arXiv:1109.3512v1 [astro-ph.CO] 16 Sep 2011
Observations of the Near- to Mid-Infrared Unidentified Emission
Bands in the Interstellar Medium of the Large Magellanic Cloud
Tamami I. Mori1, Itsuki Sakon1and Takashi Onaka1
Department of Astronomy, Graduate School of Science, The University of Tokyo
Graduate School of Science, Nagoya University
Institute of Astronomy, Graduate School of Science, The University of Tokyo
Department of Astronomy, Graduate School of Science, The University of Tokyo
Not to appear in Nonlearned J., 45.
1Department of Astronomy, Graduate School of Science, The University of Tokyo, 7-3-1
Hongo, Bunkyo-ku, Tokyo 113-0033, Japan
2Graduate School of Science, Nagoya University, Chikusa-ku, Nagoya 464-8602, Japan
3Institute of Astronomy, Graduate School of Science, The University of Tokyo, 2-21-1
Osawa, Mitaka, Tokyo 181-0015, Japan
– 2 –
We present the results of near- to mid-infrared slit spectroscopic observa-
tions (2.55–13.4µm) of the diffuse emission toward nine positions in the Large
Magellanic Cloud with the Infrared Camera (IRC) on board AKARI. The tar-
get positions are selected to cover a wide range of the intensity of the incident
radiation field. The unidentified infrared bands at 3.3, 6.2, 7.7, 8.6 and 11.3µm
are detected toward all the targets, and ionized gas signatures: hydrogen re-
combination lines and ionic forbidden lines toward three of them. We classify
the targets into two groups: those without the ionized gas signatures (Group A)
and those with the ionized signatures (Group B). Group A includes molecular
clouds and photo-dissociation regions, whereas Group B consists of H II regions.
In Group A, the band ratios of I3.3µm/I11.3µm, I6.2µm/I11.3µm, I7.7µm/I11.3µmand
I8.6µm/I11.3µmshow positive correlation with the IRAS and AKARI colors, but
those of Group B do not follow the correlation. We discuss the results in terms
of the polycyclic aromatic hydrocarbon (PAH) model and attribute the differ-
ence to the destruction of small PAHs and an increase in the recombination due
to the high electron density in Group B. In the present study, the 3.3µm band
provides crucial information on the size distribution and/or the excitation con-
ditions of PAHs and plays a key role in the distinction of Group A from B. The
results suggest the possibility of the diagram of I3.3µm/I11.3µmv.s. I7.7µm/I11.3µm
as an efficient diagnostic tool to infer the physical conditions of the interstellar
Subject headings: dust, extinction — Magellanic Clouds — galaxies: ISM — infrared:
galaxies — infrared: ISM
– 3 –
Since the discovery of the 11.3µm band in planetary nebulae in 1973 (Gillett et al.
1973), space infrared missions, and air-borne and ground-based infrared observations have
shown that the major unidentified infrared (UIR) bands appear at 3.3, 6.2, 7.7, 8.6, 11.3,
12.6 and 16.4µm together with some faint features. The UIR bands have been observed in
various astrophysical environments, including photo-dissociation regions (PDRs), reflection
nebulae, planetary nebulae (e.g., Peeters et al. 2002), the diffuse interstellar medium
(ISM) (e.g., Onaka et al. 1996; Mattila et al. 1996), nearby galaxies of various types
(e.g., Dale et al. 2006; Kaneda et al. 2008; Smith et al. 2007), and distant galaxies (e.g.,
Lutz et al. 2005; Sajina et al. 2007). The carriers of the UIR bands are generally thought to
be polycyclic aromatic hydrocarbons (PAHs) or PAH-containing carbonaceous compounds
(e.g., Sakata et al. 1984; Puget & Leger 1989; Papoular et al. 1989; Allamandola et al.
