Please Note: the e-mail address(es) and any
external links in this paper were correct when it was written in 1995,
but may no longer be valid.
Go to: Front page, Participants, Group Photograph, Preface, Contributions, Acknowledgements
High resolution studies of protostellar
condensations in NGC 2024
Helmut
Wiesemeyer
e-mail:
p713hwi@mpifr-bonn.mpg.de
Max-Planck-Institut für Radioastronomie, Bonn, GERMANY
Abstract:
We present high resolution (
) studies of the 3 mm dust continuum and
line emission from protostellar condensations in one
of the Orion B molecular cloud cores, located southwards from the ionization
front of the NGC 2024 HII region. Continuum radiative transfer
calculations under spherical symmetry and standard LVG radiative transfer
modeling of the high density gas tracing, optically thin 
transition allowed study of the conditions in and around these sources. The
results confirm that they are extremely young (``class 0'') protostars. The
emission and the dust emission show a clear lack of coincidence,
which is most likely due to freeze-out of molecules onto dust grains.
The NGC 2024 HII region is located in the Orion B molecular cloud complex
at 415 pc distance. It is surrounded by an elongated cloud core which is
believed to be a site of active star formation. The associated dust lane
appears in optical images by virtue of its high extinction. A VLA 
1.3
cm map reveals two continuum peaks bracketing the HII-region
(Gaume et al. (1992)). The relatively low gas temperatures reported by several authors
are consistent with the strong temperature gradient south from the HII
region (Schulz et al. (1991)). Mezger et al. observed the cloud core at millimeter and
submillimeter wavelengths and found several compact far infrared sources
(FIR1-FIR7, Mezger et al. (1992)). Although the true nature of these cores is still
controversial, star forming activity is evidenced by two associated outflows (a
highly collimated, unipolar flow with FIR5 and a bipolar, compact one with
FIR6, Richer et al. (1992),Richer (1990)) and a water-maser close to FIR6 (Genzel & Downes (1977)). We
believe that these cores represent extremely young protostars.
André et al. extended the classification of young stellar objects proposed
by Lada to ``class 0'' sources (André et al. (1993),Lada (1988)); the early evolutionary
status of these sources is evidenced by an envelope which is still more massive
than the central accreting core. The high optical depth towards the core
explains the absence of observable near infrared emission.
Figure 1: The observed LSB 
3 mm continuum emission (grey scale) is
shown together with 
emission (contour spacing 200
mJy, corresponding to 
) integrated from 10.9 to
(left) and 10.2 to 
(right).
The positions are taken from Mezger et al. (1992) (continuum emission) and from
Genzel & Downes (1977) (
-maser). The clean beams of the line (left) and
continuum (right) observations are shown in the lower left corner.
The observations were performed between 1991 January and 1992 February with the
IRAM three-element interferometer on the Plateau de Bure, comprising baselines
between 32 and 288 meters. The synthesized beams are 
for the line observations (corrected for missing short spatial
frequencies) and 
for the simultaneous continuum
observations. The half-power response of the primary beam of the antennas
(
at 96 GHz) included both FIR5 and FIR6. The spectral line
correlator was centered in the upper sideband on the 
transition at 96.4 GHz, covering 10 MHz of bandwidth with 128 channels,
resulting in a channel separation of 
. The bandwidth of
the continuum correlator was 
MHz. 3C84 was used as bandpass and
phase calibrator and 
as amplitude calibrator. The calibrated
visibilities were Fourier-transformed to 
pixel maps with
pixel size, using natural weighting of the data. The best results,
shown in Figure 1, were derived with the interactive, combined use of
the Cotton-Schwab and the Clark algorithms.
The high critical density (
) of
the 
transition allows for the separation of the dense cores from
the extended, low density medium. To determine the physical conditions in the
cloud cores, we decomposed the observed emission into clumps with
gaussian-shaped intensity profiles, using an iterative least-square algorithm
(Stutzki & Güsten (1990)). A drawback of this approach is that coherent structure appears
as several clumps if strong velocity gradients are present. This results in
too large brightness temperatures. Only if the 
emission at these
temperatures is optically thin will the masses derived from them be correct.
Despite this and the elongated beam shape which makes a clear deconvolution
difficult, mass determinations by means of standard large velocity gradient
modelling were possible for a few clumps of size 
. Adopting a 
abundance of 
, the clump masses of 
are consistent
with virial masses within a factor of 2. Volume averaged 
densities
are of the order of 
.
Figure 2: Left: Position-velocity cut along the ridge north of FIR6. Contour
spacings are 
, 
, 150 to 
. Right:
Velocity structure around FIR6. The size of the squares correspond to a
range from 
to 
with respect to 
. Grey pixels correspond to blueshifted emission, black
pixels to redshifted emission. The cross marks the 
position of FIR6. The offsets are given with respect to the phase reference
center.
