Dust and Nebular Emission in Star Forming Galaxies

Abstract

Star forming galaxies exhibit a variety of physical conditions, from quiescent normal spirals to the most powerful dusty starbursts. In order to study these complex systems, we need a suitable tool to analyze the information coming from observations at all wavelengths. We present a new spectro-photometric model which considers in a consistent way starlight as reprocessed by gas and dust. We discuss preliminary results to interpret some observed properties of VLIRGs.

Full text

Page 1

arXiv:astro-ph/0105495v1 29 May 2001
Dust and Nebular Emission in Star Forming Galaxies
Pasquale Panuzzo (panuzzo@sissa.it)
S.I.S.S.A., Via Beirut 4, I-34014 Trieste, Italy
Alessandro Bressan and Gian Luigi Granato
Osservatorio Astronomico, Vicolo dell’Osservatorio 5, I-35122 Padova, Italy
Laura Silva
Osservatorio Astronomico, Via G. B. Tiepolo 11, I-34131 Trieste, Italy
Luigi Danese
S.I.S.S.A., Via Beirut 4, I-34014 Trieste, Italy
Abstract.
quiescent normal spirals to the most powerful dusty starbursts. In order to study
these complex systems, we need a suitable tool to analyze the information coming
from observations at all wavelengths. We present a new spectro-photometric model
which considers in a consistent way starlight as reprocessed by gas and dust. We
discuss preliminary results to interpret some observed properties of VLIRGs.
Star forming galaxies exhibit a variety of physical conditions, from
1. Introduction: why we care about nebular emission
A common property of star forming galaxies is the presence of emission
lines in the spectrum. Emission lines are produced by gas that is ion-
ized by UV radiation emitted from young massive stars. Star-forming
regions are the sites where the nebular emission is produced, but these
regions are also dusty environments. It seems a natural goal for a model
of star-forming galaxies to treat contextually the nebular emission and
dust processing.
We present the implementation of nebular emission calculation in
the spectrophotometric code GRASIL (Silva et al., 1998). GRASIL is a
code built to simulate the SED of galaxies including a careful treatment
of starlight dust reprocessing.
The study of emission lines with GRASIL will give us many further
constraints about:
− The dust obscuration: the optical thickness of star-forming regions
can be derived from line ratio while it is very difficult to constrain
optical thickness of ISM only from continuum.
− The star formation rate: emission lines are believed to be powerful
SFR estimators. We can now compare SFR derived from nebular
lines with the results from FIR, UV etc.
ppanuzzo.tex; 1/02/2008; 18:17; p.1

Page 2

2
2. The model
In our model, the galaxy is composed by two main components: the
bulge and the disk. Dust within the disk is divided in: i) molecular
clouds (MCs), where the star formation is active, ii) diffuse medium
(or cirrus).
Young stars are supposed to be born into the MCs, and leave them
progressively as their age increases. This is modelled by considering
that a suitable fraction of the light of SSP is radiated inside the MCs.
This fraction is a function of age, parametrized by an escaping time.
So we have two populations: A) young stars inside MCs, B) older
stars outside MCs.
A molecular cloud is modelled as a thick spherical shell of dense
gas (and dust) around a central point source representing all the star
content of the cloud. The MCs, as well as old stars, are embedded in
the diffuse medium. For more details on energy transport, refer to Silva
et al. 1998 or Silva et al., this conference.
To calculate emission lines we used the photoionization code CLOUDY
94 (Ferland, 1996).
We built a library of HII regions in order to avoid the computation of
entire nebular emission via CLOUDY at every run of the code GRASIL.
The HII regions of the library are calculated for different density,
metallicity and radii. The SEDs of ionizing sources of HII regions are
obtained by integrating the light of each of the two populations for
several assumption of escaping time and IMF.
The SEDs of ionizing sources are parametrized by the number of ion-
izing photons for HI, HeI and OII (Q(H), Q(He), and Q(O)). Q(H) gives
the mass of ionized gas, while the ratios Q(He)/Q(H) and Q(O)/Q(H)
are related to the hardness of ionizing flux and to the degree of ioniza-
tion. In table I it is reported the list of computed lines.
GRASIL calculates the Qs of each population (inside and outside
the MC), interpolates on the library of HII regions, and computes the
nebular emission for each population. Then, the nebular emission is
extincted in the same way as the population that produced it.
3. An application to Very Luminous InfraRed Galaxies
VLIRGs (LIR≥ 1011.5L⊙) are the most powerful star forming galaxies
at low redshift. Recent results show that: VLIRGs do not follow the
Meurer’s relation LFIR/LFUVversus UV spectral index β (Meurer et
al., 1999; Meurer et al., 2000); SFR computed from Hα, even corrected
for extinction, is always lower than SFR derived from FIR (Poggianti
ppanuzzo.tex; 1/02/2008; 18:17; p.2

