Content uploaded by R. Gelderman
Author content
All content in this area was uploaded by R. Gelderman on Aug 11, 2014
Content may be subject to copyright.
arXiv:astro-ph/9903001v1 27 Feb 1999
Emission-Line Properties of the LMC Bubble N70
Brooke P. Skelton
Astronomy Department, University of Washington, Seattle, Washington 98195
e-mail: skelton@astro.washington.edu
William H. Waller
1,2
Raytheon STX Corporation, NASA Goddard Space Flight Center
Laboratory for Astronomy and Solar Physics, Co de 681, Greenbelt, Maryland 20771
Richard F. Gelderman
3
National Research Council & NASA Goddard Space Flight Center
Laboratory for Astronomy and Solar Physics, Co de 681, Greenbelt, Maryland 20771
Larry W. Brown and Bruce E. Woodgate
NASA Goddard Space Flight Center
Laboratory for Astronomy and Solar Physics, Co de 681, Greenbelt, Maryland 20771
Adeline Caulet
1
Astronomy Department, University of Illinois, Urbana, IL 61801
and
Robert A. Schommer
National Optical Astronomy Observatories, Cerro Tololo Inter-American Observatory
Casilla 603, L a Serena, Chile
1
Visiting Astronomer, Cerro Tololo Inter-American Observatory. CTIO is operated by
AURA, Inc. under contract to the National Science Foundation.
2
Current Address: Harvard Smithsonian Center for Astrophysics, 60 Ga r den Street,
Cambridge, Massachusetts 02138
3
Current Address: Western Kentucky University, Department of Physics and Astronomy,
– 2 –
Received ; accepted
Bowling Green, Kentucky 42101-357 6
– 3 –
ABSTRACT
We present a sp ectrophotometric imaging study of the emission bubble N70
(DEM 301) in the Large Magellanic Cloud. N70 is approximately 100 pc in size
with a nearly circular shell-like morphology. The nebular emission is powered
by an uncertain combination of EUV photons, intense winds, and supernova
shock waves from the central population of high-mass stars (the OB association
LH 114). We have obtained narrow-band images (FWHM ∼6
˚
A) of N70 in
the light of Hαλ6563, [N II]λ6584, [S II]λλ6717,6731, and [O III]λ5007, along
with the corresponding red and green continua. The resulting line fluxes and
flux ratios are used to derive ionization rates, nebular densities, volume filling
fractions, and excit ation indices. The photoionizing luminosity inferred from the
embedded stellar population is more than adequate to a ccount for the observed
hydrogen ionization rate.
We compare the emission-line photometry with that derived from similar
imaging of the Orion nebula and with data collected from the literature o n
other emission-line regions in the LMC. Compared to the Orion nebula, N70
shows much higher [S II]/Hα intensity ratios which increase smoothly with
radius — from < 0.3 near the center to > 1.0 towards the outer filamentary
shell. The measured intensity ratios in N70 more closely match the range of
excitation spanned by giant and supergiant H II shells and by some of the
sup ernova remnants observed in the LMC. The contending ionization a nd
excitation processes in the interior and outer shell of N70 are evaluated in terms
of the available data. EUV photons probably contribute most of the inner
nebula’s ionization, whereas a combination of photoionization plus collisional
ionization and excitation of sulfur atoms by low-velocity shocks seems to best
fit the emission-line luminosities and intensity ratios observed in the outer shell.
– 4 –
Considerations of the radiative and mechanical energetics tha t are involved may
indicate the need for one or two supernova explosions having occurred during
the last ∼ Myr.
– 5 –
1. Introduction
N70 (DEM 301; 5
h
43
m
16
s
−67
◦
50
′
53
′′
(J2000)) in the Large Magellanic Cloud is
an especially prominent bubble of line-emitting gas which appears to be powered by a
population of hot massive stars in its interior. Its nearly circular symmetry ha s prompted
several studies aimed a t determining which mechanisms govern the dynamical and radiative
evolution of this seemingly isolated and “simple” starburst. An early spectrophotometric
study by Dopita et al. (198 1) led to the conclusion that N70 represents a giant cavity that
has been blown into a massive H I cloud by the embedded cluster of hot, windy stars.
Ultraviolet radiation from the O stars was regarded as the dominant source of ionization in
the shell.
Lasker (1977) suggested that additional excitation by stellar winds is necessary to
explain the high [S II]/Hα line ratios that he observed. The detection of lopsided X-ray
emission from the southwestern quadrant of N70 indicated to Chu and Mac Low (1990)
that a supernova explosion occurred inside a wind-blown bubble. The energy from such a
sup ernova would have provided additional heating of the gas inside the bubble, accounting
for the higher-than-expected X-ray emission, and could have shocked the pre-existing shell
of warm gas in the visible shell, raising its [S II]/Hα line rat io .
More recently, Oey (1996a, 1996b) has attempted to explain the dynamics of N70
with standard pressure-driven bubble models; in doing so, she has examined the stellar
population of the central OB association, LH 114 (Lucke and Hodge 1970) . The current
population of stars in LH 114 has an initial mass function (IMF) slope consistent with a
Salpeter IMF slope. If only the stars interior to the edge of the nebula are considered, a
slightly flatt er slope of Γ ∼ −1.0 is determined (Γ
Sal
= −1.35). The mean age of these stars
is approximately 5 million years; three stars ar e approximately 40 M
⊙
and none are more
massive (Oey 1996a).
– 6 –
Based on this inferred age, O ey found that N70 is a “high-velocity” superbubble which
is dynamically inconsistent with the standard model—the expansion velocities observed in
N70 being too high for the observed radius and age (Oey 1996b). This, of course, assumes
that the motions observed in N70 are in fact due t o expansion, which Dopita et al. (1981)
contests. Kinematics of N70 have been observed by Dopita et al. (1981), Rosado et al.
(1981), Blades et al. (1980), Georgelin et al. (1983), and Lasker (1977), yielding velocity
dispersions of order 40 km s
−1
but differing conclusions regarding t he overall velocity field.
In this paper, we present narrow-band images of t he ionized bubble in the light of
Hα, [S II], [N II], and [O III]. These images were taken at the CTIO 1.5-m telescope with
the Rutgers Fabry Perot imaging system and Goddard Fabry Perot etalons whose tunable
narrow-band capability enables t he clean separation of the Hαλ6563 and [N II]λ6584 lines as
well as the doublet lines of [S II]λ6717 and [S II]λ6731. The resulting line flux and flux ra tio
measurements are used to derive nebular ionization rates, densities, volume filling fr actio ns,
and indices of excitation. Analysis of these spectrophotometr ic indices and comparison
with similar data on the Orion nebula and other line-emitting sources in the LMC indicates
that a mix of radiative and collisional (shock) processes is responsible for ionizing and
exciting the gas in this shell. Empirically, the N70 flux ratios span a range of values similar
to that of other LMC giant and supergiant shells except for a few regions with unusually
high [S II] t hat are more characteristic of supernova remnants. A comparison wit h shock
models (e.g. Shull and McKee 1979, Hartigan et al. 1994 ) indicates that the high [S II]/Hα
line ratios and the low, Orion-like [N II]/Hα line ratios in N70 are difficult to reconcile with
a hybrid mode of nebular photoionization and shock-excitatio n without invoking special
circumstances.
