Quantitative understanding of transport in liquids at high densities is a fundamental problem of Physical Chemistry, with applications to glassy systems, liquidsolid transitions, conformational transitions in collapsed proteins, and many more. The primary challenge is the slow, cooperative nature of molecular reconfigurations which is currently beyond the dynamic range accessible by direct molecular modeling. We will discuss recent progress in two related aspects of this problem:
We compute, on the molecular basis, the viscosity and the intrinsic ionic conductivity of supercooled melts. Both quantities follow similar, strongly non- Arrhenius temperature dependences, but sometimes “decouple” by several orders of magnitude near the glass transition. This has lead many to conclude that mechanical and electrical relaxations in ionic melts are two distinct processes that “mix” at high enough temperatures. In contrast, We demonstrate that the decoupling even as large as four orders of magnitude does not imply that distinct processes are at play but stems from the intrinsic distribution of relaxation barriers in supercooled melts.
Another signature phenomenon of molecular transport in supercooled liquids, or in crystals near melting, is the apparent universal magnitude of elemental molecular translations, known as the Lindemann criterion. This empirical criterion has also been useful in computational studies of systems ranging from conformational dynamics of collapsed proteins to vortex lattices of rotating Bose condensates. We will argue that the near universality of the Lindemann displacement arises from the collisional, strongly over-damped dynamics in dense liquids, despite system specific an-harmonicities. This work also suggests novel structural excitations are present in pre-melted layers at crystal-melt interfaces.