A Photodissociation Region study of NGC 4038", Thomas Bisbas

A photodissociation region study
of NGC4038
Thomas G. Bisbas
University College London
Collaborators: Serena Viti (UCL), Mike Barlow (UCL),
Jeremy Yates (UCL), Tom Bell, and Magda Vasta
NGC 4038
Luminosity
L ~ 7.5 x 1010 Lo
(Gao+ 2001)
They do not qualify
as luminous IR galaxies. However, they
maintain a high star
formation rate.
Distance
D ~ 19.2 Mpc
IAR
(Whitmore+ 1999)
D ~ 22±3 Mpc
(Schweizer+ 2008)
D ~ 13 Mpc
(Saviane+ 2008)
Star Formation
* All regions have high
SFR.
* IAR has the highest.
NGC 4039
* PAH emission is
coming from regions of
recent star formation.
(Brandl+ 2009)
NUMERICAL TREATMENT
* We use 3D-PDR (Bisbas+ 2012) to calculate the PDR chemistry. The code has been
used in the past in various three-dimensional applications (Offner et al. 2013, 2014).
* Although the code can treat any 3D density distribution, we consider one-dimensional
slabs of uniform density. We create a large parameter space with densities spanning four
orders of magnitude (102 < n < 106 cm-3) and UV fields spanning more than three orders of
magnitude (10 < χ < 104.5 Draines).
* The parameter space grid consists of 1400 simulations.
* Vazquez & Leitherer (2005) using Starburst99 concluded that the NGC4038/9 system is
best reproduced assuming solar chemical composition. X-ray spectra agree towards this
direction (Baldi+ 2006). We therefore adopt the following solar abundances:
H = 4x10-1; H2 = 3x10-1; He = 8.5x10-2
Mg = 3.98x10-5; C = 2.69x10-4; O = 4.9x10-4
* We use the UMIST 2012 (McElroy+ 2013) chemical network consisting of 33 species.
* We use a cosmic ray ionization rate of ζCR = 5x10-17 s-1 .The gas-to-dust ratio is taken to
be 100:1.
OBSERVATIONAL DATA
* We use observational data of various lines taken from different instruments. In particular:
CO (1-0)
CO (2-1)
CO (3-2)
CO (4-3)
CO (5-4)
CO (6-5)
CO (7-6)
CO (8-7)
SEST
CSO
CSO
CSO, Herschel
Herschel
Herschel
CSO, Herschel
Herschel
SEST: Aalto+ (1995)
CSO: Bayet+ (2006)
Herschel: Schirm+ (2014)
– All CO observations have been convolved to 43 arcsec.
[CI] 609μm
[OI] 63μm
[OI] 146μm
[CII] 158μm
CSO
ISO-LWS
ISO-LWS
ISO-LWS
CSO: Bayet+ (2006), Gerin & Phillips (2000)
ISO-LWS: Fischer+ (1996)
– All fine-structure lines have been convolved to 80 arcsec.
RESULTS
* For each simulation we obtain
the line intensity for the finestructure lines and the 12CO
ladder.
* The image shows contour plots
of the CO(2-1) intensity for
different values of density
interacting with UV radiation.
* The green stream shows
agreement with observations
(here with CSO).
CO ladder
Fine-structure lines
RESULTS
Constraining the best-fitting density range
and the number of model clouds
* We use the CO ratios from our grid of models as they are often used to constrain the
excitation conditions and chemistry of the molecular gas. We couple these ratios with
an analysis of the best-fitting intensities for each individual line.
* We introduce the scaling factor Nclouds, inferring to the beam filling factor and which
represents the number of clouds needed to reproduce the total observed intensity.
Beam size: 43''
Assumed distance: 20Mpc
The beam size covers
roughly 4 kpc of gas.
Nclouds > 1
Nclouds < 1
* We find that integrating up to Av = 10 mag lead to situation where N clouds < 1 meaning
that less than one molecular cloud would be needed – which is very unlikely!
* We therefore arbitrarily choose to integrate and plot ratios of up to Av = 2 mag only.
RESULTS
Constraining the best-fitting density range
and the number of model clouds
We estimate the number of
clouds using the relation:
If less than one cloud is
needed, then the best-fitting
model is incorrect !
For a maximum depth of
Av=2mag, we find:
Nclouds ~ 104 – 105.
We find the lowest density which corresponds to the bestfitting ratio, and the highest density which corresponds to
the area where the two individual lines intersect.
We find that for low-J
transitions, we obtain a
diagonal best-fitting region
from low densities and low
UV strengths. For high-J
lines there is less
dependence on the UV field
strength. Best fit regions are
located for ~104.5-105.2 cm-3.
RESULTS
Best-fitting UV field strengths based
on PDR fine-structure line diagnostics
* The fine-structure lines are frequently used as
diagnostics for the PDR properties (Kaufman+ 1999)
* The best fitting models for [CI]609μm/[OI]146μm
indicate the presence of a strong radiation field,
which is required to heat the gas in order to excite
the [OI] 146μm line.
* We find that the intensity of UV field is ~102.5
Draines.
