Hubble space telescope observations and geometric models of compact multipolar planetary nebulae

Abstract

We report high angular resolution Hubble Space Telescope observations of 10 compact planetary nebulae (PNs). Many interesting internal structures, including multipolar lobes, arcs, two-dimensional rings, tori, and halos, are revealed for the first time. These results suggest that multipolar structures are common among PNs, and these structures develop early in their evolution. From three-dimensional geometric models, we have determined the intrinsic dimensions of the lobes. Assuming the lobes are the result of interactions between later-developed fast winds and previously ejected asymptotic giant branch winds, the geometric structures of these PNs suggest that there are multiple phases of fast winds separated by temporal variations and/or directional changes. A scenario of evolution from lobe-dominated to cavity-dominated stages is presented. The results reported here will provide serious constraints on any dynamical models of PNs.

Full text

The Astrophysical Journal, 787:25 (19pp), 2014 May 20 doi:10.1088/0004-637X/787/1/25
C
2014. The American Astronomical Society. All rights reserved. Printed in the U.S.A.
HUBBLE SPACE TELESCOPE OBSERVATIONS AND GEOMETRIC MODELS OF COMPACT
MULTIPOLAR PLANETARY NEBULAE
Chih-Hao Hsia (), Wayne Chau (), Yong Zhang (),andSunKwok()
Department of Physics, The University of Hong Kong, Pokfulam Road, Hong Kong, China; xiazh@hku.hk,
wwlljj1314@gmail.com,zhangy96@hku.hk,sunkwok@hku.hk
Received 2013 December 5; accepted 2014 March 29; published 2014 April 30
ABSTRACT
We report high angular resolution Hubble Space Telescope observations of 10 compact planetary nebulae (PNs).
Many interesting internal structures, including multipolar lobes, arcs, two-dimensional rings, tori, and halos, are
revealed for the first time. These results suggest that multipolar structures are common among PNs, and these
structures develop early in their evolution. From three-dimensional geometric models, we have determined the
intrinsic dimensions of the lobes. Assuming the lobes are the result of interactions between later-developed fast
winds and previously ejected asymptotic giant branch winds, the geometric structures of these PNs suggest that
there are multiple phases of fast winds separated by temporal variations and/or directional changes. A scenario
of evolution from lobe-dominated to cavity-dominated stages is presented. The results reported here will provide
serious constraints on any dynamical models of PNs.
Key words: ISM: jets and outflows – planetary nebulae: general – stars: AGB and post-AGB – stars: evolution –
stars: mass-loss
Online-only material: color figures
1. INTRODUCTION
Recent high angular resolution, high dynamic range imaging
of planetary nebulae (PNs), in particular those made by the
Hubble Space Telescope (HST), have shown that many PNs
have complicated structures. Instead of simple ellipsoidal shells
or bipolar lobes, an increasing number of PNs are found to have
multiple bipolar lobes, multiple shells, arcs, rings, and extended
halos. Of particular interest are quadruple or multiple nebulae,
which have been observed in a number of PNs including M2-46
(Manchado et al. 1996), NGC 2440 (L´
opez et al. 1998), Hen
2-47, M 1-37 (Sahai 2000), NGC 6881 (Kwok & Su 2005),
NGC 6072 (Kwok et al. 2010), NGC 6644 (Hsia et al. 2010),
NGC 6058 (Gill´
en et al. 2013), and NGC 7026 (Clark et al.
2013). It would be interesting to observe a larger sample of
objects in order to determine whether multipolar structures are
isolated or part of a general phenomenon.
The existence of multipolar objects suggests that the fast
outflows responsible for the shaping of PNs are not spherically
symmetric, but are highly collimated. Multipolar structures have
been suggested as the result of simultaneous collimated outflows
in different directions, or the outflow direction has changed with
time (and the term bipolar, rotating, episodic jets, BRETS, has
been used to refer to these phenomenon; L´
opez et al. 1995).
Precession of the mass-losing star’s rotation axis and orbiting
jet with time-dependent ejection velocity due to a binary or
multiple sub-stellar companions has been suggested as one of
the possible causes (Garc´
ıa-Segura 1997). While it is commonly
accepted that the morphological structures of PNs are shaped
by wind interactions (Balick & Frank 2002), the presence of
these multiple structures implies that there must be successive
phases of stellar winds with temporal and directional variations.
In order to determine when do these dynamical events occur, it
would be useful to observe PNs that are young. In this paper, we
present the observations of 10 compact (and probably young)
PNs that have been found to have multipolar morphologies. The
main aim is to investigate nebular intrinsic structures and their
connections with other properties of PNs.
2. OBSERVATIONS AND DATA REDUCTION
Our observed sample is selected from a group of compact,
low-excitation PNs with high radio surface brightness. Since
the central stars of PNs evolve from low to high temperatures
and the nebular density decreases as the result of expansion,
these properties represent the characteristics of PNs early in
their evolution (Kwok 1990).
Observations in this study are based on data obtained by
the HST under programs 8307 (PI: S. Kwok) and 8345 (PI:
R. Sahai) using the Wide Field Planetary Camera 2. Data were
retrieved from the Space Telescope Science Archive. All objects
were observed with the Planetary Camera, which provides a
36.
8×36.
8 field of view (FOV) at a spatial resolution of
0.
045 pixel−1. These objects were imaged with three narrow-
band filters: F502N [O iii](λc=5012 Å, λ=27 Å), F656N
Hα(λc=6564 Å, λ=22 Å), and F658N [N ii](λc=6591 Å,
λ=29 Å), respectively. To allow for the imaging of both the
bright central regions and the faint outer parts of the PNs, the
actual observations were made with different exposure times
(from 20 s to 600 s). The data were processed through the
HST pipeline calibration. Standard flat-field correction and bias
subtraction were performed. All data were taken in two-step
dithered positions to enhance spatial sampling and cosmic rays
removalbyusingthetaskcrrej in the STSDAS package of
IRAF. A journal of these observations is summarized in Table 1.
3. RESULTS
The processed composite-color images of our sample objects
are shown in Figure 1. We find that most of the nebulae have
low [O iii]λ5007/Hαflux ratios, suggesting that they are low
excitation objects, consistent with they being young. Central
stars are detected in all nebulae except IC 5117.
From the Hαfalse-color and composite-color images shown
in Figure 1, it is clear that these nebulae have complex structures.
Most of them have multiple bipolar lobes, and some have central
tori, and faint diffuse halos. Sizes of the lobes and central
ionized tori are measured by fitting the ellipses to the images.
1

