Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://www.arcetri.astro.it/~lt/preprints/afgl5142/afgl5142_preprint.ps.gz Äàòà èçìåíåíèÿ: Tue Sep 11 16:56:42 2007 Äàòà èíäåêñèðîâàíèÿ: Sat Dec 22 08:57:47 2007 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: reflection nebula

A&A manuscript no.
(will be inserted by hand later)
Your thesaurus codes are:
08(02.13.3; 08.06.2; 09.08.1; 09.10.1; 09.13.2; 13.19.3)
ASTRONOMY
AND
ASTROPHYSICS
19.2.1995
A multiwavelength picture of the AFGL5142 star­forming
region
Todd R. Hunter 1 , Leonardo Testi 2 , Gregory B. Taylor 1 , Gianni Tofani 3 , Marcello Felli 3 , and Thomas
G. Phillips 1
1 California Institute of Technology, Department of Physics, Math and Astronomy, 320­47, Pasadena, CA 91125, U.S.A.
2 Dipartimento di Astronomia e Scienza dello Spazio, Universit`a di Firenze, Largo E. Fermi 5, I­50125 Firenze, Italy
3 Arcetri Astrophysical Observatory, Largo E. Fermi 5, I­50125 Firenze, Italy
Received August 17, 1994; accepted February 22, 1995
Abstract. We present molecular line, H 2 O maser, radio
continuum and near infrared maps of the bipolar outflow
source AFGL 5142. The high resolution of our molecu­
lar CO observations enables us to define the morphology
of the large­scale bipolar outflow into a two lobe struc­
ture extending for ¸ 2 0 on each side of the center. In the
perpendicular direction, we find consistent evidence for a
second, more compact (! 0:5 0 ) outflow in the form of a
spatial velocity offset in the CO map and of an H 2 jet­like
structure derived from a near infrared narrow band image.
On a smaller scale size, the radio and infrared continuum
observations reveal the engines of the molecular outflows.
The maser emission occurs near the position of the most
embedded source of the cluster, IRS1. This is located at
the center of the compact outflow and jets of shocked H 2
and coincides with an ultra compact (UC) radio contin­
uum source (most probably an ionized stellar wind). The
H 2 O cluster is composed of five spatial components: two
are within 0:2--0:3 00 from the YSO (a few hundred AU) and
three are at larger distance (1:5--2 00 , a few thousand AU).
A marginal detection of proper motion of the two more
distant masers may suggest a high expansion velocity at a
distance 4 10 3 AU from the YSO, similar to what is found
in Orion KL and W49N. The brightest NIR source of the
cluster (IRS2) is associated with an IRAS point source
and lies along the axis of the large­scale bipolar outflow.
We propose that the masers and the compact molecular
outflow are powered concurrently by the wind from the
YSO associated with IRS1, while the large scale outflow
could be the remnant from the formation of IRS2.
Key words: ISM: jets and outflows -- ISM: individual
(AFGL 5142) -- masers -- stars: formation
1. Introduction
H 2 O masers and CO molecular outflows have been known
to be associated in star forming regions for nearly two
decades (e.g. Genzel & Downes 1977; Lada 1985). Even
though these phenomena occur on two widely different
physical scales (¸ :01 pc and ¸ 1 pc, respectively), re­
cent low resolution surveys have confirmed their frequent
association (both spatial and kinematic) and suggested
that the two are powered by a single luminous YSO (Felli
et al. 1992). In a survey by Henning et al. (1992), H 2 O
emission was found to be present toward 80% of the ob­
served sources with known molecular outflows. The fact
that these phenomena tend to appear together when ob­
served at low resolution motivates a higher resolution
search for their detailed physical connection. Presumably,
the relationship between CO outflows and H 2 O masers is
governed by the sequence of events that lead to the for­
mation of a new star. Recent evidence indicates that the
bipolar molecular outflow begins during the infall phase
of protostellar evolution (see Shu et al. 1991 and refer­
ences therein) and the H 2 O masers are very good tracers
of very young stellar objects, deeply embedded in high
density molecular clumps. At the same time, recent inter­
ferometric observations have suggested that H 2 O masers
lie in much smaller scale disk­like structures around young
stellar objects such as Orion KL (Matveyenko et al. 1993)
and Orion IRc2 (Wright et al. 1990). In addition, proper
motion studies of the W49N masers reveal outflowing mo­
tion from a common center (Gwinn et al. 1992). Whether
these structures and/or motions are a common feature of
H 2 O maser sources and how they relate to the molecular
outflows are key questions in understanding the processes
that occur during the formation of massive stars and can
be answered only with improved resolution studies.
To explore the spatial correlation of molecular out­
flows with maser phenomena and to search for the ear­
liest phases of stellar formation we follow a multiwave­

2 Todd R. Hunter et al.: A multiwavelength picture of the AFGL 5142 star­forming region
length strategy (Hunter et al. 1995). We perform CO and
CS mapping of massive star­forming regions with known
H 2 O maser emission and broad CO lines. Maps with Ÿ 30 00
resolution of the molecular outflows are being obtained
in order to find the accurate position of the dynami­
cal center and the energetics of the outflow. These maps
are then compared to radio interferometric observations
(` beam ¸ 0.1 00 ) which resolve and image the H 2 O maser
spots and, in a few cases, reveal the presence of UC radio
components (Tofani et al. 1995), i.e. the unmistakable in­
dication of a YSO. Finally, to search for direct evidence
of these young stars and to obtain information on their
spectral energy distributions, in cases where UC compo­
nents are found, we obtain broad band and narrow band
near­infrared images.
Here we present new results on the AFGL 5142 star­
forming region (distance ¸1.8 kpc, Snell et al. 1988).
The IRAS source has a bolometric luminosity of 3:8 \Theta
10 3 L fi (Carpenter, Snell & Schloerb 1990). A CO outflow
associated with the IRAS source was found with moderate
resolution observations by Snell et al. (1988). H 2 O maser
and NH 3 emission were first detected toward this source
by Verdes­Montenegro et al. (1989) and OH maser emis­
sion was detected by Braz et al. (1990). Using the VLA in
C­configuration, Torrelles et al. (1992) resolved three H 2 O
spots (seven velocity components) nearly coincident with
a weak (0.69 mJy peak flux density) 8.4 GHz continuum
source with a cometary morphology. Its radio continuum
flux density is consistent with a B2 ZAMS star, assuming
optically­thin free­free emission. The continuum source
was also detected at 5 GHz using the VLA D­configuration
with a flux density between 0:5 and 1 mJy (McCutcheon
et al. 1991 and Carpenter et al. 1990). The masers and
the UC source lie approximately 30 00 to the east of the
IRAS position, just outside of the error box given in the
Point Source Catalog, so the association between the two
has been questioned (Torrelles et al. 1992). NH 3 obser­
vations of Estalella et al. (1993) with a 40 00 beam show
evidence for a dense hot core of gas centered on the H 2 O
maser and the weak radio continuum source, but removed
from the IRAS position. We present new observations of
the molecular gas, radio continuum, H 2 O maser and near­
infrared emission that answer several questions concerning
this source.
2. OBSERVATIONS
2.1. Caltech Submillimeter Observatory
The initial CSO 1 observations were performed on 14
September 1993. The source was mapped in the CO
J=2!1 transition with the 1024 channel 50 MHz band­
width acousto­optical spectrometer (AOS) as the back­
end. The CSO beamsize at this frequency is 31 00 with a
1 The Caltech Submillimeter Observatory is funded by the
National Science Foundation under contract AST­9313929.
main beam efficiency of 0.76. Two coverages of three fully­
sampled 7 \Theta 15­point maps (15 00 grid) were obtained in
the on­the­fly mapping mode with the telescope driven at
3 00 s \Gamma1 . The partially­overlapping maps were subsequently
summed and combined to form a 17 \Theta 15­point image with
at least 10 seconds of total on­source time per point. The
offset position (30 0 north) was verified to be free of emis­
sion to a level ! 0:05 K, which is well below the noise level
of the maps. The central position of the map corresponds
to the VLA H 2 O maser position. Five minute spectra of
the 13 CO J=2!1 emission were also taken with a 33 00
beam toward the central position and toward the peak
positions in the outflow lobes. A followup on­the­fly map
of the central region was taken on 2 February 1994 in the
CO J=3!2 and CS J=7!6 lines simultaneously, using
the 500 MHz AOS with a beamsize of 20 00 and a main
beam efficiency of 0.67. Finally, a 13 CO J=6!5 spectrum
(661.068 GHz) was obtained toward the central position
using the new 650 GHz SIS receiver (Kooi et al. 1994)
with an 11 00 beam and a main beam efficiency of 0.30. All
spectra were analyzed with the ``CLASS'' software pack­
age and are presented on a main beam brightness tem­
perature (Tmb ) scale. We estimate our calibration to be
accurate within 20% and our pointing to within 4 00 .
2.2. VLA
The VLA 2 observations took place on 1992 November 24
in the A­configuration as part of a larger survey of H 2 O
masers associated with CO outflows (Tofani et al. 1995).
The source was observed for 7 minutes in line mode 2AD
at 22 GHz, and for 5 minutes in continuum at 8.4 GHz.
The velocity resolution is 0.66 km s \Gamma1 and the obser­
vations with the A IF cover the velocity range ­19 to
+17 km s \Gamma1 . The D IF covers a 25 MHz bandwidth, off­
set 25 MHz from the line emission in order to search
for 22 GHz continuum emission simultaneously with the
H 2 O line observations. We obtained a synthesized beam of
0:12 00 \Theta0:11 00 (p.a.= \Gamma55 ffi ) and 0:28 00 \Theta0:27 00 (p.a.= \Gamma66 ffi )
at 22 and 8.4 GHz respectively.
2.3. NIR observations and data reduction
The near--infrared observations were carried out on 12
February 1994 (broad band) and on 17 February 1994
(narrow band), using the Arcetri NIR camera (ARNICA)
mounted at the 1:5 meter infrared telescope TIRGO 3 . AR­
NICA is equipped with a NICMOS3 256 \Theta 256 HgCdTe
panoramic detector with a scale of ¸ 0:955 arcsec=pixel
2 The National Radio Astronomy Observatory is operated
by Associated Universities, Inc., under cooperative agreement
with the National Science Foundation.
3 The TIRGO telescope is operated by the C.A.I.S.M.I.--
C.N.R. on behalf of the Ministero dell'Universit`a e della
Ricerca Scientifica e Tecnologica

