Nucleation is an omnipresent phenomenon. Nevertheless, despite many experimental studies, it is not very well understood due to its inherently momentary character and because it is usually quite hard to separate homogeneous from heterogeneous processes. Heterogeneous processes (nucleation at interfaces) usually prevail because of the lower energy barrier (even in 3D systems because of always present contaminations and the sample walls). It is for instance not yet clear how nano-scale interfacial properties (local topological and/or chemical "defects") affect nucleation and growth because of subtle confinement effects (1) and due to the specific interfacial phase behavior (2). In general, interfacial systems (e.g. thin films on surfaces) are much better suited for experimental nucleation studies because they provide localized and immobilized nucleation processes. Thus the nucleus/cluster growth can be imaged (AFM, see Figure) and nucleation-active and nucleation-inactive surface sites can be distinguished via cluster size and the recurrent, localized cluster appearance in the case of "activated" nucleation (= the 2D analogon to heterogeneous nucleation in 3D).

We present results on (interfacial) nucleation studies with molecularly thin films of long chain alkanes and fullerens at SiO2/air interfaces. We can show for instance, that a meaningful interpretation of the nucleation rate has to take into account an "effective" undercooling, which can be largely different from the bulk values due to interfacial (wetting) effects (1).

(1) D. Frenkel, Nature 443, 641 (2006)

(2) H. Riegler and R. Koehler, Nature Physics 754 (2007)

Clusters of C60 at SiO2/air surfaces imaged by AFM