Available via license: CC BY 4.0
Content may be subject to copyright.
A Study of Photoionized Gas in Two H II Regions of the N44 Complex in the LMC Using
MUSE Observations
Susmita Barman
1,2
, Naslim Neelamkodan
1
, Suzanne C. Madden
3
, Marta Sewilo
4,5
, Francisca Kemper
6,7
,
Kazuki Tokuda
8,9,10
, Soma Sanyal
2
, and Toshikazu Onishi
8
1
Department of Physics, College of Science, United Arab Emirates University (UAEU), Al-Ain, 15551, UAE; naslim.n@uaeu.ac.ae
2
School of Physics, University of Hyderabad, Prof. C. R. Rao Road, Gachibowli, Telangana, Hyderabad, 500046, India
3
Laboratoire AIM, CEA/DSM - CEA Saclay, F-91191 Gif-sur-Yvette, France
4
CRESST II and Exoplanets and Stellar Astrophysics Laboratory, NASA Goddard Space Flight Center, Greenbelt, MD 20771, USA
5
Department of Astronomy, University of Maryland, College Park, MD 20742, USA
6
European Southern Observatory, Karl-Schwarzschild-Str. 2, D-85748, Garching b. München, Germany
7
Institute of Astronomy and Astrophysics, Academia Sinica, 11F of Astronomy-Mathematics Building, AS/NTU, No.1, Sec. 4, Roosevelt Rd., Taipei 10617, Taiwan
8
Department of Physical Science, Graduate School of Science, Osaka Prefecture University, 1-1 Gakuen-cho, Sakai, Osaka 599-8531, Japan
9
Department of Physics, Graduate School of Science, Osaka Metropolitan University, 1-1 Gakuen-cho, Naka-ku, Sakai, Osaka 599-8531, Japan
10
Department of Earth and Planetary Sciences, Faculty of Sciences, Kyushu University, Nishi-ku, Fukuoka 819-0395, Japan
Received 2021 July 25; revised 2022 March 24; accepted 2022 March 29; published 2022 May 9
Abstract
We use the optical integral field observations with Multi-Unit Spectroscopic Explorer (MUSE)on the Very Large
Telescope, together with CLOUDY photoionization models, to study ionization structure and physical conditions
of two luminous H II regions in the N44 star-forming complex of the Large Magellanic Cloud. The spectral maps
of various emission lines reveal a stratified ionization geometry in N44 D1. The spatial distribution of
[OI]λ6300 emission in N44 D1 indicates a partially covered ionization front at the outer boundary of the H II
region. These observations reveal that N44 D1 is a blister H II region. The [OI]λ6300 emission in N44 C does not
provide a well-defined ionization front at the boundary, while patches of [SII]λ6717 and [OI]λ6300 emission
bars are found in the interior. The results of spatially resolved MUSE spectra are tested with the photoionization
models for the first time in these H II regions. A spherically symmetric ionization-bounded model with a partial
covering factor, which is appropriate for a blister H II region, can well reproduce the observed geometry and most
of the diagnostic line ratios in N44 D1. Similarly, in N44 C we apply a low-density and optically thin model based
on the observational signatures. Our modeling results show that the ionization structure and physical conditions of
N44 D1 are mainly determined by the radiation from an O5 V star. However, local X-rays, possibly from
supernovae or stellar wind, play a key role. In N44 C, the main contribution is from three ionizing stars.
Unified Astronomy Thesaurus concepts: H II regions (694);Large Magellanic Cloud (903);Photoionization
(2060);Interstellar line emission (844)
1. Introduction
Massive stars are the significant sources of ultraviolet (UV)
radiation in galaxies with energies high enough (>13.6 eV)to
ionize the neutral gas in the interstellar medium (ISM). A part
of this high-energy radiation is absorbed by the neutral gas and
heats the surrounding medium, creating the ionized H II
regions. A large fraction of this ionizing radiation escapes into
the diffuse medium, penetrating the molecular gas, if some part
of the H II region is optically thin. This creates an ionization
zone (H
+
), an ionization front (H
0
), and a photodissociation
region (PDR). The ionization front is at the outer boundary of
an H II region that lies inside the PDR. There have been several
studies of Galactic and extragalactic H II regions and PDRs,
e.g., Orion Nebula (Pogge et al. 1992; García-Díaz &
Henney 2007), 30 Doradus (Pellegrini et al. 2010), NGC 364
(Peimbert et al. 2000; Relano et al. 2002), dense H II regions in
IC 10 (Polles et al. 2019), and NGC 595 (Relaño et al. 2010).
The impact of ionizing radiation on the surrounding medium
and the physical properties of H II regions are normally
obtained by strong emission lines in the optical spectrum,
which is mainly populated by hydrogen recombination lines
and forbidden lines of other common elements. These gas
emission lines are sensitive to physical conditions such as
density and temperature; hence, their relative intensities can
probe the physical mechanism involved in the ionization
processes.
The Large Magellanic Cloud (LMC)is an ideal laboratory to
study the properties of H II regions and the massive star
feedback in a low-metallicity galaxy owing to its subsolar
metallicity (Z=0.5 Z
e
; Westerlund 1997), face-on viewing
angle (van der Marel & Cioni 2001), reduced extinction along
the line of sight, and a distance of 50 kpc (Pietrzyński et al.
2019), allowing the spatially resolved observations of the ISM
structures on subparsec scales. Naslim et al. (2015)have
reported on 10 PDRs in the LMC using H
2
pure rotational
transition emission obtained with Spitzer. These regions
include intense H II regions, diffuse ISM clouds, and dense
molecular clouds. We study the individual clouds in detail
using various observations to investigate the high-mass star
formation (see Naslim et al. 2018; Nayana et al. 2020)and its
impact on the ISM. To explore the impact of ionizing radiation
from massive stars on the surrounding medium, we revisit two
HII regions in a well-studied star-forming complex of the
LMC, N44. We examine the physical conditions and ionization
structure of N44D and N44C by comparing the observations
The Astrophysical Journal, 930:100 (14pp), 2022 May 10 https://doi.org/10.3847/1538-4357/ac62ce
© 2022. The Author(s). Published by the American Astronomical Society.
Original content from this work may be used under the terms
of the Creative Commons Attribution 4.0 licence. Any further
distribution of this work must maintain attribution to the author(s)and the title
of the work, journal citation and DOI.
1
with the predictions of the photoionization model, CLOUDY
(Ferland et al. 2017).
The N44 superbubble is one of the brightest star-forming
regions in the LMC, which can be clearly traced by its compact
HII regions along the main shell rim in an Hαmap (Figure 1).
The region is powered by nearly 35−38 hot stars (Oey &
Massey 1995; McLeod et al. 2019)of stellar associations
LH47, LH48, and LH49 (Lucke & Hodge 1970). McLeod et al.
(2019)report the analysis of radiative and mechanical feedback
from massive stars in the H II regions of N44 (N44 A, N44 B,
N44 C, and N44 D)using the optical integral field data from
Multi-Unit Spectroscopic Explorer (MUSE). They used the
He II λ5412 line to identify the feedback driving massive stars,
and they estimated the spectral types and luminosity classes of
these stars for determining the stellar radiative output. Using
the nebular emission-line maps of Hα,Hβ,[SII]λλ6717,
6732, [NII]λ6584, and [OIII]λ5007, they derived the electron
density from the [SII]λ6717/λ6732 ratio assuming a nebular
electron temperature of 10,000 K. They also derived the degree
of ionization using the [OII]/[OIII]ratio, kinematics, and the
oxygen abundances. In addition, they explored the role of
different stellar feedback mechanisms by estimating various
pressures and found that the H II region expansion is mainly
driven by stellar radiation pressure and ionized gas.
In this paper, we further explore the rich nebular emission
lines in the MUSE data set (McLeod et al. 2019)of N44 for a
detailed understanding of their spatial distributions, ionization
structures, and physical conditions. We compare these results
with the photoionization models to interpret the observations of
this star-forming complex. The MUSE archival data of N44
provide many iconic emission lines, such as Hα,Hβ,[SIII]
λ9069, [SII]λλ6717, 6732, [OIII]λλ4959, 5007, [OII]
λλ7318, 7329, [OI]λ6300, and [NII]λ6584, that have only
been partially utilized in McLeod et al. (2019). We exploit
various emission-line ratios to study the behavior of different
ionization zones and compare the observed line ratios with the
best-fit photoionization models to further understand the
physical process.
