Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://www.arcetri.astro.it/~lt/preprints/bd40_fig_aa/bd40_fig_aa.ps.gz Äàòà èçìåíåíèÿ: Tue Sep 11 16:56:41 2007 Äàòà èíäåêñèðîâàíèÿ: Sat Dec 22 08:45:54 2007 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: cygnus

A&A manuscript no.
(will be inserted by hand later)
Your thesaurus codes are:
3(08.16.5; 13.09.6; 08.09.2 BD+40 ffi 4124)
ASTRONOMY
AND
ASTROPHYSICS
20.6.1994
The active source in the region of the Herbig star
BD+40 ffi 4124
F. Palla 1 , L. Testi 2 , T.R. Hunter 3 , G.B. Taylor 3 , T. Prusti 4 , M. Felli 1 , A. Natta 1 and R.M. Stanga 2
1 Osservatorio Astrofisico di Arcetri, Largo Enrico Fermi 5, I­50125 Firenze, Italy
2 Dipartimento di Astronomia e Scienza dello Spazio, Universit'a di Firenze, Largo E. Fermi, 5, I­50125 Firenze, Italy
3 Department of Physics, Mathematics & Astronomy, Caltech, Pasadena, CA, 91125, USA
4 Astrophysics Division, Space Science Department, ESTEC, Postbus 299, NL­2200 AG Noordwijk, The Netherlands
Received date; accepted date
Abstract. We present new near­infrared, millimeter
and radio observations of the region associated with
BD+40 ffi 4124, a Herbig Be star located in the direction
of the Cygnus arm. BD+40 ffi 4124 is the optically bright­
est member of a small group of young emission line stars,
including V 1318 Cygni and V 1686 Cygni. Near­IR broad­
band imaging of V 1318 Cygni reveals the presence of two
distinct stellar components oriented north­south, a result
recently reported by Aspin et al. (1994). On a larger scale,
the near­IR images also indicate the presence of many
highly embedded sources concentrated in the vicinity of
the bright visible stars. In the central part of the cluster,
the near­IR sources outnumber the optically visible ones
by a factor of three.
The distribution of the high density molecular gas
traced by the CS J=5!4 emission is highly concentrated
around V1318 and V1686 Cygni, while the total column
density of gas traced by the C 18 O J=2!1 shows a ridge
morphology centered on these sources. The Herbig star
BD+40 ffi 4124 lies at the edge of both structures. From the
otpically thin C 18 O emission, we derive a total molecu­
lar mass of M core =280 M fi . A molecular outflow has been
detected in CO J=2!1. The high velocity gas is confined
to a compact region of size ú 20 00 , corresponding to a
physical length of 0.1 pc. The outflow is not well resolved,
or collimated, on the observed scales. We have also de­
tected H 2 O maser emission at 22.2 GHz at a position co­
incident with the center of the bipolar outflow. Both the
maser and the outflow are located at the position of the
southern source V 1318S which we identify as the source
responsible for the activity observed in the BD+40 region.
Key words: Stars: pre­main sequence -- Stars: individ­
ual: BD+40 ffi 4124 -- Infrared: stars -- masers -- molecular
processes -- interstellar medium: molecules
Send offprint requests to: F. Palla
1. Introduction
There is a strong tendency for stars not to form in isola­
tion. Clusters, groups, and associations with varying de­
gree of richness appear to be the rule rather the exception
in most star forming regions (Zinnecker et al. 1993; Lada
et al. 1993). This property reflects a basic outcome of the
star formation process: while the majority of low­mass pre­
main­sequence (PMS) stars are found in loose groups, the
presence of a massive star is generally accompanied by
large ensembles of lower mass stars. In this scheme, the
intermediate­mass Herbig Ae/Be stars (HAEBEs) should
represent the transitory class in which clustering effects
begin to occur at a significant level. Observational evi­
dence that this is indeed the case is now rapidly accu­
mulating, as a result of several optical and near­infrared
surveys (Barsony et al. 1991; Hillenbrand 1994; Testi et
al. 1994).
Herbig Ae/Be stars also show evidence of mass loss,
whose origin is still a matter of debate. These flows are
generally attributed to the activity of the HAEBE star it­
self, but the presence of other stellar sources in the same
area makes this association uncertain. The recent discov­
ery at 10 ¯m of a deeply embedded source near the Herbig
star Lk Hff 198 (Lagage et al. 1993) as the likely driving
source of the associated molecular outflow clearly illus­
trates the point.
In this paper, we concentrate our attention on the
small stellar cluster associated with BD+40 ffi 4124 (j
V 1685 Cygni j MWC 340), a Herbig Be star located at
a distance of 1 kpc in the direction of the Cygnus arm.
The group consists of two other optically visible, emission
line stars, V 1686 Cygni (jLkHff224) and V 1318 Cygni
(jLkHff 225) (Cohen 1972; Strom et al. 1972), plus several
embedded IR sources whose nature is still under scrutiny

2 Palla et al.:
(Testi et al. 1994). BD+40 ffi 4124 is the brightest member
of the stellar group. In a recent study, Aspin et al. (1994)
have discovered that the star V 1318 Cygni is resolved at
optical and near­infrared wavelengths in two distinct stel­
lar components plus a nebulous knot. One of these sources,
called V 1318S, is sufficiently luminous (about 1600 L fi )
to qualify as a candidate HAEBE star (Th'e et al. 1994).
Our interest in this region was triggered by the de­
tection of strong H 2 O maser emission at 22.2 GHz, dur­
ing a survey for masers associated with HAEBE stars
(Palla & Prusti 1993). The emission was rather unusual,
since it presented two components separated in velocity
by more than 100 km s \Gamma1 . Such extreme velocity differ­
ences are rare even in the case of the most powerful known
H 2 O sources associated with high luminosity HII regions.
