashed and long dashed lines show the median value for the
electron density at low-z and high-z, respectively.
Based on the derived metallicities, we can calculate the maximum difference in
O32 between high- and low-z galaxies due to metallicity differences, using the
CL01 models and by averaging over U. This gives us the expected change in O32
due to metallicity only, and we subtract this off the actually observed difference
for each galaxy.
The resulting difference can be seen in the top panel of Figure 4.4. We emphasize that since we have used an abundance calibration that assumes that changes
in O32 are due to metallicity, this correction should be the maximum possible
correction. This gives a lower limit to the difference in O32 between high- and
low-z galaxies and it is still quite sizeable. Converting this to a density difference
as done in the main text we get the lower panel in that figure. This shows that
the mean (median) electron density of the high-z galaxies is 5.5 (3.5) times higher
than the low-z galaxies with the same sSFR and mass.
111
Denser star-forming regions in the early Universe
Min O32(z)-O32(z=0)
1.0
0.8
0.6
0.4
0.2
Min n(z)/n(z=0)
0.0
8
6
4
2
0
2.0
2.5
3.0
Redshift
3.5
4.0
Figure 4.4 Top panel: The minimum difference in O32 between the high- and lowz samples when corrected for metallicity as described in the text. Bottom panel:
The resulting minimum density difference between the high- and low-z samples
when corrected for metallicity.
To test further the sensitivity of our results to metallicity differences between
our high-z galaxies and their low-z analogs, we made a low-z comparison sample
for all high-z galaxies ensuring that their metallicities were equal to within 0.3
dex, in addition to matching their stellar masses and sSFRs2 . In this case we
found that high-z galaxies show a median of ≈ 6.1 higher density compared to
their low-z analogs with similar sSFRs, masses and metallicities; a result which is
not significantly different from what we found without matching metallicities.
We also note that the densities that are measured directly from the [S ii]
doublet for the 5 high-z galaxies we selected from the SINS, are not derived using
our calibration and hence are insensitive to variations in metallicity. Yet they
have densities which are on average 8.4 times higher than their local analogs. It
also worth noting that not all SINS galaxies have detected [S ii] which is consistent with these conclusions because [S ii]/Hα decreases with increasing U at fixed
metallicity (e.g., Brinchmann et al., 2008a, their Figure 11).
In conclusion, regardless of how we correct for possible differences in metallicity
between the high- and low-z samples, the effect is minor and the main result of
the paper is robust to these corrections. Thus we conclude that differences in
metallicity can not explain the observed major offset in O32 and that systematic
differences in the ionization parameter is the main cause.
2 Note
112
that this additional metallicity constraint decreases the low-z sample sizes.
Discussion
4.6
Discussion
The observed strong evolution in the global properties such as star formation
intensity, stellar mass and size indicates that mean star formation conditions are
different in distant galaxies compared to typical galaxies today (Cowie et al., 1995;
Elmegreen & Elmegreen, 2006; Noeske et al., 2007; Daddi et al., 2007; Elbaz et al.,
2007, 2011; Tacconi et al., 2010; Genzel et al., 2010, 2011). In this work we have
however shown that even when the star formation intensity and mass are the same,
the density in the ionized gas in high- and low-z galaxies differ dramatically. This
difference would naturally imply a higher pressure in the colder ISM surrounding
the ionized gas (Dopita et al., 2006), and thence its higher density.
This could naturally occur if star formation at high-z was more concentrated
to the central regions, so to check this we compared the u-band half-light radius
for the SDSS galaxies with the half-light radius of the high-z galaxies when they
are available (for 7 galaxies). Among the high-z galaxies, only one has a smaller
size than the median size of its low-z counterparts. This is in agreement with
the findings of Lehnert et al. (2009) and can not explain the density differences
seen for any reasonable mass profile in the galaxies. We double-checked this by
2
constructing a matched low-z sample that have log SFR/πr1/2
within 0.3 dex of their
high-z counterparts where r1/2 is the half-light radius. This results in a median
density difference greater than 19 between low-z and high-z galaxies compared to
a median difference of ∼ 8 before matching SFR densities. This shows that size
differences are unlikely to be the explanation of the systematic differences. We
have not required a match in SFR density in the bulk of the paper. However,
because the size definitions are somewhat arbitrary and we do not have sizes for
all galaxies at high-z.
Assuming now that the distribution of star formation is comparable at low- and
high-z, we next assume that the H ii regions are in pressure equilibrium with their
surrounding ISM (Oey & Clarke, 1997; Dopita et al., 2006). Under this assumption
the increased pressure in the ionized regions implies a higher pressure in the cold
ISM. There are considerable uncertainties in how ionized regions expand in detail.
However, in our case it is not unreasonable to assume that those complexities
should be similar at high and low redshift. This is because the evolution of the
H ii regions is driven by the energy injection from massive stars which should be
similar at high-z and low-z, given how we selected our samples. The same applies
to cosmic ray production rates which contribute to the heating (and the pressure
support) of the ambient ISM. Note that this also means that the contribution of
radiation pressure to the equilibrium for the H ii regions (e.g., Yeh & Matzner,
2012) should be similar at low and high redshift.
It is hard to test whether the H ii regions in the high-z galaxies have reached
pressure equilibrium with their surrounding ISM. However, since the mechanical
input energy is the same at high and low redshift, and the life-times of the relevant
stellar population is also the same, it seems unlikely that the evolutionary age of
the H ii regions differs significantly between the high-z and low-z samples. This
is also supported by Verdolini et al. (2013) who used a population study to show
that the line emission of a galaxy will typically be dominated by the youngest
113
Denser star-forming regions in the early Universe
H ii regions. Verdolini et al. (2013) also show explicitly the effect of an elevated
ambient pressure on emission line ratios (their Figure 8), which is a qualitatively
similar trend to what we infer here.
Thus, the simplest explanation for the elevated density in the high-z H ii regions
is an elevated pressure in the cold ISM relative to similar galaxies nearby. This
increased pressure could arise from various sources, but in general, one would
expect a pressure-density relation, P ∝ ργ with γ > 1. Thus, the increased pressure
would correspond to an increased ISM density by an amount that depends on
the model adopted for the ISM and we do not attempt to discuss this in detail
here. The simplest model, where the ISM temperature is the same at high and low
redshift, would predict that the density difference between the ISM at high- and
low-z would be the same as that of the the ionized regions, i.e., ρhigh−z ∼ 8 ρlow−z .
This conclusion has important implications for empirical star-formation law as
well. The most popular scaling relation observed between star formation activity
and gas surface density in the local Universe is the Kennicutt-Schmidt relation
(Kennicutt, 1998),
ΣSFR ∝ Σ1.4
(4.6)
gas ,
where Σ denote surface densities. In our case ΣSFR is approximately the same in
the high- and low-z galaxies (see above), but the gas density is much higher. If
the scale-height of the gas is not significantly smaller in high-z galaxies, one can
conclude that the scaling relation in the high-z galaxies is significantly different
from what is observed in their low-z counterparts, being a factor ∼ 5–8 less efficient.
We note however that we can not distinguish between molecular and atomic
gas. Our results therefore are for the total gas and we cannot directly compare
them to molecular studies (e.g., Daddi et al., 2010; Tacconi et al., 2013) at high-z
and leave a discussion of this for future work.
4.7
Conclusion
In this work we compare the physical conditions of the ISM in high-z galaxies and
their low-z counterparts which are selected to have similar global properties as that
of high-z galaxies. This selection criteria minimize the differences between distant
and nearby galaxies due to the evolution of the global properties such as mass and
sSFR from high-z to low-z and can therefore be used to study the evolution of
intrinsic properties of the ISM.
Previous studies have already pointed out that the physical densities/properties
of the star-forming regions at high-z are very different from those in the local
Universe and we confirm this here. Using a novel approach, we have been able to
go one step further, and show that this difference can not fully be explained by an
increased star formation activity in the high redshift galaxies. Since we compare
high and low-z galaxies that are matched in sSFR, their different densities must
reflect an intrinsic difference in ISM conditions between high and low-z. We argue
that this difference is primarily due to a difference in the density of the warm
ionized gas. We have also shown that the differences between the high- and lowz galaxies can not be explained by differences in metallicity. By showing that
114
Conclusion
the high-z and low-z samples are also comparable in size, we conclude that the
relationship between star formation rate density and gas density must have been
significantly less efficient at z ∼ 2 − 3 than what is observed locally. This, in
turn, implies that most of the stars in the local Universe were formed following a
different star formation scaling relation than what is observed in normal galaxies
today.
Acknowledgments
We thank Marijn Franx for helpful discussion. We also thank Leslie Sage and the
two anonymous referees of Nature for very useful comments which improved this
paper.
115
116
RXJ1053
Cl0949
SSA22a-C30
Q0302-C131
Q0302-M80
Q0302-C171
CDFa-C9
CDFS-4414
CDFS-4417
CDFS-16767
CDFS-2528
SSA22a-M38
SSA22a-aug16M16
8oclock
Q2343-BX389
Q2343-BX610
Q2346-BX482
Name
Table 4.1 High-z sample.
R11
R11
M09
M09
M09
M09
M08
M08
M08
M08
M08
M08
M08
arc
SINSa
SINSa
SINSa
ID
2.576
2.394
3.103
3.235
3.414
3.328
3.212
3.471
3.473
3.624
3.688
3.294
3.292
2.735
2.172
2.210
2.256
z
Log Mass
M
9.620.75
−0.72
10.190.22
−0.18
10.330.31
−0.38
10.090.10
−0.33
10.070.23
−0.19
10.060.10
−0.28
10.180.40
−0.08
10.570.19
−0.22
10.290.37
−0.11
10.050.10
−0.16
9.760.09
−0.07
11.010.18
−0.41
10.290.20
−0.21
10.24−1.80
0.45
10.610.77
−2.16
11.002.70
−0.60
10.260.79
−0.46
Log sSFR
yr−1
-8.66
-9.31
-8.87
-9.09
-8.96
-9.36
-7.76
-8.52
-7.65
-8.13
-7.76
-8.95
-8.67
-7.88
−9.22
−9.22
−8.36
SFR
M yr−1
9.12.3
−2.3
7.51.5
−1.5
29.081.0
−21.0
10.06.0
−4.0
13.017.0
−8.0
2.0
5.0−2.0
265.00.0
0.0
113.00.0
0.0
438.00.0
0.0
84.00.0
0.0
101.00.0
0.0
115.00.0
0.0
0.0
42.00.0
10.0
228.0−10.0
···
···
···
ΣS FR
M yr−1 kpc−2
0.22
0.19
4.21
1.97
7.36
1.02
···
···
···
···
···
···
···
9.26
···
···
···
r1/2
kpc
3.62± 0.45
3.50± 0.88
1.48± 0.44
1.27± 0.37
0.75± 0.24
1.25± 0.39
···
···
···
···
···
···
···
2.80± 0.20
···
···
···
12 + LogO/H
8.680.11
−0.12
8.100.06
−0.05
8.160.20
−0.60
8.000.25
−0.40
8.360.15
−0.15
8.140.25
−0.45
8.100.18
−0.21
8.540.15
−0.14
8.550.09
−0.10
8.310.11
−0.17
8.070.39
−0.28
8.340.15
−0.12
7.990.26
−0.34
8.350.19
−0.19
···
···
···
Log O32
0.5890.0
−0.0
0.4070.0
−0.0
0.6300.2
−0.0
0.5150.0
−0.4
0.3720.1
−0.1
0.2930.0
−0.2
0.5000.0
−0.0
0.0380.0
−0.0
0.2330.1
−0.0
0.5800.1
−0.1
0.4460.2
−0.0
0.1880.1
−0.0
0.5640.3
−0.2
0.661
···
···
···
−3.010.02
−0.02
−3.120.02
−0.02
−2.990.02
−0.02
−3.060.02
−0.02
−3.140.02
−0.02
−3.190.01
−0.01
−3.070.02
−0.02
−3.340.01
−0.01
−3.220.01
−0.01
−3.020.02
−0.02
−3.100.02
−0.02
−3.250.01
−0.01
−3.030.02
−0.02
−2.970.02
−0.02
···
···
···
Log U
ne
cm−3
897.31145.2
−549.2
1229.41713.1
−789.8
2766.62286.8
−1623.8
2325.3
1682.9−1044.2
796.2946.5
−483.8
765.11192.4
−495.7
912.61278.7
−489.9
158.8179.9
−109.5
342.0448.9
−268.9
840.71746.3
−425.0
406.5237.1
−319.3
978.51862.6
−588.7
1698.71698.0
−934.8
391.5310.0
−310.0
1200.0700
−400
400.0700
−300
1200.0700
−400
Denser star-forming regions in the early Universe
0.150.07
−0.07
0.150.07
−0.05
0.210.05
−0.08
0.160.07
−0.05
0.170.06
−0.06
0.130.07
−0.05
0.200.07
−0.08
0.260.03
−0.10
0.190.09
−0.10
0.230.04
−0.05
0.240.04
−0.09
0.270.02
−0.09
0.220.05
−0.09
0.270.00
−0.22
0.200.05
−0.08
0.250.02
−0.09
0.220.06
−0.10
h z ia
Table 4.2 Low-z sample.
