ArticlePDF Available

The structure of the cometary globule CG 12: A high-latitude star-forming region

Authors:

Abstract and Figures

The structure of the high galactic latitude Cometary Globule 12 (CG 12) has been investigated by means of radio molecular line observations. Detailed, high signal to noise ratio maps in C18O (1-0), C18O (2-1) and molecules tracing high density gas, CS (3-2), DCO+ (2-1) and H13CO+ (1-0), are presented. The C18O line emission is distributed in a 10' long North-South elongated lane with two strong maxima, CG12 N(orth) and CG12 S(outh). In CG12 S the high density tracers delineate a compact core, DCO+ core, which is offset by 15" from the C18O maximum. The observed strong C18O emission traces the surface of the DCO+ core or a separate, adjacent cloud component. The emission in high density tracers is weak in CG12 N and especially the H13CO+, DCO+ and N2H+ lines are +0.5 km/s offset in velocity with respect to the C18O lines. Evidence is presented that the molecular gas is highly depleted. The observed strong C18O emission towards CG12 N originates in the envelope of this depleted cloud component or in a separate entity seen in the same line of sight. The C18O lines in CG 12 were analyzed using Positive Matrix Factorization, PMF. The shape and the spatial distribution of the individual PMF factors fitted separately to the C18O (1-0) and (2-1) transitions were consistent with each other. The results indicate a complex velocity and line excitation structure in the cloud. Besides separate cloud velocity components the C18O line shapes and intensities are influenced by excitation temperature variations caused by e.g, the molecular outflow or by molecular depletion. Assuming a distance of 630 pc the size of the CG 12 compact head, 1.1 pc by 1.8 pc, and the C18O mass larger than 100 Msun are comparable to those of other nearby low/intermediate mass star formation regions. Comment: 18 pages, 17 figures Accepted A&A Sep. 22 2006
Content may be subject to copyright.
arXiv:astro-ph/0609820v1 29 Sep 2006
Astronomy & Astrophysics
manuscript no. haikala˙olberg February 20, 2011
(DOI: will be inserted by hand later)
The structure of the cometary globule CG 12: a high latitude
star forming region
⋆⋆
L. K. Haikala
1,2
and M. Olberg
3
1
Observatory, PO Box 14, University of Helsinki, Finland
e-mail: haikala@astro.helsinki.fi
2
Swedish-ESO Submillimetre Telescope, European Southern Observatory, Casilla 19001, Santiago, Chile
3
Onsala Space Observatory, S 439 00 Onsala, Sweden
e-mail: michael.olberg@chalmers.se
Received ; accepted
Abstract. The structure of the high galactic latitude Cometary Globule 12 (CG 12) has been investigated by means of radio
molecular line observations. Detailed, high signal to noise ratio maps in C
18
O (1–0), C
18
O (2–1) and molecules tracing high
density gas, CS (3–2), DCO
+
(2–1) and H
13
CO
+
(1–0), are presented. The C
18
O line emission is distributed in a 10
long
North-South elongated lane with two strong maxima, CG 12-N(orth) and CG 12-S(outh). In CG 12-S the high density tracers
delineate a compact core, DCO
+
core, which is offset by 15
′′
from the C
18
O maximum. The observed strong C
18
O emission
traces the surface of the DCO
+
core or a separate, adjacent cloud component. The driving source of the collimated molecular
outflow detected by White (1993) is located in the DCO
+
core. The C
18
O lines in CG 12-S have low intensity wings possibly
caused by the outflow.The emission in high density tracers is weak in CG 12-N and especially the H
13
CO
+
, DCO
+
and
N
2
H
+
lines are +0.5 km s
1
offset in velocity with respect to the C
18
O lines. Evidence is presented that the molecular gas
is highly depleted. The observed strong C
18
O emission towards CG 12-N originates in the envelope of this depleted cloud
component or in a separate entity seen in the same line of sight. The C
18
O lines in CG 12 were analyzed using Positive Matrix
Factorization, PMF. The shape and the spatial distribution of the individual PMF factors fitted separately to the C
18
O (1–0)
and (2–1) transitions were consistent with each other. The results indicate a complex velocity and line excitation structure
in the cloud. Besides separate cloud velocity components the C
18
O line shapes and intensities are influenced by excitation
temperature variations caused by e.g, the molecular outflow or by molecular depletion. Assuming a distance of 630 pc the size
of the CG 12 compact head, 1.1 pc by 1.8 pc, and the C
18
O mass larger than 100 M
are comparable to those of other nearby
low/intermediate mass star formation regions.
Key words. clouds – ISM molecules – ISM: structure – radio lines – ISM: individual objects: CG 12, NGC5367
1. Introduction
Herschel (1847) noted that a 10
th
mag. star, now known as
h4636 or CoD –39
8581, is a binary with a separation of
3.
′′
7. In the optical, h4636 illuminates the bright reflection neb-
ula NGC 5367. Hawarden & Brand (1976) showed that it lies
in the head of an impressive Cometary Globule 12, CG 12,
with a tail stretching about one degree to the SE. With a
galactic latitude of 21
and at the distance of 630 pc esti-
mated by Williams et al. (1977), CG 12 lies more than 200 pc
above the plane. CG 12 has an associated low/intermediate
mass stellar cluster which has at least 9 members (Williams
et al. 1977). The clouds cometary structure could be due to
Send offprint requests to: L. Haikala
Based on observations collected at the European Southern
Observatory, La Silla, Chile
⋆⋆
Figure 11 and Appendix A are only available in electronic form
via http://www.edpsciences.org
the passage of a supernovablast wave. Curiously, the cometary
tail stretches towards the Galactic plane which would place
the putative supernova even farther away from the Galactic
plane than the globule. According to Maheswar et al (1996)
the head of CG 12 is pointing towards the centre of an HI
shell. Such a shell is, however, not readily evident in the
whole sky HI survey (Kalberla et al. 2005) which merges the
northern Leiden/DwingelooSurvey(Hartman & Burton 1997 )
and the southern Instituto Argentino de Radioastronomia
Survey (Arnal et al. 2000). The CG 12 cloud cometary shape
is also seen in the IRAS surface emission (Odenwald 1988).
White (1993)mapped the regionaround and south of h4636
in
12
CO (2–1) and C
18
O (2–1) lines with a spatial resolution
of 22
′′
. He found a small size C
18
O core near the binary, and
a molecular outflow with a centre close to the binary system.
Further large scale
12
CO ,
13
CO and C
18
O observations with
2.
7 resolution are presented in Yonekura et al. (1999). CG 12
contains two compact 1.2mm continuum sources, one in the
2 L. K. Haikala and M. Olberg: The structure of the cometary globule CG 12
Fig.1. Maps of integrated emission of C
18
O (1–0) and C
18
O (2–1)
R
T
A
dv from 8 km s
1
to 4 km s
1
in CG12. The lowest
contour and the contour increment are 0.3 K km s
1
. The SEST half power beam sizes are shown in the upper corners of the
panels. Offsets from the map centre position, 13
h
57
m
43.
s
1, 39
58
43.
′′
3 (J2000), are shown on the axes of the right panel. The
scale tick in the lower right corner assumes a distance of 630 pc to the cloud.
direction of the C
18
O core detected by White (1993) (Reipurth
et al. 1996) and another one two arcminutes north of it (Haikala
2006, in preparation). The centres of the continuum sources
were observed in C
18
O (3–2) by Haikala et al. (2006).
Near infrared (J, H and K) images and photometry of
stars in CG12 are available in the 2MASS survey and
Santos et al. (1998). A deeper J, H and Ks imaging study is
presented in Haikala (2006, in preparation). Far infrared emis-
sion in CG 12 is dominated by a strong point source, IRAS
13547-3944, near the binary h4636.
CG 12/NGC 5367 is an intriguing object. It has the ap-
pearance of a cometary globule, such as usually found in the
outskirts of HII regions. The linear size of cometary glob-
ules, like those in the Gum nebula, are however much smaller
(e.g., Reipurth 1983). The linear extent of CG12, 10 pc, is four
times larger than, e.g., the archetype object CG 1. A cluster of
low/intermediate mass stellar cluster is associated with CG 12.
Perhaps the most curious feature of CG 12 is its location over
200 pc above the Galactic plane with no sign of other nearby
dark clouds or star formation. CG 12 has not attracted much
interest since the identification of the stellar cluster (Williams
et al.1977) and the Hawarden & Brand (1976) cometary glob-
ule paper. The subsequent papers have either concentrated on
the Herbig AeBe binary, h4636 or the cloud has been included
in various surveys. The
12
CO (2–1) and C
18
O (2–1) observa-
tions by White (1993), though detailed, cover only the very
centre of the cloud. What has been missing is a detailed but at
the same time extended study of the dense CG 12 molecular
cloud in molecular transitions sensitive to the large scale struc-
ture (cloud envelope) and to the detailed structure of the high
density material (cores).
In this paper we report the mapping of the head of
CG12 in the C
18
O (1–0) and (2–1) and in
13
CO (1–0)
lines. The C
18
O emission maxima were further mapped in
DCO
+
(2 1), H
13
CO
+
(1 0) and CS (2–1) and (3–2).
Further pointed observations with long integration time in
CO (1–0) and (2–1) (and isotopologues), CS (2–1) and (3–2),
C
34
S (2–1), H
13
CO
+
(1–0), DCO
+
(2 1) and N
2
H
+
(1–0)
were made towards selected positions in the cloud.
Observations, data reduction and calibration procedures are
described in Sect. 2 and the observational results in Sect 3.
