Astron. Astrophys. 339, 194-200 (1998)

4. Discussion

The two spectra presented in Fig. 1 are believed to represent a typical case of mid-IR emission bands from carbon particles, and will be be taken as representative for interstellar environments with moderate radiation energy density (up to times the solar value).

4.1. Physical significance

The fact that the UIR bands are well described as wide Lorentz profiles can be related to the physics of the emission of large molecules at high temperature. We propose that the observed width, in the range of 20 to 100 cm^-1 (a fraction of a micron) is largely due to intramolecular broadening processes. When a high amount of energy is present in a large molecule, the exchange of excitation between the accessible modes is very fast (Barker et al. 1987). This process, known as Internal Vibrational Redistribution (IVR) occurs with characteristic time scales that can be shorter than 10^-13 s. The excited levels of a given vibrational mode are continuously populated and emptied by non-radiative transfer to the other modes. The width of the transition reflects the short time between such events. Note that this time is incommensurate with the radiative cooling time generally quoted for PAHs, which is of the order of seconds or more (Léger & Puget 1984).

4.2. Laboratory support

The broadening of energy levels by IVR has been demonstrated by laboratory observations. Ionov et al. (1988, see also references therein) have shown that the 935 cm-1 (10.7 µm) transition of the (CF₃)₃Cl molecule excited at 36500 and 42500 cm^-1 (4.3 and 5.3 eV) exhibits Lorentzian shapes corresponding to life-times of and . This experiment also shows that the central frequency and width of the energy transition varies with the internal temperature. Closer to the astrophysical observations, Joblin et al. (1995) have found that the bands in spectra of PAHs, such as coronene at temperatures of about 700 K, have widths in the range 10-30 cm^-1. These widths, much larger than those measured for PAHs in cryogenic matrices, increase with temperature in agreement with what is described by Ionov et al. (1988). For the 3.3 line, the extrapolation of the temperature dependence allows to connect the data of Joblin et al. to that of Shan et al. (1991), and Brenner & Barker (1992) on smaller PAHs heated by laser pulses. For coronene at a temperature of K, the 6.2 and 8.6 features widths are only a factor 2 to 3 narrower than the interstellar bands (Joblin et al. 1995). The 7.7 and features are obviously more complicated; the interstellar band appears as a combination of at least two modes (Sect. 3.2).

Since the lines of individual PAHs studied in the laboratory are spread about the interstellar bands, the interstellar spectra may result from the combination of emission from numerous molecules. However, it is most unlikely that the Lorentz shape could result from a random combination of narrower lines. We rather think that to a large extent the width of the emission bands is intrinsic to the emission of each particle and is larger than that measured, in the laboratory, for PAHs with less than a few tens of atoms.

Models built to account for the spectral energy distribution of dust emission in the near and mid-IR show that only the emission in the comes from molecules with less than atoms. For a continuous size distribution between these molecules and dust grains, the emission in the and longer wavelength features comes from particles with up to several hundred atoms, one order of magnitude larger than the PAHs studied in the laboratory (Désert et al. 1990, Schutte et al. 1993). We are not aware of any theoretical or experimental work on such large particles. In relation with our interpretation of the width of the bands, we can only speculate that for large particles, the larger number of accessible modes makes the internal exchange faster, the life-times shorter and the bands wider. Our interpretation of the spectra also requires that for such large particles the vibrational modes do not depend much on the exact shape of the particle. The positions of the emission lines of PAHs around 6.2 and depend strongly on the type of the molecule (e.g. Léger & d'Hendecourt 1987). This is due to the fact that these bands correspond to in-plane C-C vibrational modes of the hexagonal lattice, which are strongly affected by boundary effects. In small molecules, asymmetries and irregularities of the shape induce the various line shifts and splittings between 6 and 8 . In large particles, boundary effects become relatively less important and indeed, measurements by Friedel & Carlson (1971) show that a broad infrared emission feature exists around 7.7 in finely ground graphite. Raman spectroscopy of graphite by Tuinstra & Koenig (1970) reveals the same effect. Interpolating between these two extremes situations, small molecules and graphite particles, it is tempting to speculate that PAHs of a size intermediate between molecules and VSGs (Désert et al. 1990) may reconcile the observations. We note that such large particles, such as polymeres of coronene up to , and graphenes up to about 250 C-atoms, have been recently produced and detected in the laboratory (Joblin et al. 1997). The combination of the 7.6 to lines would be a remote consequence of relatively small boundary effects in those large particles.

© European Southern Observatory (ESO) 1998

Online publication: September 30, 1998
helpdesk.link@springer.de