RXJ1053
Cl0949
SSA22a-C30
Q0302-C131
Q0302-M80
Q0302-C171
CDFa-C9
CDFS-4414
CDFS-4417
CDFS-16767
CDFS-2528
SSA22a-M38
SSA22a-aug16M16
8oclock
Q2343-BX389
Q2343-BX610
Q2346-BX482
High-z ID
hLog Massi
M
9.590.20
−0.19
10.040.19
−0.11
10.180.19
−0.11
0.21
9.96−0.12
0.21
9.94−0.13
9.930.20
−0.12
10.050.15
−0.13
10.420.21
−0.10
10.100.20
−0.08
0.17
9.90−0.11
0.16
9.48−0.02
10.860.23
−0.12
10.120.19
−0.10
10.030.00
−0.07
10.440.17
−0.10
10.840.27
−0.11
10.070.13
−0.06
hLog sSFRi
yr−1
-8.81
-9.32
-9.03
-9.19
-9.08
-9.37
-8.4
-8.70
-8.0
-8.38
-7.93
-9.02
-8.84
-8.04
-9.27
-9.27
-8.55
hLog SFRi
M yr−1
0.90.2
−0.3
0.80.3
−0.3
1.20.2
−0.2
0.80.2
−0.3
0.90.2
−0.2
0.60.3
−0.3
1.742.2
−0.2
1.60.4
−0.2
1.902.0
−0.3
1.60.2
−0.1
0.1
1.7−0.3
1.80.3
−0.3
1.40.2
−0.2
1.80.0
−2.0
1.20.2
−0.2
1.50.2
−0.1
1.60.1
−0.1
hΣS FR i
M yr−1 kpc−2
1.91
0.93
3.70
1.35
1.86
0.53
17.63
12.18
25.71
14.91
22.25
3.72
6.25
17.63
2.49
5.23
12.15
hr50u−band i
kpc
1.110.41
−0.24
1.460.65
−0.42
1.220.32
−0.27
1.320.54
−0.33
1.240.41
−0.29
1.530.84
−0.46
0.950.33
−0.95
1.090.43
−0.17
0.910.39
−0.91
1.060.21
−0.29
0.830.34
−0.14
1.893.59
−0.72
1.170.28
−0.24
3.010.00
−1.95
1.400.51
−0.36
1.550.65
−0.52
1.080.27
−0.20
0.0120.163
−0.161
−0.2600.118
−0.088
−0.1570.138
−0.101
−0.2230.137
−0.100
−0.1740.143
−0.115
−0.2810.125
−0.090
0.0480.146
−0.364
−0.0450.199
−0.128
0.0190.096
−0.316
0.092
0.129−0.068
0.3090.047
−0.130
−0.2310.277
−0.029
−0.0820.149
−0.119
0.000
0.236−0.747
−0.2050.113
−0.085
−0.1970.140
−0.061
0.0840.110
−0.087
hLog O32i
h12 + LogO/Hi
M08 calib b
8.490.12
−0.05
8.700.05
−0.07
8.650.07
−0.10
8.670.07
−0.09
8.640.08
−0.10
8.710.05
−0.08
8.440.08
−8.44
8.570.13
−0.08
8.440.13
−8.44
8.450.10
−0.01
8.410.05
−0.09
8.710.04
−0.14
8.600.08
−0.12
8.780.00
−0.40
8.700.04
−0.08
8.710.04
−0.09
8.470.09
−0.03
−3.360.10
−0.10
−3.520.07
−0.05
−3.460.08
−0.06
−3.500.08
−0.06
−3.470.09
−0.07
−3.530.07
−0.05
−3.330.09
−0.22
−3.390.12
−0.08
−3.380.09
−0.16
−3.290.05
−0.04
−3.180.03
−0.08
−3.500.16
−0.02
−3.410.09
−0.07
−3.440.22
−0.22
−3.480.07
−0.05
−3.480.08
−0.04
−3.310.07
−0.05
hLog Ui
hne i
cm−3
84.6107.9
−51.8
80.0111.5
−51.4
110.291.1
−64.7
81.8113.1
−50.8
100.9
84.9−51.6
73.0113.8
−47.3
143.1200.6
−76.8
112.9128.0
−77.9
142.3186.9
−111.9
132.6275.4
−67.0
232.1135.4
−182.3
335.1
176.0−105.9
120.5120.4
−66.3
68.754.4
−54.4
112.5127
−68
153.5228
−94
143.1100
−65
Conclusion
117
REFERENCES
References
Abazajian, K. N., Adelman-McCarthy, J. K., Agüeros, M. A., et al. 2009, ApJS,
182, 543
Aihara, H., Allende Prieto, C., An, D., et al. 2011, ApJS, 193, 29
Baldwin, J. A., Phillips, M. M., & Terlevich, R. 1981, PASP, 93, 5
Brinchmann, J., Charlot, S., Kauffmann, G., et al. 2013, MNRAS,
doi:10.1093/mnras/stt551
Brinchmann, J., Charlot, S., White, S. D. M., et al. 2004, MNRAS, 351, 1151
Brinchmann, J., Kunth, D., & Durret, F. 2008a, A&A, 485, 657
Brinchmann, J., Pettini, M., & Charlot, S. 2008b, MNRAS, 385, 769
Calura, F. and Pipino, A. and Matteucci, F., 2008, A&A, 479, 669
Charlot, S., & Longhetti, M. 2001, MNRAS, 323, 887
Cowie, L. L., Hu, E. M., & Songaila, A. 1995, AJ, 110, 1576
Daddi, E., Dickinson, M., Morrison, G., et al. 2007, ApJ, 670, 156
Daddi, E., Elbaz, D., Walter, F., et al. 2010, ApJL, 714, L118
Dessauges-Zavadsky, M., Christensen, L., D’Odorico, S., Schaerer, D., & Richard,
J. 2011, A&A, 533, A15
Dopita, M. A., Fischera, J., Sutherland, R. S., et al. 2006, ApJ, 647, 244
Elbaz, D., Daddi, E., Le Borgne, D., et al. 2007, A&A, 468, 33
Elbaz, D., Dickinson, M., Hwang, H. S., et al. 2011, A&A, 533, A119
Elmegreen, B. G., & Elmegreen, D. M. 2006, ApJ, 650, 644
Elmegreen, D. M., Elmegreen, B. G., Marcus, M. T., et al. 2009, ApJ, 701, 306
Förster Schreiber, N. M., Genzel, R., Bouché, N., et al. 2009, ApJ, 706, 1364
Förster Schreiber, N. M., Shapley, A. E., Genzel, R., et al. 2011, ApJ, 739, 45
Genzel, R., Tacconi, L. J., Gracia-Carpio, J., et al. 2010, MNRAS, 407, 2091
Genzel, R., Newman, S., Jones, T., et al. 2011, ApJ, 733, 101
Hainline, K. N., Shapley, A. E., Greene, J. E., Steidel, C. C., Reddy, N. A., &
Erb, D. K. 2012, ApJ, 760, 74
Hopkins, A. M., & Beacom, J. F. 2006, ApJ, 651, 142
Ilbert, O., McCracken, H. J., Le Fevre, O., et al. 2013, arXiv:1301.3157
Jones, T. A., Swinbank, A. M., Ellis, R. S., Richard, J., & Stark, D. P. 2010,
MNRAS, 404, 1247
Kennicutt, Jr., R. C. 1998, ApJ, 498, 541
Kewley, L. J., & Dopita, M. A. 2002, ApJS, 142, 35
Kewley, L. J., & Ellison, S. L. 2008, ApJ, 681, 1183
Lara-López, M. A., Cepa, J., Bongiovanni, A., et al. 2010, A&A, 521, L53
Lehnert, M. D., Nesvadba, N. P. H., Le Tiran, L., et al. 2009, ApJ, 699, 1660
Lehnert, M. D., Le Tiran, L., Nesvadba, N. P. H., et al. 2013, A&A, 555, A72
Le Tiran, L., Lehnert, M. D., van Driel, W., Nesvadba, N. P. H., & Di Matteo,
P. 2011, A & A, 534, L4
Liu, X., Shapley, A. E., Coil, A. L., Brinchmann, J., & Ma, C.-P. 2008, ApJ, 678,
758
Maiolino, R., Nagao, T., Grazian, A., et al. 2008, A&A, 488, 463
Mannucci, F., Cresci, G., Maiolino, R., Marconi, A., & Gnerucci, A. 2010,
MNRAS, 408, 2115
118
REFERENCES
Mannucci, F., Cresci, G., Maiolino, R., et al. 2009, MNRAS, 398, 1915
Muzzin, A., Marchesini, D., Stefanon, M., et al. 2013, arXiv:1303.4409
Newman, S. F., Shapiro Griffin, K., Genzel, R., et al. 2012, ApJ, 752, 111
Newman, S. F., Buschkamp, P., Genzel, R., et al. 2013, arXiv:1306.6676
Noeske, K. G., Weiner, B. J., Faber, S. M., et al. 2007, ApJL, 660, L43
Oey, M. S., & Clarke, C. J. 1997, MNRAS, 289, 570
Penston, M. V., Robinson, A., Alloin, D., et al. 1990, A&A, 236, 53
Richard, J., Jones, T., Ellis, R., et al. 2011, MNRAS, 413, 643
Rigby, J. R., Wuyts, E., Gladders, M. D., Sharon, K., & Becker, G. D. 2011,
ApJ, 732, 59
Shirazi, M., & Brinchmann, J. 2012, MNRAS, 421, 1043
Shirazi, M., Vegetti, S., Nesvadba, N., et al. 2013, submitted to MNRAS,
arXiv:1306.6282
Swinbank, A. M., Webb, T. M., Richard, J., et al. 2009, MNRAS, 400, 1121
Tacconi, L. J., Genzel, R., Neri, R., et al. 2010, Nature, 463, 781
Tacconi, L. J., Neri, R., Genzel, R., et al. 2013, ApJ, 768, 74
Tremonti, C. A., Heckman, T. M., Kauffmann, G., et al. 2004, ApJ, 613, 898
Verdolini, S. and Yeh, S. C. C. and Krumholz, M. R. and Matzner, C. D. and
Tielens, A. G. G. M., 2013, ApJ, 769, 12
Wisnioski, E., Glazebrook, K., Blake, C., et al. 2012, MNRAS, 422, 3339
Wuyts, E., Rigby, J. R., Sharon, K., & Gladders, M. D. 2012, APJ, 755, 73
Yeh, S. C. C., & Matzner, C. D. 2012, ApJ, 757, 108
York, D. G., Adelman, J., Anderson, Jr., J. E., et al. 2000, AJ, 120, 1579
119
Chapter
5
On the spatial distribution of
star formation in distant and
nearby galaxies
Distant star-forming galaxies show more clumpy structures in the color maps that
are star formation indicators compared to nearby star-forming galaxies. At the
same time, the mean global properties of galaxies can be significantly different
between distant and nearby Universe. In this work, we study the differences
between the spatial distribution of star formation between a sample of distant
galaxies from the Hubble Ultra Deep Field and nearby star-forming galaxies in
the Sloan Digital Sky Survey with similar stellar mass and specific star formation
rate. We construct spatial maps of physical properties using multi-band imaging
data and compare the maps derived for high-redshift galaxies to those derived
for low-redshift data. In general, we find that the stellar mass is more centrally
concentrated in distant galaxies that we study here, but star formation is more extended compared to local galaxies with the same global properties. More massive
galaxies at high-redshift are more concentrated than their low-redshift counterparts while less massive galaxies show more similar concentration derived from
the specific star formation maps at low and high redshifts. On the basis of the
maps of specific star formation rate, we find that high- and low-redshift galaxies
with the same global properties also show more clumpy distributions of specific
star formation rate.
Maryam Shirazi and Jarle Brinchmann
Monthly Notices of the Royal Astronomical Society
2013, to be submitted
Spatial distribution of star-formation
5.1
Introduction
During the past decade, high resolution observations of high-redshift (high-z)
star-forming galaxies have revealed an increased fraction of clumpy actively starforming galaxies (Griffiths et al., 1994; Windhorst et al., 1995; Cowie et al., 1995;
Abraham et al., 1996; Elmegreen & Elmegreen, 2006; Elmegreen et al., 2007; Genzel et al., 2011; Wuyts et al., 2012). This finding has motivated many studies to
investigate the nature of these clumps and their implications for galaxy evolution,
from both observational and theoretical point of view.