The N44 main shell is surrounded by several Hαbright
regions (Figure 1). The compact H II region on the southwest
rim of the shell, N44 D, is the most luminous one in N44
(McLeod et al. 2019). We choose the two brightest H II regions
N44 D and N44 C for our study, which show higher degree of
ionization than N44 A and N44 B, and nearly spherical
ionization structures with different feedback characteristics
based on studies by McLeod et al. (2019). N44 D encloses
three hot stars of spectral type O5 V, O9.5 V, and O5.5 V
(McLeod et al. 2019). The second-brightest H II region, N44 C,
is adjacent to N44D on the western rim of the shell, and the
region harbors three hot stars of spectral type O5 III, O9.5 V,
and O5 III (McLeod et al. 2019). One reason for selecting
N44 D is that MUSE observations show an edge-on ionization
front, allowing a detailed study of ionization as a function of
depth into the cloud. At first glance, N44 C does not show such
Figure 1. The Hαmap of the whole N44 superbubble obtained from the Magellanic Cloud Emission Line Survey (MCELS; Smith & MCELS Team 1998). Spectral
types of hot stars (McLeod et al. 2019)are labeled on the Hαemission map, and the magenta boxes indicate the N44 D and N44 C H II regions.
2
The Astrophysical Journal, 930:100 (14pp), 2022 May 10 Barman et al.
a well-defined ionization front at the H II region boundary,
while it shows a higher photon escape fraction relative to
N44 D than reported by McLeod et al. (2019). These two
nebulae show different ionization structures; hence, physical
processes of two different types of H II regions can be
compared. Moreover, the hot star contents of N44 H II regions
are extensively studied, and high spatial resolution observa-
tions are readily available. This study allows us to directly
apply the observed stellar parameters, gas densities, and
emission-line intensities to constrain the photoionization model
without arbitrary assumption, as well as to test their influence
on the geometry of the H II regions.
2. Observations
We used the MUSE archival data of N44 C and N44 D
(program ID: 096.C-0137(A); PI: A. F. McLeod). MUSE is a
large field-of-view (FOV)integral field unit (IFU)panchro-
matic optical instrument on the European Southern Observa-
tory’s(ESO)Very Large Telescope (VLT)in Paranal, Chile.
This instrument provides high spatial resolution observations at
a pixel scale of 0.2″, with a resolving power ranging from 1770
to 3590. The observations of N44 C and N44 D have been
taken on 2015 October 21 and 2016 February 25, with the
MUSE_wfm-noao_obs_genericoffset observing template, in a
wide-field observing mode covering a wavelength range
475−935 nm. The reduced MUSE data are retrieved from the
ESO science archive.
11
The data were reduced using the
MUSE-1.6.1 pipeline. The MUSE pipeline process automati-
cally removes most of the instrumental signatures. The raw data
were preprocessed, and bias subtraction, flat-fielding, sky-
subtraction, wavelength calibration, and flux calibration were
applied. These data were not taken with the Adaptive Optics
System of MUSE, and the seeing-limited angular resolutions
0.98″and 1.30″are achieved for N44 D and N44 C,
respectively. We note that no point-spread function (PSF)
matching was applied for subsequent analysis, and all the
analyzed regions are resolved regardless of the achieved seeing.
Our analysis is based on integrated line flux maps; hence, no
PSF information is retained.
3. Emission-line Maps of N44 D and N44 C
Figure 2shows the extinction-corrected Hαline flux maps of
N44 D and N44 C obtained with MUSE. The ionized gas traced
by the Hαemission shows two H II regions in N44 D, those we
label as N44 D1 and N44 D2, and one in N44 C. Even though
the regions are nearly spherical, their boundaries cannot be
directly specified in a circular aperture. Hence, for determining
the boundaries of these Hα-bright regions based on their
surface brightness, we use the Python package ASTRODEN-
DRO (Rosolowsky et al. 2008). This algorithm identifies and
characterizes the hierarchical structures in the emission-line
map as a structure tree, where each entity is represented as an
isosurface. The local maxima represent the top level of the
dendrogram and are identified from the emission-line map with
the flux >3σ. The isosurfaces (two-dimensional contours)that
surround the local maxima are leaves, branches, and trunks.
The trunks represent parent structures that enclose the branches
connecting two leaves. Further description and methods of
using ASTRODENDRO can be found in Naslim et al. (2018).
Della Bruna et al. (2020)have recently used ASTRODENRO
to identify Hα-bright regions in MUSE maps of NGC 7793 by
applying a similar method. We define the boundary of H II
regions in the Hαmap that are identified as trunks with
ASTRODENDRO. The lower contour levels of these trunks are
taken as the boundaries of H II regions within the chosen
Figure 2. The MUSE Hαmaps of N44 D (top)and N44 C (bottom). The emission-line fluxes are extracted from the polygon regions (white)of N44 D1, N44 D2, and
N44 C for the analysis of photoionized gas in this work.
11
http://archive.eso.org/cms.html
3
The Astrophysical Journal, 930:100 (14pp), 2022 May 10 Barman et al.
observation field. These regions appear as polygons in
Figure 2.
In addition to Hαλ6562.8 emission, the MUSE spectra of
N44 D1 and N44 C show emission due to Hβλ4861, [OIII]
λλ5007, 4959, [OII]λλ7318, 7329, [OI]λ6300, [NII]λ6584,
[SII]λλ6717, 6732, [SIII]λ9069, [Ar III]λλ7135, 7751, and
many He Iand Paschen hydrogen lines (Figure 3). Figure 6
shows integrated line flux maps of Hβ,[SII]λ6717,
[OIII]λ5007, [OII]λ7318, [OI]λ6300, and [NII]λ6584 in
N44 D1, N44 D2, and N44 C. Hβand [OIII]λ5007 emission
shows similar spatial distribution to the Hαemission in both
N44 D and N44 C, while [SII],[OII],[NII], and [OI]emission
in N44 D1 shows a shell structure. The morphology of N44
D2 is irregular with pillars or filamentary structures. The
continuum subtraction is applied by creating continuum maps
from the user-defined line-free portions of the spectrum around
each emission line. The line fluxes (erg s
−1
cm
−2
)of N44 D1,
N44 D2, and N44 C regions are then extracted from the
integrated line flux maps by applying aperture photometry
within the specified regions as polygon structures obtained
from ASTRODENDRO. Only the pixel values with S/N>
10 are considered within all the polygons. These fluxes are then
point-source flux removed by subtracting the point-source
fluxes, which are also extracted within the user-defined
apertures. For uncertainties, we added in quadrature the error
in aperture photometry and an expected 20% calibration error
in every line flux measurement. The error in photometry is the
quadratically added uncertainty in measurements over all pixels
within a region.
To check the data reduction and flux calibration quality of
the data cube retrieved from the MUSE archive, we compared
the Hαline luminosities (erg s
−1
)of N44 C and N44 D1 from
the MUSE archival data to the Hαluminosities presented in
McLeod et al. (2019). In Table 1we show the comparison of
Hαluminosities obtained from the data presented in this work
and McLeod et al. (2019). We note that the line luminosities
obtained from the two data sets agree within the estimated
uncertainties; hence, we are confident to proceed with the
analysis of pipeline-reduced MUSE archival data. The
observed line luminosities (erg s
−1
)are given in Table 2.We
choose two bright regions, N44 D1 and N44 C, for further
analysis with photoionization models.
4. Emission-line Ratios: [SII]/Hα,[NII]/Hα,[OIII]/Hα, and
[OIII]/Hβ
We present the extinction-corrected [SII]λ6717/Hα,
[NII]λ6584/Hα,[OIII]λ5007/Hα, and [OIII]λ5007/Hβ
ratio maps of N44 D and N44 C in Figures 4and 5,
respectively. These ratios allow us to study the ionization
structure of the region. [SII]λ6717/Hαand [NII]λ6584/Hα
are lower at the central regions closer to the ionizing stars,
implying a higher ionization zone, while at the periphery the
values of these ratios are higher, indicating a low ionization
zone. We find that both [SII]λ6717/Hαand [NII]λ6584/Hα
maps of N44 D1 and N44 C show a shell morphology. The
central part of the N44 D1 has a lower [SII]λ6717/Hαratio
(∼0.02), and at the periphery its value is ∼0.3. N44 C shows an
[SII]λ6717/Hαvalue of ∼0.04 at the center and a value of
∼0.20 at the periphery. Similarly, the value of the
[NII]λ6584/Hαratio ranges from 0.03 to 0.20 in N44 D1
and from 0.04 to 0.15 in N44 C. The model calculations by
Allen et al. (2008)have shown that [SII]/Hαand [NII]/Hα
ratios greater than 0.39 and 0.79 would be a result of strong
contributions from shocks. The [SII]/Hαand [NII]/Hαratios
of both N44 D1 and N44 C are well below the values 0.39 and
0.79, respectively, indicating a substantial contribution from
photoionization. However, in the regions outside the boundary
of these H II regions, we find the enhanced [SII]/Hαand
[NII]/Hαratios; hence, the contribution from shocks cannot be
totally ignored.
