THE ASTROPHYSICAL JOURNAL, 544 : 881È888, 2000 December 1
( 2000. The American Astronomical Society. All rights reserved. Printed in U.S.A.
MILLIMETER OBSERVATIONS OF VIBRATIONALLY EXCITED CS TOWARD IRC ]10216 : A NEW
CIRCUMSTELLAR MASER ?
J. L. HIGHBERGER, A. J. APPONI,1 J. H. BIEGING, AND L. M. ZIURYS
Department of Astronomy, Department of Chemistry, and Steward Observatory, 933 North Cherry, University of Arizona, Tucson, AZ 85721
AND
J. G. MANGUM
National Radio Astronomy Observatory, 949 North Cherry, Tucson, AZ 85721
Received 2000 March 29 ; accepted 2000 July 24
ABSTRACT
New observations of vibrationally excited CS in its v \ 1 state have been conducted toward the circumstellar shell of IRC ]10216 using the NRAO 12 m telescope and the CSO. The J \ 3 ] 2, 6 ] 5,
and 7 ] 6 transitions in the v \ 1 level at 146, 292, and 340 GHz have been detected toward this object,
and the J \ 2 ] 1 and 5 ] 4 lines were also reobserved. After accounting for contamination from other
species, these CS transitions all exhibit line proÐles that are narrower (*V D 10 km s~1) than those
displayed by most molecules in IRC ]10216, indicating an origin in the1@2inner stellar envelope. The
J \ 3 ] 2 line is particularly narrow (*V D 1È2 km s~1), and consists of two separate velocity components at V D [26 and [22 km s~1.1@2These characteristics are suggestive of weak maser action in
LSRThe narrow line widths observed in the other transitions indicate a source size [ 0A. 5,
this transition.
which implies a CS column density in the v \ 1 state of D 2È6 ] 1016 cm~2. The ground-state (v \ 0)
column density is therefore on the order of 0.8È2 ] 1018 cm~2 for T \ 500 K. The fractional abundance of CS relative to H is consequently f D 3È7 ] 10~5 at D 14 R vib
. Such a high abundance close to
*
the stellar photosphere is 2strong evidence that CS is a ““ parent ÏÏ molecule
in IRC ]10216, as suggested
by infrared studies of the v \ 0 ] 1 vibrational transition and interferometric observations.
Subject headings : circumstellar matter È masers È radio lines : stars È stars : AGB and post-AGB È
stars : individual (IRC ]10216)
1.
INTRODUCTION
HCN, SiH , and SiO (e.g., Keady, Hall, & Ridgway 1988 ;
4
Keady & Ridgway
1993 ; Boyle et al. 1994 ; Cernicharo et al.
1996). These absorption lines arise in the inner part of the
circumstellar shell, near the dust acceleration zone. Rotational transitions of molecules in various vibrationally
excited states have been measured toward IRC ]10216 as
well, such as HCN (Ziurys & Turner 1986), CS (Turner
1987a), and SiS (Turner 1987b). Certain transitions even
appear to exhibit maser emission, as has been found in the
J \ 1 ] 0 line of HCN originating in the (02¡0) bending
state (Lucas et al. 1986). This masering transition has been
detected in six other stellar envelopes, including CIT6
(Lucas, Guilloteau, & Omont 1988).
Vibrationally excited species in IRC ]10216 are thought
to originate in the inner circumstellar envelope, as evidenced by their narrow line widths relative to ground state
species. For example, lines arising from the v \ 1 state of
SiS have zero-power line widths of *V D 20È25 km s~1
zp ]10216 values
(Turner 1987b), in contrast to the usual IRC
of *V D 30 km s~1 (Olofsson 1994). Thus, vibrationally
zp emission originates from gas that has not been
excited
accelerated to the full expansion velocity of 14.5 km s~1
(Olofsson 1994). One exception to this general case has been
vibrationally excited CS. This species was initially detected
toward IRC ]10216 by Turner (1987a), who observed the
J \ 2 ] 1 and J \ 5 ] 4 transitions in the v \ 1 state. The
J \ 5 ] 4 line proÐle measured by Turner was particularly
unusual, being broad and asymmetric in shape with an
apparent ““ dip ÏÏ toward the blueshifted side, attributed
to self-absorption. The anomalous line shape led Turner to
the conclusion that vibrationally excited CS must arise
from throughout the circumstellar envelope of IRC
]10216, and the emission must therefore be thermal.
Vibrationally excited molecules in interstellar gas are of
interest because of their potential as probes of extreme
physical conditions. Such species can be excited by collisions, in which case they trace regions of very high gas
temperatures and densities (e.g., Ziurys & Turner 1986).
Excitation by infrared radiation is also a possibility for
these molecules, and thus they may indicate the presence of
substantial infrared Ðelds (e.g., Goldsmith et al. 1983).
Various vibrationally excited species have been observed
in molecular clouds, the majority of which have been
detected via rotational transitions arising in such excited
states. For example, rotational lines of CH CN, and HC N
3 from their3v
have been identiÐed in Orion-KL originating
8
and v modes, respectively (Goldsmith et al. 1983). Vibra7
tionally excited NH in its v \ 1 state has been observed in
this object as well 3(Schilke2 et al. 1992). Such molecules
appear to undergo vibrational excitation in the ““ hot core ÏÏ
region as a result of infrared dust emission. In addition,
vibrationally excited H has been detected in numerous
2
dense clouds such as DR21
and W51 (e.g., Garden et al.