Todd R. Hunter et al.: A multiwavelength picture of the AFGL 5142 star­forming region 3
(Lisi et al. 1993 and Hunt et al. 1994a). All the data re­
duction and analysis were performed using the ARNICA
(Hunt et al. 1994c) and IRAF software packages. Absolute
position calibration was achieved using the coordinates of
a number of stars from the Hubble Space Telescope Guide
Star Catalogue that were contained in our mosaics. Due to
the least square technique used to solve for the plate con­
stants we estimate an astrometric accuracy of the order
or better than one arcsecond (Testi 1993). The full width
at half maximum of the Point Spread Function obtained
in these observations is ¸ 2--3 00 .
2.3.1. Broad band imaging
Using the three standard J, H and K broad band NIR
filters we observed the target source using a ``dithering''
technique: in each band we take images slightly chang­
ing the telescope position, in such a way that the object
of interest is inside all frames. The flat field image is ob­
tained combining a set of frames using a median algo­
rithm. The precise data acquisition and reduction proce­
dure is described in Hunt et al. (1994b,c). After reduction
the images are registered and combined in order to form
an extended map of the observed region. The final maps
are approximately 7 0 \Theta 7 0 wide. The signal to noise ratio
is not constant over the mosaic, because of the observing
technique, being higher at the center (covered by all the
frames) and poorer at the edges. Photometric calibration
was acheived observing a set of UKIRT's faint standard
stars. We estimate a photometric accuracy of about 8%.
Accurate photometry was made on all the detected point
sources using the DAOPHOT routines in IRAF. The lim­
iting magnitudes obtained at the edge of the mosaic were
17:3, 16:2 and 16:0 in J, H and K respectively.
2.3.2. Narrow band imaging
The narrow band images were taken using narrow band
interference filters with a spectral resolution of –
\Delta– ¸ 100,
centered on the rest frequencies of the Brackett fl and
H 2 S(1) 1 ! 0 lines (2:166 ¯m and 2:122 ¯m).
The observing technique is similar to that used for
broad band imaging, except for the fact that the ``dither­
ing path'' has shorter offsets so the useful field of view is
much smaller in the narrow bands (2 0 \Theta 2 0 ). The reduc­
tion technique was the same as that used for the broad
band observations. The mosaics obtained in this case are
smaller than those in the broad bands, covering 3 0 \Theta 3 0 .
The continuum emission at the frequencies of the two
narrow bands were estimated using the broad band filters.
A synthetic continuum image was built by calculating at
each position the power law spectrum which was best fit
to the broad band data. The flux calibration was made as­
suming that a set of stars present in the field do not have
line emission or absorption, and hence their flux in the
narrow band image should be the same as in the synthetic
continuum. In order to match the Point Spread Function
of the continuum image, the narrow band images were
convolved with a Gaussian function of proper width, be­
fore subtraction. Hence the narrow band images have been
scaled with a proper factor to convert counts to flux units,
then convolved and continuum subtracted.
3. RESULTS
3.1. Bipolar outflows
The CO J=2!1 and CS J=7!6 spectra of the central
position (ff = 05 h 27 m 30. s 0, ffi = 33 ffi 45 0 40. 00 0) are shown
in Fig. 1. High velocity emission is apparent in both of
these molecular tracers. The CO transition suffers absorp­
tion near the LSR velocity (\Gamma3:5 km s \Gamma1 ) with a greater
amount occuring toward redshifted velocities. Since no ab­
sorption is evident in the CS spectrum, the CO absorp­
tion is probably occuring in the cooler and less dense fore­
ground gas.
Fig. 1. The solid line is the CO J=2!1 spectrum and the
dashed line is the CS J=7!6 spectrum toward the H2O maser
position. The blueshifted (­18 to ­8 km s \Gamma1 ) and redshifted (2
to 12 km s \Gamma1 ) velocity ranges correspond to the CO outflow
map in Fig. 2.
The high­velocity CO outflow map is presented in the
upper panel of Fig. 2 overlayed on the K--band mosaic
with the blueshifted wing (­18 to ­8 km s \Gamma1 ) shown in solid
contours and the redshifted wing (2 to 12 km s \Gamma1 ) shown
in dashed contours.
In this map, each lobe has a local emission peak near
the outflow origin and near the outer end. In addition, the
blueshifted and redshifted emission peaks near the origin
are separated by 14\Sigma2 00 along a position angle +44 ffi (i.e.
approximately perpendicular to the axis of the extended
outflow). To explore the velocity structure in more detail,
channel maps of the CO emission are shown in Fig. 3. The
channel maps reveal a more compact outflow structure ori­
ented almost perpendicular to the extended outflow with
the centers of the two components displaced by a few arc­
seconds. Also plotted in Fig. 3 are the suggested axes of

4 Todd R. Hunter et al.: A multiwavelength picture of the AFGL 5142 star­forming region
DECLINATION
(B1950)
RIGHT ASCENSION (B1950)
05 27 30.20 30.15 30.10 30.05 30.00 29.95
33 45 40.5
40.0
39.5
39.0
38.5
38.0
1
2
3
4
5
Fig. 2. Upper panel: 7 0 \Theta7 0 K--band mosaic. The greyscale is logarithmic. The ellipse represent the error box of IRAS 05274+3345.
Superimposed on the greyscale is the CO J=2!1 outflow map with solid and dashed contours indicating blueshifted (­18 to
­8 km s \Gamma1 ) and redshifted (2 to 12 km s \Gamma1 ) emission (2 to 10 by 1 K km s \Gamma1 ). Lower panel: the 8.4 GHz contour map with the
maser positions marked by triangles (crosses mark Torrelles et al. 1992 positions.)

