Abstract The atomistic understanding and control of information transfer across interfaces is of primary importance for any future molecular or bioelectronic device. In a systematic approach to reach this long term goal we investigate “model systems”, i.e. thin film sandwich structures with geometrically and chemically controlled compositions, and we relate their transport properties to their spectroscopically and microscopically determined structures. Silicon is by far the most important substrate for such structures because its processing is well known from current micro- and nanotechnologies. Simple “model systems” therefore consist of organic overlayers on the substrate Si(100), but also on other substrates such as SiO 2 or Au(111). More complex “model systems” include receptor-ligand structures, transport and channel proteins, enzymes, biomolecular function units, whole cells, and tissues. Finally we discuss briefly the two alternative approaches in the development of future biomolecular devices, i.e. the use of natural biological structures or of completely synthesised biomimetic structures. We chose the “electronic nose” as a specific example to point out the importance of hybrid devices.