1

Detection of interstellar CH3

H. Feuchtgruber

Max-Planck-Institut für extraterrestrische Physik, Postfach 1603, D–85740 Garching,

Germany
arXiv:astro-ph/0005273v1 12 May 2000

F. P. Helmich

SRON, P.O. Box 800, 9700 AV Groningen, The Netherlands

E. F. van Dishoeck
Leiden Observatory, P.O. Box 9513, NL-2300 RA Leiden, The Netherlands

C. M. Wright
School of Physics, University College, Australian Defence Force Academy, UNSW,
Canberra ACT 2600, Australia

Received ; accepted

Submitted to ApJ Lett. [March 2000]

1
Based on observations made with ISO, a project of ESA with participation of ISAS and
NASA, and the SWS, a joint project of SRON and MPE with contributions from KU Leuven,
Steward Observatory and Phillips Laboratory.
 –2–

ABSTRACT

Observations with the Short Wavelength Spectrometer (SWS) onboard the

Infrared Space Observatory (ISO) have led to the first detection of the methyl
radical CH3 in the interstellar medium. The ν2 Q−branch at 16.5 µm and the
R(0) line at 16.0 µm have been unambiguously detected toward the Galactic
center SgrA∗ . The analysis of the measured bands gives a column density of
(8.0±2.4)×1014 cm−2 and an excitation temperature of (17 ± 2) K. Gaseous
CO at a similarly low excitation temperature and C2 H2 are detected for the
same line of sight. Using constraints on the H2 column density obtained from
C18 O and visual extinction, the inferred CH3 abundance is (1.3 +2.2
−0.7
) × 10−8 .
The chemically related CH4 molecule is not detected, but the pure rotational
lines of CH are seen with the Long Wavelength Spectrometer (LWS). The
absolute abundances and the CH3 /CH4 and CH3 /CH ratios are inconsistent
with published pure gas-phase models of dense clouds. The data require a mix
of diffuse and translucent clouds with different densities and extinctions, and/or
the development of translucent models in which gas-grain chemistry, freeze-out
and reactions of H with polycyclic aromatic hydrocarbons and solid aliphatic
material are included.

Subject headings: Line: identification — ISM: abundances — ISM: lines and

bands — ISM: molecules — Galaxy: center — Infrared: ISM: lines and bands
 –3–

1. Introduction

The methyl radical CH3 is an important intermediate product in the basic ion-molecule
gas-phase chemistry networks in the interstellar medium driven by cosmic-ray ionization
(Herbst & Klemperer 1973). Together with CH and CH2 , it is produced by a series of
reactions starting with C + H+ + +
3 → CH + H or the radiative association of C + H2 →

CH+
2 + hν, followed by hydrogen abstraction reactions and dissociative recombination.

Alternatively, it can be produced by photodissociation of methane (CH4 ). Subsequent

reactions of C+ with CH3 form one of the most important steps in the formation of more
complex hydrocarbons.

The ion H+
3 which initiates the chemistry in the cold gas has been detected only recently

toward the Galactic center by Geballe et al. (1999) at a surprisingly high abundance. This
unique line of sight turns out to be an extremely valuable environment to study abundances
in the cold low density interstellar medium, since even minor species like CH3 may be
detectable due to its long path.

Although CH3 is a simple species, it is difficult to obtain accurate laboratory

measurements of its molecular parameters since, as a radical, it recombines very fast with
other particles in a gas. Herzberg (1961) and Herzberg & Shoosmith (1956) were the first
to determine that the molecule is planar, but definite proof came only from measurements
of the out-of-plane bending mode ν2 at 16 µm by Yamada et al. (1981).

