First Astronomical Detection of the Cumulene Carbon Chain

Molecule H2 C6 in TMC1

W. D. Langer1 , T. Velusamy, T. B. H. Kuiper, and R. Peng

Jet Propulsion Laboratory, California Institute of Technology, MS 169-506, Pasadena, CA
91109
arXiv:astro-ph/9702183v1 21 Feb 1997

and

M. C. McCarthy2 , M. J. Travers2 , A. Kovács, C. A. Gottlieb, and P. Thaddeus2

Division of Engineering and Applied Sciences, Harvard University, 29 Oxford Street,
Cambridge, MA 02138

ABSTRACT

The cumulene carbenes are important components of hydrocarbon chemistry

in low mass star forming cores. Here we report the first astronomical detection
of the long chain cumulene carbene H2 C6 in the interstellar cloud TMC1, from
observations of two of its rotational transitions: JK,K ′ = 71,7 → 61,6 at 18.8 GHz
and 81,8 → 71,7 at 21.5 GHz, using NASA’s Deep Space Network 70 m antenna
at Goldstone, California. In addition we also observed the shorter cumulene
carbene, H2 C4 at the same position. The fractional abundance of H2 C6 relative
to H2 is about 4.7 × 10−11 and H2 C4 is about 1.1 × 10−9 . The abundance of H2 C6
is in fairly good agreement with gas phase chemical models for young molecular
cloud cores, but the abundance of H2 C4 is significantly larger than predicted.

Subject headings: interstellar: molecules — line: identification — molecular

processes — radio lines — stars: formation

1
langer@langer.jpl.nasa.gov
2
Harvard-Smithsonian Center for Astrophysics, 60 Garden Street, Cambridge, MA 02138
 –2–

1. INTRODUCTION

The detection of new cumulene carbenes is important for determining the hydrocarbon
chemistry in low mass star forming cores (cf. Bettens & Herbst 1996). The cumulene
carbene chains are also of interest in astronomy for their possible role as carriers of the
diffuse interstellar bands (McCarthy et al. 1997). Here we report the first astronomical
detection of H2 C6 (hexapentaenylidene) in the dark cloud TMC1. The identification of this
highly reactive isomer of triacetylene (HC6 H) was made possible by the recent laboratory
detection of the rotational spectrum of both H2 C5 and H2 C6 (McCarthy et al. 1997), and
the high sensitivity of NASA’s Deep Space Network antennas at cm wavelengths.
The shorter cumulene carbenes propadienylidene H2 C3 (Cernicharo et al. 1991,
Kawaguchi et al. 1991) and butatrienylidene H2 C4 (Kawaguchi et al.) were also observed
at the same position as H2 C6 , allowing comparison with abundances predicted by chemical
models. Both H2 C4 and H2 C6 have linear carbon backbones and C2v symmetry and,
because of the off-axis equivalent H-atoms, they both have ortho (K = ±1, odd) and para
(K = 0, even) rotational ladders between which both radiative and collisional transitions
are forbidden. In cold molecular clouds we expect 3/4 of the population to be in the lowest
lying ortho (K = ±1) ladders, and emission from these to be slightly stronger than from
the para (K = 0) ladder.
In addition to H2 C6 , several other complex molecules were observed in TMC1, for the
purpose of clarifying hydrocarbon chemistry in these star forming cores. We find that the
carbon chain molecules vary in abundance among the three principal velocity components.
The distribution of the cumulene carbenes is similar to those of other complex carbon
compounds, such as the cyanopolyynes HC2n+1 N and the carbon chain radicals Cn H, but
differ from the cyclopropenylidene ring, c-C3 H2 . The observed ratios of abundances in
TMC1 are compared with gas phase chemical models.