Todd R. Hunter et al.: A multiwavelength picture of the AFGL 5142 star­forming region 5
the two outflows (position angles: +34 ffi and \Gamma47 ffi ) and
the positions of two peculiar NIR sources that will be dis­
cussed in section 3:4. High velocity 13 CO emission was not
detected at the 12 CO peak positions in the outflow lobes,
even when averaged over the entire wing to an integrated
noise level of 75 mK km s \Gamma1 . Since this corresponds to
a factor of at least 80 times lower intensity than in the
12 CO line wings, the 12 CO wing emission is most likely
optically­thin. Higher resolution (20 00 ) followup observa­
tions of the central region in the CO J=3!2 line shown
in Fig. 4 confirm the presence of a compact outflow. In
this transition, which generally traces warmer gas than
the J=2!1 transition, the emission peaks are separated
by 8\Sigma1 00 along a position angle +19 ffi .
Fig. 4. A contour map of the CO J=3!2 emission with thte
blueshifted emission (­18 to ­8 km s \Gamma1 ) shown in solid contours
and the redshifted emission (2 to 12 km s \Gamma1 ) shown in dashed
contours (levels 8 to 64 by 8 K km s \Gamma1 ). The observed posi­
tions are marked by crosses.
3.2. Dense molecular core
With an excitation temperature of 66 K and a critical
density of 3:2 \Theta 10 7 cm \Gamma3 (Green & Chapman 1978),
the CS J=7!6 transition traces warm, dense gas. In
AFGL 5142, the peak of the CS emission occurs at the
center of the compact outflow. Based on our map, the peak
is unresolved, indicating that the high density molecular
gas is concentrated in a core smaller than the beamsize
(20 00 ¸ 0:17 pc). In addition, as shown in Fig. 1, the broad
wings visible in the CS spectrum provide evidence that the
dense gas in the inner region of the outflow is probably fol­
lowing the flow traced at a larger scale by the compact CO
outflow; hence the compact outflow is active also on a very
small scale.
Due to the broadening of the line, the CS linewidth
cannot be used to obtain an accurate value of the
virial mass of the core. Instead, we use the linewidths
of 13 CO J=2!1 and 6!5. Based on the full­widths
half­maximum of the 13 CO J=2!1 and 6!5 lines of
3.48\Sigma0.07 km s \Gamma1 and 4.05\Sigma0.08 km s \Gamma1 , we estimate a
virial mass within (the observed beamsize of) 16.5 00 radius
(0.14 pc) of 400\Sigma20 M fi and within 5.5 00 radius (0.048 pc)
of 180\Sigma20 M fi . Assuming a spherically symmetric density
profile, these two data points are consistent with a profile
of n(r) ¸ r \Gamma2:3 .
3.3. Water masers and radio continuum
The contour map in the lower panel of Fig. 2 shows the 8.4
GHz continuum from the inner core of AFGL 5142. The
observed H 2 O masers are marked by filled triangles and
numbered. We identify a weak continuum barely­resolved
source with peak flux density of 0.50\Sigma0.09 mJy and inte­
grated flux of 0.83\Sigma0.15 mJy at ff = 05 h 27 m 29. s 989; ffi =
+33 ffi 45 0 40. 00 110 nearly aligned with maser spots C3 and
C4. The signal to noise ratio of 6 corresponds to an abso­
lute position uncertainty of \Sigma0.03 00 and confirms the VLA
C­array detection at the same frequency of a 0.69 mJy
source by Torrelles et al. (1992) and the 5 GHz detection
of McCutcheon et al. (1991) and Carpenter et al. (1990).
The importance of our VLA A­array observations is
that they show that all the flux density (within the relative
uncertainties) is contained in a barely­resolved source at
0.3 00 resolution. We thus exclude the presence of a cometary
H ii region (evidenced in Torrelles et al. (1992) only by
the lower contours). The spectral index between 5 and 8.4
GHz is ff ¸1, consistent, within the uncertainty, with that
of an ionized wind (Panagia & Felli 1975). This source was
undetected in our observations at 22 GHz which have an
rms level of 1.9 mJy/beam. Unfortunately, this limit is too
high to confirm or disprove the wind hypothesis, since for
a spectral index ff = 0.6 the expected wind flux density
at 22 GHz would be ¸ 1.2 mJy, i.e. well below the noise.
However, an optically thick H ii region (spectral index ff
= 2) would give an expected flux density of ¸5 mJy, thus
this possibility is less likely.
Given the spectral index and the very small angular
size, we favor the interpretation of the continuum emis­
sion in terms of an ionized stellar wind rather than in
terms of a standard (i.e constant density spherical) UC
H ii region. In fact, the high density (and high internal
pressure) implied in the UC H ii region hypothesis would
predict a very short expansion life­time (unless other con­
fining agents such as blisters, or bow­shocks are present),
thus reducing the chance probability of observing such a
rapidly evolving stage. Instead, ionized stellar winds have
inherent small diameters (Panagia & Felli 1975) and do

6 Todd R. Hunter et al.: A multiwavelength picture of the AFGL 5142 star­forming region
Fig. 3. Channel maps of the CO 2­1 emission wings (one K km s \Gamma1 each) with contours beginning at 0.75 with increments of
0.75 K km s \Gamma1 . The offsets are from: ff = 05 h 27 m 30. s 0 ffi = 33 ffi 45 0 40. 00 0. The position of the NIR sources IRS1 and IRS2 are
plotted for reference. The suggested axes of the two outflows are also marked.

Todd R. Hunter et al.: A multiwavelength picture of the AFGL 5142 star­forming region 7
not suffer the above mentioned shortcoming. In the hy­
pothesis of a radiatively ionized stellar wind the free­free
emission is no longer optically thin and the required stellar
ionizing flux can be much larger than that derived in the
optically thin approximation. Consequently, the spectral
type of the exciting star can be earlier than B2.
The position, flux and isotropic luminosity of the
maser components are listed in Table 1; the corresponding
spectra are presented in Fig. 5, together with a single dish
spectrum taken with the Medicina radiotelescope (Tofani
et al. 1995) two months after the VLA run.
The bulk of the H 2 O emission occurs at velocities near
the quiescent molecular cloud velocity (¸ ­4.1 km s \Gamma1 )
and within the inner boundaries (from ­8 to 2 km s \Gamma1 )
that define the two outflow lobes. No emission is found
outside this interval in the Medicina spectra which cover
a much larger bandwidth than that shown in Fig. 5. This
confirms that the maser formation occurs in the inner re­
gion of the outflow, where the acceleration of the molecular
gas has just begun.
Maser components C3 and C4 are located very close to
the continuum source and on opposite sides of it, at a dis­
tance from the continuum source of 130\Sigma50 and 350\Sigma50
AU, respectively. These two components have broad spec­
tra, similar to the masers embedded in the UCH ii region
W75N (Hunter et al. 1994) and are markedly offset in ve­
locity. C3 extends on the red side of ­2 km s \Gamma1 , while C4
is on the blue side of 0 km s \Gamma1 . This indicates a drastic
change of the velocity pattern in the immediate surround­
ings and on opposite sides of the YSO.
Component C1 is the weakest maser component and
probably was undetected in the Torrelles et al. (1992)
study, possibly because it has a highly variable flux. Com­
ponents C2 and C5, the next two most distant spots from
the YSO (¸2300 and 3900 AU, respectively), have ve­
locities very close to the quiescent molecular cloud and
most probably represent masers moving in the plane of
the sky. They lie near the respective centroids of two Tor­
relles et al. (1992) spots (marked with a cross in Fig. 2),
but show an offset with respect to their positions. The po­
sition offsets between the two epochs (¸ 1:8 yr baseline)
can be interpreted as possible evidence for proper motion.
Both components seem to move outward from the cen­
ter position: C2 has moved southeast at 38\Sigma50 mas yr \Gamma1
while C5 has moved northwest at 49\Sigma50 mas yr \Gamma1 . Clearly
the errors (based on the absolute positional uncertainties
given by Torrelles et al. 1992) are as large as the puta­
tive proper motion and we take this only as an indication
that should be confirmed by further VLA A­array observa­
tions. However, if confirmed these proper motions imply
projected velocities of ¸370 \Sigma300 km s \Gamma1 along a +60 ffi
axis--closer in alignment with the compact outflow than
the extended outflow. Although velocities of this order of
magnitude (– 100 km s \Gamma1 ) have been previously reported
in Orion­KL (Genzel et al. 1981) and in W49N (Gwinn
et al. 1992), they are inconsistent with the observed ve­
locity of the line emission, unless both maser components
happen to be moving very nearly in the plane of the sky.
In both cases components with high proper motions are
found only at large distances from the YSO (for Gwinn
et al. (1992) the acceleration takes place in a 300 AU re­
gion at ¸ 10 4 AU from the YSO), in agreement with what
we find in AFGL5142.
The H 2 O emission is highly variable. For instance the
Medicina spectrum does not show any feature at \Gamma10
km s \Gamma1 , where C4 presents the brightest peak. Conse­
quently, this velocity component must have disappeared
on time scales of only a couple months. On a longer time
scale (years) the information available from the literature
(Verdes­Montenegro et al. 1989, Felli et al. 1992 4 , Brand
et al. 1994, Delin & Turner 1993) indicates that: 1) the
emission at positive velocities had a peak in the late 1980's
and then continuously decreased, 2) in the 1990's the more
stable component was that at ¸ ­5 km s \Gamma1 , and 3) veloc­
ity components between ­10 and 0 km s \Gamma1 appear and dis­
appear with time scales of the order of the sampling time
(approximately one year). This large variability on short
time--scales indicates energetic activity within a few thou­
sand AU around the YSO. An undesirable consequence is
that any direct comparison between maser features and
more permanent structures like the H 2 jets and the CO
outflow is highly unreliable.
3.4. The young stellar cluster
In Fig. 2 the 7 0 \Theta7 0 K--band mosaic is presented. Over 250
sources have been detected in the field, but for only 100 of
them is it possible to estimate all three J, H and K mag­
nitudes. Compared to other regions of the same type (see
for example Palla et al. 1995, Testi et al. 1994), this field
seems to have a low spatial density of sources. However,
it is evident that there exists a central cluster embedded
in a faint nebulosity.
To check quantitatively the clustering around the posi­
tion of the radio continuum source we calculate the source
densities at various radii.
In Fig. 6, the K--Band source density is plotted as a
function of the projected distance from the radio con­
tinuum source. The density is calculated at radii ranging
from 12 00 to 180 00 with increments of 12 00 . At each radius
r i the density is calculated as the number of sources de­
tected between r i\Gamma1 = r i \Gamma 12 00 and r i , divided by the area
A = ú(r 2
i \Gamma r 2
i\Gamma1 ). The density of sources shows an obvious
peak at the central position with a width of 0:3 \Gamma 0:4 pc;
for r ? 0:6 pc the density drops to an almost constant
value ten times lower than that at the center, represent­
ing the background star density. Similar curves calculated
with the central position shifted show reduced maxima
4 the velocity scale of the spectrum published in Felli et al.
1992 and the velocities given in their table 2 for this component
should be diminished by 20 km s \Gamma1 due to an offset of 1.4 MHz
discovered afterwards in the LO