Observations of the 16–16.5 µm wavelength range are strongly hampered from the
ground due to the Earth’s atmosphere. The first detections of CH3 in space have become
possible only using the Infrared Space Observatory (ISO) (Kessler et al. 1996). Bézard et
al. (1998, 1999) have recently detected CH3 in the atmospheres of Saturn and Neptune
respectively, but no previous searches for the molecule in interstellar space have been
reported.
 –4–

2. CH3 spectroscopy

Since the CH3 radical is planar and symmetric, it does not have electric dipole allowed
rotational lines which could be detected in the (sub-)millimeter wavelength range. The
planar nature also implies that the symmetric stretch ν1 is infrared inactive and the
asymmetric stretch ν3 at 3.16 µm relatively weak. Indeed, the transition dipole moment of
the ν3 band is found to be a factor of three weaker than the out-of-plane bending mode ν2
(Triggs et al. 1992, Amano et al. 1992).

To calculate the ν2 spectrum, the term energies were taken from Yamada et al. (1981).
The nuclear spin of the H-atoms can couple either to a quartet or doublet state, with
nuclear spin statistical weights of 4 and 2, respectively. Because CH3 follows Fermi-Dirac
statistics, the K = 3, 6, 9, . . . levels are quartet states, and the other K-values doublet states.
The strongest Q-branch has N = K and is located at 16.5 µm; the strongest other feature
is the Q R(0) line at 16.0 µm thanks to its favorable Hönl-London factor. The band strength
of (2.5 ± 0.8) × 10−17 cm−1 (molecule cm−2 )−1 was taken from Wormhoudt & McCurdy
(1989). The calculation of the spectrum was performed as described in Helmich (1996).

The shape of the spectrum is very sensitive to the excitation temperature (see Figure
7.14 of Helmich 1996). Besides the strong Q Q-branch and Q R(0) lines, many more features
become visible at excitation temperatures above 25 K, most notable the satellite R Q-branch
at 16.53 µm and the P (2) line at 17.60 µm.

3. Observations and Data Reduction

Observations were carried out in the SWS grating mode AOT06 (de Graauw et al. 1996)
at a spectral resolving power R = λ/∆λ ≈ 1500 − 2200. The spectral range covering the
CH3 Q-branch and the P (2) line has been measured on 1997 February 21 15:24:27-19:03:01
 –5–

UT, whereas that covering the R(0) and R(1) lines has been obtained on the same day at
19:03:45-20:55:41 UT. The SWS aperture size was 14′′ × 27′′ and has been centered on the
position of Sgr A∗ RA 17h 45m 40s .0, Dec −29◦ 00′ 28′′ .6 (J2000 coordinates), with the long
side of the slit oriented within 1 degree of the north-south direction. Due to the rather large
aperture size, the Galactic center sources IRS 1, 2, 3 and 7 also fall inside the beam (see,
e.g., Geballe et al. 1989), whereas IRS 5 and 6 are positioned just outside the slit.

Data were processed within the SWS interactive analysis system, based on the standard
ISO pipeline OLP V8.7 products. The data reduction adhered to the recommendations of
Salama et al. (1997). Raw data were rebinned to R = 5000, a value significantly larger
than the actual spectral resolving power of the SWS to avoid losing spectral detail when
convolving the observed data samples by the bin.

The absolute calibration of the SWS data has about ±20% uncertainty on average
longwards of ∼15 µm (Salama et al. 1997). However, since our analysis is entirely based
on spectra where the continuum is divided out, the actual uncertainty in the results is
determined by the noise in the data rather than the actual calibration uncertainty. The
main limitation of our analysis originates from the ±30% uncertainty in the CH3 ν2 band
strength (Wormhoudt & McCurdy 1989, Yamada & Hirota 1983).