To study the nature of these clumps observationally, the resolved color information and stellar population of their host galaxies and/or IFU observations are
needed. These have been provided by analyzing the rest-frame UV and also the
rest-frame optical emission of z ∼ 2 galaxies using Near-IR observations. Studies
of the spatial variations of the stellar populations properties on ∼ kpc scales were
started by Abraham et al. (1999) at intermediate redshifts using the Hubble Deep
Field data and continued by Thompson et al. (1999), Dickinson et al. (2000), Elmegreen et al. (2009) and Förster Schreiber et al. (2011b) for a sample of high-z
objects using NICMOS observations. Detailed analyses of the resolved colors were
also studied for a sample of clumpy galaxies in CANDELS by Guo et al. (2011a,b).
Recently, Wuyts et al. (2012) studied both the resolved colors and stellar
populations of a complete sample of 649 star-forming galaxies at 0.5 < z < 2.5
with stellar masses of > 1010 M and specific star formation rate (sSFR) of
log(sS FR) > −9.76 at z ∼ 1, and log(sS FR) > −9.51 at z ∼ 2 using deep 7-band
ACS+WFC3 imaging in the GOODS-South field. Using this high resolution data,
they showed that the inferred stellar distributions of these galaxies are less clumpy
than their star formation distributions. They argued that the clumpy rest-frame
UV morphologies and the smooth stellar mass distribution of these galaxies might
imply that the star formation history at a given galactocentric radius is uniform,
but varies spatially on timescales similar to the lifetime of OB stars.
IFU observations of clumpy star-forming galaxies at high-z have also provided
insightful information about the physical nature of these clumps and their formation and evolution. The spatially-resolved observations have revealed large-scale
outflows from individual star-forming clumps in some of these high-z galaxies (e.g.,
Genzel et al., 2011; Newman et al., 2012a,b). These powerful outflows play an important factor in setting the lifetimes of these clumps (e.g., Genzel et al., 2011).
The spatially-resolved studies of the ionized gas and molecular gas find disk-like
kinematics for many of these clumpy galaxies (Genzel et al., 2006; Shapiro et al.,
2008; Bournaud et al., 2008; Contini et al., 2012; Daddi et al., 2010; Epinat et al.,
2012; Tacconi et al., 2013).
In the low-redshift (low-z) Universe however, mergers are known to be responsible for making clumpy morphologies (Conselice et al., 2003; Lotz et al., 2004). As
only a minority of the observed high-z galaxies are in the process of merging (see
e.g., Bundy et al., 2009; López-Sanjuan et al., 2009; Williams et al., 2011; Lotz et
al., 2008, 2011a; Conselice et al., 2009), mergers can not be the dominant physical
process causing clumpy morphologies at high-z.
The formation mechanism for making the majority of these clumps is known
122
High-z sample
to be gravitational instability (e.g., Genzel et al., 2011) where these instabilities
can be caused by high fraction of dense molecular gas (Daddi et al., 2010; Tacconi
et al., 2010). From a theoretical point of view, cosmological simulations predict
that at high-z, gas accretion are playing a significant role for fueling star formation
(Kereš et al., 2005; Dekel et al., 2009). Based on this scenario, gas accretion into
the halo can continuously fuel the disk and actively star-forming clumps seen in the
Hα and rest-frame UV maps of high-z galaxies can then be caused by gravitational
instabilities within these gas-rich disks in this picture (Genzel et al., 2011).
While the morphologies of high and low-z galaxies are well studied, they can be
challenging to compare because the mean properties of the galaxies can be significantly different. Indeed, the energy input into the galaxies can be a factor of several
higher at high-z because the typical star formation rates are much higher. This
difference in energy balance could naturally lead to differences in the overall structure of star formation between high- and low-z galaxies. To compare the physical
processes that make clumpy morphologies at high-z and low-z, one should select
galaxies that have the same global properties such as mass and sSFR. Studying
the resolved color and stellar populations of galaxies with similar properties at
low-z and high-z makes it possible, at least in principle, to take out the effect of
the star formation rate (SFR) on the morphological properties of the galaxies. For
instance, whether an internal process such as SFR or external effects such as gas
accretion which is not likely at low-z are responsible.
In this work, we study the distribution of star-formation in high-z star-forming
galaxies and their low-z counterparts. We use the multi-band imaging data available in the Hubble Ultra Deep Field (HUDF) and compare this quantitatively with
low-z data from the Sloan Digital Sky Survey (SDSS). In Section 5.2, we introduce
our high-z data and infer resolved physical parameter maps for the sample. In
this section, we also contrast the properties inferred from the resolved analysis
to those inferred from modeling of the integrated photometry. In Section 5.3, we
present the sample selection of the low-z counterparts and explain the resolved stellar population modeling of our low-z sample. We compare the clumpiness (M20),
concentration (Gini) of the low-z and high-z samples in Section 5.4. Finally, we
discuss our results in 5.5 and conclude in 5.6.
5.2
High-z sample
In this study, we make use of the objects in the HUDF catalogue by Coe et al.
(2006) that have confirmed spectroscopic redshifts 1 . That amounts to a total of
57 objects with 0.127 < z < 5.819 and median redshift of 0.667, among which 36
objects have 0.5 < z < 1.5. In our analysis, we exclude 7 objects from the high-z
sample for which we infer an integrated sSFR < −11. The total HUDF catalogue
consists of 18,700 objects and comes with a segmentation map. We use the same
segmentation map as the basis for identification of our objects.
1 http://adcam.pha.jhu.edu/~coe/UDF/paper/zspec.cat
123
Spatial distribution of star-formation
5.2.1
Resolved stellar population modeling
Spatially resolved spectral energy distribution (SED) modeling of broad-band
multi-band imaging of galaxies can be have spatial biases because of the significant contribution of low signal-to-noise (S/N) pixels in the outskirts of the galaxies
(Welikala et al., 2009). Therefore, in order to have a sufficiently high S/N ratio
for resolved SED modeling, we first bin our data following the Voronoi 2D binning technique by Cappellari & Copin (2003) to have approximately constant S/N
across the images.
5.2.2
The Voronoi binning of the UDF data
As we mentioned above, to improve the color measurements in the outskirts of
the UDF galaxies where the noise is significant, we use the Voronoi tesselation
technique used for IFU data by Cappellari & Copin (2003) and later extended to
X-ray data by Diehl & Statler (2006). In this work, we use the implementation
presented by Diehl & Statler (2006) and carry out our procedure for both target
S/N of 10 and 20. Based on the Voronoi binning technique, we aggregate pixels
in approximately circular bins until the desired S/N is reached in that bin.
The SED fitting is sensitive to the colors so to avoid biases we require that the
color maps have approximately uniform S/N. To do this, we create our bins using
one color image and apply this binning to the flux images subsequently. Ideally
we should create bins using all images at the same time, but here we focus on a
single color. When the integrated S/N in the B-band image is twice the target
S/N, we use the B − V color image, otherwise we use the V − I image. This choice is
a balance between creating a reasonable number of bins per object and good S/N
per effective pixel. For the higher redshift objects the S/N in the B-band image
can be rather poor and using this for binning would lead to very few bins.
Examples of the Voronoi maps are shown in Figure 5.1 for six different objects.
Four panels that belong to one object shows the unbinned V-band image and
the Voronoi binned B − V, V − I and i0 − z0 images, as indicated in the panels.
From these maps we can see how the Voronoi binning procedure isolates regions
of different colors — which often translates into isolating regions with distinct
physical properties (e.g., star formation rates). Note for instance 3088 for which
the high-S/N color maps highlights a blue region along the major axis of the galaxy
while also showing a ring-like red feature around the galaxy. The target S/N for
this illustration is 10.
5.2.3
Fitting models to the color map of the UDF data
We fit Bruzual & Charlot (2003, , hereafter BC03) models to the B, V, i0 and
z0 fluxes for each effective pixel of the UDF data that is given by the Voronoi
binning procedure. We adopt the approach introduced by Kauffmann et al. (2003).
This relies on a set of models with variable star formation histories, as described
in detail by Gallazzi et al. (2005, 2008). For each effective pixel we calculate
the χ2 between our observed fluxes and the stochastic model grid, which gives
us a likelihood for each model, P(model|B, V, i0 , z0 ) under the assumption that the
124
High-z sample
Figure 5.1 The procedure of Voronoi binning galaxy images of the UDF data is
illustrated for six different objects. For each object the B − V, V − i0 , i0 − z0 (binned)
and V-band images are shown. The spectroscopic redshift and also the ID of each
object as presented in the Coe et al. (2006) catalogue are also indicated above each
panel. The color range in the color images is the same in each panel.
uncertainties involved are Gaussian. Then we construct a likelihood distribution of
various quantities for each effective pixel. We calculate the stellar mass (M∗ ), the
star formation rate (SFR), sSFR (sSFR= SFR/M∗ ) and the r-band weighted age,
and store the 100 most likely models and their likelihoods. As the area belongs to
each effective pixels derived from Voronoi binning might be large, we draw random
values from these likelihoods using Monte-Carlo resampling to populate each pixel
of the derived maps.
The result of this SED fitting procedure can be seen in Figure 5.2 for all 47
UDF objects studied in this work. The color images are shown in the left-most
column, where the redshift and the ID of the UDF objects are also indicated.
Other columns from left to right show the Voronoi binned B − V color map, the
stellar mass map, the SFR map, the sSFR map and the age map, respectively.
The pixel scale in these maps is 0.03 arcsec/pixel.
5.2.4
Integrated properties
In order to measure the integrated physical properties of the UDF sample, we fit
BC03 models to the integrated B, V, i0 , z0 photometry presented in the Coe et al.
(2006) catalogue. This gives us M∗ , SFR, sSFR, and the r-band weighted age for
each UDF object. We also derive the total values of the physical parameters from
the spatially resolved stellar population modeling. A comparison of global stellar
125
Spatial distribution of star-formation
Stamp
Z= 0.13
B-V bins
1"
Log Mass
1.2
1.0
Log SFR
5.0
4.5
0.8
0.6
0.4
0.2
3088
Z= 0.66
0.0
1"
4.0
3.5
3.0
Log sSFR
-8.5
-5.0
-9.0
-5.5
-9.5
-6.0
-10.0
-6.5
1.0
6.0
Log age
-4.5
9.4
9.2
9.0
8.8
8.6
8.4
-10.5
8.2
-8.5
9.0
-3.0
0.8
8.8
0.6
-3.5
-9.0
-4.0
-9.5
8.6
5.5
0.4
8.4
0.2
5.0
1375
Z= 0.74
0.0
1"
1.0
0.8
7.0
6.5
0.6
8.2
-4.5
-2.0
-8.5
-2.5
-9.0
-3.0
-9.5
-3.5
-10.0
-4.0
-10.5
6.0
0.4
0.2
8810
Z= 0.74
0.0
1"
4142
Z= 0.95
1"
6206
Z= 1.09
1"
8261
Z= 1.10
8749
5.5
1"
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
1.8
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
7.0
6.5
6.0
5.5
5.0
7.5
-1.0
-1.5
-2.0
-2.5
-3.0
-3.5
-4.0
9.6
9.4
9.2
9.0
8.8
8.6
8.4
8.2
-10.0
9.4
9.2
9.0
8.8
8.6
8.4
8.2
-10.5
9.4
-8.5
-9.0
-9.5
-3.5
-4.0
7.0
-11.0
-4.5
-5.0
-11.5
-5.5
-12.0
9.2
6.5
9.4
7.5
7.0
6.5
8.0
7.5
7.0
6.5
6.0
-1.5
-2.0
-2.5
-3.0
-3.5
-4.0
-4.5
-0.5
-1.0
-1.5
-2.0
-2.5
-3.0
-3.5
-9.0
-9.5
-10.0
-10.5
-11.0
-11.5
-12.0
-8.5
9.2
9.0
8.8
8.6
8.4
8.8
8.6
-9.0
8.4
-9.5
8.2
-10.0
8.0
Figure 5.2 The results of the SED fitting procedure for 47 UDF galaxies studied
here. The left column shows the color JPEG image of the object with the redshift
and UDF ID indicated (note that the region and orientation shown here is different
from the images in the following columns). The next columns from left to right
show the Voronoi binned B − V color map for S/N of 10, the stellar mass maps
(log M∗ /M ), the SFR maps (log SFR/M yr−1 ), the sSFR maps (log sSFR/yr−1 )
and r-band luminosity-weighted age maps (log age/yr) all on a spatial scale of 0.03
arcsec/pixel.