1989). PAHs are excited by absorbing a single UV photon and emit a number of IR
photons corresponding to vibration modes of C-C and C-H bonds. The 3.3µm band is
assigned to C-H stretching modes, the 6.2µm band to C-C stretching modes, the 7.7µm
band to blending of several C-C stretching modes and C-H in-plane bending modes, the
8.6µm band to C-H in-plane bending modes, the 11.3µm band to solo C-H out-of-plane
bending modes, and the 12.6µm band to trio C-H out-of-plane bending modes, respectively
(Allamandola et al. 1989).
Recent laboratory experiments and quantum chemical calculations suggest that the
properties of the UIR bands (e.g., shapes, center wavelengths, interband ratios, etc.) reflect
the chemical and physical properties of PAHs (e.g., molecular structure, size distribution,
ionization state, temperature, etc.), which may be altered in interstellar and circumstellar
environments (Tielens 2008). Therefore the UIR bands have a great potential to be used as
efficient diagnostic tools to infer the physical condition of the ISM even in remote galaxies.
– 4 –
Observations of the diffuse Galactic radiation and normal galaxies have shown very
little variation in the mid-infrared (MIR) UIR band spectra (6–12µm) until recently
(Chan et al. 2001; Lu et al. 2003; Sakon et al. 2004), whereas small variations in the MIR
UIR bands have been reported between the disk and halo regions or the arm and interarm
regions of galaxies (Irwin & Madden 2006; Sakon et al. 2007). Latest Spitzer and AKARI
observations clearly show distinct variations in the MIR UIR spectra in particular galaxies
(Kaneda et al. 2007; Smith et al. 2007; Galliano et al. 2008) for the first time. However,
the diagnostic for the physical conditions of the ISM as well as the chemical and physical
evolution of PAHs in galaxies by means of the UIR bands is not yet fully explored.
In this paper, we present the results of near-infrared (NIR) to MIR spectroscopic
observations of the ISM with different radiation conditions in the Large Magellanic
Cloud (LMC) with the infrared camera (IRC) onboard AKARI (Murakami et al. 2007;
Onaka et al. 2007b). The LMC is a nearby irregular galaxy and is located at the distance of
50kpc from the Milky Way (Feast 1999; Keller & Wood 2006). In addition to its proximity,
the almost face-on orientation (i ∼35◦; van der Marel & Cioni 2001; Olsen & Salyk 2002;
Nikolaev et al. 2004) provides us with a unique opportunity to investigate regions with
different physical conditions without confusion because of the spatial resolution of ∼ 5′′in
the MIR of AKARI/IRC. The IRC spectroscopy has a unique characteristic that it can
obtain a spectrum from 2.5 to 13µm simultaneously with the same slit (Ohyama et al.
2007). This has an advantage for the spectroscopy of extended objects over other space
instruments since the Short Wavelength Spectrometer (SWS) onboard the Infrared Space
Observatory (ISO) had different diaphragms from the NIR to MIR (de Graauw et al. 1996)
and the Infrared Spectrograph (IRS) on Spitzer lacks a channel in the NIR (Houck et al.
2004). Vermeij et al. (2002) report the MIR UIR band ratios of H II regions of the LMC
based on observations with the ISOPHOT/PHT-S instrument on board ISO, but do not
include the 3.3µm band in their analysis because of the low signal-to-noise ratio (S/N) in
– 5 –
the short wavelength channel. The 3.3µm band, in fact, is most sensitive to the smallest
PAHs (e.g., Schutte et al. 1993) and its relative intensity to the MIR UIR bands provides us
with significant information on the average temperature of PAHs, which depends on the size
distribution and the excitation condition of PAHs. In this paper, we investigate variations
in the relative intensity of the UIR bands in the NIR to MIR of the diffuse radiation from
regions with different radiation field conditions in the LMC and discuss them in relation to
the physical properties of PAHs.
In §2, the observation and the data reduction are described together with the selection
of the target positions. The obtained spectra are presented in §3. In §4, the observed
variations in the UIR band ratios in different radiation field conditions are investigated in
terms of the PAH model. Diagnostic of the physical conditions of the observed regions is
also discussed based on the UIR band ratios. A summary and conclusions are given in §5.