The Fabry-Perot observations of N70 and the data reduction are described in Section
2. Images of the Hα, [N II], [S II], and [O III] line emission and of their intensity ratios
– 7 –
are presented and discussed in Section 3. Comparisons with the Orion Nebula, other line
emitting regions in the LMC, and nebular models are made in Section 4, where constraints
on the total radiative and mechanical energetics are considered. A summary of our results
is presented in Section 5.
2. Fabry-Perot Observations and Reductions
Compared to the previous spectrophotometric studies of N70 (Dopita et al. 1981,
Lasker 1981), the observations presented herein represent an improvement in sensitivity
and multi-spectral coverage at Hα, [N II], [S II], and [O III], including resolution of the
[S II] doublet, thereby enabling a more detailed and comprehensive analysis of the nebular
ionization and excitation. While previous studies have presented images of N70 in the
light of Hα, none have imaged the entire nebula in [N II], [S II], or [O III] using CCD
technology or the very narrow bandwidths allowed by Fabry-Perot cameras. While the
spectroscopy employed by these studies allows one to obtain kinematic information, the
present observations are the first to afford a comprehensive look at how the emission line
fluxes and ratios change on small scales (< 1 pc). Imaging of the Orion nebula (M42, NGC
1976) was obtained to provide a check on the spectrophotometric reductions and subsequent
interpretations.
2.1. Observations
N70 and the Orion nebula wer e imaged with the CTIO 1.5 -m telescope and Rutgers
Fabry-Perot camera between 28 October a nd 1 November 1993. The imaging system
consisted of a piezoelectr ically controlled scanning etalon and a servocontroller, b oth made
by Queensgate Instruments, Ltd. (Atherton et al. 198 1), an interference filter with ≈ 100
˚
A
– 8 –
bandwidth to block out mult iple int erference orders, re-imaging optics, and a CCD detector.
Instead of the high-spectral resolution etalons (FWHM ≈ 1–2
˚
A) that are standard with
the Rutg ers Fabry-Perot camera, we installed lower-resolution etalons (FWHM ≈ 5–30
˚
A)
normally resident in the Goddard Astronomical Fabry-Perot Imaging Camera (GAFPIC,
Brown et al. 199 4). These etalons, also made by Queensgate, are optimized for imaging
nebular emission in a variety of lines rather than mapping the nebular kinematics in just
one line. The calibration of etalon spacing as a function of wavelength was obtained by
scanning the line emission from sp ectral lamps. This was done several times each night,
thereby attaining tuning accuracies of less than ±1
˚
A.
Initial observations were made with the Goddard “blue” etalon a nd CTIO’s Tek#4
512×512 CCD chip, where the image scale was 1.
′′
1 pixel
−1
and the total field of view was
circular with a diameter of 7.
′
2. The “blue” etalon was tuned to a bandwidth of 7±1
˚
A
(FWHM) for imaging the [O III] emission a nd green continuum. During the second half of
the run, observations were made with the Goddard “red” etalon and the Tek 1024 #1 chip;
the image scale was 0.
′′
96 pixel
−1
and the resulting field of view was circular with a diameter
just over 7
′
. The Goddard “red” etalon was tuned to a bandwidth of 6±1
˚
A (FWHM) for
the Hα, [N II], and [S II] observations. The average ra dial velocity of N70 with respect
to the local standard of rest is ∼ 290 km s
−1
(Dopita et al. 1981), resulting in a redshift
of the [O III]λ5006.8 line to 501 1.6
˚
A, Hαλ6562.8 to 6569.1
˚
A, [N II]λ6583.6 to 6590.0
˚
A,
[S II]λ6717.0 to 6723.5
˚
A, and [S II]λ6731.3 to 6737.8
˚
A. Atmospheric seeing averaged 2.
′′
2
(FWHM), enabling spatial resolution on the order of 0.6 pc, assuming a distance of 55 kpc
to the LMC (m − M = 18.7; Feast and Catchpole 1997). Sky transparency varied from
clear to heavy cirrus. A summary of the imaging is presented in Table 1 .
EDITOR: PLACE TABLE 1 HERE.
– 9 –
2.2. Reductions
Overscan fitting and subtraction, residual bias averaging and subtraction, domeflat
averaging and division, and skyflat smoothing and division were carried o ut using the
CCDRED routines within IRAF
4
. The standard and target frames were air-mass corrected
using the mean extinction coefficients at Cerro Tololo (Stone and Baldwin 1983). Cosmic
rays were removed with the cosmicrays task, which can be used to identify cosmic rays in
fields where only one image has been obtained.
The Hα, [S II], [N II], and red continuum images were calibrated with
spectrophoto metric standards taken close in time to the N70 observations (LTT
9239 and LTT 2415; Hamuy et al. 1992, Stone and Baldwin 1983). The standards were
chosen because they have weak Balmer absorption lines, so their flux (F
λ
) is almost constant
in t he wavelength region of interest. We interpolated between the published fluxes (in
50
˚
A bins) to arrive at the predicted flux through our 6
˚
A bandpasses. The standard star
calibration error is estimated at
∼
<15%.
The images of the Orion nebula were taken during cirrus and intermittent clouds, so
calibration with standard stars was not possible. Therefore, we bootstrapp ed our Orion
calibrations to the wide-field images at Hα, [N II], and [S II] (FWHM = 15
˚
A, 15
˚
A, 36
˚
A,
respectively) that wer e obtained by Walter et al. (1992) under photometric skies; this
procedur e should yield photometric accuracies of approximately 10%.
Because the [O III] and green continuum images of N70 were taken in cirrus, they
could not be calibrated with observations of standard stars. Instead, an approximate
4
IRAF is distributed by the National Optica l Astronomy Observatories, which is operated
by the Association o f Universities for Research in Astronomy, Inc., under contract to the
National Science Foundation.
– 10 –
calibration was constructed based on the observed spectral types of the ionizing stars in
N70. The spectral energy distributions of six stars classified between O5 and B0 (Oey
1996a) were estimated based on the Silva and Cornell (1 992) library of stellar optical
spectra. The closest spectral type in the catalog was chosen for each star, a nd the flux at
the green continuum (5031
˚
A) was predicted based on the calibrated flux at 6751
˚
A. The
green continuum imag e was calibrated to yield the correct stellar fluxes in the presence of
foreground reddening (E(B − V )=0.06), and the [O III]-band image was matched to it.
The scatter in measured 6751
˚
A to 5031
˚
A flux is about 20%, so we can determine the green
continuum calibration only to this accuracy at best. Regardless of the uncer t ainty in the
absolute calbration, this [O III] image of N70 is the first to be published, revealing unique
morphological characteristics that have been previously unrecognized.
Once the individual images were calibrated and sky background subtracted, the closest
continuum image was subtracted from the on-line image to f orm an emission image. These
emission maps are presented in the next section. One of the difficulties with Fa bry-Perot
imaging is that the central wavelength of the bandpass changes with radial distance from
the center of the field of view. The stellar spectra are essentially flat over the wavelength
range of the individual observations, so t he continuum images are unaffected. However, t he
emission-line images are subject to an increasing blueshift with field angle, amounting to
∼4
˚
A at the edge of the field (Atherton et al. 1981; Caulet et al. 1992). The net effect of
this shifting of peak transmissivity is to decrease the monochromatic sensitivity at the edge
of the field.