126
High-z sample
Z= 1.22
-1.5
1"
0.6
7.0
-2.0
0.4
6.5
-8.5
9.2
-9.0
9.0
8.8
-2.5
-9.5
-3.0
0.2
6.0
Z= 1.22
8.6
8.4
8.2
0.0
1"
8.0
0.8
7.5
0.6
7.0
0.4
6.5
0.2
6.0
4396
Z= 1.24
-10.0
-3.5
4816
-0.5
-1.0
-1.5
-2.0
-2.5
-3.0
-3.5
-9.0
9.4
9.2
-9.5
9.0
-10.0
8.8
-10.5
8.6
0.0
1"
-1.5
0.6
-8.5
7.0
9.0
-2.0
0.4
6.5
-9.0
8.8
-2.5
0.2
5.5
1829
Z= 1.24
6.0
8.6
-3.0
-3.5
-10.0
0.0
1"
0.8
0.6
-9.5
-1.0
-8.5
-1.5
-9.0
7.0
9.0
6.5
Z= 1.29
-3.5
0.0
1"
-9.5
8.8
-10.0
8.6
-10.5
8.4
-2.5
-3.0
0.2
1266
8.2
9.2
7.5
-2.0
0.4
8.4
0.8
9.2
-2.0
7.0
-9.0
0.6
9.0
-2.5
0.4
0.2
7995
Z= 1.31
-3.0
-9.5
-3.5
-10.0
8.6
6.0
8.4
0.0
1"
6188
Z= 1.32
8.8
6.5
1"
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
0.6
-3.0
7.5
-10.5
-3.5
-11.0
9.4
-4.0
7.0
-11.5
-4.5
-5.0
7.0
-2.0
0.4
6.5
-2.5
0.2
6.0
-3.0
9.2
-12.0
9.0
-9.0
8.8
8.6
8.4
7725
Z= 1.43
-9.5
0.0
1"
0.8
7.5
-1.5
0.6
7.0
-2.0
0.4
6.5
-2.5
0.2
8461
6.0
-8.5
9.0
-9.0
-3.0
8.8
8.6
-9.5
8.4
0.0
Figure 5.2 Continued.
127
Spatial distribution of star-formation
Z= 0.13
1"
5670
Z= 0.21
1"
-4.5
1.2
1.0
0.8
0.6
0.4
0.2
0.0
5.0
1.0
5.0
-5.0
-9.0
-5.5
-9.5
4.0
-6.0
3.5
-10.0
-6.5
0.4
8.6
8.4
-3.5
-4.0
8.8
-8.5
8.6
4.5
0.6
8.8
-10.5
3.0
0.8
9.2
9.0
4.5
-4.5
8.4
-9.0
4.0
-5.0
3.5
-5.5
0.0
1.0
5.0
-3.5
8.2
0.2
5620
Z= 0.21
1"
0.6
0.4
8.6
-4.5
4.0
5606
Z= 0.32
-9.0
-5.0
-5.5
0.0
1"
8.8
-4.0
4.5
0.2
Z= 0.23
8.0
9.0
-8.5
0.8
1000
-9.5
8.4
8.2
-9.5
8.0
-9.5
9.2
-10.0
9.0
-3.5
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
1"
6.0
-4.0
5.5
5.0
-4.5
-5.0
4.5
-5.5
4.0
-6.0
8.8
-10.5
6.0
9.4
-3.5
0.4
0.2
-9.0
5.5
-4.0
5.0
-4.5
4.5
-5.0
9.2
9.0
-9.5
8.8
8.6
5190
Z= 0.33
0.0
1"
7847
Z= 0.34
1"
3492
Z= 0.35
2.0
1.8
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
7.0
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
6.0
6.5
6.0
5.5
5.0
-2.0
-2.5
-3.0
-3.5
-4.0
-4.5
-5.0
-5.5
-6.0
-3.0
-10.0
9.4
-9.0
-9.5
-4.0
9.0
-10.5
8.8
-11.0
8.6
-8.5
9.0
8.8
-9.0
8.6
5.0
-4.5
4.5
-5.0
4.0
-5.5
-9.5
8.4
-10.0
8.2
-8.5
1"
9.0
0.4
-4.5
4.5
-9.0
-9.5
Figure 5.2 Continued.
128
0.0
4.0
-5.5
8.8
8.6
-5.0
0.2
4267
9.2
-10.0
-3.5
5.5
8.4
8.4
8.2
High-z sample
Z= 0.35
1"
3268
Z= 0.38
1"
8585
Z= 0.42
1"
1"
2107
Z= 0.67
1"
2607
Z= 0.67
1"
968
Z= 0.67
1"
662
Z= 0.67
6.0
-8.5
-3.5
8.8
5.5
5.0
-4.0
-9.0
8.6
8.4
4.5
-4.5
-9.5
8.2
4.0
1.2
-2.5
9.4
1.0
6.5
-3.0
0.8
6.0
-3.5
0.6
5.5
-4.0
0.4
5.0
-4.5
0.2
4.5
-5.0
7.0
-2.5
-3.0
-3.5
-4.0
-4.5
-5.0
-5.5
-6.0
-6.5
-9.0
9.2
9.0
-9.5
8.8
-10.0
8.6
0.0
900
Z= 0.53
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
1"
355
1.8
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
2.0
1.8
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
2.4
2.2
2.0
1.8
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
1.2
1.0
0.8
0.6
0.4
0.2
0.0
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
6.5
6.0
5.5
5.0
-9.5
9.4
-10.0
-10.5
-11.0
9.2
9.0
-11.5
8.8
7.5
-3.0
-9.5
9.4
7.0
6.5
-3.5
-10.0
-4.0
-10.5
9.2
-11.0
9.0
-4.5
-5.0
6.0
7.0
6.5
-5.5
-11.5
-3.0
-9.0
-3.5
-9.5
9.2
9.0
-4.0
6.0
-10.0
-4.5
5.5
-10.5
-5.0
6.5
-2.5
6.0
-3.0
-8.5
-3.5
5.0
-4.0
8.6
9.0
8.8
-9.0
5.5
8.8
8.6
8.4
-9.5
8.2
4.5
6.5
6.0
5.5
5.0
-2.0
-2.5
-3.0
-3.5
-4.0
-4.5
-5.0
-5.5
-8.5
-3.0
-9.0
9.2
-9.0
9.0
-9.5
8.8
-10.0
8.6
-10.5
8.4
-11.0
8.2
9.4
6.5
9.2
6.0
5.5
-3.5
5.0
-4.0
9.0
8.8
-9.5
8.6
-10.0
8.4
Figure 5.2 Continued.
129
Spatial distribution of star-formation
Z= 0.67
1"
53380
Z= 0.73
1"
6933
Z= 0.74
1"
2.0
1.8
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
2.0
1.8
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
9.6
-10.0
7.0
-10.5
-8.5
5.0
-4.0
1.2
7.0
-2.0
1.0
6.5
-2.5
6.0
-3.0
1"
1"
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
0.4
0.0
0.2
797
-9.5
-4.0
8.2
-2.5
-3.0
5.0
8.6
8.4
-3.5
8.2
-4.0
8.0
-1.5
-8.5
9.0
7.0
8.8
-2.0
6.5
6.0
-2.5
-9.0
-3.0
-9.5
-3.5
8.6
8.4
8.2
5.5
8.0
-1.0
7.5
-1.5
7.0
-2.0
6.5
-2.5
6.0
-3.0
5.5
-3.5
9.0
-1.5
1.2
0.4
0.2
0.0
Figure 5.2 Continued.
-9.5
5.5
9.0
8.8
-9.0
6.0
7.0
-2.0
6.5
-2.5
6.0
-3.0
5.5
-3.5
5.0
-4.0
4.5
-4.5
0.6
4394
-8.5
-2.0
0.8
1"
-1.5
6.5
1.0
4445
8.4
9.2
7.0
0.0
1"
8.8
8.6
-3.5
5.5
-8.5
0.6
0.4
130
9.0
-9.0
0.2
Z= 0.67
8.2
-9.5
1.2
0.6
1"
8.6
8.4
0.8
5417
9.0
8.8
-9.0
0.0
1.0
Z= 0.46
-3.0
5.0
3372
Z= 1.31
9.2
-2.5
6.0
0.2
Z= 1.10
9.2
-4.5
-3.5
0.4
Z= 1.00
-11.0
5.5
0.6
9.4
-4.0
6.5
0.8
2525
-3.5
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
6.5
-2.5
-3.0
6.0
-3.5
5.5
-4.0
-9.0
8.8
8.6
-9.5
-8.5
8.4
9.4
9.2
-9.0
9.0
8.8
-9.5
8.6
8.4
-10.0
-8.5
-9.0
9.0
8.8
8.6
-9.5
8.4
-4.5
5.0
-5.0
-10.0
8.2
Low-z sample
Z= 0.76
1"
1.2
1.0
-8.5
6.5
0.6
-2.5
-9.0
0.8
-3.0
6.0
-3.5
-9.5
-4.0
-10.0
0.4
0.2
8275
Z= 3.06
1"
9.5
1.0
9.0
0.8
8.5
0.6
8.0
0.4
7.5
0.2
7.0
1"
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
2.0
1.8
1.6
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
9.4
-10.0
9.2
-2.0
-10.5
-2.5
-11.0
-3.0
-11.5
9.0
8.8
6.0
-9.0
-4.0
9.2
-9.5
5.5
9.0
-4.5
-10.0
5.0
4.5
7.0
6.5
6.0
5.5
5.0
4.5
8.8
-5.0
-2.5
-3.0
-3.5
-4.0
-4.5
-5.0
-5.5
-6.0
-10.5
8.6
-9.0
9.2
-9.5
9.0
-10.0
8.8
-10.5
8.6
-11.0
8.4
-8.5
6.5
1"
-9.5
-1.5
6.5
1"
3822
Z= 1.30
1.2
0.0
8015
Z= 0.44
-4.5
0.0
865
Z= 0.28
5.5
9.4
9.2
9.0
8.8
8.6
8.4
8.2
-2.5
8.8
0.2
6.0
-3.0
-3.5
0.0
8.6
8.4
5.5
7705
-9.0
8.2
Figure 5.2 Continued.
population properties derived from the integrated photometry and the resolved
photometry for S/N of 10 are shown in Figure 5.3. As we can see from this
comparison, there is not a significant bias relative to the uncertainty estimates on
the quantities. In the next section, we use the global properties inferred from the
integrated stellar population modeling to assemble our low-z sample. The global
physical properties of the UDF sample are summarized in Table 5.1.
5.3
Low-z sample
To find a suitable comparison sample at low-z and compare their star formation
properties to that of high-z galaxies in a systematic way, we follow the same
approach that was taken by Shirazi et al. (2013). Using this procedure we can
compare galaxies that have similar star formation activities at high- and lowz. For each UDF galaxy, we select star-forming galaxies from the SDSS DR7
(Abazajian et al., 2009) that have log M∗ and log sSFR within 0.3 dex of that of
131
Spatial distribution of star-formation
Figure 5.3 A comparison between the global stellar population properties derived
from the integrated photometry and the ones derived by integrating over the resolved photometry for S/N of 10 are shown for the UDF sample. Top-left panel
shows sSFR as a function of stellar mass derived by integrating over resolved values. Top-right panel compares SFR derived using the integrated photometry and
SFR derived by integrating over the resolved photometry for S/N of 10. Bottomleft and bottom-right show the same for sSFR and stellar mass, respectively. In
all panels shaded regions show 1-σ scatter.
UDF object. Our parent sample is based on the MPA-JHU2 , however we use the
imaging from the SDSS DR9. We select star-forming galaxies based on the BPT
diagram (Baldwin, Phillips & Terlevich, 1981) and the classification presented by
Kauffmann et al. (2003).
In order to have the same resolution at high-z and low-z, we select SDSS
galaxies that have z < 0.02. The spatial resolution of ACS is 0.1” which covers
∼ 0.61 kpc at z = 0.5. The typical SDSS PSF size is 1.5” which covers the same
∼ 0.61 kpc scale at z = 0.02. In total, we found about 4000 objects as our parent
sample but in order to focus on objects that have the highest possible resolution,
whenever we have more than 40 low-redshift counterparts we use the 40 lowest
redshift ones only, which results in 577 SDSS galaxies.
2 http://www.mpa-garching.mpg.de/SDSS/DR7
132
Low-z sample
Figure 5.4 A comparison between the SDSS sample of 577 galaxies (gray squares)
and the UDF sample (colored symbols) in the mass-sSFR diagram is shown. The
scale shows the redshift of the UDF sample. We excluded two UDF galaxies that
have z > 1.5 from the UDF sample. Purple squares show the UDF objects for
which we assembled the SDSS sample. We can see that the UDF objects that
have very high sSFRs and high masses (colored symbols without purple squares
around them) are offset from low-z analog galaxies. Thus, we can not find any
low-z sample for them using our sample selection criteria.