1991). In this case, S-branch lines of the v \ 1 ] 0 vibrational transition are observed, which are thought to trace
shock waves and/or outÑows.
Vibrationally excited molecules are also present in the
circumstellar envelopes of late-type stars, especially in the
shell of the carbon-rich giant IRC ]10216. Numerous
species have been detected in absorption toward IRC
]10216 via ro-vibrational IR lines, including CO, HCCH,
1 Current address : Division of Engineering and Applied Science,
Harvard University, 29 Oxford Street, Cambridge, MA 02138.
881
882
HIGHBERGER ET AL.
Vol. 544
TABLE 1
SUMMARY OF OBSERVATIONS TOWARD IRC ]10216a
C33S (v \ 0) . . . . . . . . . .
C H (vl \ 22) . . . . . . . .
4
7
Si13CC . . . . . . . . . . . . . . .
HC N (v \ 1)g . . . . . .
3
7
SiC (3% )e . . . . . . . . . . . .
1
U......................
C34Sf,h . . . . . . . . . . . . . . . .
CNf . . . . . . . . . . . . . . . . . . .
30SiC f . . . . . . . . . . . . . . .
2
Uf,h . . . . . . . . . . . . . . . . . . .
Uf,h . . . . . . . . . . . . . . . . . . .
T *b
R
(K)
V
LSR
(km s~1)
*V
1@2
(km s~1)
26
63
43
C
0.46
0.88
0.76
26
22
24
...
...
...
...
...
...
...
...
...
...
...
...
0.47
0.36
0.67
...
...
...
...
...
...
...
...
...
...
...
...
4.00 ^ 0.10
0.003 ^ 0.001
0.054 ^ 0.010
0.020 ^ 0.010d
0.021 ^ 0.005
0.027 ^ 0.009
0.10 ^ 0.05
0.024 ^ 0.001
0.058 ^ 0.005
0.003 ^ 0.001
0.009 ^ 0.001
0.006 ^ 0.003
0.011 ^ 0.005
0.011 ^ 0.006
0.35 ^ 0.03
0.50 ^ 0.03
0.08 ^ 0.03
0.08 ^ 0.03
0.05 ^ 0.03
[25.9 ^ 1.2
[27.5 ^ 3.1
[26.2 ^ 0.3
[22.2 ^ 0.6
[25.1 ^ 1.2
[25.8 ^ 2.0
[27.2 ^ 2.7
[26.3 ^ 3.1
[25.7 ^ 1.2
[25.5 ^ 3.1
[26.1 ^ 3.1
[25.8 ^ 1.0
D [26
[26.0
[25.9 ^ 2.7
[25.5 ^ 2.7
[26.1 ^ 2.7
[26.0
[26.0
24.5 ^ 1.2
28.9 ^ 6.2
1.1 ^ 0.3
2.5 ^ 0.6
7.3 ^ 1.2
14.3 ^ 2.0
11.5 ^ 5.3
30.8 ^ 3.1
27.1 ^ 1.2
32.2 ^ 6.2
24.4 ^ 6.2
22.9 ^ 1.0
D 35
25.9 ^ 1.2
D 25
27.2 ^ 5.3
22.0 ^ 5.3
D 20
D 25
h
b
(arcsec)
J\5]4
J\2]1
J \ 3 ] 2c
244935.6
97270.9
145904.2
J \ 5 ] 4e
J\6]5
J \ 7 ] 6f
J\2]1
J\5]4
J \ 10.5 ] 9.5
J(K , K ) \ 4(1,3) ] 3(1,2)
a c
J \ 16 ] 15
J\6]5
...
J\7]6
N \ 3 ] 2, J \ 7/2 ] 5/2i
J(K , K ) \ 15(10,5) ] 14(10,4)
a c
...
...
243160.8
291782.3
340398.1
97171.8
242915.6
97244.5
97295.3
145918.2
243176.9
242970
337396.5
340247.9
340463.2
337167
337201
Transition
CS (v \ 0) . . . . . . . . . . . .
CS (v \ 1) . . . . . . . . . . . .
g
l
(MHz)
Molecule
a Measured with 2 MHz resolution unless otherwise noted. Quoted errors are 3 p.
b Assumes a unity Ðlling factor.
c Measured with 250 kHz resolution.
d Second velocity component.
e Blended lines (see text).
f Measured with Caltech Submillimeter Observatory ; resolution is 1 MHz and temperature scale is T *.
A
g Measured with 500 kHz resolution.
h From image sideband.
i Three blended hf components.
More recent studies, however, suggest that CS has a
much more conÐned distribution. Plateau de Bure interferometer (PdBI) maps show that most of the ground-state CS
emission arises from the inner envelope in a region D 10@@ in
extent, as opposed to the 40@@ outer shell where photochemistry occurs (Lucas et al. 1995). In addition, infrared observations of the v \ 0 ] 1 transition of CS near 12 km
indicate a high abundance (f D 10~6) near the star at r ¹ 12
R (Keady & Ridgway 1993), with a decreasing abundance
* to 100 R . (The outer envelope is at r D 500 R .) The
out
* of CS (v \ 1), observed in the Cernicharo,
*
J \ 3 ] 2 line
Guelin, & Kahane (2000) spectral survey of IRC ]10216,
exhibits a noticeably narrow proÐle as well, again suggesting an origin within the dust acceleration zone, and preliminary PdBI maps of the J \ 5 ] 4 (v \ 1) transition indicate
a source size \ 1@@ (Lucas & Guelin 1999).