We suggested that the maser was not associated with
BD+40 ffi 4124, but was rather due to some as yet unidenti­
fied molecular flow excited by a different source. However,
previous searches for CO outflows (Cant'o et al. 1984) and
HH objects (Poetzl et al. 1992; Goodrich 1993) associated
with BD+40 ffi 4124 had given negative results.
We present below new data at near­infrared, millime­
ter and radio wavelengths that clarify the properties of the
complex associated with BD+40 ffi 4124 (hereafter referred
to as BD+40 region) and the origin of the mass­loss activ­
ity. In particular, the high resolution VLA observations of
the maser emission show that its position coincides with
that of the bright NIR source V 1318 S. Also, millimeter
CO observations reveal the presence of a compact bipo­
lar molecular outflow centered at the same position. The
whole region is embedded in a dense molecular core of
about M core ¸ 280M fi , with the star BD+40 ffi 4124 lying
near the edge of the cloud.
2. Observations
2.1. Near infrared data
The NIR observations were carried out on 30 Septem­
ber 1993 at the Gornergrat observatory, using the Arcetri
near--infrared camera, ARNICA, on the 1:5 m TIRGO
telescope. The ARNICA detector is a NICMOS3 256 \Theta 256
pixel HgCdTe array developed by Rockwell International
(Lisi et al. 1993 and Hunt et al. 1994a,b). The scale on
the detector is roughly 0.95 00 /pixel, and the field of view
of a single frame covers about 4 0 \Theta 4 0 . The BD+40 region
was observed in the J, H, and K broad­band filters. For
each filter fourteen overlapping frames were taken with the
central 1 0 \Theta 1 0 centered on BD+40 ffi 4124 in all frames. Ab­
solute flux calibration was achieved by observing the stan­
dard star SAO070237 (spectral type A0, K­mag= 8:9). We
estimate a calibration accuracy of about 10%. The data
reduction and analysis were performed using the ARNICA
(Hunt et al. 1994c) and IRAF software packages. Sky sub­
traction and flat fielding were performed using median av­
erage images of the fourteen frames.
After the preliminary reduction, the images were regis­
tered and combined in order to produce a ¸ 7 0 \Theta 7 0 mosaic
centered on BD+40 ffi 4124. Because of the observing tech­
nique, the limiting magnitude varies in the mosaics, being
higher in the central region and poorer at the edges of
all images. Accurate photometry was performed using the
DAOPHOT routines. The 5oe limiting magnitudes in a 4
arcsec aperture obtained in the central region were J=18,
H=17 and K=17, while at the edge of the mosaic J=17.3,
H=16.1 and K=16.0.
2.2. VLA water masers and radio continuum
The VLA observations of the BD+40 ffi 4124 region took
place on 1992 November 24 as part of a larger campaign
targeting water masers associated with CO outflows (To­
fani et al. 1994). The BD+40 region was observed for 6
minutes in line mode 2AD at 1.3 cm, and for 5 minutes
in continuum at 3.6 cm. The velocity resolution is 0.66
km s \Gamma1 and the observations with the A IF cover the veloc­
ity range \Gamma51 to \Gamma88 km s \Gamma1 . The D IF covers a 25 MHz
bandwidth, offset 25 MHz from the line emission in or­
der to obtain 1.3 cm continuum data simultaneously with
the line observations. The observations were made with
the VLA in the most extended configuration (A) resulting
in a synthesised beam of 0.11 00 \Theta 0.10 00 at p.a. \Gamma67 ffi and
0.30 00 \Theta 0.25 00 at p.a. \Gamma78 ffi at 1.3, and 3.6 cm respectively.
3C48 and 3C84 were observed for the purpose of
absolute flux density and bandpass calibration. Follow­
ing phase calibration using the nearby continuum source
2005+403, the strongest channel of the maser was then
mapped and cleaned using the AIPS task MX. A single
component model was used to self­calibrate this reference
channel and the resulting solutions were applied to all
channels. The dynamic range achieved in the reference
channel was 2000. The self­calibration solutions from the
reference channel were also applied to the 1.3 cm contin­
uum.
2.3. Caltech Submillimeter Observatory CO, C 18 O, and
CS observations
Our millimeter observations were made at the Caltech
Submillimeter Observatory (CSO) on Mauna Kea. Us­
ing the facility 230 GHz SIS receiver, we observed the
CO J=2!1 transition at 230.54 GHz, C 18 O J=2!1 at
219.5604 GHz, and the CS J=5!4 transition at 244.9356
GHz. The 1024­channel 50 MHz acousto­optical spectrom­
eter (AOS) was used as the backend except in the case of
the CO line in which the broader 500 MHz AOS was used.
For each transition, the telescope was driven across the
source in the ``on­the­fly'' mapping mode which increases
the duty cycle by observing the off­position less frequently,
but for longer integrations. Since the CSO beamsize is
30 00 at the measured frequencies, we drove the telescope at
a rate of 3 00 per second while integrating for five seconds

Palla et al.: 3
and thus sampling at regular 15 00 intervals to obtain full
spatial coverage. In the case of the weaker CS and C 18 O
lines, several individual maps were observed and summed
in order to raise the signal to noise ratio of the observa­
tions. Total on­source integration times of 20 and 30 sec­
onds per point were obtained for the C 18 O and CS maps,
respectively. Overall calibration was performed using the
standard chopper wheel method. Based on the measured
CSO main beam efficiency of 76% at 230 GHz, the spectra
are presented on a main beam temperature, TMB , scale.
We estimate our calibration to be accurate to within 30%
and our pointing to within 3 00 .
3. Results
3.1. The near infrared cluster
1
Fig. 1. K­band image of the central 1 0 \Theta 1:5 0 , covering the
extent of the diffuse emission. The three emission line stars,
BD+40 ffi 4124, V 1686 and V 1318 are labelled A, B, and C,
respectively. Axes coordinates are Right Ascension and Decli­
nation for the 1950.0 epoch. The greyscale is logarithmic. A
contour plot of the V 1318 N & S complex is shown in the in­
set to the left. The designation of the two sources V 1318 N
and V 1318 S follows Aspin et al. (1994). The cross marks the
position of the H2O maser source.