The Astrophysical Journal, 787:25 (19pp), 2014 May 20 Hsia et al.
Tab l e 1
Log of Observations
Object Other Name R.A. (J2000.0) Decl. (J2000.0) Filter Exposures (s) Observation Date Program ID
H 1-54 PNG 002.1-04.2 18:07:07.24 −29:13:06.4 F656N 140 ×2, 460 ×2 1999 Aug 21 8345
Hen 2-86 PNG 300.7-02.0 12:30:30.48 −64:52:05.7 F656N 20, 300 ×2 1999 Sep 24 8345
F658N 20, 230 ×2 1999 Sep 24 8345
Hen 2-320 PNG 352.9-07.5 18:00:11.82 −38:49:52.7 F656N 20, 230 ×2 1999 Aug 23 8345
F658N 20, 300 ×2 1999 Aug 23 8345
Hen 2-447 PNG 057.9-01.5 19:45:22.16 +21:20:03.9 F502N 600 ×3 1999 Nov 09 8307
F656N 260 ×2 1999 Nov 09 8307
F658N 500 ×2 1999 Nov 09 8307
IC 5117 PNG 089.8-05.1 21:32:30.97 +44:35:47.5 F502N 160 ×2 1999 Nov 02 8307
F656N 120 ×2 1999 Nov 02 8307
F658N 300 ×2 1999 Nov 02 8307
M 1-30 PNG 355.9-04.2 17:52:58.95 −34:38:23.0 F656N 20, 300 ×2 2000 Feb 23 8345
F658N 20, 230 ×2 2000 Feb 23 8345
M 1-59 PNG 023.9-02.3 18:43:20.20 −09:04:49.1 F656N 20, 300 ×2 1999 Sep 24 8345
F658N 20, 230 ×2 1999 Sep 24 8345
M 1-61 PNG 019.4-05.3 18:45:55.12 −14:27:37.9 F502N 160 ×2 1999 Nov 08 8307
F656N 120 ×2 1999 Nov 08 8307
F658N 260 ×2 1999 Nov 08 8307
M 3-35 PNG 071.6-02.3 20:21:03.77 +32:29:24.0 F502N 520 ×2 1999 Nov 04 8307
F656N 260 ×2 1999 Nov 04 8307
F658N 600 ×3 1999 Nov 04 8307
NGC 6790 PNG 037.8-06.3 19:22:56.97 +01:30:46.5 F502N 160 ×2 1999 Oct 31 8307
F656N 80 ×2 1999 Oct 31 8307
F658N 350 ×2 1999 Oct 31 8307
The measured parameters of these features including their sizes,
position angles (P.A.s) of major axes, and derived inclination
angles (i) are listed in Table 2.
3.1. Individual Objects
Generally, these PNs show two to four pairs of lobes and a
spherical faint halo although their appearances are diverse. The
detailed descriptions of individual PNs are given as follows.
H 1-54 (PNG 002.1-04.2). Although H 1-54 seems to have
an ordinary, unspectacular structure in low-resolution images,
our HST image of this object has revealed very interesting
internal structures (Figure 1). There is a bright core, four pairs
of bipolar lobes, an outer halo, and a circular outer arc. The
largest measured dimension of the object is about 3 measured
from our HST Hαimage. Four pairs of bipolar lobes oriented
at different directions can clearly be seen. These lobes seem to
originate from a common origin within the bright central core
and extend out to the limit of the outer halo. At the outer northern
edge of the nebula is a faint arc-like filament with a radius of
about 1.
93. This arc traces out a well-defined boundary from
N to NW, suggesting that it may be a bow-shocked structure.
The possibility of this arc being the result of interaction with
the interstellar medium is supported by the direction of proper
motion of H 1-54 (P.A. =1◦; Roeser et al. 2010).
Hen 2-86 (PNG 300.7-02.0). Previous optical spectroscopy
has classified Hen 2-86 as a type I PN (de Freitas Pacheco
et al. 1992). The central star of Hen 2-86 is classified as
type WC4 (Girard et al. 2007), with a Zanstra temperature of
67,000–88,000 K (Gleizes et al. 1989). G´
orny et al. (1999)
found that nebula to have a size of 12 ×8 with strong Hα
emission. From our high-resolution HST image reveals two pairs
of bipolar lobes aligned roughly along the northeast (NE) to
southwest (SW) direction. The inner pair of lobes has a size
of 3.
25 ×1.
13 (lobe a−a) and an outer pair has a size
10.
45 ×1.
28 (lobe b−b). This structure is similar to that
of Hen 2-320 (Figure 1) and could be related to the class of
quadrupolar PNs such as M 2-46 (Manchado et al. 1996) and
IPHASX J012507.9+635652 (Mampaso et al. 2006). Around
the waist of the bipolar lobes is an ionized torus of well-defined
shape. A faint halo can also be seen.
Hen 2-320 (PNG 352.9-07.5). At first glance, Hen 2-320
seems to be a typical bipolar nebula with a size of ∼10.How-
ever, closer examination of our HST image suggests the presence
of up to four pairs of bipolar lobes aligned approximately along
the same bipolar axis (Figure 1). The origin of these multi-layer
bipolar lobes is unclear. One possibility is that they represent
episodic ionization fronts with time variation as the result of
multiple interactions by successive stellar winds with a homo-
geneous quiescent circumstellar envelope.
In the [N ii] image shown in the insert, we can see a number
of co-axial rings, two in the eastern lobe and two in the western
lobes. These four rings have been fitted by eye and are marked
as e, f, g, h in the insert gray scale panels. The sizes of the
two-dimensional (2D) rings and the P.A.s of the major axes of
these rings have also presented in Table 2. Three inner rings
(e, f, g) are close to the central part of this object, whereas
the outer ring (h) is far to the central region. Based on the
measurements and visual inspection, we can consider the ring
fto be the counterpart of g. Assuming that these rings are the
parts of tilted circles, the inclination angles (i)ofthethree
inner rings (e, f, g) are approximately 11◦(0◦being edge-on).
The measured separation between ring eand fis 0.
18. Similar
structures (bipolar lobes with multiple rings) can also be found
in the PN M 2-9, Hb 12 (Kwok & Hsia 2007) and Hen 2-104
(Corradi et al. 2001). These rings had been suggested to be
the manifestation of a time-variable, collimated fast wind along
the polar directions of bipolar lobes interact with surrounding
asymptotic giant branch (AGB) circumstellar medium.
Hen 2-447 (PNG 057.9-01.5). This object shows a bright
compact core (0.
84 ×0.
96) elongated at P.A. ∼146◦. Easily
discernible are two bright, extended pairs of lobes (labeled as
2

The Astrophysical Journal, 787:25 (19pp), 2014 May 20 Hsia et al.
Figure 1. HST Hαfalse-color and composite-color images of 10 compact, multipolar PNs. The composite images were mainly made from three bands: [O iii](shown
as blue), Hα(green), and [N ii] (red), except Hen 2-86, Hen 2-320, and M 1-30 in which blue is Hα, green is Hα+[Nii], and red is [N ii]. The right panels are the same
with the left panels but with the morphological features labeled. The positions of the central star are marked with crosses. The black dotted lines denote the position
angles of the minor axes of the bipolar structures. The locations of the 2D rings are marked with red dotted lines. All images are displayed in logarithmic scales. North
is up and east is to the left.
(A color version of this figure is available in the online journal.)
3

The Astrophysical Journal, 787:25 (19pp), 2014 May 20 Hsia et al.
Figure 1. (Continued)
a−aand b−b) with angular sizes of 4.
28 ×0.
71 and 4.
80 ×
0.
85 along the approximately northeast (NE) to southwest (SW)
direction. The waist of this nebula is surrounded by an extensive
ionized torus with strong Hαemission. We note that the major
axis and minor axis are not perpendicular for this nebula. Similar
structures can also be seen in the young PN NGC 6881 and in
some quadrupolar PNs such as M 2-46, K 3-24, and M 1-75
(Manchado et al. 1996). The origin of such asymmetries is still
a mystery, although precession of the central source has been
suggested as the cause (Guerrero & Manchado 1998;Kwok&
Su 2005).
The 5 GHz and 8.4 GHz radio images of this object both
show a bipolar morphology with a bright central torus (Aaquist
&Kwok1990; Kwok & Aaquist 1993). A slight S-shaped
structure, with the eastern extension curving toward the north
and the western extension curving to the south, can be seen.
This is probably the result of two pairs of bipolar lobes, as can
be seen clearly in the HST image.
4

The Astrophysical Journal, 787:25 (19pp), 2014 May 20 Hsia et al.
Figure 1. (Continued)
IC 5117 (PNG 023.9-02.3). This object appears to be a very
dense nebula with higher excitation (Aller & Czyzak 1983).
The nebula consists of a bright inner shell (labeled as c) with
asizeof1.
68 ×1.
16 and two pairs of bipolar lobes (marked
as a−aand b−b) (Figure 1). An extended spherical diffuse
halo can be seen in the Hαand [N ii] images. We note that a
faint, extended structure with a shocked-like boundary is aligned
along the direction of lobe a−a.
The 15 GHz radio image of IC 5117 shows a deep central
emission minimum and two bright peaks along the north and
5