In an e†ort to resolve these discrepancies, we have conducted an investigation of vibrationally excited CS toward
IRC ]10216, using the NRAO 12 m telescope and the 10 m
CSO dish. We have observed the J \ 3 ] 2, 6 ] 5, and
7 ] 6 lines of this molecule in the v \ 1 state, and have
remeasured the J \ 2 ] 1 and J \ 5 ] 4 transitions
detected by Turner. The spectra in general have narrow line
widths, indicative of an inner shell origin, and the unusual
TABLE 2
UPPER LIMITS FOR CS (v \ 1) TOWARD YOUNG STELLAR OBJECTS
d
(1950)b
V
LSR
(km s~1)
20 37 14.2
42 12 10
[3.0
IRAS 16293[2422 . . . . . . . .
16 29 20.9
[24 22 13
4.0
NGC 1333ÈIRAS 4A . . . . . .
03 26 5.0
31 03 13
7.0
Serpens-SMM4 . . . . . . . . . . . .
18 27 24.7
01 11 10
8.0
Orion A . . . . . . . . . . . . . . . . . . . .
05 32 47
[05 24 21
9.0
Source
a
(1950)a
DR21(OH) . . . . . . . . . . . . . . . . .
a Right ascension is measured in hours, minutes, and seconds.
b Declination is measured in degrees, arcminutes, and arcseconds.
Transition
rms
(mK)
Spectral Resolution
(MHz)
J\3]2
J\5]4
J\6]5
J\3]2
J\5]4
J\6]5
J\3]2
J\5]4
J\6]5
J\3]2
J\6]5
J\5]4
2.9
7.8
8.2
9.7
7.6
8.1
4.3
13.2
6.6
9.7
12.0
27.4
0.5
0.5
0.5
0.1
2.0
0.5
0.5
0.5
0.5
0.1
0.5
0.5
No. 2, 2000
A NEW CIRCUMSTELLAR MASER ?
J \ 5 ] 4 proÐle was found to be a result of a contaminating molecule. The J \ 3 ] 2 transition, moreover,
appears to consist of two anomalously narrow velocity
components and may be nonthermal in origin. In this paper
we discuss these results.
2.
OBSERVATIONS
The majority of the CS measurements toward IRC
]10216 were carried out with the National Radio
Astronomy Observatory (NRAO)2 12 m radio telescope at
Kitt Peak, Arizona, between 1996 February and 1997 May.
The receivers used were dual-channel cooled SIS mixers
operated in single sideband (SSB) mode at 1.2, 2, and 3 mm
wavelengths. The 2 and 3 mm receivers were tuned to be
SSB, while at 1 mm a Martin-Puplet interferometer was
used to reject the image sideband. The backends employed
were 256 channel Ðlter banks used in parallel mode (2 ] 128
channels) for simultaneous observations of orthogonal
polarizations. Two di†erent spectral resolutions were used
for each set of the observations. The temperature scale T *
R
was determined by the chopper wheel method, corrected for
forward spillover losses. The conversion to radiation temperature, T , is T \ T */g , where g is the corrected beam
R
R
R c transitions
c of CS (v \ 1) were
efficiency. Four
rotational
observed toward IRC ]10216 with the 12 m telescope at
97, 146, 243, and 292 GHz, in beam-switching mode. The
CS (v \ 0) J \ 5 ] 4 line was also measured at 245 GHz.
Observing frequencies, beam sizes, and beam efficiencies are
listed in Table 1 for the respective transitions. Typical Ðlter
bank resolutions employed were 1 MHz/2 MHz, used in
parallel mode, except for the J \ 3 ] 2 transition (see Table
1). The measurements were conducted toward the IRC
]10216 position a \ 9h45m14s. 8, d \ 13¡30@40@@ (1950). In
addition, a simple map of this source was made at 146 GHz
in 1997 May. Searches for vibrationally excited CS via its
J \ 3 ] 2, 5 ] 4, and 6 ] 5 transitions were also conducted
toward several molecular clouds thought to contain young
stellar objects, as noted in Table 2. These observations were
done in position-switching mode using 0.1, 0.25, 0.5, and 2
MHz resolutions, depending on the source.
The CS (v \ 1) J \ 7 ] 6 transition at 340 GHz was
measured at the Caltech Submillimeter Observatory
(CSO)3 at Mauna Kea, Hawaii on 1996 May 8. (The transition has been previously observed by Groesbeck et al.
1994). A single-channel, double sideband SIS receiver was
used for the measurements, with a 500 MHz bandwidth
AOS backend (1 MHz resolution). The CSO temperature
scale is T * such that conversion to radiation temperature is
A , where g is the main beam efficiency. Only IRC
T \ T */g
R
A
b observed
b at this frequency.
]10216 was
3.
883
Moreover, the CS (v \ 1) line widths, determined from
Gaussian Ðts to the proÐles, are narrow for this object. For
example, the J \ 7 ] 6, 6 ] 5, and 5 ] 4 lines have widths
of *V D 7È14 km s~1. (The J \ 2 ] 1 transition is an
1@2 but here the signal-to-noise was poor.) In conexception,
trast, the CS (v \ 0), J \ 5 ] 4 transition, C33S lines, and
other molecules observed (e.g., CN, SiC ) have *V D
2
1@2
23È32 km s~1 (see Table 1). These latter values are typical
for molecules observed in the circumstellar envelope where
the gas has reached the systemic velocity of 14.5 km s~1
(e.g., Olofsson 1994). The J \ 3 ] 2 transition displays an
even more peculiar proÐle, with *V \ 1.1 and 2.5 km s~1
1@2
for the two velocity components.