The new results are compared with the optical and NIR im-
L. K. Haikala and M. Olberg: The structure of the cometary globule CG 12 3
ages in Sect. 4. In Sect. 5 and Appendix A Positive Matrix
Factorization is used to analyze the C
18
O small scale struc-
ture in the globule. In the discussion part in Sect. 6 the
C
18
O column densities and the cloud mass are derived and the
observational and calculated results are summarized. The con-
clusions are drawn in Sect. 7.
2. Observations
The observations were made during various observing runs
with the Swedish-ESO-Submillimetre-Telescope, SEST, at the
La Silla observatory, Chile. The SEST 3mm dual polariza-
tion, single sideband (SSB), Schotky receiver was used for the
13
CO (1-0)and CS (2-1) mapping observations.Rest of the ob-
servations were conducted with the SEST 3 and 2 mm (SESIS)
and 3 and 1 mm (IRAM) dual SiS SSB receivers. The SEST
high resolution 2000 channel acousto-optical spectrometer
(bandwidth 86MHz, channel width 43kHz) was split into two
halves to measure two receivers simultaneously. At the ob-
served wavelengths, 3mm, 2mm and 1mm, the 43kHz chan-
nel width corresponds to 0.12 km s
1
, 0.08 km s
1
and
0.06 km s
1
, respectively.
Frequency switching observing mode was used and a sec-
ond order baseline was subtracted from the spectra after fold-
ing. Calibration was achieved by the chopper wheel method.
All the line temperatures in this paper, if not specially noted,
are in the units of T
A
, i.e. corrected to outside of the atmo-
sphere but not for beam coupling. Typical values for the effec-
tive SSB system temperatures outside the atmosphere ranged
from 200 K to 350 K. Pointing was checked regularly in con-
tinuum mode towards the nearby Centaurus A galaxy. Pointing
accuracy is estimated to be better than 5
′′
.
The observed molecular transitions, their frequencies,
SEST half power beam width, HPBW, and the telescope main
beam efficiency, η
mb
, at these frequencies are given in Table 1.
For the CO observations only the C
18
O is listed. At a distance
of 630 pc the SEST HPBW at 219 GHz corresponds to 0.07 pc.
The cloud was mapped simultaneously in C
18
O (1–0) and
C
18
O (2–1) transitions with a spacing of 20
′′
(514 positions) .
The map centre was 13
h
57
m
43.
s
1, 39
58
43.
′′
3 (J2000) which
is 5
′′
and 10
′′
in right ascension from the positions of the
IRAS 13547-3944 point source and the binary h3626, respec-
tively. The average rms of the spectra were 0.07 K and 0.09 K
for C
18
O (1–0) and (2–1), respectively. An approximately 8
by 20
area was mapped in
13
CO using 40
′′
spacing (324 po-
sitions). The C
18
O maxima were mapped in CS (2–1), (3–2),
DCO
+
(2–1) and H
13
CO
+
(1–0). Further long integration time
pointed observations in CO, CS, H
13
CO
+
, DCO
+
C
34
S and
N
2
H
+
were made.
3. Results
3.1. CO
The observed distributions of C
18
O (1–0) and C
18
O (2–1) line
emission towards CG12 are presented in Fig. 1 (the
R
T
A
dv
in the velocity range 8 km s
1
to 4 km s
1
) and Fig. 11
(C
18
O channel map). In both figures the grey/colour scales and
Table 1. Observed lines and telescope parameters
Line ν [GHz] HPBW [
′′
] η
mb
H
13
CO
+
(1–0) 86.754 57 0.75
N
2
H
+
(1–0) 93.176 54
C
34
S (2–1) 96.412 54
CS (2–1) 97.271 54
C
18
O (1–0) 109.782 47 0.70
DCO
+
(2–1) 144.077 34
CS (3–2) 145.904 34 0.66
C
18
O (2–1) 219.560 24 0.50
the contour levels are the same for the two C
18
O transitions.
The offsets from the map centre position in arc minutes are
shown on the axes of the right panel in Fig. 1.
The bulk of the molecular material traced by C
18
O emis-
sion is distributed in a narrow North-South oriented lane with
two prominentmaxima. A less intense maximum is observed to
the SW. Henceforth the three intensity maxima will be referred
to as CG 12-N, CG 12-S and CG 12-SW. CG 12-S corresponds
to the C
18
O (2–1) maximum reported in White (1993).
The morphology of the cloud is similar in the two
C
18
O transitions. However, there is a notable difference be-
tween the C
18
O maxima. The observed C
18
O (1–0) emission
is stronger than the C
18
O (2–1) emission in CG 12-N and
CG 12-SW whereas in CG 12-S the opposite is the case. In
CG 12-N the C
18
O (1–0) line remains stronger than the (2–1)
line even when expressed in the main beam brightness temper-
ature scale.
Further details can be seen if Fig. 11 which reveals that
the structure of the cloud is not as simple as Fig. 1 suggests.
The most noticeable features are the following: CG 12-N is
elongated in the North South direction at velocities 6.0
km s
1
and in the East West direction at velocities 5.4
km s
1
. The position of the maximum C
18
O (2–1) emission in
CG 12-S moves from below the map centre position (marked
with a cross in the Figure) at 7.0 km s
1
to a position 40
′′
west of this position at velocity 5.9 km s
1
. An arc like fea-
ture connects CG 12-S and CG 12-SW at velocities from 6.6
km s
1
to 6.2 km s
1
. The arc is less pronounced in the
C
18
O (2–1) transition.
The observed distribution of the
13
CO (1–0) line emis-
sion towards CG12 is presented in Fig. 2. The extent of the
C
18
O mapping and the outline of the C
18
O (2-1) emission is
indicated in the overlay. The notable difference between the
observed C
18
O and
13
CO emission is that there is only one
13
CO maximum which is offset to NE from CG 12-S. In par-
ticular there is no indication of CG 12-N in the
13
CO map.
12
CO,
13
CO and C
18
O (1–0) and (2–1) (pointed, long inte-
gration time) spectra in five selected positions in the cloud are
shown in Fig. 3. The
12
CO spectra observed at all positions in
the figure are strongly self-absorbed. Therefore the
12
CO line
peak intensity, line half width and the line integral are not phys-
ically meaningful. Line wing emission due to a molecular out-
flow (White 1993) is seen in all
12
CO spectra, also in the direc-
tion of CG 12-N which was not covered by the White (1993)
observations.
4 L. K. Haikala and M. Olberg: The structure of the cometary globule CG 12
Fig.3.
12
CO ,
13
CO and C
18
O spectra observed towards ve selected positions in CG 12. Offset from the map centre position
is shown in arc seconds in the upper left corners of each panel. Line has been used for the (1–0) and histogram for the (2–1)
transitions. The antenna temperature scale at left is for
12
CO and
13
CO spectra and the one on right for C
18
O. The velocity tick
indicates velocity of 6.5 km s
1
. The
12
CO and
13
CO spectra in position (40
′′
,160
′′
) are offset from the zero level for clarity
3.2. High density tracers, mapping
In the optically thin case the critical density of the C
18
O (1–
0) transition is 650 cm
3
and approximately ten times
this value for the (2–1) transition (Rohlfs & Wilson 1999).
These C
18
O lines are therefore excited already at low den-
sities and their emission traces column density rather than
number density. Therefore the C
18
O intensity maxima were
also mapped in the high density tracers (critical densi-
ties larger than 10
5
cm
3
) CS (2–1), H
13
CO
+
(1 0) and
DCO
+
(2 1) using 20
′′
spacing. A 4 by 4 point map with the
same spacing was obtained in CG 12-S in the CS (3–2) line.
Contour maps of the CS (2–1), CS (3–2), H
13
CO
+
and
DCO
+
line integral in CG 12 superposed on the C
18
O (2–
1) emission (grey scale) are shown in Fig. 4. CG 12-N was
not mapped in CS (3–2). The CS (2–1) emission peaks at
the position of the C
18
O (2–1) the maximum in CG 12-S.
In other molecules the maximum is shifted to SE from
the C
18
O maximum. As DCO
+
traces high density gas the
CG 12-S DCO
+
maximum will be referred to as the DCO
+
core.
3.3. C
18
O and high density tracers: pointed
observations
Pointed, long integration time CS (2–1), (3–2), C
34
S(2–1),
H
13
CO
+
(1–0), DCO
+
(2–1) and N
2
H
+
(1–0) spectra in the
same positions as in Fig 3 are shown in Fig. 5. The integra-
tion time of the CS (3–2) line at position (40
′′
,20
′′
) is only
2.5 minutes and is therefore noisier than the two other CS (3–2)
lines. N
2
H
+
(1–0) line consists of seven hyperfine components
and only the unblendedN
2
H
+
(1–0) (F
1
, F= 0, 1 1, 2) com-
ponent is shown in Fig. 5.
The C
34
S line was observed in three positions. These spec-
tra together with the CS (2–1) and (3–2) spectra observed at the
same positions are shown in Fig. 6.
The N
2
H
+
spectra showing all the hyperfine components
in the three observed positions are shown in Fig. 7. Also the
hyperfinecomponentfits (assuming optically thin emission and
one velocity component) and their residuals are shown.
L. K. Haikala and M. Olberg: The structure of the cometary globule CG 12 5
Fig.2. Map of integratedemission of
13
C0 (1–0):
R
T
A
dv from
8 km s
1
to 4 km s
1
. The lowest contour and the contour
increment is 1.2 K km s
1
. The SEST half power beam size at
the line frequency is shown in the upper corner. The extent of
the C
18
O mapping and the outline of the C
18
O (2-1) emission
is indicated in the overlay.