The Astrophysical Journal, 787:25 (19pp), 2014 May 20 Hsia et al.
Figure 1. (Continued)
Tab l e 2
Measured Parameters of Observed Features
H 1-54 Hen 2-86 Hen 2-320 Hen 2-447
Feature Size P.A. iSize P.A. iSize P.A. iSize P.A.
()(
◦)(
◦)(
)(
◦)(
◦)(
)(
◦)(
◦)(
)(
◦)
a−a2.61 ×0.53 2 ±2··· 3.25 ×1.13 49 ±3··· 5.97 ×2.02 79 ±4··· 4.28 ×0.71 44 ±2
b−b2.06 ×0.37 8 ±3··· 10.45 ×1.28 46 ±3··· 9.90 ×1.39 81 ±3··· 4.80 ×0.85 65 ±3
c−c3.15 ×0.49 134 ±2··· ··· ··· ··· 8.47 ×1.39 80 ±3··· ··· ···
d−d2.39 ×0.28 154 ±3··· ··· ··· ··· 2.52 ×1.39 75 ±3··· ··· ···
e··· ··· ··· ··· ··· ··· 1.08 ×0.22 168 ±212±3a··· ···
f··· ··· ··· ··· ··· ··· 0.91 ×0.19 168 ±212±3a··· ···
g··· ··· ··· ··· ··· ··· 0.95 ×0.16 168 ±210±3a··· ···
h··· ··· ··· ··· ··· ··· 2.08 ×0.98 164 ±328±3a··· ···
Ionized torus 2.91 ×1.58 88 ±333±3a2.66 ×1.44 137 ±3··· ··· ··· ··· ··· ···
IC 5117 M 1-30 M 1-59 M 1-61
Feature Size P.A. iSize P.A. iSize P.A. iSize P.A.
()(
◦)(
◦)(
)(
◦)(
◦)(
)(
◦)(
◦)(
)(
◦)
a−a3.03 ×0.68 88 ±2··· 4.58 ×1.40 78 ±2··· 7.75 ×3.55 15 ±3··· 2.89 ×1.13 34 ±3
b−b3.55 ×0.46 115 ±3··· 4.06 ×1.94 129 ±4··· 18.04 ×3.74 120 ±4··· 2.22 ×1.04 112 ±4
c−c··· ··· ··· ··· ··· ··· 6.78 ×2.54 130 ±3··· 5.04 ×0.98 149 ±3
d−d··· ··· ··· ··· ··· ··· ··· ··· ··· 4.80 ×0.74 162 ±3
Minor axis ··· ··· ··· ··· 17 ±3··· ··· ··· ··· ··· ···
Inner ring ··· ··· ··· 3.48 ··· ··· ··· ··· ··· ··· ···
Outer ring ··· ··· ··· 4.22 ··· ··· ··· ··· ··· ··· ···
Jet ··· ··· ··· 4.52 109 ±3··· ··· ··· ··· ··· ···
M 3-35 NGC 6790
Feature Size P.A. iSize P.A. i
()(
◦)(
◦)(
)(
◦)(
◦)
a−a2.86 ×1.24 58 ±3··· 4.15 ×0.98 118 ±3···
b−b2.30 ×0.78 170 ±4··· 4.68 ×1.06 133 ±4···
c−c0.8 60 ±3··· 1.16 ×0.89 119 ±250±3a
d−d0.69 172 ±3··· ··· ··· ···
Note. aDerived from the major–minor axis ratio assuming the orientation angle of sky plane is 90◦.
south sides of this object (Aaquist & Kwok 1991;Lee&Kwok
2005). The radio image corresponds to the central shell in the
HST image, but the two pairs of bipolar lobes and the halo are not
detected in the radio image, probably due to the low dynamic
range of the radio observations. The radio structure has been
ascribed to a prolate ellipsoidal shell projected onto the plane
of the sky (Aaquist & Kwok 1991). The extinction map of this
nebula shows a similar distribution to that of the radio map, with
two extinction peaks fall slightly outside the radio peaks, which
indicates that the major axis of the object is rotated toward us
by ∼40◦(Lee & Kwok 2005).
M 1-30 (PNG 355.9-04.2). The main structure of M 1-30
consists a double bright shells which can be interpreted as a
pair of bipolar lobes (b−b) inclined w.r.t. the plane of the sky.
6

The Astrophysical Journal, 787:25 (19pp), 2014 May 20 Hsia et al.
Figure 2. HST [N ii] images of M 1-30 after the application of an unsharp mask in order to better show the faint outer structures. The right panel has the same FOVs
as the left panel, but with morphological features are marked. The inner and outer shells are marked with red dotted and solid lines, respectively. North is up and east
is to the left. The intensity display is on a logarithmic scale and the gray scale bar is given at the bottom in units of counts per pixel.
(A color version of this figure is available in the online journal.)
The measured angular size of this pair of lobes is 4.
06 ×1.
94
oriented at P.A. =129◦. Another pair of bipolar lobes (a−a)
can be seen at P.A. =78◦with an angular size of 4.
58 ×1.
40.
Outside of the main nebula is a faint, extended halo with a size
of ∼7, which shows strong Hαemission.
After the application of an unsharp mask to the [N ii] image,
additional outer structure can be seen (Figure 2). These include
two concentric shells, an arc, and one pair of collimated jets.
Two concentric rings (hereafter called the inner shell and outer
shell) and the arc both show the apparent ridges at their edges
in the Hαand [N ii] images. The measured angular radii of
inner shell, outer shell, and the arc are 1.
74, 2.
11, and 2.
87,
respectively. Similar multiple shell structures have been seen
in other PNs such as NGC 6543, NGC 7027 (Terzian & Hajian
2000), NGC 40, NGC 1535, and NGC 3918 (Corradi et al. 2004).
These multiple shells of enhanced brightness represent results
of higher density and are probably manifestations of periodic
mass loss during the AGB. These are different from the 2D rings
observed in Hen 2-320, which are caused by time-variable fast
winds.
A pair of collimated jets with a size of 4.
52 can also be seen
in Figure 2. The orientation of this structure is at P.A. =109◦,
which is located between the two prominent lobes discussed
earlier. We also note that the major axis of the jets and the minor
axis of this nebula seem to be perpendicular. It is possible that
the pair of jets could be a third pair of lobes which are aligned
nearly along the plane of the sky.
M 1-59 (PNG 023.9-02.3). M 1-59 is usually classified as a
bipolar nebula but examination of our HST image of this object
shows a much more complex structure. There are at least three
pairs of closed-end bipolar lobes (labeled as a−a,b−b, and
c−cin Figure 1), the largest of which (b−b) has an angular
extent of ∼18. In the center of the nebula is a hollow shell (or
cavity, marked as d) with a size of 5.
42 ×4.
64. As in the case
of IC 5117, this central hollow shell could be a result of the
projection of the fourth pair of lobes which are aligned nearly
along the line of sight, or a bubble created by the shocked fast
wind from the central star.
AH
22.12 μm image of M 1-59 shows a bright central region
with emission in the lobes (Guerrero et al. 2000). The 5 GHz
radio image of the object (Aaquist & Kwok 1990)showsa
distinctly bilobate appearance with a central emission minimum
and two bright radio peaks, which indicates an elongation toward
the NW–SE direction along the major axis.
M 1-61 (PNG 019.4-05.3). M 1-61 is a clear multipolar nebula
with at least four pairs of bipolar lobes (Figure 1). Two pairs of
lobes (c−c, and d−d) are nearly aligned with each other but
the other pair (a−a, and b−b) probably represents lobes with
axes highly inclined w.r.t. the plane of the sky. A ellipsoidal
central shell (marked as e) can also be seen.
The 15 GHz radio images show a shell-like structure with
two prominent peaks with roughly the same brightness aligned
along the E–W direction (Aaquist & Kwok 1991;Lee&Kwok
2005). These two bright peaks are also apparent in the 5 GHz
map (Aaquist & Kwok 1990).
M 3-35 (PNG 071.6-02.3). M 3-35 is a medium excitation
PN with a central star temperature of ∼60,000 K (Kaler 1986).
Our HST image shows two pairs of bipolar lobes (marked as
a−aand b−bin Figure 1) with different orientations and a
bright central shell (labeled as e)ofsizeof0.
55 ×0.
64. The
orientation angles of the two bipolar lobes are measured to be
P.A. =58◦±3◦and 170◦±4◦for lobes a−aand b−b,
respectively. The two bipolar lobes have similar projected sizes
on the sky, with a−ahaving a size of ∼2.
86 and b−basize
of ∼2.
30.
In the [O iii] image (shown in the insert), two pairs of cavities
(c−cand d−d) associated with the bipolar lobes a−aand
b−bare also found in the central region. These two pairs of
7