Page 3

3
Table I. Computed lines: Hydrogen recombination lines (upper pannel),
Elium and metal lines (lower pannel)
Lyα 1216
Ly 926
H 3970
Paγ 10939
Brα 40515
Br 16808
Pf 30386
Huγ 59071
Lyβ 1025
Ly 922
H 3889
Paδ 10050
Brβ 26254
Br 16408
Pf 28724
Huδ 51277
Lyγ 972
Hα6563
H 3835
Pa 9546
Brγ 21657
Pfα 74585
Pf 27577
Hu 46716
Lyδ 949
Hβ 4861
H 3798
Pa 9229
Brδ 19447
Pfβ 46529
Pf 26746
Hu 43756
Ly 937
Hγ 4340
Paα 18752
Pa 9015
Br 18175
Pfγ 37398
Huα 123690
Hu 41700
Ly 930
Hδ 4102
Paβ 12819
Pa 8863
Br 17363
Pfδ 32964
Huβ 75011
Hu 40201
HeII 1640
HeII 2511
HeI 7065
[CII]157.7µm
[NII]6584
[NIII]57µm
[OII]3727
[OIII]4363
[NeIII]15.5µm
[NeIV]2424
[SII]6717
[SIII]9069
[ArIII]7751
HeII 1217
HeI 4471
[CI]9850
CII]2326
[NII]6548
[OI]6300
[OII]7325
[OIII]2321
[NeIII]36µm
[NeIV]4720
[SII]4070
[SIII]6312
[ArIII]5192
HeII 1085
HeI 5876
[CI]8727
CIII]1908
[NII]5755
[OI]6363
[OII]2471
[OIII]88µm
[NeIII]3869
MgII2800
[SII]4078
[SIII]3722
[ArIII]3109
HeII 4686
HeI 6678
[CI]4621
[NI]5199
[NII]122µm
[OI]5577
OIII]1663
[OIII]52µm
[NeIII]3967
[SiII]35µm
[SIII]19
[SIV]10.4µm
[ArIII]3005
HeII 3203
HeI 10830
[CI]609µm
[NI]3466
[NII]205µm
[OI]63µm
[OIII]5007
[OIV]26µm
[NeIII]3343
[SII]10330
[SIII]33.5
[ArII]69850
[ArIII]22µm
HeII 2733
HeI 3889
[CI]369µm
[NI]10400
NII]2141
[OI]145µm
[OIII]4959
[NeII]13µm
[NeIII]1815
[SII]6731
[SIII]9532
[ArIII]7135
[ArIII]9µm
& Wu, 2000). Can the above problems be solved with a correct picture
of the obscuration?
We have simulated with our code a starburst galaxy (barionic mass
of 5 · 1010M⊙) with a quiescent star formation lasting all its history,
plus a final (analitycal) burst. We supposed that, as the burst begins,
half of total (1.2·1010M⊙) gas is under molecular form, and a fraction
(from 70% to 99%) of it is converted in stars during the burst. As
the molecular gas is consumed during the burst, the MCs become more
and more transparent. This is to mimic the consumption of the gas and
the feedback of SNe. The initial optical thickness of MC is chosen to
reproduce observed eq. widht of Hα+[NII] and Hβ (-62.5˚ A and 0.7˚ A)
and the ratio LFIR/LV(88, Poggianti et al., 2001). The time evolution
of the model is shown in figure 1.
The temporal evolution in the β-FIR/FUV plane can be described
in three phases:
1) VLIRG phase: the burst is very obscured; IR emission comes from the
burst while UV comes from the disk. There is no correlation between
UV and IR because they come from different component of the galaxy.
2) UV-bright phase: the burst has almost consumed the gas and the
MCs become transparent; the model moves towards Meurer’s relation.
3) Final phase: the burst is aging and the model departs from the
relation. The galaxy can not be observationally selected as a starburst
any longer.
ppanuzzo.tex; 1/02/2008; 18:17; p.3

Page 4

4
Figure 1. Left: The time evolution of the model in β-FIR/FUV plane. Observed data
are ULIRG observations (filed squares, Meurer et al. 2000) and UV-selected starburst
(open circles, Meurer et al. 2000). Different paths correspond to different fraction
of gas converted in stars. Right: SFR derived from Hα corrected for extinction with
Balmer decrement versus SFR from FIR.
In inital VLIRG phase, Hα corrected for extinction from Balmer
decrement Hα/Hβ can be a wrong estimate of the “true” Hα (and then
of the SFR, see fig. 1 right). In fact in this phase Hβ can be dominated
by the population outside MCs affected by a low extinction; then, the
Balmer decrement is altered because of the selective extinction. This
effect can explain the low LHα0/LBolobserved in VLIRGs (Meurer &
Seibert, 2001). Concluding, we are able to interpret different UV, opti-
cal, IR properties of obscured galaxies within a unique star formation
selective (in age) extinction scenario. We found that the relation β –
FIR/FUV is not unique in the obscured phase. For the same reason, the
correction of Hα from balmer decrement can be wrong in this phase.
References
Ferland, G. J., 1996, Hazy, a brief introduction to CLOUDY, University of Kentucky,
Department of Physics and Astronomy Internal Report
Meurer, G. R., Heckman, T. M., Calzetti, D., 1999, ApJ, 521, 64
Meurer, G. K. Heckman, T. M., Seibert, M., Goldader, J., Calzetti, D., Sanders, D.,
Steidel, C. C., 2000, In Cold Gas and Dust at High Redshift, D. J. Wilner ed., in
press (astro-ph/0011201)
Meurer, G. R., Seibert, M., 2001 in Starburst Galaxies Near and Far, eds. D. Lutz
and L. Tacconi, Springer, in press (astro-ph/0101479)
Poggianti, B. M., Bressan, A., Franceschini, A., 2001, ApJ, 550, 195
Poggianti, B. M., Wu, H., 2000, ApJ, 529, 157
Silva, L., Granato, G. L., Bressan, A., Danese, L., 1998, ApJ, 509, 103
ppanuzzo.tex; 1/02/2008; 18:17; p.4