4. CH3 results

As shown in Fig. 1, the Q-branch at 16.5 µm and the R(0) line at 16.0 µm are clearly
detected. This represents the first unambiguous detection of CH3 in the interstellar medium.
The upper limit on the P (2) line at 17.60 µm provides an important constraint on the
temperature. Due to a blend with the [Ne III] 15.555 µm atomic fine structure line at the
SWS grating resolution, no information from the R(1) line at 15.54 µm could be obtained.
 –6–

Both the Q-branch and the R(0) lines are shifted with respect to their expected LSR
wavelengths by about −20 km s−1 . Although such a shift is close to the SWS wavelength
calibration accuracy (Valentijn et al. 1996), a VLSR = −30 km s−1 component of cold
molecular gas has been reported previously by several authors from observations at radio
and millimeter wavelengths with similar beam sizes (Serabyn et al. 1986 and Sutton et al.
1990, CO; Pauls et al. 1996, H2 CO; Serabyn & Güsten 1986, NH3 ; Marr et al. 1992, HCO+ ;
Güsten et al. 1987, HCN; Bolton et al. 1964, OH). In all cases several velocity components
at ∼ −50, −30 and 0 km s−1 have been observed at much higher spectral resolutions. The
relative strengths of these three components vary between the observed species with the 0
km s−1 component often the largest. At the SWS spectral resolution of ∼ 150 km s−1 it is
not possible to distinguish between these different velocity components, but our observed
shift is consistent with a mix of them. The location of the absorbing gas can therefore not
be attributed to one particular feature, but is possibly spread along the line of sight toward
Sgr A∗ among spiral arms and molecular clouds.

Fits of synthetic spectra to the data, as described in §2 and matching the SWS
resolution, were performed for different excitation temperatures (10 to 50 K), different
Doppler parameters (b between 1.5 and 30 km s−1 ) and column densities. The best fit to the
individual absorption depths (Fig. 1 a,b) as well as their ratio (Fig. 1 c) is given in Table 1.
The absorption depths are almost independent of Doppler parameter, and are mainly a
function of column density. The ratio of the depths is a strong function of the excitation
temperature. The inferred low excitation temperature of (17 ± 2) K from the 16.0/16.5
µm ratio of (0.8±0.15) is consistent with the non-detection of the P(2) line. Because CH3
has no dipole moment, the populations of the lowest rotational levels are controlled by
collisions, so that the excitation temperature is close to the kinetic temperature.

The H2 column density along the line of sight has been constrained by several sets
 –7–

of observations. First, the measured extinction of 31 mag (Rieke et al. 1989) implies
N(H2 )≈ 2 × 1022 cm−2 using NH /AV = 1.9 × 1021 cm−2 mag−1 and assuming that at least
half of the hydrogen is in molecular form. Second, the detection of at least one optically
thin C18 O line (R(0) at 4.7716 µm) in our ISO-SWS observation together with the measured
CO excitation temperature of 8-13 K implies N(C18 O)=(2 ± 0.5) × 1016 cm−2 . This is in
good agreement with the analysis by Moneti & Cernicharo (2000) on the same data. Using
16
O/18 O= 300 (Wilson & Rood 1994) and CO/H2 =10−4 implies N(H2 )≈ 6 × 1022 cm−2 .
We adopt N(H2 )=(6 ± 3) × 1022 cm−2 , leading to a CH3 abundance with respect to H2 of
x(CH3 )=(1.3 +2.2
−0.7
) × 10−8 . Note that all abundances would be increased by a factor of 3 if
the lower H2 column density derived from the extinction is used.

5. Related species: CH4 , C2 H2 and CH

The availability of the full SWS scan allows searches for other chemically related
molecules. Specifically, the ν2 /ν4 dyad of CH4 occurs around 7.7 µm and has been observed
with the ISO-SWS toward massive protostars by Boogert et al. (1998). Toward the Galactic
center, however, gas-phase CH4 is not detected. Adopting the same excitation temperature
as found for CH3 , an upper limit for its abundance of ≤1×1015 cm−2 is found. Solid
CH4 is clearly detected by Chiar et al. (2000) toward Sgr A∗ with a column density of
(3.0 ± 0.7)×1016 cm−2 . Thus, most of the CH4 is in solid form, consistent with the low
temperature.