To correct for this effect, post-observation “spectral flat fields” were made using images
of the Orion nebula that were taken with fixed-wavelength interference filters (Walter et
al. 1992). The spectral flats were created by dividing Walter et al.’s Orion images by the
corresponding Fabry-Perot images and smoothing the results. This wa s done with the
– 11 –
continuum-subtracted images in Hα, [N II], and [S II](λ6717 + λ6731). Because Walter et
al.’s [S II] image includes both [S II] doublet lines, the individual [S II] images could not be
corrected separately; however, the individual corrections should be very similar, so the line
ratio λ6717/λ6731 is basically unaffected. After correction by these spectral flats, the total
net line fluxes of N70 increased by f actors of 1.25 in Hα, 1.05 in [N II], and 2.32 in [S II].
The large increase in [S II] flux is due to the concentration of [S II] emission at the edges
of N70, near the edg e of the field. Line ratios were not as affected, changing by less than
10% at the edge of the field for [N II]/Hα and up to 50% for [S II]/Hα. No corrections were
made to the [O III] images.
Reddening corrections were determined based on Oey’s spectroscopy and photometry
of stars in LH 114 (the association inside N70); her median E(B − V )=0.06 (Oey 1996a)
was used with reddening coefficients from Mathis (1990) (R
V
= 3.1) to determine the
interstellar reddening and extinction (0.2 magnitudes in V ). When comparing line ratios, it
is worth noting that Dopita et al. (1981) used A
V
= 0.6 magnitudes, which over-corrects in
the blue.
Intensity ratio maps, presented in t he next section, were created by dividing the
dereddened emission images. The line rat io s that are tabulated and discussed are given for
specific regions in the nebulae, where the aperture photometry was done on the emissio n
line images and then the summed fluxes were ratioed. Both observed and dereddened Hα
fluxes are presented; the low level of reddening derived for N70 means that dereddened
[N II]/Hα and [S II]/Hα ratios remain essentially unchanged from the observed values.
– 12 –
3. Emission-Line Maps and Spectrophotometry
Observations of N70 over the past twenty years have revealed an ionized shell
with remarkable spectrophotometric properties. The current Fabry-Pero t data provide
observations of the main circular structure of N70 at Hα, [N II], [S II], [O III], and their
respective line ratios. The FWHM of stellar images is 2.3 pixels, or 2.
′′
2. At an LMC
distance of 55 kpc, this allows spatial resolution on the order of 0.6 pc.
3.1. Maps
Figure 1a is the final dereddened Hα image of N70. The locations of the two O stars
D301-10 05 and D301NW-8 (notation of Oey (1996a)) are marked in it and all subsequent
emission images and ratio maps. These stars are marked with the larger crosses in Figure
1b, the 6536
˚
A continuum. The labeled stars in this figure are those which Oey (19 96a)
classified; their spectral types are noted. Oey chose stars with a reddening-free index
Q = (U − B) − (B − V ) ≤ −0.70 and V ≤ 16.0; her sample should be reasonably complete
in O to early B stars. The stellar content will be discussed more fully below. Note the lack
of nebular emission in the continuum image, indicating that N70 is truly an emission-line
nebula with negligible scattered continuum light.
EDITOR: PLACE FIGURE 1 HERE.
The ot her t hree N70 emission-line images [N II]λ6584, [S II]λ6717 + λ 6731, and
[O III]λ5007 are presented in Figures 2a, 2b, and 2c, respectively. The [S II] and [O III]
images are at the original resolution, but [N II] was smoothed with a 2 pixel (1.
′′
92) Gaussian
to increase the signal-to-noise in the displayed image.
EDITOR: PLACE FIGURE 2 HERE.
– 13 –
The Hα, [S II], and [O III] images wer e combined to form the color-coded image shown
in Figure 3. Blue is [O III], green is Hα, and red is [S II]. The colors were scaled so that
features of all three emission lines could be seen. This picture, along with F ig ures 1 and 2,
shows that despite the overall circular symmetry of N70, significant and unique substructure
exists at each emission line.
EDITOR: PLACE FIGURE 3 HERE.
Most notably, strong Hα and [O III] emission is evident interior to the outer shell,
whereas the [S II] image shows no such central concentration of emission. The interior
emission is closely associated with the hottest O-type stars (labeled in Figure 1b), thus
indicating a stellar power source for the nebula sufficient to ionize oxygen twice (E > 35.1
eV). For reasons more fully explored in Section 4.3, the interior Hα and [O III] emission
is most likely due to photoionization from the hot stars and subsequent excitation of the
[O III] by nebular electrons.
Away from the centrally concentrated hot stars, the nebular emission shows filamentary
substructure in what appears to be a weblike morphology. Lozinskaya (1992) not es that this
sort of emission structure indicates shock processes at wo rk in the presence of irreg ularities
in the ambient medium. Many of the well-observed “mature” supernova remnants—such as
the Veil Nebula in Cygnus, S147 in Taurus, and the Vela and Gum nebulae (Lozinskaya
1992)—are characterized by similar filamentary structure in the same emission lines.
Figure 4 shows the central portion of the Orion Nebula in Hα, [N II], and [S II] emission.
The Trapezium stars (θ
1
Ori) as well as θ
2
Ori A and B were saturated in the long [S II]
exposure and so were not effectively subtracted out of the resulting emission-line images
(Figure 4c). Only θ
1
Ori C and θ
2
Ori A were saturated in the [N II] image, and none were
saturated in the short Hα exposure (Figure 4a), probably because of thicker cirrus during
– 14 –
that exposure. To bootstrap calibrations, all comparisons between our images and those of
Walter et al. (1992) were done in regions away from these bright stars. Ghost images of
the Trapezium stars are barely discernible in the southwest part of the emission images,
and of θ
2
A in the northwest. These are especially apparent on the [S II] image. (Only
one of the stars in N70, near the edge in the north-northwest, was bright enough to cause
a noticeable ghost image.) The Orion images provide important checks on our reduction
methods as well as spectrophotometric benchmarks for diagnosing the line emission in N70
(discussed more fully in Section 4.1). Pogge et al. (1992) have presented spectrophotometric
Fabry-Perot observations of the Orion Nebula including ratio maps very similar to those
that were constructed from our data; our data agree well with these published maps and
ratios and hence are not presented herein.
EDITOR: PLACE FIGURE 4 HERE.
Unlike the situation wit h Orion, intensity ratio maps o f N70 have not been previously
available. Figure 5 presents three intensity r atio maps useful for our analysis. Figure 5a is
[N II]/Hα, Figure 5b is [S II](λ6717+λ6731 ) /Hα, and Figure 5c is [O III]/Hα. The [S II]/Hα
map is not smoothed in order to make the change in [S II]/Hα across t he individual
filaments clearer. [N II]/Hα and [O III]/Hα do not show this fine structure, so to increase
signal-to-noise, the maps have been Gaussian smoothed (σ = 2 pixels).
EDITOR: PLACE FIGURE 5 HERE.
3.2. Spectrophotometric Results
In addition t o the intensity ratio maps, emission-line flux ratios were determined for
sections of the nebulae using polygonal aperture photometry. This method averages over
– 15 –
large areas of emission, but allows larger signal-to-noise measurements and comparisons.