The central 1 0 \Theta 1:5 0 part of the K­band mosaic is pre­
sented in Figure 1. BD+40 ffi 4124 and V 1686 Cyg are both
saturated in this image. Overall, 33 sources have been de­
tected in the NIR bands, three times more than at optical
wavelengths (see Figure 1 of Aspin et al. 1994). All the
near infrared point sources detected in this field are listed
in Table 1. For each source we report a progressive num­
ber, the absolute position and the J, H and K magnitudes.
The quoted photometric errors are 0.1 magnitude. For the
weaker sources, the photometric errors are given in paren­
theses in Table 1. Note that BD+40 ffi 4124 is saturated at
each band, while V 1686 Cyg only at K­band.
Absolute position calibration was obtained by a least
square fitting technique to solve for the plate constant,
using the position of a number of stars listed in the Hubble
Space Telescope Guide Star Catalogue (see Testi 1993 for
a complete description of the procedure). The accuracy in
Right Ascension and Declination is estimated to be better
than one arcsecond. Due to saturation problems, we could
not determine the coordinates of BD+40 ffi 4124 and those
given in Table 1 are from the Herbig & Bell Catalogue
(1988).
A comparison with the optical plate of Strom et al.
(1972) and the CCD images presented by Aspin et al.
(1994) and Goodrich (1993) shows that the optical nebu­
losity around V 1318 becomes brighter than that around
BD+40 ffi 4124 in the infrared. The (J \Gamma K) colour index
map of the same region of Figure 1 is presented in Fig­
ure 2. The darkest colours correspond to higher values of
the colour index. There is a clear increase in the colour
index at the position of the sharp edge in the diffuse op­
tical emission detected by Aspin et al. (1994) in the R­
band and by Goodrich (1993) in [SII]. This would imply
that the sources associated with V 1318 are embedded in a
denser envelope and that they are probably younger than
BD+40 ffi 4124 itself.
A contour plot of the emission at K­band from the
V 1318 region is also shown in Figure 1. The two sources
V 1318N & S are clearly resolved, but there is no evidence
of the third nebulous component detected by Aspin et al.
(see their Figure 3), probably due to our lower spatial reso­
lution. V 1318S is the source of the cluster with the largest
infrared excess (H­K=2.6 and J­H=1.3, respectively) and
it is probably the youngest and more massive member of
the cluster. As we shall see below, this source is likely re­
sponsible for the activity discovered by us (the H 2 O maser,
the CO outflow and the hot molecular gas).
The large scale, 7 0 \Theta 7 0 , K­band emission centered
on BD+40 ffi 4124 is shown in Figure 3. In this mosaic
more than 460 sources have been detected down to
a signal to noise ratio of about five. Apart from the
BD+40 ffi 4124 group, one of the most interesting objects
of the field is the very bright star at ff 1950 =20:18:52.21
and ffi 1950 =41:11:41.6. This object is almost undetectable
in the Palomar red plate (see Strom et al. 1972), is sat­
urated in our H and K bands and has a J magnitude of
10.54 (in the J­band image). As pointed out by Strom et
al. (1972), values of (V \Gamma I) ? 7 and (V \Gamma J) ? 9:5 can
be explained only in terms of: 1) a foreground M dwarf or
a background M giant, 2) a highly reddened OB star, or
3) a deeply embedded PMS star. Even though our data
are not sufficient to draw a definite conclusion, the lack
of molecular gas emission toward the direction of the star

4 Palla et al.:
Table 1. NIR sources of the BD+40 ffi 4124 cluster.
# ff (1950:0) ffi (1950:0) mJ mH mK Remarks
1 20 : 18 : 40:49 41 : 11 : 27:9 ? 18 16:9 (:14) 16:6 (:15)
2 20 : 18 : 40:66 41 : 12 : 23:1 ? 18 15:4 14:2
3 20 : 18 : 41:05 41 : 11 : 25:0 ? 18 16:7 (:16) 15:9
4 20 : 18 : 41:07 41 : 12 : 24:5 ? 18 ? 17 14:8 (:20)
5 20 : 18 : 41:25 41 : 11 : 58:4 ? 18 15:9 15:2
6 20 : 18 : 41:34 41 : 12 : 34:1 17:0 15:7 15:3
7 20 : 18 : 41:86 41 : 12 : 26:0 15:1 (:20) 14:1 (:13) 13:3 (:18)
8 20 : 18 : 41:96 41 : 12 : 00:4 12:0 12:2 11:8
9 20 : 18 : 41:98 41 : 11 : 53:5 14:0 13:3 12:6
10 20 : 18 : 42:5 41 : 12 : 20:0 BD+40 ffi 4124
11 20 : 18 : 42:64 41 : 13 : 01:1 ? 18 ? 17 16:1 (:18)
12 20 : 18 : 42:83 41 : 12 : 41:5 16:1 14:3 12:9
13 20 : 18 : 42:89 41 : 11 : 30:2 ? 18 16:3 15:1
14 20 : 18 : 42:92 41 : 12 : 24:6 14:8 (:24) 12:9 (:20) 12:3 (:18)
15 20 : 18 : 43:02 41 : 12 : 11:7 13:3 12:9 11:7 (:14)
16 20 : 18 : 43:17 41 : 12 : 50:7 ? 18 ? 17 15:1 (:17)
17 20 : 18 : 43:19 41 : 12 : 24:7 ? 18 ? 17 13:3 (:25)
18 20 : 18 : 43:33 41 : 12 : 02:8 14:7 13:4 12:8
19 20 : 18 : 43:42 41 : 13 : 04:2 ? 18 16:9 (:16) 15:9
20 20 : 18 : 43:72 41 : 11 : 54:9 11:6 9:6 V 1686
21 20 : 18 : 43:85 41 : 11 : 37:2 17:9 (:18) ? 17 ? 17
22 20 : 18 : 43:94 41 : 12 : 11:9 15:7 14:3 13:3 (:15)
23 20 : 18 : 43:96 41 : 12 : 22:2 13:7 12:6 11:8
24 20 : 18 : 44:31 41 : 12 : 57:0 ? 18 ? 17:3 16:2 (:14)
25 20 : 18 : 44:89 41 : 11 : 57:4 14:0 11:5 9:6 V 1318 N
26 20 : 18 : 44:91 41 : 11 : 53:1 13:9 12:6 10:0 V 1318 S
27 20 : 18 : 44:97 41 : 12 : 23:1 15:1 14:1 13:4
28 20 : 18 : 45:21 41 : 11 : 50:2 15:8 (:18) 14:9 (:16) ? 17
29 20 : 18 : 45:30 41 : 12 : 47:0 17:0 (:12) 16:1 15:3 (:14)
30 20 : 18 : 45:53 41 : 11 : 54:4 16:0 (:13) 14:9 (:15) 13:0 (:25)
31 20 : 18 : 45:82 41 : 11 : 41:1 ? 18 15:0 13:2
32 20 : 18 : 46:07 41 : 12 : 39:8 16:2 13:9 12:3
33 20 : 18 : 46:38 41 : 12 : 23:9 ? 18 16:6 (:12) 15:7
(see below) seems to exclude the second and third possi­
bilities.