3.3.1. CG 12-S
Near the centre of the DCO
+
core (position 20
′′
,20
′′
)
the H
13
CO
+
, DCO
+
and N
2
H
+
lines are nearly symmet-
ric and centered at velocity 6.5 km s
1
. The C
18
O lines
are skewed and the maximum intensity is redshifted from
the DCO
+
(2 1) peak velocity. Going to the West (position
40
′′
, 20
′′
) the redshifted side of the C
18
O lines becomes
more intense. Low intensity wing-like emission is observed in
the C
18
O and CS (2–1) lines in the direction of the DCO
+
core
and position (0
′′
,0
′′
).
The N
2
H
+
line in the position (20
′′
,20
′′
) can be fit with
a single velocity component centred at 6.56 km s
1
(Fig. 7,
lower panel). Due to the 0.73 km s
1
line half width the hyper-
fine componentsare blended and the hyperfine structure is only
marginally resolved. The signal-to-noise ratio of the spectrum
Fig.4. Contour maps of the observed CS (2–1), CS (3–2),
H
13
CO
+
(1 0) and DCO
+
(2 1) distributions in CG 12 su-
perposed on a grey scale map of the C
18
O (2–1) emis-
sion. The positions observed in each molecule are in-
dicated. The corresponding SEST beam size at the ob-
served frequencies is indicated in the upper left corner
of each panel. In the CS (2–1) and CS (3–2) maps the
lowest contour value is 0.5 K km s
1
and the increments
are 0.5 K km s
1
and 0.3 K km s
1
, respectively. In the
H
13
CO
+
(1 0) and DCO
+
(2 1) maps the lowest contour
value and its increment are 0.15 K km s
1
. The maximum
values of the CS (2–1), CS (3–2), H
13
CO
+
(1 0) and
DCO
+
(2 1) emission are 2.2 K km s
1
, 1.3 K km s
1
,
0.58 K km s
1
and 0.77 K km s
1
, respectively. The offset
from the map zero position is shown on the axes.
is not sufficient to make a meaningful fit of the line total optical
depth, τ
tot
. The estimated τ
tot
upper limit in this position is 5.
6 L. K. Haikala and M. Olberg: The structure of the cometary globule CG 12
Fig.6. CS (2–1), (3–2) and C
34
S (1–0) spectra observed in the
direction of CG 12-N (the two upper panels) and CG 12-S (the
lower panel)
.
3.3.2. CG 12-N
In CG 12-N (position 40
′′
,160
′′
) the C
18
O lines are nearly
symmetric but not gaussian. The emission from the high
density tracers is weak when compared to that observed in
CG 12-S.
The C
34
S line was observed in two positions in
CG 12-N (Fig. 6). The CS(2–1) and (3–2) line profiles in these
positions are not gaussian and may be composed oftwo or three
components. The peak intensity of the C
34
S (2–1) line is red-
shifted with respect to thepeakof the CS (3–2) line which itself
is redshifted with respect to the CS (2–1) line. The latter line
peaks approximately at the same velocity as the C
18
O lines.
The observedionic lines (DCO
+
(2-1), H
13
CO
+
(1-0) and
N
2
H
+
(1–0)) are offset +0.5 km s
1
in velocity with respect
to the C
18
O lines and peak at a velocity where especially the
C
18
O (2–1) line intensity is low.
N
2
H
+
(1–0) was observed in the same two positions as
C
34
S. Contrary to the CS lines, the observed N
2
H
+
lines are
identical within the noise (Fig. 7, two upper panels). The SEST
beam is the same at CS (2–1) and N
2
H
+
(1–0) frequencies.
A hyperfine component fit to the CG 12-N N
2
H
+
lines gives a
0.65 km s
1
broad line at 5.41 km s
1
. The estimated τ
tot
upper limit in position (40
′′
,160
′′
) is 2.
To rule out thatthe observedintensity/velocitystructure de-
scribed above is due to calibration, pointing or frequency set-
ting problems, the observations were carefully checked by ob-
serving the C
18
O and the DCO
+
, H
13
CO
+
and CS lines after
each other and making pointing checks before and after obser-
vations. Also, e.g., the line pair DCO
+
(2–1) and C
18
O (1-0)
could be observed simultaneously with a dual receiver,thus ex-
cluding relative pointing errors.
Fig.7. N
2
H
+
(1–0) spectra observed in the direction of
CG 12-N (the two upper panels) and CG 12-S (the lower
panel). The seven component hyperfine structure fit and the re-
sulting residuals are shown
4. Comparison with observations at other
wavelengths
The C
18
O (2–1) contour map superposed on the blue SERC-
J DSS image is shown in Fig. 8. The bright optical re-
flection nebula NGC 5367 lies in front of CG 12-S whereas
CG 12-N is in the direction of an optically heavily obscured
region which is clearly seen in the inset. A 1.2 mm source
(Haikala 2006, in preparation) coincides with the position of
the CG 12-N C
18
O maximum.
A Ks band image (Haikala 2006, in preparation) of CG 12
is shown in Fig. 9. Superposed on the image are the contours of
the
12
CO (2–1) molecular outflow line wing emission (White
1993). In the insets the C
18
O (2–1) and DCO
+
contour maps
are superposed together with the positional uncertainty ellipses
of IRAS point sources IRAS 13546-3941 and IRAS 13547-
3944. IRAS 13547-3944 does not coincide with the binary
L. K. Haikala and M. Olberg: The structure of the cometary globule CG 12 7
Fig.5. Spectra of C
18
O and molecules sensitive to number density towards the same positions as shown in Fig. 3. Only the
unblended hyperfine N
2
H
+
(1–0) (F
1
, F= 1, 0 1, 2) component is shown. The velocity tick indicates a velocity of 6.5
km s
1
. C
18
O (1–0), CS (2–1) and H
13
CO
+
lines are plotted using continuous line. The panel lower right is a blowout of the
upper right panel showing only the lines tracing high number density.
h4636 (the base of the binary star emission is saturated due
to the intensity scale which was chosen to emphasize the low
intensity surface emission).
A cone like nebulosity with a bright head is seen projected
on the DCO
+
core in Fig. 9. The apex of the cone is located
just off the tip of the red shifted and below the end of the blue
shifted, collimated CO outflow lobes of White (1993). A com-
pact 1.2 mm source (Haikala 2006, in preparation) coincides
with the DCO
+
core. The core does not have an associated
point source but this could be due to the low spatial resolution
of the IRAS satellite. A faint source would be masked by the
strong nearby source IRAS 13547-3944. The NIR cone could
be associated with the driving source of the outflow. The centre
of the molecular outflow is offset from the position of IRAS
13547-3944 by 20
′′
. It is unlikely that the point source is as-
sociated with the large outflow but it could be the driving force
of the strong redshifted outflow lobe located 1
NW of the
point source nominal position.
Small areas of the size of the SEST HPBW at the C
18
O (1–
0) frequency in the centres of CG 12-N and CG 12-S have
been mapped in the C
18
O (3–2) (Haikala et al. 2006). The
HPBW of the C
18
O (3–2) observations is 19
′′
which is simi-
lar to the SEST HPBW of 24
′′
at the C
18
O (2–1) frequency.
In CG 12-S the C
18
O (3–2) emission has a strong maximum in
the same position as the maxima in the lower C
18
O transitions.
The line profile in position (20
′′
,20
′′
) agrees with the
C
18
O (2–1) line (Fig. 5). The C
18
O (3–2) T
M B
peak line tem-
perature in this position is, however, 10 K which is nearly
twice the C
18
O (2–1) T
M B
of 5.7 K. In CG 12-N (position
(40
′′
,160
′′
) the C
18
O (3–2) line profile is similar to that of
C
18
O (1–0) but the T
M B
peak line temperature is 3 K which
is lower than the corresponding values observed in C
18
O (1–0)
and (2–1) which are 4.3 K and 3.8 K, respectively.
5. C
18
O fine scale structure
The half widths of the observed C
18
O lines in CG 12 cores
range from 1.0 km s
1
(CG 12-S) to 1.3 km s
1
(CG 12-N).
The C
18
O channel-maps (Fig. 11) and observations of other
molecules indicate, that the lines consist of more than one ve-
locity component.The division of the cloud according to its ap-
pearance in the C
18
O line integral maps (Fig. 1) into only three
components, CG 12-N, CG 12-S and CG 12-SW, is therefore
8 L. K. Haikala and M. Olberg: The structure of the cometary globule CG 12
Fig.8. Contour map of the C
18
O (2–1) emission superposed
on the optical (DSS blue) image of NGC 5367. The image is
also shown in the inset but using an intensity scale which better
brings out the surface emission. The linear size of the image is
32
by 32
.
too coarse. The fine structure within the individual maxima is
not seen because the lines are heavily blended in velocity.
5.1. Positive Matrix Factorization
Positive Matrix Factorization (PMF) has been used by
Juvela et al. (1996) and Russeil at al. (2003) in the analysis of
molecular line spectral maps. PMF assumes, that the ensem-
ble of input spectra are composed of n individual line com-
ponents (factors). Unlike the Principal Component Analysis
PMF assumes that the individual components are positive.This
makes the interpretation of the results more straightforward
than in Principal Component Analysis where the results may
contain negative components. No other assumptions are made
of the shape of the factors. Each input spectrum can be recon-
structed from the n PMF output factors by multiplying each
factor by the weight assigned to it by PMF at the position and
adding up the multiplied factors. A comparison of the results
obtained with PMF analysis with those obtained using the anal-
ysis of channel maps and by the use of Principal component
analysis is given Russeil at al. (2003).
PMF has been applied separately to the north and south
parts of CG 12 but there is a small positional overlap between
the two areas used in the analysis. The same dataset, which is
shown in Fig. 11 was used in the analysis, i.e, the C
18
O (2–1)
data is binned to the same channel width as the C
18
O (1–0)
data. The PMF analysis is presented in Appendix A.