Figures 6a and b label the polygons used in N70 and Orion, respectively. The flux in each
of these regions was summed in the individual emission-line images and then ratioed; the
Hα fluxes and line ratios with respect to [N II], [S II], and [O III] are presented in Table
2 for N70, and Hα fluxes and line ratios with respect to [N II] and [S II] are presented in
Table 3 for Orion.
The uncert ainties in the ratios due to the determination of the backgro und is estimated
to be less than 30% for [N II]/Hα, less than 15% for [S II]/Hα near the bright rim of N70,
and less than 30% for [S II]/Hα in the central region where the [S II] flux is the lowest.
The “spectral flat” correction changed the measured [N II]/Hα ratios by less than 10%
regardless of their location in the field of view. The [S II]/Hα ratios were decreased by ∼
5% at the center of the field by t he correction; the correction to the [S II]/Hα ratio t hen
increases radially up to 100% at the N70 rim. The effects are most severe for the northern
rim. For example, the [S II]/Hα ratio of regions 3, 4, and 5 are increased by 40% by the
“spectral flat” correction and region 3 by almost 100% while region 6 is increase by only
25%. However, the uncertainties in the measured ratios are by no means this large; the
errors are dominated by uncertainty in the value of the background.
EDITOR: PLACE FIGURE 6 HERE.
EDITOR: PLACE TABLE 2 HERE.
EDITOR: PLACE TABLE 3 HERE.
The emission measure for each o f the N70 apertures was calculated from the Hα surface
brightness, I
Hα
. The emission measure is defined as EM =
R
n
2
e
dl, where l is the column
– 16 –
depth of the emitting material and n
e
is its electron density. For gas with singly ionized
hydrogen and helium, n
e
≈ 1.1n
H
. The number of protons n
H
can be calculated from the
Hα flux. Assuming T
e
= 10
4
K, EM (pc cm
−6
) = 4.4× 10
17
× I
Hα
(erg cm
−2
sec
−1
arcsec
−2
).
With the definition of the emission measure and a simple spherical shell model, analytic
relationships between emission measure and root mean square (rms) electron density can
be derived. The simplest are for a line of sight through the cent er of the shell and for a
line of sight through the edge of the shell. The first of these relationships, using emission
measures from the center of the optical shell, is
n
2
rms
=
EM
center
2∆R
s
where ∆R
s
is the shell thickness. The second relationship, using emission measures fro m
the bright rim of the shell, is
n
2
rms
=
EM
rim
2R
s
"
2
∆R
s
R
s
−
∆R
s
R
s
2
#
−1/2
where ∆R
s
is the shell thickness and R
s
is the radius of the shell, measured o ut to the edge
of the optical emission.
From our N70 observations, we determine a shell radius of approximately 3.
′
2, or 51
pc, and a shell thickness of 7.
′′
5, or 2 pc. The shell thicknesses were determined from visual
examination of radial plots of the Hα emission, where the thickness was set to the full-width
quarter-max of the radial profile of emission a cross the filaments. Using (dereddened)
emission measures near the center of N70, but away from the central two knots, yields
n
rms
∼ 7.5 cm
−3
; emission measures near the rim results in n
rms
∼ 4–6 cm
−3
. The volume
filling f actor f = n
2
rms
/n
2
e
, where n
e
is the density in the “clumps” of gas in the shell and
can be determined from measurements of [S II] lines as discussed later in this section.
– 17 –
The Hα flux measurements can also be used to deduce the Lyman continuum necessary
to ionize the gas in N70. Assuming the Case B hydrogen recombination coefficient,
N
Ly c
=
L
Hα
hν
Hα
α
B
α
Hα
= 7.3 × 10
11
× L
Hα
for T = 10
4
K. The dereddened Hα luminosity of N70 is 4.3 × 1 0
37
ergs sec
−1
, so the number
of ionizing photons needed is 3.1 × 10
49
s
−1
.
The luminosity of Lyman continuum photons produced by the stars in the central cluster
can be estimated based on Oey’s (1996a) spectral classificatio ns. The spectroscopically
identified stars marked on Figure 1b include one each of O3If, O5III, O7V, O8III, O9 V,
and two O9.5V stars as well as several B stars. Using log T
eff
and M
bol
from Oey’s Table
4, radii of the stars can be calculated. (Because we choose to use a distance modulus of
18.7 instead of 18.4 a s Oey did, we subtracted 0.3 from the M
bol
in t he table.) Masses of
these stars are estimated from her H-R diagram in order to approximately determine lo g g.
Using T
eff
, log g, and assuming log
Z
Z
⊙
= −0.3, the Kurucz (1992) model atmospheres can
be used to estimate the flux of ionizing photons (cm
−2
s
−1
) from the stars. The Lyman
continuum luminosity for each star is then calculated using the radius determined above.
The total ionizing luminosity from the cluster is 7.0 × 10
49
photons s
−1
, twice as much as
is necessary to ionize the interstellar hydrogen in N70. The ultimate fates of any EUV
photons beyond what is necessary to ionize the emission nebula can include absorption by
dust, absorption by other atomic species, and escape from the nebula. As expected, the
late O and early B stars contribute very little to the ionizing radiation; the O3If and O5III
stars alone contribute over 60% of the tota l Lyman continuum. Varying the choice o f stellar
metallicity between log
Z
Z
⊙
= −1.0 and 0.0 changes the ionizing flux by less than 5%.
Another way to estimate the Lyman continuum luminosity from the cluster is with
the compilations of Vacca et al. (1996). Stellar parameters such as radius, mass, absolute
– 18 –
magnitude, radius, and Lyman continuum flux (both photons cm
−2
s
−1
and total s
−1
) are
tabulated by spectral type and luminosity class. Based on the spectral classification of the
N70 stars, the ionizing luminosity determined from Vacca et al. (1 996) is 2.2 × 10
50
photons
s
−1
, seven times higher than the hydrogen ionization rate in N70. These luminosities are
for solar-metallicity stars, but as determined above, the metallicity only slightly affects the
amount of Lyman continuum radiated by the stars.
The factor of three difference from the previous determination is because Vacca et
al. calculate radii almost twice as large as we found. The root of this difference is in the
bolometric magnitudes assumed for each spectr al type. Oey (1996b) determined bolometric
corrections in the manner described by Massey et al. (1995) using broadband colors. Vacca
et al. (1 996) discuss their calibration of M
V
with spectral type; in general they determine
brighter absolute mag nitudes. For the two hottest stars, the differences are 0.9 and 1.9
magnitudes, which translates into more luminous and thus larger stars for the same effective
temperature. The discrepa ncy between the two determinations of bo lometric magnitude
and thus ionizing luminosity shows that this can be a tricky business, and should illustrate
the large uncertainties in the calculated ionizing luminosity. However, our predictions of
ionizing luminosity are quite consistent with that recently tabulated by Oey and Kennicutt
(1997).
As discussed at length in Section 4, the line ratios in Table 2 and Figure 5 are
discordant with a straightforward photoionization model. The [N II]/Hα ratios in Table 2
are similar to those in the central regions of Orion, but the [S II]/Hα ratios are much higher
than expected for a photoionized nebula.
