arXiv:astro-ph/0003436v1 29 Mar 2000
Large Magellanic Cloud Planetary Nebula Morphology: Probing
Stellar Populations and Evolution.1
Letizia Stanghellini2,3, Richard A. Shaw
Space Telescope Science Institute, 3700 San Martin Drive, Baltimore, Maryland 21218,
Department of Astronomy, University of Washington, Seattle, Washington 98195
J. Chris Blades
Space Telescope Science Institute
2Affiliated to the Astrophysics Division, Space Science Department of ESA
3on leave, Osservatorio Astronomico di Bologna
– 2 –
Planetary Nebulae (PNe) in the Large Magellanic Cloud (LMC) offer the
unique opportunity to study both the Population and evolution of low- and
intermediate-mass stars, by means of the morphological type of the nebula. Us-
ing observations from our LMC PN morphological survey, and including images
available in the HST Data Archive, and published chemical abundances, we find
that asymmetry in PNe is strongly correlated with a younger stellar Population,
as indicated by the abundance of elements that are unaltered by stellar evolu-
tion (Ne, Ar, S). While similar results have been obtained for Galactic PNe,
this is the first demonstration of the relationship for extra-galactic PNe. We
also examine the relation between morphology and abundance of the products
of stellar evolution. We found that asymmetric PNe have higher nitrogen and
lower carbon abundances than symmetric PNe. Our two main results are broadly
consistent with the predictions of stellar evolution if the progenitors of asymmet-
ric PNe have on average larger masses than the progenitors of symmetric PNe.
The results bear on the question of formation mechanisms for asymmetric PNe,
specifically, that the genesis of PNe structure should relate strongly to the Pop-
ulation type, and by inference the mass, of the progenitor star, and less strongly
on whether the central star is a member of a close binary system.
Subject headings: Stars: AGB and post-AGB — stars: evolution — planetary
nebulae: general — Magellanic Clouds
1Based on observations made with the NASA/ESA Hubble Space Telescope, and from
the HST Data Archive, obtained at the Space Telescope Science Institute, which is operated
– 3 –
Planetary Nebulae (PNe) are produced as stars age. They are fundamental to
understanding the evolution of stars whose initial mass lies below the Supernova limit.
Classically, Galactic PNe have been divided in Population classes according to their spatial
distribution, kinematics, and chemical content (Greig 1972; Peimbert 1978; Maciel 1999).
It appears that there are clearly different PN Populations in the Galaxy, from old disk
Population PNe (Peimbert’s Type I) to extreme Pop II PNe, located in the Galactic halo or
in the bulge (Peimbert’s Type IV and V). The PN morphology varies systematically across
these classes (Peimbert 1997), and most of the Type I PNe are asymmetric in shape.
More recent studies of large Galactic PNe samples have shown that, to first
approximation, the morphology of PNe is linked to the spatial distribution within the
Galaxy, and to the mass of the progenitor star (Stanghellini, Corradi, & Schwarz 1993;
Manchado et al. 2000). The fact that most bipolar and quadrupolar PNe lie on average
closer to the Galactic plane than round and elliptical PNe, and that highly asymmetric
PNe appear to host the most massive central stars (CSs), suggests that asymmetric PNe
are the likely progeny of a younger stellar Population than symmetric PNe.4
An important and long-standing astrophysical issue is the degree to which PNe enrich
by the Association of universities for research in Astronomy, Inc., under NASA contract NAS
4In this paper we group elliptical and round PNe in the symmetric class, and bipolar
and bipolar core PNe in the asymmetric class. Quadrupolar PNe in the LMC may have
questionable morphology, thus we segregate them from these two major classes. Note that
this definition of symmetric and asymmetric PNe is identical to the elliptical and bipolar
classifications in both Stanghellini et al. (1993) and Corradi & Schwarz (1995).
– 4 –
heavy elements in the ISM. In the Galaxy, PNe supply almost an order of magnitude more
mass per year than supernovae (Osterbrock 1989). Depending on the progenitor’s mass,
PNe are expected to enrich the ISM with carbon and nitrogen (Iben & Renzini 1983).
PNe stellar progenitors undergo several dredge ups (e.g., Iben & Renzini 1983), some of
which enrich the surfaces with C and N. Subsequent winds will carry the enriched gas
into the ISM. PNe are known to account for half of the carbon and most of the nitrogen
enrichment in the solar environment (Henry, Kwitter, & Buell 1998). Therefore a study of
C and N abundances relative to the elements which are not altered during the evolution of
PNe progenitors, such as Ne, Ar, and S, provides a means for gauging the efficacy of C-N
enrichment rates by PNe.
