5 Results

5.1 K I lines

The spectrum recorded towards Star 6 exhibits very strong K  I absorption between -5 and -35 km s^-1 that is well removed from lines arising in the photosphere itself at +52.4 km s^-1 (see Fig. 2). The three vertical lines in the figure mark absorption components at -27.4, -20.4 and -9.5 km s^-1 that are repeated in both of the K  I profiles, and are centred on the heliocentric velocity of IRC +10$\degr$ 216 (-19.3 km s^-1). Moreover, the FWHM of the overall K  I profile, $\sim$28 km s^-1, is close to twice the known terminal expansion velocity of the gas in IRC +10$\degr$ 216 CSE, and consistent with the known geometry of the line of sight through the CSE.

The equivalent widths of the 7665 and 7699 Å K  I components are $\sim$630 and 525 mÅ respectively. Because of the low interstellar reddening in this region of the sky, $E_{ B-V}\sim 0.03$, we can therefore exclude the possibility that these very strong K  I lines are due to absorption by foreground or background diffuse interstellar matter. More precisely, for this low interstellar reddening, one would expect K  I equivalent widths of the order of only 8 mÅ, within a factor of 2, based the survey of interstellar K  I by Chaffee & White (1982).

Consequently, both the strengths and the velocity components of these K  I lines provide clear evidence for an origin in the IRC +10$\degr$ 216 circumstellar envelope.

Figure 3 shows the results of the spectral synthesis in the region of the two K  I components, indicating that there is no significant contamination of the circumstellar signature by photospheric lines of any species. Furthermore, Fig. 4 shows the 7665 Å profile (lower panel) compared to a pure telluric spectrum (upper panel). Both spectra are uncorrected for the heliocentric velocity, i.e. they are both at the rest wavelength of telluric lines. The telluric spectrum is of a very metal-poor halo subdwarf F0 star, G64-12, obtained from the UVES archive, for which $\rm [Fe/H] = -3.03$ (Axer et al. 1994). It can be seen that there is no telluric contamination of the circumstellar K  I 7665 component.

Figure 3: K  I lines in the VLT spectrum of Star 6, shown corrected for the stellar radial velocity. The dashed lines are a spectral synthesis (for details see text). The CS K  I components are clearly unaffected by other photospheric lines.

5.2 Na I lines

Figure 5 shows the region of the Na D lines in the Star 6 spectrum recorded with VLT/UVES. As for K  I, very strong absorption components are seen at velocities between -5 and -35 km s^-1, shifted away from the strong photospheric lines. Since these absorption components are saturated, there is little information in the profile shape, and no comparison has been performed between the profiles as was done for K  I. The components are centred on the velocity of IRC +10$\degr$ 216, and have a width commensurate with the gas expansion velocity, and it is likely that a large part of the absorption arises in the CSE of IRC +10$\degr$ 216. However interstellar contamination cannot be ruled out: even at E_B-V = 0.03 some interstellar absorption may be expected. Although the spectral synthesis of the Na lines is poor, probably because NLTE effects are not taken into account, it shows that there is again no significant additional component arising from other species in the Star 6 photosphere.

Figure 6 shows spectra of Na  I D lines from WHT/UES for comparison with the VLT K  I and Na  I Star 6 data. The signal-to-noise ratios of these spectra are $\sim$220 and $\sim$50 for $\nu$ Leo and Star C respectively. For Star C, the line profiles are very reminiscent of Star 6. Lying 153$\arcsec$ from IRC +10$\degr$ 216, within the CO and dust envelope radii, it is likely that these spectra are also probing CS Na  I in the IRC +10$\degr$ 216 CSE. The heliocentric velocity is near 0 km s^-1 (as measured using H$\beta$), so the profiles may be contaminated by photospheric Na  I on the red wing. Unfortunately, the WHT spectra of Star C do not cover the K  I resonance lines.

By contrast, the Na D lines in $\nu$ Leo are very weak (note the ordinate scale). With a spectral type of B9IV, there is no possibility of photospheric Na  I in this object. The lines observed are interstellar in origin and are likely to be typical of the ISM in this direction, at least up to a distance of $\sim$60 pc. ($\nu$ Leo lies 2.5$\degr$ from IRC+10$\degr$ 216).

Figure 4: The 7665 Å  K  I region in the VLT spectrum of Star 6 (lower panel) compared to a pure telluric spectrum (top panel - see text). Both spectra are at telluric rest. The CS K  I component is unaffected by telluric contamination.