Fig.9. Ks–band image of CG-12 (Haikala 2006, in prepa-
ration). Contours of the blue (white contours) and red-
shifted (black)
12
CO (2–1) line wing emission (White 1993)
are also shown. The two insets show the CG 12-S and
CG 12-N regions in detail. Superposed are the contours of
C
18
O (2–1) (black) and DCO
+
(white) emission. The po-
sitional uncertainty ellipses of the IRAS point sources IRAS
13547-3944in CG 12-S andIRAS 13546-3941in CG 12-N are
also shown.
6. Discussion
6.1. C
18
O as a tracer of molecular material
C
18
O emission is generally considered a good tracer of
large scale structure of dark clouds and globules. However,
C
18
O photodissociation in the cloud envelopes and molecular
depletion in the dense and cold cloud cores may restrict the
number density interval where C
18
O emission can be used as
a direct measure of H
2
column density (especially if the LTE
approximation is used). Unlike
12
CO, the C
18
O molecules are
not shielded against photodissociation by strong H
2
lines and
further, the C
18
O self-shielding is weaker than that of
12
CO.
For C
18
O the self-shielding is most efficient for the two low-
est rotational levels with the largest populations. According to
Warin et al. (1996), the consequence is that the J=1–0 transi-
tion becomes thermalized, whereas the higher transitions re-
main subthermally excited in the cloud envelopes. In the other
extreme (pre-stellar cloud cores and protostellar envelopes)
the CO molecule may vanish from gas phase because of de-
pletion onto dust grains. Further complications in interpreting
C
18
O data are the possible strong gradients in the CO excita-
tion temperature on the line of sight due to heating of the gas
by newly born stars and protostellar objects. Time and tem-
perature dependent chemistry like, e.g., deuterium fractiona-
tion, also complicates the comparison of C
18
O data with that
of other molecular species.
Depletion of molecules on dust grains takes place in the
dense, quiescent and cold molecular cloud cores. CO and CS
are among the first molecules to disappear from the gas phase
L. K. Haikala and M. Olberg: The structure of the cometary globule CG 12 9
but nitrogen bearing species like, e.g., N
2
H
+
and ammonia,
seem to be able to resist depletion (Tafalla et al. 2002). The
N
2
H
+
and DCO
+
lines in CG 12-N peak at a velocity of
5.4 km s
1
where the C
18
O line emission is weak (Sect. 3.3).
It is likely, that strong CO depletion has taken place in the gas
traced by these two molecules in CG 12-N .
In CG 12-S the C
18
O maximum is spatially offset from the
DCO
+
core. Deuterium fractionation reactions are favored in
cold gas (e.g., Herbst 1982) and therefore the relative abun-
dance of DCO
+
is enhanced in cold cloud cores. The relatively
strong DCO
+
emission observed in CG 12-S, the DCO
+
core,
could indicate that fractionation has indeed taken place. If the
molecules in CG 12-S were heavilydepleted in the cold DCO
+
core one would not expect CS (3–2) or H
13
CO
+
to outline the
core like they do (also DCO
+
will be finally depleted).
The observed
12
CO (1–0) and (2–1) line temperatures
in CG 12-S (Fig. 3) are high. The lines are self reversed
so the actual line peak temperatures must be even higher.
This indicates that the CO excitation temperature is in ex-
cess of 30 K in the part of the cloud which is traced by
the observed
12
CO emission. White (1993) suggests that ei-
ther h4636 or IRAS 13547-3944 is the heating source. The
12
CO and
13
CO optical depth is high and unlike C
18
O these
CO isotopoloques are likely to trace only the surface of
the molecular cloud associated with CG 12. The observed
C
18
O (3–2) T
M B
peak temperature of 11 K in CG 12-S also
points at a high C
18
O excitation temperature (Haikala et al.
2006). However, the observed relative C
18
O (2–1) and (1–0)
line intensities in CG 12-S are not compatible with C
18
O T
ex
values higher than 20 K.
6.2. PMF: The interpretation
One should be cautious in interpreting the PMF results. Even
though the fit to the C
18
O (1–0) and (2–1) data is good, all
the PMF factors do not necessarily describe real cloud compo-
nents. The structure of the cloud can, however, be discussed
with some confidence when the PMF fit results are consid-
ered with the information provided by other available molec-
ular line, mm continuum and NIR-FIR data.
6.2.1. CG 12-S
The location of the CG 12-S C
18
O (2–1) PMF factor maxima
relative to the NIR cone (and the mm continuum source) in the
centre of the DCO
+
core is shown in Fig. 10. From East to
West the factors are 1s, 2s and 3s. The centres of the maxima
differ in position by about one SEST beam size at 220 GHz,
24
′′
.
The data would seem to implicate that CG 12-S is frag-
mented into three individual cores. If this were the case one
would expect that the 2s factor would coincide with the DCO
+
core because the 2s centre of line velocity is the same as that of
DCO
+
. Even though the 3s factor is the strongest in CG 12-S it
is detected only as an asymmetry in the DCO
+
and CS lines in
Figs. 5 and 6. CO is ubiquitous and easily excited and therefore
it can be detected at much lower densities than the high density
Fig.10. The CG 12-S C
18
O (2–1) 1s, 2s and 3s (from East to
West) factor peak emission positions (0.8 K km s
1
contours)
and DCO
+
core position (white) superposed on the SOFI Ks
band image. Offsets from the map zero position are shown in
arcminutes. The SEST beam at the C
18
O (2–1) frequency is
0.37 arcminutes
tracers, DCO
+
and CS (3–2),which trace the DCO
+
core. The
bulk of the observed C
18
O emission may have its origin in the
part of the cloud where the density is lower and the excitation
temperature higher than in the DCO
+
core.
This leads to a model where most of the observed
C
18
O emission traces rather the surface of the dense and cold
cloud core (the DCO
+
core) or a separate, adjacent cloud com-
ponent,than the core itself. The observed CS (2–1) distribution,
which is similar to C
18
O in CG 12-S (Fig. 4), is in accord with
this model. Much of the observed CS (2–1) emission is known
to originate in cloud envelopes and not in the dense cores. The
CS (3–2) effective critical density is higher than that of CS (2–
1) and therefore it traces the dense gas deeper in the cloud than
CS (2–1).
The possible interaction of the collimated molecular out-
flow detected by White (1993) with the cloud core further com-
plicates the interpretation of the molecular line data. The centre
of the outflow is located in the direction of the DCO
+
core and
it is not unlikely that the outflow can produce weak C
18
O line
wings and that it can raise the CO excitation temperature lo-
cally. Weak line wings are also observed in the CS (2–1) line in
Fig. 5. It is argued that the weak C
18
O red and blue shifted line
wing emission near the DCO
+
core is produced by the interac-
tion of the collimated molecular outflow with the parent cloud.
The line wings would be described by the PMF factors 1s and
4s. For the latter factor this would consists only of the emission
immediately to the west of the DCO
+
core.
The CG 12-S C
18
O (3–2) map in Haikala et al. (2006) cov-
ers only the very centre of CG 12-S. The C
18
O (3–2) line peak
10 L. K. Haikala and M. Olberg: The structure of the cometary globule CG 12
velocity and the intensity distribution is in accordance with the
elongated shape of the 3s component in Fig. 10. The maximum
of the 3s factor is compact and lies on the outflow axis only
20
′′
awayfrom the apex of the NIR cone (Fig. 9), the putative
position of the outflow driving source. This suggest, that the
high 3s C
18
O line intensities and the molecular outflow could
be connected.
6.2.2. CG 12-N
PMF produces a seemingly straightforward solution for
CG 12-N. However, the available data from molecules other
than C
18
O and the 1.2 mm continuum emission deny such a
simple solution. As argued in Sect. 6.1 it is likely that the cloud
component traced by DCO
+
and N
2
H
+
is heavily depleted.
If this is the case C
18
O only probes the undepleted part of
CG 12-N.
6.3. Cloud physical properties
A straightforward division of the CG 12 molecular cloud
into three homogeneous components, CG 12-N, CG 12-S and
CG 12-SW, is not possible. The good velocity resolution and
the high signal to noise ratio of the data allows one to divide the
C
18
O data into separate components. However, the derivation
of the cloud or core physical properties calls for a detailed three
dimensional non-LTE model which takes into account both the
molecular depletion, the varying density and excitation con-
ditions. Such a model is beyond the scope of this paper and
will be left for the future. The LTE-approximation approach is
used instead of a sophisticated model to make a zeroth order
estimate of the masses of the three C
18
O maxima, CG 12-N,
CG 12-S and CG 12-SW. An average C
18
O excitation temper-
ature for each maximum is estimated using the observed rela-
tive C
18
O (2–1) and (1–0) line intensities.
6.3.1. Mass and column density estimation
The observed T
A
line temperatures were converted into main
beam temperatures using the beam efficiencies in Table
1. The C
18
O excitation temperatures were estimated from
the C
18
O (1–0) and (2–1) data assuming optically thin
C
18
O emission and LTE. This assumes that the observed emis-
sion in both transitions originates in the same volume of gas
at a constant excitation temperature. The observed C
18
O line
T
M B
(2–1)/ T
M B
(1–0) ratio is 1.8 in CG 12-S, compatible
with an excitation temperature near 15 K. In CG 12-N and
CG 12-SW this ratio is one or less indicating excitation tem-
peratures of the order of 10 K.
The masses of CG 12-N, CG 12-S and CG 12-SW are
calculated without dividing them into smaller components.