Abundance studies as a function of Population class and morphological type have
been carried out for Galactic PNe. However, since asymmetric PNe lie close to the plane
where foreground extinction is severe, and since such nebulae are also formed from the
most massive progenitors, Galactic PNe suffer a serious selection bias for understanding
enrichment rates by PNe of different types. Accordingly, we have begun a comprehensive
survey of Magellanic Cloud PNe in order to investigate these and other parameters in a
large and well-understood sample. We are using HST and STIS in direct imaging and
slitless spectroscopy mode for this survey, and our first observational results are described
in Shaw et al. (2000). For our purposes the advantage of LMC PNe is their relatively
low foreground extinction (i.e., their low selection bias) as well as their well determined
distances. That is, abundance studies tend to be complete to a limiting nebular luminosity,
which is far less of a bias than a brightness limited survey in the presence of highly variable
extinction with Galactic latitude.
In the study presented here, we have selected a subsample of LMC PNe based solely
on the availability of morphological information (from HST images) and relative chemical
– 5 –
abundances (from ground-based spectroscopy in the literature). The HST data come from
Shaw et al. (2000), from the HST Data Archive, and from the literature. In §2 we describe
the database and the data analysis, and discuss the results; and in §3 we discuss the
uncertainties and possible systematic effects.
2. PN Morphology, Stellar Populations, and Stellar Evolution
The morphological data used in this paper consist of all LMC PNe observed in imaging
mode with HST. The sample includes the pre-refurbishment data set of PC1 and FOC
images (see Stanghellini et al. 1999), the unpublished WFPC2 images in the Hubble Data
Archive (HST Program ID: 6407, PI: Dopita, Cycle 6), and the set of 27 LMC PNe observed
to date within our HST/STIS snapshot survey (Shaw et al. 2000).
The abundance data set was compiled from papers by Leisy & Dennefeld (1996), Monk,
Barlow, & Clegg (1988), Dopita et al. (1997), and Dopita & Meatheringham (1991a,b). We
have used abundances of carbon, nitrogen, oxygen, argon, neon, and sulfur where available.
However, some of the quoted abundances are model dependent (Dopita & Meatheringham
1991a,b); abundances therein were used only when the model independent values were not
available. A discussion of the effects of the data set inhomogeneity is presented in the next
section, along with a critical analysis of the published abundances, including uncertainties
and systematic effects that could affect our conclusions.
We classified PN morphologies homogeneously according to their shape in narrow-band
[O III] λ5007 images: Round, elliptical, bipolar, bipolar core (Stanghellini et al. 1999),
quadrupolar, and point-symmetric PNe are included in the sample. While classifying the
images, we realized that in a few cases the morphologies are uncertain. These objects are
not included in the numerical analysis and discussion, nor in the figures.
– 6 –
Figures 1 and 2 show the distribution of the morphological types with respect to neon,
argon, and sulphur abundances, which have presumably remain unchanged since the birth
of the progenitor star. As such, these elements should be good indicators of the Population
type of the progenitor stars. The crossed large circles in these figures marks the location of
the average abundance for LMC H II regions (Leisy & Dennefeld 1996). These and other
young population LMC abundances are given in Table 1, discussed below.
The segregation of morphological type with respect to S and to Ne (Fig. 1) is striking;
Ar does not appear to be as good a discriminant (Fig. 2). Specifically, the overwhelming
majority of PNe with log Ne + 12 > 7.6 are asymmetric, and none is round. Furthermore,
all PNe with log Ne + 12 < 7.6 are symmetric, with the exception of one quadrupolar PN.
Evidently, PN morphology is a good indicator of progenitor Population in the LMC. By
comparing the Ne abundance in PNe with the average for H II regions (which gives a good
indication of the abundance of the young stellar Population), it is evident that most of the
asymmetric PNe are enriched with respect to the young Population, while the opposite
holds for symmetric PNe. Sulfur is related to nebular shape in a similar way. In fact, only
one of the 13 symmetric PNe has S larger than the LMC H II region average, and three
asymmetric PNe are under-abundant with respect to the H II regions.
To quantify the importance of the neon, sulphur, and argon overabundance in
asymmetric PNe, in Table 1 we list the average abundances for our PN sample, together
with the abundances of other significant LMC objects, namely, H II regions and Supernova
remnants: column (1) lists the atom; in columns (2) and (3) we give the average elemental
abundances for the symmetric and asymmetric PNe of our samples, in the usual form
12+log (N/H), with the sample size in parenthesis; column (4) gives the LMC H II region
average from Leisy & Dennefeld (1996), as used in our Figures; columns (5) and (6) give the
abundance ranges found by Russel & Dopita (1990) respectively in LMC H II regions and
– 7 –
LMC Supernova remnants. The ranges are only indicative of the abundance distribution on
the LMC, since they include a few observed objects for each diagnostics.
