Detection of C5 in the Circumstellar Shell of IRC+10216

The C5 molecule has been identified in the infrared spectrum of the prototypical obscured carbon star, IRC+10216. In addition to their astrophysical importance, pure carbon chain molecules such as C5 are of interest in the chemistry of flames and propellants.

molecules, including chains, rings, and sheets, as well as spheroidal filllerenes (1,2) such as C60. These molecules are believed to play an important role in the formation of soot in flames (1,2). In astrophysics, they may be both the building blocks and photofragments of carbonaceous materials.
Recently we detected the C3 molecule in the high-resolution circumstellar absorption spectrum of a carbon star, IRC+ 10216 (3). The C3 absorption features were very sharp (<0.014 cm-' wide, full width at halfmaximum) and remarkably strong (almost 50% absorbing). Inspired by these results, we searched for the "next" member of the carbon chain series, Cs.
We skipped searching for C4 in favor of C5 for a variety of reasons: (i) C5 is predicted to be linear and symmetric with a closedshell X'Yg electronic ground state similar to the C3 molecule (4-8); (ii) the C4 molecule is known to be less abundant than C5 (9,10) in high-temperature carbon vapors (although this is not necessarily the case in a circumstellar shell); and (iii) the previous infrared spectrum of C4 (11) recorded near 2164 cm in inert gas matrices was recently reassigned to C5 (12).
The C5 molecule has no dipole moment and thus cannot be monitored by pure rotational microwave spectroscopy. The optical spectrum of Cs is not known with any degree of confidence, although some published experimental (13) and theoretical (7) suggestions are available.
The infrared vibration-rotation transitions are ideal for the detection ofC5. Ofthe seven vibrational modes of Cs, V3 (a,+, symmetry) near 2164 cm-' (12) is predicted to have the largest transition dipole moment (0.48 D) (8). The V3 mode of C5 is an "antisymmetric stretching mode" similar to V3 (Ou+) of C3 near 2040 cm-' (3). The V3 mode of C3 is also predicted to have a remarkably strong transition dipole moment of 0.44 D (14).
Our search for the high-resolution vibration-rotation spectrum of C5 was successful. Guided by our preliminary results, C5 has been detected in the laboratory by two reseach groups. Our astronomical measure-ments are in excellent agreement (15,16) with the subsequent laboratory measurements.
On 17 January 1989, spectra ofthe prototypical obscured carbon star IRC+ 10216 and the moon were observed with an unapodized resolution of 0.010 cm-' (1.4 km s-1, full width at half-maximum) by means ofthe Kitt Peak National Observatory 4-meter telescope and the Fourier transform spectrometer located at the coude focus (17). The 218-min integration on IRC+ 10216 resulted in a peak signal-to-noise (1 SD) ratio of 360. The spectrum was limited to 2115 to 2195 cmn' by a cold blocking filter. The moon, which is an intrinsically featureless source in the 2100 to 2200 cm-' region, was observed at the same air mass as IRC+ 10216. We used the ratio ofthe lunar spectrum, which had a peak signal-to-noise ratio of 460, to the IRC+ 10216 spectrum to remove the telluric spectrum.
The observed 2170-cm' line positions of Cs needed to be corrected for the Doppler shift caused by the motion ofthe circumstellar shell relative to the Fourier transform spectrometer. To do this we assumed, in analogy to other molecules observed in the infrared with similar line shape and excitation temperature (see below), that the lines exist in the circumstellar shell at the terminal expansion velocity, -14 km s-1, relative to the center of mass (18). We used this velocity and corrected for a local standard-of-rest velocity of -26 km s-' for IRC+10216 (19) and for the earth's velocity at the time of observation to measure the frequencies.
Telluric ozone lines (20) were used to obtain the frequency zero point.
In the spectrum (Fig. 1), a series of six   (4)(5)(6)(7)(8). This series oflines turned out to be a P branch (AJ = -1, J the rotational angular momentum quantum number). The R branch lines (AJ = +1) were then assigned by prediction from the P branch with the use of the usual rotational energy level expression F(J) = BJ(J + 1) (1) where B is the rotational constant.
At this point in our analysis, the attribution ofthe spectrum to C5 was secure, as was the relative rotational numbering between the P and R branch lines. However, the absolute rotational assignment and the location of the band origin was not certain. Fortunately, by this time Heath et al. (15) had detected the first lines, R(0) and P(2), in their cold laboratory spectrum ofC5. This work provided us with the correct absolute rotational quantum number assignments for our lines. The lowJ lines near the origin in our spectrum are very weak and, in the case of R(0) and P(2), are obscured by atmospheric absorption.
The 16 observed line positions of C5 (Table 1) were reduced to three molecular constants ( Table 2) with Eq. 1. The observed band origin of 2169.442 cm-' agrees with the argon matrix value of 2164 cm-' (11,12) and the neon matrix value of 2167 cm-l (11).
The rotational constants of Table 2 agree well with the ab initio predictions, for example Be = 0.0866 cm-' (6), although we measure vibrationally averaged constants, whereas the quantum chemical calculations usually provide equilibrium Be values. These calculations (4-8) also predict that C5 has a cumulene type ofstructure with all ofthe C-C bonds almost identical in length. If the bonds are identical in length, then from our ground-state rotational constant (Table 2) we predict an ro value for the bond lengths of 1.283 A. The corresponding ro for C3 iS