In addition, there is a definite increasing trend in the [S II]/Hα intensity ratio across
the individual filaments on the face of N70 in addition to the overall increasing [S II]/Hα
intensity ratio with increasing radius from the center of the nebula. This can be seen in
– 19 –
Figures 2b and 5b. Figure 7 displays cuts across three filaments in N70; the [S II]/Hα ratio
rises with distance from the center of N70. A fourth plot shows a cut across the Orion
Bar; the [S II]/Hα ratio is mult iplied by a factor of ten in order to see it on this scale. As
discussed below, the Ba r is an ionization front in the Orion nebula seen edge on. The cut
across the bar has similar shape as cuts across N70 filaments, but the [S II]/Hα ratios are
much smaller. Also the physical scale is much smaller in Orion; 16 pixels total only 0.03 pc
while 16 pixels totals 4.1 pc linear distance in N70.
EDITOR: PLACE FIGURE 7 HERE.
The discussion of the reduction of the N70 [S II] images focussed on the sum of the
individual [S II]λ6717 and [S II]λ6731 images. Although the “spectral flat fields” could
not be individually derived for these two line images, we can assume that the radially
decreasing “gain” in each of these images is similar and thus measure the density-dependent
[S II]λ6717/λ6731 line ratio (Osterbrock 1989). As Table 2 shows, these ratios range from
1.4 to 1.9; for T = 10
4
K, the maximum λ6717/λ6731 line ratio in models is about 1.4 at
the low density limit. Therefore, the density of gas in N70 is low, probably < 100 cm
−3
, in
all parts of the optical nebula. Because we ca n not constrain n
e
any further, we constrain
the filling factor in the shell only to be > 0 .002–0.006 using the rms densities determined
from the emission measure of the gas. If n
e
were only 10 cm
−3
, the filling factor in the shell
would be ∼ 0.2–0.6.
4. Comparison with Orion, LMC Emission-Line Regions, and Nebular Models
– 20 –
4.1. Orion
When comparing N70 to the Orion Nebula, it is important to remember how much
more detail we can see in nearby objects—Orion is ∼ 450 pc away (Goudis 1982), 120
times clo ser than N70 . At the distance of the LMC, the part of the Orion Nebula seen in
our images would be contained in 3.5 pixels! In addition, our coarse resolution (pixel scale
0.
′′
96/pixel and seeing ∼2.
′′
2) limits the amount of fine structure we would be able to see in
our N70 images, especially when compared with Orion and LMC emission regions which
have been observed with WFPC2.
As previously mentioned, Pogge et al. (1992) present two-dimensional images of the
central region of Orion. The complexity of these images is apparent; the central bright
regions are certainly ionized by the Trapezium stars, but there are a lso Herbig-Haro
objects, the “Dark Bay” (very high absorption), and the “Bar” . Wen and O’Dell (1995)
have constructed a three-dimensional model of this region of the Orion Nebula. In their
model, Orion is a “blister” on the edge of the g iant molecular cloud OMC-1 rather than a
classical Str¨omgren sphere. Wen and O’Dell (1995) find that most of the radiation from
the Orion Nebula arises from a relatively thin surface layer of ionized material, and have
used that information to model the geometry near the Trapezium (θ
1
Ori). They show
that the emission enhancement seen just to the west of the Trapezium is due to the “hilly”
structure of the ionized layer; the bright emission comes from an area which is closer to the
ionizing stars than average. The Dark Bay is an area where the “lid”, or the ionization front
between an Earth observer and the Trapezium, is thicker than average and has increased
absorption by neutral gas and dust. The Ba r , on the other hand, is the ionization front
seen edge on. This is consistent with the thinness of the Bar in [N II] and [S II] compared
to Hα (see Figure 4) and with the enhancements in the [N II]/Hα and [S II]/Hα that would
arise in such a transitional layer where sulfur and nitrogen have yet to reach their highest
– 21 –
ionization stages (c.f. Petuchowski and Bennett 1995). The [O III] image of Pogge et a l.
(1992) shows that the higher-ionization material is mainly on the northwest side of the Ba r ,
closer to the ionizing stars. The [N II]- and [S II]-bright objects which are to the southeast
of the Bar near θ
2
Ori A are the shock-excited Herbig-Haro objects M42-HH3 (closer to the
Bar) and M42-HH4 (a little further to the southeast).
Figure 8 is a plot of the emission line ratios for Orion and N70 from Tables 2 and 3.
The differences between the two emission regions are obvious: the Orion Nebula, even in
the shock-excited Herbig-Haro objects, has a moderate range of [N II]/Hα and very low
[S II]/Hα throughout, while N70 has consistently lower [N II]/ Hα by ∼ 50% but a wide
range of [S II]/Hα ratios. As seen in Table 3 and Figure 6b, the [N II]/Hα ratios across the
face of Orion are lower than those in the Bar or in the Herbig-Haro objects. The regions
of N70 with low [S II]/Hα have [N II]/Hα most similar to that seen in Orion; Table 2 and
Figure 6a show that these regions are close to the ionizing stars and hence are most likely
photoionized like the centr al regions of Orion. This is corroborated by Figure 2c, which
shows that [O III], a higher-ionization state, is quite elevated nea r the central stars that
power N70. There is disagreement over whether or not Orion has lower metallicity than
solar (Walter et al. 1992), but in either case the average LMC nitrogen abundance of 0.4
solar (Russell and Dopit a 1992) is significantly less than that of Orion, which can explain
the lower [N II]/Hα ratios seen in N70.
EDITOR: PLACE FIGURE 8 HERE.
The enhanced [S II]/Hα emission-line r atios across Orion’s Bar and the outer filaments
in N70 lead one to wonder whether the filament s in N70 are actually just ionization fronts
like the Or io n Bar. However, as mentioned above, the physical scale and relative amount of
[S II] emission is very different in N70 than in Orion. An ionization front in N70 would occur
– 22 –
over physical lengths too small to be distinguished due to the effects of seeing. Another
difference is apparent when comparing Figures 1a and 2b with Figures 4a and 4c: while the
width of the filaments in N70 is approximately the same whether measured in Hα or [S II],
the same is not true for the Orion Bar. The Bar’s Hα emission extends over a much larger
area than its [S II] emission, consistent with the [S II] tracing the transitional ionization
front. The similar Hα and [S II] widths in the filaments of N70 indicate that the elevated
[S II]/Hα ratios ar e not due to multiple ionization fronts within the nebula. Instead, the
displacement in Hα and [S II]/Hα peaks observed in N70 is consistent with an outward
propagating shock front that is “back illuminated” by the embedded cluster of OB stars (see
Figures 5b and 7). The back illumination would be responsible for the inwar d-facing Hα
peak, while the shock front would explain the enhanced [S II] emission.
The special ionization and excitation mechanisms r equired for the elevation of the
[S II]/Hα intensity ratios can be constra ined by comparison with other emission nebulae
and with theoretical models.
4.2. LMC Emission-Line Regions
When comparing N70 with objects in the Large Magellanic Cloud, differences in
metallicities can be assumed to be small. Observed emission-line ratios of various obj ects in
the LMC were collected from the literature and plotted in F igure 9. The supernova remnants
are from Danziger and Leibowitz (198 5) with the exception of two from Westerlund and
Mathewson (1966). The H II region line ratios are those observed by Wilcots (1992).
Hunter (1 994) conducted a survey of ionized shells and supershells in the LMC; her “giant
shells” have radii 50 to 300 pc, and her “supergiant shells” have radii greater than 3 00 pc.