Typical spectra for vibrationally excited CS in IRC
]10216 are shown in Figure 1. Here the J \ 7 ] 6, 6 ] 5,
and 2 ] 1 transitions are presented. The J \ 6 ] 5 and
7 ] 6 lines show symmetric, parabolic line shapes, as
opposed to the square and Ñat-topped or cusped proÐles
normally observed for optically thin molecular emission in
this object. These shapes suggest that CS (v \ 1) emission is
optically thick and unresolved in the 22@@È24@@ telescope
beams at 292 and 340 GHz ; therefore, it does not likely
arise from the outer envelope, whose extent is 40@@È60@@
RESULTS
The results of the observations of IRC ]10216 are summarized in Table 1. Besides rest frequencies and telescope
parameters, Table 1 lists the line intensities (in units of T *),
R
LSR velocities, and line widths (at half-power) for the
observed CS transitions, as well as for other features
detected in the bandpasses. Typical intensities observed for
the CS (v \ 1) lines are on the order of 5È100 mK, with LSR
velocities near the expected [26 km s~1 for IRC ]10216.
2 The NRAO is operated by Associated Universities, Inc., under cooperative agreement with the National Science Foundation.
3 The CSO is funded by NSF grant AST96-15025.
FIG. 1.ÈSpectra of the J \ 2 ] 1, J \ 6 ] 5, and J \ 7 ] 6 transitions
of CS in its v \ 1 state at 97, 292, and 340 GHz, measured toward IRC
]10216. The J \ 2 ] 1 and J \ 6 ] 5 lines were observed with the 12 m
telescope using 2 MHz resolution and the J \ 7 ] 6 line was measured
using the CSO. (The temperature scale for this transition is T *.) The
individual transitions are indicated by arrows. The J \ 7 ]A6 and
J \ 6 ] 5 line proÐles appear parabolic in shape, suggesting an unresolved, optically thick source.
884
HIGHBERGER ET AL.
(Guelin, Lucas, & Neri 1997). (The signal-to-noise on the
J \ 2 ] 1 transition is not sufficient to extract reliable line
shape information.)
Figure 2 presents the J \ 5 ] 4 line of CS (v \ 1), which
exhibits an asymmetric line shape toward the blue side of
the proÐle, as also observed by Turner (1987a). However,
this asymmetry is readily attributable to another line
blended in with that of CS. As the Ðgure illustrates, this
proÐle can be Ðtted with two Gaussian curves, one at the
frequency of CS, and another at 16 MHz higher in frequency, 243.176 GHz. This is the exact frequency of the
J \ 6 ] 5 transition of SiC in its ) \ 1 ladder (Cernicharo
et al. 1989). (It should be noted that the SiC frequencies
were not known at the time of TurnerÏs observations.)
Although the lowest lying spin orbit component of SiC corresponds to ) \ 2, the J \ 4 ] 3 lines in the ) \ 1 ladder
have been observed by Cernicharo et al. (2000) at 2 mm
toward IRC ]10216. Hence it is expected that the
J \ 6 ] 5 () \ 1) lines should be present in this source. As
the line proÐle and its corresponding parameters in Table 1
show, subtracting the SiC ““ shoulder ÏÏ leaves a narrow CS
feature similar to the other observed transitions.
Figure 3 displays the spectrum observed for the
J \ 3 ] 2 transition of CS (v \ 1). This line also has been
detected in the 2 mm survey of Cernicharo et al. (2000).
However, because these authors used 1 MHz resolution, the
unusual line proÐle shown here at 250 kHz resolution was
not apparent in their data. This line was originally observed
at the 12 m in 1996 February, with coarser resolution,
where it appeared to be asymmetric in shape. It was reobserved in 1996 December, with the 250 kHz Ðlters, such that
Vol. 544
FIG. 3.ÈSpectrum of the J \ 3 ] 2 line of CS (v \ 1) near 146 GHz
observed toward IRC ]10216 with 250 kHz resolution using the 12 m
telescope. This line proÐle is extremely narrow (1È2 km s~1) and consists of
two velocity components, one at the systemic velocity of [26 km s~1 and
another near V \ [22 km s~1.
LSR
the two velocity components could be resolved. As Figure 3
shows, the stronger component is extremely narrow and has
V \ [26 km s~1, while the weaker feature is redshifted
LSRabout 2 MHz and is almost as narrow. The second
by
feature could be due to another molecule ; however, to our
knowledge, there is no feasible candidate at this frequency.
Furthermore, the second feature is almost as narrow as the
Ðrst, indicating a similar origin. Such small line widths
suggest non-LTE conditions, which also could result in two
velocity components. Maser emission from the (011c0)
vibrationally excited states of HCN shows multiple velocity
components in the J \ 2 ] 1 line (Lucas & Cernicharo
1989). Therefore, it is quite probable that both features are
attributable to CS (v \ 1) emission.
A crude Ðve point map was made of the J \ 3 ] 2 lines,
using the 12 m. The emission appeared to be a point source
with respect to the 43@@ beam at 146 GHz.