2. OBSERVATIONS

The observations were made with NASA’s Deep Space Network (DSN) 70 meter
antenna at Goldstone, CA during October and November of 1996. In addition to H2 C6 ,
five other molecules (HC9 N, C5 H, H2 C3 , H2 C4 , and c-C3 H2 ) were observed as part of an
ongoing effort to study the carbon chemistry in TMC1. The relevant transitions and line
frequencies are listed in Table 1.
The observations were made with a broad band (17.5 - 26 GHz) cooled HEMT
receiver with typical system temperatures of 50 - 65 K and the 2 million channel Wide
 –3–

Band Spectrometer Analyzer (WBSA). The spectra were smoothed to resolutions in the
range of 0.03 to 0.15 km s−1 with a total spectral range of 20 to 40 MHz (cf. Langer et
al. 1995). All observations, except that of the 18.4 GHz transition of H2 C6 , were made
using position switching. In this mode the spectra were obtained by taking the difference
between successive ON and OFF source integrations. To suppress baseline irregularities of
instrumental and atmospheric origin, the right ascension of the OFF source position was
also chosen to be 6 minutes in time to the west or east of the ON source position such
that each ON-OFF pair had the same mean elevation. The presence of several stronger
lines in the passband of the 21.5 GHz line of H2 C6 and the 20.8 GHz line of H2 C3 , allowed
us to check our calibration during each observing session. Following the detection of the
21.5 GHz line of H2 C6 , a frequency switching option was implemented for the DSN-WBSA
system. We used this mode for the observations of the 18.8 GHz transition of H2 C6 using
an LO offset of 0.5 MHz between the signal and reference frequencies.
The detections reported here were made toward TMC1 at a position [α(1950) =
4 38 41s and δ(1950) = 25o 35′40′′ ] corresponding to the strongest HC7 N and HC9 N
h m

emission observed at vLSR = 5.8 km s−1 in our DSN 70 m high spectral resolution maps
(Velusamy et al. 1997). The ON source integration times for the H2 C6 lines were 23 and 18
hours for the 21.5 and 18.8 GHz transitions, respectively, and yielded a corresponding rms
of 2.5 mK and 3.5 mK at spectral resolutions of 0.14 and 0.12 km s−1 .

3. RESULTS

In Figure 1a we show the spectra for the two transitions of H2 C6 , and those of C5 H
and HC9 N. The H2 C6 21.5 GHz transition, HC9 N, and C5 H were observed simultaneously
in the same passband. The spectrum of the 20,2 → 10,1 transition of H2 C4 at 17.9 GHz in
TMC1 is shown in Figure 1b, along with the lines of H2 C3 , C6 H, and c-C3 H2 . The H2 C4
and H2 C6 spectra peak at ∼ 6 km s−1 with antenna temperatures TA = 180 mK and 9
mK, respectively. Previous high spectral resolution studies of carbon chain molecules in
TMC1, such as CCS and HC7 N (Langer et al. 1995), revealed that there are three velocity
components at 5.7, 5.9, and 6.1 km s−1 with very narrow linewidths ∼ 0.15 to 0.20 km s−1 .
Our detection of H2 C6 in TMC1 supports the suggestion of a tentative assignment of a
weak unblended and a blended 3 mm line in IRC+10216 (Guélin et al. 1997).
In Figure 1 it appears that the strongest H2 C4 and H2 C6 emission is at 5.9 km s−1 .
The slight asymmetry in the line profiles is most likely due to a second weaker (partially
blended) component at 6.1 km s−1 , exactly what is seen in the velocity structure of C5 H
and probably C6 H. There is no evidence in H2 C4 and H2 C6 for a component at 5.7 km s−1 .
 –4–

In contrast, c-C3 H2 and its linear isomer H2 C3 show strong emission at 5.7 and 5.9 km s−1 ,
and only slightly less emission at 6.1 km s−1 . The presence of strong c-C3 H2 and H2 C3
emission at 5.7 km s−1 and the apparent absence of H2 C6 or H2 C4 emission at that velocity
indicates that there is a significant chemical inhomogeneity in the hydrocarbon chemistry
in TMC1. Furthermore, the c-C3 H2 emission is much stronger than that of H2 C3 (TA = 2.0
K versus 0.12 K).