The data plotted are for a variety of positions within three giant shells and three supergiant
shells rather than the average line ratio across these shells. The N70 ratios in Table 2 are
– 23 –
also plotted. As shown in Figur e 9, the [N II]/Hα and [S II]/Hα flux ratios found in N70
span a range of values similar to that of giant and supergiant shells. However, a significant
population of regions show even higher [S II]/Hα flux ratios that are more characteristic of
sup ernova remnants (SNRs).
EDITOR: PLACE FIGURE 9 HERE.
Unlike the comparison with the Orion Nebula, the line ratios in the central regions of
N70 are very similar to those of H II regions in the LMC. The H II regions in Wilcots’ (1992)
sample are “classical” H II regions with “interesting” morphologies. Their Hα luminosities
are in the range of 10
36
to 10
37
erg s
−1
, a bit lower than N70. However, the Hα luminosity
of the two central knots (1 and 2 in Figure 6a) is 5.3 × 10
36
erg s
−1
, similar to Wilcots’ H II
regions. The line ratios of these regions of N70 are also similar to some of Hunter’s (1 994)
giant shells; she identifies these as H II r egions within the shells. We believe the same is
true for the region of N70 directly encircling the ionizing sta rs; these are volumes where
photoionization is the main influence on the gas. Of course the massive sta r s are blowing
winds into the gas, but there is nothing to distinguish these regions from o t her nebulae
interacting with massive main-sequence stars.
The line ratios of N70 share some properties with t he sample of SNRs; however, many
of the SNRs have higher [N II]/Hα but similar [S II]/Hα. As discussed below, these regions
perhaps have higher shock velocities than those important in N70. This comparison shows
that the very high [S II]/Hα ratios measured in N70 are not unknown in LMC emission
regions. The exception is the point at [S II]/Hα = 2.06, which is the north rim of the nebula,
and an area where the shape of the rim is not circular. It is possible that this is a volume
expanding into lower density material, or that there is some off-center effect of winds or a
sup ernova.
– 24 –
The more intriguing comparison is with the giant and supergiant shells (Hunter 1994).
Like N70, these shells show a range of both [N II]/Hα and [S II]/Hα ratios. The trends in
these line ratios are very similar for all of the shells; some regions of the shells a r e more
like H II regions, and others are more like SNRs. In her paper, Hunter presents many plots
of diagnostic line ratios. Although the data presented here does not include many of the
line ratios that Hunter examined, t his spectrophotometric info r matio n can be collected
from the liter ature on N70 (Lasker 1977, Lasker 1981, Dopita et al. 1981) and compared
with the giant and supergiant shells. For example, N70 line ratios such as [O III]/ Hβ and
[O II]/[O III] are also consistent with the values measured for Hunt er’s sa mple.
4.3. Nebular Models
A complete modeling of the photoionization and shock conditions in N70 is beyond t he
scope of t his paper. However, comparison with existing models can yield significant insights
on the respective roles played by stellar photons and shock waves.
Shull and McKee (1979) constructed theoretical models of interstellar sho cks moving
through a low-density medium. Their models pre-ionize the gas in front of the shock with
UV flux created by the shocked gas itself. In slow shocks, such as those to be considered
below, the gas flowing toward the shock is only partially pre-ionized. The strength of
the Balmer lines are sensitive to the pre-ionization, but metal-line strengths depend more
strongly on collisional excitation by electrons behind the shock. With solar abundances,
pre-shock density n
0
= 10 cm
−3
, and shock velocity of 40 km s
−1
, the Shull and McKee
(1979) models predict [N II]/Hα = 0.02 and [S II]/Hα = 1.24. One of the models also
explores the effect of depleted abunda nces on line ratios, predicting that [O I], [N I], and
[S II] are strengthened while [O II], [O III], and [N II] are weakened relative to Hα due to
diminished cooling and a larger hydrogen recombinatio n zone.
– 25 –
Hartigan et al. (1994) examine slower shocks, down to 15 km s
−1
, but with higher
pre-shock densities (n
0
= 10
2
, 10
3
, and 10
4
cm
−3
). These models show the [S II]/Hα flux
ratio increasing with decreasing shock velocity, reaching a peak value of ∼2 at a shock
velocity of 25 km s
−1
in the 10
2
cm
−3
model. In addition, [N II]/Hα decreases and [N I]/Hα
increases with decreasing shock velocity, as would be expected if the shock no longer has
the energy to ionize the nitrogen. (S
0
→ S
+
requires only 10.4 eV while N
0
→ N
+
requires
14.5 eV). Extending these results to the lower pre-shock densities estimated for N70 (0.1 –
0.7 cm
−2
; Rosado et al. 1981; Meaburn 1978) would suggest further enhancement of [S II]
as collisional de-excitation can be ignored.
The high [S II]/Hα ratios seen in N70 can be explained with the aforementioned
models with shock velocities of 25 to 40 km s
−1
, consistent with the reported expansion
velocities of 20 – 40 km s
−1
(Lasker 1977, Blades et al. 1980) but less than the expansion
velocity of 70 km s
−1
found by Rosado et al. (1981).
5
The low [N II]/Hα ratios in the outer
parts of N70 are also consistent with this interpretation. However, the [O III]λ 5007/Hβ
and [O II]λλ3 726,3729/Hβ measured by Lasker (1981) and Dopita (1981) as well as the
[O III]/Hα presented here are higher than the flux ratios predicted by the low-velocity
shock models. We therefore propose that N70 has distinct regions of photoionization
augmented to varying degree by shock ionization and excitation. This composite powering
is qualitatively apparent by comparing the morphology o f the [S II] emission with the Hα,
[N II], and [O III] emission morphologies, which are more similar to each other than to the
shock-excited [S II].
The unusually bright [O III] emission seen in the southern edge of N70 cannot be
5
Dopita et al. 1981 argues that the velocity field in N70 cannot be interpreted as simple
expansion. Even without coherent expansion, t he observed velocity dispersion is sufficient
to provide collisional (shock) excitation consistent with the models.
– 26 –
explained by the proposed photoionization plus slow-shock model. Perhaps a stronger shock
has excited the gas on the southern rim. X-ray data from the Einstein satellite (Chu and
Mac Low 1990) are sugg estive of an off-cent er supernova within the bubble, as their model
explains. The detection is marginal, but the enhanced [O III] emission could come from a
higher velocity shock than that exciting the [S II] on the western and northern rims of N70.
4.4. Radiative and Mechanical Energetics
In her study o f ionized bubbles in the LMC, Oey (1996b) found that those with high
X-ray luminosities and large expansion velocities have discrepantly small radii relative
to their expansion ages and stellar wind powering. She concluded that the dynamical
discrepancy is probably caused by recent sup ernova events accelerating the shells to higher
expansion velocities than would be obtained by stellar winds alone. She also suggested that
additional energy sink mechanisms might explain the anomalously smaller radii. Because
her sample of “superbubbles” included N70, further constraints o n the r elevant energetics
can be obtained by considering both the mechanical and radiative luminosities that are
involved.