A similar effect is found in the [OIII]λ5007/Hβand
[OIII]λ5007/Hαratios. The values are higher in the regions
Figure 3. The identified emission lines are labeled in the MUSE spectra extracted from a 1.0″-radius circular region close to the crosscut indicated as a red line in
Figure 6. This is to show the rich emission lines available for analysis in the MUSE observation of N44 D1 and N44 C. The continua of the spectra are normalized to
1, and the peaks of certain strong lines are cut out of the scale for the weaker lines to be visible in the plot properly.
Table 1
Comparison of HαLuminosities Obtained from the MUSE Archival Pipeline
Data (This Work)with the Data Obtained from McLeod et al. (2019)
N44 D1 N44 C
a
L
obs
(erg s
−1
)
b
L
obs
(erg s
−1
)
a
L
obs
(erg s
−1
)
b
L
obs
(erg s
−1
)
(×10
37
)(×10
37
)(×10
37
)(×10
37
)
1.47 ±0.16 1.47 1.39 ±0.15 1.51
Notes.
a
MUSE archival data used in this work.
b
McLeod et al. (2019).
4
The Astrophysical Journal, 930:100 (14pp), 2022 May 10 Barman et al.
closer to the central ionizing stars, showing a high degree of
ionization, while the ratios are lower at the outer regions,
indicating a low degree of ionization. We note that the
integrated [OIII]λ5007/Hαratio of N44 D1 (2.52)is much
higher than those of N44 D2 (1.67)and N44 C (0.55), which
indicates the hardness of radiation field in N44 D1. The
[OIII]λ5007/Hβratio is also higher in N44 D1, representing
the effect of high effective temperature of the ionizing star of
spectral type O5 V. N44 D1, N44 D2, and N44 C show an
[OIII]λ5007/Hβratio of 7.89, 5.18, and 2.07, respectively. In
N44 C, the high values of [OIII]λ5007/Hα(∼1.0–0.3)and
[OIII]λ5007/Hβ(∼3.0—1.0)are near the O5 III star, and the
values decrease toward the edge of the bubble.
5. Hαand HβEmission
5.1. Extinction Correction
The line luminosities are corrected for extinction using the
intensity ratios (Hα/Hβ)
obs
. Since the Hα/Hβratio is
relatively sensitive to temperature, it can be used as a reliable
reddening indicator. This ratio is compared with the theoreti-
cally expected value of the Balmer decrement (Hα/Hβ)
exp
for
case B recombination (Osterbrock & Ferland 2006). Any
deviation from the expected value of the Hα/Hβratio for a
particular electron temperature can be associated with extinc-
tion. We estimate the nebular emission-line color excess
E(B−V)from the Hα/Hβratio using the equation from
Domínguez et al. (2013),
()
() ()
()
()()
⎡
⎣
⎢⎤
⎦
⎥
EB V kk
2.5 log HH
HH .1
HH
10
obs
exp
ll
ab
ab
-= -
ba
The expected value of the (Hα/Hβ)
exp
flux ratio is ∼2.86
(Osterbrock & Ferland 2006). This value is obtained by
assuming case B recombination at an electron temperature of
10,000 K and a density of 100 cm
−3
. Then, following the
extinction curve estimated by Calzetti et al. (2000),
() ( ) ()kR2.659 1.857 1.040 2
V
ll=-+ +
for λ=0.63−2.2 μm and
() (
)()
k
R
2.659 2.156 1.509 0.198
0.011 3
V
2
3
lll
l
=-+ -
++
for λ=0.12−0.63 μm.
Table 2
Luminosities of Observed Emission Lines
Emission Lines N44 D1 N44 D2 N44 C
L
obs
(erg s
−1
)
a
L
int
(erg s
−1
)
b
L
obs
(erg s
−1
)L
int
(erg s
−1
)L
obs
(erg s
−1
)L
int
(erg s
−1
)
(×10
36
)(×10
36
)(×10
36
)(×10
36
)(×10
36
)(×10
36
)
Hα14.70 ±1.62 17.5 ±1.93 5.36 ±0.58 5.45 ±0.69 13.9 ±1.53 21.5 ±1.86
Hβ4.68 ±0.47 6.12 ±0.61 1.71 ±0.24 1.76 ±0.31 3.69 ±0.55 7.24 ±0.75
[NII]λ6584 0.80 ±0.13 0.96 ±0.16 0.29 ±0.07 0.30 ±0.08 0.97 ±0.16 1.50 ±0.20
[OI]λ6300 0.32 ±0.07 0.39 ±0.08 0.07 ±0.03 0.07 ±0.03 0.13 ±0.04 0.21 ±0.05
[OII]λ7318 0.16 ±0.05 0.19 ±0.06 0.06 ±0.03 0.06 ±0.03 0.17 ±0.05 0.24 ±0.06
[OII]λ7329 0.15 ±0.04 0.17 ±0.04 0.05 ±0.02 0.05 ±0.02 0.14 ±0.04 0.20 ±0.04
[OIII]λ4959 12.40 ±0.64 16.1 ±2.13 2.98 ±0.38 3.06 ±0.50 2.61 ±0.04 5.02 ±0.56
[OIII]λ5007 36.90 ±3.85 48.5 ±5.06 8.87 ±0.99 9.11 ±1.30 7.66 ±1.09 15.1 ±1.48
[SII]λ6717 1.11 ±0.16 1.31 ±0.18 0.34 ±0.07 0.35 ±0.08 0.94 ±0.15 1.43 ±0.18
[SII]λ6732 0.83 ±0.13 0.98 ±0.15 0.26 ±0.06 0.26 ±0.07 0.68 ±0.11 1.04 ±0.14
[SIII]λ9069 1.48 ±0.21 1.62 ±0.23 0.60 ±0.10 0.61 ±0.11 1.45 ±0.18 1.82 ±0.20
Notes.
a
L
obs
is the observed luminosity.
b
L
int
is the reddening-corrected luminosity.
Figure 4. [SII]λ6717/Hα,[NII]λ6584/Hα,[OIII]λ5007/Hα, and [OIII]λ5007/Hβratio maps of N44 D1 and N44 D2 regions. The white polygons represent the
regions taken for analysis in this work, and blue crosses are the locations of hot stars.
5
The Astrophysical Journal, 930:100 (14pp), 2022 May 10 Barman et al.
Here k(λ
Hα
)and k(λ
Hβ
)are the extinction curves at Hαand
Hβwavelengths, respectively.
Assuming the ratio of total to selective extinction
R
V
(=A
V
/E(B−V)) =3.1, which is valid at optical wavelength
for the LMC (Gordon et al. 2003), we get k(λ
Hα
)=2.38 and
k(λ
Hβ
)=3.65.
Using the color excess E(B−V), the extinction in
magnitude is obtained by
()( ) ()AkEBV 4
HH
l=-
aa
for the Hαline and
()( ) ()AkEBV 5
HH
l=-
bb
for the Hβline. Then, extinction-corrected Hαluminosity, L
(Hα),is
() () ()LLHH10, 6
A
obs 0.4 H
aa=a
and the extinction-corrected Hβluminosity, L(Hβ),is
() () ()LLHH10. 7
A
obs 0.4 H
bb=b
Here L(Hα)
obs
and L(Hβ)
obs
are the observed luminosities of
Hαand Hβemission, respectively.
The values of the color excess E(B−V)are 0.08 for N44 D1
and N44 D2 and 0.20 for N44 C. The extinction toward
N44 D1 is A
V
=0.25 mag. Our value agrees with the
calculations by Garnett et al. (2000)and Lopez et al. (2014)
for N44. N44 C has a significantly higher extinction, A
V
=0.62
mag.
We applied the same method of extinction correction to
other line emission, and the extinction-corrected luminosities
(erg s
−1
)are given in Table 2.