The ratio of the average neon abundance for asymmetric to symmetric PNe in the
LMC, 1.7, is similar to that found in the Galaxy, 1.6 (Corradi & Schwarz 1995). Note,
however, that the Galactic average is based on a sample that is markedly biased toward
the symmetric PNe (see §1). A similar overabundance in asymmetric PNe is found for
sulfur and argon, although in these cases this includes a few PNe that do not follow the
Figure 3 shows the oxygen distribution among different Population PNe. Oxygen
yield should be invariant with respect to progenitor mass or other parameters, at least in
the LMC (van den Hoek & Groenewegen 1997). Data shown in Figure 3 are consistent
with the expectation that the oxygen abundance are within the same range in symmetric
and asymmetric PNe. These results from our limited sample suggest that progenitors
of asymmetric PNe were formed in an enriched environment that is typical of a very
young stellar Population, while precursors of symmetric PNe formed in a medium that
was under-abundant with respect to the LMC H II region average. PNe split into two
distinct Population groups according to alpha-element (neon, argon, sulphur) abundance
In order to compare our findings with the analysis of Galactic PNe, we looked for
spatial segregation of symmetric and asymmetric PNe. Although the sample size is modest,
we did not find such a strong spatial segregation of LMC PNe across morphological type,
or of their location with respect to the LMC bar or any other particular region, with the
exception of a very slight predominance of asymmetric PNe along, rather than perpendicular
to, the bar. This result is consistent with the short dynamical mixing time of the LMC, and
we should not expect a marked segregation of stellar Populations as is found, for example,
– 8 –
in spiral galaxies.
A number of theoretical models (Renzini & Voli 1981; Iben & Renzini 1983; van den
Hoek & Groenewegen 1997) predict that lower-mass PN precursors typically go through
the carbon stars phase, while the hot-bottom burning (HBB) takes place in higher mass
precursors and prevents carbon star formation. This aspect of stellar evolution does not
depend dramatically on the initial composition, and in particular for LMC PNe one expects
the same carbon depletion and nitrogen enrichment above ≈4 M⊙of main sequence stars
(van den Hoek & Groenewegen 1997). Observations of Galactic PNe support the theory,
in that there seem to be two AGB star types characterized by distinct carbon-enriched or
depleted dust (Trams et al. 1999).
Then how do carbon and nitrogen abundances correlate with morphology in the LMC?
We plot nitrogen against carbon abundance in Figure 4. It is clear that symmetric PNe are
well confined in such a plot, and all symmetric PNe are carbon-enriched with respect to the
LMC H II region average. The situation for asymmetric PNe is rather different: they are
al nitrogen-enriched, yet three of them are also carbon-enriched with respect to the H II
region average. The figure could be interpreted as follows: low mass stars (< 4 M⊙on main
sequence) go through the carbon star phase, and do not produce asymmetric PNe. The
high mass stars do not go through the carbon star phase, they suffer HBB on the AGB,
and they produce asymmetric PNe. Some of the low-mass stars producing carbon stars also
end up as asymmetric PNe, perhaps through the common envelope phase.
A discussion of the formation mechanisms for asymmetric PNe is in order, in light
of our findings. If asymmetry in PNe were due uniquely to common envelope evolution,
or binary evolution in general, we would not expect to find any of the separations among
morphological classes that we show in Figures 1 though 3. That is, we do not expect the
incidence of close binaries to vary as a function of the mass of the PN progenitor. Though
– 9 –
this is a qualitative statement, it is substantiated by other observations (i. e., the existence
of bipolar PNe that do not show an equatorial ring, such as the Galactic PN Hubble 5), and
studies of the initial mass function. On the other hand, there may be a small fraction of
asymmetric PNe that are developed as a consequence of close binary evolution, and this is
also supported by observations. It may be that we are very far from understanding how the
morphology of PNe is created, as there are a large number of variables in this game. What
we can conclude with some certainty is that asymmetry in PNe is related to the Population
type, and by inference the mass of the progenitor star. Any model for the formation of PNe
that predicts the morphology must be consistent with this relationship.
3. Abundance Uncertainties and Systematic Effects
If the results of the previous sections are borne out by additional observations and
analysis, the effect on subsequent interpretations of PN formation and evolution could
be substantial. In this section we take a critical look at the derivation of the chemical
abundances that are the underpinning of the conclusions presented here.