Table 2 presents upper limits obtained for vibrationally
excited CS in searches toward several molecular clouds with
known star formation. The J \ 3 ] 2, 5 ] 4, and 6 ] 5
transitions were searched for toward DR 21 (OH), IRAS
16293[2422, and NGC 1333, and a subset of these lines in
Serpens and Orion A. No emission was detected down to
levels of 3È30 mK (1 p), with typically 20 hr of integration.
Vibrationally excited CS is therefore not as readily excited
in molecular clouds as other molecules such as HCN
(Ziurys & Turner 1986).
FIG. 2.ÈSpectrum of the J \ 5 ] 4 line of CS (v \ 1) near 243 GHz
observed toward IRC ]10216 with 2 MHz resolution using the 12 m
telescope. The J \ 6 ] 5 () \ 1) line of SiC appears as a contaminating
““ shoulder ÏÏ on the CS (v \ 1) proÐle. The solid line is a Ðt of the spectrum
with two Gaussians, one centered at the CS frequency and the other at the
SiC frequency.
4.
ANALYSIS AND DISCUSSION
4.1. Origin of CS (v \ 1) Emission
As shown in Figure 4, the v \ 1 state of CS lies 1830 K in
energy above the v \ 0 state (Yamada & Hirota 1979). The
No. 2, 2000
A NEW CIRCUMSTELLAR MASER ?
FIG. 4.ÈQualitative energy level diagram illustrating the rotational
transitions observed for CS (v \ 1), marked by arrows, and showing the
relevant radio and infrared (v \ 1 ] 0) Einstein A coefficients. The bolder
arrows indicate transitions previously observed by Turner (1987a).
Einstein coefficient for the vibrational decay is A D 16 s~1,
while the A coefficients for the rotational transitions
detected within the v \ 1 level (shown by arrows in the
Ðgure), range from 2 ] 10~5 to 8 ] 10~4 s~1. Therefore,
the decay to ground state is 5 orders of magnitude faster
than any rotational transition observed. Consequently,
extreme conditions are required to sustain sufficient population in the v \ 1 level to enable pure rotational transitions
to occur. Such conditions (high temperatures and densities,
large IR Ðelds) are found in the inner envelope rather than
throughout the entire shell.
The origin of CS (v \ 1) emission can be found by
examining the J \ 6 ] 5 and J \ 7 ] 6 transitions. These
lines both appear to be free of contamination from other
molecules, exhibit symmetric, parabolic proÐles, and have
reasonable signal-to-noise. The J \ 6 ] 5 and J \ 7 ] 6
transitions have line widths at half-power of *V \ 14.3
1@2
and 11.5 km s~1, respectively.
Most line widths of molecular spectra in IRC ]10216
reÑect the full expansion velocity of the gas (V \
exp
14.5 km s~1), and presumably arise from beyond the acceleration and dust-forming zone (Olofsson 1994). In contrast,
the CS (v \ 1) lines are sufficiently narrow to suggest that
they arise from a region where the gas is not fully accelerated. Comparable line widths have been found in features
arising from other vibrationally excited molecules such as
SiS (v \ 1) (Turner 1987b ; Cernicharo et al. 2000), and
885
HCN (011c0) (Lucas & Cernicharo 1989). It has been concluded that these species arise from the inner shell.
The line widths of the CS (v \ 1) spectra can be used to
estimate the source size of this emission. Based on infrared
line proÐles, Keady et al. (1988) and Keady & Ridgway
(1993), modeled the gas expansion velocity for IRC ]
10216 as a function of distance from the star. Their calculations suggest the presence of three steep acceleration
regions, one located at D 5 R , a second at D 10 R , and
*
*
the Ðnal zone at D 14 R , respectively, where R is the
*
*
stellar radius. At 5 R , the gas increases in velocity from
*
D 4 to 11 km s~1, and near 10 R is accelerated from 11 to
*
14 km s~1. The terminal velocity of 14 km s~1 is achieved at
º 14 R . Therefore, using the Keady et al. models, the
*
upper limit to the source radius of CS (v \ 1) emission is
[ 14 R , corresponding to a source size h [ 28 R .
* a distance of 150 pc (Crosas & Menten
s
Assuming
1997* ;
Weigelt et al. 1998), this size corresponds to an angular
extent of 0A. 63, using an average stellar radius of 5 ] 1014
cm (Keady et al. 1988 ; Cernicharo et al. 1996 ; Lucas &
Guelin 1999).
Another estimate of the vibrationally excited CS source
extent can be obtained from Plateau de Bure interferometer
maps of the J \ 5 ] 4 (v \ 1) transition (Lucas & Guelin
1999). These measurements indicate a source size for CS
(v \ 1) emission of roughly 0A. 35 ( D 10 R ), or D 7 ] 1014
cm at a distance of 150 pc. Unfortunately,*these authors did
not take into account the fact that this transition is contaminated by SiC ; however, they likely resolved out the
more extended SiC emission. On the other hand, their value
of D 0A. 35 is remarkably consistent with the source size
independently determined from the CS (v \ 1) line widths
and the Keady et al. expansion proÐle of IRC ] 10216.
Given the uncertainties involved, a reasonable estimate of
the source size for CS (v \ 1) is h [ 0A. 5, based on both
PdBI measurements and the CS linesproÐles.
