Abstract: Liquid Crystals (LC) are anisotropic fluids characterized by long range orientational order and pair correlations. Mesoscale models, based on the drastic simplification of representing molecules as simple rigid objects such as (LC) theories and computer simulations. While these approaches are still very valuable in obtaining the general properties of complex LC one of the most important current challenges is to relate a realistic molecular structure to spherocylinders or ellipsoids or even spins on a lattice have been the cornerstone of the first generation of liquid crystal physical observables and predict properties such as morphologies, order parameters, and phase-transition temperatures. 
Atomistic molecular dynamics (MD) simulations, consisting in the numerical solution of Newton equations of motion for all the atoms in the system now start to make this possible, also allowing the test of classical theories for bulk LC (e.g. Maier-Saupe or Onsager). However, for most practical applications LC are not used in bulk but in thin films where the LC is aligned with the help of surface interactions, so it is somewhat surprising that surface effects are still described only empirically, while little is known on their molecular origin. In the talk we shall show that computer simulations start to shed some light on the interfacial behavior of liquid crystals and show examples for the prediction of the alignment and anchoring of  LC at the interface with different solid surfaces e.g. silicon or crystalline and glassy silica with different roughness (see figure). Simulations show in various cases that molecular organizations at the interface differ radically from those in the bulk, showing either discontinuities or broad distributions of orientations rather than the simple Dirichlet type boundary conditions assumed by many continuum type theories. In the talk an introduction to these systems and a discussion of some open problems will be presented.