8 Todd R. Hunter et al.: A multiwavelength picture of the AFGL 5142 star­forming region
­20 ­10 0 10 20
0
20
40
v (km/s)
Medicina
01/26/93
AFGL 5142
AFGL5142
DECLINATION
(B1950)
RIGHT ASCENSION (B1950)
05 27 30.2 30.1 30.0 29.9 29.8
33 45 42
41
40
39
38
1
2
3
4
5
1
Fig. 5. The VLA spectra of the brightest maser spots are shown. For comparison the Medicina single dish spectrum taken after
the VLA run is also shown.

Todd R. Hunter et al.: A multiwavelength picture of the AFGL 5142 star­forming region 9
Table 1. AFGL 5142 H2O Maser Components.
Comp. Coordinates (1950) Peak Flux Integ. Flux LSR Vel. (km s \Gamma1 ) Luminosity Dist. from YSO
R.A. Decl. (Jy) (Jy km s \Gamma1 ) vpeak vmin vmax (L fi ) (AU)
1 05 h 27 m 29. s 941 33 ffi 45 0 38. 00 487 0.62 \Sigma 0.09 0.64 \Sigma 0.07 ­8.3 ­9 ­8.3 2.9 \Theta10 \Gamma8 3100
2 05 27 29.985 33 45 38.871 18.8 \Sigma 1.29 35.7 \Sigma 1.09 ­5 ­7.7 ­2.4 1.6 \Theta10 \Gamma6 2300
3 05 27 29.986 33 45 40.196 10.3 \Sigma 0.53 23.4 \Sigma 0.49 ­1.7 ­9 7.4 1.0 \Theta10 \Gamma6 130
4 05 27 30.002 33 45 39.928 23.1 \Sigma 1.16 59.0 \Sigma 1.21 ­9.6 ­10.3 6.7 2.6 \Theta10 \Gamma6 350
5 05 27 30.165 33 45 40.164 7.63 \Sigma 0.39 11.1 \Sigma 0.34 ­5.7 ­7 ­3.1 5.0 \Theta10 \Gamma7 3900
Fig. 6. K--Band source density as a function of the distance
from the radio continuum source. The horizontal scale is the
projected linear radius on the plane of the sky assuming a
distance of 1:8 kpc (1 pc= 115 00 ).
and are broader, implying that the center of the cluster
is within ¸ 10 00 from the position of the radio continuum
source. Therefore, IRS1 lies closer than IRS2 to the center
of the cluster.
The result that young stars with infrared excess tend
to be confined in sub--parsec scale clusters has been
obtained also for other H 2 O--maser/CO--outflow regions
studied, such as the BD+40 ffi 412 (Palla et al. 1995) and
NGC 3576 (Persi et al. 1994). The cluster sizes are 0:35 pc
in AFGL 5142, 0:19 in the BD+40 ffi 412 region and 0:58 pc
in NGC 3576.
Contour plots of the cluster in the three J, H and K
broad bands are shown in Fig. 7, the region covered by
the cluster is 1:5 0 \Theta 1:5 0 wide. In Table 2 absolute posi­
tions and magnitudes of the sources detected in this field
are given. When the quoted magnitude is a lower limit
it is indicated with a ? symbol. Unless otherwise noted,
the quoted photometric errors are 0:1 magnitudes. The
source labelled IRS1 is coincident, within our position un­
certainties, with the radio continuum source detected at
8:4 GHz. The source labelled IRS2 is the brightest source
and nearly coincident with the center of the error ellipse
of the far infrared source IRAS 05274 + 3345.
Fig. 8. (H \Gamma K;J \Gamma H) color--color plot for the sources with good
measurements in the three broad bands. Filled circles represent
sources with infrared excess, as defined in the text, crosses
sources without. Solid line labelled MS marks the position of
unreddened main sequence stars, the two long dashed lines the
position of the reddening belt for main sequence stars. The
short dashed line is the line used to discriminate stars with
and without infrared excess.

10 Todd R. Hunter et al.: A multiwavelength picture of the AFGL 5142 star­forming region
2 8
2 9
3 0
3 1 s
3 2 s
2 7 3 3
m
+ 3 3 4 5
'
o
+ 3 3 4 6
'
o
RA 1950
DEC
1950
H­Band
2 7 s
2 8
2 9 s
3 0
3 1 s
3 2
5 2 7 3 3
h m s
+ 3 3 4 5 0 0 "
o
1 0 "
2 0 "
3 0 "
4 0 "
5 0 "
+ 3 3 4 6 0 0 "
o
1 0 "
2 0 "
RA 1950
DEC
1950
3 2
5 2 7
h m
+ 0 0 "
1 0 "
2 0 "
3 0 "
4 0 "
5 0 "
+ 0 0 "
1 0 "
2 0 "
RA 1950
DEC
1950
J­Band
2 7 s
3 1
3 2
2 7 3 3
h m
+ 0 0 "
2 0 "
4 0 "
+ 0 0 "
2 0 "
RA 1950
DEC
1950
K­Band
1
Fig. 7. From left to right: contour plots of the central cluster in J, H and K and greyscale image of the J \Gamma K color index map.
The lowest contour correspond to 0:01, 0:02 and 0:03 milli--Jansky per square arcsec in the three bands respectively. The other
contours are spaced by 0:01 milli­Jansky per square arcsecond.
In Fig. 8 the (H \Gamma K; J \Gamma H) color--color diagram for
all 100 sources with good measurements is presented. As
discussed in Lada & Adams (1992), this diagram provides
a powerful tool to identify different stages of evolution of
stellar objects, even though the J, H and K near infrared
colors alone may not allow one to draw definitive conclu­
sions on low mass objects (Aspin & Barsony 1994).
We define a line ([J \Gamma H] = 1:75 \Theta [H \Gamma K] \Gamma 0:35)
parallel to the reddening law on the color--color plane in
order to separate qualitatively the sources with infrared
excess from the unreddened main--sequence and reddened
main--sequence stars. This line takes into account 1­sigma
uncertainties in the observed magnitudes. In principle all
the stars that lie on the right side of the reddening belt in
Fig. 8 should have an infrared excess, but the scatter of
the measurements around the main sequence may produce
spurious infrared excess for the sources near the reddening
belt. To avoid this effect all the sources on the right of the
defined line should have a bona--fide infrared excess. Hence
the number of stars with infrared excess that we measure
is a lower limit to the real value in the field. Even with
this restriction, a large fraction of the sources show a near
infrared excess (¸ 29%).
The clustering of sources around the central position
is even more evident when looking only at sources which
show infrared excess. In fact, in the central region the frac­
tion of sources with infrared excess is higher than the mean
value for the entire region. Out of a total of 25 sources
listed in Table 2 with good measurements in all the three
bands, 11 (44%) show infrared excess, with the criterion
described before. IRS1 is the source with the highest color
index, while IRS2, even if it exhibits an infrared excess, is
not the ``reddest'' source of the cluster (see also Fig. 8).