4.2. Inner Shell Abundances : CS as a Parent Molecule
Based on a source size of D 0A. 5, brightness temperatures
in the v \ 1 level of CS can be estimated. These quantities
are listed in Table 3. As shown in the table, the temperatures
range from T D 50 to 340 K, and steadily increase with J
B
quantum number,
with the exception of the J \ 3 ] 2 transition. The line intensity here is D 526 K, and appears
anomalously high relative to the other temperatures, suggesting some non-LTE e†ects in this transition. (This intensity is based on the component at V \ [26 km s~1). The
brightness temperatures in general,LSR
however, are consistent
with other measurements. Lucas & Guelin (1999), for
example, estimated T D 900 K for the J \ 5 ] 4 line of CS
B
TABLE 3
CS (v \ 1) BRIGHTNESS
TEMPERATURESa
Transition
T
B
(K)
J\2]1......
J\3]2......
J\5]4......
J\6]5......
J\7]6......
54
526b
121
145
344
a Assumes h D 0A. 5.
b If h ¹ 1 sR , T D
*
B
105 K. s
886
HIGHBERGER ET AL.
(v \ 1), which may include some SiC emission. Turner
(1987b) found line intensities in the rotational lines of SiS
(v \ 1) of T Z 600 K, assuming a similar source size.
B
Moreover, the CS rotational temperature at 14 R , based
*
on the v \ 0 ] 1 ro-vibrational transition, is T D 500 K
rot
(Keady & Ridgway 1993). Finally, estimates of gas kinetic
temperatures are T D 1000 K at r \ 6 ] 1014 cm
k
(Cernicharo et al. 1996) and D 500 K at 20 R (Crosas &
*
Menten 1997).
From these brightness temperatures, a column density in
the v \ 1 state can be estimated. If these lines are optically
thin, a column density can be computed from a rotational
diagram. This analysis yields N D 6 ] 1015 cm~2 and
tot 5, results in an unconT D 40 K, but, as shown in Figure
rot
vincing Ðt. However, the parabolic line proÐles of the
uncontaminated J \ 6 ] 5 and J \ 7 ] 6 transitions indicate that the CS (v \ 1) lines may be somewhat optically
thick, as also suggested by Lucas & Guelin (1999). The e†ect
could explain the nonlinearity of the rotational diagram. If
an opacity of 1 is assumed for the J \ 6 ] 5 transition, then
T \ 230 K for this line and N (v \ 1) D 1.9 ] 1016 cm~2.
Ifexq D 1 is estimated for the J \ 7 ] 6 transition, then T \
540 K and N (v \ 1) D 5.9 ] 1016 cm~2. In both cases,exit is
assumed T D T for the partition function in the total
rot
ex
column density
calculation.
Larger opacities (q D 10È100)
can be disregarded because they result in CS (v \ 1) column
densities greater than 1017 cm~2, or CS (v \ 0) column
densities of D 1019 cm~2, comparable or greater than that
of CO (Crosas & Menten 1997).
The v \ 1 column density for CS is therefore on the order
of 2È6 ] 1016 cm~2. From the v \ 1 value, that of the v \ 0
level can be derived if a vibrational temperature is assumed.
Keady & Ridgway (1993) estimate T D 500 K near 14 R .
vib
*
If this temperature is used, then the column
density of CS in
the v \ 0 state is 0.7È2.3 ] 1018 cm~2. This value is consistent with the v \ 0 column density derived from the C33S :
J \ 2 ] 1 observations (see Table 1). This line suggests
N D 2 ] 1018 cm~2, assuming T D 50 K. These results
rot
are summarized in Table 4.
The column density derived for the v \ 0 state of CS
implies a fractional abundance of this species of 2.4È
7.6 ] 10~5, relative to H , at a radius of D 14 R . (The
2 density was taken from*Keady
molecular hydrogen column
& Ridgway (1993), who calculate M0 D 3 ] 10~5 M yr~1).
_
In comparison, the infrared observations of CS suggest
an
FIG. 5.ÈA rotational diagram for the observed CS (v \ 1) transitions
(Table 1) based on the optically thin assumption for all lines and a source
size of 0A. 5. The solid line is the linear least-squares Ðt to the data. The diagram suggests that some transitions are optically thick. Error bars are 3 p.
Vol. 544
TABLE 4
CS COLUMN DENSITIES AND ABUNDANCES
N
N
N
tot
tot
or f
tot
(v \ 1) (cm~2) . . . . . .
(v \ 0) (cm~2) . . . . . .
Value
6 ] 1015a
1.9È5.9 ] 1016b
0.7È2.3 ] 1018c
2.0 ] 1018d
2.4È7.6 ] 10~5
f(CS/H ) . . . . . . . . . . . . . . . . .
2
a From rotational diagram analysis,
assuming q > 1.
b Based on J \ 6 ] 5 and J \ 7 ] 6 transitions, assuming q D 1.
c Optically thick assumption for v \ 1 (see
text).
d From C33S calculations.
abundance of 4 ] 10~6, at roughly the same distance from
the star, while the model of Lucas et al. (1995) predict
f D 10~6, based on CS interferometer measurements. These
abundance estimates are all in reasonable agreement, considering the di†erent methods used to obtain them. Interestingly, at roughly comparable stellar radii, the fractional
abundance of SiS is calculated to be D 7 ] 10~6 (Lucas &
Guelin 1999) and that found for HCN is 3 ] 10~5
(Cernicharo et al. 1996). Hence, large abundances of CS,
SiS, and HCN appear to exist at small stellar radii in IRC
]10216.