In Figure 4, we plot the location of the sources detected
in the large field in the (J \Gamma H,H \Gamma K) color--color plane. We
include only sources with detections in all the NIR bands.
In order to characterize the distribution of young sources,
we arbitrarily define sources with infrared excess as those
which lie in the color­color plane to the right of the line
defined by (J \Gamma H) ! 1:75 \Theta (H \Gamma K) \Gamma 0:35. It is a well
known result that the location of young stellar objects on
the JHK diagram is determined by their evolutionary sta­
tus. For example, Lada & Adams (1992) and Hillenbrand
et al. (1992) have shown that Herbig Ae/Be stars, as well
as luminous protostars, tend to lie well to the right of the
color­color relationship for main­sequence reddened stars,
and they have interpreted the observed near­infrared ex­
cesses in terms of emission from circumstellar disks with
inner holes. Unlike these objects, the colors of the ma­
jority of classical T Tauri stars are more similar to those
of heavily reddened main­sequence stars and can be ac­
counted for by standard disk models (e.g. Adams et al.
1988). Although the disk interpretation has been ques­
tioned in the case of the Herbig stars (Hartmann et al.
1993), it is clear that measurements of broad­band indices
are a powerful tool for identifying young stellar objects.
We also note that a significant number of sources lie to the
left of the upper reddening vector, in a region that should
be devoid. Several effects, such as atypical reddening laws,
the presence of extended circumstellar envelopes or pho­
tometric errors, can be invoked to explain the location of
the sources (see the extensive discussion in Aspin & Bar­
sony 1994), but the uncertainty on which is the dominant
one still remains.
In our case, 40 out of the 177 sources (23%) plotted
in figure 4 show an IR excess. This percentage increases
drastically if we consider only the sources found in the
central 1 0 \Theta 1:5 0 . Then, 10 of the 17 sources detected at
J, H and K (59%) do show an excess. This indicates the
strong tendency of the youngest sources to be found in the
immediate vicinity of the Herbig star. Of course, follow up

Palla et al.: 5
Fig. 3. K­band mosaic of the large scale emission, 7 0 \Theta 7 0 , centered around BD+40 ffi 4124. The corners of the field are not
covered by the mosaic. The greyscale is logarithmic.
spectroscopic and photometric observations are required
in order to ascertain their true evolutionary status. In­
terestingly, the source with the largest infrared excess in
figure 4 is V 1318 S.
3.2. The H 2 O maser
The H 2 O maser spectrum, shown in Figure 5, dis­
plays a smooth, symmetric profile with an integrated
flux of 59.1 Jy km s \Gamma1 and a peak velocity VLSR =
\Gamma80:5 km s \Gamma1 . The position of the high velocity H 2 O
maser (ff 1950 =20:18:44.90; ffi 1950 =41:11:53.00) is clearly
not coincident with BD+40 ffi 4124, although it is at the
apparent center of the newly discovered molecular outflow
(see below). The velocity interval covered by the VLA ob­
servations was chosen based on the results of the Medicina
observations which showed multiple emission features and
large flux variability, with the velocity of the most intense

6 Palla et al.:
4 1 s
4 2 s
4 3 s
4 4 s
4 5 s
2 0 1 8 4 6
h m s
2 0 "
4 0 "
+ 4 1 1 2
' 0 0 "
o
2 0 "
4 0 "
+ 4 1 1 3
' 0 0 "
o
RA 1950
DEC
1950
Fig. 2. (J \Gamma K) colour index map of the BD+40 ffi 4124 clus­
ter. The darkest grey tones correspond to higher values of
(J \Gamma K). The extended emission around the V 1318 complex has
a colour index almost 3 magnitudes larger than that around
BD+40 ffi 4124, as indicated by the gray scale with (J \Gamma K) val­
ues.
Fig. 4. The NIR color­color plot of the sources in the 7 0 \Theta 7 0
region. Only those sources detected in all the wavebands are
plotted. Plus symbols refer to objects without infrared excess
(see text), while black diamonds to objects with infrared ex­
cess. Circles are for sources in the inner 1 0 \Theta 1:5 0 with (filled)
and without (open) infrared excess, respectively. Superposed
on this plot is the main sequence relation of Bessell & Brett
(1988; solid line) and the reddening vectors computed for a
standard reddening law (Koornneef 1983; dashed line).
component varying between \Gamma70 and \Gamma84 km s \Gamma1 . The
VLA spectrum of Figure 4 shows that we have missed
some of the other components. However, our aim was the
determination of the exact position of the maser source,
and not the analysis of its intrinsic properties.
Fig. 5. The spectrum of the H2O maser in V 1318S. The ver­
tical scale gives the flux density in Jy.