2. Observations and Data Reduction
The present study employs the datasets of eight pointed observations (observation IDs:
1400330, 1400346, 1400318, 1402426, 1400324, 1402422, 1400334 and 1400320) collected as
part of the AKARI mission program ”ISM in our Galaxy and Nearby Galaxies” (ISMGN;
Kaneda et al. 2009). All of the observations were performed with the slit spectroscopic
mode with the choice of the grism for the NIR disperser (AOT04 b:Ns; Ohyama et al. 2007).
The NIR spectrum was taken with the grism, NG (2.5–5.0µm, λ/∆λ ∼ 100), in the NIR
channel and the MIR spectra were taken with two grisms, SG1 (4.6–9.2µm, λ/∆λ ∼ 50)
and SG2 (7.2–13.4µm, λ/∆λ ∼ 50), in the MIR-S channel. The NIR and MIR spectra
were obtained simultaneously by means of the beam splitter, which enabled us to get a
– 6 –
continuous NIR to MIR spectrum from 2.5 to 13.4µm of the same slit area of 1′length by
5′′width (Ohyama et al. 2007). The observation parameters are summarized in Table 1.
The accurate slit position is determined from the 3.2µm image (N3), which is taken during
each pointed observation (see §2.3.1), by referring to the positions of point sources in the
2.2. Target Selection
The targets are selected by taking account of the CO mapping data (Mizuno et al.
2001) and the IRAS colors of I25µm/I12µmand I60µm/I100µm, which indicate local star
formation activities (Boulanger et al. 1988; Onaka et al. 2007a). Sakon et al. (2006) have
shown that the extremely large IRAS colors of I25µm/I12µmand I60µm/I100µm(∼3 and
∼0.6, respectively) in the diffuse emission in the LMC can be accounted for by a large
contribution from nearby young (< 30Myr) clusters to the incident interstellar radiation
field and that the small IRAS colors of I25µm/I12µmand I60µm/I100µm(∼1 and ∼0.3,
respectively) are those expected from the heating by the incident radiation field inside
quiescent molecular clouds (e.g., Miville-Deschˆ enes et al. 2002). According to these results,
we select several infrared bright positions with different IRAS colors as the targets of the
present study, where molecular clouds are recognized on the CO maps (Mizuno et al. 2001;
Fukui et al. 2008). Since the beam size of the IRAS data is larger than the size of the slit of
the AKARI/IRC, we also derive the AKARI color of IL24/IS11from the dataset of Ita et al.
(2008), where IL24is the flux density at the L24 (24µm) band and IS11is that at the S11
band (11µm) of the AKARI/IRC, to obtain the local radiation field conditions. The flux
density is measured over an aperture of 5′′in diameter around the central position of the
slit. The IRAS and AKARI colors of the present targets are summarized in Table 2. The
trend of the AKARI color of IL24/IS11is very similar to that of the IRAS I25µm/I12µmcolor,
– 7 –
whereby we confirm that the selection based on the IRAS colors is in fact relevant to the
purpose of the present study. Hence, the target positions cover a wide range of incident
radiation field conditions, including molecular clouds, PDRs and H II regions.
In Figure 1, the slit positions of those datasets are shown over the false color image
of the LMC obtained by the AKARI IRC LMC survey program (Ita et al. 2008). The S11
band images of a 10′× 10′area including each slit position are also shown in Figures 2a–h.
The slit is positioned at regions without apparent point sources in all the positions except
for Position 8. At Position 8, a bright point-like source is recognized in part of the slit.
We split the spectrum of Position 8 into 2 parts: the one including the point-like source
(Position 8-1) and the other without the point-like source (Position 8-2). For the other
positions, the spectrum is extracted over the almost entire slit length of 40–50′′.
The AKARI color of IL24/IS11and the IRAS colors of I25µm/I12µmand I60µm/I100µmat
Positions 1, 2, 3, 4, 5 and 6 exhibit only a limited range of 0.5–1.8, 1.0–1.9, and 0.3–0.5,
respectively. These targets are not associated with SWB 0 or I type star clusters in
Bica et al. (1996), indicating that they are surrounded by relatively quiescent environments
(Sakon et al. 2006). The AKARI and IRAS colors at Positions 7, 8-1 and 8-2 exhibit large
values. These targets are located in regions associated with OB star clusters as well as
Herbig Ae/Be star clusters (N158-O1 and N158-Y1 at Position 7 and N159-Y4 at Position
8; Nakajima et al. 2005), confirming that the infrared colors manifest recent star-formation
activities in the ISM.
2.3. Data Reduction
The present data reduction basically follows the standard toolkit for the IRC
spectroscopy. However, most of the present targets are faint and require careful data
– 8 –
processing. Thus, some part of the process is carried out separately from the toolkit with
special care. Details of the data reduction process are described in the following.
2.3.1. Slit spectroscopy with AKARI IRC/NIR
During a single pointed observation with AOT04 grism mode, eight to nine exposure
frames of NIR grism spectroscopic (NG) data and one exposure frame of 3.2µm imaging
(N3) data are taken with the IRC/NIR (Ohyama et al. 2007). The dark current is measured
in one frame each at the first and last parts of the pointed observation. A single exposure
frame consists of one short-exposure and one long-exposure images. In the present study,
only the long exposure data are used. The dark image for each NIR observation is obtained
by averaging three long-exposure images of the dark current, which are collected from the
adjacent pointed observations including itself to correct for any high-energy ionizing particle
(hereafter cosmic-ray) effects by a 1.5-σ-clipping method. In the present analysis, only the
dark current data measured in the first part of each pointed observation are used to avoid
the latent image effects and subtracted from the observation images.
We recognize small shifts in position due to the pointing instability by at most ∼ 5′′
during each pointed observation except for Position 5, where an extraordinary large shift
of ∼ 15′′is found. The shift in the direction parallel to the slit is corrected so that the
spectra of the same area of the sky are extracted. Because the shift in the orthogonal
direction is uncorrectable, the exposure frames shifted in the direction perpendicular to the
slit by more than a pixel (∼ 1.5′′) as well as those affected by severe artifacts are discarded,
except for Position 5 (see §2.3.3). The remaining images are averaged taking account of the
shift in positions in the direction parallel to the slit by a 1.5-σ-clipping method to remove
cosmic-ray events and the artifacts.
– 9 –
The most critical part in the data-reduction of NG spectra is the removal of artificial
patterns as well as the foreground components originating from the zodiacal light and
diffuse Galactic emission. In NIR observations, pixels saturated by cosmic-ray hits or
extremely luminous objects often produce artificial line-like patterns, sometimes termed as
”column pull-down” and ”multiplexer bleed” (Pipher et al. 2004). Particularly an artificial
line-like pattern in the direction parallel to the slit mimics emission or absorption features in
the slit spectrum. The position of the artificial line-like structure sometimes differs among
different exposures even in the same pointed observation. When an artificial line structure
emerges in a certain exposure image, it is removed by replacing the data of the affected
pixels with those of the unaffected pixels of other exposure images at the same position.
To estimate the foreground components from the zodiacal and diffuse Galactic emission,
we employ the datasets obtained at positions off the LMC with AOT04 a;Ns (observation
IDs:1500719 and 1500720). In these observations the NIR spectra are taken with the prism
mode (NP) instead of the grism mode (NG). Because the emission at the off-LMC positions
are too faint for the observations with the NG mode, we use the data with the NP mode to
obtain a reliable foreground spectrum. The slit positions of both observations are centered
at (α2000,δ2000) = (06h00m00.s0,−66◦36′30′′), which is almost at the same ecliptic latitude
(β ∼ −90◦) as that of the LMC (∼ −85◦), but is at ∼ 9.7◦away from the center of the
LMC. The observation log of the off-LMC position is also given in Table 1.
2.3.2.Slit spectroscopy with AKARI IRC/MIR-S
The data reduction procedures including the dark-current subtraction and the
cosmic-ray correction for MIR-S spectroscopic observations are basically the same as those
for NIR spectroscopic observations. During a single pointed observation with AOT04, four
exposure frames of SG1 data, four to five exposure frames of SG2 data, and one exposure
– 10 –
frame of 9.0µm imaging (S9W) data are taken with the MIR-S channel. The dark current
is measured in one frame each in the first and the last parts of the pointed observation. A
single exposure frame consists of one short-exposure image and three long-exposure images.
The dark image for each MIR-S observation is obtained by averaging three long-exposure
images of the dark current by a 1.5-σ-clipping method to correct for the cosmic-ray effects.
In this process, only the dark current data measured in the first part of each pointed
observation sequence are used to avoid the latent image effects. The same shifts in position
as those recognized in the NIR data are expected in the MIR-S observations because the
same field-of-view is shared by the NIR and MIR-S channels and they are corrected in the
same manner as in the NIR data. The shift in position among three long-exposure images
taken in a single frame is negligible in most cases and, therefore, they are averaged by a
1.5-σ-clipping method to remove the cosmic-ray effects.
The subtraction of the foreground components (zodiacal light and diffuse Galactic
emission) is a more serious problem in the data reduction of MIR spectroscopy than in NIR.
In addition, the MIR detector suffers scattered light originating in the scattering within
the detector array (e.g., Sakon et al. 2007). To estimate the foreground emission and the
scattered light component, the SG1 and SG2 spectra collected at the position off the LMC
(Observation IDs: 1500719 and 1500720) are used. The spectrum at off-position is obtained
by averaging the two observations and is subtracted from the spectra of the target positions.
The MIR-S spectra are basically dominated by the zodiacal light and thus the scattered
light component of the off-position spectrum is almost the same as that in the spectra of
the target. Therefore, the subtraction of the off-position spectrum works effectively not
just to remove the foreground emission but also to correct for possible artifacts and greatly
improves the resultant spectra.
– 11 –
2.3.3. Continuous spectra from NIR to MIR
Using the spectral response function of each module, we obtain NG, SG1 and SG2
segmental spectra at the same area of the sky except for positions 5 and 8-1 (see below).
Each segmental spectrum is truncated at the wavelengths where the S/N becomes low:
NG is truncated at 2.55µm and 4.9µm, SG1 at 5.5µm and 7.9µm and SG2 at 7.9µm
and 13.4µm. Then, the correction for the slit efficiency for extended sources (Sakon et al.
2008) is applied and continuous spectra from 2.55 to 13.4µm are obtained with a small
gap between 4.9 and 5.5µm. Because of the severe artifacts (column-pulldown) the NG
spectrum is truncated at 3.8µm and 4.5µm for Positions 2 and 5, respectively. Note that
although there is a small gap between the NG and SG1 segments, all the NG, SG1, and
SG2 segmental spectra are smoothly connected to each other without scaling, suggesting
that the subtraction procedure of the foreground emission and scattered light works well
and reliable spectra are obtained.
For Positions 5 and 8-1, we cannot obtain segmental spectra precisely at the same
region of the sky between SG1 and SG2 because the SG1 and SG2 observations are not
carried out simultaneously during a pointed observation. A relatively large positional shift
(∼ 15′′) recognized during the pointed observation for Position 5 prevents us from obtaining
the SG1 and SG2 spectra from the same region of the sky. For Position 8-1, the positional
stability during the pointed observation is almost the same as the other observations.
However, a small shift in the position of the point-like source in the slit of 5′′width between
the SG1 and SG2 observations changes the source flux to some extent and makes a small
difference in the flux level between the SG1 and SG2 spectra. Since the positional stability
was better when the SG2 spectrum was taken than the SG1 spectrum for the observations
of both positions, only the NG data taken simultaneously with the SG2 are used for the
data at both positions. Then, the SG1 spectrum is scaled to match with the SG2 spectra
– 12 –
in the spectral region of 7.3 to 7.9µm. The scaling factors are 0.9 and 0.8 for Position 5
and Position 8-1, respectively. The scaling of the SG1 spectrum is taken into account in
the derivation of the band intensity ratios (§3) and does not make serious effects on the
The resultant spectra toward our target positions are shown in Figure 3. The UIR
bands are clearly seen at 3.3, 6.2, 7.7, 8.6 and 11.3µm in every spectrum. A weak feature
around 5.70µm is also seen at Positions 7 and 8-A (Boersma et al. 2009). The hydrogen
recombination line of Brα 4.05µm is detected in four of the spectra, and the hydrogen
recombination lines of Brβ 2.63µm, Pfγ 3.75µm, and Pfβ 4.65µm as well as the forbidden
lines of [Ar II] 6.98µm, [Ar III] 8.99µm, [S IV] 10.51µm and [Ne II] 12.81µm are detected
in three of the spectra. We note that a small bump seen around at 9.6µm is an artifact,
originating from the latent of the S9W exposure frame taken just before the SG2 exposure
frames. Therefore, the spectral data from 9.4 to 9.8µm are not used in the following model
fit, and we cut 9.4–9.8µm from spectra in Figure 3.
To derive the intensity of each UIR band and emission line, we fit the observed spectra
(λ − λkl)2+ (γkl/2)2+
(λ − λkg)2
kg/(8 · ln2)
where λ is the wavelength. The first term represents the continuum, which is modeled with
a polynomial function of the 5thorder and is constrained to be non-negative. The second
and third terms correspond to the UIR bands and the emission lines, respectively. For the
UIR bands, we include 17 components centered at 3.30, 3.41, 3.46, 3.51, 3.56, 5.70, 6.22,
6.69, 7.60, 7.85, 8.33, 8.61, 10.68, 11.23, 11.33, 11.99, 12.62µm according to Smith et al.
– 13 –
(2007). Except for the 3.46, 3.51 and 3.56µm components the UIR band components have
band widths larger than or similar to the spectral resolution of AKARI/IRC and are thus
modeled with Lorentzian profiles (the second term), where λklis the center wavelength,
γklis the FWHM and bklis the height of each component. As for the 3.30, 3.41, 5.70,
6.22, 6.69, 7.60, 7.85, 8.33, 8.61, 10.68, 11.23, 11.33, 11.99, 12.62 components, γklis fixed
to the best-fit value obtained for the spectrum that has the highest S/N ratio with the
spectral resolution of AKARI/IRC as the minimum value. The adopted value of γklfor
each component is summarized in Table 3. Only the height bklis left as a free parameter.
The integrated intensity of each component is calculated as πbkl· γkl/2. In the following
analysis, the 7.7µm band is defined as a combination of the 7.60 and 7.85µm components,
and the 11.3µm band as a combination of the 11.23 and 11.33µm components. We note
that the 12.6µm band is defined as one component of the 12.62µm component, because
the red-wing component, such as the 12.69µm component suggested by Smith et al. (2007),
is not detected at a significant level in all the spectra due to the poor S/N and the low
The 3.46, 3.51 and 3.56µm components as well as the emission lines, which have
the band widths smaller than the spectral resolution of AKARI/IRC, are modeled with
Gaussian profiles (the third term), where λkgis the center wavelength, γkgis the FWHM
and ckgis the height of each component. In the fit, γkgis fixed to match with the spectral
resolution of AKARI/IRC at the corresponding segment. Only the height ckgis a free
parameter in the fitting. The integrated intensity of each Gaussian component is given by
(π/ln2)1/2·ckg·γkg/2. The adopted values of γkgfor the 3.46, 3.51 and 3.56 µm components
and the emission lines are summarized in Tables 4 and 5, respectively.
The best-fit model spectra given by Eq. (1) are plotted together with the observed
spectra in Figure 4. The residual spectra are also plotted in the lower panel of each plot.
– 14 –
The derived intensities of the major UIR bands at 3.3, 6.2, 7.7, 8.6, 11.3, and 12.6µm
and the emission lines of Brα , Brβ , Pfγ , Pfβ, [Ar II]6.98, [Ar III]8.99, [S IV]10.51, and
[Ne II]12.81 are summarized in Tables 6 and 7, respectively, where those detected with more
than 2σ are indicated. The uncertainties in the intensity are estimated from the fitting
errors taking account of the observational uncertainties.