Figure 5.6 The SDSS parent sample (577 low-z galaxies) is shown in the BPT
diagram. The color coding shows the sSFR. Note that the sSFR decreases with
increasing [N ii] /Hα (metallicity), and for a given metallicity, it increases with
increasing [O iii] /Hβ.
133
Spatial distribution of star-formation
Figure 5.5 The median values of the mass and sSFR of the SDSS sample (gray
square) selected for each UDF galaxy are compared with that of the UDF galaxies
that have local comparison samples (colored symbols).
We run SExtractor (Bertin & Arnouts , 1996) on the SDSS r-band images
to produce segmentation maps. The default setting for SExtractor was to use a
DETECT THRESH = 0.61, DETECT MINAREA = 40, ANALYSIS THRESH = 0.61,
DEBLEND MINCONT = 0.03 and DEBLEND NTHRESH = 32. For individual
objects this was sometimes adjusted e.g. when a bright nearby star affects the
background level. To create the mosaic images we made use of the routines3
developed by Zibetti et al. (2009) to download and align the SDSS images.
We compare our low-z and high-z sample of galaxies in the stellar mass - sSFR
plane in Figure 5.4. As we can see, the UDF objects that have very high sSFR
and mass are offset from the low-z analog sample. Therefore, we cannot find
low-z analogs for all 47 UDF objects. Purple squares show the UDF objects for
which we could assemble a SDSS comparison sample. In Figure 5.5, we show the
median values of the mass and sSFR of the SDSS sample (gray square) selected
for each UDF galaxy in comparison with that of the UDF galaxies that have local
comparison samples (colored symbols).
We show the emission line properties of our SDSS sample based on the BPT
diagram in Figure 5.6. This diagram is used extensively to classify galaxies in terms
of their main sources of ionization/excitations. Note that our star-forming galaxies
are all distributed below the classification lines shown by dashed line derived by
Kauffmann et al. (2003). The x-axis shows [N ii] /Hα which is a metallicity tracer
and the y-axis show [O iii] /Hβ which is sensitive to ionization properties of the
main ionizing sources in emission line galaxies, the color coding in this diagram
shows the sSFR. From this figure we can see that the sSFR decreases with increas3 Downloaded from http://www.arcetri.astro.it/~zibetti/Software/SDSSmosaic.html and
updated to use SDSS DR9.
134
Low-z sample
Figure 5.7 The procedure of Voronoi binning galaxy images from the SDSS illustrated for six different objects. For each object the u − g, g − r, r − i (binned)
and g-band images are shown. The spectroscopic ID and the redshift are also
indicated.
ing [N ii] /Hα (metallicity), and for a given metallicity, it increases with increasing
[O iii] /Hβ in our low-z sample.
5.3.1
Resolved stellar population modeling of low-z sample
We use the same technique used for the UDF data to bin the SDSS color images.
This has been done for target S/N of 5, 10, 20 in the u − g color image and S/N of
5 in the g − r and r − i color images. An example of the Voronoi maps is shown in
Figure 5.7 for six SDSS galaxies. Each group of panels belongs to a single object
and the four panels show the unbinned g-band image, the Voronoi binned u − g,
g−r and r −i images (as indicated on the maps). The target S/N for this particular
illustration is S/N of 10 in the u − g color.
5.3.2
Fitting models to the color map of the SDSS data
We follow the same approach which was used for the UDF data and fit models
to the u, g, r, i and z flux of effective pixels of the SDSS data, derived from the
Voronoi binning procedure. An example of the fitting procedure for one SDSS
galaxy (PlateID-MJD-FiberID: 2746-54232-104) is shown in Figure 5.8. The r
band image is shown in the left-most column where the redshift of the SDSS
galaxy is also indicated. Thereafter follows the Voronoi binned u − g color map,
135
Spatial distribution of star-formation
the stellar mass map, the SFR map, the sSFR map and the r-band luminosityweighted age map. Each row, from the top to the bottom, shows the result of SED
fitting for target S/N of 5, 10, 20 in the u−g color and S/N of 5 in the g−r and r −i
color images, respectively. By comparing the SED fitting results that we get using
different Voronoi binning, we conclude that generally using u-g binning identifies
structures very well, while a binning using redder colors washes out the structures
completely. The reason might be because of the lower S/N in the u-band than g
or r, so a cut at S/N of 5 for g − r results in a very low S/N in u − g and hence the
SED fits will be poor. We can also see (by comparing the first three rows) that a
S/N of 5 in u-g is sufficiently good to derive galaxy properties. In the following,
we use the results of the SED fitting on maps that were Voronoi binned requiring
a target S/N of 5 for the u-g SDSS images. The clumpy structure of the galaxy
can be seen in the SFR and the sSFR maps but we note that the mass maps have
a much more smoothed distribution. This is consistent with Wuyts et al. (2012)
who showed the stellar mass maps derived for high-z galaxies are more smoothed
than their color maps.
5.3.3
Integrated properties
In order to test the result of our resolved stellar population modeling, we integrate
over the inferred resolved physical properties and compare them with the results
that are derived from the integrated photometry. Figure 5.9 shows a comparison
between the global stellar population properties derived from the integrated photometry, and those derived by integrating over the resolved photometry for S/N
of 5 for the SDSS sample. Top-left panel shows sSFR as a function of stellar
mass derived by integrating over resolved values. Top-right panel compares SFR
derived using the integrated photometry and SFR derived by integrating over the
resolved photometry. Bottom-left and bottom-right show the same for sSFR and
stellar mass, respectively. From these comparisons we can see that the integrated
and the summed resolved properties for the quantities used to match to the high-z
sample, the stellar masses and sSFRs, are in good agreement (within 0.1 dex) for
galaxies with M? > 108.5 M .
5.4
Structural parameters
Galaxy morphologies can be parameterized using a variety of indicators such as
the concentration index (Abraham et al., 1996; Conselice et al., 2003), the Gini
coefficient (0 < Gini < 1, a more concentrated morphology corresponds to a higher
Gini, see, e.g., Abraham et al. 2003; Lotz et al. 2004; Förster Schreiber et al.
2011a) and the M20 index (M20< 0, a more clumpy morphology corresponds to a
higher M20, see Lotz et al. 2004).
We use galSVM code4 (Huertas-Company et al., 2008, 2009, 2011) to measure
Gini and M20 parameters on the mass, SFR and sSFR maps inferred for the
UDF and SDSS objects and show the results in Figure 5.10 and Figure 5.11.
4 http:www.lesia.obspm.fr/~huertas/galsvm.html
136
Structural parameters
r-band
Z=0.008
u-g bins
10 "
Log Mass
0.8
0.6
Log SFR
6.0
5.5
0.4
0.2
u-g voronoi sn5
0.0
Z=0.008
10 "
0.8
5.0
4.5
6.0
0.6
5.5
0.4
0.2
5.0
u-g voronoi sn10
0.0
Z=0.008
10 "
0.8
6.0
0.6
0.4
5.5
0.2
5.0
u-g voronoi sn20
0.0
Z=0.008
10 "
0.8
0.6
0.4
0.2
g-r voronoi sn5
Log sSFR
-3.5
-4.0
-4.5
-5.0
-5.5
-6.0
-6.5
-3.5
-4.0
-4.5
-5.0
-5.5
-6.0
-6.5
-3.5
-4.0
-4.5
-5.0
-5.5
-6.0
-6.5
6.0
5.5
5.0
4.5
4.0
3.5
3.0
-3.5
-4.0
-4.5
-5.0
-5.5
-6.0
-6.5
-7.0
6.0
5.5
5.0
4.5
4.0
3.5
3.0
-3.5
-4.0
-4.5
-5.0
-5.5
-6.0
-6.5
-7.0
Log age
-9.5
9.4
-10.0
9.2
-10.5
9.0
-11.0
8.8
-9.5
9.4
-10.0
9.2
-10.5
-11.0
10 "
0.8
0.6
0.4
0.2
r-i voronoi sn5
0.0
8.8
9.6
-9.5
9.4
-10.0
-10.5
-11.0
9.2
9.0
-11.5
8.8
-9.0
9.4
-9.5
-10.0
9.2
9.0
8.8
8.6
0.0
Z=0.008
9.0
-9.0
-9.5
9.4
9.2
9.0
-10.0
8.8
8.6
Figure 5.8 An example of the fitting procedure for one SDSS galaxy (PlateIDMJD-FiberID: 2746-54232-104). The Voronoi binning has been done for target
S/N of 5, 10, 20 in the u − g color image and S/N of 5 in the g − r and r − i color
images. The result of SED fitting for each one has been shown in each row from the
top to the bottom. The left column shows the r-band image of the object with the
redshift indicated. The next columns show the Voronoi binned u − g color map, the
stellar mass maps (log M∗ /M ), the SFR maps (log SFR/M yr−1 ), the sSFR maps
(log sSFR/yr−1 ) and r-band luminosity-weighted age maps (log age/yr), respectively. By comparing the SED fitting results that we get using different Voronoi
binning (shown in each row), we see that binning in u − g is able to identify structures very well, while a binning using redder colors washes out the structures
completely. The reason might be because of the lower S/N in the u-band than g or
r, so a cut at S/N of 5 for g − r results in a very low S/N in u − g and hence the SED
fits will be poor. Using this we conclude that generally using u-g binning provides
us with better results. We can also see (by comparing the first three rows) that
S/N of 5 in u-g is sufficiently good to derive physical properties. We can see the
clumpy structure of the galaxy in the SFR and the sSFR maps. Note that those
clumps have younger ages in the age maps. The stellar mass maps show more
smoothed structure.
137
Spatial distribution of star-formation
Figure 5.9 A comparison of global stellar population properties derived from the
integrated photometry and the resolved photometry inferred for S/N of 10 are
shown for the SDSS sample. Top-left panel shows sSFR as a function of stellar
mass derived by integrating over resolved values. Top-right panel compares SFR
derived using the integrated photometry and SFR derived by integrating over the
resolved photometry. Bottom-left and bottom-right show the same for sSFR and
stellar mass, respectively. From these comparisons we can see that the derived
stellar masses and sSFRs are in good agreement (within 0.1 dex) for galaxies with
M? > 108.5 M .
We adopt the stellar mass-weighted centers for all maps on which the structural
measurements are measured.
The color scales in Figure 5.10 shows the inferred Gini parameters. In all panels we show the sSFR as a function of stellar mass. The left panels show the Gini
parameters derived from the mass maps, the middle ones show the Gini parameters derived from the SFR maps and the right panels show the Gini parameters
inferred from the sSFR maps. Top panels show the results for the UDF galaxies
and the bottom panel show the median values of the Gini parameters derived for
the SDSS samples of UDF objects. We can see that using the mass maps we get
a higher Gini values than SFR map and sSFR maps. This shows that the mass
maps are more concentrated (higher Gini). The SDSS counterparts show less concentrated morphologies in all maps but the SFR map. While there are also regular
trends seen in the SDSS plots we caution that the sample here is constructed for
differential analysis against the high-z data and is not complete in any statistical
138
Structural parameters
sense for making statements about trends in the low-z Universe.
The color scales in Figure 5.11 shows the inferred M20 parameter for the UDF
galaxies on the top panels and the SDSS galaxies on the bottom panels. In all
panels we show the sSFR as a function of mass. The left panels show the M20
parameters derived from the mass maps, the middle one show the M20 parameters
derived from the SFR maps and the right panels show the M20 parameters inferred
from the sSFR maps. We can see that using the mass maps we get a higher M20
values than SFR map and sSFR maps. This shows that the mass maps are less
clumpy or smoother than the other maps. We can also note that SDSS galaxies
show less clumpy morphology (lower M20) than UDF objects.
To show the comparison between low-z and high-z data more quantitatively,
in Figure 5.12 we compare the differences in the Gini parameters that we infer for
the UDF objects and that of their SDSS counterparts. The left panels show the
∆Gini (∆Gini = GiniUDF − GiniSDSS ) parameters derived from the mass maps, the
middle ones show the ∆Gini parameters derived from the SFR maps and the right
panels show the ∆Gini parameters inferred from the sSFR maps. In all top panels
on the x-axis we show the integrated sSFRs measured for the UDF galaxies and
in all bottom panels on the x-axis we show the integrated stellar masses measured
for the UDF galaxies. We plot the median values of the ∆Gini by solid circles and
show 1-σ scatter by error bars, the results for the UDF galaxies at z > 0.5 and
z < 0.5 are shown by blue and gray colors, respectively. The most striking result
in the figure is the positive ∆Ginimass for all sSFRs and masses and on average
a negative ∆GiniSFR . However, the figure shows at most a very weak correlation
between the ∆Gini parameters and the Log sSFR. A similar picture is presented
by the stellar mass but there is a tentative correlation between ∆GinisSFR and
stellar mass. This shows that galaxies that are more massive at high-z are more
concentrated than their local analogs compared to the one that are less massive
at high- and low-z. While the UDF galaxies are more concentrated in their stellar
content but their star formation is more extended compared to galaxies with the
same global properties in the local Universe.