4. DISCUSSION

To estimate the abundances of H2 C6 and H2 C4 in TMC1 we adopted a simple model in

which we treat each rotational ladder as a separate linear rotor. The excitation of the ortho
and para states can be treated separately, because the radiative and collisional transitions
between them are negligible. For purposes of estimating abundances, the excitation of
the ortho ladder can be treated as two equivalent linear rotors. The energy levels of the
K = ±1 ladders are nearly equal and can be estimated within 1% by treating them as a
linear rotor with B0 = (B + C)/2 (see McCarthy et al.). Collisional rate coefficients for
these species have not been calculated; we approximate their J-dependent values by using
those derived for the linear chain molecule HC3 N (Green and Chapman 1978). We assume
that the rotational population of H2 C6 has a normal ortho:para ratio of 3:1. Furthermore,
reactive collisions with protonating ions can produce H2 C6 H+ which, upon decomposition
to H2 C6 + H, will preserve the nuclear spins (Herbst, 1977).
H2 C6 is highly polar, with a calculated dipole moment of 6.2 D (Maluendes & McLean
1992). The J = 8 ortho levels lie ∼ 5 K above the ortho ground state (J = 1) and these
levels have an Einstein A coefficient of about 2 × 10−6 s−1 . This yields a critical density
n(H2 ) ∼ 5 × 103 cm−3 , implying that the lower rotational levels are easily thermalized at
the density in the cores of TMC1. Figure 2 shows an excitation calculation for the K = −1
ortho ladder of H2 C6 and the K = 0 para ladder of H2 C4 (using a dipole moment of 4.5 D)
for a range of fractional abundance gradients (in km s−1 pc−1 ). For convenience, calculations
were done with a large velocity gradient (LVG) code which is essentially equivalent to
an LTE calculation for optically thin lines, as appropriate for the weak emission seen in
H2 C4 and H2 C6 . We assumed typical conditions found in the TMC1 clumps: Tkin = 10 K
and n(H2 ) = 104 cm−3 (Langer et al.). The velocity gradient is estimated from the line
width, ∼ 0.3 km s−1 , divided by the size of the emission region (∼ 0.06 pc) estimated from
our HC7 N map. The total fractional abundances of H2 C4 and H2 C6 , after accounting for
molecules in the other ladders, a velocity gradient of 5 km s−1 pc−1 , and a beam efficiency of
0.7, are about 1.1 × 10−9 and 4.7 × 10−11 , respectively. The column density for H2 C4 derived
 –5–

by Kawaguchi et al. (1991) from several transitions at a nearby position corresponds to a