The spectra between declination offsets (including the lim-
its) 120
′′
and +60
′′
were assigned to CG 12-S. The spectra
above and below these limits were assigned to CG 12-N and
CG 12-SW, respectively. An average C
18
O excitation temper-
ature of 10 K is now assumed for CG 12-N and CG 12-SW and
15 K for CG 12-S. The adopted C
18
O abundance is 2.0 10
7
.
This results in calculated total masses of 34, 96, and 110
M
for CG 12-SW, CG 12-S and CG 12-N, respectively, when
C
18
O (1–0) data is used. If C
18
O (2–1) data is used the masses
are 14, 49 and 53 M
. The maximum T
M B
line integrals can
be used to estimate the H
2
column densities in the correspond-
ing beams. The calculated maximum column densities in the
C
18
O (1–0) beam are 1.3 10
22
, 1.8 10
22
and 2.7 10
22
cm
2
for CG 12-SW, CG 12-S and CG 12-N, respectively. For the
C
18
O (2–1) beam the corresponding values are 6.4 10
21
, 1.6
10
22
and 1.8 10
22
cm
2
.
The masses calculated from the C
18
O (1–0) data are ap-
proximately twice higher than from the (2–1) data. The LTE
approximation assumes that all the C
18
O rotational states are
thermalized. However, according to Warin et al. (1996) only
the lowest C
18
O rotational state is thermalized in dark clouds
and the higher states are subthermally excited. The subthermal
excitation would be stronger in the relatively low density cloud
envelope than in the more dense core. The subthermal level
population leads to an underestimation of the cloud mass when
calculated from the C
18
O (2–1) data. This could be a partial
explanation for the large discrepancy between the masses cal-
culated using C
18
O (1–0) and (2–1) data.
Strong C
18
O (3–2) emission (maximum T
M B
11 K)
distributed similar to the 3s PMF factor was detected by
Haikala et al. (2006) in the centre of CG 12-S . The C
18
O (3–
2) data was modelled with a compact (60
′′
to 80
′′
diam-
eter) and hot (80 K .T
ex
. 200 K) optically thin clump
of 1.6 M
. The high temperature was derived from the
T
M B
(C
18
O (3–2))/T
M B
(C
18
O (2–1)) ratio. However, T
M B
(C
18
O (2–1))/T
M B
(C
18
O (1–0)) ratio for the CG 12-S PMF
factor 3s is not compatible with such a high temperature. This
could be due to, e.g., contribution from a subthermally ex-
cited cool cloud envelopeto the C
18
O (1–0)emission. The LTE
mass calculated from the C
18
O (2–1) 3s component within the
area modelled in Haikala et al. (2006) is 4.1 M
when a T
ex
of 15 K is assumed. The discrepancy in the calculated ex-
citation temperatures and masses highlights the uncertainties
when LTE approximation is used and only data of the two low-
est C
18
O transitions are available. In contrast to CG 12-S the
C
18
O T
ex
and the LTE mass of CG 12-N correspond well to
the modelled values using the the three transitions (Haikala et
al 2006).
6.4. Summary
The analysis of the spectral lines presented above deliver a
complicated picture of CG 12. Even though the spectral lines
are relatively narrow the analysis reveals a rich structure in line
shape and velocity. Probable depletion of molecules on dust
grains in CG 12-N further complicates the interpretation. The
cloud parameters derived using the LTE analysis are highly un-
certain. However, if the molecular line observations are com-
bined with the available optical, NIR, FIR and mm continuum
data the cloud structure can be discussed with some confidence.
L. K. Haikala and M. Olberg: The structure of the cometary globule CG 12 11
6.4.1. CG 12-N
CG 12-N harbors a compact, cold mm source and the detected
relatively weak molecular emission in high density tracers is
probably associated with this source. Much of the molecular
material associated with this cloud component is likely to be
highly depleted onto dust. The observed strong C
18
O emission
towards CG 12-N originates therefore in the envelope of this
depleted core or in a separate entity seen in the same line
of sight.
12
CO observed in the direction of CG 12-N has line
wing emission indicating molecular outflow. It is not known if
this outflow is connected to the outflow in CG 12-S or if it is
local and originates in CG 12-N .
6.4.2. CG 12-S
The bulk of the observed C
18
O emission does not trace the gas
associated with the compact cloud core detected in H
13
CO
+
,
DCO
+
and CS (3–2) lines. Most likely the C
18
O emission
traces only the surface of this core. The moderate C
18
O line
wing emission, probably due to interaction of the highly col-
limated molecular outflow with the surrounding gas, further
complicates the interpretation.
The molecular line data presented in this paper combined
with the molecular outflow data from White (1993), the NIR
imaging and mm dust continuum data (Haikala 2006, in prepa-
ration) shows that the outflow centre coincides with the mm
continuum source and the NIR cone in the centre of the DCO
+
core. The strong point source IRAS 13547–3944 is offset from
the mm source and from h4636. This point source could, how-
ever, be the driving source for the strong redshifted outflow
lobe north of CG 12-S.
6.4.3. CG 12-SW
CG12-SW is inconspicuouswhen comparedtothe two stronger
C
18
O maxima. No signs of star formation have been found in
its direction and it is therefore either still in pre-star forma-
tion phase or its density-temperature structure is such that no
star formation will take place. The arc like feature which con-
nects CG 12-S and CG 12-SW at velocities from 6.6 km s
1
to 6.2 km s
1
in Fig. 11 suggests that CG 12-SW might be
connected with CG 12-S.
7. Conclusions
We have performed a detailed, high signal-to-noise ratio, mm
line study of CG 12 in various molecular transitions, princi-
pally of C
18
O (1–0) and C
18
O (2–1), as well as in molecular
lines probing dense material, and have obtained the following
results:
1. The C
18
O line emission is distributed in a 10
North-
South elongated lane with two strong, compact maxima,
CG 12-N and CG 12-S, and a weaker maximum, CG 12-SW.
2. High density tracers CS (2–1), (3-2), DCO
+
and
H
13
CO
+
are detected in both strong C
18
O maxima. Emission
from these molecules is weak in CG 12-N but in CG 12-S it
defines a compact core (referred to as DCO
+
core) which is
spatially offset from the C
18
O maximum.
3. The emission from the high density tracers in
CG 12-N takes place at a velocity where emission from
C
18
O is weak. The molecules associated with the cloud com-
ponent detected in high density tracers are likely to be heavily
depleted.
4. Positive Matrix Factorization was applied to study
the cloud C
18
O fine scale structure. The observed strong
C
18
O emission in CG 12-N (PMF factor 2n) originates in the
envelope of the depleted cloud component or in a separate en-
tity seen in the same line of sight. In CG 12-S the most intense
PMF factor 3s traces warm gas on the surface of the DCO
+
core or a separate adjacent cloud component.
5. The driving source of the collimated molecular outflow
detected by White (1993) lies in the DCO
+
core.
6. The average C
18
O LTE mass is 80 M
for
CG 12-N, 70 M
for CG 12-S and 20 M
for CG 12-SW.
These numbers can only be considered as a zeroth or-
der estimate because of the uncertainty in defining the
C
18
O excitation temperature and possible molecular depletion
in the C
18
O maxima.
7. If the distance to CG 12, 630 pc, is correct the linear
size and the mass of this cometary globule approaches that of a
typical low mass star forming region like eg. Chamaeleon I.
Acknowledgements. We thank Bo Reipurth and the A&A Letters ed-
itor, Malcolm Walmsley for critically reading the manuscript and for
very useful comments thathelped to improve this paper. We also thank
Mika Juvela and Kalevi Mattila for helpful discussions and Pentti
Paatero for providing the PMF code to us. Data retrieved from the
Canadian Astronomy Data Centre were used to produce the CO out-
flow contours in Fig. 9. Canadian Astronomy Data Centre is operated
by the Herzberg Institute of Astrophysics, National Research Council
of Canada.
References
Arnal, E. M., Bajaja, E., Larrarte, J. J., Morras, R. & P¨oppel, W. G. L.
2000, A&AS, 142, 35
Haikala, L.K., Harju, J., Mattila, K. & Toriseva, M., 2005 A&A, 431,
149
Haikala, L.K., Juvela, M., Harju, J., Lehtinen, K., Mattila, K. &
Dumke, M., 2006 A&A, 454, L71
Hartman, D., & Burton, W. B. 1997, Atlas of Galactic Neutral
Hydrogen (Cambridge University Press), ISBN 0521471117
Hawarden, T.G., & Brand, P.W.J.L. 1976, MNRAS, 175, 19P
Herbst, E., 1982, A&A 111, 76
Herschel, J.F.W. 1847, Results of Astronomical Observations made
during the years 1834, 5, 6, 7, 8 at the Cape of Good Hope, Smith,
Elder (London)
Juvela, M., Lehtinen, K. & Paatero, P. 1996, MNRAS, 280, 616
Kalberla, P.M. W., Burton, W. B., Hartman, D., Arnal, E. M., Bajaja,
E., Morras, R. & P¨oppel, W. G. L. 2005, A&A, 440, 775
Maheswar, G., Manoj, P., & Bhatt, H.C. 2004, MNRAS, 355, 1272
Odenwald, S.F. ApJ, 325, 320
Rawlings, J.M.C, Taylor, S.D., & Willimas, D.A 2000, MNRAS, 313,
461
Reipurth,B.1983, A&A, 117, 183
Reipurth,B., Nyman, L-
˚
A, & Chini, R. 1996, A&A, 314, 258
12 L. K. Haikala and M. Olberg: The structure of the cometary globule CG 12
Rohlfs, K. & Wilson. T.L., 1999, Tools of Radio Astronomy, (Berlin:
Springer)
Russeil, D., Juvela, M., Lehtinen, K., Mattila, M., & Paatero, P. 2003,
A&A, 409, 135
Tafalla, M., Myers, P. C., Caselli, P., Walmsley, C. M., & Conito, C.