In Figure 5.13 we compare the differences in the M20 parameters that we infer
for the UDF objects and that of their SDSS counterparts. The left panels show the
∆M20 (∆M20 = M20UDF − M20SDSS ) parameters derived from the mass maps, the
middle ones show the ∆M20 parameters derived from the SFR maps and the right
panels show the ∆M20 parameters inferred from the sSFR maps. In all top panels
on the x-axis we show the integrated sSFRs measured for the UDF galaxies and
in all bottom panels on the x-axis we show the integrated stellar masses measured
for the UDF galaxies. We show the median values of the ∆M20 derived for the
UDF galaxies at z > 0.5 by solid blue circles and the ones at z < 0.5 by solid
gray circles. 1-σ scatter in the ∆M20 is shown by error bars. The figure shows a
correlation between the ∆M20mass and the Log sSFR and also stellar mass. This
shows more massive galaxies are more clumpier than their local analogs compared
to less massive galaxies at high-z and low-z. On average high-z galaxies are more
clumpier using ∆M20sSFR maps.
We show the correlation between ∆M20sSFR and ∆Ginimass in Figure 5.14, where
blue circles show the results for the UDF galaxies at z > 0.5 and gray circles show
139
Spatial distribution of star-formation
them for the UDF galaxies at z < 0.5. In general, more concentrated galaxies
at high-z are more clumpier. The average ∆M20sSFR for the UDF galaxies with
z < 0.5 is 0.06 and for the UDF galaxies with z > 0.5 is 0.12. However, the average
∆Ginimass changes from 0.36 to 0.38 from z < 0.5 to z > 0.5.
5.5
Discussion
Figure 5.13 showed us that galaxies with the same global properties (e.g., stellar
mass and sSFR) at high- and low-z show slightly different distribution in their
sSFR maps with an evolution in clumpiness at high-z. We also tested this results
by measuring the structural parameters on the color maps that are correlated with
star formation activity, e.g., B − V for the UDF and u − g for the SDSS. The results
agree well with the parameters that we inferred using sSFR maps at high- and
low-z.
Our high-z galaxies show a median of ∼ 0.3 dex higher Ginimass than their local
counterparts. There is also a correlation between Ginimass and stellar mass showing
that galaxies that are more massive at high-z are more concentrated than their
local analogs compared to the ones that are less massive. This has been shown
also for passive galaxies at z & 1.5 which show more concentration compared to
local ellipticals with the same stellar mass (e.g., van Dokkum et al., 2008)
There is a possibility of getting more concentration if galaxies at high-z have
AGN. We know that there is an evolution in X-ray luminosity function of AGN
up to redshift ∼ 1.2 (e.g., Aird et al., 2010). However, as we do not use emission
lines, we do not expect our SED fitting results are affected by presence of AGNs.
As we mentioned above, high-z galaxies that we study here are slightly more
clumpy in their star formation distributions than their local analogs. Based on
our selection criteria, the energy injection from the massive star populations is
expected to be approximately the same in our high- and low-z galaxies. Therefore,
in order to have more clumpy morphology at high-z, they need to have more
surface density of the disk compared to their local analogs. These higher surface
densities at high-z that we would postulate being responsible for the slightly more
unstable disks can be caused by cold gas accretion. Then clumps can be formed
from gravitational instability within these gas-rich disks at high-z (see e.g., Genzel
et al., 2011).
However, as cold gas accretion seems unlikely at low-z, this can not be a physical explanation for clumpy morphology in the local Universe. Galaxies that we
compare at high- and low-z have the same global properties, including the same
star formation activity, but the high-z galaxies show more extended star formation
distributions in SFR maps compared to their low-z analogs. It is beyond the scope
of this study to discuss what mechanisms can cause very extended star formation
at high-z and not at low-z in galaxies with the same global properties.
We note that also the low-z data here is constructed for differential analysis
against the high-z data and is not complete in any statistical sense for making
statements about trends in the low-z Universe.
140
Discussion
Figure 5.10 Top: sSFR-mass map for the UDF objects. Bottom: the same for the
median of the SDSS counterparts. Color scales show the Gini parameter measured
on the mass map, SFR map and sSFR map from left to right, respectively. We
can see that using the mass maps we get a higher Gini values than SFR map and
sSFR maps. This shows that the mass maps are more concentrated (higher Gini).
The SDSS counterparts show less concentrated morphologies in all maps but SFR
map.
141
Spatial distribution of star-formation
Figure 5.11 Top: sSFR-mass map for the UDF objects. Bottom: the same for the
median of the SDSS counterparts of UDF galaxies. Color scales show the M20
parameter measured on the mass map, SFR map and sSFR map from left to right,
respectively. We can see that using the mass maps we get a higher M20 values than
SFR map and sSFR maps. This shows that the mass maps are less clumpier or
more smoother than the other maps. We see also SDSS galaxies show less clumpy
morphology (lower M20) than UDF objects.
142
Discussion
Figure 5.12 We compare the differences in the Gini parameters that we infer for
the UDF objects and their SDSS counterparts as a function of the UDF sSFRs on
the top and as a function of the UDF stellar masses on the bottom panels. We
plot the median values of the ∆Gini by solid circles and show 1-σ scatter by error
bars. Blue color marks the results for the UDF galaxies at z > 0.5 and gray color
shows ∆Gini for the UDF galaxies at z < 0.5. Note the positive ∆Ginimass for all
sSFRs and masses and a negative ∆GiniSFR . We see also a very weak correlation
between the ∆Gini parameters and the Log sSFR. There is a tentative correlation
between ∆GinisSFR and stellar mass.
143
Spatial distribution of star-formation
Figure 5.13 We compare the differences in the M20 parameters that we infer for the
UDF objects and their SDSS counterparts as a function of the UDF sSFRs on the
top and as a function of the UDF stellar masses on the bottom panels. We show
the median values of ∆M20 by solid circles and show 1-σ scatter in ∆M20 by error
bars. Blue color marks the results for the UDF galaxies at z > 0.5 and gray color
shows ∆M20 for the UDF galaxies at z < 0.5. We see a correlation between these
∆M20 and sSFR and stellar mass that shows more massive galaxies (low sSFR) are
more clumpier than their local analogs compared to less massive galaxies at high-z
and low-z. Note that the UDF galaxies and SDSS counterparts agree better in the
M20 parameter inferred from the sSFR maps compared to the M20 inferred from
mass or SFR maps. On average high-z galaxies are more clumpier in all maps.
144
Conclusion
Figure 5.14 This plot shows ∆M20sSFR as a function of ∆Ginimass . Blue circles show
the results for the UDF galaxies at z > 0.5 and gray circles show them for the UDF
galaxies at z < 0.5. Error bars show 1-σ scatter. We see that there is correlation
between ∆M20sSFR and ∆Ginimass showing more concentrated galaxies at high-z are
more clumpier. The average ∆M20sSFR for the UDF galaxies at z < 0.5 is 0.06 and
for the UDF galaxies at z > 0.5 is 0.12. However, the average ∆Ginimass changes
from 0.36 to 0.38 from z < 0.5 to z > 0.5.
5.6
Conclusion
We investigated the differences between the spatial distributions of star formation
at high-redshift and low-redshift Universe for galaxies with similar global properties. We studied the resolved stellar populations of these galaxies and compared
the spatial distributions of star formation and mass by measuring the structural
parameters for high-z galaxies and their low-z counterparts. Galaxies at high-z that
we study here have more concentrated stellar content but their star formation is
more extended compared to galaxies with the same global properties at z∼0. In
general, these galaxies are also more clumpier in their star formation distributions
than their local analogs. Therefore, in order to have more clumpy morphology,
high-z galaxies need to have more surface density of the disk compared to their
local analogs.
145
Spatial distribution of star-formation
ID
z spec
Log M∗ [M ]
Log SFR [M yr−1 ]
Log sSFR [yr−1 ]
Log age [yr]
03088
57290
01375
08810
04142
06206
08261
08749
04816
04396
01829
01266
07995
06188
07725
08461
05670
01971
05620
01000
05606
50001
05190
07847
03492
04267
03268
08585
00900
04929
02107
02607
00968
00662
00355
53380
06933
02525
03372
05417
00797
04445
04394
08275
09397
00865
08015
03822
07705
0.13
0.22
0.66
0.74
0.74
0.95
1.09
1.10
1.22
1.22
1.24
1.24
1.29
1.31
1.32
1.43
0.13
0.15
0.21
0.21
0.23
0.24
0.32
0.33
0.34
0.35
0.35
0.38
0.42
0.44
0.53
0.67
0.67
0.67
0.67
0.67
0.73
0.74
1.00
1.10
1.31
0.46
0.67
0.76
1.22
3.06
0.28
0.44
1.30
8.00.0
−0.0
8.50.1
−0.1
9.10.1
−0.1
10.60.0
−0.3
10.30.0
−0.0
0.2
10.8−0.1
0.2
10.5−0.3
10.30.0
−0.0
9.80.5
−0.3
10.30.2
−0.1
9.10.3
−0.2
11.00.0
−0.1
10.70.3
−0.4
0.3
10.7−0.3
0.3
10.5−0.4
10.60.4
−0.1
8.90.0
−0.1
0.1
9.0−0.0
7.50.1
−0.1
7.60.1
−0.1
9.30.0
−0.1
10.00.1
−0.1
8.80.2
−0.0
0.1
10.0−0.1
9.80.0
−0.0
7.80.1
−0.1
9.00.0
−0.0
10.20.4
−0.0
10.40.0
−0.0
10.50.0
−0.0
10.20.1
−0.1
10.70.0
−0.0
0.0
10.2−0.2
0.3
9.5−0.0
9.40.1
−0.0
10.30.1
−0.1
9.40.4
−0.2
0.5
9.8−0.2
10.60.0
−0.0
10.30.0
−0.0
10.80.2
−0.2
10.10.0
−0.0
0.0
10.4−0.2
0.0
10.4−0.3
10.11.0
−0.2
11.20.4
−0.4
0.1
9.4−0.0
0.0
10.7−0.0
8.90.3
−0.3
−2.00.1
−0.1
−0.80.3
−0.3
0.2
0.5−0.2
0.2
1.3−0.0
1.70.0
−0.0
−0.10.6
−0.7
0.3
1.7−2.2
0.1
2.2−0.0
1.40.1
−0.5
−0.10.9
−0.2
0.20.2
−0.1
0.70.3
−0.0
1.20.1
−0.2
−0.10.4
−0.9
0.1
1.0−0.3
0.2
1.2−0.7
−1.40.1
−0.0
−0.50.2
−0.1
−0.70.1
−0.2
−0.70.2
−0.2
−0.60.1
−0.1
−0.80.8
−1.5
0.0
0.3−1.1
0.4
−0.2−0.7
1.30.0
−0.0
−1.00.2
−0.3
0.1
0.6−0.0
0.8
−0.1−0.0
1.10.1
−0.2
1.60.0
−0.0
−0.50.5
−0.3
2.00.0
−0.0
0.0
1.7−0.2
1.20.2
−0.1
0.60.3
−0.3
−0.00.7
−0.5
0.2
1.0−0.1
0.1
1.4−0.2
2.10.0
−0.0
2.00.1
−0.2
1.50.3
−0.0
1.10.0
−0.0
0.0
1.8−0.0
0.2
1.6−0.0
1.80.1
−0.0
1.60.3
−1.0
−0.60.0
−0.1
0.0
1.7−0.0
−0.10.2
−0.1
−10.10.1
−0.2
−9.40.3
−0.3
0.2
−8.7−0.2
−9.40.4
−0.0
−8.60.0
−0.0
−10.90.5
−0.9
−8.90.3
−2.0
0.1
−8.1−0.1
−8.50.3
−0.9
−10.40.7
−0.2
−9.00.3
−0.4
−10.40.4
−0.0
−9.50.4
−0.5
−10.90.3
−0.7
−9.60.4
−0.5
0.4
−9.7−0.5
−10.30.1
−0.0
−9.50.1
−0.1
−8.30.2
−0.3
−8.40.2
−0.3
−10.00.1
−0.1
−10.90.8
−1.5
0.0
−8.5−1.3
−10.30.4
−0.6
−8.70.0
−0.0
0.2
−8.8−0.3
−8.50.0
−0.1
−10.30.4
−0.0
−9.40.2
−0.2
−8.90.0
−0.0
−10.70.4
−0.4
−8.80.0
−0.0
0.1
−8.5−0.0
−8.40.1
−0.1
−8.90.4
−0.3
−10.30.6
−0.4
−8.40.3
−0.4
0.3
−8.5−0.5
−8.60.0
−0.0
−8.30.1
−0.2
−9.40.4
−0.1
−9.10.0
−0.0
−8.60.3
−0.1
−8.90.5
−0.1
−8.40.2
−1.0
−10.00.8
−0.9
−10.10.1
−0.1
0.0
−9.0−0.1
−9.10.4
−0.4
9.20.1
−0.1
8.80.2
−0.2
8.20.2
−0.2
9.40.0
−0.3
8.60.0
−0.0
9.20.3
−0.2
8.40.6
−0.3
7.70.0
−0.1
8.40.8
−0.6
9.40.1
−0.3
8.50.4
−0.3
9.70.0
−0.4
9.20.2
−0.4
9.40.2
−0.2
9.20.4
−0.5
9.30.2
−0.3
9.40.0
−0.1
9.00.1
−0.1
7.90.6
−0.3
8.00.5
−0.3
9.20.1
−0.2
9.50.2
−0.2
8.70.9
−0.0
9.30.2
−0.3
8.80.0
−0.0
8.40.2
−0.2
8.00.1
−0.1
9.70.0
−0.1
8.80.2
−0.1
8.80.0
−0.0
9.40.2
−0.2
8.20.0
−0.0
8.50.0
−0.6
8.00.6
−0.1
8.50.2
−0.4
9.30.2
−0.3
8.30.6
−0.6
8.40.7
−0.6
8.40.0
−0.0
7.90.1
−0.1
9.10.2
−0.3
9.00.0
−0.0
8.70.0
−0.4
8.90.0
−0.5
8.10.9
−0.4
9.00.4
−0.4
9.20.2
−0.1
8.50.0
−0.0
8.60.4
−0.4
Table 5.1 The integrated properties of the UDF sample derived from the SED
fitting to the integrated photometry.
