The two-dimensional freezing transition is very different from its three-dimensional counterpart. In contrast, the glass transition is usually assumed to have similar characteristics in two and three dimensions. Using computer simulations we show tha
t glassy dynamics in supercooled two- and three-dimensional fluids are fundamentally different. Specifically, transient localization of particles upon approaching the glass transition is absent in two dimensions, whereas it is very pronounced in three dimensions. Moreover, the temperature dependence of the relaxation time of orientational correlations is decoupled from that of the translational relaxation time in two dimensions but not in three dimensions. Lastly, the relationships between the characteristic size of dynamically heterogeneous regions and the relaxation time are very different in two and three dimensions. These results strongly suggest that the glass transition in two dimensions is different than in three dimensions.
Recently, ultrastable glasses have been created through vapor deposition. Subsequently, computer simulation algorithms have been proposed that mimic the vapor deposition process and result in simulated glasses with increased stability. In addition, r
andom pinning has been used to generate very stable glassy configurations without the need for lengthy annealing or special algorithms inspired by vapor deposition. Kinetic and mechanical stability of experimental ultrastable glasses is compared to those of experimental glasses formed by cooling. We provide the basis for a similar comparison for simulated stable glasses: we analyze the kinetic and mechanical stability of simulated glasses formed by cooling at a constant rate by examining the transformation time to a liquid upon rapid re-heating, the inherent structure energies, and the shear modulus. The kinetic and structural stability increases slowly with decreasing cooling rate. The methods outlined here can be used to assess kinetic and mechanical stability of simulated glasses generated by using specialized algorithms.
We combine computer simulations and analytical theory to investigate the glassy dynamics in dense assemblies of athermal particles evolving under the sole influence of self-propulsion. The simulations reveal that when the persistence time of the self
-propelled particles is increased, the local structure becomes more pronounced whereas the long-time dynamics first accelerates and then slows down. These seemingly contradictory evolutions are explained by constructing a nonequilibrium mode-coupling-like theory for interacting self-propelled particles. To predict the collective dynamics the theory needs the steady state structure factor and the steady state correlations of the local velocities. It yields nontrivial predictions for the glassy dynamics of self-propelled particles in qualitative agreement with the simulations.
We use a long, all-atom molecular dynamics (MD) simulation combined with theoretical modeling to investigate the dynamics of selected lipid atoms and lipid molecules in a hydrated diyristoyl-phosphatidylcholine (DMPC) lipid bilayer. From the analysis
of a 0.1 $mu$s MD trajectory we find that the time evolution of the mean square displacement, [delta{r}(t)]^2, of lipid atoms and molecules exhibits three well separated dynamical regions: (i) ballistic, with [delta{r}(t)]^2 ~ t^2 for t < 10 fs; (ii) subdiffusive, with [delta{r}(t)]^2 ~ t^{beta} with beta<1, for 10 ps < t < 10 ns; and (iii) Fickian diffusion, with [delta{r}(t)]^2 ~ t for t > 30 ns. We propose a memory function approach for calculating [delta{r}(t)]^2 over the entire time range extending from the ballistic to the Fickian diffusion regimes. The results are in very good agreement with the ones from the MD simulations. We also examine the implications of the presence of the subdiffusive dynamics of lipids on the self-intermediate scattering function and the incoherent dynamics structure factor measured in neutron scattering experiments.