2002, ApJ, 569, 815
Santos, N.C., Yun, J.L., Santos, S.A. &Marreiros, R.G. 1998, AJ,116,
1376
Warin, S., Benayoun, J.J. & Viala, Y.P., 1996 A&A, 308, 535
White G. 1993 A&A, 274, L33
Williams, P.M., Brand, P.W.J.L., Longmore, A.J., Hawarden, T.G.
1977, MNRAS, 181, 179
Yonekura, Y., Hayakawa, T., Mizuno, N., Mine, Y., Mizuno, A.,
Ogawa, H. & Fukui, Y. 1999, PASJ, 51, 837
List of Objects
‘h4636’ on page 1
‘CoD –39
8581’ on page 1
‘NGC 5367’ on page 1
‘Cometary Globule 12’ on page 1
‘CG 12’ on page 1
‘IRAS 13547-3944’ on page 2
‘13546-3941’ on page 6
L. K. Haikala and M. Olberg: The structure of the cometary globule CG 12, Online Material p 1
Online Material
L. K. Haikala and M. Olberg: The structure of the cometary globule CG 12, Online Material p 2
List of Objects
Appendix A: Positive Matrix Factorization
analysis of CG 12
A.1. CG 12-N
The results of the PMF, assuming three factors, to observed
C
18
O (1–0) and (2–1) data in CG 12-N are shown in Figs. A.1
and A.2. The uppermost panels show the three fitted PMF fac-
tors. The area under each factor is normalized to one. The his-
togram superposed on the rightmost factor in Fig A.2 indicates
the velocity resolution of the input data. The intensity distri-
bution of the PMF factors in CG 12-N are shown in the lower
three panels.
The PMF reproduces well the velocity structure displayed
in Fig. 11. The fits to the C
18
O (1–0) and (2–1) data are inde-
pendentof each other. Therefore it is encouraging that the PMF
output factors for both C
18
O transitions are similar in velocity
and in spatial distribution. Because of the unambiguity of the
fits for both transitions the factors, centered at 6.6, 5.9 and
5.4 km s
1
, are referred to in the following as 1n, 2n and 3n,
respectively.
The PMF factors, 1n, 2n and 3n, are distributed in a north-
south oriented ridge, in a pear shaped body and a narrow
east-west ridge, respectively. The 3n factor lies at the ve-
locity where the emission from the H
13
CO
+
, DCO
+
, and
N
2
H
+
is at maximum. Even though this factor is not read-
ily evident in the individual spectra PMF finds it in both
C
18
O transitions. Factors 2n and 3n are symmetric but fac-
tor 1n has blue and redshifted wing like emission, especially
in the (1–0) transition. According to the PMF fit the high
T
A
(C
18
O (1–0))/T
A
(C
18
O (2–1)) ratio observed in the direc-
tion of CG 12-N is due to the factor 2n.
If the number of factors in the PMF is chosen to be four
PMF divides in essence the factor 1n into two, leaving the two
other factors untouched. The spatial distribution of the split up
factors is similar to that of the 1n in the three-factor PMF. If the
number of factorsis further increased to vethe resultis similar
to that with four factors plus a fifth factor which has nearly
zero intensity, ’empty’ field, in the map. The three factors are
therefore sufficient to produce the velocity structure evident in
Fig. 11 and increasing the number of factors does not improve
the fit significantly.
A.2. CG 12-S and CG 12-SW
The results of the PMF fits, assuming four factors, to both ob-
served C
18
O transitions in CG 12-S are shown in Figs. A.3 and
A.4. The location of the DCO
+
core is shown with a dashed
contour in the figures. The factors, centered at 6.8, 6.4,
6.2 and 5.8 km s
1
, will be referred to as 1s to 4s, respec-
tively.
The most redshifted factor 4s covers the very northern
part of the region and is close in velocity to the 2n factor in
CG 12-N. It is natural to consider that this factor is due to emis-
sion extending from CG 12-N to CG 12-S. Also the very faint
Fig.A.1. Upper panel: the basic spectral profiles (factors) cal-
culated by the PMF in CG 12-N (three factor fit), C
18
O (1–
0) data. Lower panels: maps of the weights of the basic fac-
tors. The lowest contour level and increment are 0.2 K km s
1
.
Offsets from the centre position of the original C
18
O map are
shown in arcminutes on the x and y axes.
emission observedin the northernpart of the fieldin other three
factors is at least partly due to emission from CG 12-N.
The emission from the most blueshifted factor1s is concen-
trated just below the (0,0) position. The 2s factor represents the
arc seen in Fig. 11 which connects CG 12-S and CG 12-SW .
There is a small size local maximum in the C
18
O (2–1) 2s fac-
tor north of the DCO
+
core. Even though this local maximum
is not seen in the C
18
O (1–0) 2s factor, there is extended emis-
sion at its location. The southern part of the arc is more intense
in the (1–0) transition than in the (2–1). The maximum of the
southern extension coincides with CG 12-SW. The emission
from the 3s factor peaks west of the DCO
+
core and is seen in
both transitions.
L. K. Haikala and M. Olberg: The structure of the cometary globule CG 12, Online Material p 3
Fig.11. T
A
channel map of the C
18
O (1–0) emission (upper three rows) and C
18
O (2–1) (lower three rows). The colour scale is
the same for both transitions. The LSR velocity is indicated in the upper left corners of the panels. Each pixel corresponds to a
single observed position but the C
18
O (2–1) data is binned in this figure to the same channel width as the (1–0) transition, 0.116
km s
1
. The highest intensity in the panels is 3.0 K. The cross in the panels is located in the map centre position.
L. K. Haikala and M. Olberg: The structure of the cometary globule CG 12, Online Material p 4
Fig.A.5. The PMF fit to the C
18
O lines in the selected positions positions in Fig. 5. The observed C
18
O (2–1) and (1–0) lines
are plotted with a heavy histogram and line, respectively. The green and red lines show the fitted PMF C
18
O (2–1) and (1–0)
factors, respectively. The residuals of the fits are shown in the lower part of each panel.
A.3. Individual line profiles
The PMF fits to the C
18
O lines in the same selected positions
as in Fig. 5 are shown in Fig. A.5. The residuals obtained by
subtracting the added up PMF factors from the observed lines
are shown below the lines in the figure. A grey scale map of
all the PMF fit residuals is shown in Fig. A.6. One should
stress that, unlike a gaussian multicomponent fit to a single
spectrum, PMF fits, in this case 3 (CG 12-N) or 4 (CG 12-S)
factors into all the input spectra simultaneously. Neither the
shape nor the velocity of the factors is fixed in the fit. The two
C
18
O transitions were fit separately and PMF could have used
factors differing in profile and velocity for the two transitions.
The fitted PMF factors are, however, similar both in shape and
velocity. The residuals shown in Figs. A.5 and A.6 are small
and demonstrate that PMF produces a good fit to the data.
The largest residual in Fig. A.5 takes place in the
blue shifted side of the C
18
O (1–0) line in the centre of
CG 12-N (position 40
′′
,160
′′
). This would imply that the
blue shifted wing of factor C
18
O (1–0) 1n is too strong. A
closer comparison of the input data and PMF results reveals
that similar deviations take place for three other C
18
O (1–
0) spectra around the position above. The C
18
O (1–0) line
profiles farther away from the CG 12-N centre actually do
have a more prominent blue shifted wing than the four above
mentioned spectra. The large number of spectra with good
fits outweighs the less optimal fit for the spectra in the cen-
tre of CG 12-N. The residuals shown in Fig. A.5 for the re-
maining four spectra in CG 12-S are smaller than in posi-
tion (40
′′
,160
′′
).
L. K. Haikala and M. Olberg: The structure of the cometary globule CG 12, Online Material p 5
Fig.A.2. As Fig. A.1 but for C
18
O (2–1) data. Factor 3n in
the upper panel has been plotted also as a histogram to indicate
velocity resolution of the input/output spectra.
Fig.A.3. Upper panel: the basic spectral profiles calculated
by the PMF in CG 12-S (four factor fit), C
18
O (1–0) data.
Lowerpanels: the maps of the intensity of the basic factors. The
factor numbers correspond to the numbers in the upper panel.
The lowest contour level and increment is 0.2 K km s
1
. The
dashed contour outlines the DCO
+
core.
L. K. Haikala and M. Olberg: The structure of the cometary globule CG 12, Online Material p 6
Fig.A.4. As Fig. A.3 but for C
18
O (2–1) data.
Fig.A.6. The PMF fit residuals in the K km s
1
scale.
... They show ★ E-mail: s.piyali16@gmail.com high number densities, 10 4 − 10 5 cm −3 Haikala & Olberg (2007); Vilas-Boas et al. (1994); Bourke et al. (1995), with a mass range of 10 − 100 M (Lefloch & Lazareff 1994;Haikala & Olberg 2007). A number of CGs are found to be sites of ongoing low-mass star formation (e.g., Reipurth 1983;Santos et al. 1998;Ogura et al. 2002;Maheswar et al. 2007;Getman et al. 2008;Rebull et al. 2013). ...
... They show ★ E-mail: s.piyali16@gmail.com high number densities, 10 4 − 10 5 cm −3 Haikala & Olberg (2007); Vilas-Boas et al. (1994); Bourke et al. (1995), with a mass range of 10 − 100 M (Lefloch & Lazareff 1994;Haikala & Olberg 2007). A number of CGs are found to be sites of ongoing low-mass star formation (e.g., Reipurth 1983;Santos et al. 1998;Ogura et al. 2002;Maheswar et al. 2007;Getman et al. 2008;Rebull et al. 2013). ...