Todd R. Hunter et al.: A multiwavelength picture of the AFGL 5142 star­forming region 11
Table 2. Sources detected in the cluster region.
# ff (1950) ffi (1950) mJ eJ mH eH mK eK name
1 5 : 27 : 27:13 33 : 45 : 28:7 15:7 15:0 14:3
2 5 : 27 : 27:23 33 : 45 : 03:8 13:9 13:0 12:2
3 5 : 27 : 27:55 33 : 45 : 17:8 17:4 (:26) ? 16:5 14:8 (:13)
4 5 : 27 : 27:89 33 : 45 : 48:9 15:3 13:9 13:2
5 5 : 27 : 28:01 33 : 45 : 40:1 12:6 11:6 10:5 IRS2
6 5 : 27 : 28:94 33 : 45 : 46:5 16:7 (:17) 15:1 (:15) 13:7
7 5 : 27 : 29:01 33 : 45 : 16:9 15:1 13:8 13:1
8 5 : 27 : 29:10 33 : 45 : 59:0 16:2 (:11) 15:4 (:19) 14:7 (:28)
9 5 : 27 : 29:15 33 : 45 : 22:3 16:3 14:8 13:9
10 5 : 27 : 29:17 33 : 46 : 05:8 17:1 (:2) 15:5 (:17) 14:8 (:20)
11 5 : 27 : 29:41 33 : 45 : 38:8 16:7 (:13) 14:4 12:9
12 5 : 27 : 29:46 33 : 45 : 28:3 16:8 (:19) 14:8 (:11) 13:4
13 5 : 27 : 29:53 33 : 45 : 47:1 ? 17:5 15:9 (:26) 13:8
14 5 : 27 : 29:71 33 : 45 : 17:8 16:3 14:5 13:6
15 5 : 27 : 29:75 33 : 45 : 56:2 ? 17:5 ? 16:5 14:5 (:16)
16 5 : 27 : 29:80 33 : 45 : 22:9 16:8 (:14) 14:9 13:8
17 5 : 27 : 29:83 33 : 45 : 32:9 14:7 13:9 13:0
18 5 : 27 : 29:91 33 : 45 : 36:2 15:4 13:7 12:9
19 5 : 27 : 29:92 33 : 45 : 40:2 17:5 (:3) 15:1 (:14) 13:2 IRS1
20 5 : 27 : 29:97 33 : 46 : 07:1 16:1 14:7 14:0
21 5 : 27 : 31:09 33 : 45 : 32:6 ? 17:5 15:8 (:15) 14:1 (:12)
22 5 : 27 : 31:17 33 : 45 : 24:1 16:4 15:4 (:12) 14:8
23 5 : 27 : 31:23 33 : 45 : 04:2 17:7 (:2) 16:5 (:23) 15:2
24 5 : 27 : 31:45 33 : 45 : 32:7 16:8 (:12) 15:3 14:5
25 5 : 27 : 31:67 33 : 45 : 50:1 16:5 16:1 (:21) ? 16
26 5 : 27 : 31:70 33 : 45 : 22:2 16:9 (:11) 15:9 (:17) 15:3 (:14)
27 5 : 27 : 32:20 33 : 45 : 41:9 15:9 15:1 14:8
28 5 : 27 : 32:52 33 : 45 : 45:3 14:8 14:4 14:1
29 5 : 27 : 32:81 33 : 45 : 39:7 17:1 (:13) 15:2 14:2
30 5 : 27 : 32:90 33 : 45 : 18:0 16:4 15:5 (:11) 15:2
Note: The limiting magnitudes are 17:5, 16:5 and 16:0 in J, H and K.
The J\GammaK color index map of the cluster region (Fig. 7),
shows that the highest values of the color index are as­
sociated with the IRS1 region. Since both IRS2 and the
diffuse region surrounding it have similar, lower values of
the color index, the diffuse emission is likely to be a re­
flection nebula. Somewhat surprisingly the peak of the
IRAS source does not coincide with the source exhibit­
ing the highest infrared excess (IRS1), the H 2 O masers
and the continuum radio source, but with the brightest
source (IRS2) which also (most probably) illuminates the
diffuse emission. This positional distinction between IRAS
sources and NIR sources with high color excess has been
found in other regions of the same type (Testi et al. 1994).
We conclude that the IRAS source traces the region of the
extended nebulosity, usually associated with the most lu­
minous source.
3.5. Narrow band images
In Fig. 9 contour plots of the Brackett fl images of the
central cluster are shown. On the left is the image before
convolution and subtraction; on the right is the convolved
image with the synthetic continuum subtracted.
Obviously the subtraction procedure has failed to com­
pletely remove the IRS2 source. Unfortunately ARNICA
does not have continuum narrow band filters adjacent to
the lines which would improve the subtraction. In spite of
this fact the subtraction procedure successfully removes
most of the stars that are likely to have no line emission
or absorption detectable; thus the applied algorithm is re­
liable for our purposes.
The Brackett fl image does not show any line emis­
sion feature, even at the position of the radio contin­
uum source (coincident with IRS1), down to ¸ 1:6 \Theta
10 \Gamma15 erg s \Gamma1 cm \Gamma2 arcsec \Gamma2 . For the emission associated
with IRS1, taking into account the continuum subtraction
noise, we can estimate an upper limit to the Brackett fl
flux of ¸ 10 \Gamma14 erg s \Gamma1 cm \Gamma2 . This value can be used to
extimate a lower limit to the extinction to IRS1 in the
case that the radio continuum emission comes from a stel­
lar wind. In fact, in this hypothesis (Simon et al. 1983):
F(Brfl) ¸ 1:6 \Theta 10 \Gamma12 S 8:5 GHz(mJy) erg s \Gamma1 cm \Gamma2 . From

12 Todd R. Hunter et al.: A multiwavelength picture of the AFGL 5142 star­forming region
7 s
2 8
2 9
3 0
3 1
3 2
5 2 3
h m s
+ 3 3 4 5
' 0 "
o
2 0 "
4 0 "
+ 3 3 4 6
' 0 "
o
2 0 "
RA 1950
DEC
1950
2 s
2 8 s
2 9 s
0 s
3 1
3 2
5 2 7 3 3
h m
+ 3 3 4 5 0 0
o
2 0
4 0
+ 3 3 4 6 0 0
o
2 0
RA 1950
DEC
1950 B
A
C
D
E
F
HH­190
1
Fig. 9. Contour plots of the H2 image. On the left: before convolution and continuum subtraction. On the right: after convo­
lution and continuum subtraction. Solid contours: positive values; dashed contours: negative values. The first contour (positive
and negative) corresponds to three times the noise (¸ 1:6 \Theta 10 \Gamma16 erg cm \Gamma2 s \Gamma1 00 \Gamma2 ), all the other contours are spaced by
1:6 \Theta 10 \Gamma16 erg cm \Gamma2 s \Gamma1 00 \Gamma2 . The detected knots are labelled for reference with Table 3. The position of the HH object discov­
ered by Eiroa et al. (1994) is also marked.
the measured radio continuum flux we would obtain
F(Brfl) ¸ 1:3 \Theta 10 \Gamma12 erg s \Gamma1 cm \Gamma2 , hence, for the visual
extinction toward the source we find AV – 50.
On the other hand, the H 2 image shows a rather com­
plex line emission structure. Even before the continuum
subtraction, when compared to the K--Band image of
Fig. 7, a jet--like structure is evident to the north--east
of IRS1. After continuum subtraction a number of other
knots appear more clearly. In Fig. 9 the knots are labelled
from A to F and in Table 3 their position and fluxes are
reported. Since the jet--like structure does not show any
continuum emission, the fluxes of components A and B are
more accurate (8--10%). For all the other clumps the flux
estimates are less accurate due to continuum subtraction
noise (¸ 20%). The F component has a measured flux
that is of the order of three times the noise induced by
the continuum subtraction and hence is only marginally
detected.
The jet--like feature A--B is pointing outward from
IRS1 (see Fig. 10) at a position angle ¸ 46 ffi , and the pos­
sible counter--jet C--D at ¸ \Gamma72 ffi . The eastern component
of the jet is aligned to within few degrees of the posi­
tion angle of the compact outflow discussed in section 4.1,
see also Fig. 3. This almost perfect alignment between the
H 2 --jet and the compact molecular outflow suggests that
they are physically related.

