1 The breakdown of the shear modulus at the glass transition

The glass transition is described in terms of thermally activated local structural rearrangements, the secondary relaxations of the glass phase. The interaction between these secondary relaxations leads to a much faster and much more dramatic breakdown of the shear modulus than without interaction, thus creating the impression of a separate primary process which in reality does not exist. The model gives a new view on the fragility and the stretching, two puzzling features of the glass transition. The mode coupling theory of the glass transition (Götze and Sjögren 1992) postulates a crossover in the flow mechanism at the critical temperature T c to thermally activated hopping on the low-temperature side. This postulate has received strong support by recent simulations, which showed that the system passes from the saddle points of the energy landscape to the minima at exactly this critical temperature (for a review see Buchenau 2003). Thus one indeed expects thermally activated flow between T c and the calorimetric glass transition temperature T g , where the system freezes. However, if one looks at the temperature dependence of the flow process below T c , it is dramatically faster than that of a thermally activated process, particularly for the so-called fragile glass formers (Böhmer et al 1993). This fragility seems to be linked to a strong stretching of the shear stress relaxation, extending over several decades in time. Both phenomena have not yet found a generally accepted explanation. The present paper proposes to describe the glass transition in terms of the energy taking the interaction between different thermally activated jumps into account. There is increasing evidence (Richert 2002) for a heterogeneous energy landscape dynamics.