Article
Full-text available
The Lynds’ Dark Nebula (LDN) 1615/1616 and CB 28 (hereafter L1616) together form a cometary globule located at an angular distance of about 8° west of the Orion OB1 association, aligned roughly along the east–west direction, and showing a distinct head–tail structure. The presence of massive stars in the Orion belt has been considered to be responsible for the radiation-driven implosion mode of star formation in L1616. Based on the latest Gaia Early Data Release 3 (EDR3) measurements of the previously known young stellar objects (YSOs) associated with L1616, we find the distance to this cloud to be 384 ± 5 pc. We present optical polarimetry towards L1616 that maps the plane-of-sky component of the ambient magnetic field (BPOS) geometry. Based on the proper motion of the YSOs associated with L1616, we investigate their plane-of-sky motion relative to the exciting star ϵ Ori. Using the Gaia EDR3 measurements of the distances and proper motions of the YSOs, we find two additional sources comoving with the known YSOs. One comoving source is HD 33056, a B9 star, and the other might be a young pre-main-sequence star not reported in previous studies. The mean direction of BPOS is found to follow the cloud structure. This could be the effect of dragging of the magnetic field lines by the impact of the ionizing radiation from ϵ Ori. Based on the pressure exerted on L1616, and the ages of the associated YSOs, we show that it could possibly be the main source of ionization in L1616, and thus the star formation in it.
... They show ★ E-mail: s.piyali16@gmail.com high number densities, 10 4 − 10 5 cm −3 Haikala & Olberg (2007); Vilas-Boas et al. (1994); Bourke et al. (1995), with a mass range of 10 − 100 M (Lefloch & Lazareff 1994;Haikala & Olberg 2007). A number of CGs are found to be sites of ongoing low-mass star formation (e.g., Reipurth 1983;Santos et al. 1998;Ogura et al. 2002;Maheswar et al. 2007;Getman et al. 2008;Rebull et al. 2013). ...
... They show ★ E-mail: s.piyali16@gmail.com high number densities, 10 4 − 10 5 cm −3 Haikala & Olberg (2007); Vilas-Boas et al. (1994); Bourke et al. (1995), with a mass range of 10 − 100 M (Lefloch & Lazareff 1994;Haikala & Olberg 2007). A number of CGs are found to be sites of ongoing low-mass star formation (e.g., Reipurth 1983;Santos et al. 1998;Ogura et al. 2002;Maheswar et al. 2007;Getman et al. 2008;Rebull et al. 2013). ...
Preprint
LDN 1615/1616 and CB 28 (hereafter, L1616) together form a cometary globule located at an angular distance of about 8 degrees west of the Orion OB1 association, aligned roughly along the east-west direction, and showing a distinct head-tail structure. The presence of massive stars in the Orion belt has been considered to be responsible for the radiation driven implosion mode of star formation in L1616. Based on the latest Gaia EDR3 measurements of the previously known young stellar objects (YSOs) associated with L1616, we find the distance to this cloud of 384$\pm$5 pc. We present optical polarimetry towards L1616 that maps the plane-of-sky component of the ambient magnetic field (B$_{POS}$) geometry. Based on the proper motion of the YSOs associated with L1616, we investigate their plane-of-sky motion relative to the exciting star $\epsilon$ Ori. Using the Gaia EDR3 measurements of the distances and proper motions of the YSOs, we find two additional sources comoving with the known YSOs. One comoving source is HD33056, a B9 star and the other might be a young pre-main sequence star not reported in previous studies. The mean direction of B$_{POS}$ is found to follow the cloud structure. This could be the effect of dragging of the magnetic field lines by the impact of the ionizing radiation from $\epsilon$ Ori. Based on the pressure exerted on L1616, and the ages of the associated YSOs, we show that it could possibly be the main source of ionization in L1616, and thus the star formation in it.
Article
The cometary globule CG 12 and its associated young cluster NGC 5367 is located at a Galactic latitude of b = 21°, and the origin of this high-latitude star formation event has been unclear. Getman et al. identified a nearby group of B stars and suggested they could be the remnant of a group of massive stars of which one exploded as a supernova. We here use Gaia EDR3 data to measure the accurate distance of NGC 5367 and the B-stars. We find a mean distance of 670 pc for NGC 5367 and 675 pc for the B-stars, strongly supporting the suggestion of Getman et al.
Article
Full-text available
We present results of our R-band polarimetry of a cometary globule, LBN 437 (Gal 96-15, ℓ = 96°, b = −15°), to study the magnetic field geometry of the cloud. We estimated a distance of 360 ± 65 pc to LBN 437 (also one additional cloud, CB 238) using a near-IR photometric method. The foreground contribution to the observed polarization values was subtracted by making polarimetric observations of stars that are located in the direction of the cloud and with known distances from the Hipparcos parallax measurements. The magnetic field geometry of LBN 437 is found to follow the curved shape of the globule head. This could be due to the drag that the magnetic field lines could have experienced because of the ionization radiation from the same exciting source that caused the cometary shape of the cloud. The orientation of the outflow from the Herbig A4e star, LkHα 233 (or V375 Lac), located at the head of LBN 437, is found to be parallel to both the initial (prior to the ionizing source was turned on) ambient magnetic field (inferred from a star HD 214243 located just in front of the cloud) and the Galactic plane.
Article
Context. Cometary globule 12 is a relatively little investigated medium- and low mass star forming region 210 pc above the Galactic plane. Aims: This study sets out to discover the possibly embedded members of the CG 12 stellar cluster, to refine the NIR photometry of the known member stars and to study the star formation activity in CG 12 and its relation to the distribution of molecular gas, dust and mid- to far-infrared emission in the cloud. Methods: NIR J, H, and Ks imaging and stellar photometry is used to analyse the stellar content and the structure of CG 12. Results: Several new members and member candidates of the CG 12 stellar cluster were found. The new members include in particular a highly embedded source with a circumstellar disk or shell and a variable star with a circumstellar disk which forms a binary with a previously known A spectral type cluster member. The central source of the known collimated molecular outflow in CG 12 and an associated “hourglass”-shaped object due to reflected light from the source were also detected. The maximum visual extinction in the cloud, based on observations of background stars, is ~20m, but this is only a lower limit for the extinction through the two dense cloud cores. HIRES-enhanced IRAS images are used together with JHKs imaging to study the two associated IRAS point sources, 13546-3941 and 13547-3944. Two new 12 μ m sources coinciding with NIR excess stars were detected in the direction of IRAS 13546-3941. The IRAS 13547-3944 emission at 12 and 25 μ m originates in the Herbig AeBe star h4636n and the 60 and 100 μ m emission from an adjacent cold source. Based on observations collected at the European Southern Observatory, La Silla, Chile.Table 1 is only available in electronic form at the CDS via anonymous ftp to cdsarc.u-strasbg.fr (130.79.128.5) or via http://cdsweb.u-strasbg.fr/cgi-bin/qcat?J/A+A/510/A1
Article
Full-text available
The mysterious high galactic latitude cometary globule CG12 has been observed with the ACIS detector on board the Chandra X-ray Observatory. We detect 128 X-ray sources; half are likely young stars formed within the globule's head. This new population of >50 T-Tauri stars and one new embedded protostar is far larger than the previously reported few intermediate-mass and two protostellar members of the cloud. Most of the newly discovered stars have masses 0.2-0.7Mo, and 9-15% have K-band excesses from inner protoplanetary disks. X-ray properties provide an independent distance estimate consistent with CG12's unusual location >200pc above the Galactic plane. The star formation efficiency in CG12 appears to be 15-35%, far above that seen in other triggered molecular globules. The median photometric age found for the T-Tauri population is ~4Myr with a large spread of <1-20Myr and ongoing star formation in the molecular cores. The stellar age and spatial distributions are inconsistent with a simple radiation driven implosion (RDI) model, and suggest either that CG12 is an atypically large shocked globule, or it has been subject to several distinct episodes of triggering and ablation. We report a previously unnoticed group of B-type stars northwest of CG12 which may be the remnants of an OB association which produced multiple supernova explosions that could have shocked and ablated the cloud over a 15-30Myr period. HD120958 (B3e), the most luminous member of the group, may be currently driving an RDI shock into the CG12 cloud. Comment: 58 pages, 12 figures, 5 tables, submitted to ApJ 05/08/07
Article
Full-text available
We have conducted a search for molecular clouds toward southern IRAS point sources at intermediate-to-high galactic latitude (|b| ≥ 10° in 12CO (J = 1−0) emission with the NANTEN telescope. The main purpose of the survey was to find a new sample of star-forming clouds unknown to date. Of the 29 targets, we detected CO emission toward 5 IRAS sources, i.e., 04591–0856, 05044–0325, 05050–0614, 06345–3023, and 13543–3941. Among 5 detections, the molecular cloud associated with 06345–3023 (G 239.2–16.3) was detected for the first time. If we assume the distance as being 1.5 kpc, this\ source is ∼ 430 pc away from the galactic plane. Three sources 04591–0856 (G 208.3–28.4), 05044–0325 (G 203.5–24.7), and 05050–0614 (G 206.4–25.9) are located west to the Orion molecular clouds, and 13543-3941 (G 316.4+21.2) is located in the cometary globule CG 12. Four of the 5 clouds show a cometary shape with a “head-tail” distribution, and three of them have the IRAS source located at the edge opposite to the tail. We find that all of the CO clouds are associated with the optical nebulosities, which are likely to be reflection nebulae. A comparison of the present CO clouds with those in Ophiuchus, Taurus, Chamaeleon, and L 1333 indicates that star-forming clouds tend to have a high column density, as well as a smaller ratio of the virial mass to the LTE mass.