Todd R. Hunter et al.: A multiwavelength picture of the AFGL 5142 star­forming region 13
Table 3. Integrated fluxes of the H2knots.
Knot Coordinates (1950) Dist. a Int. Flux
R.A. Decl. (pc) (erg cm \Gamma2 s \Gamma1 )
A 5 : 27 : 30:91 33 : 45 : 55:6 0:172 2:6 \Theta 10 \Gamma14
B 5 : 27 : 31:60 33 : 46 : 02:6 0:268 5:4 \Theta 10 \Gamma14
C 5 : 27 : 29:41 33 : 45 : 42:4 0:059 3:5 \Theta 10 \Gamma14
D 5 : 27 : 29:02 33 : 45 : 42:3 0:100 4:0 \Theta 10 \Gamma14
E 5 : 27 : 30:17 33 : 46 : 10:6 0:267 1:6 \Theta 10 \Gamma14
F 5 : 27 : 29:74 33 : 45 : 35:7 0:044 3:1 \Theta 10 \Gamma14
a Distances are measured from IRS1
4. Discussion
4.1. The nature of two outflows
The CO J=2!1 channel maps (Fig. 3, clearly indicate a
large scale bipolar outflow but with additional structure in
the perpendicular direction at the origin. There are three
possible interpretations for this smaller scale structure: 1)
a rotating circumstellar disk collimating the large scale
outflow; 2) part of a single poorly­collimated conical out­
flow in which only the limb­brightened edges are seen;
and 3) a more compact outflow, unrelated to the large
scale one. For several reasons, we favor the explanation of
a more compact outflow near the origin of the extended
flow. First, the structure is too large to be considered a
circumstellar disk since the virial mass required to bind it
(M ¸ R¯v 2 =G ¸ 2000M fi ) is nearly three orders of mag­
nitude larger than the measured mass and a factor of 5
larger than the virial mass of the cloud core estimated
from the 13 CO J=2!1 linewidth. Second, the observed
geometry of the lobes--the large difference in their lengths
and their nearly perpendicular alignment--requires a large
opening angle (¸ 90 ffi ) coupled with asymmetric expan­
sion if it is to be explained by a single edge­brightened
outflow. Third, our discovery of a near infrared H 2 jet in
the same direction of the offset and on a scale size of ¸
20 00 (0.17 pc) supports the existence of a more compact
outflow associated with IRS1. Also, the presence of mul­
tiple outflows with different size scales in the same star­
forming region has been reported or speculated in several
cases (e.g. DR21 Garden et al. 1991, and Mon R2 Wolf
et al. 1990). For these reasons, we believe the presence of
two outflows to be the simpler and more likely interpreta­
tion of the CO emission.
Proceeding with the assumption of two distinct out­
flows, the emission in the CO channel maps has been spa­
tially partitioned into two components: an extended out­
flow in the NW/SE direction, and a compact outflow in
the NE/SW direction. Emission from the 73 sampled posi­
tions closest to the NE/SW axis of the map was assigned
to the compact outflow while emission from the rest of
the 182 sampled positions was assigned to the extended
outflow. The column density, mass and energetics of the
outflowing gas were computed using the beam­filling fac­
tor approach detailed by Garden et al. (1991) assuming
CO to be optically­thin based on the 13 CO data. The de­
rived parameters of the outflows are presented in Table 4.
The characteristic velocity, ¯ v, is the intensity­weighted
radial velocity measured from line center (­3.5 km s \Gamma1 )
for the redshifted (­1 to 12 km s \Gamma1 ) and blueshifted (­6
to ­18 km s \Gamma1 ) emission. An estimate of the dynamical
timescale is derived from the physical length divided by
the characteristic velocity. The mass ejection rate is then
computed by dividing the total mass by the dynamical
timescale. In the channel maps nearest the LSR veloc­
ity, the extended outflow shows emission in both lobes,
indicating a large inclination angle. We have estimated
the inclination of the two outflows using the method de­
scribed by Cabrit & Bertout (1990) and computed the
geometrically­corrected values of the physical parameters.
For the extended outflow, spectra at the outer emission
peaks, offsets (­75 00 ,+30 00 ) and (+60 00 ,­90 00 ) relative to the
map center, indicate that the CO line emission is non­
zero over the velocity ranges: ­21.0 to +0.5 km s \Gamma1 , and
­7.0 to +14.0 km s \Gamma1 , respectively. These values yield an
average ratio (R) of 4.69 relative to the LSR velocity (­
3.5 km s \Gamma1 ). Based on the half­power width at the end of
the outflow (31 00 ) and the length (105 00 ), the maximum
opening half­angle of the outflow is estimated to be 8:4 ffi .
In the formula of Cabrit & Bertout (1990), these numbers
imply an inclination of 84:5 ffi (nearly in the plane of the
sky). A similar analysis was carried out for the compact
outflow, but only the spectrum at offset (+15 00 ,+15 00 ) was
considered due to contamination of the blueshifted wing
by the extended outflow. Here the emission extends from
­12.0 to +14.0 km s \Gamma1 for a ratio of 2.06, and based on
the width (43 00 ) and length (40 00 ), the maximum opening
half­angle of the outflow is 28:3 ffi . These parameters im­
ply an inclination of 79:5 ffi . The correction factors applied
in Table 4 are quite large because both outflows present
high inclinations; thus the corrected parameters should
be considered only as approximate values. Based on the
inclination­corrected parameters, it is remarkable that in
spite of the difference in the observed length of the two
outflows (the large scale outflow is nearly a factor of 3
longer than the compact one), the timescales and, in turn,
the mass ejection rates, are nearly identical. This agree­
ment suggests that the basic driving mechanism is the
same for both outflows. The gas in the extended outflow
is simply being ejected at a higher velocity, perhaps be­
cause its origin lies further from the center of the dense
core of the molecular cloud. In support of this interpre­
tation, the axis of the two outflows, as shown in Fig. 3,
seem to indicate that there are two centers displaced by
several arcseconds: the compact outflow is clearly centered
on IRS1, while IRS2 lies almost exactly on the axis of the
extended outflow.