Detection of a blend of the pure rotational lines of CH at 149.09 and 149.39 µm towards
the Galactic center has been reported by White et al. (1999). We have re-analyzed the LWS
observations carried out on 1998 February 20 10:11:34-11:06:44 in the LWS01 grating mode
(Clegg et al. 1996) at a resolution of ∼ 1500 km s−1 (Fig. 2 a). The LWS data reduction
has been based on OLP V8.7 products and has been carried out within ISAP (Sturm et al.
 –8–

1997). Outliers have been removed by iterative sigma clipping and the different scans have
been flatfielded to their mean value by applying a second order polynomial offset to each
individual scan. The fringing in the LWS data, present in all extended source observations
has been removed by the dedicated module within ISAP. The inferred equivalent width for
the unresolved doublet is (85 ± 5) km s−1 . Using the formulae from Stacey et al. (1987) and
assuming Tex =17 K, this leads to N(CH)≈ (1.1 ± 0.1) × 1015 cm−2 .

Finally, the ν5 band of gas-phase C2 H2 at 13.7 µm is clearly detected (Fig. 2 b).

Following the analysis of Lahuis & van Dishoeck (2000), we find N(C2 H2 )=(5.5 ± 0.8) × 1014
cm−2 assuming Tex = 17 K. Table 1 summarizes the results obtained from ISO observations.
Note that the relative abundances of the molecules have much smaller error bars than the
absolute values since the uncertainty in the H2 column density cancels out.

6. Chemistry

The absolute and relative abundances of the observed molecules have been compared
with a wide variety of models, including time- and depth-dependent models. None of the
published pure gas-phase dense cloud models can reproduce the observations of all species
(e.g., Millar et al. 1997, Lee et al. 1996, see Table 2). In general, the model CH3 abundances
are too low and the CH4 abundances too high. Also, the model abundance of C2 H2 is
significantly smaller than that of CH4 , in contrast with observations. The only models which
come close to matching the absolute and relative abundances of CH3 and CH3 /CH and
CH3 /CH4 are low-density translucent cloud models with n(H2 )≈ 103 cm−3 and AV ≈few
mag, so that photodissociation of CH4 to CH3 and CH2 is still efficient. Table 2 lists recent
model calculations by Terzieva & Herbst (priv. communication) and results based on the
models by van Dishoeck & Black (1986) and Jansen et al. (1995) using updated branching
ratios for the dissociative recombination of the hydrocarbon ions (Andersen et al. 2000).
 –9–

Low metal abundances are favored, to prevent destruction of the hydrocarbons by oxygen
and by sulfur atoms and ions. Note, however, that even though the abundances at AV ≈ 3
may match the data within a factor of a few, the CH3 column density in such models is
only 1 × 1013 cm−2 , nearly two orders of magnitude below observations. At the same time,
the CH column density of 1.4 × 1014 cm−2 is a factor of 10 below observations. Because of
the strong depth dependence of the CH, CH3 and CH4 abundances, it is not possible to
reproduce the column density ratios with these same models. The large observed H+
3 column

density suggests that there are several clouds along the line of sight. Some combination of
low-density diffuse clouds to produce the CH and denser clouds to account for the solid
CH4 may explain those data, but the mix would have to be tailored very specifically to
simultaneously approach the large column densities of CH3 and C2 H2 .

An alternative suggestion is to invoke turbulent chemistry, in which a high CH+

abundance leads to enhancements of other hydrocarbons by 1-2 orders of magnitude (e.g.
Hogerheijde et al. 1995, Joulain et al. 1998). However, the relative ratios of CH and CH3
are unlikely to change in such models.

Given the detection of solid CH4 and the low inferred temperatures, it is plausible
that gas-solid interactions and grain-surface chemistry also play a role in producing the
hydrocarbons. In this respect, the situation for CH3 may be similar to that for NH in
diffuse clouds (Mann & Williams 1984, van Dishoeck 1998). Conversion of atomic carbon to
small hydrides on grain surfaces may be significant, but no model results exist yet for these
conditions. Such models should also explain the C2 H2 abundances and lack of complete
C2 H2 freeze-out. Alternatively, reactions of atomic H with PAHs and solid aliphatic
hydrocarbon material, known to be present toward Sgr A∗ from the 3.4 µm absorption
feature, may lead to CH3 . Shock chemistry is not likely to be important for this line of
sight because of the low temperatures.
 – 10 –

Future high spectral resolution observations of CH3 toward Sgr A∗ to constrain the
velocity structure, as well as observations of CH3 and other molecules in different types of
diffuse clouds are needed to constrain the basic hydrocarbon chemistry.