146
Conclusion
UDF ID
Log M∗ [M ]
Log SFR [M yr−1 ]
Log sSFR [yr−1 ]
Log age [yr]
03088
57290
01375
08810
04142
06206
08261
08749
04816
04396
01829
01266
07995
06188
07725
08461
05670
01971
05620
01000
05606
50001
05190
07847
03492
04267
03268
08585
00900
04929
02107
02607
00968
00662
00355
53380
06933
02525
03372
05417
00797
04445
04394
08275
09397
00865
08015
03822
07705
8.20.1
−0.1
8.60.1
−0.1
9.00.2
−0.2
10.30.1
−0.1
—
10.70.1
−0.1
—
—
9.70.1
−0.1
10.20.1
−0.1
0.1
9.0−0.1
10.80.1
−0.1
10.50.1
−0.1
10.60.1
−0.1
10.30.1
−0.1
10.50.1
−0.1
8.90.1
−0.1
0.1
8.9−0.1
7.60.1
−0.1
7.60.1
−0.1
0.1
9.2−0.1
10.00.1
−0.1
8.70.2
−0.1
10.10.1
−0.1
9.60.1
−0.1
7.80.1
−0.1
9.00.2
−0.2
10.10.1
−0.1
10.20.1
−0.1
—
10.10.1
−0.1
—
—
9.70.1
−0.1
9.30.3
−0.1
10.20.1
−0.1
0.1
9.1−0.1
0.3
9.5−0.2
—
—
—
9.90.1
−0.1
—
—
—
—
9.30.1
−0.1
—
8.90.1
−0.1
−1.60.3
−0.3
−0.90.3
−0.2
0.00.4
−0.1
0.2
0.8−0.1
—
−0.10.2
−0.3
—
—
0.70.1
−0.3
−0.20.3
−0.2
−0.20.2
−0.2
0.1
0.4−0.2
0.70.1
−0.1
−0.20.3
−0.3
0.60.2
−0.1
0.60.1
−0.1
−1.20.3
−0.3
−0.50.2
−0.2
0.2
−0.7−0.1
0.2
−0.8−0.2
−0.60.2
−0.2
−0.70.3
−0.3
0.3
−0.3−0.2
−0.30.3
−0.3
0.70.1
−0.2
−1.00.3
−0.2
0.30.3
−0.4
−0.10.3
−0.2
0.80.1
−0.1
—
−0.40.3
−0.3
—
—
0.70.1
−0.3
0.00.3
−0.2
−0.20.3
−0.2
0.1
0.7−0.1
0.1
−0.2−0.0
—
—
—
−1.20.1
−0.1
—
—
—
—
−0.60.3
−0.2
—
−0.40.3
−0.1
−9.90.3
−0.3
−9.50.2
−0.2
−8.60.3
−0.4
−9.60.2
−0.1
—
−10.80.3
−0.3
—
—
−8.60.2
−0.3
−10.40.3
−0.2
−9.20.3
−0.2
−10.40.3
−0.2
−9.80.1
−0.1
−10.80.3
−0.3
−9.80.2
−0.2
−9.80.1
−0.2
−10.30.3
−0.4
−9.50.3
−0.2
0.2
−8.3−0.1
−8.40.2
−0.2
−9.90.3
−0.3
−10.80.4
−0.3
−8.70.4
−0.3
−10.40.3
−0.3
−8.90.2
−0.2
−8.80.3
−0.2
−8.60.3
−0.3
−10.30.3
−0.3
−9.60.2
−0.1
—
−10.60.3
−0.4
—
—
−8.60.2
−0.3
−9.00.3
−0.4
−10.30.3
−0.2
−8.50.1
−0.1
0.3
−8.8−0.3
—
—
—
−9.20.5
−0.4
—
—
—
—
−10.00.3
−0.3
—
−9.20.3
−0.3
9.00.2
−0.2
8.90.2
−0.2
8.90.2
−0.4
9.50.1
−0.2
—
9.60.2
−0.2
—
—
9.30.2
−0.2
9.40.2
−0.2
8.80.3
−0.3
9.50.1
−0.2
9.60.1
−0.2
9.50.2
−0.2
9.40.2
−0.2
9.50.2
−0.2
9.30.2
−0.2
8.90.2
−0.2
7.90.3
−0.2
8.00.3
−0.2
9.20.2
−0.2
9.50.2
−0.2
8.30.3
−0.3
9.40.2
−0.2
8.80.2
−0.2
8.30.3
−0.3
8.90.2
−0.4
9.40.2
−0.2
9.40.2
−0.1
—
9.50.2
−0.2
—
—
9.30.2
−0.2
8.80.3
−0.3
9.40.2
−0.2
8.90.2
−0.2
8.70.4
−0.4
—
—
—
9.50.0
−0.3
—
—
—
—
9.20.2
−0.2
—
8.70.2
−0.2
Table 5.2 The integrated properties of the analog sample of the UDF galaxies
derived from the SED fitting to the integrated photometry.
147
REFERENCES
References
Abazajian, K. N., Adelman-McCarthy, J. K., Agüeros, M. A., et al. 2009, ApJS,
182, 543
Abraham, R. G., van den Bergh, S.,& Nair, P. 2003, ApJ, 588, 218
Abraham, R. G., Ellis, R. S., Fabian, A. C., Tanvir, N. R., & Glazebrook, K.
1999, MNRAS, 303, 641
Abraham, R. G., Tanvir, N. R., Santiago, B. X., Ellis, R. S., Glazebrook, K.,&
van den Bergh, S. 1996, MNRAS, 279, 47
Aird, J., Nandra, K., Laird, E. S., et al. 2010, MNRAS, 401, 2531
Baldwin, J. A., Phillips, M. M., & Terlevich, R. 1981, PASP, 93, 5
Bertin, E.,& Arnouts, S. 1996, A&AS, 117, 393
Bournaud, F., Daddi, E., Elmegreen, B. G., et al. 2008, A&A, 486, 741
Bouché, N., Dekel, A., Genzel, R., et al. 2010, ApJ, 718, 1001
Brinchmann, J., Charlot, S., et al. 2004, MNRAS, 351, 1151
Bruzual, G,& Charlot, S. 2003, MNRAS, 344, 1000
Bundy, K., Fukugita, M., Ellis, R. S., et al. 2009, ApJ, 697, 1369
Cappellari, M.,& Copin, Y. 2003, MNRAS, 342, 345
Coe, D., Benı́tez, N., Sánchez, S. F., et al. 2006, AJ, 132, 926
Conselice, C. J., Yang, C.,& Bluck, A. F. L. 2009, MNRAS, 394, 1956
Conselice, C. J., Bershady, M. A., Dickinson, M., & Papovich, C. 2003, AJ, 126,
1183
Contini T., et al., 2012, A&A, 539, 91
Cowie, L. L., Hu, E. M., & Songaila, A. 1995, AJ, 110, 1576
Daddi, E., Dickinson, M., Morrison, G., et al. 2007, ApJ, 670 156
Dekel, A., Birnboim, Y., Engel, G., et al. 2009, Nature, 457, 451
Dickinson, M., Hanley, C., Elston, R., et al. 2000, ApJ, 531, 624
Diehl, S., & Statler, T. S. 2006, MNRAS, 368, 497
Elbaz, D., Daddi, E., LeBorgne, D., et al. 2007, A&A, 468, 33
Elmegreen, B. G., Elmegreen, D. M., Fernandez, M. X.,& Lemonias, J. J. 2009,
ApJ, 692, 12
Elmegreen, M. E., Elmegreen, B. G., Ravindranath, S.,& Coe, D. A. 2007, ApJ,
658, 763
Elmegreen, B. G., & Elmegreen, D. M. 2006, ApJ, 650, 644
Elmegreen, B. G.,& Elmegreen, D. M. 2005, ApJ, 627, 632
Epinat, B., Tasca, L., Amram, P., et al. 2012, A&A, 539, 92
Gallazzi, A., Brinchmann, J., et al., 2008, MNRAS, 383, 1439
Gallazzi, A., Charlot, S., et al., 2005, MNRAS, 362, 41
Griffiths, R. E., Casertano, S., Ratnatunga, K. U., et al. 1994, ApJ, 435, 19
Förster Schreiber, N. M., Shapley, A. E., Erb, D. K., et al. 2011a, ApJ, 731, 65
Förster Schreiber, N. M., Shapley, A. E., Genzel, R., et al. 2011b, ApJ, 739, 45
Förster Schreiber, N. M., Genzel, R., Bouché, N., et al. 2009, ApJ, 706, 1364
Förster Schreiber, N. M., Genzel, R., Lehnert, M. D., et al. 2006, ApJ, 645, 1062
Förster Schreiber, N. M., Genzel, R., Lutz, D.,& Sternberg, A. 2003, ApJ, 599,
193
Genzel, R., Newman, S., Jones, T., et al. 2011, ApJ, 733, 101
148
REFERENCES
Genzel, R., Tacconi, L. J., Gracia-Carpio, J., et al. 2010, MNRAS, 407, 2091
Genzel, R., Burkert, A., Bouché, N., et al. 2008, ApJ, 687, 59
Genzel, R., Tacconi, L. J., Eisenhauer, F., et al. 2006, Nature, 442, 786
Guo, Y., Giavalisco, M., Cassata, P., et al. 2011a, ApJ, 735, 18
Guo, Y., Giavalisco, M., Ferguson, H. C., et al. 2011b, (arXiv1110.3800)
Huertas-Company, M., Aguerri, J. A. L., Bernardi, M., Mei, S., & Sánchez Almeida, J. 2011, A&A, 525, A157
Huertas-Company, M., Tasca, L., Rouan, D., et al. 2009, A&A, 497, 743
Huertas-Company, M., Rouan, D., Tasca, L., Soucail, G., & Le Fèvre, O. 2008,
A&A, 478, 971
Kauffmann, G., Heckman, T.M., et al. 2003, MNRAS, 346, 1055
Keres, D., Katz, N., Fardal, M., Davé, R.,& Weinberg, D. H. 2009, MNRAS, 395,
160
Kereš, D., Katz, N., Weinberg, D. H., & Davé, R. 2005, MNRAS, 363, 2
López-Sanjuan, C., Le Fèvre, O., de Ravel, L., et al. 2009, A&A, 501, 505
Lotz, J. M., Jonsson, P., Cox, T. J., Croton, D., Primack, J. R., Somerville, R.