14 Todd R. Hunter et al.: A multiwavelength picture of the AFGL 5142 star­forming region
Table 4. Derived Parameters of the Bipolar Outflows in AFGL 5142.
Wing Component ¯ v Length Timescale Mass Momentum Luminosity Mass Loss
( km s \Gamma1 ) ( 00 ) (pc) (10 5 yr) (M fi ) (M fi km s \Gamma1 ) (L fi ) (10 \Gamma5 M fi yr \Gamma1 )
Red Extended 5.7 120 1.05 1.6 1.6 9.3 0.036 1.0
Blue Extended 4.6 95 0.79 2.0 1.5 6.9 0.017 0.7
Total Extended 5.2 215 1.84 1.8 3.1 16.2 0.053 1.7
Corrected i = 84:5 ffi 54 217 1.86 0.17 3.1 170 5.8 18
Red Compact 5.2 40 0.35 0.66 1.2 6.0 0.053 1.8
Blue Compact 4.6 40 0.35 0.75 1.2 5.5 0.036 1.6
Total Compact 4.9 80 0.70 0.70 2.4 11.5 0.089 3.4
Corrected i = 79:5 ffi 27 82 0.72 0.13 2.4 63 2.7 18
The region around IRS1/IRS2 has been recently ob­
served by Eiroa et al. (1994) in the I broad band (0:9 ¯m)
and in the Hff and S[II] optical narrow bands, revealing the
presence of several red stars and an Herbig--Haro object
(HH 190). The broad band CCD image (Fig. 2 in their pa­
per) shows a cometary shaped reflection nebulosity to the
west of their ``star--1'', which is coincident with our IRS2
source. No diffuse emission is evident on the east side of
the source and no emission is detected at the position of
IRS1 and of the other peaks in the (J \Gamma K) color index map
of Fig. 7. In the narrow band images they detect the pres­
ence of HH 190 located 7 00 to the west of IRS2, but they
do not detect any emission near IRS1 where we found the
H 2 jet--like structure. This is probably due to the fact that
the region around IRS1 is more heavily extincted. The po­
sition of HH 190 is marked in Fig. 9. In the H 2 image there
is a small excess of emission at the position of the HH ob­
ject, but the subtraction noise induced by the presence of
the bright IRS2 source a few arcseconds away makes any
estimate of the flux extremely uncertain.
The comparison between our infrared images and the
optical CCD frames confirms the presence of a cluster of
young embedded sources. The fact that the region around
IRS2 is less extincted than that around IRS1 suggests that
IRS1 lies deeper inside the molecular clump, while IRS2
may be at the edge of it, consistent with the interpreta­
tion from the CO outflow maps. The extended CO out­
flow is aligned with IRS2, suggesting that this young star
could be the powering source of the flow. The compact
outflow is, however, centered at the position of IRS1, and
it is aligned with the H 2 jet axis and the marginally de­
tected H 2 O maser spots proper motion. Thus we believe
the extended outflow has been driven by the processes of
the formation of IRS2, while the compact outflow is being
driven by the IRS1 source.
Regarding the FIR emission detected by IRAS, we ex­
pect that IRS2 dominates the mid infrared emission from
the region, while the most embedded sources are proba­
bly responsible for the bulk of the emission in the 60 and
100 ¯m bands. In order to test this idea we obtained 5
the calibrated IRAS images of the AFGL 5142 region.
The emission from the region is unresolved in all the four
bands, but the peak of the point source seems to be slightly
displaced in the four bands, being near IRS2 at 12 and
25 ¯m, and nearly coincident with the IRS1 position at
100 ¯m. Hence, all the information from the infrared data
seem to tell us that the bright red source IRS2 is indeed
a young object but not the most embedded and youngest
of the region, while the faint NIR source IRS1 is probably
one of the most embedded and youngest sources. The same
conclusion has been obtained with our radio and molecu­
lar data which indicate that the H 2 O masers, the UC ra­
dio continuum source, the peak of the high­density CS gas,
and the center of the compact CO outflow all coincide with
IRS1. It is clearly not possible to establish precisely the
fraction of the FIR luminosity associated with each source,
nevertheless we expect that a consistent fraction is indeed
associated with IRS1. In the following we assume that the
FIR luminosity of IRS1 is of the order of 10 3 L fi . Com­
bining this luminosity with the derived compact outflow
parameters in Table 4, we obtain values for the outflow
mechanical luminosity to stellar radiant luminosity ratio
(– 3 \Theta 10 \Gamma3 ) and the outflow momentum flux to stellar
radiant momentum flux ratio (45) (see eq. 19 and 20 of
Garden et al. 1991). These values are consistent with the
well­documented trends of increasing wind mechanical lu­
minosity and wind force with increasing stellar bolometric
luminosity (Lada 1985). This result suggests that central
driving mechanism of the outflow in AFGL5142 is physi­
cally similar to outflows in other star­forming regions.
5 The IRAS data were obtained using the IRAS data base
server of the Space Research Organisation of the Netherlands
(SRON) and the Dutch Expertise Center for Astronomical
Data Processing funded by the Netherlands Organisation for
Scientific Research (NWO). The IRAS data base server project
was also partly funded through the Air Force Office of Scientific
Research, grants AFOSR 86­0140 and AFOSR 89­0320.

Todd R. Hunter et al.: A multiwavelength picture of the AFGL 5142 star­forming region 15
2 s
8
2 s
3 0
3 1 s
5 2 7 3
h m s
1 "
2 "
3 "
4 "
5 "
+ 3 4 6 0 0
o
1 "
R.A. (1950)
DEC.
(1950)
Fig. 10. AFGL 5142 central cluster. In greyscale is the K--band
image, the contours are the H2 emission, the ellipse the IRAS
Point Source error box, the crosses represent the positions of
the H2O maser spots (the crosses are much larger than the
position uncertainties) and the two concentric circles mark the
peak of the radio continuum emission.
In Fig. 10 we present an overlay of the molecular hy­
drogen image (contours) on the K--band (greyscale). On
the same figure we give also the error box of the IRAS
Point Source (ellipse), and the position of the H 2 O maser
spots and that of the radio continuum peak (white square,
see Fig. 2. The role of IRS1 as the most active source of
the region is evident from the figure. Unfortunately from
the J, H and K broad band data alone it is not possible
to model the spectral energy distribution of the embed­
ded source reliably and to determine the spectral type
of the YSO, because the presence of the NIR excess pre­
vents us from using the observed color indexes to derive
the extinction of the central source. From the two J and
H measurements alone, assuming that in these two bands
the stellar photosphere emission extincted by the circum­
stellar material dominates the measured flux, we would
estimate a visual extinction greater than 15 magnitudes.
Nevertheless it is likely that even at 1:25 ¯m and 1:65 ¯m
the emission of the circumstellar material is important and
this estimate of the extinction is very rough.
4.2. The association of the masers with the compact out­
flow
It is of interest to compare the energies (or the luminosi­
ties) involved in the different phenomena that occur in the
surroundings of IRS1. As already explained it is difficult
to quote a FIR luminosity for IRS1; a rough estimate of ¸
10 3 L fi can be used as indicative. The luminosity required
to create the maser spots can be derived from the J­type
shock collisional pumping model of Elitzur et al. (1989).
Using their equation 4.5 the mechanical luminosity of the
wind needed to create the conditions for 5 maser spots is
of the order of 70 L fi . This value is rather uncertain since
it depends on several model parameters (e.g. the size and
aspect ratio of the maser) which are totally unknown, and
is also independent of distance. However, it should be cor­
rect as an order of magnitude. A mechanical wind lumi­
nosity between 5 to 10% of the total FIR luminosity tells
us that a considerable fraction of the stellar energy must
go into the shock compression to create the masers. Most
of this energy will be reprocessed into FIR continuum by
heated dust. In fact, the mechanical luminosity of the out­
flow is only 2.7 L fi . This means that only 4% of the energy
needed to produce the maser goes into the acceleration of
the gas in the molecular outflow lobes, with a rather low
efficiency, while the rest is dissipated into heat. The de­
tection of a high excitation molecule such as CS J=7!6
with Tmb = 8 K may suggest high kinetic temperature
in the inner core, as also indicated by the high temper­
ature found in the NH 3 line (Estalella et al. 1993). From
the H 2 fluxes of Table 3 the integrated luminosity of the
H 2 jets in the S(1) 1 ! 0 line is ¸0.02 L fi . Interestingly,
this value is two orders of magnitude smaller than the me­
chanical luminosity of the CO outflow; since the jet is a
factor of 5 smaller in size, the outflow and the jet contain
approximately equal mechanical luminosity per unit vol­
ume. Finally, the luminosity of all the H 2 O masers is 5.7
10 \Gamma6 L fi . The ratio of this value with the total FIR lumi­
nosity is ¸ 5 10 \Gamma9 , in very good agreement with the aver­
age value found by Felli et al. (1992) from a large sample.
However, the present analysis points out that only a small
fraction of the total FIR luminosity is directly related to
the creation of the masers and of the H 2 jets­molecular
outflow. The correlation found between maser luminosity
and total FIR luminosity over a large range of luminosi­
ties suggests that this fraction should not vary appreciably
with the luminosity. The ratio of H 2 O maser luminosity
to mechanical luminosity of the CO outflow is ¸ 2.1 10 \Gamma6 ,
nearly identical to the average value found by Felli et al.
(1992) for a large sample of regions.
Although there is no simple velocity pattern observed
across the more distant maser components (C1, C2 and
C5), they lie along an axis that is roughly parallel to the
jet observed in H 2 and to the compact bipolar molecu­
lar outflow. The proper motions suggested for C2 and C5
also lie outward along this axis. To summarize, the phys­
ical picture that emerges from these observations, is that
the ionized stellar wind (detected in the 8.4 GHz con­
tinuum) ultimately generates shocks in the gas surround­
ing the YSO (IRS1) where the density is high enough (as
traced by the CS J=7!6 transition) to collisionally­pump
the H 2 O maser line, while at the same time drives the
highly­collimated H 2 jet and the compact bipolar molec­
ular outflow seen at larger distances. The picture might
be different for C3 and C4, the ones within a few hundred
AU of the YSO, where we may see the interaction of the
accreting disk material with the stellar wind.

