1 Introduction

Diffuse interstellar bands (DIBs) have been observed in the visible spectra of reddened OB stars for over seventy years; over 200 are now known, but the identification of the DIB carrier(s) remains a major challenge in molecular astrophysics (Herbig 1995). Many diverse origins have been advanced, such as carbon chains (Douglas 1977; Thaddeus et al. 1993), polycyclic aromatic hydrocarbons (PAHs) (Léger et al. 1987; Salama et al. 1996), fullerenes (Kroto & Jura 1992; Foing & Ehrenfreud 1997) or even H₂ (Sorokin & Glownia 1996; Ubachs et al. 1997), among many others. None has been confirmed with certainty. In addition to the intrinsic importance of identification of this ubiquitous material, DIB carriers may provide crucial information on the nature of interstellar matter, including dust composition and extinction properties, gas-phase and dust based chemistry and physics and, conceivably, astrobiology (Herbig 1995; Williams 1996; Miles & Sarre 1993).

A key issue is to discover where DIB carriers are formed, and there are several reasons to consider the dense outflows of asymptotic giant branch (AGB) carbon stars as a likely source. First, these stars inject into the ISM a large amount of carbon-rich material. Carbon chains (polyynes, C₃, C₅), to which the DIB carriers may be related, are known to be present in the circumstellar envelope (CSE) of IRC $+10\degr$ 216, the nearest dusty, molecule-rich carbon star with a high mass-loss rate. PAH molecules or fullerenes may also be present. A further strong argument is the fact that DIBs have been detected (in emission) in the circumstellar matter of two carbon-rich objects; The Red Rectangle (HD 44179, Sarre et al. 1995), which is believed to be in a post-AGB evolutionary phase, and the H-deficient R CrB star V854 Cen at minimum light (Rao & Lambert 1993).

The first search for diffuse bands in absorption in a circumstellar envelope was attempted by Snow & Wallerstein (1972) and Snow (1973), and no such signature was found. A more detailed search for diffuse band absorption in carbon-rich CSE was performed by Le Bertre (1990), and extended by Le Bertre & Lequeux (1992,1993). In the majority of cases, these authors did not find evidence of diffuse circumstellar bands, and concluded that the band carriers should be, in general, depleted in carbon-rich circumstellar matter (with the exception of a very few objects such as CS 776, which we reconsider in Sect. 6). Nor did they find more intense diffuse bands in sources with strong unidentified infrared bands (UIBs), such as NGC 7027. However, they proposed that nitrogen may be an important constituent of the carriers, or play a role in their formation. Spectra recorded towards the central star of IRC + $10\degr$ 216 revealed the presence of circumstellar C₂, but no diffuse band absorptions were reported (Bakker et al. 1997).

More recently, effort has been concentrated on carbon-rich proto-planetary nebulae with F- or G-type spectra (Zacs et al. 1999; Klochkova et al. 1999), demonstrating that diffuse band absorptions can be observed toward targets of later spectral type with more crowded stellar spectra, than the reddened OB stars typically used for ISM studies of this nature. Zacs et al. (1999) have observed a feature on the red wing of the $\lambda$5780 DIB, seen towards the G-type post-AGB object HD 178921, which they claim may arise in the CSE. They also propose possible circumstellar diffuse band absorptions at $\lambda\lambda$5780, 6597 and 6196 towards another G-type evolved object, SAO 34504 (IRAS 22272+5435). Finally, Klochkova et al. (2000b) (see also Klochkova et al. 1999; Klochkova et al. 2000a) proposed a circumstellar origin for diffuse band features in the spectra of IRAS 23304+6147, IRAS 04296+3429, IRAS 22223+4327 and the Egg Nebula, via velocity arguments.

In this study, we report the first results of a search for DIBs directly in the CSE of IRC $+10\degr~$216, by observing background stars seen through the envelope. Section 2 describes the targets that were selected and their characteristics. The observations, which were made with the ESO VLT/UVES and the La Palma WHT/UES spectrographs, are described in Sect. 3. A model atmosphere analysis of our principal target has been performed, and the results are given in Sect. 4 together with details in the Appendix. Our main results on circumstellar K  I, Na  I, and our search for DIBs are presented in Sect. 5, and their implications are discussed in Sect. 6. We conclude in Sect. 7 with a summary of our findings.