The first step is to give some information on the abundance uncertainties as derived
from the original papers. None of the references that we have used for abundance cite
individual errors. Dopita et al. (1997) do not discuss errors, but from the errorbars in their
Fig. 7 it is possible to infer that the N, O, and C abundances are good to 0.1 dex, while
the errors on the alpha-elements are about 0.08 dex. We should infer similar uncertainties
for Dopita & Meatheringham (1991a,b), since the three papers use the same abundance
determination method. Monk, Barlow, & Clegg (1988) evaluate that their oxygen and
neon abundances are good to 0.1 or 0.15 dex, while neon and argon abundances are more
uncertain, up to 0.2 dex. Finally, Leisy & Dennefeld (1996) determine that most elemental
abundances are good to 0.1 dex, with the exception of nitrogen, whose uncertainty is up to
– 10 –
0.2 dex or more.
The potential for systematic errors from using different sources of abundance in the
literature is worrisome enough to warrant some scrutiny. We compared the abundances from
different bibliographic sources, and found that the values from Dopita & Meatheringham
(1991a,b) and Monk, Barlow, & Clegg (1988) agree, within the quoted errors, with the
abundances by Leisy & Dennefeld (1996). Plotting the results of pairs of references shows
only a scatter in the final results for most elements with no systematic differences. The only
element that is worrisome is nitrogen. Leisy & Dennefeld (1996) give a nitrogen abundance
that is systematically about 0.3 dex higher than other authors. However, omitting the
work of any one paper does not change the qualitative results of Fig. 4. In particular, any
discrepancies do not correlate with the ionization states of the nebulae5.
Most abundances are derived from observations using some form of the ionization
correction method to convert from measured ionic abundances to total abundances.
An ionization correction “factor” (ICF) for unseen ionization states is required in this
conversion. Alexander & Balick (1997) found that the ICFs are very large in the case of
low ionization nebulae, giving, for example, artificially high neon or sulphur abundances
that in principle could alter the effects seen in Figures 1 though 3. We explored the
original line intensity lines used for abundance calculation by Leisy & Dennefeld (1996),
and Dopita & Meatheringham (1991a,b), and we related the ionization level from the [N II]
to Hα line ratio, to the morphological type. The question becomes whether those classes
of PNe with segregated abundances have systematically low or high ionization. We found
no correlation between ionization level and morphology in our combined sample. Most of
the nebulae show high ionization, and the few low-ionization objects are equally distributed
5Low ionization PNe are those in which O+is the dominant ionization state, while in
high ionization PNe the O++state dominates.
– 11 –
among symmetric and asymmetric PNe. The only exception to this rule concerns extremely
bipolar PNe (indicated with a filled square in the Figures). Among four of such PNe, three
are low ionization. Eliminating these extreme bipolars from the plots would not change
the conclusions of this paper. It is worth mentioning that Alexander & Balick (1997)
compared ICF abundances to those of model computation, and they find that Ne/H is the
most reliable measure of the abundance of primordial elements since the O++ and Ne++
volumes are very similar and the corrections for unseen ionization states are relatively small.
In general, LMC PNe are point-sources when observed from the ground, and the line
intensities quoted in the literature typically refer to the global volume of the nebulae, thus
the ICF problem here is minimal (Alexander & Balick 1997). In fact, often Magellanic
Cloud PNe are discovered via [O III] imaging (or [O III] must have been present in an
objective prism spectrum), so fewer PNe in our sample have low ionization.
One last concern comes from the possibility that the line intensities used for abundances
calculation suffer from alteration due to the presence of a shock front. This is particularly
worrisome for the sulphur abundances. In fact, artificially high sulphur abundances may
derive from excessively high intensities of the low-ionization states of sulphur. We have
checked the line intensities for all PNe with 12+log(S/H)>7 to obtain diagnostic ratios for
shocks, as explained by Veilleux & Osterbrock (1987). In our sample, only two PNe have
log [O I](λ6300)/Hα and log [S II](λ6716+6731)/Hα close to the limit for shock fronts (see
the dashed-dotted lines in Veilleux & Osterbrock (1987)’s Figs. 4 through 6). We conclude
that only a negligible fraction of the asymmetric PNe in Figure 1 may have high sulphur
abundance due to the presence of a shock front. The overall scientific results that we
describe in the previous chapter holds even if this was the case.