Willacy & Cherne† (1998) suggest that CS is a ““ parent ÏÏ
molecule for chemical reactions occurring in the shell of
IRC ]10216, i.e., it is a major starting material for the
formation of other species throughout the envelope. They in
fact derive a ““ TE ÏÏ abundance for CS of 1.3 ] 10~5, close to
the range indicated by our v \ 1 observations. They also
predict a signiÐcant decrease in the abundance of this molecule due to shock-driven reactions in the inner envelope.
The details of this chemistry cannot be tested by our observations ; however, CS does undergo a signiÐcant decrease in
abundance in the outer envelope (Lucas et al. 1995). It may
be converted into longer chains such as C S or C S (e.g.,
2
Cernicharo et al. 1987). In any case the v \ 13 measurements
are additional evidence that CS is indeed a parent molecule
with a very large abundance relatively close to the star.
4.3. Excitation of ““ Thermal ÏÏ CS (v \ 1) Emission
Collisional and/or radiative (infrared) mechanisms may
be responsible for the excitation of the v \ 1 level of CS.
The collisional rate for excitation to the v \ 1 state can be
estimated from the experimentally derived formula of Millikan & White (1963), which depends on the gas kinetic temperature and the reduced mass of colliding particles (in this
case CS and H ). The temperature of IRC ]10216 at the
photosphere is 2 T D 2000 K (Groenewegen 1997), which
puts an upper limit on the gas kinetic temperature. The
Crosas & Menten (1997) model suggests T D 500 K at 20
k K at 6 ] 1014
R , while Cernicharo et al. (1996) use D 1000
*
cm (D14 R ). Considering the 300È1000 K range, the col*
lisional excitation
rate varies from 3 ] 1014 cm3 s~1 to
2 ] 10~13 cm3 s~1, barely 1 order of magnitude change.
For an Einstein A coefficient of 16.1 s~1, the molecular
hydrogen column density required to equate the collisional
rate with the spontaneous de-excitation rate (i.e., A D C)
is n(H ) D 1È5 ] 1014 cm~3. Such densities are several
orders 2of magnitude higher than those found in the shell of
IRC ]10216. For example, n(H ) is estimated to be D 1011
2
No. 2, 2000
A NEW CIRCUMSTELLAR MASER ?
cm~3 at 1 R , decreasing to 108 cm~3 at 10 R , and 107
*
*
cm~3 at 20 R (Keady et al. 1988).
*
Radiative trapping could signiÐcantly reduce the radiative decay rate by the amount A/q , where q is the
IR
IR
optical depth in the v \ 0 ] 1 transition at 12 km.
However, q in this case would have to be on the order of
IR
D 104È105 to sufficiently lower the gas densities. Such opacities would require a ground-state CS column density of N
(CS v \ 0) D 1020 cm~2, i.e., a number that is close to the
H column density. This abundance of CS is impossible. In
2
fact, we estimate q D 4È170 in the v \ 0 ] 1 transition.
IR
Hence, there is not sufficient trapping in the 12 km transition of CS to reduce the density required for collisional
excitation to make it a feasible mechanism. The densities
required are simply too large.
The excitation for the v \ 1 level of CS must consequently be radiative in nature. Because the v \ 0 ] 1 transition of
CS was observed in absorption against the infrared background toward IRC ]10216, at similar stellar radii where
the rotational lines in v \ 1 originate (Keady & Ridgway
1993), there must be sufficient 12 km radiation for this excitation. Moreover, IRC ]10216 is known to be a strong
emitter at D 12 km, and the size of the dust photosphere is
roughly that of the estimated CS (v \ 1) source size (Bieging
& Tafalla 1993). Calculations of the dust thermal temperature (Mangum & Wootten 1993) also indicate the presence of considerable 12 km infrared radiation.
4.4. A CS (v \ 1) : J \ 3 ] 2 Maser ?
The extremely narrow line proÐle and the presence of a
second, unusual velocity component in the J \ 3 ] 2 line of
CS (v \ 1) are suggestive of weak maser emission. The
observed intensity of this transition is a further indication of
non-LTE e†ects. Based on the thermal J \ 5 ] 4 and
J \ 2 ] 1 transitions of CS (v \ 1), the brightness temperature in the J \ 3 ] 2 line should be on the order of 90
K for h D 0A. 5, or an antenna temperature of T * D 9 mK in
s
R
a 43@@ beam.
Both observed components are considerably
stronger, with T * D 54 and 20 mK, respectively. In fact,
R in the J \ 3 ] 2 spectrum, the thermal
given the noise level
emission should not be detectable (see Fig. 3). Moreover,
the line widths of both components are so narrow that these
transitions must originate very close to the star, near 1 R ,
*
or perhaps even in the stellar atmosphere itself, as proposed
for the HCN maser (Lucas & Cernicharo 1989). At 1 R , the
* the
brightness temperature T is on the order of 105 K in
B
[26 km s~1 component, indicating substantial maser gain.
The second component lies higher than the systematic
velocity at a slightly redshifted value of [22 km s~1. There-
887
fore, this component may originate in gas accelerating away
from the line of sight, but there could be complications
because of atmospheric dynamics (e.g., shocks, pulsations).
The (011c0) : J \ 2 ] 1 maser line of HCN also has a second
velocity component, but at a more negative velocity than
the systematic one (Lucas & Cernicharo 1989). The two
HCN features of this transition, however, have line widths
of *V D 10 km s~1 ; therefore, they are not as narrow as
1@2J \ 3 ] 2 lines. The broader proÐles may be partly
the CS
due to nitrogen hyperÐne structure.