To estimate the gas mass around FIR6, we propose a model which consists of a
rotating and expanding streamer - see Figure 2 - of inner radius
and outer radius 
. The rotation is
Keplerian in nature (
) and the expansion
is decelerated (
). From the column density
per velocity bin, we determined the line temperature on a rectangular grid at
spacing. The intrinsic half power linewidth has been chosen to be
, which exceeds the thermal linewidth expected for the
35 K gas by a factor of 2.7, to take into account the line broadening due to
microturbulence. The results are then convolved with the clean beam to compare
with the observations. For the chosen parameter set, the total mass is 
. The inclination of the rotation axis of the streamer is consistent
with that of the compact outflow axis, estimated by comparison of the lobe
shape (Richer (1990)) with outflow models (Cabrit & Bertout (1986)). In this scenario, the
streamer extends perpendicularly to the outflow axis.
A least square fit to the observed dust continuum visibilities, allowing for
several source components, yielded the deconvolved source size with relative
errors of 20-30%. The error in determining the peak intensity is governed by
the intrinsic calibration uncertainty which is also 
. Total
fluxes at 
3 mm are 250 mJy (FIR5, both components) and 90 mJy (FIR6).
The FIR5 main component is 
AU in size (FWHP), FIR6 is more compact
(460 AU). The nature of the other FIR5 component with its east-west structure
(beam deconvolved size 
) elongated perpendicularly
to the outflow axis is suggestive of a disk envelope. These structures are not
unexpected for young stellar objects at this stage of evolution, although a
clear identification is yet to be confirmed.
Figure 3: Spectral energy distribution as calculated by radiative transfer
modeling under spherical symmetry for the given model specifications. Errorbars
show the estimated flux uncertainty. From left to right: VLA 
(upper limit), this work 
, IRAM 30-m 
and 
. Upper limits for the near infrared emission are
also shown. For reference, an isothermal envelope (
) with
constant density is given by the dotted line.
The FIR5 and FIR6 protostellar envelopes are assumed to have a temperature
gradient perhaps maintained by an accretion shock. Thus, we performed radiative
transfer calculations, using a code developed by Yorke which solves
self-consistently for the equation of radiative transfer in spherical geometry,
given a user-specified dust model (Yorke (1980)). We explored a set of
solutions constrained by the observational results: The VLA 
map (Gaume et al. (1992)) does not show any evidence for the dust condensations,
thus excluding the possibility of a flat spectral index and confirming the
thermal nature of the continuum emission. It does not exclude, a priori, a
luminous source in the envelope; such an object is likely to be heavily
embedded such that its ionized region would be restricted in size. At 
, the visibilities measured at Plateau de Bure yield the total
source fluxes and the envelope size estimate (see previous section). The fluxes
at 
and 
were determined from observations
with the IRAM 30m telescope (Mezger et al. (1992)). In this context, one has to avoid
the contribution from the extended emission to which the interferometer is
insensitive. Near infrared observations lack any evidence of the central
sources (Meyer (1995)). Far infrared observations (Thronson et al. (1984)) detected hot
dust associated with the HII region, but there is no evidence for
dominant emission from FIR5. Furthermore, we used a three component dust model
(Preibisch et al. (1993)) as a realistic approach to the dust grain properties. It
consists of amorphous carbon (aC) grains smaller than 30nm and astronomical
silicates of up to 
in size. The size distribution (above a lower
cut-off) follows the power law 
(Mathis et al. (1977)). At
temperatures below 2000 K, 60% of the ISM carbon abundance condenses out as aC
grains. Astronomical silicate grains are processed from the full Si abundance
at temperatures below 1500 K. Below 125 K, they are covered by ice coatings
which are polluted by aC grains enclosed in the ice mantles (20 Vol.%). The aC
grains are not expected to develop ice coatings (for details see Preibisch et al. (1993)).
The flux measurements cited above are not sufficient to constrain the exponent
of the density fall-off in the envelope. We adopted a power law 
which is appropriate for matter behind an inside-out
expanding collapse wave (Shu (1977)), although a flat density gradient cannot
be ruled out (e.g. André et al. (1993)). Keeping the outer radius of the envelope
fixed at 
cm (1060 AU), the density profile is determined by
the total envelope mass, the inner radius which is given by the aC sublimation
radius, and the density power law index. The radiation field of the central
source shining on the inner edge of the dust envelope is the inner boundary
condition for the outwards directed intensity component. The measured fluxes
can be explained in terms of internal heating only, taking as outer boundary
condition to the inwards directed intensity component a 3 K undiluted blackbody
radiation field (case A in Table 1). To consider the more realistic
case of both external and internal heating, we used a blackbody radiation field
at 
K diluted by a factor 
(case B). As the dust composition
is dependent on the equilibrium temperature and governs the spectral energy
distribution (SED) of the radiation field (which in turn determines the
equilibrium temperature), the code has to work iteratively to solve for the
dust sublimation radii. One of the results is presented in Figure 3.
Table 1: Results from spherical-symmetric radiative transfer (FIR5, see text
for cases A and B). 
is the central source bolometric luminosity, and for
the dust models 
is core size, 
is mantle outer radius, and 
is
coating thickness.