S.,& Stewart, K. 2011a, ApJ, 742, 103
Lotz, J. M., Davis, M., Faber, S. M., et al. 2008, ApJ, 672, 177
Lotz, J. M., Primack, J.,& Madau, P. 2004, AJ, 128, 163
Newman, S. F., Shapiro Griffin, K., Genzel, R., et al. 2012, APJ, 752, 111
Newman, S. F., Genzel, R., Förster-Schreiber, N. M., et al. 2012, APJ, 761, 43
Noeske, K. G., Weiner, B. J., Faber, S. M., et al. 2007, ApJ, 660, 43
Overzier, R. A., Heckman, T. M., Kauffmann, G., et al. 2008, APJ, 677, 37
Overzier, R. A., Heckman, T. M., Schiminovich, D., et al. 2010, APJ, 710, 979
Renzini, A. 2009, MNRAS, 398, 58 Robaina, A. R., Bell, E. F., Skelton, R. E.,
et al. 2009, ApJ, 704, 324 Rodighiero, G., Daddi, E., Baronchelli, I., et al. 2011,
ApJ, 739, 40
Shapiro, K. L., Genzel, R., Förster Schreiber, N. M., et al. 2008, ApJ, 682, 231
Shapley, A. E., Steidel, C. C., Erb, D. K., et al. 2005, APJ, 626, 698
Shirazi, M., Brinchmann, J., & Rahmati, A. 2013, arXiv:1307.4758
Teplitz, H. I., McLean, I. S., Becklin, E. E., et al. 2000, ApJ, 533, L65
Thompson, R. I., Storrie-Lombardi, L. J., Weymann, R. J., et al. 1999, AJ, 117,
17
van Dokkum, P. G., Franx, M., Kriek, M., et al. 2008, ApJL, 677, L5
Welikala, N., Connolly, A. J., Hopkins, A. M., & Scranton, R. 2009, APJ, 701,
994
Williams, R. J., Quadri, R. F.,& Franx, M. 2011, ApJ, 738, L25
Windhorst, R. A., Fomalont, E. B., Kellermann, K. I., et al. 1995, Nature, 375,
471
Wuyts, S., Förster Schreiber, N. M., Genzel, R., et al. 2012, ApJ, 753, 114
Wuyts, S., Förster Schreiber, N. M., Lutz, D., et al. 2011, ApJ, 738, 106
Zibetti, S., Charlot, S., & Rix, H.-W. 2009, MNRAS, 400, 1181
149
Nederlandse samenvatting
Sterrenstelsels in al hun verschijningen huisvesten tijdens hun leven miljarden sterren. Het is de aanwezigheid van deze schijnende sterren, die het mogelijk maakt
dat wij ze kunnen waarnemen door de kosmische tijd. Hoewel we sterrenstelsels
voornamelijk door middel van sterlicht observeren, kunnen we deze sterren niet
individueel oplossen, tenzij ze heel dichtbij staan. Daarom wordt al het licht van
de miljarden sterren bij elkaar opgeteld, geanalyseerd met gebruik van sterpopulatiemodellen om informatie over de evolutie van sterrenstelsels eruit af te kunnen
leiden. Sterlicht bereikt ons niet zonder door het interstellair medium (ISM) te
reizen, dat wolken van gas en stofdeeltjes bevat. Gas en stof kunnen het licht
van de sterren absorberen en opnieuw uitzenden, of het richting ons verstrooien,
wat het interpreteren van wat we observeren in sterrenstelsels erg gecompliceerd
maakt. Ondanks al deze moeilijkheden kunnen we, door alleen het totale licht van
sterrenstelsels te analyseren, de globale eigenschappen van sterrenstelsels, zoals
stermassa, stervormingssnelheid en leeftijd, binnen een goede marge vaststellen,
met behulp van sterpopulatiemodellen. Door de sterpopulatiemodellen en photoionisatiemodellen te combineren, kunnen we het emissielijnspectrum van stervormende sterrenstelsels dat ontstaat in het geı̈oniseerde gas rond jonge sterren, verder
analyseren, wat ons een schat aan informatie biedt over de eigenschappen van sterrenstelsels op kleine schaal, bijvoorbeeld het ISM. Dit proefschrift is een poging om
de relatie tussen deze eigenschappen op kleine schaal en de globale eigenschappen
van stervormende sterrenstelsels te begrijpen door de kosmische tijd, door middel
van sterpopulatiemodellen en photo-ionisatiemodellen. Stervorming in sterrenstelsels kan niet alleen door middel van sterren direct worden getraceerd, maar ook
door het effect van sterren op het omringende gas te bestuderen.
Hieronder volgt een vereenvoudigde samenvatting van dit proefschrift. Voor
meer informatie verwijs ik graag naar het desbetreffende hoofdstuk.
Dit proefschrift
De straling van sterren ioniseert het omringende gas en produceert nebulaire emissielijnen doordat het geı̈oniseerde gas recombineert. We kunnen de stervormings151
Nederlandse samenvatting
snelheid meten, samen met andere eigenschappen van sterrenstelsels, met behulp
van emissielijnen van zowel nabijgelegen als verafgelegen stelsels. Voor dit proefschrift maak ik veel gebruik van nebulaire emissielijnen om de intrinsieke eigenschappen van sterrenstelsels in het nabije en verafgelegen Universum te meten. Ik
gebruik ook nebulaire emissielijnen om massieve sterren indirect te traceren en de
eigenschappen van het ISM op kleine schaal te onderzoeken.
Hoofdstuk 2:
De evolutie van massieve sterren is een complex en nog niet goed begrepen proces. Hoewel we bij het observeren van het stellair continuüm gelimiteerd zijn door
interstellaire absorbtie bij λ < 228 Å, kunnen we de nebulaire He ii emissielijn van λ
4686 gebruiken om waardevolle informatie te krijgen over het hoge-energie gedeelte
van stellaire spectrale energie verdelingen. Slechts de meest extreme stervormende
sterrenstelsels vertonen nebulaire He ii emissie en het is algemeen gedacht dat
Wolf-Rayet (WR) sterren de benodigde ioniserende straling hiervoor produceren.
In Hoofdstuk 2 bestuderen we de fysieke eigenschappen van emissielijnstelsels
in de SDSS, die He ii vertonen. Wanneer we ons baseren op deze gegevens, vinden we dat de He ii niet geassocieerd is met WR kenmerken in een groot aantal
stervormende sterrenstelsels met zodanige emissie, bij lage metalliciteit. Het gebrek aan WR sterren heeft belangrijke implicaties voor de evolutie van de meest
massieve sterren bij lage metalliciteit. Niet-homogene sterevolutiemodellen (bijv.
Yoon & Langer, 2006) en ruimtelijke offset tussen de WR sterren en de He ii gebieden (bijv. Kehrig et al. 2008) kunnen twee mogelijke verklaringen zijn voor
deze discrepantie. We tonen ook dat huidige sterpopulatiemodellen de gevonden
ratio tussen He ii en Hβ in gebieden met lage metalliciteit niet kunnen produceren.
Dit resultaat heeft gevolg voor het interpreteren van de waarnemingen van stelsels op hoge roodverschuiving, waar de metalliciteit typisch lager wordt verwacht.
Een ander belangrijk resultaat in deze studie is het definiëren van een nieuw diagnostisch diagram door middel van de He ii /Hβ ratio, welke kan worden gebruikt
om de AGN contributie in stervormende sterrenstelsels die He ii emissie vertonen,
binnen een marge vast te stellen.
Hoofdstuk 3:
De gedetailleerde analyse van verafgelegen sterrenstelsels, wordt gelimiteerd
door hun kleine schijnbare diameter aan de hemel en hun zwakke schijnbare magnitude. Beide beperkingen kunnen worden overkomen door stelsels te observeren
die door zwaartekrachtlenzen optisch zijn vergroot. In Hoofdstuk 3 analyseren
we opgeloste data van de zogenaamde ”8 o’clock arc”, een gelensd Lyman break
sterrenstelsel, samen met HST beelden van dit stelsel, waarop het lensmodel van
dit stelsel is gebaseerd. Met dit lensmodel kunnen we ontrafelen hoe het stelsel
eruitziet zonder het lenseffect en de Hβ emissie, snelheid en snelheidsdispersie
in kaart brengen. We tonen aan dat een eenvoudig roterend schijfmodel niet
in staat is het snelheidsveld van het sterrenstelsel te reproduceren en dat we
een complexer snelheidsveld nodig hebben. Het Hβ profiel van het stelsel toont
een brede blauwverschoven vleugel, wat een uitstroom van 200 km/s suggereert.
De geschatte oppervlaktedichtheid en gasmassa van de ”8 o’clock arc”toont een
152
gasinhoud dat 2.5 tot 7 maal hoger is dan in vergelijkbare stelsels in de SDSS.
Hoofdstuk 4:
De meeste sterren die ons tegenwoordig omringen zijn enige miljarden jaren
geleden gevormd, toen er een piek was in de stervormingsactiviteit van het Universum. De omstandigheden waaronder deze sterren geboren werden is van groot
belang, maar het is erg moeilijk deze te bestuderen door de beperkte observationele resolutie van vergelegen objecten. In Hoofdstuk 4 presenteren we een nieuwe
benadering om direct de dichtheid in stervormende gebieden van sterrenstelsels die
zich dichtbij de piek in stervormingsactiviteit van het Universum bevinden, te vergelijken met die van nabijgelegen stelsels. Om indirect het ISM op hoge roodverschuiving te traceren, gebruiken we emissielijnintensiteiten van vergelegen stelsels.
We kalibreren een nieuwe relatie tussen de [O iii]λ5007/[O ii]λ3727 emissielijnratio
en de ionisatieparameter om het verschil tussen de ionisatieparameters tussen de
samples van hoge en lage roodverschuiving te schatten. We analyseren de ionisatieeigenschappen van een sample van stelsels met hoge roodverschuiving tussen 2.6
en 3.4, waaronder de ”8 o’clock arcën vergelijken die met stelsels met vergelijkbare fysieke eigenschappen in het lokale Universum. We tonen aan dat, nadat we
rekening hebben gehouden met alle verschillen in eigenschappen op grote schaal,
zoals massa en specifieke stervormingssnelheid, de dichtheid in stervormingsgebieden acht maal hoger was in het verleden. Dit impliceert dat de meerderheid van
sterren in het Universum zijn ontstaan in gas dat aan heel verschillende schalingsrelaties voldeed, dan wat we in het hedendaagse Universum zien. Dit is een treffend
resultaat dat sterke beperkingen oplegt aan de omstandigheden van stervorming
in normale sterrenstelsels in het vroege Universum.
Hoofdstuk 5:
In Hoofdstuk 5 bestuderen we de verschillen tussen de ruimtelijke verdeling
van stervorming op hoge en lage roodverschuiving, voor stelsels met vergelijkbare
globale eigenschappen (bijv. stermassa en stervormingssnelheid). We gebruiken
multi-band imaging gegevens die beschikbaar zijn in de HUDF en vergelijken deze
kwantitatief met de gegevens van lage roodverschuiving van de SDSS. Hiermee
bestuderen we de fysieke processen die klonterige stervormingsverdelingen veroorzaken in stelsels met vergelijkbare stervormingsactiviteit op zowel hoge als lage
roodverschuiving. We vergelijken de opgeloste sterpopulaties van deze stelsels door
de structurele parameters van stelsels op hoge en lage roodverschuiving te meten.
We tonen aan dat stelsels op hoge roodverschuiving een meer geconcentreerde sterinhoud hebben, maar dat hun stervorming meer wijd verspreid is in vergelijking
met stelsels met dezelfde globale eigenschappen in het lokale Universum. We laten
zien dat stelsels op hoge roodverschuiving klonteriger zijn qua stervormingsdistributie, dan hun lokale equivalenten. Deze klonterige morfologie wekt de suggestie
dat stelsels op hoge roodverschuiving een grotere dichtheid in de schijf hebben,
dan hun lokale equivalenten.
153
Publications
1. On the spatial distribution of star formation in distant and nearby galaxies
Maryam Shirazi & Jarle Brinchmann
2013, MNRAS, to be submitted.
2. Stars were born in significantly denser regions in the early Universe
Maryam Shirazi, Jarle Brinchmann & Alireza Rahmati
2013, ApJ, submitted, arXiv:1307.4758.
3. The physical nature of the 8 o’clock arc based on near-IR IFU spectroscopy
with SINFONI
Maryam Shirazi, Simona Vegetti, Nicole Nesvadba, Sahar Allam, Jarle
Brinchmann, Douglas Tucker
2013, MNRAS, submitted, arXiv:1306.6282.
4. Strongly star forming galaxies in the local Universe with nebular He II 4686
emission
Maryam Shirazi & Jarle Brinchmann
2012, MNRAS, 421, 1043.
5. Accelerating universe in brane gravity with a confining potential
Maliheh Heydari-Fard, Maryam Shirazi, Shahram Jalalzadeh, Hamid Reza
Sepangi
2006, Physics Letters B, 640, 1.
155
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement