Neon and Sulfur Abundances of Planetary Nebulae in the Magellanic Clouds

Abstract

The chemical abundances of neon and sulfur for 25 planetary nebulae (PNe) in the Magellanic Clouds are presented. These abundances have been derived using mainly infrared data from the Spitzer Space Telescope. The implications for the chemical evolution of these elements are discussed. A comparison with similarly obtained abundances of Galactic PNe and HII regions and Magellanic Clouds HII regions is also given. The average neon abundances are 6.0x10(-5) and 2.7x10(-5) for the PNe in the Large and Small Magellanic Clouds respectively. These are ~1/3 and 1/6 of the average abundances of Galactic planetary nebulae to which we compare. The average sulfur abundances for the LMC and SMC are respectively 2.7x10(-6) and 1.0x10(-6). The Ne/S ratio (23.5) is on average higher than the ratio found in Galactic PNe (16) but the range of values in both data sets is similar for most of the objects. The neon abundances found in PNe and HII regions agree with each other. It is possible that a few (3-4) of the PNe in the sample have experienced some neon enrichment, but for two of these objects the high Ne/S ratio can be explained by their very low sulfur abundances. The neon and sulfur abundances derived in this paper are also compared to previously published abundances using optical data and photo-ionization models. Comment: 13 pages, 4 tables, 5 figures. Accepted for publication in ApJ

Full text

Page 1

arXiv:0709.3292v1 [astro-ph] 20 Sep 2007
DRAFT VERSION FEBRUARY 2, 2008
Preprint typeset using LATEX style emulateapj v. 6/22/04
NEON AND SULFUR ABUNDANCES OF PLANETARY NEBULAE IN THE MAGELLANIC CLOUDS
J. BERNARD-SALAS1, S. R. POTTASCH2, S. GUTENKUNST,1, P. W. MORRIS3, J. R. HOUCK1
Draft version February 2, 2008
ABSTRACT
The chemical abundances of neon and sulfur for 25 planetary nebulae (PNe) in the Magellanic Clouds are
presented. These abundances have been derived using mainly infrared data from the Spitzer Space Telescope.
The implications for the chemical evolution of these elements are discussed. A comparison with similarly
obtained abundances of Galactic PNe and HII regions and Magellanic Clouds HII regions is also given. The
average neon abundances are 6.0×10−5and 2.7×10−5for the PNe in the Large and Small Magellanic Clouds
respectively. These are ∼1/3 and 1/6 of the average abundances of Galactic planetary nebulae to which we
compare. The average sulfur abundances for the LMC and SMC are respectively 2.7×10−6and 1.0×10−6. The
Ne/S ratio (23.5) is on average higher than the ratio found in Galactic PNe (16) but the range of values in both
data sets is similar for most of the objects. The neonabundances foundin PNe and HII regionsagree with each
other. It is possible that a few (3-4) of the PNe in the sample have experienced some neon enrichment, but for
two of these objects the high Ne/S ratio can be explained by their very low sulfur abundances. The neon and
sulfur abundances derived in this paper are also compared to previously published abundances using optical
data and photo-ionizationmodels.
Subject headings: Infrared: general — ISM: abundances — Magellanic Clouds — planetary nebulae: general
1. INTRODUCTION
Stars of low- and intermediate-mass (∼1-8 M⊙) become
planetary nebulae after they evolve off the Asymptotic Giant
Branch(AGB)(Iben&Renzini1983). InthePNphasethehot
central star ionizes the previously ejected material which then
emits copious amounts of emission lines of different ions.
These emission lines are ideal to study the chemical compo-
sition of the gas. The abundances of elements such as carbon,
nitrogen and oxygen can be used to give information on the
nucleosynthesishistoryoftheprogenitorstar. Otherelements,
such as neon, sulfur and argon are not supposed to be altered
in the course of evolution of low- and intermediate-mass stars
and are therefore indicators of the chemical composition at
the epoch of formation (Marigo et al. 2003).
It is for these reasons that PNe have been the subject of
many spectroscopic studies over the years. Due to limita-
tions in the observations, the bulk of this spectroscopic work
has been focused on analysing PNe in the Milky Way (MW).
However, observations of PNe outside the Galaxy are very
important because one can probe different metallicity regions
and, unlike Galactic PNe, the distance is known which al-
lows one to relate the abundanceto the centralstar luminosity.
During the last years several papers (e.g. Magrini et al. 2001,
2003; Corradi et al. 2005, and references therein) have been
devoted to identifying PNe outside the Galaxy. As a conse-
quence, the number of PNe known in the Local Group keeps
increasing. The Large and Small Magellanic Clouds (here-
after LMC and SMC respectively) are ideal candidates to ob-
tain spectroscopic observations of PNe.
Aller et al. (1981) and Aller (1983) published optical spec-
troscopic data for 6 PNe in the LMC and 7 in SMC respec-
tively. In a follow-uppaperAller et al. (1987)presentedultra-
violet data from the IUE satellite of 12 PNe in the Magellanic
1Center for Radiophysics and Space Research, Cornell University, 222
Space Sciences Building, Ithaca, NY 14853-6801, USA.
2Kapteyn Astronomical Institute, 9700 AV, Groningen, The Netherlands.
3NASA Herschel Science Center, IPAC/Caltech, MS 100-22, Pasadena,
CA 91125.
Clouds (MC) and derived abundances for several elements.
Ultraviolet data are essential in order to derive abundances of
elements such as carbon and nitrogen. Optical spectroscopy
and abundances of 71 MC PNe were presented by Monk
et al. (1988). In the early 90’s Meatheringham and Dopita
(1991a,b) obtained optical spectroscopy of over a hundred
PNe in the MC. Morgan & Parker (1998) presented FLAIR
spectroscopy of 97 PNe in the LMC which included fluxes
of the [O III], [S II], [N II], and He II lines. Using optical and
IUE data, Peña et al. (1997)studied a sample of MC PNe with
WR nuclei and found that the distribution of spectral type
was different from those of Galactic WR-PNe. Stanghellini
et al. (2002, 2003) have characterised optically a large num-
ber of PNe in the Magellanic Clouds using HST observations.
And more recently Leisy & Dennefeld (2006) have derived
the abundances of several elements for a large sample of PNe
in the MC using optical data.
Despite their importance, infrared spectroscopic studies of
PNe in the MC are scarce in the literature. This is mainly
because full integrated spectra in the infrared can only be
achieved from space. IRAS detected several PNe in the MC
(Zijlstra et al. 1994), mainly at 12 and 25 µm, but some of
the identifications are dubious. The SWS spectrograph (de
Graauw et al. 1996) on board ISO did an excellent job study-
ing nearby PNe, but it did not have enough sensitivity to al-
low the study of PNe outside the Galaxy. The Spitzer Space
Telescope (Werner et al. 2004) with its increased sensitivity
enables us to observe PNe outside the Milky Way (Bernard-
Salas et al. 2006, 2004). The importance of using infrared
lines when deriving abundances has been highlighted by Ru-
bin et al. (1988) and we summarize these reasons here. In-
frared lines are little affected by extinction as opposed to op-
tical or UV lines. Uncertainties in the electron temperature
or fluctuations in the temperature within the nebula are not
important when using infrared lines because they originate
from levels very close to the ground level. Additionally many
ions emit in the infrared, and therefore the use of ionization
correction factors (ICFs) can be greatly reduced by including
infrared observations. This is especially true for neon, sul-

Page 2

2 Bernard-Salas et al.
fur and argon. Finally, and while not discussed in this paper,
emission of dust can be studied in this part of the electromag-
netic spectrum.
In this paper we present Spitzer high-resolution spectro-
scopic observations of 25 PNe in the MC (18 in the LMC and
7 in the SMC). This paper focuses on the emission of fine-
structure lines and their use in the abundance determination.
The measured lines are used to derive abundances for sulfur
and neon. These abundances are mainly compared to Galac-
tic PNe abundancesfromPottasch &Bernard-Salas(2006),as
well as Galactic, MC, M33 and M83 HII regions abundances
fromMartín-Hernándezetal.(2002),Vermeij&vanderHulst
(2002), and Rubin et al. (2007). All of these abundances have
been derived using infrared data and in a similar way to that
presented in this paper. Dust features present in the spectra
such as Polycyclic Aromatic Hydrocarbons (PAHs) and sili-
cates will be discussed in a future paper.
2. OBSERVATIONS AND DATA REDUCTION
The observations were made using the Infrared Spectro-
graph (IRS) (Houck et al. 2004) on board the Spitzer Space
Telescope and resulted in high- and low-resolution spectra
of 25 PNe. These observations were part of the GTO pro-
gram (ID 103) and were taken between March and November
2005. The object name and AORkey numbers for each obser-
vation are given in the first and second columns of Table 1.
The nomenclaturegiven by Sanduleak et al. (1978)is adopted
through the paper. In addition, the analysis includes data on
SMP LMC 31, and abundances on SMP LMC 83 derived by
Bernard-Salas et al. (2004). These data were taken during In
Orbit Check-Out (IOC). There are two high-resolution mod-
ules in the IRS, named Short-High and Long-High (SH and
LH respectively). Together, they cover the wavelength region
between 10 and 37 µm at a resolution of 600. The reader
should refer to the paper by Houck et al. (2004) for more in-
formation on the IRS instrument. We used coordinates given
by Stanghelliniet al. (2002,2003)andLeisy et al. (1997),and
performedbluePeak-Upacquisitionona nearbystar to obtain
accurate pointing (0.4′′). Figure 1 shows the position of the
LMC PNe on an IRAC image from the SAGE team (Meixner
et al. 2006).
The data were processed through a copy of the S13.2 ver-
sion of the Spitzer Science Center’s pipeline which is main-
tained at Cornell and using a script version of Smart (Higdon
et al. 2004). The reduction started from the droop images.
These are equivalentto the most commonlyused bcd data and
only lack the flatfield and stray-cross-light removal (which is
only important for bright sources). Rogue pixels which are
especially notorious in the LH module were removed using
the irsclean4tool. The rogue pixels were first flagged us-
ing a campaign mask and then removed. If different cycles
(repetitions) were present for a given observation these were
combinedusing the mean to improvethe S/N. The 2D-images
were extracted using full aperture extraction. The calibration
was performed by dividing the resultant spectrum by that of
the calibration star ξdra (extracted in the same way as the tar-
get) and multiplying by its template (Cohen et al. 2003, and
Sloan et al. in prep). Finally, glitches which were not present
in both nod positions or in the overlapping region between
orders were removed manually.
There is an expected mismatch between the SH and LH
spectra. This mismatch is due to differences in the back-
4Thistool isavailable from the SSC website: http://ssc.spitzer.caltech.edu
TABLE 1
ADOPTED PARAMETERS FOR ABUNDANCE DETERMINATION.
ObjectAORkey Log(FHβ)a,b
CHβa
Te(K)a
Ne(cm−3)a
3000c
5500
6200
3800
2000
6800
1600
3000c
9800
1100
4000
20000
26000
4400
13600
4300
31400
1900
SMP LMC 02
SMP LMC 08
SMP LMC 11
SMP LMC 13
SMP LMC 28
SMP LMC 31
SMP LMC 35
SMP LMC 36
SMP LMC 38
SMP LMC 40
SMP LMC 53
SMP LMC 58
SMP LMC 61
SMP LMC 62
SMP LMC 76
SMP LMC 78
SMP LMC 85
SMP LMC 87
4946944
15902464
4947712
4947968
4948224
7459584
4948736
4949248
12633600
4949504
15902720
4950784
12633856
4951040
4951296
15902208
4952320
4952576
-13.18
-13.74
-13.94
-12.88
-13.57
-12.92
-12.81
-12.72
-12.62
-13.25
-12.62
-12.54
-12.48
-12.30
-12.54
-12.60
-12.42
-12.91
0.06
0.23
0.31
0.09
0.32
0.54
0.04
0.41
0.21
0.20
0.13
0.11
0.22
0.07
0.34
0.21
0.26
0.25
11600
11000
25000
13600
10000
12800
13300
15000
13000
13900
13700
12100
10800
15800
11600
14200
10500
19200
SMP SMC 01
SMP SMC 03
SMP SMC 06
SMP SMC 11
SMP SMC 22
SMP SMC 24
SMP SMC 28
4953088
4953600
4954112
15902976
4954624
15901952
4955136
-12.85
-13.13
-12.80
-12.87
-12.94
-12.66
-13.18
0.287
0.000
0.385
0.82
0.165
0.047
0.200
11000
13800
15300
17600
18800
12700
20300
9600
5600
14700
1100
2500
1300
8800
aValues taken from the following references (see § 3.2): Leisy & Dennefeld
(2006); Meatheringham et al. (1998); Meatheringham and Dopita (1991a,b);
Shaw et al. (2006); Stanghellini et al. (2002, 2005); Villaver et al. (2003, 2004);
Wood et al. (1987).
bFlux in units of erg s−1cm−2.
cAssumed electron density.
87
13
02
40
35
85
83
36
28
58
08
38
31
76
78
11
62
53
N
E
FIG. 1.— IRAC four-band mosaic of the LMC (SAGE) with the positions
of the PNe overlaid on it (circles). The dynamical center of the LMC as given
byKim et al. (1998) is indicated by the square close to the center of the figure.

Page 3

Ne and S Abundances of PNe in the Magellanic Clouds3
groundcontributionthat falls into theslits becausetheslit size
of the LH module is about 4.6 times larger in area than the
SH slit. We do not scale the spectra because we are interested
in the line fluxes and the nebulae are contained in both slits.
These PNe are not extended at such a distances. Diameters of
PNe in the LMC given by Shaw et al. (2001) and Stanghellini
et al. (1999) are usually less than 1′′, and only in a very few
cases does the diameter reach 3′′(still smaller than the SH slit
width of 4.5′′). Representative examples of the full extracted
high-resolutionspectra from10-37µm are shown in Figure 2.
3. ANALYSIS
3.1. Line emission
Figure 3 shows an inset of the most relevant lines for the
abundance determination present in the spectra of all the ob-
jects. The sample ranges from PNe showing high excitation
to low excitation lines (e.g. SMP LMC 85, SMP LMC 62).
Most of the spectra show features (PAHs) characteristics of
carbon-rich material (e.g. SMP LMC 36, SMP SMC 11 in
Fig. 2), except for SMP LMC 53 and SMP LMC 62 (Fig. 2)
which show amorphous silicates in emission at 9 and 18 µm
which is usually an indicator of an oxygen-rich environment.
The line fluxes are listed in Table 2. In addition to the lines
listed in this table, other lines such as [Mg V] at 13.52 µm
have been measured for the PNe with higher S/N. The [S III]
33.48µm line is always detected whenthe 18.7µm is present,
but because the spectrum is noisier at the long wavelength
end of the LH module (see Fig.2) we preferred to use the
18.71 µm line flux for the abundance determination of this
ion which also has a larger transition probability and critical
density. Similarly we favored the use of the [Ne III] 15.55 µm
line instead of the 36.0 µm line when both were measured.
Table 2 also includes (mostly) upper-limits on the argon line
at 13.1 µm. Although not shown in this paper, the [Ar III]
line at 8.99 µm is detected in 9 objects in the low-resolution
spectra, and the [Ar II] 6.99 µm line in 2 objects. The line
fluxes were measured using the Gaussian line-fitting routine
in Smart. These were derived for each nod independently.
The uncertainty in the fluxes was assumed to be the largest of
either the difference between the flux in the nod positions, or
the uncertainty in the fit. These errors are given as footnotes
in the Table. The upper-limits were calculated from a Gaus-
sian fit with height 3 times the rms and a FWHM as given by
the resolution of the instrument.
3.2. Assumed parameters
Table 1 lists the Hβ flux, extinction, electron density (Ne)
and electron temperature (Te) assumed to calculate the abun-
dances. These values were compiled from a large list of ref-
erences which are given in the footnote of the table. When
several values were given by different authors an average was
used. Only in the cases where several values differed by a
large amount (i.e. SMP LMC 87, SMP SMC 11), we as-
sumed those we estimate are more accurate. The infrared
lines are little affected by uncertainties in the Teand extinc-
tion corrections but the abundances relative to hydrogen are
derived using optical measurements of Hβ which are affected
by these factors. The extinction law used is that of Fluks et
al. (1994). In two cases no Newas reported and we assumed
a value of 3000 cm−3which seems a reasonable value in view
of the other measurements.
3.3. Abundance determination
Using the above parameters, the ionic abundances were
computed from the infrared line intensities using Eq. 1 of
Bernard-Salas et al. (2001). The results are shown in Table 3.
The total abundancesare also given in this table, where some-
times a correction due to missing ionization stages is neces-
sary.
For the sulfur abundance the addition of S+, S+2and S+3
is sufficient for those nebulae for which no O IV is observed.
The S+abundance was determined using the optical lines of
S+measured by Meatheringham and Dopita (1991a,b) and
Stanghellini et al. (2002, 2003). The contribution from this
ion is usually small and in the order of what it is found in
Galactic PNe. For the PNe which show the O IV line we must
take into account the possibility that S+4is present. An esti-
mate may be obtained by looking at the two photo-ionization
models of Me2-1 (Surendiranath et al. 2004) and NGC 6886
(Pottasch & Surendiranath 2005). Both of these PNe are ex-
cited by high temperature stars (with Tef f between 140000
and 180000 K) and both show O IV lines. In addition, Me2-
1 has lower average abundances compared to Galactic PNe
which are closer to the nebulae studied here. Both models
give similar results and show that S+4contributes between 7
and 23% of the total sulfur abundance based on the S+4/S+3
ratio. We have used these numbers to correct the sulfur abun-
dances in Table 3.
In the case of neon no Ne IV line has been observed. For
those PNe which do not show an O IV line it is unlikely that
there is any Ne+3because it requires a higher energy radi-
ation field than does O IV. For those PNe which do show
the O IV line (7 in the LMC and 2 in the SMC) a correc-
tion must be made. This can be done in two ways. First,
using the same models as for sulfur we obtain a contribution
of Ne+3that varies from 2 to 33% of the total neon abun-
dance depending on the strength of the Ne V line. A second
way of determining the correction could be done by looking
at the neon abundances in the sample studied by Pottasch &
Bernard-Salas (2006). This study made use of the same in-
frared lines as used here and had both Ne III and Ne V lines,
but the ultravioletlines ofNe IV at 2422Å were also observed
sothat Ne+3couldalso bedetermined. Forthese PNe theNe+3
abundance was on average 21% of the total neon abundance
and 35% of the sum of Ne+2and Ne+4. To be consistent with
the sulfur abundance we have used the first method, but the
results using the second method would result in the same val-
ues within the expected errors. The total neon abundancesare
given in Table 3.
TheuncertaintyinthevaluesofboththeNe+3andS+4abun-
dances is probably not more than a factor of 2. This leads to
a maximum error for sulfur and neon of about 30% from this
source. In addition uncertainties of measurement of the other
ions in the infrared are about 10% with only a few excep-
tions (see footnote in Table 2). The error in the optical lines
we used is of the same order. In addition an error in the ex-
tinction affects the Hβ flux and uncertainties in the electron
temperature dominate the uncertainty in our abundance de-
termination. By comparing the optical measurements for the
same objects by different observers in the literature, which
usually agree, we estimate that the total error remains within
50% except for SMP LMC 08 and SMP SMC 11. The abun-
dances of SMP LMC 08 are uncertain probably because of
the assumed Hβ flux. (see §5.3). In SMP SMC 11 the uncer-
tainty is dominatedby the inconsistent Hβ flux and extinction
quoted in the literature and for the purpose of this paper the

Page 4

4 Bernard-Salas et al.
FIG. 2.— SH and LH spectra of a handful of PNe. The jump around 19.5µm is due to the larger background contribution that falls in the LH slit compared to
the smaller SH slit (see §2).
most recent values given in the literature have been adopted.
4. COMPARISON SOURCES
This section describes the sources to which the MC PN
abundances are compared in §5.2. These comparison sources
include Galactic PNe, Galactic, MC, M33 and M83 HII re-
gions, and the solar abundance.
4.1. Solar values
The solar carbon, oxygen, sulfur, argon and neon abun-
dances have been subject to significant changes, especially
during the last seven years. These changes reflect in some
way the difficulty in derivingsolar abundances. The neon and
argon abundances are especially troublesome because there
are no lines of these elements in the solar photosphere and
their abundances must be derived from coronal lines. As-
plund et al. (2005) quoted a neon value of 6.9×10−5using
the oxygen solar abundance and assuming a ratio of the Ne/O

Page 5

Ne and S Abundances of PNe in the Magellanic Clouds5
TABLE 2
LINE FLUXESaOF THE OBSERVED LMC AND SMC PNE.
Object [S IV]
(10.51µm)
[Ne II]
(12.89µm)
[Ar V]
(13.10µm)
[Ne V]
(14.31µm)
[Ne III]
(15.55µm)
[S III]
(18.73µm)
[O IV]
(25.89µm)
SMP LMC 02
SMP LMC 08
SMP LMC 11
SMP LMC 13
SMP LMC 28
SMP LMC 31
SMP LMC 35
SMP LMC 36
SMP LMC 38
SMP LMC 40
SMP LMC 53
SMP LMC 58
SMP LMC 61
SMP LMC 62
SMP LMC 76
SMP LMC 78
SMP LMC 85
SMP LMC 87
<0.72
2.95
<1.08
12.08
0.44
<0.75
10.29
12.75b
12.96b
2.96
20.55
2.92
7.59
27.80
4.05
25.63
1.85
11.39
1.84
7.48
1.89
0.22b
1.96
22.45
0.48
1.34
2.57
0.87
0.90
2.06
6.11
1.41
2.28
1.92
13.15
3.63
<0.32
<0.38
<0.77
<0.19
<0.30
<0.31
<0.19
0.52
<0.24
<0.24
<0.23
<0.24
<0.46
1.24
<0.29
0.94
<0.27
0.75
<0.26
<0.38
<0.69
0.94
0.32
<0.17
0.20b
25.37
<0.31
2.59
<0.26
<0.38
<0.35
28.80
<0.23
26.83
<0.40
12.53
<0.66
34.19
0.67
12.74
3.55
0.38b
16.53
36.35
48.67
6.29
26.17
20.06
30.09
32.06
20.66
46.13
16.53
11.28
0.82
1.98
<0.68
2.10
0.51
0.99
1.87
1.83
3.80
1.33
4.57
1.10
6.36
6.47
2.33
5.83
2.33
4.58
<0.75
<0.63
<1.38
55.54
<1.15
<1.09
27.22
22.31b
<1.53
24.04
<1.49
<2.16
<0.59
38.91
<0.42
47.39
<0.86
39.53
SMP SMC 01
SMP SMC 03
SMP SMC 06
SMP SMC 11
SMP SMC 22
SMP SMC 24
SMP SMC 28
<0.53
2.11
4.41
3.76
1.67
1.52
1.53b
8.10
<0.35
0.95
13.44
1.30
1.82
0.43
<0.29
<0.32
<0.23
<0.30
<0.30
<0.33
<0.24
<0.34
<0.24
<0.36
<0.37
3.60
<0.25
2.53
2.89
3.09
14.87
8.19
2.53
7.46
1.70
0.63
<0.48
0.92
12.67
1.25b
2.28
0.47
<0.66
<1.25
<0.65
<1.18
7.44
<0.84
1.29b
aFluxes in units of ×10−14erg cm−2s−1. Unless otherwise indicated, the uncertainties are less than 10% for all
the lines except the [S IV] line flux which has an uncertainty between 10 and 20%.
bThese lines have uncertainties in the flux between 20 and 30%, except for the [O IV] line in SMP SMC 28
where the error is 42%.
of 0.15. Previously, Feldman & Widing (2003) using coro-
nal line measurements found a neon abundance of 1.2×10−4.
This value is much higher than the value by Asplund et al.
(2005) but agrees better with the earlier values reported by
Grevesse & Sauval (1998). This discrepancy is important
in the debate over the consistency of the helioseismological
measurements and the solar model (Antia & Basu 2005; Bah-
call et al. 2005). The neon abundance derived by Pottasch
& Bernard-Salas (2006) in a sample of Galactic PNe is more
consistent with the higher neon value of Feldman & Widing
(2003). Very recently Landi et al. (2007) derived a value of
1.29×10−4, again higher than the value given by Asplund et
al. (2005) and in very good agreement to the previous values
reported Feldman & Widing (2003) and Grevesse & Sauval
(1998). The quoted value of the solar sulfur abundance has
been decreasing in the last years. The sulfur abundance de-
rived by Grevesse & Noels (1993)is 1.4×10−5while Asplund
et al. (2005)find0.94×10−5. Giventhese discrepancies,in the
rest of the paper instead of assuming a certain value we will
refer and compare to the above range of solar values.
4.2. PNe and HII regions
For comparisonpurposes we have selected a sample of PNe
andHII regionsforwhich abundanceswerealso derivedfrom
infrared data and in a similar way to the PNe presented in this
paper. The Galactic PNe abundances in Pottasch & Bernard-
Salas (2006) using ISO data have been complemented with
the Spitzer derived abundances of IC2448 (Guiles et al.
2007),M1-42(Pottaschet al.2007),andNGC2392(Pottasch
et al., in prep). Galactic and MC HII regions were taken from
Martín-Hernández et al. (2002) and Vermeij & van der Hulst
(2002) respectively. They include ISO derived abundances
from 26 HII regions in the Milky Way, 13 in the LMC, and 3
in the SMC. The Spitzer abundances in Lebouteiller et al.(in
prep)of thegiant HIIregionsNGC3603(in theMW), 30 Do-
radus (LMC), and NGC346 (SMC) are also included. While
Lebouteilleret al. (in prep) derive abundances at several posi-
tions in each region, the calculated abundances are very sim-
ilar for a given region and here we adopt their average value.
Rubin et al. (2007) derived recently the Ne/S abundance ratio
of HII regions in M83 using Spitzer data but the absolute val-
ues are not given. The same authors are working on a study
of HII regions in M33 and we use their neon and sulfur abun-
dance ranges5in Figure 5.
5. DISCUSSION
5.1. Neon, Sulfur and the Ne/S ratio
5These abundances were presented in the Xiang Shan workshop in 2007
(http://ast.pku.edu.cn/ xs2007/).