Article
This paper reports on a high sensitivity lambda 21-cm neutral hydrogen survey of the sky south of delta <= -25o. A total of 50980 positions lying on a galactic coordinate grid with points spaced by (bigtriangleup l, bigtriangleup b) = (0\fdg5/cos b, 0\fdg5), were observed with the 30 m dish of the Instituto Argentino de Radioastronomía (IAR). The angular resolution of the survey is HPBW = 30\arcmin, and the velocity coverage spans the interval -450 km s-1 to +400 km s-1 (LSR). The velocity resolution is 1.27 km s-1 and the final rms noise of the entire database is ~ 0.07 K. The brightness temperature scale is accurate to 5%.
Article
Tailed, comet-like objects found in the outskirts of the Gum nebula (and elsewhere) are shown to be $\gt\,{10}^{5}$ yr old. This implies that the H II region is much older than the Vela Pulsar 0833 – 45. Tailed structure found in NGC 5367 suggests the presence of ionized hydrogen falling into the galactic plane near l = 320°.
Article
The high latitude translucent molecular cloud L1642 has been mapped in the J=1-0 and J=2-1 transitions of 12CO, 13CO and C18O using the SEST radio telescope. We have analysed the morphology and velocity structure of the cloud using the Positive Matrix Factorization (PMF) method. The results show that L1642 is composed of a main structure at radial velocity 0.2 km s-1 while the higher velocity components at ~ 0.5 and 1.0 km s-1 form an incomplete ring around it, suggesting an expanding shell structure. Fainter emission extends to the north with a still higher velocity of up to 1.6 km s-1. Such a velocity structure suggests an elongated morphology in the line of sight direction. The physical properties of the cloud have been investigated assuming LTE conditions, but non-LTE radiative transfer models are also constructed for the 13CO observations. We confirm that L1642 follows an r-1 density distribution in its outer parts while the distribution is considerably flatter in the core. The cloud is close to virial equilibrium. In an Appendix the PMF results are compared with the view obtained through the analysis of channel maps and by the use of Principal Component Analysis (PCA). Both PMF and PCA present the observations as a linear combination of basic spectral shapes that are extracted from the data. Comparison of the methods shows that the PMF method in particular is able to produce a presentation of the complex velocity that is both compact and easily interpreted. Based on observations collected at the European Southern Observatory. Appendices A and B are only available in electronic form at http://www.edpsciences.org
Article
The authors present an analysis of 12CO spectra of the Thumbprint Nebula using a new method, positive matrix factorization (PMF). The complexity of the observed spectra and especially the non-Gaussian nature of the profiles make traditional analysis using fitted Gaussians impractical. Good results are obtained, however, with the use of PMF. PMF has already been applied to a large variety of problems and the authors present the first application of this method to the analysis of molecular emission line measurements. The method assumes the spectra to be the sum of a few basic components in velocity space and can determine the shapes as well as the intensities of the components. The only restriction is the positivity of both the basic profiles and their intensities. This prerequisite is sufficient to limit the range of possible solutions and to guarantee results that are easier to interpret than those determined by similar traditional methods, principal component analysis and factor analysis. PMF is also able to point out the existence of features that are contrary to the model assumptions, e.g. the existence of velocity gradients. In this article the authors explain the principles of PMF and illustrate its use with some simulated data before turning to the analysis of the 12CO spectra of the Thumbprint Nebula.
Article
We have developed a detailed interstellar cloud model that takes into account the chemistry among simple carbon and oxygen bearing molecules, as well as ^13^C and ^18^O isotopic compounds. The rotational population of H_2_, CO, ^13^CO and C^18^O and the fine structure excitation of C, C^+^ and O are controlled by chemical processes, including selective photodissociation which depends on the rotational quantum number J, collisional and radiative processes allowing for photon trapping of CO millimeter and sub-millimeter lines. The gas and grain temperature distributions are obtained by solving a thermal equilibrium equation for each component. Comparison with the previous model calculations of van Dishoeck & Black (1988) and Le Bourlot et al. (1993) are made throughout the paper. We have computed series of models for three kinds of quiescent interstellar clouds: diffuse, translucent and dense dark clouds. The variations of the photodissociation rates are mainly controlled by self-shielding effects leading to a rotational selective photodissociation, i.e. a strong dependence on the rotational quantum number J: the more populated the level the more efficient is the self-shielding and the more rapid is the decrease of its photodissociation rate. This implies a strong coupling between photodissociation and rotational excitation, which has for main consequence to overpopulate low-lying levels and to under populate excited levels with respect to an LTE calculation. Under the combined influence of the rotational photodissociation and the inefficiency of collisions to populate excited levels, at the densities and temperatures prevailing in most clouds, the rotational population of CO and its isotopes is sub-thermally excited, except for the first two rotational levels of the three isotopes and the levels J=2 and 3 of the main isotope which are thermalized in the inner part of dark clouds. Photo-electric emission of electrons from grains is the dominant heating process in regions where UV photons are present, i.e. in diffuse and translucent clouds and in the outer parts of dark clouds. It is replaced by cosmic ray ionization of atomic and molecular hydrogen as a dominant heating process in the core of dark clouds. The cooling efficiency is entirely correlated to the cloud composition, with cooling dominated by fine structure excitation of C^+^ and rotational excitation of CO in diffuse and dark clouds, respectively; fine structure excitation of C contributes to the cooling of translucent clouds. The strong dependence of the temperature distribution in interstellar clouds on the external UV radiation field is emphasized. Our model calculations have been confronted with selected CO surveys of molecular clouds by Frerking et al. (1982), Cernicharo & Guelin (1987) and Lada et al. (1994). For that purpose, we have run series of model with various visual extinctions and densities covering a large range of values, from diffuse to dense cloud conditions. The temperature distribution was obtained by solving the thermal balance equation. The model predicts emissivities of the 1->0 line of CO well below the observations with the exception of the C^18^O(1->0) line in the the molecular clouds surveyed by Frerking et al. (1982) for which agreement between theory and observations is very good. Although the model predicts too low abundances of both ^13^CO and C^18^O, the column density ratio R(^13^CO/C^18^O) normalized to the isotopic ratio that we compute is in fairly good qualitative agreement with the one derived by Lada et al (1994) in the IC 5146 molecular cloud. The model calculations have also been used to predict the emissivities of the CO, ^13^CO, C^18^O, C and C^+^ lines under various interstellar conditions.
Article
Optical and far-IR studies have been conducted on 14 high Galactic latitude clouds with comet-like or filamentary appearances. Their physical characteristics are deduced and found to be comparable to those of small, low-mass molecular clouds found in the solar neighborhood in recent CO and H I surveys. The morphologies of five of the clouds are found to be consistent with objects moving supersonically through the interstellar medium. These clouds are also active in forming B-type stars in their nuclei where the dynamic pressures are expected to be highest. The filamentary clouds have shapes consistent with subsonic, low Reynolds number hydrodynamics, a feature they may share with the Draco molecular cloud. A simple qualitative scenario is also presented, describing a possible evolutionary link among these clouds as a population.
Article
ABSTRACTA method for determining distances to dark clouds and Bok globules based on broad-band optical and near-infrared photometry is presented. In this method, intrinsic colour indices of stars projected towards the direction of a cloud are computed by dereddening the observed colour indices using various trial values of extinction AV and a standard extinction law. The computed intrinsic colour indices for a star are then compared with the intrinsic colour indices of normal main-sequence stars and a spectral type is assigned to the star for which the computed colour indices best match the standard intrinsic colour indices. Distances (d) to the stars are determined using the AV and absolute magnitudes (MV) corresponding to the spectral types thus obtained. A plot of AV against d undergoes a sharp rise at a distance corresponding to the distance to the cloud. Using this method, we have determined a distance of 550 pc to the cometary globule CG 12. The distance of 550 pc and a Galactic latitude of b= 21° imply that CG 12 is at a height of ∼200 pc above the galactic mid-plane. The star formation efficiency in this cloud is found to be relatively high (≳16 per cent, to as large as ∼33 per cent). The existence of an H i shell centred at l= 315°, b= 30° with CG 12 near its boundary and its tail pointing away from the centre of the shell supports the suggestion of a supernova explosion, near the centre of the H i shell, being responsible for the cometary morphology and the triggering of star formation. Thus CG 12 is a rare example of triggered high-mass star formation at relatively large galactic height.
Article
Anomalous molecular line profile shapes are the strongest indicators of the presence of the infall of gas that is associated with star formation. Such profiles are seen for well-known tracers, such as HCO+, CS and H2CO. In certain cases, optically thick emission lines with appropriate excitation criteria may possess the asymmetric double-peaked profiles that are characteristic of infall. However, recent interpretations of the HCO+ infall profile observed towards the protostellar infall candidate B335 have revealed a significant discrepancy between the inferred overall column density of the molecule and that which is predicted by standard dark cloud chemical modelling. This paper presents a model for the source of the HCO+ emission excess. Observations have shown that, in low-mass star-forming regions, the collapse process is invariably accompanied by the presence of collimated outflows; we therefore propose the presence of an interface region around the outflow in which the chemistry is enriched by the action of jets. This hypothesis suggests that the line profiles of HCO+, as well as other molecular species, may require a more complex interpretation than can be provided by simple, chemically quiescent, spherically symmetric infall models. The enhancement of HCO+ depends primarily on the presence of a shock-generated radiation field in the interface. Plausible estimates of the radiation intensity imply molecular abundances that are consistent with those observed. Further, high-resolution observations of an infall-outflow source show HCO+ emission morphology that is consistent with that predicted by this model.