Due to the high extinction towards the HII region, it appears unlikely
that external heating by the near-infrared sources plays an important role,
unless the material is very clumpy (Schulz et al. (1991)). But according to model
results, even moderate external heating can considerably lower the luminosity
of the central source which is necessary to explain the SED. The central source
luminosities and envelope masses depend on the assumed dust model. All models
show that 90-98% of the envelope volume contain cool dust (
), corresponding to 70-90% of the envelope mass. The most striking
feature in Figure 1 is the anticorrelation between 
dust and 
emission. It is most likely due to a freeze-out of
elements which are important for 
synthesis onto the surface of
dust grains (or as ice incorporated in fluffy grains). Furthermore, our model
results give evidence that submillimeter observations suffer from optically
thick conditions and do not penetrate deeper than into the cool outer parts of
the envelope. On the other hand, optically thin emission at millimeter
wavelengths onwards traces the cold dust component. Thus, submillimeter fluxes
can be described by an isothermal, cool envelope (
). Two questions remain to be answered: what drives the FIR5 and FIR6
outflows, and what is the nature of the embedded sources? If FIR5 is indeed a
``class 0'' object, a low 
ratio is not unexpected.
Centrifugally driven MHD winds may be good candidates for the driving source of
the CO outflow instead of radiation pressure, provided that a disk has already
formed. Conclusions concerning the embedded source are difficult to make, as
the effective temperature of the central source (we assumed 5000 K) has no
influence on the spectrum within reasonable limits; the dust grains convert the
radiation field at the inner radii rapidly into one at lower temperature. More
conclusive is that all the models have, as common feature, high envelope masses
and low luminosities. Having this result in mind, the status of the embedded
source can be estimated in the following way. Half of the accretion energy is
radiated away at the accretion shock (Pringle (1981)). Assuming that the
conditions in the envelope are such that the free-fall time scale corresponds
to the sound-crossing time (see Stahler et al. (1980) for reference), it can be readily
shown that the mass of the protostellar core (for constant central source
temperature and constant outer envelope radius) scales with the ratio of the
central source luminosity to the envelope mass as 
. It follows from the low 
ratio that the core of FIR5 must be an extremely young object which is yet to
gain most of its mass by accretion. If the assumption of spherical geometry is
a valid description, FIR5 must be a ``class 0'' source. To constrain the
ratio, the turnover of the spectrum which becomes
evident at 
, and the lack of detectable infrared emission, are
of crucial importance. In the optically thin regime of the SED, a lower
envelope mass and higher central source luminosity can explain the observed
fluxes as well, but not in the optically thick regime (i.e. the 
flux). Final conclusions certainly have to wait until submillimeter
observations better constrain the 
ratio.
This project was initiated and supervised by Rolf Güsten (Bonn) and carried
out in collaboration with Jörn E. Wink (Grenoble). The author owes Harold W.
Yorke and Thomas Preibisch (Würzburg) a debt of gratitude for assistance with
the radiative transfer calculations and the dust model. IRAM is a joint
collaboration between the German MPG, the French CNRS, and the Spanish INSU.
- Andre P., Ward-Thompson D. & Barsony M., 1993,
ApJ, 406, 122.
- Cabrit S., Bertout C., 1986,
ApJ, 307, 313.
- Gaume R.A., Johnston K.J. & Wilson T.L.W., 1992,
ApJ, 388, 489.
- Genzel R., Downes D., 1977,
A&AS, 30, 145.
- Lada C.J., 1988,
in Formation and Evolution of Low Mass Stars,
eds Dupree A.K. & Lago M.T.V.T., p.93, (Kluwer, Dordrecht).
- Mathis J.S., Rumpl W. & Nordsieck K.H., 1977,
ApJ, 217, 425.
- Meyer M., 1995,
in preparation.
- Mezger P.G., Sievers A.W., Haslam C.G.T., Kreysa E., Lemke,
R., Mauersberger R. & Wilson T.L.W., 1992,
A&A, 256, 631.
- Preibisch, Th., Ossenkopf V., Yorke H.W. & Henning, Th., 1993,
A&A, 279, 577.
- Pringle J.E., 1981,
ARAA, 19, 137
- Richer J.S., 1990,
MNRAS, 245, 24P.
- Richer J.S., Hills R.E. & Padman R., 1992,
MNRAS, 254, 525.
- Schulz A., Güsten R., Zylka R. & Serabyn E., 1991,
A&A, 246, 570.
- Stahler S.W., Shu F. & Taam R., 1980,
ApJ, 241, 637.
- Stutzki J., Güsten R., 1990,
ApJ, 356, 513.
- Thronson H.A., Lada C.J., Schwartz P.R., Smith H.A., Smith,
J., Glaccum W., Harper D.A. & Loewenstein R.F., 1984,
ApJ, 280, 154.
- Shu F.H., 1977,
ApJ, 214, 488.
- Yorke H.W., 1980,
A&A, 86, 286.
YERAC 94 Account
Wed Feb 22 23:27:27 GMT 1995