5.3 Diffuse band absorptions

In diffuse ISM spectra, the strongest reasonably narrow DIBs lie at 5780 Å, 5797 Å, 6284 Å and 6614 Å. Owing to an instrumental gap in the data, the first two regions were not observed. However, the 6284 Å  and 6614 Å regions have been closely examined. In Fig. 7, the region of the DIB at 6284 Å is shown, together with a Kurucz synthesis (upper panel) and the pure telluric spectrum of G64-12. A feature can be seen at 6283.80 Å, indicated by a vertical line at this wavelength. This wavelength is very close to the 6283.86 Å central wavelength of the 6284 DIB, derived from ISM studies (Herbig 1995). While this feature is unaffected by telluric lines, the situation is complicated by the Kurucz synthesis prediction of a line coinciding with the possible DIB. The line is Fe  I $\lambda$6283.729. While the synthesis clearly greatly overestimates the strength of this line, presumably owing to uncertainty in the log gf value, an atomic contribution to the DIB profile cannot be ruled out.

The DIB is observed independently in all 4 spectra of this wavelength region, hence an instrumental origin is ruled out. Its equivalent width is 27 mÅ. Using the 6283.86 Å value as a rest wavelength, a heliocentric velocity of -2.9 km s^-1 is derived, which is far from the heliocentric velocity of IRC +10$\degr$ 216. Using the well known correlation of DIB strength to E_B-V, it is possible to estimate the reddening from the DIB strength, assuming an origin in the ISM, which seems likely. For HD 183143 ( E_B-V = 1.28), $W_{\lambda}$for this DIB is 1945 mÅ (Herbig 1995) while for the star $\mu$ Sgr ( E_B-V = 0.25), a well characterised ISM line of sight, $W_{\lambda}$ is 167 mÅ (unpublished results of the author). Simple scaling yields E_B-V=0.018 and 0.04 respectively, values that are close to that suggested by the H  I maps and galaxy counts of Burnstein & Heiles (1982) for this region.

Figure 5: Na   I profiles in the Star 6 VLT spectrum, together with spectral syntheses (dashed lines). The left panel shows Na D1, the right Na D2. The spectra are in velocity space, and the model spectra have been shifted to the photospheric radial velocity.

Figure 6: Na   I profiles in the WHT spectra of $\nu$ Leo (left panel) and Star C (right panel). D1 is above, D2 below. Note the similarity of the Star C profiles to those in Star 6, and the weakness of the profiles in $\nu$ Leo. Note the ordinate scales.

This argument reinforces the suggestion that the $\lambda$6284 feature arises in the ISM, not the IRC +10$\degr$ 216 CSE. Moreover, its heliocentric velocity is near that of the bluer of the two components seen in Na D in the spectra of $\nu$ Leo (see Fig. 6), that are likely to be interstellar. It is probable that these Na D components and the $\lambda$6284 DIB observed towards Star 6 arise in the same diffuse cloud, which is far closer than Star 6 and even closer than IRC +10$\degr$ 216, the distance of $\nu$ Leo being only 55 pc.

The other strong DIB searched for is $\lambda$6614 (see Fig. 8). This wavelength region is free of photospheric and telluric lines. No DIB is observed. An upper limit to the equivalent width of this DIB can be obtained. We assume a width (FWHM) of about 0.35 Å, as is shown by the high resolution profiles through clouds with especially narrow Na  I lines given by Walker et al. (2000). Then, by applying the method of Cowie & Songaila (1986; their Eq. (3.12)), with R=50 000 and S/N=58, we find a 3$\sigma$ upper limit of 16 mÅ. By scaling the case of HD 183143 given by Herbig (1995), this $W({\rm 3\sigma})$ corresponds to $A_{\rm v}=0.18$ mag or $N_{\rm H}=3.8\times10^{20}$ cm^-2. Consequently, our 3$\sigma$ upper limit on the strength of a diffuse circumstellar band at $\lambda$6614 is about 5 times smaller than what would be expected from the circumstellar medium, if it obeyed the interstellar relation between $N_{\rm H}$ and the DIB strengths.

We have also examined the regions of the DIBs at 6196, 6203, 6270, 6379, 6993, and 7224 Å. Blends with photospheric lines occur for $\lambda$6196, $\lambda$6379 and $\lambda$7224, but the other three regions are clear, and no diffuse feature is seen with upper limits comparable but not better than the one derived above for $\lambda$6614. DIBs have also been sought in the WHT/UES spectrum of Star C. The strongest DIBs in the highest signal-to-noise ratio spectral regions are $\lambda$6284 ( $S/N \sim 35$) and $\lambda$5780 ( $S/N \sim 50$). No DIBs are observed.

Figure 7: The region of the 6283 DIB in the Star 6 spectrum. The possible DIB feature is indicated by the vertical line. All spectra are at the heliocentric velocity. The observational data (solid line) is compared to a Kurucz synthesis (top panel) and a pure telluric spectrum (bottom panel).

Figure 8: The region of the 6614 DIB in the Star 6 spectrum. The rest wavelength of the DIB feature is indicated by the vertical line. All spectra are at the heliocentric velocity. The observational data (solid line) is compared to a Kurucz synthesis (dashed line).