5.2. Lyman Continuum Photon Flux
The O-type stars in H II regions are the prominent sources of
Lyman continuum photons. These stars deposit a bulk of their
high-energy photons into the surrounding H II region within the
Stromgren radii. If the gas is optically thick in the Lyman
continuum, we expect all the ionizing photons emitted by the star
to be absorbed. However, a significant fraction of these photons
can escape on a larger scale outside of the H II region into the
ISM. This fraction of photon leakage from H II regions needs to
be measured to understand the overall energy budget and to
probe whether the dominant hot stars in the region are responsible
for the photoionization in the surrounding medium. We calculate
the number of Lyman continuum photons (Q)absorbed in the
region surrounding a hot star corresponding to Hαluminosities
by assuming the case B recombination for electron temperature
T
e
=10,000 K and density n
e
=100 cm
−3
. The number of
Lyman continuum photons related to Hαluminosity is obtained
Figure 5. [SII]λ6717/Hα,[NII]λ6584/Hα,[OIII]λ5007/Hα, and [OIII]λ5007/Hβratio maps of the N44 C region. The white polygons represent the regions
taken for analysis in this work, and blue crosses are the locations of hot stars.
6
The Astrophysical Journal, 930:100 (14pp), 2022 May 10 Barman et al.
by Q(Hα)=7.31 ×10
11
L(Hα)photons s
−1
(Kennicutt 1998;
Osterbrock & Ferland 2006).
The numbers of ionizing photons derived from the Hα
luminosity (Q)for N44 D1, N44 D2, and N44 C are tabulated
in Table 3. To calculate the photon escape fraction, we also
need to know the number of total Lyman continuum photons
emitted by the ionizing stars. We adopt the model calculations
(Q
0
)for hot stars of appropriate spectral types from Martins
et al. (2005). The Q
0
values of the only O5 V star in N44 D1; a
combination of three ionizing stars of spectral types O5 III,
O8 V, and O9.5 V in N44 C; and the O5.5 V star in N44 D2 are
given in Table 3.
Using the Qand Q
0
values, we calculate the photon escape
fraction,
()fQQ
Q.8
esc
0
0
=-
McLeod et al. (2019)reported f
esc
∼0.37 and 0.68 for
N44 D1 and N44 C, respectively, using this method. We verify
their determination and find that f
esc
for N44 D1, N44 D2, and
N44 C are 0.36, 0.71, and 0.70, respectively. These values
imply that about 36% of the ionizing photons escape from
N44 D1, 71% from N44 D2, and 70% from N44 C. N44 C
shows a larger amount of photon leakage than N44 D1 and is
more optically thin to the ionizing photon. This observation is
consistent with the study of H II regions in the LMC by
Pellegrini et al. (2012)that boundaries of optically thick
regions are generally characterized by stratification in the
ionization structures. The ionization structure of N44 D1 shows
a well-defined nebular boundary where [OI]emission and [SII]
emission peak at the outer boundary of the Hαemission and
[OIII]emission. The [SII]/Hαratio map clearly shows the
transition region in the ionization structure. N44 C does not
show such an ionization stratification, while a slightly extended
shell structure is found in the [SII]/Hαmap.
Using the number of ionizing photons Q, we can also
estimate the average electron density of emitting gas in an H II
region, 〈n
e
〉. Assuming the spherical nebula where H is fully
ionized, the recombination balance equation is
()QnR
4
3.9
e
B2
HII
3
pa=
Here α
B
is the case B recombination coefficient
∼2.59 ×10
−13
cm
3
s
−1
for gas at T=10,000 K. R
H
II
is the
mean radius of the H II region. Then, the average electron
density from Hαemission is obtained by
⟨⟩ ()n177 . 10
e
Q
R
48
HII
3
=
Here Q
48
(=Q/10
48
)is the number of Lyman continuum
photons derived from the Hαluminosity and R
H
II
is the radius
in parsecs. The 〈n
e
〉derived from Hαemission of N44 D1,
N44 D2, and N44 C are 31, 26, and 38 cm
−3
, respectively.
5.3. Electron Density
Electron density (n
e
)and electron temperature (T
e
)are two
important physical parameters for characterizing an H II region.
The n
e
can be determined from the observed line intensities of
two different energy levels with nearly equal excitation energy
of the same ion. Their line ratios are generally not sensitive to
T
e
. The forbidden line ratio, [SII]λ6717/λ6732, is usually
used to determine n
e
, where [SII]λλ6717 and 6732 emission is
relatively strong in the ionized nebula. Their corresponding
critical density is ∼10
3
cm
−3
, hence probing the low-density
regimes. McLeod et al. (2019)have reported n
e
∼152 ±42
cm
−3
for N44 C and ∼143 ±42 cm
−3
for N44 D1, applying
the analytical solution given in McCall (1984)and assuming a
T
e
of 10,000 K. Toribio San Cipriano et al. (2017)have derived
n
e
∼200 ±150 cm
−3
for a 3.0 ×9.4 arcsec
2
region closer to
the ionizing star in N44 D1 using the [SII]λ6717/λ6732 ratio
obtained from the VLT-UVES spectrum. Lopez et al. (2014)
have reported a relatively low value of n
e
∼60 cm
−3
for the
entire N44 using the flux density of the free–free emission at
3.5 cm. Garnett et al. (2000)derived n
e
<160 cm
−3
for N44D1
using the [SII]λ6717/λ6732 ratio obtained with the 0.9 m
telescope at Cerro Tololo Inter-American Observatory.
McLeod et al. (2019)emphasize that densities derived from
radio emission by Lopez et al. (2014)are expected to be
smaller than those derived from the ratio of collisionally
excited lines (Peimbert et al. 2017). These studies show a
discrepancy in derived values of n
e
for N44 D1 and N44 C;
hence, we calculate the electron densities using [SII]λ6717/
λ6732 ratios of N44 D1 and N44 C derived in our analysis of
MUSE spectra. Electron temperature can be obtained by the
forbidden line ratios [SIII]λ6312/λ9069 and [NII]λ5755/
λ6384; however, the MUSE observations of N44 D1 and
N44 C show very weak [SIII]λ6312 and [NII]λ5755 emis-
sion, which cannot be extracted from the data cube within a
5σdetection threshold in most of the pixels inside the
defined polygons. We calculate the electron density as in
McLeod et al. (2015)by applying the analytical solution in
Table 3
Emission-line Properties
Regions QQ
o
f
esc
〈n
e
〉
a
n
e
[SII]
b
n
e
[SII]
c
(s
−1
)(s
−1
)(cm
−3
)(cm
−3
)(cm
−3
)
N44 D1 1.07 ×10
49
1.66 ×10
49
0.36 31 132 ±50 141 ±43
N44 D2 0.36 ×10
49
1.26 ×10
49
0.71 26 115 ±45 121 ±37
N44 C 1.02 ×10
49
3.37 ×10
49
0.70 38 66 ±40 92 ±35
Notes.
a
〈n
e
〉is the average electron density from Hαemission.
b
n
e
[SII]is the electron density derived using PYNEB.
c
n
e
[SII]is the electron density derived from Equation (11).
7
The Astrophysical Journal, 930:100 (14pp), 2022 May 10 Barman et al.
McCall (1984),
()nR
R
1.49
12.8 5.6713 10 cm . 11
eSII
SII
43
=-
´- ´-
Here R
S
II
is the [SII]λ6717/λ6732 ratio, and electron
temperature is assumed to be 10,000 K as in McLeod et al.
(2019). The derived n
e
for N44D1, N44D2, and N44C using
this method are 141 ±43 cm
−3
, 121 ±37 cm
−3
, and 92 ±35
cm
−3
, respectively. We also estimate n
e
using the publicly
available Python-based package PYNEB (Luridiana et al.
2013)for a comparison. This algorithm includes FIVEL (De
Robertis et al. 1987)and NEBULAR (Shaw & Dufour 1995)
packages for analyzing nebular emission lines. The package
calculates the physical conditions (T
e
and n
e
)for a given set of
emission-line intensities and returns the diagnostic plots. We
use the density-sensitive [SII]λ6732/λ6717 line ratio to
determine n
e
using the diags.getTemDen task in PYNEB for
a given T
e
of 10,000 K. The estimated electron densities from
PYNEB for N44D1, N44D2, and N44C are 132 ±50 cm
−3
,
115 ±45 cm
−3
, and 66 ±40 cm
−3
, respectively (Table 3).