fractional abundance of 8 × 10−10 (Bettens et al. 1995), which is in very good agreement
with our value. We estimate that H2 C4 is ∼ 25 more abundant than H2 C6 .
Recent gas phase chemical models have included reaction networks for calculating
abundances of long carbon chains (Millar et al. 1997, Bettens, Lee & Herbst 1995) and very
large hydrocarbons (Thaddeus 1994, and Bettens & Herbst 1996). Furthermore, laboratory
data and theoretical calculations have begun to clarify the mechanisms for forming the
simplest ring structure cyclopropynylidyne, c-C3 H, via neutral carbon insertion into
acetylene, C + C2 H2 (Kaiser et al. 1996). To compare qualitatively our observations with
chemical models we use the results of the gas phase model of Millar et al. (1997), which is
based on one of the standard chemical data bases used for modeling cloud chemistry. It
includes hydrocarbon production up to 8 or 9 carbons, but does not include the carbon
insertion reactions.
In Figure 3 we plot these model abundances for the carbon chains C2n H, C2n+1 H,
H2 C2n , H2 C2n+1 , and HC2n+1 N, as well as a point for c-C3 H2 , for the physical conditions
relevant to this core of TMC1. Figure 3 uses the early time results of Millar et al. as
only this evolutionary stage has enough neutral carbon to produce a successful complex
carbon chemistry with large hydrocarbon abundances. The late times, or steady state,
solutions have too little carbon and complex carbon species to explain the observations.
This result is not surprising as other tracers, such as CCS, confirm the ages of these dense
core components to be ≤ a few ×105 years (Kuiper et al. 1996, Velusamy et al. 1997).
Note that the Cn H radicals are predicted to be more abundant than the corresponding
H2 Cn cumulene carbenes, as observed. However, the Millar et al. model also predicts that
H2 C6 is more abundant than H2 C4 which is exactly the opposite from what is observed.
The measured H2 C6 abundance agrees within a factor of 2 with their model calculations,
while the predicted H2 C4 fractional abundance is too low by a factor ∼ 25. On the other
hand, the observed H2 C4 /H2 C6 ratio is consistent with the slope of the abundance versus
chain length in the Cn H radicals. However, some caution is in order regarding absolute
abundances, because this model does not distinguish between different isomers, such as
triacetylene (HC6 H) and H2 C6 .
Bettens et al. explored the effects of various assumptions about the neutral-neutral
reactions on the production of complex carbon molecules in interstellar clouds. Bettens
& Herbst (1996) recommend using two of their chemical models which best explain
the polyatomic species: the new standard model (NSM) and model 4 (M4). NSM is a
modified version of the standard ion-molecule chemical scheme; in M4, rapid neutral-neutral
reactions play a critical role. A key feature of M4 is that the reactions O + Cn →
 –6–

Cn−1 + CO are assumed negligible for n > 2. Model M4 of Bettens et al. provides the
best agreement with our H2 C4 observations, but the calculated abundance is still too
high by a factor of five. The models appear to be sensitive to assumptions about the
neutral-neutral reaction rate coefficients at low temperature. H2 C6 and H2 C4 are important
diagnostics to discriminate among models of hydrocarbon chemistry in dense cores, however,
more laboratory measurements and astronomical observations are needed to resolve the
hydrocarbon chemistry in dense cores.
We thank the staff of the DSN at Goldstone for their assistance in making these
observations and Eric Herbst for useful comments. The research of the JPL group was
conducted at the Jet Propulsion Laboratory, California Institute of Technology, under
contract with the National Aeronautics and Space Administration. WDL would also like to
thank the Smithsonian Astrophysical Observatory and the Center for Astrophysics for their
hospitality during two visits while this work was in progress.
 –7–

TABLE 1

MOLECULAR TRANSITIONS OBSERVED IN TMC1

Species Transition Frequency Reference

MHz
H2 C3 10,1 → 00,0 20792.640 Lovas et al.
c-C3 H2 11,0 → 10,1 18343.143 Pickett et al.
H2 C4 20,2 → 10,1 17863.810 Travers et al.
C5 H J = 9/2 → 7/2, F = 5 − 4b 21484.710 Lovas
C5 H J = 7/2 → 7/2, F = 4 − 3b 21485.262 Lovas
C6 H J = 15/2 → 13/2, F = 8 − 7f 20792.872 Lovas
H2 C6 71,7 → 61,6 18802.235 McCarthy et al.
H2 C6 81,8 → 71,7 21488.255 McCarthy et al.
HC9 N 37 → 36 21498.181 Pickett et al.
 –8–

REFERENCES

Bettens, R. P. A. & Herbst, E. 1996, ApJ, 468, 686

Bettens, R. P. A., Lee, H.-H., & Herbst, E. 1995, ApJ, 443, 664

Cernicharo, J., Gottlieb, C. A., Guélin, M., Killian, T. C., Paubert, G., Thaddeus, P. &
Vrtilek, J. M. 1991, ApJ, 368, L39

Green, S. & Chapman, S, 1978, ApJS, 37, 169

Guélin, M., Cernicharo, J., Travers, M. J., McCarthy, M. C., Gottlieb, C. A., Thaddeus, P.,
Ohishi, M., Saito, & Yamamoto, S. 1997, A&A, 317, L1