The Astrophysical Journal, 787:25 (19pp), 2014 May 20 Hsia et al.
Figure 3. HST image (left), three-dimensional mesh (center), and rendered image (right) of H 1-54. The mesh of the model is oriented to best display the three-
dimensional morphology and is not related to the actual orientation of the PN with respect to the plane of the sky. The scattering halo is not shown in the mesh. A
Gaussian blur has been applied to the rendering.
(A color version of this figure is available in the online journal.)
cavities have symmetry axes that lie at the P.A.s of 60◦±3◦and
172◦±3◦, which are closely aligned with the orientations of the
two pairs of bipolar lobes.
The 15 GHz radio continuum images clearly resolve the
central region of this nebula (Aaquist & Kwok 1991;Lee
&Kwok2005). A deep central emission minimum and a
double-peaked structure with roughly the same brightness are
aligned along the minor axis of the a−abipolar lobes in the
optical image. The radio structure can be modeled by a prolate
ellipsoidal shell with the major axis rotated toward us by ∼60◦
(Aaquist & Kwok 1991). Lee & Kwok (2005) also found that the
extinction map of this object has similar distribution to that of the
radio map, which shows the locations of high extinction regions
coinciding with the double-peaked structure. A comparison of
the optical [O iii] enlarged image and dust extinction map (Lee
&Kwok2005) of this nebula, we can see that the cavity d−d
and two high-extinction regions have approximately the same
locations. It is possible that the presence of lobe b−bcould
be the result of illumination through the cavity d−din the
equatorial torus.
NGC 6790 (PNG 037.8-06.3). NGC 6790 is another nebula
seen with multi-layer shells. There are two outer bipolar lobes
(labeled as a−aand b−b) with axes along the NW–SE
direction. The two inner shells (marked as cand c)areshown
in more detail in the insert [N ii] image. The symmetry axes of
lobes a−aand b−bhave P.A.s of 118◦±3◦and 133◦±4◦,
respectively. The major axes of the two inner shells are similar,
both oriented at P.A. =119◦±2◦. Similar structures can be seen
in other multipolar PN such as NGC 6644 (Hsia et al. 2010) and
He 2-47 (Sahai 2000). These structures probably represent the
presence of another pair of lobes aligned nearly along the line
of sight. If the projection of the third pair of lobes (referred to
as lobe c−c) is indeed composed of these two inner shells,
then we can derive an inclination angle of about 50◦for the lobe
c−c(assuming the orientation of sky plane is 90◦).
While the HST image shows an elongated structure with a
major axis dimension of ∼5, the 5 GHz and 15 GHz ratio
images are roughly circular with a diameter of ∼2. The slight
elongation along the NW–SW direction has been suggested to
be the result of the nebula being density bounded along this
direction (Aaquist & Kwok 1990,1991). This is consistent
with the present HST observations, where the nebula is seen
to be density bounded along the bipolar axis but ionization
bounded elsewhere. A faint, extended halo with a well-defined
spherical shape (with a size of ∼6.
54) can also be seen, which
are consistent with the size measured by Hua et al. (1993).
4. THE MODELS
It is clear from Section 3that the structures of these PNs
are complex. Because of the presence of emission lines, young
PNs are much brighter optically than proto-PNs, which shines
by dust-scattered starlight. It is therefore much easier to see
the fainter structures in young PNs than proto-PNs. We would
like to take advantage of the excellent quality of these HST
images to discern the intrinsic structures of the nebulae. In
order to achieve this goal, we have performed three-dimensional
(3D) modeling using the software package Shape (Steffen
et al. 2011) and compare the rendered model images with
observed images. Using the observed images as a guide, we
have created models of multiple bipolar lobes as a starting point.
The sizes and orientations of the lobes are given in Table 3.
The surface brightness of the simulated images represented
integrated emission measures (n2
ed) along the line of sight.
H 1-54. As discussed in Section 3,theHST image of
H 1-54 (shown in the left panel of Figure 3), shows four pairs
of bipolar lobes, an ionized torus in the central region, a halo
component, and an arch-shaped feature. To model this PN, it has
been assumed that all the bipolar elements in the PN are iden-
tical and the differences in their apparent lengths are entirely
attributed to the projection effect. Since no shell-like structure
is seen in the observed image, the elements in the model are
given a uniform density distribution. The arc feature is emu-
lated using a thin, incomplete spherical shell that spans 80◦.
A 3D mesh is also provided in the center panel of Figure 3to
aid visualization of the morphology. As seen from the rendered
image in the right panel of Figure 3, in which a Gaussian blur
has been applied to account for the telescope beam, the model
is able to generate images that closely resemble the observed
image.
Hen 2-86. From the HST image displayed in the left panel
of Figure 4, three elements can be identified in the PN: an
outer pair of bipolar lobes, an inner pair of bipolar lobes and
an ionized torus. Under the assumption that both the outer and
inner bipolar elements lie more or less in the plane of the sky,
the elements are created according to measurements listed in
8

The Astrophysical Journal, 787:25 (19pp), 2014 May 20 Hsia et al.
Tab l e 3
Parameters of Shape Modelsa,b
Object Name Component Component Morphology Absolute Dimensions Density Distribution P.A. Inc.
(l ×w/re,r
i)(
◦)(
◦)
H1-54 a−aBipolar 3.20 ×0.50 Uniform 2 34
b−bBipolar 3.20 ×0.50 Uniform 8 −49
c−cBipolar 3.20 ×0.50 Uniform −44 0
d−dBipolar 3.20 ×0.50 Uniform −26 −41
ISM Interaction Incomplete shell 3.52, 3.36 Uniform −30 0
Ionized Region Toroidalc0.96, 0.32 Uniform −50
Hen 2-86 aBipolar (55% squeeze) 3.22 ×1.12 Uniform 46 0
bBipolar (50% squeeze) 10.4 ×1.25 Uniform 49 0
Ionized Region Toroidalc1.56, 0.52 Uniform 47 0
Hen 2-320 aBipolar (50% squeeze) 5.94 ×2.08 Shell 79 0
bBipolar (50% squeeze) 8.47 ×1.39 Shell 80 0
cBipolar (50% squeeze) 9.90 ×1.39 Shell 81 0
dBipolar (75% squeeze) 1.58 ×0.99 Shell 90 0
Hen 2-447 a−aBipolar (20% squeeze) 4.80 ×0.85 Shell 44 −63
b−bBipolar (20% squeeze) 4.80 ×0.85 Shell 80 0
IC 5117 a−aBipolar (20% squeeze) 3.60 ×0.72 Shell 88 0
b−bBipolar (20% squeeze) 3.60 ×0.72 Shell 115 31
cBipolar 1.80 ×0.95 Uniform −75 30
M1-30 a−aBipolar (50% squeeze) 4.60 ×1.41 Shell 78 0
b−bBipolar (20% squeeze) 4.09 ×1.95 Shell −51 0
M1-59 a−aBipolar 18.0 ×3.53 Shell 15 65
b−bBipolar 18.0 ×3.73 Shell −60 0
c−cBipolar 18.0 ×2.51 Shell −50 68
d−dBipolar 18.0 ×3.73 Shell −50 80
Ionized region Toroidalc4.50, 1.80 Uniform −10 −70
M1-61 a−aBipolar 5.00 ×0.73 Shell 34 70
b−bBipolar 5.00 ×1.03 Shell −68 64
c−cBipolar 5.00 ×0.97 Shell −31 0
d−dBipolar 5.00 ×1.12 Shell −18 −18
eBipolar (pole-on) 5.00 ×0.73 Shell −2−82
M3-35 a−aBipolar 2.70 ×1.17 Shell 59 0
b−bBipolar 2.70 ×1.17 Shell −10 36
e Bipolar 2.70 ×1.17 Shell 60 74
NGC 6790 a−aBipolar 4.70 ×0.98 Shell −62 −27
b−bBipolar 4.70 ×1.07 Shell −47 0
c−cBipolar 1.81 ×0.89 Shell −61 50
Notes.
aEmission intensity of the Shape models are proportional to n2,wherenis the number density of emitters. Unless otherwise
specified, emitters are distributed on the surface of the listed components in a shell of finite thickness. For certain objects, the
distribution of emitter may also follow an angular dependency as specified in the text.
bP.A. is defined from the north rotating eastward and Inc. referenced to the plane of the sky.
cP.A. and Inc. of toroidal components referenced to the axis of rotational symmetry. Absolute dimensions of toroidal components
are specified by reand ri, which are respectively the internal and external radius.
Table 2. Also added to the bipolar lobes is a slight tightening at
their waists—a property that is evident in the HST image. Like
H 1-54, no shell structure is apparent; hence, the elements in
the model are given a uniform density distribution. A 3D mesh
of the model is presented in the center panel of Figure 4with
the numerical details of the model provided in Table 3. We can
see that the Shape rendered image in the right panel of Figure 4
shows a good resemblance to the observed image.
Hen 2-320. The lobes of this nebula show shell-like structures
and we assume a thin (0.
1) shell geometry for all the four bipolar
lobes identified in the left panel of Figure 5. All four pairs of
lobes are assumed to all reside on the plane of the sky, and a
waist tightening has been applied to the model. Parameters of
the model are derived from the measurements in Table 2are
listed in Table 3.
Hen 2-447. Like Hen 2-320, the bipolar lobes of Hen 2-447
appear to have a shell-like structure as seen in the HST im-
age displayed in the left panel of Figure 6. Initial inspection
of the observed image suggests three bipolar elements, which
can easily be mimicked using three pairs of identical bipolar
lobes oriented at different P.A.s and inclinations. Interestingly,
further investigation reveals that two bipolar components are
already sufficient to produce a rendering (right panel of Fig-
ure 6) that is consistent with observations. The bright region in
between the two lobes a−aand b−bcan be accounted for by
the intersecting regions of the two bipolar lobes. This model is
9

The Astrophysical Journal, 787:25 (19pp), 2014 May 20 Hsia et al.
Figure 4. HST image (left), three-dimensional mesh (center), and rendered image (right) of He 2-86. The scattering halo is not shown in the mesh. A Gaussian blur
has been applied to the rendering.
(A color version of this figure is available in the online journal.)
Figure 5. HST image (left), three-dimensional mesh (center), and rendered image (right) of He 2-320. The scattering halo and the innermost bipolar element are not
shown in the mesh. A Gaussian blur has been applied to the rendering.
(A color version of this figure is available in the online journal.)
Figure 6. HST image (left), three-dimensional mesh (center), and rendered image (right) of He 2-447. The scattering halo is not shown in the mesh. A Gaussian blur
has been applied to the rendering.
(A color version of this figure is available in the online journal.)
particularly enlightening as it demonstrated that features inter-
preted as a separate component can easily arise from the overlap
of other components. Signification simplification in PN models
can result from similar interpretation of observed feature.
IC 5117. Clearly identifiable from the HST image of IC 5117
in the left panel of Figure 7are two pairs of bipolar lobes. As
they display a shell-like appearance, these bipolar features are
likely to be cavities. In the model, these cavities are assumed
10