A (3).
We assumed an infrared transition dipole moment for C5 of 0.48 D (8) and used an observer's frame radiation transport code (18) to synthesize the C5 lines. The spectrum synthesis indicates that the observed rotational strength distribution is consistent with a rotational temperature of approximately 40 K. There is no a priori reason to expect a Boltzmann population distribution, and a range of temperatures, 20 to 70 K, was required for C3 (3). However, the Cs data are not of sufficiently high signal-tonoise ratio to determine if a range of temperatures is also required for Cs.
The derived column density of Cs is n(Cs) = 0.9 + 0.25 x 1014 cm-2, implying a fractional abundance relative to molecular  Fig. 1. A short section of the spectrum of the moon, IRC+ 10216, and the ratio IRC+ 10216/moon showing most of the observed C5 P branch. Also identified are CO lines of various isotopic species. The CO lines are formed throughout the circumstellar shell and have a different line shape and excitation temperature from the C5. The frequency scale has not been corrected for the earth's velocity. See text to convert the C5 frequencies to laboratory values. Note that the intensity scale for the ratioed spectrum has been expanded by a factor of 4.
Thus Cs is a factor of 11 less abundant than C3. The Cs temperature (40 K) and abundance are similar to those of the C6H [n= 3 x 10'4 cm-2 and T= 30 K, (21,22)] and C5H [n = 1.5 x 1014 cm-2 and T = 20 K (22,23)]. The detection limits now available in infrared astronomy compare favorably with those achieved in radio astronomy.
The production mechanisms for Cs in IRC+ 10216 are unclear. Like C3, the rotational lines of Cs are very sharp (<0.010 cm-1, full width at half-maximum) and cold only in the cold, outer part ofthe circumstellar shell, unlike CO, C2H2, HCN, and CS (18,24), which occur abundantly throughout the stellar envelope. Our discussion in the C3 paper (3) also applies to C5 and, indeed, it is possible that C3 and C5 originate from the same mechanisms.
It is very likely that C5 is a direct or iridirect result of photochemistry in the outer part of the envelope. The ultraviolet flux present in the interstellar medium can only penetrate into the outer portions of the dusty shell of IRC+10216. The C5 could result directly from the photolysis of larger molecules, such as C5H and C6H, or of carbonaceous grains. Alternately, C5 could be produced through ion-molecule chemistry, although no ions have been detected yet in IRC+ 10216. The necessary parent ions such as C2H2h (25,26) would also be produced by the interstellar ultraviolet flux, supplemented by cosmic ray ionization.
A new spectroscopic experiment has been developed in which rovibrational transitions of supersonically cooled carbon clusters, which were produced by laser vaporization of graphite, have been measured by direct-absorption diode-laser spectroscopy. Thirty-six sequential rovibrational lines of the V3 band of the C5 carbon cluster have been measured with Doppler-limited resolution. The absorption spectrum is characteristic of a linear molecule with a center of symmetry. Least-squares analysis of the spectrum indicates an effective carbon-carbon bond length of 1.283 angstroms, in good agreement with ab initio quantum chemical calculations. This work confirms the detection of C5 in IRC+ 10216 reported in the accompanying paper.
T HE INVESTIGATION OF CLUSTERS OF refractory elements has been one of the fastest growing fields in chemistry and physics during the past 6 to 7 years. Carbon clusters in particular have received much attention because of their relevance to interstellar chemistry and combustion chemistry (1)(2)(3)(4). The first theoretical calculations on carbon clusters date to the pioneering work of Pitzer and Clementi (5) and Hoffman (6). Those calculations postulated that carbon clusters of n atoms were either linear (for n c 10) or formed monocyclic rings (for n > 10). Odd-numbered linear clusters were predicted to have 'Y ground states and the even-numbered clusters were to have 3 ground states. Several recent theoretical papers have not qualitatively changed these results, but instead have given more detailed structural and spectroscopic predictions (7)(8)(9). One major new result from high-level quantum chemistry is the calculated existence of low-lying cyclic isomers of small even-numbered clusters (n = 4, 6, or 8) (10)(11)(12).
Complementary high-resolution spectroscopy experiments that can rigorously test these theories have not been possible. Prior to this work, the largest carbon cluster to be studied by high-resolution spectroscopy is the relatively stable cluster C3. Spectroscopy of C3 dates to cometary detection in 1881 Departmnent of Chemistry, University of California, = eIey3 CA 94720. (13), laboratory detection by Herzberg in 1942 (14), and the spectral assignment by Douglas in 1951 (15). Recently, the 2040cm-1 V3 band of C3 has been observed in space by Hinlde et al. (16) and in the laboratory by Matsumura et al. (17)