4.4.1. Sources
As shown in Figure 1b, N70 contains 17 massive stars with spectral types ranging
from B2.5V to O3If (Oey 1996a). The total wind luminosity from this population wa s
modeled by Oey (1996b) to be no more tha n 10
37
erg s
−1
(see her Figure 1). A total
mechanical luminosity o f 1.5 × 10
37
erg s
−1
is obtained using the spectral classifications in
Oey (19 96a), the corresponding luminosities (for Galactic stars) listed in Leitherer (1997),
and adjustments for the LMC’s lower metallicity. Here, we used the theoretical prediction
– 27 –
that
˙
M ∝ Z
0.8
and v
∞
∝ Z
0.13
, where L
w
=
˙
Mv
∞
2
/2 (Leitherer 1997). Given this source
of power, is it sufficient to explain the currently observed X-ray luminosity, enhanced [S II]
emission, and expanding motions?
4.4.2. Sinks
N70 is noted for having a high X-ray luminosity relative to predictions based on its size
and expansion velocity (Oey 1996b, Chu and Mac Low 1990). Observations by the Einstein
observatory yield L
x
(Einstein) = 1.8 × 10
35
erg s
−1
(Chu and Mac Low 1990), which when
scaled up by 3 to the 0.1–2.4 keV ROSAT bandpass becomes L
x
(ROSAT ) = 5.4 × 10
35
erg
s
−1
, or a bout 7 times higher than is predicted from the nebular dynamics (Chu et al. 1995).
For our purposes, it is worth no ting that the total X-ray luminosity comprises a negligible
fraction of the total mechanical power that is available from the stellar winds.
Another, more important, radiative sink o f input mechanical power is the excess [S II]
emission that we measure. From Table 2, the ratio of summed [S II] and Hα fluxes is 0.77,
with a total [S II] flux of 9.24 × 10
−11
erg s
−1
cm
−2
, yielding a total [S II] luminosity of
3.37× 10
37
erg s
−1
. If photoionization typically produces flux ratios of f([S II])/f(Hα) ≤ 0.4
(see Figure 9), then ≥50% of the [S II] emission must result from other, more mechanical,
ionization/excitation processes. The required powering of L([S II]) ≥ 1.7 × 10
37
erg s
−1
would be mult iplied by about 1.5, if the excess cooling by [O I] is included (Dopita et al.
1981). These power requirements are marginally higher than those provided by the stellar
winds, leaving little “wiggle room” fo r any ot her energy sinks such as nebular expansion.
A variety of techniques can be used t o estimate the kinetic energies and mechanical
luminosities associated with expanding shells of gas (cf. Tenorio-Tagle and Bodenheimer
1988, Lozinskaya 1992). All of t hese techniques are critically sensitive to the density and
– 28 –
structure of the surrounding medium as well as the powering timeline, and hence are
fraught with uncertainties. The disparities in size, age, and expansion velocity found by
Oey (1996b) underscore these difficulties. Nevertheless, the expansion energetics can be
significant and hence are worth estimating.
In their kinematic study Rosado et a l. (1981) obtain a swept up mass of 2.3 × 10
3
M
⊙
and an expansion velocity of 70 km s
−1
, thus deriving a kinetic energy of 1.1 × 10
50
erg.
A more representat ive expansion velocity is about 35 km s
−1
(Chu and Kennicutt 1988),
resulting in kinetic energy of 2.8 × 10
49
erg. Averag ing this energy over the 5 My lifetime of
the cluster would then yield a mechanical luminosity of 2 × 10
35
erg s
−1
—comfortably less
than the 10
37
erg s
−1
available from the winds.
Slightly higher estimates of mechanical energy (E
m
= [1 – 7.5] × 10
50
erg) are obtained
with a momentum-conserving model for the expansion (Tenorio-Tagle and Bodenheimer
1988), where
E
m
= 5.3 × 10
43
n
1.12
0
R
3.12
v
1.4
,
the expansion velocity is v ≈ 35 km s
−1
and the ambient density is assumed to be
n
0
≈ 0.1– 0 .5 cm
−3
.
We conclude that the radiative and mechanical sinks of energy collectively exceed
the input wind power by factors of ∼ 2, the observed radiative sink of [S II] alone being
dominant. Allowing for other radiative sinks such a s [O I], [O II], and [O III] in the optical
(Dopita et al. 1981) and by C II, C II], C III, and C III] in the UV (cf. Dopita et al. 1984)
would further exacerbate the observed disparity in energetics. One or two recent supernovae
with individual energ ies of 10
51
ergs would be sufficient to make up the difference. Recent
sup ernova activity would also help to explain the anomalously high expansion velocity and
X-ray luminosity.
– 29 –
5. Summary
N70 is a fascinating emission-line region in the Large Magellanic Cloud whose spherical
symmetry belies its complex powering. The data presented here cannot solve the mystery
of N70’s dynamic history, but can provide new insights on the nebular energetics based
on diagnostic emission line ratios such as [N II]/Hα and [S II]/Hα. Our conclusions are as
follows:
• Although N70’s dynamics cannot be well explained by a standard pressure-driven
bubble model (Oey 1996b; note that the high luminosity half of her sample of bubbles
are inconsistent with the model), its emission-line ratios—[N II]/Hα and [S II]/Hα
from our data and [O III]/Hβ and [O II]/[O III] fro m the literature—match well with
the ratios of other LMC giant and supergiant shells in the LMC (Hunter 1994).
6
• N70’s central regions emit emission lines with flux ra tio s similar to those of
photoionized H II regions, while the rim of the N70 shows elevated [S II] emission.
The ionization of all of the hydro gen can be attributed to stellar EUV photons, but
additional processes such as slow shocks are necessa ry to explain the combinatio n of
high [S II] emission and low [N II] emission, especially in the northeast and so uthern
parts of the nebula.
• The energetics associated with the stellar winds, expanding shell, and radiating [S II]
are best reconciled if one or two supernova explosions have occurred within N70 in
the past ≈ 10
6
years. The enhanced [O III] emission and marginal X-ray detection to
the south also indicate higher velocity sho cks from recent supernovae.
6
The samples of Oey (1996b) and Hunter (1994) do not overla p; a useful endeavor would
be to collect emission-line diagnostics for Oey’s sample for comparison with Hunter’s data,
as well as to investigate the dynamics of Hunter’s sample using Oey’s model.
– 30 –
A wealth of information about the small-scale details of ionization and shock fronts has
been gained about o ther emission regions with HST and WFPC2 (e.g. Hester et al. 1995,
1996); some of the remaining questions about N70 could be answered with higher resolution
images, especially a finer-scale mapping of line ratios across its filament s.
We would like to thank Don Walter for sharing his calibrated images of the O rion
nebula. We also thank the referee, You-Hua Chu, for her guidance and patience as we
wrestled with the nebular energetics. BPS would like t o thank Paul Hodge, Gene Magnier,
and Eliot Malumuth for helpful advice and conversations while reducing and a na lyzing
this data. WHW is grateful to NASA for funding under the Astrophysics Data Program
(#071-96adp), to the UIT Science Team led by Ted Stecher, and to the Goddard SNR and
ISM lunch bunch led by Robin Shelton a nd Jonathan Keohane for intellectual stimulation
and supp ort.