These density values are comparable to the density derived
from Equation (11)within the estimated uncertainties.
6. Structure of Ionized Gas
To investigate the structure of ionization zones in N44 D and
N44 C, we compare the spatial distribution of Hβ,[OIII],[OII],
[OI],[NII], and [SII]emission-line maps (Figure 6). The
spatial distribution of [OIII],[OII], and [OI]of N44 D1 in
Figure 6(a)shows a clear stratification from ionization zones
where [OIII]λ5007 emission peaks around the O5 V star to the
[OI]λ6300 emission at the ionization front. O’Dell & Wen
(1992)have reported that [OI]λ6300 emission arises at the
ionization front behind the PDR, between the ionization zone
and molecular cloud. The majority of the [OI]λ6300 emission
in H II regions is due to the collisional excitation by thermal
electrons and atomic hydrogen via the charge exchange; hence,
the intensity of [OI]λ6300 emission can be a measure of
Figure 6. (a)Spatial distributions of N44 D1 and N44 D2 are shown in [OIII]λ5007 (blue),[OII]λ7318 (red), and [OI]λ6300 emission (green).(b)N44 D1 and
N44 D2 are shown in Hβ(blue),[SII]λ6717 (red), and [NII]λ6584 emission (green).(c)N44 C is shown in [OIII]λ5007 (blue),[OII]λ7318 (red), and [OI]λ6300
emission (green).(d)N44 C is shown in Hβ(blue),[SII]λ6717 (red), and [NII]λ6584 emission (green). The ionization zone is traced by [OIII]and Hβ, the partially
ionized zone is traced by [OII]and [NII], and the ionization front is traced by [OI]λ6300 and [SII]λ6717. The observed spatial profiles in Figures 7and 8are taken
along the crosscuts indicated as red lines in these maps.
8
The Astrophysical Journal, 930:100 (14pp), 2022 May 10 Barman et al.
neutral hydrogen content. [OI]λ6300 in PDRs can also be
collisionally excited by electrons ejected by dust grains and
polycyclic aromatic hydrocarbons (PAHs)that absorb far-UV
radiation emitted by the massive stars. In Figure 6(b),we
compare the spatial distribution of Hβ(blue)and [NII]
λ6584 (red)with [SII]λ6717 (green).[SII]emission is
expected to peak at the ionization front. [NII]emission is
found to be cospatial with [OII]emission that peaks at the
partial ionization zone and concentrated in the outer boundary
of the Hβand [OIII]emission. The structure of N44 D1 is
nearly spherical, with only one important source of ionizing
photons.
In Figure 7we show the spatial profiles of various nebular
emission lines by taking a crosscut along the northwestern edge
of N44 D1. [OII]and [NII]emission is in a thin layer at the
partial ionization zone outside [OIII], but slightly interior to the
[OI]emission. The [OI]emission is concentrated in a thin zone
∼4–7 pc located in the outer boundary of the H II region. Hα,
Hβ,[OIII], and He Iare found to be cospatial in the ionization
zone. In the outer layer of the ionization front in this H II
region, we expect a well-defined PDR with a layer of C II, then
H
2
emission, and a molecular cloud. High spatial resolution
spectroscopic observations in infrared and submillimeter
wavelengths are required for further studies of PDR properties.
A similar ionization structure is reported in the Orion Nebula
HII region. The [OI]λ6300 and [SII]λ6717 emission appears
to peak along a bright bar at the ionization front in the outer
boundary, which forms a thin transition layer of thickness
∼10
15
–10
16
cm between the ionization zone and PDR. The
emission from higher ionization species [OIII]arises away
from the ionization front and close to the ionizing star θ
1
Ori C
(O’Dell et al. 2017;O’Dell 2001; Hester 1991;O’Dell &
Wen 1992).
In N44 C, the [OIII]λ5007 and Hβemission appears to peak
in the interior of the bubble near the ionizing star O5 III
(Figures 6(c)and (d)). We show the spatial profiles of [OIII],
[OII],[SII],[NII], and Hαemission in N44 C, taking a
crosscut from the position centroid of three ionizing stars to the
northeast of N44 C (Figure 7). The spatial distributions of
[OIII]λ5007 and Hβare similar and cospatial in the fully
ionized zone. [OI]λ6300, [OII]λ7318, [NII]λ6584, and
[SII]λ6717 emission does not form a well-defined outer
boundary but appears as patches of emission bars within the
[OIII]λ5007 and Hβemission region.
Figure 7. The spatial profiles of various emission lines in MUSE observations of N44 D1 are shown for comparison with the constant-density model and constant-
pressure model. (a)The observed spatial profiles of [OI]λ6300, [OII]λ7318, [OIII]λ5007, [NII]λ6584, and [SII]λ6717 emission of N44 D1 obtained by taking a
crosscut (red line)shown in Figure 6.(b)The emission-line profiles of [OI]λ6300, [OII]λ7318, [OIII]λ5007, [NII]λ6584, and [SII]λ6717 emission from the
constant-density model of N44 D1 are shown for comparison. (c)The spatial profiles of [OI]λ6300, [OII]λ7318, [OIII]λ5007, [NII]λ6584, and
[SII]λ6717 emission from the constant-pressure model of N44 D1 are shown for comparison. The best-fit model line ratios for these models are given in
Table 4. All the line emissivities are normalized to 1.0.
9
The Astrophysical Journal, 930:100 (14pp), 2022 May 10 Barman et al.
7. Comparison with Photoionization Model
The MUSE observations of N44 in the LMC show that the
two bright H II regions N44 D1 and N44 C have different
ionization geometries and physical characteristics. N44 D1 has
Hαand Hβsurface brightness values 0.03 dex and 0.1 dex
higher than N44 C. [OIII]/Hβand [OIII]/Hαratios in N44 D1
are considerably larger than in N44 C, indicating a higher
degree of ionization. N44 D1 shows a higher [OI]λ6300/Hβ
ratio with a well-defined [OI]λ6300 emission at the H II region
outer boundary, indicating an ionization front.
In the ideal case, the [OI]λ6300 emission dominates at the
outer boundary of the H II region, where the neutral hydrogen
density dominates; hence, we expect most of the Lyman
continuum photons are absorbed by the nebula, and are
ionization bounded. This is an optically thick case, where we
find a shell of [OI]λ6300 emission at the ionization front,
indicating the border of the H II region. In the density-bounded
HII regions, there is no well-defined ionization front
surrounding highly ionized gas, and most of the Lyman
continuum photons leak from the cloud contributing to the
diffuse ionized gas outside of the cloud. In such cases, there is a
weak [OI]λ6300 emission condensation within the H II region
or no shell structure at the boundary. In a blister H II region,
there is a partial ionization front at the boundary, which does
not cover the nebula completely; hence, the photons escape in
certain directions. Our study of the [OI]λ6300 emission in
N44 D1 and N44 C reveals these different observational
properties of H II regions. Moreover, the measurement of the
photon leakage using the Hαemission indicates that N44 D1
and N44 C have photon escape fractions of 36% and 70%,
respectively. These calculations are in agreement with the study
of the photon leakage by McLeod et al. (2019). The remaining
ionizing photons are trapped within the H II region itself and
affect the overall ionization balance. Pellegrini et al.
(2011,2012)reported that the ionization-bounded H II regions
are constrained to an escape fraction <0.6 and those with
density-bounded regions are >0.6.
To further interpret these observations, we compute various
photoionization models for comparing the emission-line ratios
and the geometry of ionization structure. We model the
ionization structure of N44 D1 and N44 C using the photo-
ionization code CLOUDY (Ferland et al. 2017). We derive the
emissivities of prominent ionic species across the H II region,
from the illuminated face of the H
+
region through the partially
ionized zone to the neutral ionization front, where ionizing
radiation has been attenuated and becoming neutral to the
molecular zone. We develop various photoionization models
for N44 D1 and N44 C, and we test which model can better
match the observed ionization geometry of the cloud and
emission-line ratios. To compare the ionization geometry, we
use the spatial profiles of the line emissions along the H II
regions (Figures 7and 8). The emission-line ratios we use
for tests are given in Tables 4and 5. The [SII]λ6717/
λ6732 line ratio is sensitive to electron density, and [OII]
(λ7318+λ7329)/[OIII]λ5007 to the ionization parameter.
