Modelling of fracture processes is useful on many different scales and may reveal new fracture phenomena or help to understand long-known phenomena. Atomistic modeling and the accurate representation of bond breaking processes for example play a particularly critical role in brittle fracture, stress corrosion cracking or the evolution of surface roughness. I will give several examples of such phenomena ranging from hydrogen induced fracture in silicon, via crystallographic aspects of cleavage to the multiplication of dislocations at crack tips in metals. The dissociative chemisorption of H2 molecules at the edges of hydrogen-induced platelet defects in Silicon crystals which destabilizes the highly strained silicon bonds. The resulting stress-corrosion fracture process yields atomically smooth cleavage of Si crystals along otherwise unstable (100) planes. Cleavage along the known cleavage planes of Silicon is also interesting from an atomistic point of view, where cleavage anisotropy and the generation of roughness pose interesting questions. The fracture toughness of brittle metals and the brittle to ductile transition depend critically on the state of predeformation and how it changes plastic response near the crack tip. Atomistic simulations demonstrate that individual pre-existing dislocations may lead to the generation of large numbers of dislocations at the crack tips. Analysis of the forces acting on the dislocations allows to determine which dislocations multiply and the slip systems they activate. At last I will scale up by many orders of magnitude and demonstrate the natural occurrence of anticracks (negative mode I fracture) and their role in the generation of snow slab avalanches.