Herbst, E. 1977, private communication

Kaiser, R. I., Ochsenfeld, C., Head-Gordon, M., Lee, Y. T., & Suits, A. G. 1996, Science,
274, 1508

Killian, T. C., Vrtilek, J. M., Gottlieb, C. A., Gottlieb, E. W., & Thaddeus, P. 1990, ApJ,
365, L89

Kuiper, T. B. H., Langer, W. D., & Velusamy, T., 1996, ApJ, 468, 773

Kawaguchi, K. et al. 1991, PASJ, 43, 607

Langer, W. D., Velusamy, T., Kuiper, T. B. H., Levin, S., Olsen, E. & Migenes, V. 1995,
ApJ, 453, 293

Lovas, F. J. 1992, J. Chem. Phys. Ref. data, 21, 181

Lovas, F. J., Suenram, R. D., Ogata, T., & Yamamoto, S. 1992, ApJ, 399, 325

Maluendes, S. A. & McLean, A. D. 1992, Chem. Phys. Lett., 200, 51

Millar, T. J., Farquhar, P. R. A., & Willacy, K. 1997, A&AS, 121, 139

Pickett, H., Poynter, R., & Cohen, E. 1991, Submillimeter, Millimeter, and Microwave Line
Catalog (3d ed.; JPL Tech. Rep. 80-23)

Thaddeus, P. 1994, in Molecules and Grains in Space, ed. I. Nenner and L. Trojanowski,
American Institute of Physics Conf. Proc. 312, (New York: Am. Inst. Phys.), 711

Travers, M. J., Chen, W, Novick, S. E., Vrtilek, J. M., Gottlieb, C. A., & Thaddeus, P.
1996, J. Mol. Spectrosc., 180, 75
 –9–

Velusamy, T., Kuiper, T. B. H., Langer, W. D., & Peng, R. 1997, in preparation

Vrtilek, J. M., Gottlieb, C. A., Gottlieb, E. W., Killian, T. C., & Thaddeus, P. 1990, ApJ,
364, L53

This preprint was prepared with the AAS LATEX macros v4.0.
 – 10 –

Fig. 1.— (a) DSN 70 m detections of H2 C6 at 21.5 and 18.8 GHz in TMC1. The spectra for
H2 C6 at 21.5 GHz, HC9 N, and C5 H, were taken within the same passband, while the 18.8
GHz line of H2 C6 was observed separately. The source position is α(1950) = 4h 38m 41s and
δ(1950) = 25o35′ 40′′ , with vLSR = 5.8 km s−1 . (b) The spectrum for H2 C4 towards TMC1 is
shown along with those of C6 H (2 hyperfine components), H2 C3 , and c-C3 H2 . The vertical
line marks 5.9 km s−1 .

Fig. 2.— The antenna temperatures for the ortho levels of H2 C6 (solid lines) and para
levels of H2 C4 (dashed lines) calculated using a LVG excitation model and approximating
the energy levels with a simple linear rotor (see text) as a function of J. The parameters are
Tkin = 10 K, n(H2 ) = 104 cm−3 , and ∆V/∆L = 1 km s−1 pc−1 . Curves for three fractional
abundances (in units of km s−1 pc−1 ) are shown for each molecule. The observed TA for
H2 C4 (filled triangle) and H2 C6 (filled circles) are marked in the figures along with their
1σ error bars. Total fractional abundances need to include a factor for the ortho and para
fractions, the velocity gradient, and antenna efficiency (see text).

Fig. 3.— The fractional abundances of five carbon chains, C2n H, C2n+1 H, H2 C2n+1 , H2 C2n ,
and HC2n+1 N as a function of number of carbons Cn . These are taken from the early time
solutions for the gas phase chemical model of Millar et al. for a cold cloud core with Tkin =
10K and n(H2 )=104 cm−3 . The observed fractional abundances of H2 C4 and H2 C6 , assuming
∆V/∆L = 5 km s−1 pc−1 , are marked by filled triangles.