[OIII]λ5007/Hα,[OIII]λ5007/Hβ,[OII]/Hβ,[NII]/Hβ,
and [OI]/Hβgive the behavior of different ionization zones
and are also sensitive to the metallicity. These line ratios are
very sensitive to the adopted input parameters. A built-in
optimization program based on the PHYMIR algorithm (van
Hoof 1997; van Hoof et al. 2013; Ferland et al. 2013)is used to
obtain the best-fit model, which applies a χ
2
minimization to
determine the goodness of fit by varying the input parameters.
We vary the total hydrogen density (hden),Φ(H),brems, and
filling factor to find the best agreement between the model with
the observed line ratios and geometry. Sometimes, for a given
set of constraints, some observables are optimized very well
compared to the other sets, and the best-fit model is obtained
by the overall χ
2
. Therefore, we manually fine-tuned certain
parameters until the best-matching line ratios with the
observations are obtained. Finally, the model line ratios are
compared with the observed line ratios, and the best-fit model
was determined by calculating the χ
2
as (Mondal et al. 2017;
Pavana et al. 2019)
() ()MO .12
i
n
iii
2
1
22
/
å
cs=-
=
Here, the number of observed lines is n,M
i
is the model line
ratio, O
i
is the observed line ratio, and σ
i
is the error in the
observed flux ratio. The best optimized model line ratios,
observed line ratios, and χ
2
values are given in Tables 4and 5.
The basic input parameters to CLOUDY require the
geometry of the cloud, intensity of incident ionizing photon
flux, elemental abundances, and gas density.
Figure 8. (a)The observed spatial profiles of [OI]λ6300, [OII]λ7318, [OIII]λ5007, [NII]λ6584, and [SII]λ6717 emission of N44 C obtained by taking a crosscut
(red line)shown in Figure 6.(b)The spatial profiles of [OI]λ6300, [OII]λ7318, [OIII]λ5007, [NII]λ6584, and [SII]λ6717 emission from the constant-density
model obtained for N44 C are shown for comparison. The best-fit model line ratios for this model are given in Table 5.
10
The Astrophysical Journal, 930:100 (14pp), 2022 May 10 Barman et al.
7.1. Geometry
Observations show a clear ionization stratification in
N44 D1. The ionized gas is traced by [OIII]λ5007 and Hβ
emission, and the ionization front is traced by [OI]λ6300 and
[SII]λ6717 emission. The line ratio maps show a nearly radial
symmetry around the ionizing star. This structure is quite
simple to model, as observation shows that there is only a
prominent source of the ionizing photon at the cloud’s interior.
For N44 D1, we calculate an optically thick spherical model
with a covering factor 0.64. We choose an outer radius
R=7.4 pc based on the observed geometry (Figure 7). McLeod
et al. (2019)report a radius containing 90% of the Hα
emission, R
90
=7.4 pc for N44 D1, which agrees with the
adopted radius in our model. The model constitutes the exciting
star at the center of a spherical cloud, surrounded by the layers
of the H II region and PDR.
MUSE observation of N44 C does not show a clear
ionization stratification as in N44 D1. However, observations
indicate that the majority of photons escape from N44 C and
the region harbors three ionizing stars (see Figure 1). There-
fore, we adopt an optically thin open geometry for N44 C with
a covering factor 0.3 and a radius of R=6.93 pc.
7.2. Stellar Continuum
We use OSTAR TLUSTY models in CLOUDY for defining
the stellar continuum. We choose a model with an effective
temperature (T
eff
)of 41,540 K, gravity (logg)of 3.92, and
metallicity of 0.5 Z
e
, which is appropriate for a spectral class
O5 V star (McLeod et al. 2019)in N44 D1. The radiation field
from this O5 V star corresponds to an incident flux of ionizing
photons, log Φ(H)=9.98 photons s
−1
cm
−2
at the ionization
front of N44 D1 traced by the peak of [OI]λ6300 emission.
Here Φ(H)=Q(H)/4πr
2
, where the number of ionizing
photons per second Q(H)is 1.82 ×10
49
photons s
−1
, and ris
the distance from the ionizing star to the cloud illumination
face. We note that the flux of ionizing photons Φ(H)plays a
significant role in the shape of spatial profiles and lines
strengths; hence, we test the models with varying Φ(H)until the
best-fit model is obtained. For N44 D1 we vary the Φ(H)in a
range 9.98−10.60 photons s
−1
cm
−2
.
Observations show that N44 C encloses three hot stars of
spectral types O5 III, O8 V, and O9.5 V; hence, we chose three
hot star model atmospheres from TLUSTY. We adopt models
with T
eff
=39,500 K and logg=3.69 for spectral type O5 III,
T
eff
=33,400 K and logg=3.92 for an O8 V star, and
T
eff
=30,500 K and logg=3.92 for an O9.5 V star with a
metallicity of 0.5 Z
e
.
We expect the X-rays to significantly affect the ionization
balance of the gas, since the observed total X-ray luminosity is
in a comparable range to that of the observed Hαluminosity in
both N44 D1 and N44 C. Chu et al. (1993)have presented the
global X-ray emission in N44 using observations with the
ROSAT satellite. They reported a diffuse X-ray luminosity of
(0.29−3.5)×10
37
erg s
−1
at a characteristic temperature
(1.6−2.5)×10
6
K. We vary the bremsstrahlung temperature
(brems)in a range (1.6−2.5)×10
6
K.
7.3. Abundances
We follow the abundances provided by Toribio San Cipriano
et al. (2017)for C and O and Garnett et al. (2000)for He, N,
Ne, S, and Ar. For the remaining species, we use the standard
values included in CLOUDY for H II regions (Baldwin et al.
1991)and adopt an overall gaseous metallicity of 0.5 Z
e
. Since
Table 4
Model Line Ratios Compared with Observations for N44 D1
Line Ratios Observed Pressure Model χ
2
Density Model χ
2
ratios Φ(H)
10.15
Φ(H)
10.14
[SII]λ6717/λ6732 1.32 ±0.40 1.28 0.01 1.32 0.00
[OII](λ7318+λ7329)/[OIII]λ5007 0.008 ±0.002 0.004 4.0 0.006 1.00
[OIII]λ4959/λ5007 0.33 ±0.07 0.33 0.00 0.33 0.00
[OIII]λ5007/Hβ7.89 ±1.60 6.76 0.50 7.10 0.24
[SIII]λ9069/Hβ0.31 ±0.08 0.36 0.40 0.37 0.56
[OII]λ7329/Hβ0.03 ±0.01 0.02 1.0 0.02 1.00
[SII]λ6717/Hβ0.24 ±0.10 0.36 1.44 0.34 1.00
[NII]λ6584/Hβ0.17 ±0.05 0.14 0.36 0.14 0.36
[OI]λ6300/Hβ0.07 ±0.05 0.15 2.56 0.14 1.96
[OIII]λ5007/Hα2.52 ±0.50 2.40 0.06 2.51 0.00
[SII]λ6717/Hα0.08 ±0.02 0.13 6.30 0.12 4.00
[NII]λ6584/Hα0.05 ±0.02 0.05 0.00 0.05 0.00
T
e
over radius (K)12,491 12,410
n
e
over radius (cm
−3
)175 136
Table 5
Model Line Ratios Compared with Observations for N44 C
Line Ratios Observed ratios
Model
ratios χ
2
[SII]λ6717/λ6732 1.37 ±0.4 1.35 0.002
[OII](λ7318+λ7329)/
[OIII]λ5007
0.04 ±0.02 0.02 1.00
[OIII]λ4959/λ5007 0.34 ±0.01 0.33 1.00
[OIII]λ5007/Hβ2.07 ±0.6 1.75 0.28
[SIII]λ9069/Hβ0.39 ±0.1 0.28 1.20
[OII]λ7329/Hβ0.03 ±0.02 0.02 0.25
[SII]λ6717/Hβ0.25 ±0.08 0.30 0.39
[NII]λ6584/Hβ0.26 ±0.08 0.19 0.76
[OI]λ6300/Hβ0.03 ±0.01 0.03 0.00
[OIII]λ5007/Hα0.55 ±0.14 0.63 0.32
[SII]λ6717/Hα0.07 ±0.02 0.10 2.25
[NII]λ6584/Hα0.07 ±0.02 0.07 0.00
T
e
over radius (K)10 709
n
e
over radius (cm
−3
)84
11
The Astrophysical Journal, 930:100 (14pp), 2022 May 10 Barman et al.
the dust contributes significantly to heating and the overall
equilibrium of the photoionized gas in the cloud, we also
include the dust grains with a metallicity scaled to half solar.