3.3. The radio continuum
No 3.6 cm radio continuum emission was detected at
a 3oe upper limit of 0.24 mJy/beam associated with
BD+40 ffi 4124 or any of the other near­infrared sources
listed in Table 1. A weak point source with a flux
density at 3.6 cm of 0.64 mJy/beam was detected at
ff 1950 =20:18:46.297, ffi 1950 =+41:12:31.92. This source was
also detected by Skinner et al. (1993) at 6 cm, who's
more sensitive VLA observations also tentatively detected
a continuum source of flux 0.3 mJy/beam at 3.6 cm near
the optical position of BD+40 ffi 4124. No sources were de­
tected at 1.3 cm stronger than the 3oe upper limit of 6
mJy/beam.
3.4. C 18 O emission
The low isotopic abundance of C 18 O makes it an excellent
tracer of total molecular column density. Indeed, the nar­
row linewidth (FWHM=1.5 km s \Gamma1 ) and Gaussian line­
shape of the C 18 O spectra confirm that the emission is
mostly likely to be optically thin. Shown in figure 6, a map
of the integrated C 18 O J=2­1 emission (5 to 12 km s \Gamma1 )
reveals that the molecular gas in the BD+40 region forms

Palla et al.: 7
a ridge structure. The star BD+40 ffi 4124 itself lies near the
edge of the cloud, over 30 00 from the column density peak
of 7.0 \Theta 10 22 cm \Gamma2 , computed assuming an excitation tem­
perature of 12 CO of 28 K equal to the peak main beam
brightness temperature in the map of optically thick 12 CO
emission. On the other hand, the H 2 O maser lies within
a few arc seconds of the column density peak . Assum­
ing a C 18 O/ 12 CO terrestrial abundance ratio of 490 and
a H 2 /CO ratio of 10 4 (e.g. Blake et al. 1987), the total
molecular gas mass of the region mapped is M core =280
M fi for an assumed distance to the cloud of 1 kpc.
Fig. 6. Contour map of the integrated C 18 O emission super­
posed on the K­band map. The map covers a 1:5 0 \Theta 1:5 0 field
centered on BD+40. The C 18 O emission is in the velocity in­
terval 5 to 12 km s \Gamma1 . Contours go from 1 to 7 K km s \Gamma1 ,
increasing by unity.
3.5. CS emission
A map of the integrated CS J=5!4 emission (6 to 12
km s \Gamma1 ) is presented in Figure 7. The CS peak is offset
by about 7.5 00 to the north of the maser peak. The CS
emission region is quite compact, with very little emission
at the position of BD+40 ffi 4124 itself.
The CS line profile at the central position of the map
is shown in Figure 8, Hanning­smoothed to a velocity res­
olution of 0.12 km s \Gamma1 . A Gaussian fit to the line profile is
overlayed. We estimate the virialized mass of the central
region to be M vir =106\Sigma10 M fi , based on the observed
velocity dispersion oe = 2:5 \Sigma 0:1 km s \Gamma1 (FWHM).
The spatial correlation of the CS and H 2 O maser emis­
sion is consistent with the theory of Elitzur et al. (1989)
which predicts the strongest maser amplification to occur
Fig. 7. Contour map of the integrated CS (J=5! 4) emission
superposed on the K­band map. The map covers a 3 0 \Theta 3 0
field centered on the maser position. The CS emission is in the
velocity interval 6 to 12 km s \Gamma1 . Contours go from 1 to 4 K
km s \Gamma1 , increasing by 0.5.
Fig. 8. The CS J=5!4 line profile at 0.12 km s \Gamma1 resolution
at the maser position with a Gaussian fit overlayed.
in regions where the preshock densities are of the order of
3 \Theta 10 7 cm \Gamma3 . By comparison, based on CS collision rates,
the critical density to excite the 5!4 transition is 1.5 \Theta
10 7 cm \Gamma3 at T kin =20­40 K (Green & Chapman 1978).
3.6. High­velocity CO emission
The line profiles of the CO J=2!1 and C 18 O J=2!1
transitions at the position of the maser are shown in Fig­
ure 9. This figure presents convincing evidence for the
presence of asymmetries in several of these profiles. A
molecular outflow is detected in CO J=2!1. Figure 10
displays the outflow superposed on the K­band map. The

8 Palla et al.:
dynamical center of the outflow coincides with the H 2 O
maser position, which is south of the CS emission peak
by a quarter of the beamsize. Furthermore, the axis of
the outflow lies nearly perpendicular to the high column
density ridge structure detected in C 18 O.
Fig. 9. The 12 CO and C 18 O line profiles at the maser position
with velocity resolutions of 0.74 and 0.13 km s \Gamma1 , respectively.
A Gaussian fit to the C 18 O line is overlayed.
Fig. 10. The CO outflow superposed on the K­band map.
Shown is the distribution of the 12 CO (J=2!1) emission from
the dense molecular core. The velocity intervals are from \Gamma10
to 5 km s \Gamma1 (blue wing; solid lines) and 15 to 30 km s \Gamma1 (red
wing; dashed lines). Contours start at 8 K km s \Gamma1 and increase
by 4 K km s \Gamma1 . The cross marks the position of the H2O
maser.
C 18 O emission was not detected in neither the red
nor the blue wings at the central position of the map
to a 3­sigma level of 0.15 K km s \Gamma1 . Because this is a
factor of over 370 times weaker than the corresponding
12 CO integrated intensity, we assume that 12 CO is opti­
cally thin in the line wings. Since the outflow is not well­
resolved, we apply the beam filling factor analysis in de­
riving the column densities of the molecular gas (Garden
et al 1991). The total molecular gas mass of the blue and
red­shifted components within the central 1.5 arcmin is
M blue =21 and M red =14 M fi , respectively. Along the out­
flow direction, the separation of the half­power contours of
the blue and red wing emission is roughly 30 00 , which rep­
resents an upper limit to the actual length. Since a solid
lower limit is given by the peak­to­peak separation of the
blue and red wing emission, about 10 00 , we adopt a value
of 20\Sigma10 00 which corresponds to a physical length of 0.1
pc. The integrated median velocity of the outflow is 7.9
km s \Gamma1 . We then derive a dynamical timescale of 1.2\Theta10 4
yr and a molecular mass ejection rate of —
M flow =2.9\Theta10 \Gamma3
M fi /yr. The momentum and the mechanical luminosity
estimated for the flow are 275 M fi km s \Gamma1 and 17 L fi , re­
spectively. The main outflow properties are listed in Table
2.