As shown in Table 7, the 3.3, 6.2, 7.7, 8.6, 11.3, and 12.6µm bands are detected at
every target position except for the 12.6µm band at Position 5, which shows the faintest
emission. The 5.70µm band are detected only at Positions 7 and 8-A. On the other hand,
as shown in Table 7, all of the hydrogen recombination lines of Brα 4.05µm, Brβ 2.63µm,
Pfγ 3.75µm, and Pfβ 4.65µm and the fine structure lines from the ionized gas of [Ar II]
6.98µm, [Ar III] 8.99µm, [S IV] 10.51µm and [Ne II] 12.81µm are detected at Positions 7,
8-1 and 8-2, whereas none of them are detected at Positions 1, 2, 3, 4, 5 and 6 except for
Brα detected barely at Position 5. Taking account of the high ionization potentials of 27.63
eV, 34.83 eV and 21.56 eV to form Ar2+, S3+and Ne+(Allen 1973), respectively, Positions
7, 8-1 and 8-2 are exposed to the hard incident radiation field powered by young massive
stars and are associated with H II regions. This view is consistent with the radiation field
conditions suggested by the IRAS and AKARI colors. Based on the characteristics of
the observed spectra, we classify the targets into two groups: ”Group A”, which includes
Positions 1, 2, 3, 4, 5 and 6 and ”Group B”, to which Positions 7, 8-1 and 8-2 belong. The
members of Group A are supposed to be exposed to incident radiation fields of weak to
moderate intensities and consist mostly of molecular clouds and PDRs. Group B members
are all associated with H II regions.
We investigate the effects of extinction on the spectra observed at the present target
positions according to Dobashi et al. (2008). The visual extinction AV toward the present
target positions ranges from 0.0 to 2.5 as shown in the last row of Table 1. We assume
– 15 – Download full-text
the ”LMC avg” extinction curve provided by Weingartner & Draine (2001) to estimate
the infrared extinction. We also estimate the value of AV from the observed ratio of Brβ
to Brα at Positions 7, 8-1, and 8-2, assuming the Case B condition of Te= 104K and
ne= 104cm−3(Storey & Hummer 1995). The UIR band intensities corrected with AV
provided by Dobashi et al. (2008) differ from those corrected with AV estimated from the
Case B condition by less than 10% at Positions 7, 8-1, and 8-2. We adopt those corrected
with AV provided by Dobashi et al. (2008) at these positions for consistency with the
other target positions. The effect of extinction correction on the UIR band ratios is small
(< 15%), and does not affect the following discussion.
Next, we evaluate a contribution from the unresolved emission line Pfδ at 3.30µm
to the 3.3µm band, and that from Pfα at 7.46µm to the 7.7µm band. We assume that
the intensities of Pfδ and Pfα are equal to 9.3% and 30.1% of that of Brβ, respectively,
according to the Case B condition and subtract them from the intensity of the 3.3µm and
7.7µm bands. In the present target positions, the contribution from Pfδ to the 3.3µm band
is less than 10% and that from Pfα to the 7.7µm band is less than 3%, both of which are
similar to the measurement uncertainties and thus do not affect the results. The intensity
for which these corrections are applied is also listed in the lower row for each position in
Tables 6 and 7. The values of the corrected UIR band ratios are listed in Table 8. The
effect of these corrections on the values of the UIR band ratios is less than ∼ 10% for all
the targets. We estimate the uncertainties in I6.2µmand I7.7µmas the difference between
the intensities with and without scaling, which dominates over the fitting errors.
Figures 5a–f show the plots of the corrected UIR band ratios of the I3.3µm/I11.3µm,
I6.2µm/I11.3µm, I7.7µm/I11.3µm, I8.6µm/I11.3µm, I12.6µm/I11.3µm, and I6.2µm/I7.7µmagainst the
IRAS color of I25µm/I12µm. Figures 5g–l plot those ratios against the AKARI color of
IL24/IS11. Group A forms a sequence with a positive slope in the plots of I3.3µm/I11.3µm,