The Astrophysical Journal, 787:25 (19pp), 2014 May 20 Hsia et al.
Figure 7. HST image (left), three-dimensional mesh (center), and rendered image (right) of IC 5117. The scattering halo is not shown in the mesh. A Gaussian blur
has been applied to the rendering.
(A color version of this figure is available in the online journal.)
Figure 8. HST image (left), three-dimensional mesh (center), and rendered image (right) of M 1-30. The scattering halo is not shown in the mesh. A Gaussian blur has
been applied to the rendering.
(A color version of this figure is available in the online journal.)
to be identical to one another and their apparent differences
are attributed to the projection effect. The interpretation of
the central feature, on the other hand, is much less straight
forward. While there exist ostensible variations in the surface
brightness, it is difficult to assert with any degree of certainty that
it is shell-like. After experimenting with a couple of possible
morphologies, including ones in which the central feature is
shell-like and more or less perpendicular to the plane of the
sky, we found that the model in which the central element
has a bipolar structure with an uniform density distribution
produces renderings that most closely resembles the HST
observations. The 3D mesh and the Shape rendered image of
the model is displayed in the center and right panel of Figure 7
respectively.
M 1-30. M 1-30 consists of two pairs of bipolar lobes. The
observed complex S-shaped feature is emulated by the inter-
secting region of two bipolar elements. The bipolar components
are constructed following the measured dimensions and P.A.
listed in Table 2and are assumed to lie on the plane of the sky.
The waist of each bipolar cavity is then tightened according to
the parameters provided in Table 3. Rendered from the 3D mesh
shown in the center panel of Figure 8, the image displayed in
the right panel successfully captures the characteristic S-shaped
feature in the HST image.
M 1-59. M 1-59 consists of five bipolar elements that appear
to exhibit shell-like features (left panel of Figure 9). We assume
that the bipolar components are of equal lengths but with
different widths. Differences in the apparent lengths seen in
the observed images arise solely from the projection effect.
Inclination of each element is derived from the measurements
listed in Table 2. Specific parameters used in the model are given
in Table 3. Using the 3D mesh in the center panel of Figure 9,
we are able to simulate the observed structure of this nebula
(right panel of Figure 9).
M 1-61. Four pairs of bipolar lobes can be identified from
the HST image of M 1-61. Similar to M 1-59, the these bipolar
components are assumed to be of identical absolute length but
different widths. Their projected lengths, provided in Table 2,
are used to derive the inclination of each element. The bipolar
lobes in this PN are once again likely to be cavities; hence,
they are modeled using a shell structure. The central feature,
not shown in the 3D mesh in the center panel of Figure 10
to improve visibility, is emulated by carving out another pair
of bipolar lobes from the halo component. Shown in the right
panel of Figure 10,theShape rendered image of M 1-61 display
strong resemblance to the HST image. Features arising from the
intersecting regions of the different components are also clearly
visible in the rendering.
11

The Astrophysical Journal, 787:25 (19pp), 2014 May 20 Hsia et al.
Figure 9. HST image (left), three-dimensional mesh (center), and rendered image (right) of M 1-59. A Gaussian blur has been applied to the rendering.
(A color version of this figure is available in the online journal.)
Figure 10. HST image (left), three-dimensional mesh (center), and rendered image (right) of M 1-61. The scattering halo is not shown in the mesh. A Gaussian blur
has been applied to the rendering.
(A color version of this figure is available in the online journal.)
Figure 11. HST image (left), three-dimensional mesh (center), and rendered image (right) of M 3-35. A Gaussian blur has been applied to the rendering. The gray-scale
insert shows the central cavity.
(A color version of this figure is available in the online journal.)
M 3-35. Two bipolar cavities are identified in the HST
image of M 3-35. They are replicated in the Shape model using
identical bipolar shell at the orientations listed in Table 3.
Although the narrow band [Oiii] image seems to suggest
another two pairs of bipolar lobes, the addition of such structure
introduces inconsistencies in the rendering. To replicate the
central feature, however, another bipolar component, identical
to the other two, is introduced and is assumed to align on the
line of sight. The resulting rendering, displayed in right panel
of Figure 11 is in good agreement with the observation. Since a
shell structure is not particularly obvious in the central feature, a
model in which a uniform density distribution is assigned to the
12

The Astrophysical Journal, 787:25 (19pp), 2014 May 20 Hsia et al.
Figure 12. HST image (left), three-dimensional mesh (center) and rendered image (right) of NGC 6790. The scattering halo is not shown in the mesh. The mesh of
the model is oriented to best display the three-dimensional morphology and is not related to the actual orientation of the PN with respect to the plane ofthesky.A
Gaussian blur has been applied to the rendering. The gray-scale insert shows the central cavity.
(A color version of this figure is available in the online journal.)
third bipolar component has also been created. The rendered
image such a model, nonetheless, does not offer significant
improvement as some surface brightness variation is visible
in the central region of the HST image.
NGC 6790. As can be seen in the left panel of Figure 12,the
outer region of the PN composes of two bipolar components.
They are modeled using identical bipolar shells whose orien-
tations are derived from measurements provided in Table 2.
Elongation of the central feature rules out the possibility that it
is formed by a single bipolar element that is oriented slightly
off the line of sight. The central features must, therefore, be
modeled using multiple elements. While the outer central fea-
ture is replicated using a third, smaller bipolar component, the
innermost central feature is emulated by carving out a spherical
cavity from the halo component. The rendered image in the right
panel of Figure 12 not only shows strong resemblance with the
HST image, features that arise from the intersection of the outer
bipolar elements are also successfully replicated.
In our models, the lobes of each PN have approximately
equal width-to-length ratios. If we assume that the width-to-
length values are uniquely determined by the velocity of the
outflow respect to the local sound speed, we can conclude that
the outflows at different directions have nearly equal velocities.
5. DISCUSSIONS
Although at first glance many of the objects in our sample of
compact PNs have simple elliptical or bipolar structures, careful
examinations of the HST images have revealed new internal
structures. Based on these observations, we would classify
H 1-54, Hen 2-447, IC 5117, M 1-30, M 1-59, M 1-61, and
M 3-35 (hereafter Class I) as multipolar nebulae with multiple
pairs of bipolar lobes oriented at different directions. The other
three objects: Hen 2-86, Hen 2-320, and NGC 6790 (hereafter
Class II) all have multiple lobes aligned along almost the same
direction. All objects possess point symmetry—meaning that
all the multiple lobes have a common point of origin. The
morphology of the Class I objects gives the impression that the
lobes are the result of a precessing phenomena, and the Class II
objects may be the result of consecutive ejections collimated by
a similar mechanism.
All the Class I multipolar PNs exhibit a central cavity, which
is even more pronounced than the lobes for some PNs (e.g.,
M 1-30). Central cavity of this type has not been seen in proto-
PNs (see, e.g., Sahai et al. 2007) and the existence of these
cavities may be the result of stellar evolution. After the stellar
wind ceases, the increasing thermal pressure of the photoionized
region becomes dominant. The expansion of the photoionized
gas leads to the destruction of the multipolar lobes and other
small-scale structures (Garc´
ıa-Segura et al. 2006).
A slightly different scenario has been presented by Huarte-
Espinosa et al. (2012), who successfully modeled the morpho-
logical transition from bipolar to elliptical by assuming that the
systems begin with a jet flow (during the proto-PN phase) but
are followed by a spherical fast wind (during the PN phase).
Therefore, our observations may represent an evolution during
the young PN age—the transition from a lobe-dominant to a
cavity-dominant phase. Figure 13 shows a schematic diagram
of the morphology transition. This is consistent with the ob-
servations that many young PNs show bipolar or multipolar
structures while evolved PNs are mostly round or elliptical.
The observed lobes show sharp boundaries at their tips (e.g.,
H 1-54 and M 1-61) or/and along the lateral edges (e.g., Hen
1-320 and NGC 6790), suggesting that the lobes are confined
by an external medium. The observed morphology of the lobes
suggested that they represent low-density cavities carved out by
colliminated fast winds.
Recent hydrodynamic simulations have shown that multipolar
structures and secondary lobes can be naturally produced
through the interaction between a fast wind and a highly
inhomogeneous and filamentary shell (Steffen et al. 2013), and
as a consequence, there is no need for a collimation mechanism
of the outflow. We find that some of the PNs closely resemble
the modeled results. For example, M1-59 and M1-61 display a
pair of main lobes with less well-defined boundaries and a few
secondary lobes and filamentary structures in the central regions.
These morphological structures are similar to the modeled
images of Steffen et al. (2013, Figures 6–11). The filamentary
shell may be formed from instabilities of the photodissociation-
shock front in the pre-PN stage, as suggested by Garc´
ıa-Segura
(2010). If this is case, the directions of the secondary lobes
should be nearly random. The models are apparently not able to
interpret the coaxial lobes of the Class II objects.
From the spectral energy distributions (SEDs) of these nebu-
lae (Appendix), we do not find obvious difference between the
SEDs of Class I and Class II objects. However, the SEDs of
13