4. Discussion
4.1. Comparison with other observations
The molecular cloud associated with BD+40 ffi 4124 was
first studied in 12 CO and 13 CO by Loren (1977) who
found complex emission at several velocities extending in
the north­south direction. The strongest component at
VLSR = 8 km s \Gamma1 was found to occur at a position cen­
tered on BD+40 ffi 4124, with a large uncertainty due to the
poor angular resolution of the observations (2. 0 6). A higher
resolution map of the 8 km s \Gamma1 component by Cant'o et al.
(1984) confirmed this result and showed an overall match­
ing of the molecular emission to the optical appearance
of the system, suggesting that the cloud represents an ex­
tension of the larger molecular cloud located to the north
of the star. Whether the other emission components at
2, 13 and 14 km s \Gamma1 are physically associated with the
BD+40 ffi 4124 region it is still not clear.
More recently, Fuente et al. (1990) have mapped in
NH 3 a region 3 0 \Theta 5 0 centered on BD+40 ffi 4124 (with a
beam resolution 42 00 FWHM ). Ammonia emission was de­
tected only at 8 km s \Gamma1 . The emission showed an elon­
gation in the north­south direction, with the size of the
filaments of 40 00 x 110 00 . Two maxima along the ridge
stand out, one of them coincident with the position of
BD+40 ffi 4124, and the other one with that of V 1318.
From the position­velocity diagram Fuente et al. found
that the radial velocity does not vary along the fila­
ment, but that the linewidth increases from 0.8 km s \Gamma1 at
BD+40 ffi 4124 to 1.5 km s \Gamma1 40 00 south. They suggest that

Palla et al.: 9
Table 2. CO outflow parameters.
Total mass (M fi ) : : : : : : : : : : : : : : 35
Average velocity ( km s \Gamma1 ) : : : : 7.9
Length (pc) : : : : : : : : : : : : : : : : : : : 0.1
Dynamical timescale (yr) 1.2\Theta 10 4
Mass loss rate (M fi yr \Gamma1 ) 2.9\Theta10 \Gamma3
Momentum (M fi km s \Gamma1 ) : : : : : 275
Flow force (M fi km s \Gamma1 yr \Gamma1 ) : 0.02
Mechanical luminosity (L fi ) : : : : 17
the broader profiles are due to two velocity components.
Our C 18 O spectra confirm the linewidth increase from 1.2
at BD+40 ffi 4124 to 2.1 km s \Gamma1 at the (+30, \Gamma30 00 ) offset
(nearest to V 1318 S). Also, we do detect a radial veloc­
ity increase from 7.71 at the position of BD+40 ffi 4124 to
8.34 km s \Gamma1 at the same offset. Clearly, the turbulence
generated by the molecular outflow can be responsible for
the observed increase of the nonthermal component of the
linewidth both in NH 3 and C 18 O.
The distribution of the submillimeter dust emission at
800 ¯m from the BD+40 ffi 4124 region has been presented
by Aspin et al. (1994). There is a strong peak at the posi­
tion of V 1318S (with an uncertainty of \Sigma2 00 ), while prac­
tically no emission is found from BD+40 ffi 4124 itself. A
secondary weaker peak is also present in the northern part
of the complex. Aspin et al. conclude that V 1318S is the
main source of the sub­mm emission and note that it is
clearly extended at 800 ¯m.
Our molecular line observations show that the high
density gas is concentrated in the vicinity of the south­
ern sources, V 1318 and V 1686, while it is marginally
present at the location of BD+40 ffi 4124. The morphology
of the dense core appears rather different depending on the
molecular tracer used. The two distinct cores observed by
Fuente et al. (1990) in NH 3 are not present in the C 18 O
and CS integrated intensity maps, which, on the contrary,
show a very steep drop away from V 1318. The north­south
elongation of the emission, however, especially in C 18 O re­
sembles that traced in ammonia. The total mass derived
from column density estimates (several 10 2 M fi ) is typical
of other dense cores found around intermediate luminos­
ity objects, such as the NGC 7129 complex (Ladd et al.
1991) or LkHff198 (Natta et al. 1992). This large mass
concentration, if compared to that of low­mass star form­
ing regions, is likely to be the defining characteristic that
allows more massive stars to form in such regions. From
the similarity of this and the virial mass, the difference
of a factor of two being within the uncertainties, we can
conclude that there is enough internal pressure to support
the cloud against further collapse. As to the origin of this
support, an obvious possibility is in the interaction of the
core with the molecular outflow.
4.2. The molecular outflow
One of the most interesting results of our observations
is the determination of the exact location of the central
source driving the molecular outflow. In this respect, the
coincidence of the location of both the strongest high ve­
locity CO emission and the H 2 O maser on V 1318S pro­
vides a direct answer to the problem. Additional support
is given by the 2¯m spectra of the region obtained by
Aspin et al. (1994). The spectra include several H 2 vi­
brational emission lines, the CO bandhead emission and
the Brfl line. The strongest emission lines are detected on
V 1318 S, while weaker emission is observed on V 1318N.