In conclusion, we feel that the results shown in this paper are quite sound, in spite
of the inhomogeneous abundances. This work will be extended to the SMC in which the
– 12 –
metal abundances are considerably lower than in the LMC. In addition, we are extending
the ground-based spectroscopy to a far larger sample of LMC and SMC PNe. Both the
numbers of targets and the accuracy of the data will be greatly improved as a consequence.
Thanks to Max Mutchler for his work on our LMC images, to Karen Kwitter for
carefully reading the manuscript, and to an anonymous referee for important comments.
Support for this project was provided by NASA through grant number GO-08271.01-97A
from Space Telescope Science Institute, which is operated by the Association of Universities
for Research in Astronomy, Incorporated, under NASA contract NAS5–26555.
– 13 –
Alexander, J., & Balick, B. 1997, AJ 114, 713
Corradi, R. L. M., & Schwarz, H. E. 1995, A&A 293, 871
Dopita, M. A., & Meatheringham, S. J. 1991a, ApJ 367, 115
Dopita, M. A., & Meatheringham, S. J. 1991b, ApJ 377, 480
Dopita, M. A., et al. 1997, ApJ 474, 188
Greig, W. E. 1972, A&A18, 70
Henry, R. B. C., Kwitter, K. B., & Buell, J. 1998, Rev. Mex. A. Conf. 7, 30
Iben, I., Jr. & Renzini, A. 1983, ARA&A21, 271
Leisy, P., & Dennefeld, M. 1996, A&AS 116, 95
Maciel, W. J. 1999, astro-ph/9911090
Manchado, A., Stanghellini, L., Villaver, E., Guerrero, M. A. 2000, ApJ, to be submitted
Monk, D. J., Barlow, M. J., & Clegg, R. E. S. 1988, MNRAS 234, 583
Osterbrock, D. E. 1989, Astrophysics of Gaseous Nebulae and Active Galactic Nucei (Mill
Valley, CA: University Science Books)
Peimbert, M. 1978, IAU Symp. 76, ed. Y. Terzian, Reidel, 215
Peimbert M. 1997, IAU Symp. 180, eds. H. Habing and H. Lamers, Kluwer, 175
Renzini, A. & Voli, M. 1981, A&A94, 175
Russel, S. C., & Dopita, M. A. 1990, ApJS 74, 93
– 14 –
Shaw, R. A., Stanghellini, L., Blades, J. C., Balick, B., & Mutchler, M. 2000, ApJ,
Stanghellini, L., Corradi, R. L. M., & Schwarz, H. E. 1993, A&A 279, 521
Stanghellini, L., Blades, J. C., Osmer, S. J., Barlow, M. J., & Liu, X.-W. 1999, ApJ 510,
Trams, N. R., et al. 1999, A&A 346, 843
van den Hoek, L. B., & Groenewegen, M. A. T. 1997, A&AS 123, 305
Veilleux, S., & Osterbrock, D. E. 1987, ApJS 63, 295
This manuscript was prepared with the AAS LATEX macros v5.0.
– 15 –
Table 1: Average elemental abundances in the LMC
Atom Symm. PNe Asymm. PNeH II Reg.H II Reg.SNR
He 11.0 (19)11.0 (18)10.9 10.91-11.03n.a.
C 8.76 (11)8.23 (7)7.87 n.a.7.66
N 7.83 (19)8.26 (18) 6.976.85-7.277.26-7.45
O 8.30 (20)8.41 (18)8.38 8.18-8.60 8.10-8.54
Ne7.49 (19)7.73 (18)7.64 7.56-7.78 7.11-7.95
S7.04 (15) 7.15 (13)6.67 6.68-7.06.4-7.0
Ar 6.08 (17)6.32 (15)6.205.8-6.376.51-6.65
Note. — Abundances are given as 12+log(N/X). Numbers in parenthesis in columns (2) and (3) indicate
the available PN sample for a given statistics. H II region abundances in column (4) are from Leisy &
Dennefeld (1996). Abundance ranges for H II Regions and Supernova Remnants in columns (5) and (6) are
from Russel & Dopita (1990).
– 16 –
Fig. 1.— S vs. Ne abundance of LMC PNe for morphological types Round (open circles),
elliptical (asterisks), quadrupolar (filled triangles), bipolar core (filled circles), and bipolar
(filled squares). The large crossed circle represents the average for LMC H II regions (see
Table 1 for abundance ranges in H II regions and SNR in the LMC).
Fig. 2.— Argon vs. neon abundance, symbols as in Fig. 1.
Fig. 3.— Oxygen vs. neon abundance, symbols as in Fig. 1.
Fig. 4.— Nitrogen vs. carbon abundance, symbols as in Fig. 1.
Page 20 Download full-text