7.4. Density
For gas density at the illuminated face of the cloud, we use
the range of electron density values obtained from the Hα
luminosity and [SII]λ6717/λ6732 line ratio (30 cm
−3
n
e
180 cm
−3
)for N44 D1. For N44 C, we varied the density
between 40 and 100 cm
−3
and finally obtained a value
consistent with the observed electron density obtained from
the [SII]λ6717/λ6732 ratio (66 cm
−3
). We compute models
with constant density and constant pressure in a time-steady
hydrostatic cloud. When we use a filling factor of 1, we find
that most of the predicted line ratios differ from the observed
values, and the ratios considerably change when we use a
filling factor <1. We estimate an approximate range of filling
factors using the relation N
2
e
(rms)=ò
N
e
2(local)(Relano et al.
2002). Here N
e
(rms)is taken as the average electron density
derived from Hαflux and N
e
(local)is the electron density
obtained from the [SII]λ6717/λ6732 ratio.
We note that the total gas pressure and radiation pressure
vary with the position and width of the ionization front. This
can also be a result of varying density at the ionization front as
pressure changes. If density increases, the ionization front
pushes the interior to the cloud, increasing the line emissivities.
We therefore tested two models: one with constant-pressure
distribution and the other with constant-density distribution.
7.5. Constant-pressure and Constant-density Distribution
In the constant-pressure model, the total pressure is kept
constant throughout the cloud. The sum of gas pressure, line
radiation pressure, turbulent pressure, and the outward pressure
of starlight remains constant. At any particular region on the
cloud, the resulting forces are due to the various contributions
to the pressure balance. Therefore, the cloud remains in
hydrostatic equilibrium. The hydrostatic equilibrium model
would balance the pressure gradient due to the kinetic energy
and momentum carried by stellar photon flux, with the thermal
gas pressure exerted by ionized gas. However, this model
significantly changes the gas density and width of the
ionization front. The constant density represents the total
hydrogen density constant throughout the nebula, but electron
and molecular fractions vary with depth. In Figure 8,we
compare the spatial profiles of various emission lines,
[OI]λ6300, [OII]λ7318, [OIII]λ5007, [NII]λ6584, and
[SII]λ6717, obtained for N44 D1 with constant-density and
constant-pressure models. We note that constant-pressure
models provide narrower ionization fronts than constant-
density models. Pellegrini et al. (2007,2009)applied
constant-pressure models to the Orion bar and M17 PDRs for
a self-consistent simulation of H
+
,H
0
, and H
2
regions,
including additional turbulent pressure and magnetic field. In
that model, the physical depth, the separation of H
0
and H
2
, and
overall geometry depend on the gas density. In our constant-
pressure model, we include additional turbulence of 10 km s
−1
.
This provides turbulent line broadening in addition to the
thermal line broadening and slightly increases the widths of the
spatial profiles in the ionization front. Pellegrini et al.
(2007,2009)found that the presence of a magnetic field
significantly increases the physical width of the PDR, as a
result of reduced gas density and increased photon path length.
We calculate the constant-pressure models consisting of gas
pressure and turbulent pressure. We did not include the
pressure due to the magnetic field, as currently there is no
observational evidence of a magnetic field in N44. In Table 4
we show the diagnostic line ratios obtained for both constant-
pressure and constant-density models, along with the observed
line ratios of N44 D1. We note that the constant-pressure model
describes most of the line ratios; however, we obtain reduced
χ
2
for most of the line ratios in the constant-density model. We
also note that the geometry of the [OIII]λ5007 spatial profile
matches reasonably well with the observation in the constant-
density model. The constant-pressure model predicts a
relatively larger electron density than observed. Our studies
show that the depth of the H
+
region, the width of the
ionization front, and the overall geometry of the emission-line
profiles with depth are largely dependent on the thermal gas
pressure and stellar radiation pressure. Within H
+
regions, the
density can be constant with the depth; hence, the predicted line
ratios have a minimal effect on the chosen equation of state.
7.6. Photoionization Model of N44 D1
Even though observations give a photon escape fraction of
0.36 for N44 D1, it does not mean that the H II region is
completely ionization bounded. We expect the photon leakage
from some part of the cloud; hence, that direction can be
density bounded, and the remaining part of the cloud can be
ionization bounded. This model can be a blister-type H II
region as suggested by Pellegrini et al. (2011). It is also
interesting to note that N44 D1 shows relatively large [OIII]/
Hβratios, implying a large ionization parameter, which can be
a result of the high effective temperature of the only ionizing
(O5 V)star in N44 D1. However, to obtain a high [OIII]/Hβ
ratio, we include an additional contribution from a brems-
strahlung component with the temperature in a range
10
6.2
–10
6.4
K. This plasma temperature is inferred from the
studies of the X-ray emission by Chu et al. (1993). These
authors have reported an excess of X-ray emission in the N44 D
region. This excess X-ray emission can be shock driven owing
to massive stellar winds or the off-center supernova remnant.
We find that partial ionization-bounded geometry with
constant-density models can reproduce reasonably well the
observed geometry and line ratios of N44 D1 for a log Φ(H)=
10.14 photons s
−1
cm
−2
, electron density 136 cm
−3
, and
bremsstrahlung temperature 1.67 ×10
6
K.
7.7. Photoionization Model of N44 C
The observed geometry of N44 C does not give a well-
defined ionization front at the H II region boundary, and the
patches of [SII]and [OII]concentrations are found closer to
the cloud centroid. We obtain an optimal model for N44 C by
applying an optically thin constant-density distribution. We
note that the shape of the incident radiation mainly depends on
three fixed stellar atmosphere models, which are appropriate for
three enclosed hot stars O5 III, O8.5 V, and O9.5 V. The
[OIII]/Hβratio in N44 C is relatively low compared to
N44 D1, indicating low ionization parameter, and to obtain this
gas ionization, we did not include any additional bremsstrah-
lung component as in N44 D1. Since the observations show
that 70% of photons escape from N44 C, we use an optically
thin open geometry where the Lyman continuum optical depth
12
The Astrophysical Journal, 930:100 (14pp), 2022 May 10 Barman et al.
(τ
912
)is found to be very low (<1); hence, the majority of the
cloud is open to the Lyman photons. We note that this model
can reasonably well describe the observed geometry and the
line ratios of N44 C. Figure 8shows the model emission-line
profiles as a function of depth in the nebula, and the line ratios
are given in Table 5. Comparison of our models with the
observed geometry and line ratios indicates that N44 C has an
optically thin geometry and the region is being energized
mainly by the outward momentum carried by the radiation
pressure from three ionizing stars.
8. Conclusions
We carry out a detailed analysis of two H II regions in N44
using the integral field optical spectroscopic observations
obtained with MUSE. Comparing these observations with the
photoionization models computed with CLOUDY, we describe
the spatial distribution of emission-line geometry and the
physical conditions. Our results are summarized as follows:
1. Our analysis reveals that the spatial distribution of
various spectral lines in N44 D1 provides a stratified
ionization geometry. The central ionizing star is covered
by a fully ionized hydrogen gas, and at the periphery
there is a well-defined transition zone from O
++
through
O
+
to the neutral zone O
0
.Hα,Hβ, and [OIII]emission
is cospatial and peaks at the fully ionized zone, while
[OII],[NII],[SII], and [OI]emission peaks at the outer
boundary. This region provides an excellent site for
modeling an ideal H II region with a stratified ionization
geometry. The [OI]λ6300 emission of the N44 D1
region reveals a clear boundary/transition zone in the
outer boundary of [OIII]λ5007 and [OII]λ7329 emis-
sion, which does not cover the entire nebula completely,
indicating a partial ionization front. Comparing these
studies with Pellegrini et al. (2011,2012), we suggest that
the N44 D1 is a blister H II region.
2. The spatial distributions of various spectral lines in
N44 C do not show a stratified ionization front at the
boundary. However, it shows the condensations of [SII]
and [NII]emission within the H II region. The
[OI]λ6300 emission is relatively weak in N44 C and
does not show a well-defined outer boundary as in
N44 D1. These observations support relatively higher
photon escape fraction reported by McLeod et al. (2019),
suggesting that N44 C is a density-bounded optically thin
HII region.