The emission extends about 40 00 north and 15 00 south of
V 1318 S, while none is found at a significant level in
coincidence of BD+40 ffi 4124 and V 1686. From the ra­
tio of the v=1­0 S(1) and Q(3) H 2 lines, Aspin et al.
derive a visual extinction on V 1318S of A v ¸25 mag­
nitudes, assuming an extinction law A – / – \Gamma1:8 . Such
high extinction is similar to the value we obtain us­
ing our C 18 O observations. Following the empirical de­
termination N (H 2 ) ú 10 21 (A v =mag)cm \Gamma2 (Frerking et
al. 1982), or the more recent prescription N (C 18 O) ¸
2 10 14 (A v =mag)cm \Gamma2 (Lada et al. 1994), we find that at
the peak column density of C 18 O the line­of­sight extinc­
tion is A v ¸50 mag. This explains the lack of optical emis­
sion around V 1318 (cf. the optical R image in Aspin et
al., their Fig. 1) and the powerful emission in the submil­
limeter continuum originating from V 1318 S.
The simultaneous detection of H 2 emission lines and
the H 2 O maser at the origin of the outflow is strongly
suggestive of shock, rather than UV excitation conditions
in this region. Current models of maser emission do indeed
predict the formation of molecules, sensitive to drastically
different physical conditions, in the cooling postshock gas
(e.g. Elitzur et al. 1989). In this scheme, H 2 molecules
are excited in a thin layer of hot (T kin ¸ 2000K) gas im­
mediately following the strong shock; H 2 O molecules pro­
duce maser emission in a plateau region of intermediate
(T kin ¸ 400K) temperature, and then CO emission is pro­
duced in a relatively cool (T kin ! 100K) shell of diffuse gas
which has had enough time to cool down. It would be in­
teresting to compare detailed contour maps of CO and H 2
emission to check the relative position of the peaks, with

10 Palla et al.:
the H 2 maximum expected closer to the exciting star. As
for the positional coincidence of the H 2 O maser at the cen­
ter of the outflow, the recent studies by Felli et al. (1992)
and Tofani et al. (1994) have shown that it is a common
property among known sources.
Considering the outflow energetics, it is well known
that there exists a general correlation between the me­
chanical luminosity of the outflow and the bolometric lu­
minosity of the driving source. The mechanical luminos­
ity of the flow derived in section 4 is L mech ¸ 17L fi .
With regard to the bolometric luminosity, we assume
that V 1318S is the driving source of the outflow and
we adopt the estimate of L \Lambda ¸ 1600L fi obtained by As­
pin et al. (1994) integrating the spectral energy distri­
bution from the optical to the millimeter and assigning
to the star all the sub­mm flux. The largest source of
uncertainty in the derivation of L \Lambda comes from the far­
infrared flux, since it is hard to properly partition the
observed IRAS fluxes to the various sources within the
BD+40 complex. Weaver & Jones (1992) and Aspin et al.
(1994) assign to V 1318S the majority of the flux at all
wavelengths; a maximum entropy run of the IRAS data
performed by us gives a result in good agreement with
this assumption. With the values quoted above, the ra­
tio L mech =L \Lambda ¸ 1:1 \Theta 10 \Gamma2 is in excellent accord with
the standard relation found for a large sample of molecu­
lar outflows from embedded and optically visible sources,
L mech =L \Lambda =0.01 (L \Lambda =1000L fi ) \Gamma0:2 =9:6 \Theta 10 \Gamma3 (Levreault
1988; Cabrit & Bertout 1992). Thus, it appears that the
energy of the radiation field is sufficient to drive the flow.
On the other hand, the ratio of the observed flow force
to the radiative force, L \Lambda =c, where c is the speed of light,
gives a values of ¸ 7, showing that radiation pressure alone
is insufficient to drive the outflow. We can then conclude
that the outflow originating from V 1318S is not anoma­
lous relative to the other known cases. The unusual aspect
is the lack of radio continuum emission at the position
of the H 2 O maser both at 1.3 and 3.6 cm. The detec­
tion of ionized winds in the centimetric range is a com­
mon property of molecular outflow sources, and Cabrit
& Bertout (1992) have discussed their connection to the
molecular flows and the implications on the driving mech­
anism. Missing this piece of information, it is hard to ar­
gue what kind of stellar wind/outflow interaction, energy
or momentum conserving, is at work in the present case.
Finally, let us consider the support provided by the
outflow to the cloud core. From the mass estimate of the
cloud, M core = 280M fi , we derive a gravitational energy
of E grav = 7 \Theta 10 46 ergs, whereas the kinetic energy of
the outflow provides E k = 3 \Theta 10 46 ergs. Thus, although
the outflow can contribute significantly to the internal dy­
namics of the core, it does not provide full support against
further gravitational collapse.
5. Conclusions
We have obtained near­infrared JHK images, maps in CO,
C 18 O and CS line emission, and water maser spectra of
the BD+40 region. From these observations we can draw
the following conclusions:
1. The presence of water maser emission at VLSR=­
80 km s \Gamma1 is confirmed by our VLA observations. The
position of the maser coincides with that of V 1318, one
of the optically visible emission line stars associated with
BD+40 ffi 4124, and not with the Herbig star itself. The
water maser is also at the center of a newly discovered
bipolar outflow.
2. Near­IR broad­band imaging of V 1318 reveals the
presence of two distinct components with a stellar appear­
ance, oriented north­south, a result also reported by Aspin
et al. (1994). The southern component, V 1318 S, coincides
with the position of the H 2 O maser and we attribute to it
the origin of the mass­loss activity observed in the whole
complex.
3. Near­IR imaging of a large
area around BD+40 ffi 4124 reveals the presence of many
highly embedded sources concentrated in the vicinity of
the bright visible stars. In the central part of the cluster,
the near infrared sources outnumber the optically visible
ones by a factor of six.
4. The distribution of the high density molecular gas
traced by the CS J=5!4 emission is highly concentrated
around V 1318 and V 1686, while the total column den­
sity of gas traced by the C 18 O J=2!1 shows a ridge
morphology centered on these sources. BD+40 ffi 4124 it­
self lies at the edge of both structures. From the optically
thin C 18 O emission, we derive a total molecular mass of
M core = 280M fi , for an assumed distance to the complex
of 1 kpc.