The Astrophysical Journal, 787:25 (19pp), 2014 May 20 Hsia et al.
Figure 13. Schematic diagram of the morphology transition from lobe-dominant to cavity-dominant. The evolved PN NGC 7293 apparently exhibits vestiges of
previous outflows.
(A color version of this figure is available in the online journal.)
Hen 2-86 and Hen 2-320 exhibit a steeper rise in the wavelength
range from 5 to 30 μm than those of Class I PNs (Figure 13).
This is conceivable in that differing from optical emission, the
IR emission of the two PNs dominantly originates from a nearly
edge-on torus, which is very optically thick so that the inner
warm regions are largely obscured. The central torus could play
an important role in collimating the outflows in Hen 2-86 and
Hen 2-320. It has been suggested that many PNs with mixed
dust chemistry exhibit point-symmetric structures (Stanghellini
et al. 2012). This proposed correlation needs to be explored
further by infrared spectroscopic observations.
6. CONCLUSIONS
Although a number of PNs with multipolar morphology have
been reported recently (Manchado et al. 2011; Sahai et al.
2011), the results of this paper suggest that multipolar nebu-
lae are much more common than previously thought. In the
past, many multipolar PNs have been misclassified as “bipo-
lar” or/and “elliptical.” Examples of such misclassification in-
clude NGC 2440, NGC 6072 (Corradi & Schwarz 1995) and
NGC 6644 (Stanghellini et al. 1993), but these objects are actu-
ally multipolar (L´
opez et al. 1998;Kwoketal.2010;Hsiaetal.
2010). It is possible the present sample of multipolar nebulae
only represent the tip of the iceberg, as other bipolar/elliptical
PNs may also be multipolar when deeper imaging reveals fainter
structures.
Our sample of PNs are generally considered as young objects.
It would be interesting to explore whether the multinebular
properties are related to age. If the multinebular property is
more common in young PNs, it could be the result of either
the multipolar lobes are washed out/smoothed over by later
dynamical processes, or the decreased surface brightness of the
lobes in evolved PNs make them more difficult to detect.
From 3D simulations, we are able to reproduce the observed
structures with a set of multiple bipolar lobes. One issue of
creating 3D models from 2D images is its inherent degeneracy.
As a result of the projection effect, different 3D structures
can lead to the same apparent image. A way to break this
degeneracy is to make use of position–velocity diagrams from
imaging-spectroscopic observations. Unfortunately, there exist
no kinematic data for most objects in this study. Even for the
PNs with spatially resolved spectra available, the resolution is
so poor that they can all be reproduced using uniform spheres
with a Hubble flow. In spite of the possible non-uniqueness
of these models, they do illustrate that the observed nebular
structures of these objects need to be explained by multiple
sets of geometric/kinematic components. These components
are probably the result of interactions between multiple phases
of mass loss, separated by time and/or a change in ejection
directions. We hope that these observations will motivate further
theoretical work on the dynamics of PNs.
We thank Nico Koning for helpful discussions on the SHAPE
modeling. Some of the data presented in this paper were obtained
from the Multimission Archive at the Space Telescope Science
Institute (MAST). STScI is operated by the Association of
Universities for Research in Astronomy, Inc., under NASA
contract NAS5-26555. Support for MAST for non-HST data
is provided by the NASA Office of Space Science via grant
NAG5-7584 and by other grants and contracts. This work was
partially supported by the Research Grants Council of the Hong
Kong Special Administrative Region, China (project no. HKU
7031/10P.).
APPENDIX
SPECTRAL ENERGY DISTRIBUTION
One of the characteristics of young PNs is strong infrared
excess. Since the dust component of PNs expands and cools
as PN ages, the degree of infrared excess is largest in PNs
14

The Astrophysical Journal, 787:25 (19pp), 2014 May 20 Hsia et al.
Figure 14. SEDs of 10 multipolar PNs in the HST sample. The filled squares are Band Vphotometry of the central stars. The filled and open triangles, filled circles,
open diamonds, inverse filled triangles, open squares, open circles, and asterisks are from the DENIS, 2MASS, GLIMPSE, WISE,MSX, MIPSGAL, IRAS,andAKARI
survey, respectively. The light asterisks represent the uncertain AKARI detections. Note that some fluxes measured from IRAS 12 and 100 μm are upper limits. When
available, the ISO and Spitzer IRS spectra are also plotted. Some of the spectral features are marked. The three BB-like curves (from left to right) represent the central
star, and the two dust components. The dotted lines are model curves for the nebular continuum emission. The total flux from all components are plotted as solid lines.
in their early stages of stellar evolution. Analysis of the SED
of young PNs suggests that a significant fraction of their total
energy output is in the infrared (Zhang & Kwok 1991). We
have therefore constructed the SEDs of our sample objects
using archival data and the results are shown in Figure 14.In
the UV, we have taken data from the International Ultraviolet
Explorer (IUE) low dispersion spectrograph. The IUE spectra
of two objects (IC 5117 and NGC 6790) are extracted from
the IUE Data Analysis Center (IUEDAC). In the infrared, the
Spitzer Infrared Spectrograph (IRS) measurements and Infrared
15

The Astrophysical Journal, 787:25 (19pp), 2014 May 20 Hsia et al.
Figure 14. (Continued)
Space Observatory (ISO) observations are used. A journal of
the IUE,Spitzer IRS, and ISO observations of these nebulae are
given in Table 4. For photometric measurements, the optical B
and Vmagnitudes of central stars of the nebulae are obtained
from Shaw & Kaler (1985) and Tylenda et al. (1989,1991),
respectively. In the infrared, data are taken from the Deep Near-
Infrared Southern Sky Survey (DENIS), Two Micron All Sky
Survey (2MASS), Midcourse Space Experiment (MSX), AKARI
(Tajitsu & Tamura 1998) and Infrared Astronomical Satellite
(IRAS) catalogs. We have also made used of the a mid-infrared
data from the Spitzer and Wide-field Infrared Survey Explorer
(WISE; Wright et al. 2010) surveys to measure the integrated
fluxes of the sample PNs using method described in Zhang
&Kwok(2009) and Hsia & Zhang (2014). A summary of
archival data used is given in Table 5. We note that some of the
filters are broad, therefore color corrections and source aperture
calibrations may be needed. The values for the color corrections
and aperture calibration factors used in Figure 14 and Table 5
can be found in Wright et al. (2010) and Reach et al. (2005).
The nebular physical parameters of these objects are taken
from the literature. The interstellar extinction coefficients are
taken from Tylenda et al. (1989,1991,1992), Cahn et al.
(1992), Shaw & Kaler (1985), and Girard et al. (2007). The
Hβfluxes are from Acker et al. (1992). The nebular electron
densities (ne) and electron temperatures (Te)aretakenfrom
Tylenda et al. (1991), Cahn et al. (1992), Shaw & Kaler (1985),
Stanghellini & Kaler (1989), and Zhang & Kwok (1991).
Tab l e 4
IUE,IRSandISO Observations
PN Name Instrument Exposures (s)
IUE spectra
G 037.8-06.3 NGC 6790 SWP 15446 2100
SWP 31831 3600
SWP 15411 5400
LWR 08709 1200
LWR 11935 1500
LWR 11922 3600
G 089.8-05.1 IC 5117 SWP 25835 1800
SWP 31825 9000
LWR 05883 1800
LWR 05884 17700
IRS observation
G 019.4-05.3 M 1-61 AORkey 25837824 12.58
G 071.6-02.3 M 3-35 AORkey 18183424 14.68
G 300.7-02.0 Hen 2-86 AORkey 25853440 12.58
G 352.9-07.5 Hen 2-320 AORkey 11326464 12.58
ISO spectra
G 037.8-06.3 NGC 6790 TDTkey 53400762 766
TDTkey 13401107 1062
TDTkey 13401608 1330
G 089.8-05.1 IC 5117 TDTkey 36701824 1140
TDTkey 36701822 1858
16