The authors are grateful to the SWS instrument teams, to E. Herbst, R. Terzieva and
D.J. Jansen for updated model results on CH3 and to W. Duley for inspiring discussions.
This work was supported by DARA grants no 50 QI9402 3 and 50 QI 8610 8 and by
NWO grant 614.41.003. CMW acknowledges receipt of an ARC Australian Postdoctoral
Fellowship.
 – 11 –

REFERENCES

Amano, T., Bernath, P.F., Yamada, C., Endo, Y., Hirota, E. 1982, J. Chem. Phys. 77, 5284

Andersen, L. et al., 2000, in IAU Symposium 97, Astrochemistry: From molecular clouds
to Planetary Systems, eds. Y. C. Minh & E.F. van Dishoeck (ASP), in press

Bézard, B., Feuchtgruber, H., Moses, J.L., Encrenaz, T., 1998, A&A 334, L41

Bézard, B., Romani, P.N., Feuchtgruber, H., Encrenaz, T., 1999, ApJ 515, 868

Bolton, J.G., Gardner, F.F., McGee, R.X., Robinson, B.J., 1964, Nature 204, 30

Chiar, J.E., Tielens, A.G.G.M., Whittet, D.C.B., Schutte, W.A., Boogert, A.C.A., Lutz,
D., van Dishoeck, E.F., Bernstein, M.P., ApJ, in press

Clegg, P.E., Ade, P.A.R., Armand, C., et al., 1996, A&A 315, L38

Geballe, T.R., Baas, F., Wade, R., 1989, A&A 208, 255

Geballe, T.R., McCall, B.J., Hinkle, K.H., Oka, T., 1999, ApJ 510, 251

de Graauw Th., Haser, L.N., Beintema, D.A., et al., 1996, A&A 315, L49

Güsten, R., Genzel, R., Wright, M.C.H., Jaffe, D.T., Stutzki, J., Harris, A.I., 1987, ApJ
318, 124

Helmich, F. P., 1996, PhD Thesis, University of Leiden

<http://www.strw.leidenuniv.nl/∼fph/thesis.ps.gz>

Herbst, e., Klemperer, W., 1973, ApJ, 185, 505

Herzberg, G., 1961, Proc. Royal Soc. London A 262, 291

Herzberg, G., Shoosmith, J., 1956, Can. J. Phys. 34, 523

 – 12 –

Hogerheijde, M.R., de Geus, E.J., Spaans, M., van Langevelde, H.J., van Dishoeck, E.F.,
1995, ApJ, 441, L93

Jansen, D.J., van Dishoeck, E.F., Black, J.H., Spaans, M., Sosin, C. 1995, A&A, 302, 223

Joulain, K., Falgarone, E., Pineau des Forets, G., Flower, D., 1998, A&A, 340, 241

Kessler, M.F., Steinz, J.A., Anderegg, M.E., et al., 1996, A&A 315, L27

Lee, H.-H., Bettens, R.P.A., Herbst, E. 1996, A&AS, 119, 111

Mann, A.P.C., Williams, D.A. 1984, MNRAS, 209, 33

Marr, J.M., Rudolph, A.L., Pauls, T.A., Wright, M.C.H., Backer, D.C., 1992, ApJ, 400, L29

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

Moneti, A., Cernicharo, J., Proc. of Workshop “ISO beyond the Peaks”, ESA SP-456, in
press

Pauls, T., Johnston, K.J., Wilson, T.L., 1996, ApJ 461, 223

Rieke,G.H., Rieke, M.J., Paul, A.E., 1989, ApJ, 336, 752

Salama, A., Feuchtgruber, H., Heras, A., et al., 1997, First ISO Workshop on Analytical
Spectroscopy, ESA SP-419, 17