Infrared pumping likely excites the CS v \ 1 level.
However, the J \ 3 ] 2 line alone appears to experience
maser ampliÐcation. This e†ect may be attributable to
spontaneous decay rates, as illustrated in Figure 4. Infrared
excitation populates high rotational levels (J [ 30), such as
found by Keady & Ridgway (1993). Some fraction of excited
molecules cascade down the rotational ladder and the Einstein A coefficients decrease accordingly with rotational
transition. At J \ 3, the decay rate (A D 10~5 s~1) may
have decreased sufficiently relative to the A coefficients of
higher transitions that a ““ bottleneck ÏÏ occurs at this rotational level ; that is, the J \ 3 level cannot deplete its population into the J \ 2 level at a rate fast enough to balance
the incoming rate. Therefore, a population inversion occurs.
Such an e†ect might also explain why the J \ 2 ] 1 transition is weak relative to the other lines.
5.
CONCLUSION
New observations of rotational transitions of CS in its
v \ 1 excited vibrational state in IRC ]10216 indicate that
the species arises far closer to the star than past radio measurements suggested. In fact, CS emission arises from a
region before the circumstellar shell has accelerated to its
terminal velocity. The fractional abundance of CS in this
part of the envelope, based on the v \ 1 measurements, is
quite high (CS/H D 10~5), and is additional evidence that
2
this molecule functions
as a ““ parent ÏÏ species for the chemistry in this object. The excitation of vibrationally excited
CS is likely due to infrared radiation at 12 km, as densities
are not high enough to produce the observed population in
the v \ 1 level. While four of the Ðve observed transitions
appear to be thermal, the J \ 3 ] 2 transition may be a
weak maser. Additional observations of CS in its v \ 1 and
v \ 2 states are being conducted towards IRC ]10216 and
other AGB stars to test this hypothesis.
This research is supported by NSF grants AST-98-20576,
AST 95-03274, and AST-96-18523.
REFERENCES
Bieging, J. H., & Tafalla, M. 1993, AJ, 105, 576
Keady, J. J., Hall, D. N. B., & Ridgway, S. T. 1988, ApJ, 326, 832
Boyle, R. J., Keady, J. J., Jennings, D. E., Hirsch, K. L., & Wiedemann, G.
Keady, J. J., & Ridgway, S. T. 1993, ApJ, 406, 199
R. 1994, ApJ, 420, 863
Lucas, R., & Cernicharo, J. 1989, A&A, 218, L20
Cernicharo, J., et al. 1996, A&A, 315, L201
Lucas, R., & Guelin, M. 1999, IAU Symp. 191, Asymptotic Giant Branch
Cernicharo, J., Gottlieb, C. A., Guelin, M., Thaddeus, P., & Vrtilek, J. M.
Stars, ed. T. LeBetre, A. Lebre, & C. Waelkens (San Francisco : ASP),
1989, ApJ, 341, L25
305
Cernicharo, J., Guelin, M., Hein, H., & Kahane, C. 1987, A&A, 181, L9
Lucas, R., Guelin, M., Kahane, C. Audinos, P., & Cernicharo, J. 1995,
Cernicharo, J., Guelin, M., & Kahane, C. 2000, A&AS, 142, 181
Ap&SS, 224, 293
Crosas, M., & Menten, K. M. 1997, ApJ, 483, 913
Lucas, R., Guilloteau, S., & Omont, A. 1988, A&A, 194, 230
Garden, R. P., Geballe, T. R., Gatley, I., & Nadeu, D. 1991, ApJ, 366, 474
Lucas, R., Omont, A., Guilloteau, S., & Rieu, N-Q. 1986, A&A, 154, L12
Goldsmith, P. F., Krotov, R., Snell, R. L., Brown, R. D., & Godfrey, P.
Mangum, J. G., & Wootten, A. 1993, ApJS, 89, 123
1983, ApJ, 274, 184
Millikan, R. C., & White, D. R. 1963, J. Chem. Phys., 39(12), 3209
Groenewegen, M. A. T. 1997, A&A, 317, 503
Olofsson, H. 1994, in Circumstellar Media in the Last Stages of Stellar
Groesbeck, T. D., Phillips, T. G., & Blake, G. A. 1994, ApJS, 94, 147
Evolution, ed. R. E. S. Clegg, I. R. Stevens, & W. P. S. Meikle
Guelin, M., Lucas, R., & Neri, R. 1997, IAU Symp. 170, CO : Twenty Five
(Cambridge : Cambridge Univ. Press), 246
Years of Millimeter-Wave Spectroscopy, ed. W. B. Latter et al.
Schilke, P., GuŽsten, R., Schulz, A., Serabyn, E., & Walmsley, C. M. 1992,
(Dordrecht : Kluwer), 359
A&A, 261, L5
888
HIGHBERGER ET AL.
Turner, B. E. 1987a, A&A, 182, L15
ÈÈÈ. 1987b, A&A, 183, L23
Weigelt, G., Balega, Y., BloŽcker, T., Fleischer, A. J., Osterbart, R., &
Winters, J. M. 1998, A&A, 333, L51
Willacy, K., & Cherchne†, I. 1998, A&A, 330, 676
Yamada, C., & Hirota, E. 1979, J. Molec. Spec., 74, 203
Ziurys, L. M., & Turner, B. E. 1986, ApJ, 300, L19