The Astrophysical Journal, 787:25 (19pp), 2014 May 20 Hsia et al.
Tab l e 5
Photometric Measurements
Objects
Filters H 1-54 Hen 2-86 Hen 2-320 Hen 2-447 IC 5117 M 1-30 M 1-59 M 1-61 M 3-35 NGC 6790
Central star
B(mag) 15.7(1) 19.2(2) 14.4(1) ··· 17.5(3) 16.6(1) ··· 17.1(2) ··· 16.3(2)
V(mag) 15.4(1) 18.4(2) 14.3(1) ··· 16.7(3) 16.4(1) ··· 16.8(2) ··· 15.5(2)
Nebulaa
DENIS Ib(mag) 13.614 13.912 12.103 ··· ··· 14.042 13.700 11.602 ··· 11.963
DENIS Jb(mag) 12.668 11.942 11.330 ··· ··· 12.921 12.250 11.529 ··· 10.598
DENIS Kb(mag) 11.886 10.746 10.526 ··· ··· 12.133 11.258 10.649 ··· 9.449
2MASS Jc(mag) 12.703 12.155 11.516 12.803 11.297 13.163 12.328 12.785 11.844 10.610
2MASS Hc(mag) 12.571 11.853 11.476 12.282 10.716 13.058 12.205 11.544 11.277 10.452
2MASS Ksc(mag) 11.851 10.789 10.720 11.211 9.488 12.184 11.225 10.594 10.105 9.469
SPITZER F3.6μm (mJy) 11.18 ··· ··· ··· ··· ··· ··· ··· ··· ···
SPITZER F4.5μm (mJy) 12.33 ··· ··· ··· ··· ··· 62.26 ··· ··· ···
SPITZER F5.8μm (mJy) 11.59 ··· ··· ··· ··· ··· ··· ··· ··· ···
SPITZER F8.0μm (mJy) 63.54 ··· ··· ··· ··· ··· 610.73 ··· ··· ···
SPITZER F24μm (mJy) ··· 22.17 ··· ··· ··· ··· ··· ··· ··· ···
WISE F3.4μm (mJy) 10.12 43.84 40.25 37.46 278.43 9.97 42.91 92.34 138.26 229.31
WISE F4.6μm (mJy) 13.42 83.96 52.51 65.56 522.21 13.74 70.71 86.14 202.32 497.63
WISE F12μm (mJy) 446.53 2275.12 2859.34 1892.31 13056.11 385.33 2042.26 2311.37 6959.20 14935.59
WISE F22μm (mJy) 4680.31 21158.78 26295.30 10468.79 38714.54 3998.31 5294.27 21080.05 20374.48 27866.67
MSX F8.28μmd(Jy) 0.168 0.990 ··· 0.858 6.137 0.148 0.909 ··· 3.361 ···
MSX F 12.13 μmd(Jy) ··· 2.298 ··· 2.728 12.86: ··· ··· ··· 6.855 ···
MSX F 14.65 μmd(Jy) 1.216 6.542 ··· 3.453 17.26: 0.923 4.399 ··· 10.86 ···
MSX F 21.3 μmd(Jy) 4.628 18.79 ··· 8.853 33.14 3.428 3.77 ··· 20.08 ···
AKARI F9μme(Jy) 0.192 1.102 1.441 0.951 6.447 ··· 0.899 ··· 3.437 6.636
AKARI F18μme(Jy) 3.092 12.943 16.842 6.286 26.937 ··· 3.844 13.886 16.454 21.477
AKARI F65μme(Jy) ··· 18.480 14.692 ··· 13.535 5.714 ··· 11.311 4.837 7.792
AKARI F90μme(Jy) ··· 14.073 11.152 ··· 11.489: 5.292 ··· 9.759 3.650 6.359
AKARI F 140 μme(Jy) ··· 3.369: 4.286 ··· 3.659: ··· ··· ··· ··· 2.105
AKARI F 160 μme(Jy) ··· ··· 3.226: ··· 0.543: ··· ··· 1.401: ··· 2.493:
IRAS F12μmf(Jy) <2.43 2.23 3.54 2.37 13.58 <0.50 1.61 2.46 7.60 13.81g
IRAS F25μmf(Jy) 4.16 29.82 31.21 13.71 46.15 6.05 7.79 27.73 20.96 34.90g
IRAS F60μmf(Jy) 1.31 20.74 16.82 6.82 19.69 7.81 24.26 13.84 5.23 13.10g
IRAS F 100 μmf(Jy) <64.19 <15.21 7.17 <43.52 8.04 <34.55 26.69 5.29 <36.02 3.40g
Notes.
aThe colon represents uncertain detection. For some IRAS measurements, the fluxes measured from 12 μm and 100 μm are upper limits.
bFrom DENIS database.
cFrom 2MASS point source catalog.
dFrom MSX infrared point source catalog.
eFrom AKARI all sky survey point source catalog.
fFrom Tajitsu & Tamura (1998), the color- and diameter-corrected IRAS fluxes were given.
gFrom IRAS point source catalog.
References. (1) Tylenda et al. 1991; (2) Tylenda et al. 1989; (3) Shaw & Kaler 1985.
The abundances of He2+/H+and He/H of these PNs are taken
from Preite-Martinez et al. (1989), Cahn et al. (1992), Shaw
& Kaler (1985), and Zhang & Kwok (1991). After correcting
their UV spectrum and corresponding optical measurements by
the extinction values (which accounts for both circumstellar and
interstellar extinction), the emergent fluxes of these objects can
be then fitted by using the same expressions as Hsia et al. (2010).
The SEDs of these PNs corrected for extinction are shown in
Figure 14.
We can see that these SEDs clearly separate the contributions
from different components including photospheric continuum
of the central star (except Hen 2-447, M 1-59, and M 3-35),
the nebular continuum from ionized gas, and the dust thermal
emission. The observed dust emission is too broad to fitted
by a single blackbody and the observed spectrum is therefore
fitted by two modified blackbodies of different temperatures,
corresponding to a warm (Twd) and a cold (Tcd ) dust components.
Using the assumed distances (D) listed in Column 13 of Table 6,
we can derive a number of physical parameters such as minimum
luminosity of the central star (L∗), the total mass of dust
components (Md), and the mass of ionized gas (Mi). The dust
mass is derived from the infrared flux emitted by the dust
component, and the ionized mass is derived based on a fitting
of the bound–free and free–free continuum. The dust particles
are assumed to have an emissivity of Qλ=Q0(λ/λ0)−α, where
λ0=1μm, Q0=0.1, and αhas values between 1 and 1.3.
The kinematic ages of these PNs can also be estimated by
their apparent sizes of the longest lobes (assuming the lobes
lie close to the plane of the sky), distances, and expansion
velocities (Vexp). A summary of the derived properties of these
objects is given in Table 6. The names of these PNs are listed
in Column 1. The derived effective temperatures of central
17