* For the [OI] 146μm the model line emission shows
little sensitivity to the density. From this model, we
find a UV field of 101.5 – 102.5 Draines. If this
emission is diluted by the beam, this would suggest
an even stronger UV field, leading to a better
agreement with the [CI]/[OI] ratio. We find that the
dilution factor is ~10-2.
* The [CII] 158μm is matched by a very low radiation
field (<10 Draines ! ): if the PDR surfaces cover the
region and fill the beam, then the lower value cannot
be a result of beam dilution. Therefore either PDRs
are not uniformly distributed or their layers are very
thin, less than 2mag.
RESULTS
Unified picture
RESULTS
Unified picture
* Our findings are generally in agreement
with those of other authors.
* Fischer+ (1996) found typical cloud
properties of PDRs corresponded to
densities of (0.25-1.5) x 104 cm-3 interacting
with UV fields ~ 200-2500 Draines.
* Nikola+ (1998) found that the average
density is ~105 cm-3 with a UV radiation
field of ~500 Draines. The bulk of the [CII]
158μm comes from PDRs that only a minor
fraction of the line emission can come from
ionized gas in HII regions.
* Gilbert+ (2000) found an average density of 104-105 cm-3 and UV field of ~5000 Draines.
* Bayet+ (2006) found a best-fitting density of ~3.5 x 105 cm-3 and UV of ~105.4 Draines.
* Schulz+ (2007) found that CO emission arises from small and dense clumps (<5x104 cm-3)
interacting with 500-3000 Draines.
* Schirm+ (2014) found a warm gas component corresponding to ~1000 Draines, and a
cold component corresponding to ~100 Draines. These intensities are consistent with all
three regions of the Antennae (NGC4038/9, IAR).
RESULTS
X-factor
The contours correspond to XCO values
from the PDR model, whereas the thick
solid line to the Milky Way canonical
value of ~ 2 x 1020 cm-2 (K km s-1)-1 (Strong
& Mattox 1996; Dame+ 2001).
The ellipse corresponds to our best-fit
region found in the 'unified picture' plot.
We find that the best XCO value for NGC
4038 is ~ 3 x 1019 cm-2 (K km s-1)-1. This is
in agreement with the XCO factors derived
for other AGN and starburst galaxies (see
Bell+ 2006, 2007).
The CO emission is related to the column
density of molecular hydrogen according to:
where N(H2) is the column density of H2
along the line of sight. The denominator
corresponds to the integrated antenna
temperature of CO (1-0) line.
** We also note that the atomic carbon
(CI) might be a potentially powerful tracer
for H2. See Papadopoulos+ (2004), Bell+
(2007), Offner+ (2014).
First paper discussing CI as H2 tracer:
Papadopoulos, Thi, Viti, 2004, MNRAS,
251, 147
DISCUSSION
* The mass of a spherical cloud corresponding to a given one-dimensional slab model is
determined by the equation
where R is the maximum depth reached in the PDR model, nH is the H-nucleus density,
mH the hydrogen mass, and μ=1.36 is a mass correction factor to account contribution
from Helium and heavier elements.
* Adopting Avmax = 2 mag, we find masses for individual clouds between <10Mo up to 104
Mo, whereas the total mass of all clouds is 103-1010 Mo. The range of total masses may
form two extremes of the ISM conditions.
* The low-mass component could indicate that the emission we are reproducing with our
models comes from the surface layers of numerous larger clouds, in which case these
masses are those of the total gas contained in PDRs spread across the telescope beam.
* The upper end of the total mass range is comparable to estimates of the total virial
mass within NGC 4038 (Wilson+ 2000). Here, our model clouds would represent the
entirety of the molecular gas contained within this region, implying that all molecular gas
resides within smaller, low Av clouds that fill the beam.
* For the best-fit region, we find a total mass of ~ 3 x 10 4 Mo, corresponding to the
warmer PDR component only, and not to the total gas mass including cold (~10K) gas.
CONCLUSIONS
Key publication: Bisbas et al. 2014, MNRAS, 443, 111
1. We use 3D-PDR to produce 1,400 PDR models of different densities interacting
with different UV radiation fields. We compare our models with observations.
2. The number of cloud surfaces needed to match the different CO J transitions
varies, as different transitions will be tracing different gas components.
3. For low-J CO transitions we obtain a diagonal-best fitting region extending from
low densities interacting with low UV fields, to high densities with high UV fields.
4. For high-J CO transitions, the best fitting region is around 10 4.5 – 105.2 cm-3.
5. We argue that using a single component model to explain all results is less
meaningful, as it leads to large uncertainties in the derived properties.
6. The fine-structure lines show a UV intensity of >10 2.5 Draines and a beam dilution
factor of ~ 10-2 needed to reproduce the observed line emission.
7. The total mass in all clouds is estimated to be from 103 Mo up to 1010 Mo. The low
mass end indicates emission from surfaces of large clouds that suffer beam dilution,
while the high mass end indicates that molecular gas resides within smaller clouds
that fill the beam. Best fitting mass corresponds to ~ 3 x 104 Mo.
8. The CO-to-H2 XCO factor value is less than the canonical Milky Way value, which is
in agreement with the values found for other starburst galaxies and AGNs.

Download Report