16 Todd R. Hunter et al.: A multiwavelength picture of the AFGL 5142 star­forming region
5. Conclusions
Our new observations provide a clearer picture of the star
forming region AFGL 5142 and a better comprehension
of the effects generated by the interaction of the energy
sources (the YSOs) with the surrounding molecular envi­
ronment (i.e. the water masers, the regions of shocked H 2
and the CO molecular outflows).
A cluster of IR sources is found in the center of the
region, with a radius of 0.3 pc and with a K­band source
density about 10 times that over the rest of the observed
field. Many of the sources in the cluster show a strong IR
excess, typical of YSOs still surrounded by their parent
dust/molecular cloud. Two of the cluster members, sepa­
rated by about 30 00 , have received more attention: IRS1,
which has the strongest IR excess and IRS2, the brightest
member at K band (2.7 magnitudes more luminous than
IRS1). IRS2 seems to be at the center of an extended and
well collimated CO outflow. It also coincides with an IRAS
point source. Several pieces of evidence suggest that IRS2
may represent a YSO in an advanced phase of evolution:
there are no traces of UC H ii regions or of maser emission
in IRS2, an HH object has been found in its proximity, and
it lies away from the peak of the high density molecular
tracers. All these facts imply that the region of the molecu­
lar cloud from which IRS2 formed has been dispersed, the
extinction is smaller and that the only remnant of the star
formation process is the large scale collimated outflow.
In contrast, IRS1 possesses all the characteristic as­
pects of a YSO in an early evolutionary phase. It is at the
center of a compact CO bipolar outflow, it coincides with
an almost unresolved thermal continuum source (most
probably an ionized wind), it is at the center of a cluster of
H 2 O masers (two of which are within a few hundred AU
and two others with indication of proper motion in the
outward direction), it coincides with a small and dense
molecular cloud core and it is at the origin of H 2 jets.
The H 2 O maser emission occurs only in the immediate
surroundings of the YSO, where a substantial fraction of
the YSO energy required for their formation is available.
This tight spatial connection between masers and YSOs
is on the same lines of that found by Testi et al. (1994)
in a large sample of H 2 O masers. The maser velocities are
always much lower than those of the two outflow lobes,
indicating that they occur in the region where the accel­
eration begins. No maser formation occurs either in the
region where the H 2 is shock excited, or at the outer edges
of the outflow where the high velocity molecular gas hits
the surrounding material, most probably because there is
no longer sufficient energy to create J­type shocks.
Higher resolution, single dish and interferometric
molecular line observations will be important to explore
this dense molecular core as well as to better separate
the two bipolar outflows. Furthermore, additional high­
resolution H 2 O maser observations and single dish studies
are needed to confirm the proper motions implied by the
two epochs analyzed here and to study the dependence of
variability on the distance of the maser from the YSO.
Acknowledgements. We thank the CSO staff for their ongo­
ing assistance during this observing project and offer our deep
though posthumous appreciation to Larry Strom for his assis­
tance at the summit over the past few years. GBT acknowl­
edges support from NSF under grant AST­9117100.
References
Aspin, C. & Barsony, M., 1994, A&A, 288, 849
Brand, J., Cesaroni, R., Caselli, P., Catarzi, M. & Codella, C.
et al. 1994, A&AS, 103, 541
Braz, M.A., Lepine, J.R.D., Sivagnanam, P. & Lesqueren, A.M.
1990, A&A, 236, 479
Cabrit, S. & Bertout, C. 1990, ApJ, 348, 530
Carpenter, J.M., Snell, R.L. & Schloerb, F.P. 1990, ApJ, 362,
147
Delin, Xiang & Turner, B. 1993, NRAO preprint 93/169
Eiroa, C., Torrelles, J.M., Miranda, L.F., Anglada, G. & Es­
talella, R. 1994, A&AS, 108, 73
Elitzur, M., Hollenbach, D.J., & McKee, C.F. 1989, ApJ, 346,
983
Estalella, R., Mauersberger, R., Torrelles, J.M., et al. 1993,
ApJ, 419, 698
Felli, M., Palagi, F. & Tofani, G. 1992, A&A, 255, 293
Garden, R.P., Hayashi, M., Gatley, I., Hasegawa, T. & Kaifu,
N. 1991, ApJ, 374, 540
Genzel, R. & Downes, D. 1977, A&AS, 30, 145
Genzel, R., Reid, M.J., Moran, J.M. & Downes, D. 1981, ApJ,
244, 884
Green, S., & Chapman, S. 1978, ApJSS, 37, 169
Gwinn, C.R., Moran, J.M. & Reid, M.J. 1992, ApJ, 393, 149
Henning, T., Cesaroni, R. & Walmsley, M. 1992, A&AS, 93,
525
Hunt, L.K., Maiolino, & R., Moriondo, G., 1994a, Technical
Report 2/94, Arcetri Astrophysical Observatory
Hunt, L.K., Maiolino, R., Moriondo, G., & Testi, L., 1994b,
Technical Report 3/94, Arcetri Astrophysical Observatory
Hunt, L.K., Testi, L., Borelli, S., Maiolino, R., & Moriondo, G.,
1994c, Technical Report 4/94, Arcetri Astrophysical Obser­
vatory
Hunter, T.R., Taylor, G.B., Felli, M., & Tofani, G. 1994, A &
A, 284, 215
Hunter, T.R., et al. 1995, in preparation
Kooi, J.W., Walker, C.K., Leduc, H.G., Schaffer, P.L., Hunter,
T.R., Benford, D.J., & Phillips, T.G. 1994, Int J IR&mm
Waves, 15, 477
Lada, C.J. 1985, ARAA, 23, 267
Lada, C., & Adams, F.C., 1992, ApJ, 393, 278
Lisi, F., Baffa, C., & Hunt, L.K., 1993, in SPIE Vol. 1946 ``In­
frared Detectors and Instrumentation'', pp. 594
McCutcheon, W.H., Dewdney, P.E., Purton, C.R. & Sato, T.
1991, AJ, 101, 1435
Matveyenko, Diamond, P., Graham & Junor 1993, p. 176 of
Astrophysical Masers: proceedings of a conference held in
Arlington, VA, 9­11 March 1992, ed. A.W. Clegg, Berlin:
Springer­Verlag
Palla, F., Testi, L., Hunter, T.R., Taylor, G.B., Prusti, T., et
al. 1995, A&A, 293, 521

Todd R. Hunter et al.: A multiwavelength picture of the AFGL 5142 star­forming region 17
Panagia, N. & Felli, M. 1975, A&A, 39, 1
Persi, P., Roth, M., Tapia, M., Ferrari--Toniolo, M., & Marenzi,
A.R. 1994, A&A, 282, 474
Shu, F.H., Ruden, S.P., Lada, C.J. & Lizano, S. 1991, ApJ,
370, L31
Simon, M., Felli, M., Cassar, L., Fischer, J. & Massi, M. 1983,
ApJ, 266, 623
Snell, R.L., Huang, Y.L., Dickman, R.L., & Claussen, M.J.
1988, ApJ, 325, 853
Testi, L., 1993, Technical Report 10/93, Arcetri Astrophysical
Observatory
Testi, L., Felli, M., Persi, P., & Roth, M. 1994, A&A, 288, 634
Tofani, G., Felli, M., Taylor, G.B. & Hunter T.R. 1995, A&AS,
in press
Torrelles, J.M., Gomez, J.F., Anglada, G., Estalella, R.,
Mauersberger, R., & Eiroa, C. 1992, ApJ, 392, 616
Verdes­Montenegro, L., Torrelles, J.M., Rodriguez, L.F.,
Anglada, G., Lopez, R., Estalella, R., Canto, J. & Ho,
P.T.P. 1989, ApJ, 346, 193
Wolf, G.A., Lada, C.J. & Bally, J. 1990, AJ. 100, 1892
Wright, M.C.H., Carlstrom, J.E., Plambeck, R.L. & Welch,
W.J. 1990, AJ, 99, 1299
This article was processed by the author using Springer­Verlag
L A T E X A&A style file L­AA version 3.