3. Our studies reveal that the [OIII]/Hαand [OIII]/Hβline
ratios give a clear indication of a higher degree of
ionization in the regions closer to the ionizing stars and
the ratios are lower toward the boundary of N44 D1 and
N44 C. [SII]/Hαand [NII]/Hβline ratio maps show a
shell structure in both N44 D1 and N44 C. [OIII]/Hαand
[OIII]/Hβof N44 D1 are much higher than N44 C,
indicating a harder radiation field. The effective temper-
ature of the hot star plays a key role here because N44 D1
has a hotter ionizing star (O5 V)than N44 C (O5 III).
4. We use our results of spatially resolved MUSE spectra to
explore the photoionization models with CLOUDY that
can well describe the observed geometry and emission-
line ratios. We find that the constant-density model gives
better geometry and line ratios than the constant-pressure
model in N44 D1. An ionization-bounded geometry with
a partial covering factor can well reproduce the observed
geometry and line ratios, indicating that N44 D1 is a
blister H II region. The spatial profile of [OIII]λ5007
matches very nicely with the observation. Model
calculations reveal that a significant amount of X-ray
emission takes part in shaping the geometry of the
emission-line profiles, in addition to the ionizing
radiation from the O5 V star. The electron density and
temperature values from this model are also consistent
with our measurements from the observed values.
5. An optically thin and open geometry model has been
applied to reproduce the observed geometry and line
ratios in N44 C. The modeling results show that the
N44 C region is mainly energized by the radiation from
three ionizing stars. Our studies indicate that the
ionization structure and physical conditions in N44 D1
and N44 C are set by the stellar radiation pressure and gas
thermal pressure.
This research has been supported by the United Arab
Emirates University (UAEU)through UAEU Program for
Advanced Research (UPAR)grant G00003479 and start-up
grant G00002964. This paper makes use of the following
MUSE (VLT)data: program ID: 096.C-0137(A). F.K.
acknowledges the Ministry of Science and Technology of
Taiwan for the grant MOST107-2119-M-001-031-MY3 and
Academia Sinica Investigator Award, AS-IA-106-M03. M.S.
acknowledges the NASA award, 80GSFC21M0002 (M.S.).
K.T. acknowledges Grants-in-Aid for Scientific Research
(KAKENHI)of Japan Society for the Promotion of Science
(JSPS; grant No. 21H00049).
ORCID iDs
Naslim Neelamkodan https://orcid.org/0000-0001-
8901-7287
Suzanne C. Madden https://orcid.org/0000-0003-
3229-2899
Marta Sewilo https://orcid.org/0000-0003-2248-6032
Francisca Kemper https://orcid.org/0000-0003-2743-8240
Kazuki Tokuda https://orcid.org/0000-0002-2062-1600
Toshikazu Onishi https://orcid.org/0000-0001-7826-3837
References
Allen, M. G., Groves, B. A., Dopita, M. A., Sutherland, R. S., & Kewley, L. J.
2008, ApJS,178, 20
Baldwin, J. A., Ferland, G. J., Martin, P. G., et al. 1991, ApJ,374, 580
Calzetti, D., Armus, L., Bohlin, R. C., et al. 2000, ApJ,533, 682
Chu, Y.-H., Mac Low, M.-M., Garcia-Segura, G., Wakker, B., &
Kennicutt, R. C. J. 1993, ApJ,414, 213
De Robertis, M. M., Dufour, R. J., & Hunt, R. W. 1987, JRASC, 81, 195
Della Bruna, L., Adamo, A., Bik, A., et al. 2020, A&A,635, A134
Domínguez, A., Siana, B., Henry, A. L., et al. 2013, ApJ,763, 145
Ferland, G. J., Chatzikos, M., Guzmán, F., et al. 2017, RMxAA, 53, 385
Ferland, G. J., Porter, R. L., van Hoof, P. A. M., et al. 2013, RMxAA, 49, 137
García-Díaz, M. T., & Henney, W. J. 2007, AJ,133, 952
Garnett, D. R., Galarza, V. C., & Chu, Y.-H. 2000, ApJ,545, 251
Gordon, K. D., Clayton, G. C., Misselt, K. A., Landolt, A. U., & Wolff, M. J.
2003, ApJ,594, 279
Hester, J. J. 1991, PASP,103, 853
Kennicutt, R. C. J. 1998, ARA&A,36, 189
Lopez, L. A., Krumholz, M. R., Bolatto, A. D., et al. 2014, ApJ,795, 121
Lucke, P. B., & Hodge, P. W. 1970, AJ,75, 171
Luridiana, V., Morisset, C., & Shaw, R. A. 2013, PyNeb: Analysis of Emission
Lines, Astrophysics Source Code Library, ascl:1304.021
Martins, F., Schaerer, D., & Hillier, D. J. 2005, A&A,436, 1049
13
The Astrophysical Journal, 930:100 (14pp), 2022 May 10 Barman et al.
McCall, M. L. 1984, MNRAS,208, 253
McLeod, A. F., Dale, J. E., Evans, C. J., et al. 2019, MNRAS,486, 5263
McLeod, A. F., Dale, J. E., Ginsburg, A., et al. 2015, MNRAS,450, 1057
Mondal, A., Anupama, G. C., Kamath, U. S., et al. 2017, MNRAS,474, 4211
Naslim, N., Kemper, F., Madden, S. C., et al. 2015, MNRAS,446, 2490
Naslim, N., Tokuda, K., Onishi, T., et al. 2018, ApJ,853, 175
Nayana, A. J., Naslim, N., Onishi, T., et al. 2020, ApJ,902, 140
O’Dell, C. R. 2001, ARA&A,39, 99
O’Dell, C. R., Kollatschny, W., & Ferland, G. J. 2017, ApJ,837, 151
O’Dell, C. R., & Wen, Z. 1992, ApJ,387, 229
Oey, M. S., & Massey, P. 1995, ApJ,452, 210
Osterbrock, D. E., & Ferland, G. J. 2006, Astrophysics of Gaseous Nebulae
and Active Galactic Nuclei (Singapore: University Science Books)
Pavana, M., Anche, R. M., Anupama, G. C., Ramaprakash, A. N., &
Selvakumar, G. 2019, A&A,622, A126
Peimbert, M., Peimbert, A., & Delgado-Inglada, G. 2017, PASP,129, 082001
Peimbert, M., Peimbert, A., & Ruiz, M. T. 2000, ApJ,541, 688
Pellegrini, E. W., Baldwin, J. A., Brogan, C. L., et al. 2007, ApJ,658, 1119
Pellegrini, E. W., Baldwin, J. A., & Ferland, G. J. 2010, ApJS,191, 160
Pellegrini, E. W., Baldwin, J. A., Ferland, G. J., Shaw, G., & Heathcote, S.
2009, ApJ,693, 285
Pellegrini, E. W., Oey, M. S., Winkler, P. F., et al. 2012, ApJ,755, 40
Pellegrini, E. W., Oey, M. S., Winkler, P. F., Smith, R. C., & Points, S. 2011,
BSRSL, 80, 410
Pietrzyński, G., Graczyk, D., Gallenne, A., et al. 2019, Natur,567, 200
Pogge, R. W., Owen, J. M., & Atwood, B. 1992, ApJ,399, 147
Polles, F. L., Madden, S. C., Lebouteiller, V., et al. 2019, A&A,622, A119
Relaño, M., Monreal-Ibero, A., Vílchez, J. M., & Kennicutt, R. C. 2010,
MNRAS,402, 1635
Relano, M., Peimbert, M., & Beckman, J. 2002, ApJ,564, 704
Rosolowsky, E. W., Pineda, J. E., Kauffmann, J., & Goodman, A. A. 2008,
ApJ,679, 1338
Shaw, R. A., & Dufour, R. J. 1995, PASP,107, 896
Smith, R. C. & MCELS Team 1998, PASA,15, 163
Toribio San Cipriano, L., Domínguez-Guzmán, G., Esteban, C., et al. 2017,
MNRAS,467, 3759
van der Marel, R. P., & Cioni, M.-R. L. 2001, AJ,122, 1807
van Hoof, P. A. M. 1997, PhD thesis, Utrecht Univ.
van Hoof, P. A. M., Van de Steene, G. C., Exter, K. M., et al. 2013, A&A,
560, A7
Westerlund, B. E. 1997, The Magellanic Clouds (Cambridge: Cambridge Univ.
Press)
14
The Astrophysical Journal, 930:100 (14pp), 2022 May 10 Barman et al.