5. A molecular outflow has been detected in CO
J=2!1. The high velocity gas is confined to a compact
region of size ú 20 00 , corresponding to a physical length
of 0.1 pc. The outflow is not well resolved, or collimated,
on the observed scales and it is not likely to be close to
the plane of the sky. Conversely, the outflow might be
very young. The derived outflow parameters yield values
for the force and luminosity of the flow that agree well
with the standard correlations found in many other out­
flow sources.
6. The presence of maser emission and the molecular
outflow at the position of V 1318 S, together with strong
H 2 line emission reported by Aspin et al. 1994, clearly
indicate the presence of shock excitation conditions in the
region.
Acknowledgements. It is a pleasure to thank the TIRGO
staff for making ARNICA available to us during the com­
missioning period of the instrument. This work was partly
supported by the ASI­92­RS­54 grant. GBT acknowledges

Palla et al.: 11
support from NSF under grant AST 9117100. Research at
the CSO is supported by NSF contract AST­9313929.
References
Adams F.C., Lada C.J., Shu F.H. 1988, ApJ, 326, 865
Aspin C. & Barsony M. 1994, A&A, in press
Aspin C., Sandell G., Weintraub D.A. 1994, A&A, 282, L25
Barsony M., Schombert J.M., Kis­Halas K. 1991, ApJ, 379, 221
Bessell M.S. & Brett J.M. 1988, PASP, 100, 1134
Blake G.A., Sutton E.C., Masson C.R., Phillips T.G. 1987,
ApJ, 315, 621
Cabrit S. & Bertout C. 1992, A&A, 261, 274
Cant'o J., Rodriguez L.F., Calvet N., Levreault R.M. 1984,
ApJ, 282, 631
Cohen M. 1972, ApJ, 172, L61
Elitzur M., Hollenbach D.J., Mc Kee, C.F. 1989, ApJ, 346, 983
Felli M., Palagi F., Tofani G. 1992, A&A, 255, 293
Frerking M.A., Langer W.D., Wilson R.W. 1982, ApJ, 262, 590
Fuente A., Martin­Pintado J., Cernicharo J., Bachiller R. 1990,
A&A, 237, 471
Garden R.P., Hayashi M., Gatley I., Hasegawa T., Kaifu N.
1991, ApJ, 374, 540
Goodrich R.W. 1993, ApJS, 86, 499
Green S. & Chapman S. 1978, ApJS, 37, 169
Hartmann L., Kenyon S.J., Calvet N. 1993, ApJ, 407, 219
Herbig G.H. & Bell K.R. 1988, Lick Obs. Bull., No. 1111
Hillenbrand L. 1994 in The Nature and Evolutionary Status of
Herbig Ae/Be Stars, P.S. Th'e, M.R. Per'ez & J.P.E. van den
Heuvel eds., A.S.P. Conf. Ser., in press
Hillenbrand L., Strom S.E., Vrba F.J., Keene J. 1992, ApJ,
397, 613
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, Techni­
cal Report 3/94, Arcetri Astrophysical Observatory
Hunt L.K., Testi L., Borelli S., Maiolino R., Moriondo G.
1994c, Technical Report 4/94, Arcetri Astrophysical Ob­
servatory
Koornneef J. 1983, A&AS, 51, 489
Lada C.J. & Adams F.C. 1992, ApJ, 393, 278
Lada C.J., Lada E.A., Clemens D.P., Bally J. 1994, ApJ, in
press
Lada E.A., Strom K.M., Myers P.C. 1993, in Protostars and
Planets III, E.H. Levy & J.I. Lunine eds. (Tucson: The Uni­
versity of Arizona Press), p.245
Ladd E.F., Fuller G.A., Myers P.C., Stacy J.G., Benson P.J.,
Goodman A.A. 1991 in Atoms, Ions and Molecules, Proc.
of the 3 rd Haystack Conf., A.D. Haschick & P.T.P. Ho eds.,
A.S.P. Conf. Ser. 16, 297
Lagage P.O., Olofsson G., Cabrit S., Cesarsky C.J., Nordh L.,
Rodriguez Espinosa J.M. 1993 ApJ 417, L79
Levreault R.B. 1988, ApJ, 330, 897
Lisi, F., Baffa, C., and Hunt, L. K. 1993, in Infrared Detectors
and Instrumentation, SPIE Vol. 1946, p. 594
Loren R.B. 1977, ApJ, 218, 716
Natta A., Palla F., Butner H.M., Evans N.J.II, Harvey P.M.
1992, ApJ, 805, 816
Palla F. & Prusti T. 1993, A&A, 272, 249
Poetzl L.S., Ray P.T., Mundt R. 1992, A&A, 262, 229
Skinner S.L., Brown A., Stewart R.N. 1993, ApJS, 87, 217
Strom K.M., Strom S.E., Breger M., Brooke A.L., Yost J.,
Grasdalen G., Carrasco L. 1972, ApJ, 173, L65
Testi L. 1993, Technical Report 10/93, Arcetri Astrophysical
Observatory
Testi L., Stanga R.M., Natta A., Palla F., Prusti T., Baffa
C., Hunt L.K., Lisi F. 1994 in The Nature and Evolution­
ary Status of Herbig Ae/Be Stars, P.S. Th'e, M.R. Per'ez &
J.P.E. van den Heuvel eds., A.S.P. Conf. Ser., in press
Tofani G., Felli M., Taylor G.B., Hunter T.R. 1994, A&A, in
press
Th'e P.S., de Winter D., Per'ez M.R. 1994, A&A, in press
Weaver W.B. & Jones G. 1992, ApJS, 78, 239
Zinnecker H., McCaughrean M.J., Wilking B.A. 1993, in Proto­
stars and Planets III, E.H. Levy & J.I. Lunine eds. (Tucson:
The University of Arizona Press), p.429
This article was processed by the author using Springer­Verlag
L a T E X A&A style file L­AA version 3.