Serabyn, E., Güsten, R., Walmsley, C.M., Wink, J.E., Zylka, R., 1986, A&A 169, 85

Serabyn, E., Güsten, R., 1986, A&A 161, 334

Stacey, G.J., Lugten, J.B., Genzel, R. 1987, ApJ, 313, 859

Sturm E., Bauer O.H., Lutz D., et al. 1998, ASP Conf. Ser. 145: Astronomical Data
Analysis Software and Systems VII, 7, 161
 – 13 –

Sutton, E.C., Danchi, W.C., Jaminet, P.A., Masson, C.R., 1990, ApJ 348, 503

Triggs, N.E., Zahedi, M., Nibler, J.W., DeBarber, P., Valentini, J.J. 1992, J. Chem. Phys.
96, 1822

Valentijn, E.A., Feuchtgruber, H., Kester, D.J.M., et al., 1996, A&A 315, L60

van Dishoeck, E.F., Black, J.H. 1986, ApJS, 62, 109

van Dishoeck, E.F. 1998, in The Molecular Astrophysics of Stars and Galaxies, eds. T.W.
Hartquist and D.A. Williams (OUP), p. 53

White, G.J., Smith, H.A., Stacey, G.J., Fischer, J., Spinoglio, L., Baluteau, J.P., Cernicharo,
J., Bradford, C.M., 1999, Proceedings of the Conference “The Universe as seen by
ISO”, ESA SP-427, 787

Wilson, T.L., Rood, R. 1994, ARA&A, 32, 191

Wormhoudt, J., McCurdy, K.E., 1989, Chem. Phys. Letters 156, 47

Yamada, C., Hirota, E., Kawaguchi, K., 1981, J. Chem. Phys. 75, 5256

Yamada, C., Hirota, E., 1983, J. Chem. Phys. 78, 669

This manuscript was prepared with the AAS LATEX macros v4.0.
 – 14 –

Fig. 1.— (a),(b): Data and synthetic spectra of the CH3 observations; (c): Determination
of the CH3 excitation temperature Tex from the integrated absorption ratio of the 16.0 µm
R(0) line and 16.5 µm Q-branch.

Fig. 2.— ISO spectra of Sgr A∗ : (a) LWS spectrum of the CH pure rotational line doublet;
(b) SWS spectrum around the C2 H2 ν5 Q-branch and the synthetic spectrum.
 – 15 –

Table 1: Summary of results

Molecule Col. Dens. Tex Obs. Modea

[cm−2 ] [K]

CH (1.1±0.1)×1015 17b LWS01

CH3 (8.0±2.4)×1014 (17±2) SWS06

CH4 ≤1×1015 17b SWS06

C2 H2 (5.5±0.8)×1014 17b SWS06

C18 O (2.0±0.5)×1016 8-13 SWS06

a
For details of the observing modes see de Graauw et al. (1996) and Clegg et al. (1996)
b
Assumed excitation temperature
 – 16 –

Table 2: Comparison of observed abundances with modelsa

Model CH CH3 CH4 C2 H2

Obs. 18 +22
−7
(-9) 13 +22
−7
(-9) <17(-9) 9 +21
−4
(-9)

LBH96b 0.26(-9) 0.20(-9) 180(-9) 15(-9)

TH00c 44(-9) 8.6(-9) 5.6(-9) 9.2(-9)

VDB00d 3.2(-9) 6.4(-9) 18(-9) 2.0(-9)

a
All abundances with respect to H2
b
Lee et al. (1996) new standard model with nH = 104 cm−3 and T =10 K at steady state
with low metal abundances
c
Terzieva & Herbst (priv. comm.) translucent cloud model with nH = 2 × 103 cm−3 , T =10
K and AV =3 mag at steady state with low metal abundances
d
Updated models of van Dishoeck & Black (1986) and Jansen et al. (1995) for nH = 2 × 103
cm−3 and AV =3 mag at steady state with low metal abundances