– 31 –
REFERENCES
Atherton, P. D., Reay, N. K., Ring, J., & Hicks, T. R. 1981, Opt. Eng., 20, 806
Blades, J. C., Elliott, K. H., & Meaburn, J. 1980, MNRAS, 192, 101
Brown, L. W., Woodgate, B. E., Ziegler, M. M., Kenney, P. J., & Oliversen, R. J. 1994,
Rev. Sci. Instrum., 65 , 3611
Brunet, J. P. 1975, A&A, 43, 245
Caulet, A., Woodgate, B. E., Brown, L. W., Gull, T. R., Hintzen, P., Lowenthal, J. D.,
Oliversen, R. G., & Ziegler, M. M. 1992, ApJ, 388, 301
Chu, Y.-H., Chang, H.-W., Su, Y.-L., & Mac Low, M.-M. 1995, ApJ, 450, 157
Chu, Y.-H., & Kennicutt, R. C. 1988, AJ, 95, 1111
Chu, Y.-H., & Mac Low, M.-M. 1990, ApJ, 365, 510
Danziger, I. J., & Leibowitz, E. M. 1985, MNRAS, 216, 365
Dopita, M. A., Ford, V. L., McGregor, P. J., Mathewson, D. S., & Wilson, I. R. 1981, ApJ,
250, 103
Dopita, M. A., Binette, L., D’Odorico, S., & Benvenuti, P. 1984, ApJ, 276, 653
Feast, M. W., & Catchpole, R. M. 1997, MNRAS, 286, L1
Georgelin, Y. M., Georgelin, Y. P., Laval, A., Monnet, G., & Rosado, M. 1983 , A&AS, 54,
459
Goudis, C. 1982, The Orion Complex: A Case Study o f Interstellar Matter (Reidel,
Dordrecht)
– 32 –
Hamuy, M., Wa lker, A. R., Suntzeff, N. B., Gigoux, P., Heathcote, S. R., & Phillips, M. M.
1992, PASP, 104, 533
Hartigan, P., Morse, J. A., & Raymond, J. 199 4, ApJ, 436, 125
Hester, J. J., et al. 1995, ApJ, 448, 240
Hester, J. J., et al. 1996, AJ, 111, 2349
Hunter, D. A. 1994, AJ, 10 7, 565
Isserstedt, J. 1975, A&A, 41, 175
Kurucz, R. L. 1992, in IAU Symposium 149, The Stellar Populations of Galaxies, ed. by B.
Barbuy and A. Renzini (Kluwer, Dordrecht), 225
Lasker, B. M. 1977, ApJ, 212 , 390
Lasker, B. M. 1981, PASP, 93, 422
Leitherer, C. 1997, in Stellar Astrophysics for t he Local Group, ed. by A. Aparicio, A.
Herrero, and A. Sanchez (Cambridge University Press, Cambridge), in press
Lozinskaya, T. A. 1992, Supernovae and Stellar Wind in the Interstellar Medium (AIP
Press, New Yor k)
Lucke, P. B., & Hodge, P. W. 1 970, AJ, 7 5, 171
Massey, P., Lang, C. C., DeGioia-Eastwood, K., & Garmany, C. D. 1995 , ApJ, 438, 188
Mathis, J. 1990, ARA&A, 28, 37
Meaburn, J. 1 978, Ap&SS, 59, 193
Oey, M. S. 1 996a, ApJ, 465, 2 31
– 33 –
Oey, M. S. 1 996b, ApJ, 467, 666
Oey, M. S., & Kennicutt, Jr., R. C. 1997, MNRAS, 291, 827
Osterbrock, D. E. 198 9, Astrophysics of Gaseous Nebulae and Active Galactic Nuclei (Univ.
Science Press, Mill Valley)
Petuchowski, S. J., & Bennett, C. L. 1995, ApJ, 438, 735
Pogge, R. W., Owen, J. M., & Atwood, B. 19 92, ApJ, 399, 147
Rosado, M., Georgelin, Y. P., Georgelin, Y. M., Laval, A., & Monnet, G. 1981, A&A, 97,
342
Russell, S. C., & Dopit a, M. A. 1992, ApJ, 384, 508
Shull, J. M., & McKee, C. F. 1979, ApJ, 227, 131
Silva, D. R., & Cornell, M. E. 1992, ApJS, 81, 865
Stone, R. P. S., & Baldwin, J. A. 1983, MNRAS, 204, 347
Tenorio-Tagle, G. & Bodenheimer, P. 1988, ARA&A, 26, 145
Vacca, W. D., Garmany, C. D., & Shull, J. M. 1996, ApJ, 460, 914
Walter, D. K., Dufour, R. J., & Hester, J. J. 1992, ApJ, 397, 196
Wen, Z., & O’Dell, C. R. 1 995, ApJ, 438, 784
Westerlund, B. E., & Mathewson, D. S. 1966, MNRAS, 131, 371
Wilcots, E. M. 1992, Ph.D. thesis, University of Washington
This manuscript was prepared with the AAS L
A
T
E
X macros v4.0.
– 34 –
Fig. 1 .— (a) Dereddened Hα emission-line image of N70 taken with a Fabry-Perot imaging
system with FWHM∼6
˚
A. (b) 6536
˚
A continuum; stars are labeled with spectral types from
Oey (19 96a). The field of view has a diameter of about 7
′
, or 110 pc at the distance of the
LMC.
Fig. 2.— Dereddened emission-line images of N70: (a) [N II] (smoothed with a 2 pixel
Gaussian), (b) [S II], and (c) [O III]. The FWHM is ∼6
˚
A for [N II] and [S II], and ∼7
˚
A for
[O III]. The field of view has a diameter of about 7
′
.
Fig. 3.— Color-coded image of N7 0 constructed from emission-line images. Blue is [O III],
green is Hα, and red is [S II]. Not e t hat yellow = green + red.
Fig. 4.— Emission-line images of the central region of Orion, approximately centered on the
Trapezium: (a) Hα, (b) [N II], (c) [S II]. The field of view has a diameter of about 7
′
, or less
than 1 pc.
Fig. 5 .— Emission-line ratio maps of N70: (a) [N II]/Hα, ( b) [S II]/Hα , and (c) [O III]/Hα.
The [N II], [O III], and Hα maps were smoothed before constructing the [N II]/Hα and
[O III]/Hα ratios. The [O III]/Hα map does not account f or the radially decreasing sensitivity
at [O III] (see text).
Fig. 6.— Hα images of N70 and Orion showing the polygonal apertures used for aperture
photometry of various regions of emission. Fluxes and ratios are listed in Tables 2 and 3.
Fig. 7.— [S II]/Hα intensity ratio s across three filaments in N7 0, o ne in the southeast, one
in the south, and another in the west. These filaments are characteristic of many of the
filaments in N70. The [S II]/Hα rat io across the Orion Bar is also plotted (multiplied by a
factor of ten so that it could be seen more easily). Each of the four cuts begins on the side
of the filament closer to the ionizing stars (LH 11 4 for N70 and the Tra pezium for Orion)
– 35 –
and extends 16 pixels (15.
′′
4) radially away from the center. 15.
′′
4 corresponds to 4.1 pc at
the distance of the LMC and 0.0 34 pc at the distance of Orion.
Fig. 8.— [N II]/Hα vs. [S II]/Hα for regions in N70 and the Orion Nebula. The ratios
plotted are those in Ta bles 2 and 3 from the polygonal apertures shown in Figure 6.
Fig. 9.— [N II]/Hα vs. [S II]/Hα for a variety of line-emissio n regions in the Large
Magellanic Cloud. The line ratios of N70 span the excitation domain populated by giant
shells, supergiant shells, and even SNRs.