The Astrophysical Journal, 787:25 (19pp), 2014 May 20 Hsia et al.
Tab l e 6
Physical Properties of the Sample Nebulae
Object T∗aT∗,ref TwdaTcd alog L∗aMdaMiaMd/MiClassbθcVexpdDAgee
(103K) (103K) (K) (K) (L)(M)(M)(
)(kms
−1) (kpc) (yr)
H1-54 43±5 41.3(1),39
(2) ··· 136 3.77 9.90 ×10−50.04 2.48 ×10−3O(5) 1.58 22 8.01(10) 2730
Hen 2-86 72.5 ±8.7 67(2) ,88
(2) 312 96 3.74 7.14 ×10−40.15 4.76 ×10−3M(6) 5.23 22 3.58(10) 4040
Hen 2-320 53 ±749
(2) 208 97 3.76 8.51 ×10−40.09 9.46 ×10−3M(7) 4.95 22 4.44(10) 4750
Hen 2-447 ··· ··· 298 103 ··· 6.04 ×10−40.10 6.04 ×10−3M(8) 2.40 22 6.30(10) 3260
IC 5117 88.5 ±7.5 76(3) ,90
(3) 335 117 3.84 4.43 ×10−40.16 2.77 ×10−3C(9) 1.78 21.5 3.83(10) 1510
M1-30 56±7.3 48(2),61
(2) 203 84 3.75 9.18 ×10−40.09 1.02 ×10−2··· 2.29 22 5.58(11) 2760
M1-59 ··· ··· 211 68 ··· 3.29 ×10−40.02 1.65 ×10−2··· 9.02 13f1.23(12) 4050
M1-61 69±6.2 67(2), 72.5(1) 217 96 3.87 7.98 ×10−40.58 1.38 ×10−3M(6) 2.52 22 4.73(10) 2570
M3-35 ··· ··· 323 129 ··· 1.30 ×10−40.33 3.93 ×10−4M(9) 1.43 30 4.44(10) 1370
NGC 6790 80.5 ±7.6 76(3) ,85
(4) 381 154 3.88 9.74 ×10−50.14 6.96 ×10−4M(8) 2.34 22 3.18(10) 1610
Notes.
aDerived from SED fitting.
bO: oxygen-rich dust, M: mixed-chemistry dust, C: carbon-rich dust.
cThe angular radius of the longest lobe of this object assuming all lobes have similar projected sizes on the sky.
dExpansion velocity measured from [N ii]λ6584 (Acker et al. 1992).
eDue to the inclination effect in a 2D projection model and global geometry of the expanding structures, the derived kinematic ages here are the upper limits.
fExpansion velocity measured from [O iii]λ5007 (Acker et al. 1992).
References. (1) Zhang & Kwok 1991; (2) Gleizes et al. 1989; (3) Kaler & Jacoby 1991; (4) Aller et al. 1996; (5) Casassus et al. 2001a; (6) Stanghellini et al. 2012;
(7) This study; (8) Casassus et al. 2001b; (9) Rinehart et al. 2002; (10) Zhang 1995; (11) Phillips 2005; (12) Acker et al. 1992.
stars (T∗) from SED fittings and those obtained from the
references for comparison (T∗,ref) are given in Columns 2 and
3. Columns 4–6 give the temperatures of warm and cool dust
components, and the derived luminosities of the central stars of
these PNs, respectively. The total mass of the warm and cold dust
components, the derived ionized mass, and the dust-to-ionized
gasmassratios(Md/Mi) are given in Columns 7–9. Column 10
gives their dust grain classes based on spectral information. The
angular radii of the longest lobes (θ) of these objects measured
from Hαimages are listed in Column 11. Columns 12–14 give
their expansion velocities, distances, and the derived kinematic
ages. Because of the uncertain extinction correction of each
object for the visible photometry of the central star, and the fact
that these photometric data lie in the Rayleigh–Jeans side of
the blackbody curve, the temperature of central star can not be
determined precisely. Even so, our estimates of T∗are consistent
with values previously reported in the literature (see Table 6).
The derived values of Md/Mifor our sample objects are gen-
erally within the range of dust-to-gas mass ratios (10−2–10−3)
typically found in PNs (Stasi´
nska & Szczerba 1999). We note
that these values should be considered as the maximum for the
actual dust-to-gas ratios since part of the gas in the nebulae
is likely to be in molecular (non-ionic) form. Considering the
uncertain determinations of projected sizes and measured ex-
pansion velocities of the PNs caused by the inclination effect in
a 2D projection model and global geometry of the expanding
structures, our derived kinematic ages used here are just rough,
order-of-magnitude estimates. We find that most of these PNs
are quite young, with a median age of 2740 yr. The expansion
velocities of multipolar PNs in the lobes are usually found to be
larger than that in the central region, thus the derived kinematic
ages here are likely to be the upper limits.
REFERENCES
Aaquist, O. B., & Kwok, S. 1990, A&AS, 84, 229
Aaquist, O. B., & Kwok, S. 1991, ApJ,378, 599
Acker, A., Ochsenbein, F., Stenholm, B., et al. 1992, Strasbourg-ESO Catalog
of Galactic Planetary Nebulae, Parts I and II (Garching: European Southern
Observatory)
Aller, L. H., & Czyzak, S. J. 1983, ApJS,51, 211
Aller, L. H., Hyung, S., & Feibelman, W. A. 1996, PASP,108, 488
Balick, B., & Frank, A. 2002, ARA&A,40, 439
Cahn, J. H., Kaler, J. B., & Stanghellini, L. 1992, A&AS, 94, 399
Casassus, S., Roche, P. F., Aitken, D. K., et al. 2001a, MNRAS,327, 744
Casassus, S., Roche, P. F., Aitken, D. K., et al. 2001b, MNRAS,320, 424
Clark, D. M., L´
opez, J. A., Steffen, W., et al. 2013, AJ,145, 57
Corradi, R. L. M., Livio, M., Balick, B., et al. 2001, ApJ,553, 211
Corradi, R. L. M., S´
anchez-Bl´
azquez, P., Mellema, G., et al. 2004, A&A,
417, 637
Corradi, R. L. M., & Schwarz, H. E. 1995, A&A, 293, 871
de Freitas Pacheco, J. A., Maciel, W. J., & Costa, R. D. D. 1992, A&A,
261, 579
Garc´
ıa-Segura, G. 1997, ApJL,489, L189
Garc´
ıa-Segura, G. 2010, A&A,520, L5
Garc´
ıa-Segura, G., L´
opez, J. A., Steffen, W., Meaburn, J., & Manchado, A.
2006, ApJL,646, L61
Girard, P., K¨
oppen, J., & Acker, A. 2007, A&A,463, 265
Gleizes, F., Acker, A., & Stenholm, B. 1989, A&A, 222, 237
G´
orny, S. K., Schwarz, H. E., Corradi, R. L. M., et al. 1999, A&AS, 136, 145
Guerrero, M. A., & Manchado, A. 1998, ApJ,508, 262
Guerrero, M. A., Villaver, E., Manchado, A., et al. 2000, ApJS,127, 125
Guill´
en,P.F.,V
´
azquez, R., Miranda, L. F., et al. 2013, MNRAS,432, 2676
Hsia, C.-H., Kwok, S., Zhang, Y., et al. 2010, ApJ,725, 173
Hsia, C.-H., & Zhang, Y. 2014, A&A,563, A63
Hua, C. T., Grundseth, B., & Maucherat, A. J. 1993, A&AS, 101, 541
Huarte-Espinosa, M., Frank, A., Balick, B., et al. 2012, MNRAS,424, 2055
Kaler, J. B. 1986, ApJ,308, 322
Kaler, J. B., & Jacoby, G. H. 1991, ApJ,372, 215
Kwok, S. 1990, MNRAS, 244, 179
Kwok, S., & Aaquist, O. B. 1993, PASP,105, 1456
Kwok, S., Chong, S.-N., Hsia, C.-H., et al. 2010, ApJ,708, 93
Kwok, S., & Hsia, C.-H. 2007, ApJ,660, 341
Kwok, S., & Su, K. Y. L. 2005, ApJ,635, 49
Lee, T.-H., & Kwok, S. 2005, ApJ,632, 340
L´
opez, J. A., Meaburn, J., Bryce, M., et al. 1998, ApJ,493, 803
L´
opez, J. A., V´
azquez, R., & Rodr´
ıguez, L. F. 1995, ApJL,455, L63
Manchado, A., Garc´
ıa-Hern´
andez, D. A., Villaver, E., & Guironnet de Massas,
J. 2011, in ASP Conf. Ser. 445, Why Galaxies Care About AGB Stars II, ed.
F. Kerschbaum, T. Lebzelter, & B. Wing (San Francisco, CA: ASP), 161
Manchado, A., Stanghellini, L., & Guerrero, M. A. 1996, ApJ,466, 95
Mampaso, A., Corradi, R. L. M., Viironen, K., et al. 2006, A&A,458, 203
Phillips, J. P. 2005, MNRAS,362, 847
Preite-Martinez, A., Acker, A., K¨
oppen, J., et al. 1989, A&AS, 81, 309
Reach, W. T., Megeath, S. T., Cohen, M., et al. 2005, PASP,117, 978
Rinehart, S. A., Houck, J. R., Smith, J. D., et al. 2002, MNRAS,336, 66
Roeser, S., Demleitner, M., & Schilback, E. 2010, AJ,139, 2440
Sahai, R. 2000, ApJL,537, L43
18

The Astrophysical Journal, 787:25 (19pp), 2014 May 20 Hsia et al.
Sahai, R., Morris, M., S´
anchez Contreras, C., & Claussen, M. 2007, AJ,
134, 2200
Sahai, R., Morris, M. R., & Villar, G. G. 2011, AJ,141, 134
Shaw, R. A., & Kaler, J. B. 1985, ApJ,295, 537
Stanghellini, L., Corradi, R. L. M., & Schwarz, H. E. 1993, A&A, 279, 521
Stanghellini, L., Garc´
ıa-Hern´
andez, D. A., Garc´
ıa-Lario, P., et al. 2012, ApJ,
753, 172
Stanghellini, L., & Kaler, J. B. 1989, ApJ,343, 811
Stasi´
nska, G., & Szczerba, R. 1999, A&A, 352, 297
Steffen, W., Koning, N., Esquivel, A., et al. 2013, MNRAS,436, 470
Steffen, W., Koning, N., Wenger, S., Morisset, C., & Magnor, M. 2011, IEEE
Trans. Vis. Comput. Graphics, 17, 454
Tajitsu, A., & Tamura, S. 1998, AJ,115, 1989
Terzian, Y., & Hajian, A. R. 2000, in ASP Conf. Ser. 199, Asymmetrical
Planetary Nebulae II: From Origins to Microstructures, ed. J. H. Kastner,
N. Soker, & S. Rappaport (San Francisco, CA: ASP), 33
Tylenda, R., Acker, A., Gleizes, F., et al. 1989, A&AS, 77, 39
Tylenda, R., Acker, A., Raytchev, B., et al. 1991, A&AS, 89, 77
Tylenda, R., Acker, A., Stenholm, B., et al. 1992, A&AS, 95, 337
Wright, E. L., Eisenhardt, P. R. M., Mainzer, A. K., et al. 2010, AJ,
140, 1868
Zhang, C. Y. 1995, ApJS,98, 659
Zhang, C. Y., & Kwok, S. 1991, A&A, 250, 179
Zhang, Y., & Kwok, S. 2009, ApJ,706, 252
19